Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Protein tyrosine phosphatase PRL-3 expression and transcriptional regulation in cancer and metastasis Wong, Patrick Chun Wei 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_2005-0701.pdf [ 14.26MB ]
Metadata
JSON: 831-1.0092194.json
JSON-LD: 831-1.0092194-ld.json
RDF/XML (Pretty): 831-1.0092194-rdf.xml
RDF/JSON: 831-1.0092194-rdf.json
Turtle: 831-1.0092194-turtle.txt
N-Triples: 831-1.0092194-rdf-ntriples.txt
Original Record: 831-1.0092194-source.json
Full Text
831-1.0092194-fulltext.txt
Citation
831-1.0092194.ris

Full Text

P R O T E I N T Y R O S I N E P H O S P H A T A S E PRL-3 EXPRESSION A N D TRANSCRIPTIONAL R E G U L A T I O N IN C A N C E R A N D METASTASIS By  PATRICK C H U N WEI W O N G B . M . L . S c , The University o f British Columbia, 2003  A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S FOR THE D E G R E E OF  MASTER OF SCIENCE In  THE FACULTY OF GRADUATE STUDIES (Pathology and Laboratory Medicine) THE UNIVERSITY OF BRITISH COLUMBIA June 2005 © Patrick Chun Wei Wong, 2005  ABSTRACT Elevated expression o f protein tyrosine phosphatase P R L - 3 is implicated in cell migration, invasion, and metastasis, although the molecular targets o f this phosphatase remain unidentified. P R L - 3 expression is strikingly upregulated in metastatic colon carcinoma, compared with much lower expression i n primary colon tumours and virtually undetectable expression in normal colon epithelia. Increased transcriptional activity is suggested to account for increased P R L - 3 expression i n 75% o f the metastatic samples. W e used bioinformatics analyses and cell-based promoter assays to investigate the tumour- and metastasis-specific transcriptional activity o f the ~4.3 kb promoter region upstream o f the P R L - 3 transcription start site.  Through bioinformatics analyses, sequence conservation (human/mouse) and multiple putative transcription factor binding sites were identified in various regions upstream o f the P R L - 3 transcription start site. A 518bp C p G island (CpG-64) was predicted to have high regulatory potential. The transcriptional activities o f the P R L - 3 promoter were analyzed by luciferase reporter assays i n two cell lines derived from (i) a primary colon adenocarcinoma (SW480) and (ii) a metastasis o f the same tumour (SW620). These assays revealed that the full-length promoter region (which contains the CpG-64 island) was 4.5-fold more transcriptionally active i n the metastatic colorectal carcinoma cells than i n the primary colorectal carcinoma tumour cells. The C p G island itself directed 2.3-fold higher transcriptional activity in the metastatic colorectal carcinoma cells than in the primary colorectal tumour cells. In prostate cancer cells, this CpG-64 island was  8.7-fold more active i n the highly invasive androgen-independent D U I 4 5 cells compared to the less invasive androgen-dependent L N C a P cells. However, the full length promoter was 3.6-fold more active in the L N C a P cells than i n the D U I 4 5 cells. Additionally, this CpG-64 island was significantly more active in malignant muscle cells ( R D and R M S 13) than in normal muscle cells (C2C12). These results suggest that the CpG-64 island contains regulatory elements responsible for the tumour- and metastasis-specific transcriptional upregulation o f P R L - 3 . Identification o f these elements and their interacting factors may provide insight into the molecular pathways that are involved i n the tumour-to-metastatic transition.  U s i n g tissue-microarray technology, P R L - 3 expression was detected in colorectal, uveal melanic, breast, and certain pediatric tumours. Immunohistochemical localization o f P R L - 3 appeared to be associated with tissue differentiation status. M o r e specifically, nuclear P R L - 3 was mainly present i n poorly differentiated tissues, whereas cytoplasmic P R L - 3 was predominantly found in well differentiated tissues. A l s o , high levels o f P R L 3 in clinical tumour samples correlated with an unfavorable trend in uveal melanic and breast cancer patient prognosis. In conclusion, the evidence from both the basic science and the clinical sides argue strongly for the importance o f P R L - 3 as a tumourigenic and/or metastatic biomarker. Thus, it is imperative to study this phosphatase at the molecular level to understand the clinical>'ramifications o f its activities.  iv  TABLE OF CONTENTS  Abstract  ii  Table o f Contents  iv  List o f Tables  vii  List o f Figures  viii  List o f Abbreviations Acknowledgements  x xi  CHAPTER 1: INTRODUCTION  1  1.1  The Protein Tyrosine Phosphatase Superfamily  1  1.2  The Protein i n Regenerating Liver ( P R L ) Family  4  1.3  P R L - 3 : A Metastasis-Associated Phosphatase  8  1.4  Hypotheses and A i m s  18  CHAPTER 2: MATERIALS AND METHODS  19  2.1  Bioinformatics  19  2.2  Molecular Cloning  22  2.2.1  Polymerase Chain Reaction ( P C R )  22  2.2.2  TA-vector subcloning  24  2.2.3  Luciferase expression vector subcloning  25  2.2.4  P R L - 3 promoter constructs  26  Transfection and Luciferase Assay  28  2.3.1  C e l l line propagation  28  2.3.2  Transfection  29  2.3  2.3.3 2.4  2.5  Luciferase Assay  29  Western Blot  30  2.4.1  Antibodies  30  2.4.2  C e l l line treatments  30  2.4.3  SDS-PAGE  32  Immunohistochemistry  33  2.5.1  Polyclonal anti-PRL-3 antibody specificity  33  2.5.2  Immunohistochemistry staining  33  2.5.3  Tissue Microarray  34  C H A P T E R 3: R E S U L T S  35  3.1  Transcriptional Analysis o f the P R L - 3 Promoter  35  3.1.1  P R L - 3 promoter activity i n human embryonic kidney 293T cells  35  3.1.2  Auto-transcriptional regulation o f the P R L - 3 promoter i n human embryonic kidney 293T cells  3.1.3  37  P R L - 3 promoter activity in primary and metastatic colorectal cancer cell lines  40  3.1.4  P R L - 3 promoter activity in metastatic prostate cancer cell lines  43  3.1.5  P R L - 3 promoter activity i n normal muscle and rhabdomyosarcoma cell lines  45  3.2  Bioinformatics Analyses o f the P R L - 3 Promoter Sequence  48  3.3  Endogenous Snail and P R L - 3 expression  56  3.4  P R L - 3 expression in clinical tumour samples  58  3.4.1  58  Polyclonal anti-PRL-3 antibody specificity  vi  3.4.2  P R L - 3 expression in Colorectal Carcinoma  61  3.4.3  P R L - 3 expression in Uveal Melanoma  64  3.4.4  P R L - 3 expression in Breast Carcinoma  68  3.4.5  P R L - 3 expression i n Pediatric Tumours  72  CHAPTER 4: DISCUSSION 4.1  4.2  76  Importance o f the CpG-64 island i n the P R L - 3 promoter  76  4.1.1  CpG-64 island: The only or alternative promoter to P R L - 3 ?  77  4.1.2  CpG-64 island and cancer-associated c.s-elements  82  4.1.2A  Max  83  4.1.2B  N-myc  84  4.1.2C  Mef-2  85  4.1.2D  Snail  87  Poor prognosis and functional localization o f P R L - 3  92  4.2.1  Colorectal Carcinoma  92  4.2.2  Uveal Melanoma  93  4.2.3  Breast Carcinoma  95  4.2.4  Pediatric Tumours: Neuroblastoma and Rhabdomyosarcoma  98  CHAPTER 5: SUMMARY  100  5.1  Conclusions  100  5.2  Future Directions  102  REFERENCES  104  VII  LIST OF TABLES Table 1  Bioinformatics Programs  21  Table 2  P C R Primers  24  Table 3  Putative T F B S within the C p G - 6 4 island o f the P R L - 3 promoter  51  Table 4  Putative T F B S within ~1.5kb o f the T S S o f the P R L - 3 promoter  52  Table 5  Colorectal Carcinoma Frequency Table  63  Table 6  Uveal Melanoma Frequency Table  67  Table 7  Cytoplasmic P R L - 3 Breast Carcinoma Frequency Table  70  Table 8  Nuclear P R L - 3 Breast Carcinoma Frequency Table  71  Table 9  Pediatric Neuroblastoma Frequency Table  74  Table 10  Pediatric Rhabdomyosarcoma Frequency Table  75  Table 11  Pediatric Neuroblastoma 2 X 2 Contingency Table  75  Table 12  Pediatric Rhabdomyosarcoma 2 X 2 Contingency Table  75  viii  LIST OF FIGURES Figure 1  The P T P Superfamily  3  Figure 2  P R L - P T P family member identity  5  Figure 3  Ras Protein prenylation  7  Figure 4  P R L - 3 promoter constructs  27  Figure 5  P R L - 3 promoter activity i n H E K 2 9 3 T cells  36  Figure 6  P R L - 3 protein induction in HEK-FLA  38  Figure 7  P R L - 3 auto-transcriptional role i n H E K 2 9 3 T cells  39  Figure 8  P R L - 3 promoter activity in Colorectal Carcinoma cell lines  42  Figure 9  P R L - 3 promoter activity in Prostate Carcinoma cell lines  44  Figure 10  P R L - 3 promoter activity i n Rhabdomyosarcoma cell lines  47  Figure 11  T F B S and orthologue alignment o f the P R L - 3 promoter  50  Figure 12  Transcript generation, regulatory potential, and  G - P R L - 3 cells  cis-regulatory  modules analyses ofthe P R L - 3 promoter  55  Figure 13  Endogenous Snail and P R L - 3 expression  57  Figure 14  P R L - 3 antibody specificity analysis using immunoblot  59  Figure 15  P R L - 3 antibody specificity analysis using I H C  60  Figure 16  Immunohistochemistry using pre-immune rabbit serum  61  Figure 17  Colorectal Carcinoma T M A Immunohistochemistry  63  Figure 18  Uveal Melanoma T M A Immunohistochemistry  66  Figure 19  Uveal Melanoma T M A Kaplan-Meier Survival Curve  67  Figure 20  Breast Carcinoma T M A Immunohistochemistry  69  ix  Figure 21  Cytoplasmic P R L - 3 Breast Carcinoma T M A Kaplan-Meier Survival Curve  Figure 22  Figure 23  70  Nuclear P R L - 3 Breast Carcinoma T M A Kaplan-Meier Survival Curve  71  Pediatric Tumour T M A Immunohistochemistry  74  X  LIST OF ABBREVIATIONS  Ab:  Antibody  AR:  Androgen Receptor  ARMS:  Alveolar Rhabdomyosarcoma  CHO:  Chinese Hamster Ovary  CRM:  Cz's-Regulatory Modules  EGF:  Epidermal Growth Factor  EMT:  Epithelial-Mesenchymal Transition  ERMS:  Embryonal Rhabdomyosarcoma  EST:  Expression Sequence Tags  HEK:  Human Embryonic Kidney  IGF-1:  Insulin-like Growth Factor 1  fflC:  Immunohistochemistry  MEF2:  Myocyte Enhancer Factor 2  MMP:  Matrix Metalloproteinase  PBS:  Phosphate Buffered Saline  PCR:  Polymerase Chain Reaction  PRL:  Protein in Regenerating Liver  PTK:  Protein Tyrosine Kinase  PTP:  Protein Tyrosine Phosphatase  RMS:  Rhabdomyosarcoma  TFBS:  Transcription Factor Binding Sequence  TMA:  Tissue Microarray  TSS:  Transcriptional Start Site  UCSC:  University o f California, Santa Cruz  UTR:  Untranslated Region  xi  ACKNOWLEDGEMENTS  I would like to take this opportunity to first and foremost send m y sincere gratitude to my research supervisor, Dr. Catherine J. Pallen, for providing excellent mentorship to me these past years, being so thorough and concerned about all aspects o f my research, allowing me to explore various ideas that came to m y mind, supervising the preparation o f this thesis, and giving me plenty o f freedom i n the lab to be myself. Secondly, I want to thank all my unforgettable lab members, for always being available when I needed help, sharing laughs together, staying late in the lab, and always laughing at my jokes even when they were not even logical. I also want to take this opportunity to thank everyone at the Genetic Pathology Evaluation Centre ( G P E C ) , especially Dr. Blake Gilks and Maggie Cheang for being so generous with their time and with various tissue microarray sections. In addition, I want to thank Dr. Valerie White for allowing me to investigate the uveal melanoma tissue microarray. I also want to thank: Dr. Fergall Magee for taking the time to teach me about pediatric oncology, Dr. Maureen O'Sullivan for collaborating with us on the pediatric array project, and all the Pathology laboratory technicians at B C Children's Hospital for allowing me to use their space and being so helpful. In addition, I want to thank my supervisory committee: D r . D a v i d Walker (Chair), D r . Paul Rennie, and D r . Wyeth Wasserman for overseeing my research progress and giving me valuable and constructive ideas / suggestions when needed. A special thank-you goes out to Dr. Wyeth Wasserman and D a v i d Arenillas for teaching and providing me with the necessary tools to complete the bioinformatics aspect o f this project. Lastly, I am forever indebted to my mother for single-handedly raising me, giving me encouraging words when I felt down, but mostly, for providing a suitable home for me to pursue further education. Thanks mom.  1 CHAPTER 1  INTRODUCTION  1.1  THE PROTEIN TYROSINE PHOSPHATASE SUPERFAMILY  Protein tyrosine phosphatases (PTPs) belong to a large superfamily o f enzymes observed in eukaryotes, prokaryotes, and viruses. They have been classified into several subfamilies based on their substrate specificity and structural features (Figure 1). In regards to substrate specificity, PTPs fall into two groups. The tyrosine-specific PTPs have an absolute specificity for phosphotyrosine substrates, while the dual specificity P T P s can dephosphorylate both phosphotyrosine and phosphoserine / phosphothreonine residues. In terms o f their structures, PTPs fall into three major subfamilies: receptor-like PTPs, non-receptor intracellular PTPs, and the low molecular weight P T P s which share minimal sequence identity with other members o f the superfamily. A l l PTPs are characterized by a highly conserved active site signature motif C X 5 R , where the invariant cysteine residue is essential for catalysis. However, no P T P s share any sequence similarity with the protein serine/threonine phosphatases or the alkaline phosphatases.  Protein phosphorylation is a common mechanism that a cell utilizes to control biological processes. A t a given time, up to 30% o f the total proteins i n the cell are phosphorylated and most o f these phosphorylated proteins comprise o f either serine or threonine phosphorylation (Roach, 1991; Hunter, 1995). Tyrosine phosphorylation occurs to a  2  much lesser extent but it plays a key role in many cell processes such as cell proliferation and differentiation, cell cycle control, cytoskeletal organization, cell migration, signal transduction, transcriptional regulation, ion-channel activity, and the immune response (Hunter, 1995; Neel, 1997). Increased cellular protein tyrosine phosphorylation is often associated with transformation and oncogenesis (Blume-Jensen and Hunter, 2001). The central role played by a constitutively activated protein tyrosine kinase ( P T K ) in oncogenesis was first illustrated by the discovery o f the transforming v-src P T K (Hunter and Sefton, 1980). Immediately after, a large number o f viral and cellular oncogenes that encode P T K s were identified (Blume-Jensen and Hunter, 2001). The observation that many P T K s are proto-oncogenes and growth factor receptors suggests that tyrosine phosphorylation is controlled by P T K s to activate signal transduction pathways. After growth factor stimulation or oncogenic transformation, the phosphotyrosine content o f the cell can increase by as much as 100-fold to - 1 - 2 % o f the total protein phosphorylation in the cell (Levinson et al, 1980; Sefton et al, 1981.) In resting cells, protein tyrosine phosphatases (PTPs) maintain a basal level o f tyrosine phosphorylation on P T K s and also P T K substrates to ensure the proper functioning o f the cell (Hunter, 1995; Fischer et al, 1991). In addition to negative regulatory roles, PTPs also play key roles as positive effectors o f signal transduction pathways (Tonks and Neel, 1996; Sun and Tonks, 1994).  The tyrosine-specific PTPs include all the receptor-like PTPs and some o f the nonreceptor PTPs, as well as the low molecular weight PTPs. Both receptor-like P T P s and non-receptor P T P s share a conserved catalytic domain encompassing about 240 amino acids. A m o n g these PTPs, the active site sequence conservation can be extended from  the minimal C X R motif to H C X A G X X R [ S / T ] .  In addition, each P T P contains  5  heterogeneous non-catalytic sequences flanking the conserved catalytic domains. These sequences are important for subcellular localization and regulation / interaction with other proteins, such as the formation o f a P T P / P T K complex.  •  Catalytic domain  u  •  o•o  MAM CAH-like  •  PEST  FN-III Ig-like  SH2  n u  •  m  CH2A/B CAAX motif  Band 4.1 C X R motif 5  §1 I I Extracellular  J mi  P T P i ; PTPa/e CD45  LAR PTP5  PTPn  I  _  SHP1/2  PTP1B  VH-1 I—I  P T P - Cdc25 M E G MKP-1  LMW • I M\M P TP . PRL  •  PTEN  PTPPEST  Tyrosine Specific  Receptor-like P T P s  Non-receptor P T P s  Protein Tyrosine Phosphatase Superfamily : Active Site Motif  •  LMW- I PRL PTP  CXsR  Figure 1. The P T P Superfamily. Schematic diagram o f structures o f selected members of the P T P superfamily. The inset box shows various structural motifs that are found in the P T P s (FN-III: Fibronectin type-Ill; M A M : Meprin/A5/PTP|a domain; l g : Immunoglobulin; C A H - l i k e : Carbonic anhydrase-like; C X R : P T P signature motif; Band 4.1: ezrin-band 4.1-merlin-radixin protein; SH2: Src homology 2; P E S T : Pro-Glu-Ser-Thr motif; C H 2 A / B : sequence homology found in Cdc25 and M K P - 1 ; C A A X : prenylation motif). 5  4  1.2  T H E P R O T E I N I N R E G E N E R A T I N G L I V E R (PRL) F A M I L Y  The protein tyrosine phosphatase (PTP) superfamily is made up o f a large and diverse family o f enzymes, including a subclass o f three closely related enzymes called the P R L PTPs ( P R L - 1 , -2, and -3). P R L - 1 was first discovered as an early immediate gene expressed in mitogen-stimulated cells and i n regenerating liver, hence the term "Protein in Regenerating L i v e r " ( P R L ) was coined (Mohn et al, 1991). P R L - 1 is undetectable i n quiescent liver but its expression is elevated after 70% hepatectomy (Diamond et al, 1994). P R L proteins have been identified and cloned from rat, human and mouse (Diamond et al, 1994; Montagna et al, 1996; Cates, et al, 1996; Zhao et al, 1996; Zeng et al, 1998). P R L - 1 has little sequence identity with other known PTPs outside its P T P signature motif and is thought to represent a novel group o f PTPs.  Three P R L proteins have been reported to date and termed P R L - 1 (also known as P T P c a a x l ) , P R L - 2 (also known as P T P 4 A , O V - 1 and PTPcaax2) and P R L - 3 (also known as P T P 4 A 3 ) . A s shown i n Figure 2 A , mouse P R L s share high sequence identity with one another. A s demonstrated i n Figure 2 B , P R L - 2 is 87.4% and 75.4% identical to P R L - 1 and P R L - 3 respectively, while P R L - 3 is 75.7% identical to P R L - 1 (Zeng et al, 1998). The P R L - 1 amino acid sequence is identical among human, rat and mouse. The P R L - 2 amino acid sequence is identical between human and mouse with the substitution o f only two residues (Carter, 1998). The P R L - 3 amino acid sequence is more diverse between human and mouse with the substitution o f six residues (Matter et al, 2001). Interestingly, a P R L has been identified as a putative gene product i n Drosophila ( d P R L -  5  1) and C. elegans ( c P R L - 1 ) . d P R L - 1 and c P R L - 1 have over 5 0 % sequence identity with human P R L - 1 , -2, and -3, which implies that P R L proteins are evolutionarily conserved and that the different isoforms o f P R L proteins o f human, mouse, and rat may come from the same precursor.  A PRL-1 MARMNRPAPVEVTYKNMRFLITHNPTNATLNKFIEELKKYGVTTIVRVCEATYDTTLV PRL-2  M  NRPAPVEISYENMRFLITHNPTNATLNKFTEELKKYGVTTLVRVCDATYDKAPV  PRL-3  MARMNRPAPVEVSYRHMRFLITHNPSNATLSTFIEDLKKYGATTWRVCEVTYDKTPL  PRL-1  EKEGIHVLDWPFDDGAPPSNQIVDDWLSLVKIKFREEPGCCIAVHCVAGLGRAPVLVA  PRL-2  EKEGIHVLDWPFDDGAPPPNQIVDDWLNLLKTKFREEPGCCVAVHCVAGLGRAPVLVA  PRL-3  EKDGITVVDWPFDDGAPPPGKVVEDWLSLLKAKFYNDPGSCVLVHCVAGLGRAPVLVA  PRL-1  LALIEGGMKYE DAVQFIRQKRRGAFNSKQLLYLEKYRPKMRLRFKDSNGHRNNCCIQ  PRL-2  LALIECGMKYEDAVQFIRQKRRGAFNSKQLLYLEKYRPKMRLRFRDTNGH  CCVQ  PRL-3 LALIESGMKYEDAIQFIRQKRRGAINSKQLTYLEKYRPKQRLRFKDPHTHKTRCCVM  B |  PRL-2 H H  Active Site  1 CAAX box  H N 75^4% s  75.7%  PRL-1 Figure 2. PRL-PTP family member identity. A,  PRL-3 alignment o f mouse P R L - 1 , -2, and -3  amino acid sequences, identical residues are highlighted i n yellow, prenylation motifs are underlined; B, percentage identity between the three members o f the P R L family.  6  Expression o f P R L - 1 is associated with proliferation in liver but with differentiation in intestine, suggesting that PRL-1 may target different substrates in different tissues (Diamond et al, 1996). In addition, over-expression o f either PRL-1 or P R L - 2 alters cell growth and induces a transformed phenotype in transfected mouse fibroblasts and hamster pancreatic epithelial cells, as well as promoting tumour growth i n nude mice (Diamond et al, 1994; Cates et al, 1996). B y and large, the cellular substrates and functions o f the P R L - P T P s are still unknown.  M a n y prenylated proteins are important regulators o f signal transduction pathways. Protein prenylation is a post-translational modification involving the covalent attachment of a farnesyl (C15) or geranylgeranyl (C20) isoprenoid lipid group to a cysteine residue at or near the C-terminus. Prenylation can be followed by other modifications such as proteolysis, methylation and palmitoylation (Glomset et al, 1990; Zhang and Casey, 1996; Sinensky, 2000). One example o f a protein that is post-translationally prenylated is the Ras family GTPase (Casey et al, 1989; Schafer et al, 1989; Hancock et al, 1989) as seen in Figure 3. Protein prenylation is specified by the amino acid motif C A A X , C C X X , C X C or C C at the C-terminus and mediated by C A A X prenyltransferases and geranylgeranyltransferase II (Zhang and Casey, 1996; Sinensky, 2000). A m o n g P T P superfamily members, the P R L - P T P s are unique in possessing a C-terminal prenylation motif (Cates et al, 1996; Zeng et al, 1998), suggesting that the P R L proteins are potentially modified by prenylation. Human PRL-1 and P R L - 2 have been shown to be farnesylated by farnesyltransferase in vitro (Cates et al, 1996), and P R L - 2 has been shown to be farnesylated in vivo (Si et al, 2001).  7  FT CAAX  F  F  Protease  C  CAAX FPP  Methylase  t Ras  F C—OMe  Figure 3. Ras protein prenylation. Newly synthesized Ras is prenylated (FT: farnesyltransferase; F P P : farnesyl-pyrophosphate) and undergoes prenylation-dependent proteolysis and carboxylmethylation in the endoplasmic reticulum.  Protein prenylation can be important for targeting proteins to intracellular membranes and in protein-protein interactions (Zhang and Casey, 1996; Marshall, 1993; and Seabra, 1998). P R L - 1 was reported to be a nuclear protein tyrosine phosphatase (Diamond et al., 1994), which is atypical for a prenylated protein because most prenylated proteins are associated with the cytoplasmic side o f various intracellular membranes. P R L proteins are prenylated in vivo and their prenylation is required for their association with the plasma membrane and the early endosome (Zeng et al., 2000). Inhibition o f prenylation, by treatment o f cells with a farnesyltransferase inhibitor, results in nuclear localization o f newly synthesized P R L proteins (Zeng et al., 2000). The prenylation motifs o f the P R L s ( C C I Q , C C V Q , and C C V M ) conform to either the C A A X motif preferentially recognized by farensyltransferase or to the C C X X motif recognized by geranylgeranyltransferase II, although their relocalization in response to farensyltransferase inhibition indicates that they are likely farnesylated proteins in vivo. The prenylation-dependent subcellular localization o f the P R L s suggests that regulated prenylation could be a mechanism to control the access o f these PTPs to early endosomal or nuclear substrates.  8  1.3  PRL-3: A METASTASIS-ASSOCIATED P H O S P H A T A S E  The P R L - P T P s are implicated in cell proliferation and tumourigenesis, and more recently have also been linked to metastasis. Both P R L - 1 and P R L - 2 have been shown to possess oncogenic activity (Diamond et al, 1994; Cates et al, 1996). P R L - 1 is highly expressed in some tumour cell lines, particularly i n melanoma cells (Wang et al, 2002). Stable expression o f heterologous P R L - 1 , -2, or -3 increases cell growth rates (Diamond et al, 1994; Cates et al, 1996; Matter et al, 2001). Furthermore, over-expression o f P R L - 1 i n NTH 3T3 mouse fibroblasts, and P R L - 1 or -2 in D27 hamster pancreatic ductal epithelial cells caused cell transformation, and i n the latter case conferred the ability to form tumours in nude mice (Diamond et al, 1994; Cates et al, 1996).  P R L - 2 is among eight genes found to be differentially expressed i n benign prostatic fibroblast cells after stimulation with tumourigenic L N C a P cell-conditioned media (Wang et al, 2002). P R L - 2 expression is up-regulated 4-fold i n benign fibroblast cells and 9fold in tumour fibroblast cells upon incubation with the conditioned media for 1 day. A m o n g the prostate cancer cell lines examined, androgen-independent cell lines such as P C - 3 and D U 1 4 5 have higher P R L - 2 expression than the androgen-dependent L N C a P cell line and normal prostatic epithelial cells. Notably, the clinical transition from the androgen-dependent state to the androgen-independent state is crucial i n prostate cancer development (Stearns and McGarvey, 1992; Fossa, 1994; Newling, 1996). Once this transition occurs, the options for effective treatment are substantially reduced. The over-  9  expression o f P R L - 2 in two out o f the three samples further implicates P R L - 2 i n advanced prostate cancer.  Saha et al. (2001) found that P R L - 3 expression is upregulated i n metastatic colorectal carcinoma. Eighteen metastatic colorectal cancer samples were profiled for gene expression analysis. O f the 144 genes that were found to be upregulated, only P R L - 3 was consistently over-expressed (50-450 fold) in all 18 metastatic specimens. It was almost undetectable in normal colon tissue, and primary colorectal tumours had intermediate P R L - 3 expression levels. Later, Zeng et al. (2003) demonstrated that the over-expression o f P R L - 3 in Chinese Hamster Ovary ( C H O ) cells resulted i n enhanced motility compared to non-PRL-expressing control C H O cells i n a wound-healing assay. Similarly, the P R L 3 expressing C H O cells displayed increased invasiveness in a Matrigel assay. The expression o f P R L - 3 promoted cell metastasis in vivo as well. This study provided initial evidence for a causal role o f P R L - 3 in promoting cell motility, invasion, and metastasis, which are all essential characteristic properties o f the metastatic process.  P R L - 3 is highly expressed in metastatic melanoma B 1 6 - B L 6 cells but not i n the parental B16 cell line with low metastatic ability ( W u et al, 2004). B 1 6 cells transfected with P R L - 3 c D N A displayed morphological transformation from epithelial-like to fibroblastlike shaped cells, much like the highly metastatic B 1 6 - B L 6 melanoma cells. These P R L 3 over-expressing cells showed much higher migratory ability, which was reversed by the presence o f specific P R L - 3 anti-sense oligodeoxynucleotide. These observations provide convincing evidence that P R L - 3 plays a causal role in tumour metastasis.  10  P R L - 3 has also been implicated in gastric cancer. M i s k a d et al. (2004) found that 87.5% o f Reverse-Transcriptase P C R analyzed gastric cell lines expressed P R L - 3 . Ninety-four cases o f primary gastric and fifty-four cases o f matching nodal metastases were also evaluated for immunohistochemical (IHC) P R L - 3 expression. It was found that 81.5% o f the cases with nodal involvement expressed P R L - 3 , whereas 50% ofthe cases without nodal involvement expressed P R L - 3 . Moreover, 68% o f all the primary gastric tumour samples also expressed P R L - 3 . P R L - 3 expression was closely associated with lymphatic invasion, extent o f lymph node metastasis, and tumour stage. Furthermore, the distribution pattern o f P R L - 3 at the cell membrane may be correlated with the metastatic ability ofthe tumour cells (Miskad et al., 2004). This finding is consistent with previous reports that localization o f P R L - 3 at the plasma membrane enhanced the migration and invasion properties o f cells. These observations strongly suggest that P R L - 3 expression plays a significant role i n invasion and metastasis o f gastric carcinoma.  Since the striking observation was first made that P R L - 3 is implicated in colorectal metastasis (Saha et al., 2001), many similar studies have been performed. Peng et al. (2004) found, using immunohistochemistry (IHC) to investigate normal, primary, and metastatic colorectal tumour samples, that P R L - 3 expression intensity and positivity were not significantly different between normal colorectal epithelium and colorectal cancer, or between metastatic lymph nodes and liver metastasis. However, P R L - 3 expression was significantly higher in metastases when compared with normal colorectal epithelia and primary colorectal cancers. Kaplan-Meier survival curve analysis showed that P R L - 3  11  expression was correlated with shorten survival time o f colorectal cancer patients. These results suggest that P R L - 3 is a predictor o f liver metastasis and poor prognosis for patients with colorectal cancer (Peng et al, 2004). Kato et al. (2004) looked at 177 cases o f primary colorectal cancer, and 92 cases o f metastasis. Ofthe patients with any types o f metastases, - 8 0 % ofthe primary colorectal cancer tissues displayed increased P R L - 3 expression, and ofthe patients with no lung or liver metastases, - 4 0 % o f the primary colorectal cancer tissues had increased P R L - 3 expression. However, more than 90% o f the liver or lung colorectal metastases had increased P R L - 3 expression. This again demonstrates that high expression o f P R L - 3 is closely correlated with liver and lung metastasis in colorectal cancer. Additionally, P R L - 3 s i R N A treatment o f D L D - 1 cells (a colorectal cancer cell line with relatively abundant P R L - 3 transcript) caused a 32%> decrease in migration o f these cells, accompanied with morphological alterations, but had no effect on proliferation (Kato et al, 2004). These findings exemplify the importance o f P R L - 3 in colorectal cancer progression.  What is the molecular basis underlying the P R L ' s role i n oncogenic transformation and metastasis? In the past three decades, numerous studies have focused on identifying mechanisms o f cancer development and progression, developing novel therapeutic strategies, and improving surgical techniques. Cancer progression is a multi-step process from a localized primary tumour mass to an invasive secondary metastasis resulting from the interaction o f tumour cells with the organ microenvironment (Kassis et al, 2001; Price and Collard, 2001; Ellenrieder et al, 1999; Portera et al, 1998). This complicated process involves progressive changes from the primary tumour to the distal metastasis.  12  Cells from the primary tumour must migrate, settle in a remote organ, and grow in the secondary site. The essential steps in the development o f metastasis are: 1) neoplastic transformation, proliferation and angiogenesis; 2) detachment o f primary tumour cells from the lesion and invasion o f the adjacent stroma; 3) aggregation o f tumour cells into the circulation; 4) adherence o f tumour cells to the subendothelial basement membrane; and 5) invasion into the remote organ, and proliferation and angiogenesis. In addition, the newly formed metastases can be the source o f further metastasis, another characteristic o f progressive aggressiveness. Multiple properties have been observed i n metastatic cells, including cytoskeletal changes, loss o f adhesion, enhanced motility, and expression o f matrix metalloproteinases ( M M P s ) that degrade the basement membrane.  Cytoskeletal remodeling is required for cancer cell mobility (Ridley, 2000). The cytoskeleton controls cell characteristics and processes that are fundamental to carcinogenesis, such as morphology, cell growth, migration, differentiation, and gene expression (Varedi et al, 1997). Phosphorylation and dephosphorylation o f cytoskeletal proteins can regulate the shape o f cultured cells (Oliver et al, 2002; Mancini et al, 2002; Jackson et al, 1997). Thus, the changes i n cell shape o f the P R L - 3 over-expressing clones might be related to the cytoskeletal rearrangement induced by increased P R L - 3 activity. This suggests that P R L - 3 could play a key role in the cytoskeletal remodeling that is required for cancer cell motility (Kato et al, 2004). Moreover, protein prenylation is important in targeting proteins to intracellular membranes and i n protein-protein  13  interactions (Seabra, 1998; S i et al, 2001). The metastatic properties o f P R L - 3 may be dependent on prenyltransferase activity (Zeng et al, 2000 and 2003). Signaling pathways that control the cytoskeletal changes or the expression and activation o f proteases and their inhibitors i n metastasis are largely unknown.  P T P s were initially thought to have the potential for tumour suppression ( D i Cristofano and Pandolfi, 2000; Brown-Shimer et al, 1992). However, further studies have shown that some P T P s might have positive roles i n carcinogenesis. For example, P T P a activates Src family kinases by dephosphorylating the inhibitory C-terminal pTyr-527, thereby promoting cell transformation (Su et al, 1999; Ponniah et al, 1999; Zheng et al, 1992). P T P s and P T K s play key roles in the regulation o f cellular growth, differentiation, cell cycle, cell-cell communication, and other activities through the dephosphorylation and phosphorylation o f tyrosine (Hunter, 2000; Neel and Tonks, 1997; Tan, 1993). Deregulation o f phosphorylation is known to result i n neoplastic or non-neoplastic diseases (Streuli, 1996). Furthermore, PTPs are as important as the well-studied P T K s because phosphorylation is a dynamic and reversible process (Tonks and Neel, 1996). The study o f P R L - 3 promises to provide new insights into the biochemical pathways controlling various aspects o f metastasis. Stable expression o f wild-type active P R L - 3 dramatically enhanced cell migration, whereas the catalytically inactive P R L - 3 (C104S) mutant greatly reduced this ability (Zeng et al, 2003). These results indicate that the phosphatase activity o f P R L - 3 is required to promote cell migration.  14  For patients with solid tumours, the biggest threat to survival is tumour cell metastasis. Metastasis is the main cause o f death i n cancer patients because o f its resistance to conventional chemotherapy or radiotherapy (Susan and Schumacher, 2001). A representative metastasis model is hard to develop because it requires cell-cell and cellmicroenvironment interactions. It has become a major challenge, therefore, to identify the underlying molecular switches o f the metastatic state. H i g h levels o f P R L - 3 expression were generally not maintained upon in vitro culture o f cancer cells (Bardelli et al., 2003). P R L - 3 m R N A level i n metastatic cancer cell lines was significantly lower than generally observed i n metastatic cancer cells in vivo. In addition, gene expression patterns i n cancers i n their natural setting can be considerably different than those o f cell lines grown in artificial environments (Zhang et al., 1997). Thus, this truly presents some difficulties i n studying the comprehensive relation o f P R L - 3 and metastasis in vitro.  In contrast, there have been many advances i n understanding the earlier stages o f tumourigenesis revolving around cell autonomous processes that determine the rate o f cell division and cell death (Kinzler and Vogelstein, 1998). The earlier phases o f tumourigenesis have been associated with cell-specific patterns o f gene expression or mutation (Velculescu et al., 1999; Giordano et al., 2001; Dennis et al., 2002). P R L - 3 m R N A expression was elevated i n nearly all metastatic lesions derived from colorectal cancer, regardless o f the site o f metastasis (Bardelli et al., 2003). Conversely, the s i R N A knock-down o f the expression o f P R L - 3 in D L D - 1 cells suppressed metastatic tumour formation in vivo (Kato et al, 2004). This indicates that P R L - 3 can regulate not only cell motility but also the formation o f metastatic lesions. To fully understand the putative role  15  o f P R L - 3 in metastasis, P R L - 3 needs to be evaluated i n relation to known key players o f metastasis such as E-cadherin and Snail.  Metastasis is a process determined by cell-cell interactions and specific microenvironments (Weiss, 2000; Fidler, 2001; Chambers, 2002). Epithelialmesenchymal-transition ( E M T ) is a process whereby epithelial cells acquire fibroblastlike properties and exhibit reduced cell-cell adhesion and increased motility ( W u et al., 2004). Furthermore, E M T implies a dramatic phenotypic change that includes the loss o f epithelial markers (i.e. cytokeratin), the gain o f mesenchymal markers (i.e. vimentin), and changes i n cell shape (Vega et al, 2004). The process o f metastasis is associated with cell type specific gene expression, and E M T has been advocated to be a causative mechanism for the suppression o f E-cadherin and tumour progression (Yokoyama et al, 2003). The loss o f E-cadherin expression is a major characteristic o f highly invasive and metastatic cancers. Moreover, reduced expression o f E-cadherin is frequently associated with dedifferentiation, invasion, lymph node or distant metastasis, and poor prognosis in various human malignancies (Jiang, 1996; Jiao et al, 2001; M u k a i et al, 2001; Tanaka et al, 2002).  Snail is known to play a vital role i n E M T as it suppresses E-cadherin expression and transcriptionally upregulates M M P s . In particular, over-expression o f Snail has been shown to induce higher levels o f M M P - 2 expression and activity. Luciferase reporter assays revealed that the M M P - 2 promoter activity was induced by Snail (Yokoyama et al, 2003). M M P - 2 is involved in the degradation o f the basement membrane and the  16  extracellular matrix, which further facilitates the E M T process. Proteases that degrade the basement membrane are likely to be essential for invasion (Egeblad and Werb, 2002). This indicates that Snail has a fundamental role i n E M T through its suppression o f E cadherin and upregulation o f M M P - 2 . Since E M T is a complex and intricate process, there are possibly many other targets o f Snail waiting to be discovered.  Interestingly, P R L - 3 signals were also observed i n a subset o f endothelial cells o f some tumours. Since angiogenesis is a major component o f the process o f establishing a metastasis, P R L - 3 expression i n tumour vasculature could be relevant to this. Moreover, high P R L - 3 expression in primary colorectal cancer significantly correlated with venous invasion (Kato et al, 2004). Neo-angiogenesis is required at the metastatic site allowing micrometastases to grow into macrometastatic lesions. In contrast to other tumour types, invasive breast tumour vasculature exhibited high expression o f the Snail transcription factor (Parker et al, 2004). However, aside from its enhanced expression in breast tumour epithelium (Blanco et al, 2002; Cheng et al, 2001), Snail has not been linked to the transcriptional regulation o f genes important i n angiogenesis. P R L - 3 also appears to be expressed predominantly in the vasculature o f invasive breast cancers as well. Moreover, adenoviral-expressing P R L - 3 human microvascular endothelial cells were also found to have enhanced migratory ability in vitro (Parker et al, 2004).  Does P R L - 3 function as one o f the key players in the epithelial-mesenchymal-transition? The specific molecular changes that promote the spread o f tumour cells from the original site to other parts o f the body are still primarily unclear. However, nude mice injected  17  with P R L - 3 expressing cells exhibited tumour formation with metastasis (Zeng et al., 2003; Pathak et al, 2002), and the growth o f the tumours was markedly inhibited by the anti-protozoa drug pentamidine [l,5-di(4-amidinophenoxy)pentane] at a tolerable dose (Pathak et al, 2002). Although the specific protein substrate(s) for P R L - 3 has not yet been identified, blocking or reducing the functions o f P R L - 3 i n metastasis might be achieved by the inhibition o f prenylation and/or the catalytic inactivation o f P R L - 3 . Upregulated P R L - 3 i n cancer provides an excellent target for developing novel therapeutics, especially for tumours which are intractable due to cancer metastasis.  The mechanism o f P R L - 3 upregulation in colorectal tumour metastases was investigated by Saha et al. (2001). They demonstrated that P R L - 3 upregulation was due to gene amplification i n 25% ofthe metastases, and suggested that increased transcriptional activity likely accounted for increased P R L - 3 expression i n the rest ofthe cases. Since P R L - 3 is upregulated i n metastatic cells, it is possible that this protein is important for primary cancer cells to spread. Although many genetic alterations have been attributed to neoplastic development, very few have been associated with metastasis. Thus, P R L - 3 could be a potential molecular marker for clinical estimation o f tumour aggressiveness and a novel therapeutic target.  18  1.4  H Y P O T H E S E S A N D ATMS  The evidence presented above led to my working hypothesis that the upstream region o f the P R L - 3 gene contains regulatory elements that mediate its tumour and metastasisspecific transcriptional upregulation. Our first aim is to characterize regulatory elements in the P R L - 3 promoter that mediate P R L - 3 upregulation in primary and metastatic tumour cell lines using a combination o f software predictive and wet-lab based methods. A luciferase reporter assay system and bioinformatics w i l l be used to identify active and recognizable regulatory elements, active and novel elements/regions, and regions responsible for tumour and metastatic-specific regulation o f P R L - 3 . Future identification ofthe corresponding transcription factors that interact with these elements w i l l provide insight into cellular signals that regulate P R L - 3 expression and the processes o f tumourigenesis and metastasis.  More than 90% o f cancer-related deaths are due to metastatic disease. Since metastasis is often attributed to a poor prognosis and since P R L - 3 has been implicated in metastasis, we also hypothesize that over-expression o f P R L - 3 i n clinical tumour samples is correlated with an unfavorable prognosis. Thus, our second aim is to define tumour/metastatic tissues in which P R L - 3 is over-expressed, using Tissue Microarray technology and immunohistochemistry, and correlate these findings with clinical data to infer prognostication.  19  CHAPTER 2  MATERIALS AND METHODS  2.1  BIOINFORMATICS  The P R L - 3 gene is located on chromosome 8q24.3 (Bardelli et al, 2003). The sequence (~4.6kb) ofthe P R L - 3 promoter was obtained from the University o f California Santa Cruz ( U C S C ) Genome Browser (Wasserman and Sandelin, 2004; Karolchik et al, 2003) website (Human A p r i l 2003 Assembly; chr8:142,l 11,086-142,115,888; 5'CGGGCGCCCT  G A G G C G C C A T - 3 ) . The sequence was downloaded as a F A S T A r  file format for use in primer design, luciferase reporter plasmid construction, and subsequent bioinformatics analyses.  The web-based program Consite (Wasserman and Sandelin, 2004; Lenhard et al, 2003) was used to investigate the P R L - 3 promoter sequence for orthologue conservation (between mouse and human) and putative transcription factor binding sites (TFBSs). Transcription factors generally have distinct preferences towards specific target sequences. However, T F B S prediction alone is not sufficient as potential binding sites do not necessarily mean functional importance. Therefore, it is imperative to first consider orthologue conservation prior to T F B S analysis. Consite eliminates - 9 0 % o f predictions while retaining ~70-80%> o f experimentally validated sites (Wasserman and Sandelin,  20  2004). The combination o f orthologue conservation and putative T F B S searches greatly reduces the chance o f false predictions with only a slight reduction i n sensitivity.  To analyze the P R L - 3 promoter for potential transcript generation and plausible regulatory potential, the appropriate tracks were selected on the U C S C Genome Browser website. Data generated from expression sequence tags (ESTs) were used to align possible transcript positions. Regulatory potential is a new method to investigate promoter regulatory sites. It analyses the patterns o f nucleotide identity in subregions o f the interspecies alignment and classifies conserved regions as coding or regulatory (Elnitski et al, 2003). This regulatory potential algorithm is based i n part on the pattern of observed nucleotide identity.  Another web-based program, C I S T E R (Frith et al, 2001), was used to examine the possibility o f certain trans-activating factors working together i n a module to regulate P R L - 3 promoter activity. Biochemical specificity o f transcription is generated i n the cell nucleus by combinatorial interactions between transcription factors (Davidson, 2001). Thus, looking into m-regulatory modules can predict substantially better specificity than analysis o f isolated sites.  21  Analysis  Program  PRL-3 Promoter Sequence (April 2003 Assembly) Transcription Factor Binding Sites  UCSC Genome Browser  Consite  http://mordor.cgb.ki.se/cgibin/CONSITE/consite/  Conservation  Consite  http ://mordor.cgb .ki. se/cgibin/CONSITE/consite/  Transcripts (July 2003 Assembly) Regulatory Potential (July 2003 Assembly)  UCSC Genome Browser UCSC Genome Browser  http://www.genome.ucsc.edu/cgibin/hgGateway  CisRegulatory Module  CISTER  http ://zlab .bu.edu/~mfrith/cister. shtml  Website Address http://www.genome.ucsc.edu/cgibin/hgGateway  http://www.genome.ucsc.edu/cgibin/hgGateway  References Wasserman and Sandelin, 2004; Karolchik et al, 2003 Wasserman and Sandelin, 2004; Lenhard et al, 2003 Wasserman and Sandelin, 2004; Lenhard etal, 2003 Wasserman and Sandelin, 2004 Wasserman and Sandelin, 2004; Kolbe et al, 2004 Frith et al, 2001  Table 1. Bioinformatics Programs. A list o f the web-based programs and corresponding website addresses used for the bioinformatics analyses o f the P R L - 3 promoter.  22  2.2  MOLECULAR CLONING  2.2.1  Polymerase C h a i n Reaction ( P C R )  A l l P R L - 3 promoter regions were amplified using normal human genomic D N A (Clontech) as a template, with a final concentration o f 8ng//xl. Primer sequences are shown i n Table 2. A l l P C R reactions were carried out i n a final volume o f 25/xl.  Long Range (-4173 / +336), full length P R L - 3 promoter The P C R master m i x contained 1 0 % D M S O , l . O m M M g A O c , 160/*M d N T P , 160/iM forward and reverse primers each, 0.7 units o f Taq polymerase (Invitrogen), and 1.25 units of Accurase polymerase with the included P C R buffer ( B I O / C A N Scientific). The thermocycling steps were as follows: 94°C-2mins, 40 cycles (94°C-20s, 62°C-30s, 68°C3.5mins), and 68°C-7mins.  5 - U T R and M i d d l e Range sequences together (-3656 / +3361 The P C R master m i x contained 1 0 % D M S O , 1.5mM M g A O c , 160/iM d N T P , 160/xM forward and reverse primers each, 1 unit o f Taq polymerase (Invitrogen), and 1.25 units o f Accurase polymerase with the included P C R buffer ( B I O / C A N Scientific). The thermocycling steps were as follows: 94°C-2mins, 40 cycles (94°C-20s, 60°C-45s, 68°C3.5mins), and 68°C-7mins.  23  5 ' - U T R (-2027/+336) The P C R master m i x contained 1.5mM o f M g C l , 200/xM d N T P , 160/xM forward and 2  reverse primers each, and 1.5 units o f Taq polymerase with the included P C R buffer (Invitrogen). The thermocycling steps were as follows: 95°C-3mins, 40 cycles (95°C-30s, 7 0 ° C - l m i n , 7 2 ° C - l m i n ) , and 72°C-10mins.  M i d d l e Range (-3656 / -2027) The P C R master mix contained 1 0 % D M S O , 1.5mM M g A O c , 160/xM d N T P , 1 6 0 / M forward and reverse primers each, 1 unit o f Taq polymerase (Invitrogen), and 1.25 units o f Accurase polymerase with the included P C R buffer ( B I O / C A N Scientific).  The  thermocycling steps were as follows: 94°C-2mins, 40 cycles (94°C-20s, 60°C-30s, 68°C1.75mins), and 68°C-7mins.  CpG-64 (-4173 /-3656) The P C R master mix contained 1 0 % D M S O , 1.5mM M g A O c ,  \60LM  d N T P , 160piM  forward and reverse primers each, and 1.25 units o f Accurase polymerase with the included P C R buffer ( B I O / C A N Scientific). The thermocycling steps were as follows: 94°C-2mins, 40 cycles (94°C-20s, 60°C-30s, 6 8 ° C - l m i n ) , and 68°C-7mins.  24  Region  Coorc inates From To  Long Range (full length promoter) Middle Range + 5'-UTR  -4173  +336  -3656  +336  5'-UTR  -2027  +336  Middle Range  -3656  -2027  CpG-64  -4173  -3656  Primer Sequences  Primer Length / Tm  Forward: 5 '-CC A G G G G G A G C T C C G A G G T T C - 3 ' Reverse: 5 '-ATGGCGCCTCCCGACGGG-3' Forward: 5 -CGCAGCTTACTTCTCCTGAGACCCG-3' Reverse: 5 '-ATGGCGCCTCCCGACGGG-3' Forward: 5'-GACACCCACCGCCC ATTTTAACCTT-3' Reverse: 5 '-ATGGCGCCTCCCGACGGG-3' Forward: 5'-GGGCCTCC AGCCTCCTCC-3' Reverse: 5'-GGTGTCAGGCGC AGTGGCT-3' Forward: 5 '-CTGGCTGGGCTTCCTCTTCC-3' Reverse: 5 -TCTGGGCGACTCGGGGC-3'  21 / 68.79°C 18/ 70.48°C 25/ 69.61°C 18/ 70.48°C 25/ 70.14°C 18/ 70.48°C 18/ 63.78°C 19/ 63.97°C 20/ 63.45°C 17/ 64.62°C  T a b l e 2. P C R P r i m e r s . A list o f primer sequences used i n P C R to amplify specific regions o f the P R L - 3 promoter for eventual subcloning into the luciferase expression vector. The nucleotide position where transcription starts is designated as +1. Sequences upstream o f this transcriptional start site (TSS) have a negative (-) sign.. Sequences downstream o f this T S S have a positive (+) sign.. A l l primer solutions were made at a stock concentration o f 2 0 / i M .  2.2.2  T A - v e c t o r subcloning  The P C R products were subcloned into either the p C R 2 . 1 - T A or the X L - T A vectors using a TOPO-isomerase kit (Invitrogen). The multiple cloning sites in these two vectors are identical. Blue/white selection was used to identify the positive colonies. Insert  25  orientation was validated using internal restriction cut sites i n the P C R insert o f nonequivocal lengths. The positive and correctly oriented plasmids were then purified using a mini-prep kit ( Q I A G E N ) . The purified plasmids were cut with Kpn I and Xho I restriction enzymes (New England BioLabs). The digest-released promoter inserts were gel-purified using a gel-extraction kit ( Q I A G E N ) .  2.2.3  Luciferase expression vector subcloning  The pGL3-Basic (Promega) vector was transformed into competent D H 5 a (Invitrogen) cells, and were grown overnight in L B media with the appropriate antibiotics. The p G L 3 Basic vector was purified from the bacterial culture using a maxi-prep kit (Sigma). Due to D N A supercoiling, the Xho I site is hidden i n this vector, therefore a spiking double digest was performed. Briefly, Kpn I (New England BioLabs) was incubated with the pGL3-Basic vector at 37°C for 5hrs. Then, Kpn I and Xho /restriction enzymes (New England BioLabs) were added together to this digest mixture and incubated at 37°C overnight. The next day, the linearized vector was dephosphorylated using calf intestinal phosphatase (Invitrogen) at 37°C for lhr.  The purified promoter fragments were ligated, using T4 ligase (Invitrogen), with the above linearized and dephosphorylated empty luciferase vector overnight at 4°C. The insert to vector ratio was approximately 3:1 in the ligation mixture. The ligated plasmids were transformed into competent D H 5 a (Invitrogen) cells, and were grown overnight in L B media with the appropriate antibiotic. The P R L - 3 promoter plasmids were purified  26  from the bacterial culture using a maxi-prep kit (Sigma). Restriction enzyme digests and sequence analyses were carried out to confirm the validity o f the promoter constructs.  2.2.4  P R L - 3 promoter constructs  Reporter assays were conducted to assess P R L - 3 promoter activity. This method involves measuring the ability o f a D N A sequence (putative promoter) to stimulate transcription o f a 3' gene, usually luciferase, when a plasmid containing these sequences is introduced into cultured cells. The p G L 3 vector series (Promega) contains the Firefly luciferase gene. A l l o f the promoter regions (Figure 4) were cloned into the multiple cloning site o f the pGL3-Basic vector, which contains the Firefly luciferase gene and lacks any promoter elements. To normalize transfection efficiency, the p h R L - T K plasmid, which contains the Renilla luciferase gene, was co-transfected with the p G L 3 plasmids. The Firefly and Renilla luciferases have dissimilar enzyme structures and substrate requirements because o f their distinct evolutionary origins. These differences make it possible to selectively discriminate between their respective bioluminescent reactions. W i t h the DualLuciferase Reporter Assay kit (Promega), the luminescence from the Firefly luciferase reaction can be measured, and then quenched while simultaneously activating the luminescent reaction o f Renilla luciferase.  27  Transcription Start Site ( T S S )  pGL3-  Middle Range  -4173  -3656  r  5'-UTR  Luciferase  I  -2027  +1  I +336  Transcription Start Site ( T S S ) P  GL3-.  Middle Range  I  5'-UTR  Luciferase  I  -2027  -3656  Iranscription Start Site ( T S S )  pGL3-  +1  I +336  5'-UTR -2027  pGL3-  Luciferase +1  I  +336  Middle Range  Luciferase  I pGL3  'I  CpG-64  -4173  L . .  Luciferase  -3656  pGL3-Control, positive control  I sv4o Promoter  Luciferase  pGL3-Basic, empty vector phRL-TK, internal control  Luciferase T K Promoter  |  Luciferase  F i g u r e 4. P R L - 3 promoter constructs. Various regions o f the P R L - 3 promoter were cloned into the p G L 3 - B a s i c luciferase reporter plasmid. The full-length P R L - 3 promoter was divided into three discrete regions (top). The C p G - 6 4 island, M i d d l e Range, and the 5 ' - U T R sequences are represented by green, blue, and red boxes, respectively. The black box represents the Firefly luciferase gene. These plasmids were transfected into selected cancer cell lines to evaluate their activity and subsequently define a region(s) o f regulatory importance. The p h R L - T K vector was used as an internal control to normalize transfection efficiency. The grey box represents the Renilla luciferase gene.  ]  28  2.3  TRANSFECTION AND LUCIFERASE ASSAY  2.3.1  Cell line propagation  The HEK293T (derived from human embryonic kidney cells transformed with adenovirus 5 D N A , A T C C # : CRL-1573),  PRL-3-HEK293T (PRL-3 stably expressing  H E K 2 9 3 T cells created by D . Bessette in the Pallen lab),  C2C12 (derived from mouse  normal myoblastic cells, A T C C # : CRL-1772), RD (derived from human embryonal rhabdomyosarcoma cells, A T C C # : C C L - 1 3 6 ) , and  RMS13 (derived from human alveolar  rhabdomyosarcoma cells isolated from metastatic bone marrow, A T C C # : CRL-2061) cell lines were maintained i n D M E M media, supplemented with 10% F B S . The SW480 (derived from human colorectal adenocarcinoma cells, A T C C # : C C L - 2 2 8 ) and  SW620  (derived from human colorectal adenocarcinoma cells isolated from metastatic lymph node, A T C C # : C C L - 2 2 7 ) cell lines were maintained in D M E M media with 1.5g/L sodium bicarbonate, supplemented with 10%> F B S . The  LNCaP (derived from human  prostate carcinoma cells isolated from metastatic left supraclavicular lymph node, A T C C # : CRL-1740) cells were grown i n RPMI-1640 media with 2 . 0 m M L-glutamine and phenol red, supplemented with 15%> F B S . Lastly, the  DU145 (derived from human  prostate carcinoma cells isolated from metastatic brain, A T C C # : H T B - 8 1 ) cells were grown i n M E M media with Earle's salts (BSS), 2 . 0 m M L-glutamine, O . l m M nonessential amino acids, and l . O m M sodium pyruvate, supplemented with 10%> F B S . Propagation methods were employed as described by the American Type Culture Collection ( A T C C ) . A l l cell lines were grown in a 5 % C 0 incubator at 37°C. 2  29  2.3.2  Transfection  Cells were plated onto 6-well plates (Falcon) 24hrs prior to transfection. Each well contained 3.0mL o f growth media and 2 . 0 X 1 0 or 4 . 0 X 1 0 cells per well. Cells were 5  5  transfected by the cation-lipid method. A l l transfections were done according to the manufacturer's (Invitrogen) protocol. Briefly,  1.5/Ug o f  the Firefly luciferase p G L 3 -  plasmids (pGL3-Control, pGL3-Basic, or pGL3-promoter-constructs) and 0.5/ig o f the Renilla p h R L - T K plasmid (Promega) were mixed with 5juL o f Lipofectamine reagent (Invitrogen). The D N A - l i p i d complexes were then incubated with the selected cell lines in serum-free media for 5-6hrs. The cells were then switched to full growth media and allowed to grow for 24, 48, or 72hrs.  2.3.3  Luciferase Assay  Luciferase assays were performed using the Dual-Luciferase Reporter Assay System (Promega), following the manufacturer's protocol. A t confluency, transfected cells were washed with room temperature phosphate buffered saline (PBS) prior to addition o f Passive Lysis Buffer (Promega). Cells were lysed at room temperature for 30-40mins with orbital shaking. After addition o f the appropriate luminescent substrates, the signal was detected directly i n the 6-well plates using a PerkinElmer Victor-3 Multiplate reader.  30  2.4  WESTERN BLOT  2.4.1  Antibodies  A 9-amino-acid peptide ( D P H T H K T R C ) that is unique to P R L - 3 was used to immunize a rabbit for antibody production. This unique region is found near the C-terminal C A A X box o f P R L - 3 . After initial injection, three more injections o f the antigen were administered at 14-day intervals. Antibodies specific for P R L - 3 were purified from the immunized rabbit serum using a recombinant P R L - 3 affinity column. The final polyclonal anti-PRL-3 antibody concentration was about l m g / m l , and was used at a 1:5000 dilution. This antibody was produced by m y colleague Dr. Jing Wang. In addition, a monoclonal anti-PRL-3 antibody was generously donated to us by Dr. Q i Zeng (Institute o f Molecular and C e l l Biology, Singapore), and was used at a 1:2000 dilution. The anti-Snail antibody was purchased from Santa Cruz Biotechnology, Inc., S N A I 1 (E18): sc-10432, and was used at a 1:200 dilution. The a n t i - F L A G antibody was purchased from Sigma, and was used at a 1:1000 dilution. The anti-c-rayc antibody was purchased from Santa Cruz, and was used at a 1:1000 dilution.  2.4.2  Cell line treatments  A l l cells were harvested in RIP A lysis buffer ( l O m M sodium phosphate, 0 . 1 5 M sodium chloride, I m M E D T A , 5 0 m M sodium floride, 1% NP-40, 0.1% S D S , 10^g/ml aprotinin, I m M sodium orthovanadate, and O . l m M P M S F ) .  31  PRL-3-Stable-HEK293T  cells:  A P R L - 3 stably expressing H E K 2 9 3 T cell line was created by m y colleague M r . Darrell Bessette. P R L - 3 expression can be induced upon appropriate antibiotic treatment. Briefly, at - 8 0 % confluency, these cells were treated with l/i,g/ml o f Doxycycline i n full growth media, and harvested after 24hrs.  MDA-MB-435  cells:  These cells were derived from human ductal breast carcinoma cells isolated from metastatic pleural effusion, and were maintained in D M E M media with 1.5g/L sodium bicarbonate, supplemented with 10% F B S . These cells were used to visualize endogenous Snail protein as described in Zhou et al. (2004). Briefly, at - 8 0 % confluency and immediately after a media change (full growth media), the cells were treated with 4 0 m M L i C l (Sigma) together with 10/nM M G 1 3 2 (Sigma) for 5hrs prior to harvest.  HEK293T  cells:  Western Blot analysis determined that the affinity-purified anti-PRL-3 antibody specifically recognized P R L - 3 and not P R L - 1 or -2 (Dr. Jing Wang, personal communication). Moreover, the ability o f the immunizing P R L - 3 peptide to block recognition o f P R L - 3 protein by the antibody was tested. First, H E K 2 9 3 T cells were transfected with a plasmid encoding myc-tagged-PRL-3, and incubated in full growth media for 24hrs prior to harvest. The polyclonal anti-PRL-3 antibody was pre-incubated with the unique P R L - 3 peptide that was used to immunize the rabbit for antibody  32  production. The antibody-peptide solution was left to rotate at 4°C for 2hrs prior to use for Western Blot analysis o f P R L - 3 detection i n the cell lysate.  2.4.3  SDS-PAGE  A l l whole cell lysates were resolved on a 12.5% polyacrylamide gel and transferred onto a P V D F membrane at room temperature for l h r using an electrotransfer apparatus. The membrane was then blocked with 5%(w/v)-skim milk powder for l h r at room temperature. The primary antibody solution (in l%(w/v)-skim milk powder) was incubated with, shaking, the membrane at 4°C overnight. The next day, the primary antibody solution was removed and the membrane was washed with P B S at room temperature (shaking), 3 X lOmins. The membrane was then probed with the corresponding secondary antibody (also in l%(w/v)-skim milk powder) at room temperature for lhr, shaking. After 3 X lOmins P B S washes (shaking), the membrane was visualized with E C L solution on exposable film.  33  2.5  IMMUNOHISTOCHEMISTRY  2.5.1  Polyclonal anti-PRL-3 antibody specificity  To further confirm specificity, the polyclonal anti-PRL-3 antibody was pre-incubated with the same unique P R L - 3 peptide that was used to immunize the rabbit for antibody production. The antibody-peptide solution was left to rotate at 4°C for 2hrs prior to use for immunohistochemistry.  2.5.2  Immunohistochemistry staining  Formalin-fixed and paraffin-embedded tissue sections were deparaffinized and hydrated via conventional methods. After a brief tap water rinse, the slides were placed in hot (70°C) l O m M sodium citrate buffer at p H 6.0 for 40mins for antigen retrieval/pretreatment. After another tap water rinse, endogenous peroxidase activity was blocked by incubation with 3% H2O2 for 30mins at room temperature. The slides were then rinsed with phosphate-buffered-saline ( P B S ) at p H 7.4. Endogenous proteins were blocked with avidin and biotin blocking solutions (Vector) for 15mins each at room temperature with P B S washes i n between. Excessive hydrophobicity in the tissues was blocked by a 3% B S A solution (ID Labs) incubated for 30mins at room temperature. The solution was then drained, and without washing the slides, the primary antibody solution (1:200 dilution) was added to the sections for 30mins at room temperature. After a quick P B S wash, the slides were incubated with the secondary antibody (ID Labs) for 30mins at  34  room temperature. After another P B S wash, the labeling solution (Streptavidin-horseradish peroxidase, ID Labs) was added to the slides and incubated for 30mins at room temperature. After another P B S wash, 0 . 1 M sodium acetate solution at p H 5.2 was incubated with the sections for l m i n . The solution was drained and without washing, the chromagenic solutions were added to the sections. A t the desired staining intensity, the slides were placed in tap water to stop the colour change reaction and were later counterstained with hematoxylin. Slides were then finally mounted using an aqueous mounting media.  2.5.3  Tissue M i c r o a r r a y  The analysis o f many sample core sections mounted on a single slide, known as tissue microarray ( T M A ) , is efficient and reduces staining variability compared to analysis o f single sections on many slides. In addition, multiple-tumour tissue arrays are an efficient method o f assessing the sensitivity and specificity o f antibodies used in determining origin or cell type o f human tumours (Hsu et al, 2002). P R L - 3 expression in various T M A s was evaluated as follows.  After immunohistochemistry (IHC) staining, the staining intensity o f each core was assigned a score. A score o f 0 corresponded to 0-5% staining, a score o f 1 corresponded to 5-50%) staining, and a score o f 2 corresponded to >50%> staining. Generally, a score o f 0, 1, and 2 represented negative, weak, and strong staining, respectively. Uninterpretable cases were assigned an X .  35  CHAPTER 3  RESULTS  3.1  T R A N S C R I P T I O N A L A N A L Y S I S OF T H E PRL-3 P R O M O T E R  3.1.1  PRL-3 promoter activity in human embryonic kidney 293T cells  H E K 2 9 3 T cells exhibit high transfection efficiency and are suitable cells to be used as a model system to initially confirm plasmid expression and to optimize cell splitting, transfection, and luciferase assay techniques. Thus, H E K 2 9 3 T cells were transfected with the reporter constructs to evaluate the functionality o f the P R L - 3 promoter plasmids.  Luciferase activities measured in the panel o f transfected cells are shown i n Figure 5. A s expected, the pGL3-Control luciferase activity was significantly greater than pGL3-Basic and thus validated the luciferase assay. Cells transfected with the 5 ' - U T R (-2027 / +336), Middle Range (-3656 / -2027), and 5 ' - U T R + Middle Range (-3656 / +336) all exhibited no significant luciferase activity. However, the full-length promoter (-4137 / +336) resulted in ~6.8-fold (p<0.031) higher luciferase activity than observed with the empty vector (pGL3-Basic). Moreover the CpG-64 island (-4137 / -3656) alone induced -131.5-fold (p<0.01) higher luciferase activity than did the empty vector. When comparing the fold difference in activity between the full-length promoter (-4137 / +336) and the CpG-64 island (-4137 / -3656), the latter sequence was ~19.3-fold (p<0.03) more  36  effective in promoting luciferase expression. The CpG-64 island thus appeared to be an important region for P R L - 3 transcriptional regulation. A significant level o f promoter activity is only seen with the presence o f this CpG-64 island (alone or in the full-length promoter).  r  pGL3-Control pGL3-Bas ic  0.4  0.6  0.8  1.2  Relative Light Units  F i g u r e 5. P R L - 3 promoter activity i n H E K 2 9 3 T cells, l u g o f the pGL3-Control (Firefly), pGL3-Basic (empty vector), or p G L 3 - P R L - 3 promoter plasmids were cotransfected with l u g o f the p h R L - T K (Renilla) vector. After 24hrs i n full media, the cells were lysed and luciferase activity was measured. Three independent duplicate experiments were performed. The fold-increase in luciferase activity o f certain plasmids relative to p G L 3 - B a s i c was calculated, and the asterisk (*) depicts p-values <0.05 (student's t-test). The CpG-64 island (green box) and the full-length promoter showed a significant induction o f luciferase activity in H E K 2 9 3 T cells. The 5 ' - U T R (red box), Middle Range (blue box), and 5 ' - U T R + Middle Range sequences (red + blue boxes) exhibited no significant induction o f luciferase activity.  37  3.1.2  Auto-transcriptional regulation ofthe PRL-3 promoter in human embryonic kidney 293T cells  To investigate the possibility that P R L - 3 expression might be regulated v i a auto-activated transcription, the promoter activity ofthe p G L 3 - P R L - 3 promoter constructs was assessed in H E K 2 9 3 T cells with inducible expression o f heterologous P R L - 3 protein. These cells were stably transfected (Mr. Darrell Bessette, Pallen Lab) with a plasmid that expresses P R L - 3 protein in response to doxycycline treatment (Figure 6). Thus, the ability o f P R L 3 protein to interact with P R L - 3 promoter elements and thus regulate transcription was tested using HEK-FZ-v4G-PRL-3 cells treated with or without doxycycline.  Western blot analysis was performed to demonstrate that P R L - 3 is expressed in this inducible system. A s shown in Figure 6, P R L - 3 expression is evident upon antibiotic treatment (+). Conversely, no P R L - 3 can be detected without antibiotic treatment (-). The induced P R L - 3 protein has an N-terminal FLAG-tag; also probed with anti-FLAG  antibody. A FLAG-tagged  therefore, the cell lysates were protein was only detected in the  antibiotic treated cells (Figure 6), and thus further substantiating this P R L - 3 inducible system. This experiment was performed in conjunction with the luciferase reporter analysis o f the auto-transcriptional role o f P R L - 3 (Figure 7).  38  Ami-/ /.. 1(7 7  € M k  F i g u r e 6. Inducible expression of P R L - 3 in stably transfected H E K - F L ^ G P R L - 3 cells. Cells were untreated (-) or treated (+) with 1/ig/ml o f Doxycycline for 24hrs and harvested. C e l l lysates were probed for P R L - 3 expression using anti-PRL-3 (top panel) and a n t i - F L ^ G (bottom panel) antibodies.  Luciferase activities measured i n the panel of transfected cells that were induced to express P R L - 3 or were left un-induced are shown i n Figure 7. The pGL3-Control luciferase activity was significantly greater than that expressed by the pGL3-Basic plasmid in both P R L - 3 expressing and non-expressing cells, and thus validates the luciferase assay. Cells transfected with the 5 - U T R (-2027 / +336), M i d d l e Range (-3656 / -2027), and 5 - U T R + Middle Range (-3656 / +336) plasmids exhibited no significant luciferase activity either with or without induced P R L - 3 expression. However, in the uninduced cells, the full-length promoter (-4137 / +336) stimulated ~3.7-fold (p<0.003) higher luciferase expression than did the empty vector. In the P R L - 3 induced cells, the full-length promoter (-4137 / +336) had ~4.6-fold (p<0.001) higher transcriptional activity than the empty vector. There was no statistical difference when comparing luciferase activities induced by the full-length promoter (-4137 / +336) in the un-induced cells and the cells induced to express the P R L - 3 protein. This suggests that P R L - 3 has no direct auto-transcriptional role.  39  ,  1  1  pGL3-Control  1  —  ,  1 *  1 *  pGL3-Basic  ri—i  g . . v  H  . H i • . 0.05  PRL-3 Non-induced  ~)l.2X 0.1  0.15  0.2  0.25  Relative Light Units  PRL-3 Induced  F i g u r e 7. P R L - 3 auto-transcriptional role i n H E K 2 9 3 T cells. 1 ixg o f the p G L 3 Control (Firefly), pGL3-Basic (empty vector), or p G L 3 - P R L - 3 promoter plasmids were co-transfected with l u g o f the p h R L - T K (Renilla) vector. After 24hrs i n full media with and without l u g / m l o f Doxycycline, the cells were lysed and luciferase activity was measured. T w o independent duplicate experiments were performed. The fold-increase in luciferase activity o f the plasmids relative to pGL3-Basic was calculated, and the asterisk (*) depicts p-values <0.05 (student's t-test). The 5 ' - U T R (red box), M i d d l e Range (blue box), and 5 ' - U T R + Middle Range sequences (red + blue boxes) induced no significant level o f luciferase activity. A significant level o f activity from the full-length promoter (which contains the CpG-64 island) is observed in both P R L - 3 induced and non-induced conditions. However, there is no significant difference between P R L - 3 induction and non-induction on the promoter construct activities, suggesting that P R L - 3 has no direct auto-transcriptional role.  40  3.1.3  PRL-3 promoter activity in primary and metastatic colorectal cancer cell lines  P R L - 3 expression is upregulated in metastatic colorectal and gastric carcinomas (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al> 2004; M i s k a d et al, 2004), as well as in melanoma ( W u et al, 2004). T o investigate whether the P R L - 3 promoter activates transcription i n response to metastasis-specific factors, promoter activity was determined i n a matched pair o f tumour and metastatic cancer cell lines. The SW480 cell line was derived from the primary tumour o f a colorectal cancer patient. The SW620 cell line was derived from a metastatic site (lymph node) o f the same patient.  The luciferase activities measured i n panels o f pGL3-Control, pGL3-Basic, and p G L 3 P R L - 3 promoter plasmid transfected cells are shown i n Figure 8. The luciferase activity in the SW480 and SW620 cells transfected with the pGL3-Control plasmid was significantly greater than in the cells transfected with the p G L 3 - B a s i c plasmid, thus validating the luciferase assay. A s expected, both pGL3-Control transfected SW480 cells and SW620 cells exhibited similar luciferase activities, since both o f these cell lines were derived from the same patient. This further validates the reliability o f this luciferase assay. Luciferase expression was not significantly affected by the 5 ' - U T R (-2027 / +336) or the M i d d l e Range (-3656 / -2027) sequences i n either cell lines. The Middle Range and 5 ' - U T R sequences together (-3656 / +336) increased luciferase expression by ~3-fold (p<0.04) compared to pGL3-Basic (empty vector) i n the SW480 cells, but no difference was detected i n the SW620 cells. However, the additional presence o f the CpG-64 (-4137  41  / -3656) sequence with the M i d d l e Range (-3656 / -2027) and the 5 ' - U T R (-2027 / +336) sequences (full-length promoter) caused a dramatic and significant ~18-fold (p<0.002) increase i n luciferase activity i n the metastatic SW620 cells. A l s o , the full-length promoter induced a ~4-fold (p<0.01) elevation in luciferase activity in the SW480 cells. Together, these results suggest that the CpG-64 island i n the P R L - 3 promoter is active in colorectal carcinoma cells. The full-length promoter activity was ~4.5-fold (p<0.007) higher in the SW620 cells than i n the SW480 cells. The CpG-64 (-4137 / -3656) sequence alone induced a ~13-fold (p<0.004) increase i n luciferase expression in the SW480 cells and a.~30-fold (p<0.0001) increase i n the SW620 cells compared to that observed in these cells transfected with the pGL3-Basic vector. The CpG-64 promoter activity was ~2.3-fold (p<0.01) higher i n the SW620 cells than in the SW480 cells. This strongly argues for the transcriptional importance ofthe CpG-64 island in the P R L - 3 promoter, and suggests that P R L - 3 may potentially play a vital role in the metastatic process.  U s i n g interphase Florescent In-Situ Hybridization analysis, Bardelli et al. (2003) concluded that classic gene amplification was not invariably (only 25% o f cases) the cause o f P R L - 3 over-expression i n colorectal cancer metastasis. Since gene amplification is not the major cause o f increased P R L - 3 expression, then increased transcriptional activity is most likely responsible for P R L - 3 upregulation. This further substantiates the merits o f our investigation o f the P R L - 3 promoter.  42  pGL3-Control pGL3-Basic  fl8X*_ 4.5X*  0.02  SW480 (primary)  0.04  0.06  0.08  0.1  0.12  0.14  0.16  H30Xc  ] 2.3) (  0.18  0.2  Relative Light Units  SW620 (metastasis)  F i g u r e 8. P R L - 3 promoter activity i n S W 4 8 0 p r i m a r y a n d S W 6 2 0 metastatic colorectal carcinoma cell lines. 1.5ug o f the pGL3-Control (Firefly), p G L 3 - B a s i c (empty vector), or p G L 3 - P R L - 3 promoter plasmids were co-transfected with 0.5ug o f the p h R L - T K (Renilla) vector. After 72hrs in full media, the cells were lysed and luciferase activity was measured. Three independent duplicate experiments were performed. The fold-increase i n luciferase activity induced by various plasmids relative to pGL3-Basic was calculated, and the asterisk (*) depicts p-values <0.05 (student's t-test). The 5 - U T R (red box) and the Middle Range (blue box) did not induce significant levels o f luciferase activity in either cell line. However, the 5'-UTR + Middle Range sequences (red + blue boxes) induced a slight elevation in luciferase activity in the S W 4 8 0 cells, but not in the SW620 cells. Moreover, a significant level of activity from the full-length promoter (which contains the CpG-64 island) is observed in both cell lines with the full-length promoter sequence being more (~4.5-fold) active in the S W 6 2 0 cells than in the SW480 cells. A l s o , a significant level o f activity from the C p G - 6 4 island (green box) alone is observed in both cell lines with the CpG-64 sequence being more (~2.3-fold) active in the SW620 cells than in the SW480 cells.  43  3.1.4  PRL-3 promoter activity in metastatic prostate cancer cell lines  To compare P R L - 3 promoter activity in cell lines with different invasive properties, a less invasive androgen-dependent L N C a P cell line and a more invasive androgen-independent D U I 45 cell line (Ghosh et al, 2005) were used. These cell lines are o f prostate carcinoma origin, and L N C a P cells were derived from a nodal metastasis o f a patient, whereas the D U 1 4 5 cells were derived from the brain metastasis o f another patient.  The luciferase activities measured in the panels o f transfected L N C a P and D U I 4 5 cells are shown i n Figure 9. The 5 ' - U T R (-2027 / +336) or 5 ' - U T R and M i d d l e Range together (-3656 / +336) did not induce statistically significant increases in luciferase activity in either cell lines, but a slight elevation is apparent. However, transfection of the fulllength promoter (-4173 / +336) containing the additional CpG-64 sequence (-4173 / 3656) resulted i n a dramatic ~30.6-fold (p<0.003) increase in luciferase expression in L N C a P cells compared to that observed upon transfection with the empty vector. In the D U 1 4 5 cells, the full-length P R L - 3 promoter (-4137 / +336) stimulated a lesser ~8.5-fold (p<0.003) increase i n luciferase activity. Thus, the full-length P R L - 3 promoter is ~3.6fold (p<0.02) more active in the L N C a P cells than in the D U I 4 5 cells. Although not statistically significant (p>0.05), the CpG-64 island alone induced a ~6-fold increase (compared to pGL3-Basic) i n luciferase activity in the L N C a P cells. This is quite surprising as the CpG-64 island is usually very active. Comparatively, the same CpG-64 island i n D U 1 4 5 cells exhibited a -52.3-fold ( p O . O O l ) higher promoter activity  44  compared to p G L 3 - B a s i c . Thus, the CpG-64 promoter region is about ~8.7-fold (p<0.0004) more active in DU145 cells than in L N C a P cells.  0.2 L n C A P (node metastasis)  0.3  0.4  0.5  0.6  0.7  Relative Light Units  DU145 (brain metastasis)  Figure 9. PRL-3 promoter activity in less invasive androgen-dependent LNCaP cells and in more invasive androgen-independent DU145 cells. 1.5ug o f the pGL3-Control (Firefly), pGL3-Basic (empty vector), or p G L 3 - P R L - 3 promoter plasmids were cotransfected with 0.5ug o f the p h R L - T K (Renilla) vector. After 48hrs in full media, the cells were lysed and luciferase activity was measured. Three independent duplicate experiments were performed. The fold-increase in luciferase activity induced by the various plasmids relative to pGL3-Basic was calculated, and the asterisk (*) depicts pvalues <0.05 (student's t-test). The 5 ' - U T R (red box) or the 5 ' - U T R + Middle Range (red + box boxes) sequences did not induce significant levels o f luciferase activity in either cell lines. However, a significant level o f activity stimulated by the full-length promoter (which contains the CpG-64 island) was observed in both cell lines with the full-length promoter sequence being more (~3.6-fold) active in the L N C a P cells than in the D U 1 4 5 cells. A l s o , a statistically significant level of activity from the C p G - 6 4 island (green box) alone was observed in the DU145 cells but not in the L N C a P cells w i t h the C p G - 6 4 sequence being more (~8.7-fold) active in the DU145 cells than in the L N C a P cells.  45  3.1.5  PRL-3 promoter activity in normal muscle and rhabdomyosarcoma cell lines  P R L - 3 is naturally expressed i n striated muscle cells (Matter et al, 2001). Thus, the C 2 C 1 2 (normal mouse myoblast) cell line was used to examine P R L - 3 promoter activity. In addition, to evaluate the effect o f malignant transformation o f muscle cells on P R L - 3 promoter activity, the p G L 3 - P R L - 3 promoter plasmids were transfected into cancerous muscle cell lines as well. R D cells were derived from a well-differentiated embryonal rhabdomyosarcoma ( E R M S ) tumour, whereas R M S 13 cells were derived from a poorlydifferentiated alveolar rhabdomyosarcoma ( A R M S ) tumour.  The luciferase activities measured in transfected C 2 C 1 2 , R D , and R M S 13 cells are shown in Figure 10. M i n i m a l luciferase activity was induced by the 5 ' - U T R (-2027 / +336), 5 U T R and M i d d l e Range together (-3656 / +336), or the full-length promoter (-4173 / +336) in C 2 C 1 2 cells. Although not statistically significant, these same three P R L - 3 promoter constructs induced a higher luciferase activity in the rhabdomyosarcoma cell lines than in the C 2 C 1 2 cells. More measurements may confirm whether the possible trends in the observations are due to chance or real tendencies. The levels o f promoter activity observed i n the R D and R M S 13 cells from each o f these three constructs were similar. Moreover, the 5 ' - U T R and the Middle Range sequences together (-3656 / +336) promoted slightly lower luciferase expression than did the 5 - U T R region alone in R D and R M S 13 cells. Luciferase activity induced by the full-length P R L - 3 promoter i n the R D cells was about ~9.4-fold (p<0.02) higher than the empty vector, whereas it was about ~13.4-fold (p<0.002) higher i n the R M S 1 3 cells. The CpG-64 island (-4173 / -3656) was  46  the only sequence that showed a significant level (~14-fold, p<0.003) o f promoter activity in the C 2 C 1 2 cells. Since P R L - 3 is normally expressed i n striated muscle cells, and since no other promoter regions show a significant elevation in promoter activity, this strongly suggests that this CpG-64 island is essential for P R L - 3 transcriptional regulation. This same region exhibited ~91-fold (p<0.002) and ~55-fold (p<0.0003) higher activity i n the R D cells and R M S 13 cells, respectively. The CpG-64 sequence was ~6.5-fold (p<0.004) more active i n R D cells than i n C 2 C 1 2 cells, and ~3.9-fold (p<0.003) more active in R M S 13 cells than i n C 2 C 1 2 cells. Although not statistically significant, the CpG-64 promoter region was more active in the R D cells than in the R M S 13 cells. M o r e measurements are needed to validate this trend.  47  0.15 C 2 C 1 2 (normal myoblast)  0.2  0.25  0.4  Relative Light Units  R D (embryonal) RMS13 (alveolar)  Figure 10. PRL-3 promoter activity in normal muscle and rhabdomyosarcoma cell lines. 1.5ug o f the pGL3-Control (Firefly), pGL3-Basic (empty vector), or p G L 3 - P R L - 3 promoter plasmids were co-transfected with 0.5ug o f the p h R L - T K (Renilla) vector. After 48hrs in full media, the cells were lysed and luciferase activity was measured. Three independent duplicate experiments were performed. The fold-increase in luciferase activity induced by various plasmids relative to p G L 3 - B a s i c was calculated, and the asterisk (*) depicts p-values <0.05 (student's t-test). The 5 - U T R (red box) or the 5 ' - U T R + Middle Range (red + box boxes) exhibited no significant induction o f luciferase activity in C 2 C 1 2 cells, but a slight elevation in luciferase activity was effected by these promoter regions in both R D and R M S 13 cells. However, a significant level o f activity induced by the full-length promoter (which contains the CpG-64 island) was observed in R D and R M S 13 cells, but not in C 2 C 1 2 cells. A l s o , a significant level o f activity from the C p G 64 island (green box) alone was observed in all three cell lines with the C p G - 6 4 sequence being more active in the R D (~6.5-fold) and R M S 1 3 (~3.9-fold) cells than i n the C 2 C 1 2 cells.  48  3.2  BIOINFORMATICS A N A L Y S E S OF T H E PRL-3 P R O M O T E R S E Q U E N C E  Orthologue conservation analysis between mouse and human P R L - 3 promoters indicated that the CpG-64 island contains regions that are most highly conserved (Figure 11). A s an assumption, mutations within functional regions o f genes w i l l accumulate more slowly than mutations within non-functional regions. So from an evolutionary perspective, the comparison o f sequences from orthologous genes can indicate segments that might direct transcription (Wasserman and Sandelin, 2004). Sequence similarity that results from selective pressure during evolution is the foundation for many bioinformatics methods (Ureta-Vidal et al, 2003; Frazer et al, 2003). One key assumption i n the application o f phylogenetic footprinting is that the orthologous genes are regulated by the same mechanisms i n different species. Although the definition o f orthology does not include retained functionality, for the purpose o f phylogenetic footprinting one assumes that orthologous genes are under common evolutionary pressures.  Trans-activating factors generally have distinct preferences towards specific cis target sequences. However, identification o f putative transcription factor binding sites (TFBSs) does not necessarily translate into functional importance. Potential functionality is greatly enhanced by analyzing T F B S s i n highly conserved regions. To more accurately reflect the characteristics at each position o f a T F B S , a matrix that contains the number o f observed nucleotide at each position is created, called the position frequency matrix. After conversion o f this frequency matrix to a log-scale, a position weight matrix ( P W M ) is created. A quantitative score for any D N A sequence can now be generated by  49  summing the values that correspond to the observed nucleotide at each position.  Consite  eliminates - 9 0 % o f predictions while retaining -70-80%) o f experimentally validated sites (Wasserman and Sandelin, 2004). Therefore, the combination o f phylogenetic footprinting and P W M searches applied to orthologous human and mouse gene sequences reduces the rate o f false predictions by an order o f magnitude with only modest reduction in sensitivity. This analysis o f the P R L - 3 promoter region found four discrete regions, arbitrarily designated regions A - D , i n which all c/s-elements were located (Figure 11). Region A , in which the CpG-64 island is located, contains multiple cancer-associated ciselements, such as n-myc, M a x , Mef-2, and Snail (Table 3). There are also many other types o f cz's-elements in regions B , C , and D (Table 4). However, region A possesses the most abundance o f cancer-associated c/s-elements. Moreover, the CpG-64 island also exhibits high orthologue conservation as indicated by the peaks i n Figure 11.  50  C  CFI-USP Max USF ARNT ARNT Snai 1 USF n-MVC n-MYC  TFBS cutoff 8 0 %  D  Bsap S0X17 Bsap RORalfa-1 NF-kappaB HMG-IY SQUA HFH-3 CF2-II Sox-5 S0X17 ATHB5 ATHB5 HFH-2 HNF-3beta Athb-1 CF2-II Hunchback  Human i Mouse _  4637  Constructs  Transcription Start Site (TSS)  Middle Range  CpG-64  I -4173  -3656  I -2027  H +1  +336  Conservation  n u c 1 e o t l d e p o s i t i on  F i g u r e 1 1 . Putative T F B S s a n d ortbologue conservation alignment o f the h u m a n protein tyrosine phosphatase P R L - 3 promoter. The human and mouse P R L - 3 promoter sequences were submitted to Consite for orthologue conservation analysis and alignment. A window size o f 50 nucleotides and a conservation cut-off o f 70% were chosen. The sequences with 70% or more conservation were analyzed for putative transcription factor binding sites (TFBSs) with a cut-off o f 80%. There are many cancerassociated c/s-elements and high orthologue conservation within the P R L - 3 promoter, especially in the C p G - 6 4 island.  51  Region A (-4099 to -3886), 213bp Transcription Number of Coordinates Factor cw-elements Myf ARNT Snail  1 4 5  n-myc  4  Max  2  CREB USF CFI-USP bZIP910 TBP  AGL3 CF2-H HFH-3 MEF2  2 2 1 2 3  2 2 1 2  -4099 -4093 -4045 -4093 -4045 -3902 -3891 -4093 -4045 -4048 -4046 -4046 -4045 -4046 -4045 -4044 -4044 -4042 -4035 -4033 -4028 -4028 -4028 -4028 -4028  to -4088 to -4088 to -4040 to -4088 to -4040 to -3897 to -3886 to -4088 to -4040 to -4039 to -4037 to -4035 to -4034 to -4040 to -4039 to -4035 to -4038 to -4036 to -4021 to -4019 to -4014 to -4019 to -4019 to -4017 to -4019  Strand +  +/+/+ +/-  + +/+/+  +  +  +  + +/+/-  +/-  Table 3. Putative c/s-elements identified within the CpG-64 island of the PRL-3 promoter. Sequences upstream o f the transcriptional start site (TSS) are designated by a negative (-) sign. Region A is within the CpG-64 island (-4173 / 3656) ofthe P R L - 3 promoter. The P R L - 3 promoter sequences identified as T F B S s have at least 70% orthologue conservation (human / mouse) and 80% matching homology to known cz's-sequences.  52  Region B (-1341 to -1317), 24bp Transcription Number of Coordinates Factor c/s-elements CFI-USP Max USF ARNT Snail n-myc  1 1 2 2 1 2  -1341 -1325 -1323 -1322 -1322 -1322 -1322  to-1332 to-1316 to-1317 to-1316 to-1317 to-1317 to-1317  Region C (-407 to -13), 394bp Transcription Number of Coordinates Factor cis-elements Bsap  2  SOX17 RORalfa-1 NF-kB HMG-rY SQUA HFH-3  1 1 1 1 1 1  Transcription Factor  -407 -331 -398 -205 -149 -28 -26 -24  to -388 to -312 to -390 to -196 to-140 to -13 to-13 to-13  Region D (+159 to +177), 18bp Number of Coordinates c/s-elements  CF2-U  2  Sox-5 SOX17 ATHB5 HFH-2 HNF-3beta Athb-1 Hunchback  1 1 2 1 1 1 1  +159 +168 +163 +164 +166 +166 +166 +167 +168  to+166 to+177 to+169 to+172 to+174 to+177 to+177 to+174 to+177  Strand -  + -  + +/+/-  Strand -  + + .+ -  +  Strand + + + +/-  + + -  Table 4. Putative c/s-elements within ~1.5kb ofthe transcription start site of the PRL-3 promoter. Sequences upstream o f the transcriptional start site (TSS) are designated by a negative (-) sign, and those downstream o f the T S S are designated by a positive (+) sign. The P R L - 3 promoter sequences identified as T F B S s have at least 70% orthologue conservation (human / mouse) and 80%o matching homology to known c/s-sequences.  53  Transcript data were obtained from alignments between human expressed sequence tags (ESTs) i n GenBank and the genome. E S T s are single-read sequences, typically about 500 bases i n length that usually represent fragments o f transcribed genes. In general, the 3'E S T s mark the end o f transcription reasonably well, but the 5'-ESTs may end at any point within the transcript. To be considered spliced, an E S T must show evidence o f at least one canonical intron (i.e. one that is at least 32 bases i n length and has G T / A G ends). Evidence o f transcripts potentially starting at the CpG-64 island (spliced and unspliced human ESTs) is apparent and suggests that the first P R L - 3 exon may originate from this region (Figure 12).  A new method analyses the patterns o f nucleotide identity i n subregions ofthe alignment and classifies conserved regions as coding or regulatory (Elnitski et al, 2003). This 'regulatory potential' algorithm is based on the pattern o f observed identical nucleotides between species. Coding regions tend to vary at the third codon position and have insertion/deletion lengths that are multiples o f three. However, non-coding regulatory sequences tend to have more frequent insertions/deletions and variations occur i n distinct blocks that are separated by segments o f high similarity (Wasserman and Sandelin, 2004). When homologous sequences from more than two species are available, blocks o f strongly conserved sequences can effectively predict binding sites for transcription factors (Gumucio et al, 1992; Hardison et al, 1997b). A three-way (human / mouse / rat) regulatory potential analysis was used to evaluate the P R L - 3 promoter (Figure 12). A broad peak is present only within the CpG-64 island. This signifies this region as having  54  transcriptional importance. Although there are sharp discrete peaks i n the Middle Range (-3656 / -2027) region, this area lacks significant nucleotide identity; thus, the functional implication is limited.  Biochemical specificity o f transcription is generated in the cell nucleus by combinatorial interactions between transcription factors (Davidson, 2001). Thus, gene regulation that is mediated by cooperative interactions between transcription factors that bind to clusters o f sites within cw-regulatory modules ( C R M s ) can be defined using computational software to improve performance. Compared with the rate o f predictions o f individual T F B S s , the focus on C R M s eliminates - 9 9 % o f false T F B S predictions while retaining 60%> functional regions (Wasserman and Sandelin, 2004). Analyzing groups o f cw-elements in a module substantially improves the prediction specificity over analyzing isolated sites. Analysis o f C R M s within the P R L - 3 promoter reveals that only the CpG-64 island contains such modules. Mef-2 has the highest posterior probability, on either strand, for cooperative binding. Mef-2 is a trans-activating factor responsible for myogenesis and is a muscle-specific marker (Chen et al, 2001; Molkentin et al, 1995; Naidu et al, 1995; Black and Olson, 1998). Since P R L - 3 is naturally expressed in striated muscle, then it follows that muscle-specific c/s-elements exist within the P R L - 3 promoter. Moreover, only the CpG-64 island contains such myogenic c«-sequences, suggesting that this promoter region is transcriptionally essential for P R L - 3 expression.  55  Constructs  Transcription Start Site  (TSS)  5'-UTR  Middle Range  r  1  1  Transcripts  Human ESTs That Have Been Spliced  • •  Human ESTs Including Unspliced  •  ••••••  Regulatory Potential  6,91 _ 3-Way Regulatory Potential - Human, Mouse (Feb, £883/mm3), Rat (June 2883/rn3)  C;'s-Requlatorv M o d u l e s  500  i  1 •K  0-8 H  ~  0.2 H  -8 *  o  b  1000  position in sequence Cbases> 1500 2000 2500  3000  a 0.6 US  3  0.8  cis-elements:  F i g u r e 12.  cluster probability TATA Spl CRE  NF-1  E2F  CCAAT  AP-1  T r a n s c r i p t generation, regulatory potential, and c/s-regulatory  module  a n a l y s e s o f t h e h u m a n p r o t e i n t y r o s i n e p h o s p h a t a s e PRL-3 p r o m o t e r . This figure displays spliced and unspliced human ESTs starting from the C p G - 6 4 island. Regulatory potential analysis defined a broad band over the CpG-64 island. Moreover, many cismodulators exist within the CpG-64 island o f the P R L - 3 promoter. Mef-2 and T A T A binding proteins appear to have the highest probability o f cooperative binding.  56  3.3  E N D O G E N O U S SNAIL A N D PRL-3 E X P R E S S I O N  Bioinformatics analyses o f the P R L - 3 promoter revealed that putative Snail binding sites were the most abundant predicted czs-element (Figure 11 and Table 3). A l l these sites were located within the CpG-64 island. In addition, Snail transcription factor activity is involved i n tumourigenesis and metastasis (Vega et al, 2004), suggesting an interesting potential link to P R L - 3 . Thus, I wished to determine i f Snail expression induced P R L - 3 expression. In cell lines, Snail expression is usually low as it is highly unstable, with a short half-life o f about 25min (Zhou et al, 2004). GSK-3(3 binds to and phosphorylates Snail at two consensus motifs to dually regulate the subcellular localization and function of this protein (Zhou et al, 2004). In the hypo-phosphorylated state, Snail is active and is present i n the nucleus where it functions as a transcription factor. U p o n phosphorylation by GSK-3P, phospho-Snail is relocated into the cytoplasm. A n additional phosphorylation o f Snail by GSK-3P tags this hyper-phosphorylated Snail for proteosomal degradation. Thus, in order to study the effects o f endogenous Snail, its degradation needs to be inhibited. A s previously reported by Zhou et al. (2004), L i C l and M G 1 3 2 were used to treat M D A 4 3 5 invasive breast ductal carcinoma cells to inhibit GSK-3(3 and proteosomal activities, respectively. Western Blot analysis demonstrated that these treatments enhance Snail protein level (Figure 13). Notably, probing the same cell lysates for P R L - 3 revealed that increased P R L - 3 expression is seen with endogenous Snail accumulation (Figure 13). In the absence o f the L i C l and M G 1 3 2 inhibitory treatments, no Snail was seen and very little P R L - 3 was detectable. This could reflect a positive regulatory effect o f Snail on P R L - 3 transcription. B y inhibiting  57  (hyper)phosphorylation o f Snail, active Snail accumulates i n the nucleus. A t the same time, P R L - 3 protein level also increases as Snail protein is stabilized. Thus, this suggests a correlation between Snail and P R L - 3 .  No Treatment  +LiCl +MG132  Figure 13. Enhanced PRL-3 protein level is seen with inhibition of endogenous Snail degradation in breast cancer cells. Invasive ductal breast carcinoma cells ( M D A - M B - 4 3 5 ) at - 8 0 % confluency were either left untreated or treated with 4 0 m M L i C l (lithium chloride) and 10/xM M G 1 3 2 for 5hrs prior to harvesting i n R I P A lysis buffer. Untreated cells have no Snail expression and extremely low P R L - 3 expression. When Snail degradation is inhibited by L i C l and M G 1 3 2 , P R L - 3 protein increases drastically. I thank m y colleague, M s . H o a Le, for her assistance with the P R L - 3 Western blot.  58  3.4  PRL-3 EXPRESSION IN CLINICAL T U M O U R S A M P L E S  3.4.1  Polyclonal anti-PRL-3 antibody specificity  The following studies o f P R L - 3 protein expression i n patient tumours relied upon recognition o f P R L - 3 b y P R L - 3 polyclonal antibody. It was thus important to demonstrate antibody specificity for P R L - 3 to ensure that signal detected upon probing tumour samples really represented P R L - 3 . Specificity was initially confirmed to heterologous P R L - 3 expression in cultured cells. The polyclonal anti-PRL-3 antibody was pre-incubated with the same unique P R L - 3 peptide that was used to immunize the rabbit for antibody production. The antibody-peptide solution was left to rotate at 4°C for 2hrs prior to use for immunohistochemistry. If this antibody is specific only to this peptide, then pre-iricubating the antibody and peptide together before use w i l l neutralize the function o f this antibody. This was tested by probing lysates o f H E K 2 9 3 T cells that had been transfected with a myc-tagged P R L - 3 expression plasmid. Western blot using the polyclonal anti-PRL-3 antibody showed a tight band at the expected weight o f P R L - 3 (22KDa) and no other non-specific bands (Figure 14). Pre-incubating the P R L - 3 antibody with the competitive peptide at 5 X the concentration o f the antibody considerably reduces the signal to a faint band. A t 10X blocking peptide concentration, the P R L - 3 band is completely abolished. This demonstrates the specificity o f the antibody to the epitope that it was raised against. In addition, probing with the c-myc antibody confirms that the P R L - 3 band was produced from the transfected construct.  59  Coomassie P R L - 3 P R L - 3 P R L - 3 c-myc Blue Stain A b only A b + 5 X A b + l O X Ab peptide peptide  Figure 14. PRL-3 antibody specificity analysis using immunoblot. H E K 2 9 3 T cells were transfected with a myc-tagged P R L - 3 expression plasmid. Transfected cells were harvested after growth in full media ( D M E M , 10%FBS) for 24hrs. The whole cell lysate was resolved by S D S - P A G E and stained with Coomassie Blue to visualize total protein, or transferred to a membrane and probed with the polyclonal anti-PRL-3 antibody alone or antibody pre-incubated with different blocking peptide concentrations. The mti-c-myc antibody was used to verify mycP R L - 3 expression and detection. The positions o f protein standards are shown to the left (KDa).  Immunohistochemical detection o f P R L - 3 using the polyclonal anti-PRL-3 antibody was initially performed and optimized by staining cardiac tissue, because P R L - 3 is naturally expressed in striated muscle (Matter et al, 2001). P R L - 3 expression (red) was detectable in the cytoplasm as well as in the nucleus o f cardiac muscle cells (Figure 15). Preincubation o f the P R L - 3 antibody with the antigenic peptide used to raise this antibody abolished antibody recognition o f P R L - 3 protein (Figure 15). This demonstrates that the immunohistochemical signal detected in these tissue sections is not due to non-specific binding o f the anti-PRL-3 antibody.  60  Figure 15. PRL-3 immunohistochemical staining specificity in cardiac muscle. 128.5X. Human cardiac tissues were stained with the polyclonal anti-PRL-3 antibody alone or after pre-incubation with the blocking peptide at either 5 X or 1 0 X the concentration o f the antibody. The A E C chromagenic substrate was used to visualize P R L - 3 expression (red), and the sections were counterstained with hematoxylin (blue).  To rule out the possibility o f non-specific serum antibody contamination during the affinity purification process, the cardiac sections were also probed with pre-immune rabbit serum (Figure 16). Some faint background staining was observed with the preimmune serum but this was less intense than that observed using the P R L - 3 antibody. This is generally expected with polyclonal antibodies, but more so with serum. However, upon addition o f the P R L - 3 peptide, the staining pattern or intensity did not change. This suggests that the red staining pattern was not due to P R L - 3 expression and represents a non-specific reactivity that is not present in the purified anti-PRL-3 antibody preparation.  61  Pre-immune serum  + 5 X PRL-3 peptide  + 10X PRL-3 peptide  Figure 16. Cardiac muscle immunohistochemical staining with pre-immune rabbit serum. 128.5X. Human cardiac tissues were stained with pre-immune rabbit serum before or after pre-incubation with 5 X or 10X P R L - 3 blocking peptide. The A E C chromagenic substrate was used to visualize staining (red), and the sections were counterstained with hematoxylin (blue).  3.4.2  PRL-3 expression in Colorectal Carcinoma  Colorectal cancer is the second leading cause o f cancer-related death in both men and women in the United States. The incidence o f colorectal cancer has remained stable in the Western world for the last 100 years (Damjanov, 2000). Colorectal cancers are twice as common in men as in women, but the incidence does not reflect gender discrimination.  The study conducted by Saha et al. (2001) showed that only the P R L - 3 gene was consistently over-expressed in all metastatic colorectal carcinoma samples analyzed. A s an initial test and a proof-of-principle, a colorectal cancer T M A (Imgenex) was used to initially optimize P R L - 3 I H C staining conditions. Ten normal colon epithelia, forty primary carcinomas, and ten matching metastatic cases were examined. Figure 17 shows that the normal colon tissues showed no I H C staining for P R L - 3 . In contrast, more than  Q2 half o f the primary carcinoma cases exhibited strong P R L - 3 staining, and roughly a quarter o f the cases showed weak P R L - 3 expression, with the remaining cases either being negative for P R L - 3 expression or deemed uninterpretable (Table 5). In accordance with other studies (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004), 70% o f colorectal metastatic cases showed greatly elevated expression o f P R L - 3 (Table 5). These findings are consistent with others as Bardelli et al. (2003) reported that over 80%) o f colorectal cancer metastatic cases over-express P R L - 3 . In addition, Peng et al. (2004) discovered that over 60% o f colorectal cancer metastases display overexpression o f P R L - 3 , with detectable P R L - 3 expression in primary carcinomas and minimal detection in normal colon epithelium. Similarly, Kato et al. (2004) found that more than 90% o f colorectal metastatic cases over-express P R L - 3 . A s seen here, our findings are consistent with previous observations, thus confirming that P R L - 3 is indeed over-expressed in colorectal metastases, and validating our method o f investigation.  63  Primary Tumour  Metastasis  Figure 17. Immunohistochemical detection of PRL-3 in a colorectal carcinoma tissue-microarray. 88.5X. The basophilic appearance o f the crypt cells suggests that P R L - 3 is not expressed i n normal colon epithelium (arrows: normal colon). Prominent P R L - 3 staining is seen in the primary tumours o f patients #1 and #3, but only weakly in patient #2. The metastatic lesions are generally more intense in staining compared to their matching primary tumours, especially in patients #2 and #3. P R L - 3 localization appears to be mainly in the cytoplasm (arrows: cancerous lesions).  IHC Score (Staining) 0 1 2  (Negative) (Weak) (Strong)  X  (uninterpretable)  Normal  Primary Tumour  Metastasis  100% 0 0 0  10% 22.5% 52.5% 15%  0 0 70% 30%  Table 5. Frequency table of PRL-3 expression in colorectal carcinoma TMA. A tissue microarray containing cores o f 10 normal colon samples, 40 cases o f primary colorectal carcinoma, and 10 cases o f colorectal metastasis was analyzed for P R L - 3 expression.  64  3.4.3  PRL-3 expression in U v e a l M e l a n o m a  Intraocular melanomas are the most common primary ocular malignancy i n Caucasians (Griffin et al, 1998). Compared to cutaneous melanoma with a 1:90 to 1:100 chance o f developing disease over a lifetime for the Caucasian population, uveal melanoma has an annual expected frequency o f 6 cases per million in the adult Caucasian population (Osterlind, 1987). In 1998, White et al. (along with Horsman and White, 1993; Sisley et al, 1990) reported that certain cytogenetic findings were consistent among uveal melanoma samples. In particular, isochromosome 8q was associated with a poor prognosis due to death by metastatic disease (White et al, 1998). Moreover, the gain o f 8q was significantly associated with poor overall survival (Aalto et al, 2001). The P R L - 3 gene is situated on chromosome 8q24.3, and P R L - 3 over-expression has been associated with metastatic disease (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004; M i s k a d et al, 2004; W u et al, 2004). P R L - 3 expression in uveal melanoma was explored to elucidate the potential o f P R L - 3 as a prognostic marker for uveal melanoma.  The uveal melanoma T M A was obtained from Dr. Valerie White. It contained duplicate cores o f 184 formalin-fixed, paraffin embedded samples from eyes enucleated for large primary non-irradiated uveal melanoma. Diameter cores o f 0.6mm were consecutively taken from tumour tissue and implanted into an empty paraffin block (in duplicates), with 1.0mm between each core. Figure 18 shows representative I H C detection o f P R L - 3 in uveal melanoma. A s seen i n Table 6, a majority o f the cases had weak P R L - 3 expression.  65  The distribution o f P R L - 3 localization between cytoplasm and nucleus was similar. Only a small portion o f the cases on this uveal melanoma T M A had strong P R L - 3 staining. Kaplan-Meier survival curve analysis showed a positive trend (Figure 19). Although not statistically significant, patient samples that were positive for P R L - 3 had a poorer survival prediction than those patient samples that were negative for P R L - 3 . Nonetheless, more patient samples need to be evaluated before drawing a statistically relevant conclusion.  66  Figure 18. Immunohistochemical detection of PRL-3 in a uveal melanoma tissuemicroarray. 191.5X. A, the case seen here was heavily pigmented with melanin (brown) and showed little P R L - 3 staining (red), it was thus assigned a score o f 0. Although this tumour contained a lot o f melanin, the basophilic nuclei were still seen; B, this case had many positively staining nuclei, but not more than 50% o f these nuclei were red, hence the score o f 1; C , this case is a Spindle type B lesion and the localization o f P R L - 3 stain was mainly i n the cytoplasm and not in the nucleus, a score o f 2 was assigned to this patient due to evidently strong stain; D, this case is an epitheliodal tumour, P R L - 3 staining was predominantly localized i n the cytoplasm as seen by the circumferential staining pattern (arrow), a score o f 2 was also assigned to this patient.  67  Percentage of Cases  IHC Score (Staining) 0 1 2 X  18.5% 60.9% 13.0% 7.6%  (Negative) (Weak) (Strong) (uninterpretable)  Table 6. Frequency table of PRL-3 expression in a uveal melanoma TMA.  1.0  -r  0.90.80.7-  Negative  !*)£-  10.5co 0.4-  Positive  0.30.20.10.0-  —i—i—i—i—i—i—i—r  0 1  2  3  4  5  6  7  0  1  9  1 1—I 10 11 12 13 -  Time (years) Figure 19. Kaplan-Meier survival curve for uveal melanoma patient samples with negative and positive IHC staining for PRL-3. This is a disease-specific survival plot with 107 patients analyzed. Scores o f 0 and 1 were grouped as 'negative', whereas a score o f 2 was considered as 'positive'. This plot shows no statistically significant correlation between I H C o f P R L - 3 and survival (p=0.11). However, a positive trend is observed as patients with positive staining cores have poorer survival outcome compared to those with negative staining scores.  68  3.4.4  PRL-3 expression in Breast Carcinoma  Breast cancer is the most common cancer in women, surpassed i n lethality only by lung cancer. A n estimated 1 i n 14 woman w i l l develop breast cancer during their life span (Damjanov, 2000). A t least 180,000 new cases o f breast cancer are diagnosed each year. Unfortunately, the number o f new cases is increasing steadily. Approximately 46,000 women die o f breast cancer every year in the United States.  A T M A with 495 cases o f breast carcinoma was obtained from D r . Blake Gilks and stained for P R L - 3 . Localization o f P R L - 3 I H C staining is found to be cytoplasmic as well as nuclear (Figure 20). The majority ofthe cases were negative for cytoplasmic P R L - 3 , with only a small group o f cases having high levels o f cytoplasmic P R L - 3 staining (Table 7). Cases with high cytoplasmic P R L - 3 expression had a poor clinical prognosis compared to cases with no / low cytoplasmic P R L - 3 expression (Figure 21). Conversely, majority o f the cases were positive for nuclear P R L - 3 I H C staining (Table 8). Cases with high nuclear P R L - 3 expression had no correlation with clinical outcome o f the patients (Figure 22). Carcinoma in-situ lesions have 100% ofthe epithelial cells classified as neoplastic, exhibit prominent features o f dysplastic changes, and typically have basement membrane involvement. Stromal invasion is not seen i n carcinoma insitu. Invasive / infiltrating lesions have all the characteristics o f carcinoma in-situ but with the additional presences o f typical 'Indian files'. This single-file cell arrangement pattern is found in the surrounding stroma, and is a good indicator o f disease aggressiveness.  69  Figure 20. Immunohistochemical detection of PRL-3 in a breast carcinoma tissuemicroarray. 191.5X. This figure shows the commonest form o f breast cancer, ductal carcinoma, which accounts for roughly 7 5 % o f all carcinomas o f the breast. Nests o f cells and retained tube-like structures are seen in ductal carcinomas. A high level o f P R L - 3 staining (red) was observed in both o f these cases. Strong P R L - 3 staining was localized to the cytoplasm as well as in the nucleus o f these cancerous cells. A, the panel on the left depicts a carcinoma in-situ lesion (arrow), the neoplastic cells that are surrounded / encompassed by the myoepithelial cells exhibit typical dysplastic features o f a high-grade carcinoma lesion, and stromal involvement is not seen this is case; B, the panel on the right shows a high-grade invasive ductal carcinoma lesion. The metastatic / invasive cells appearing as a single file are seen wandering in the stroma (arrow). Stromal involvement is a typical feature o f invasive carcinoma.  70  I H C Score (Staining) 0 (Negative) 1 (Weak) 2 (Strong) X  (uninterpretable)  Percentage of Cases 53.1% 26.3% 2.63% 17.97%  Table 7. Frequency table of cytoplasmic P R L - 3 expression i n breast carcinoma T M A . P R L - 3 expression was analyzed by immunohistochemical staining o f 495 carcinomas. The staining intensity was scored from 0-2 as shown.  No PRL-3 |>0.6H  Low PRL-3  1 0.5H  High PRL-3  6o 0 . 4 -  0.30.20.10.0  0  ->  1  r  10  20  Time (years) F i g u r e 21. K a p l a n - M e i e r survival curve of cytoplasmic P R L - 3 expression i n breast carcinoma. N o , low, and high I H C staining for cytoplasmic P R L - 3 expression was plotted against patient outcome data (n=405). This is a disease-specific survival plot. A score o f 0 was considered to be null for P R L - 3 expression. Scores o f 1 or 2 represented low and high staining intensity, respectively. This plot shows a significant correlation (p=0.0441) between I H C expression o f cytoplasmic P R L - 3 and survival. Patients with high levels o f P R L - 3 have a poorer prognosis than those with no / l o w levels of P R L - 3 .  71  I H C Score (Staining) 0 (Negative) 1 (Weak) 2 (Strong) X  (uninterpretable)  Percentage of Cases 13.54% 21.21% 47.07% 18.18%  Table 8. Frequency table of nuclear P R L - 3 expression i n breast carcinoma T M A . P R L - 3 expression was analyzed by immunohistochemical staining o f 495 carcinomas. The staining intensity was scored from 0-2 as shown.  High PRL-3  g>0.6H  I 0.5co 0.4-  No PRL-3 Low PRL-3  0.30.20.10.0  0  10  20  Time (years) F i g u r e 22. K a p l a n - M e i e r s u r v i v a l curve of nuclear P R L - 3 expression i n breast carcinoma. N o , low, and high I H C staining for nuclear P R L - 3 expression was plotted against patient outcome data (n=406). This is a disease-specific survival plot. A score o f 0 was considered to be null for P R L - 3 expression. Scores o f 1 or 2 represented low and high staining intensity, respectively. This plot shows no significant correlation (p=0.16) between I H C expression o f nuclear P R L - 3 and survival.  72  3.4.5  P R L - 3 expression in Pediatric T u m o u r s  Cancer i n childhood and adolescents represents 2-3% o f all cancers, but is the leading cause o f non-violent deaths i n the 1-15 year age group i n developed countries. However, ofthe 200,000 new cases o f cancer that arise every year i n this age group, 85%> occur in developing countries, with the expectation that this proportion w i l l soon rise to 90%. Childhood cancers differ markedly from adult cancers in type, biology, genetic features, and diagnostic requirements.  A s a preliminary survey, I H C was performed on 60 cases o f the commonest types o f pediatric tumours on a T M A (14 cases o f Primitive Neuroectodermal Tumour ( P N E T ) / E w i n g ' s sarcoma, 13 cases o f neuroblastoma, 13 cases o f rhabdomyosarcoma, 7 cases o f fibromatosis, 6 cases o f congenital fibrosarcoma, and 7 cases o f miscellaneous rare tumours). The majority o f the cases exhibited high levels o f P R L - 3 expression compared to control tissue (fetal muscle).  Neuroblastoma accounts for 6% o f all pediatric cancers and is the most common cancer i n children under 2 years o f age. Neuroblastoma is a peripheral nervous system tumour and typically arises from the sympathetic nervous system, mainly from the adrenal medulla. The poorly differentiated form o f this tumour is the blastic neuroblastoma, and the well differentiated form is called ganglioneuroma. Patients with ganglioneuromas typically have a better clinical picture compared to patients with neuroblastomas. Rhabdomyosarcoma ( R M S ) accounts for 4%> o f all pediatric cancers and is the most  73 common soft tissue sarcoma in childhood. Rhabdomyosarcoma arises from skeletal muscle and is analogous to fetal muscle at 10-12 weeks gestation. The more primitive form o f R M S with little muscle differentiation is called alveolar R M S , and the typically more differentiated form is called embryonal R M S . Alveolar R M S is usually more aggressive than embryonal R M S (Kouraklis et al, 1999). In general, poorly differentiated neuroblastomas and rhabdomyosarcomas are classically more prognostically unfavorable. Due to the associations o f n-myc and Mef-2 with neuroblastoma and rhabdomyosarcoma respectively, and the presence o f these ciselements within the P R L - 3 promoter, I H C expression for P R L - 3 in neuroblastoma and rhabdomyosarcoma was considered (Figure 23). Although the cohort was small, strong P R L - 3 I H C staining is evident in more than 60% o f the neuroblastoma and rhabdomyosarcoma cases (Tables 9 and 10). One interesting observation is that P R L - 3 tends to be more localized i n the nucleus o f cells i n the poorly differentiated tumours, such as blastic neuroblastoma and alveolar rhabdomyosarcoma, and localized in the cytoplasm o f cells in the well differentiated tumours, such as ganglioneuroma and embryonal rhabdomyosarcoma. For neuroblastoma, this observation is statistically valid as all o f the poorly differentiated (blastic) neuroblastoma tumours exhibit nuclear P R L - 3 I H C staining and all o f the well differentiated (gangliocytic) neuroblastoma tumours exhibit cytoplasmic P R L - 3 I H C staining (Table 11). A s for rhabdomyosarcoma, there was no significant statistical correlation between tissue type and P R L - 3 localization (Table 12). More cases o f both neuroblastoma and rhabdomyosarcoma need to be investigated to further confirm the relationship between P R L - 3 I H C localization and differentiation type i n pediatric tumours.  74  Figure 23. Immunohistochemical detection of PRL-3 in a pediatric tumour tissuemicroarray. Neuroblastoma, 71.8X. Rhabdomyosarcoma, 143.6X. This figure shows two fairly common forms o f solid tumours found in children. Neuroblastoma is typically a poorly-differentiated tumour with a predominant blastic appearance. Ganglioneuroma is a well-differentiated form o f neuroblastoma. Rhabdomyosarcoma can be generally divided into two categories; alveolar (typically less differentiated) and embryonal (typically more differentiated). This figure shows that P R L - 3 is generally localized i n the nucleus o f poorly differentiated tumour cells and is predominantly localized i n the cytoplasm o f well-differentiated tumour cells.  IHC Score (Staining) 0 1 2 X  (Negative) (Weak) (Strong) (uninterpretable)  Percentage of Cases 7.7% 23.0% 61.5% 7.8%  Table 9. Frequency table of PRL-3 expression in pediatric neuroblastoma. The T M A contained 13 cases o f neuroblastoma.  75  IHC Score (Staining) 0 (Negative) 1 (Weak) 2 (Strong) X (uninterpretable)  Percentage of Cases 15.5% 23.0% 61.5% 0  Table 10. Frequency table of PRL-3 expression in pediatric rhabdomyosarcoma. The T M A contained 13 cases of rhabdomyosarcoma.  Fisher's Exact 2-Tail Test pO.Ol Localization Cytoplasmic Nuclear Total  Differentiation / Type Poor / Blastic W e l l / Gangliocytic 0 3 8 0 8 3  Total 3 8 11  Table 11. Pediatric Neuroblastoma 2X2 Contingency Table. X - statistical analysis 2  was used to test for independence between I H C o f P R L - 3 localization and tumour differentiation status o f neuroblastoma. This table shows a significant correlation between the two variables ( p O . O l ) as all o f the poorly differentiated (blastic) neuroblastoma tumours exhibit nuclear P R L - 3 I H C staining and all o f the well differentiated (gangliocytic) neuroblastoma tumours exhibit cytoplasmic P R L - 3 I H C staining.  Fisher's Exact 2-Tail Test p>0.05 Localization  Cytoplasmic Nuclear Cytoplasmic/Nuclear Total  Differentiation / Type Poor / Alveolar W e l l / Embryonal 0 1 4 2 3 1 7 4  Total 1 6 4 11  Table 12. Pediatric Rhabdomyosarcoma 2X2 Contingency Table. X - statistical 2  analysis was used to test for independence between I H C o f P R L - 3 localization and tumour differentiation status o f rhabdomyosarcoma. This table shows no significant correlation between the two variables (p=0.43). Cases with both cytoplasmic and nuclear stains were excluded from statistical analysis on this table.  76  CHAPTER 4  DISCUSSION  4.1  I M P O R T A N C E OF T H E CPG-64 I S L A N D I N T H E PRL-3 P R O M O T E R  P R L - 3 is a relatively new phosphatase that has received much attention over the past few years. P R L - 3 has been shown to be significantly over-expressed i n colorectal (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004), gastric (Miskad et al, 2004), and melanoma metastasis ( W u et al, 2004). P R L - 3 over-expressing C H O cells exhibit enhanced metastatic properties such as migratory abilities and invasiveness (Zeng et al, 2003). However, very little is known about the molecular actions o f this protein, let alone the genetic properties that regulate its transcription. Since P R L - 3 is'overexpressed i n metastatic colorectal lesions, and increased gene copy number o f P R L - 3 is not the primary cause o f this over-expression (Bardelli et al, 2003), I investigated the transcriptional regulation o f P R L - 3 expression by characterizing the P R L - 3 promoter. The P R L - 3 upstream sequence (-4173 / +336) was evaluated for its overall and regionspecific promoter activity, with the particular aim o f identifying sequences active i n promoting transcription in metastatic cells or cells with metastatic properties. O f the promoter regions evaluated via reporter assays and bioinformatics, the CpG-64 sequence is generally the most active, and most probable as a promoter o f P R L - 3 expression.  77  I established that the P R L - 3 promoter has differential transcriptional activities in various cell lines exhibiting assorted levels o f invasiveness, metastatic state, malignancy, and cellular differentiation. Luciferase reporter assays confirm that the CpG-64 island is an important region for P R L - 3 transcriptional regulation. The CpG-64 island is more active in colorectal metastatic cells than in primary tumour cells, more active i n invasive prostate androgen-independent tumour cells than i n androgen-dependent tumour cells, and more active i n malignant muscle cells than in normal muscle cells. Bioinformatics analyses revealed that this CpG-64 island has multiple cancer-associated cz's-elements, high orthologue conservation, potential transcript generation, and detectable regulatory potential. Other regions i n the P R L - 3 promoter have some, but not all, o f these above characteristics. Together, these findings suggest that the CpG-64 island is a candidate alternative promoter region that is differentially utilized i n various stages o f cancer development and progression.  4.1.1  CpG-64 island: The only or alternative promoter to PRL-3?  C p G islands are short D N A sequences / regions that are typically a few hundred base pairs in length and are highly enriched with the nucleotides cytosine and guanine. Methylation on cytosine silences transcription. Within regulatory sequences, C p G s remain unmethylated, whereas up to 80% o f C p G s i n other regions are methylated on a cytosine (Wasserman and Sandelin, 2004). Approximately 60% o f all human promoters are situated proximally to C p G islands (Gardiner-Garden and Frommer, 1987). Functional regulatory regions that control transcription rates are usually proximal to the  78  transcription start site(s). The current set o f laboratory-annotated regulatory sequences indicates that sequences near a T S S are more likely to contain functionally important regulatory controls than those that are more distal; however, the exact position o f a T S S can be difficult to locate due to alternative start sites (Wasserman and Sandelin, 2004). Computationally, the C G dinucleotide imbalance can be a powerful tool for finding regions i n genes that are likely to contain promoters (Hannenhalli and Levy, 2001).  In addition, C R M analysis also showed that only the CpG-64 island o f the P R L - 3 promoter contains a T A T A box. The posterior probability o f this T A T A box is the second highest o f the positive strand binding trans-activating factors. This again argues for the fact that the CpG-64 island can be a promoter for P R L - 3 . T A T A / G A T A regions are typically used for induction o f genetic expression, whereas G C - r i c h regions are mainly utilized for maintenance and housekeeping genetic expression (Mura et al, 2004). Having a T A T A box within a C G - r i c h region (CpG-64 island) suggests that P R L - 3 has differential levels o f expression. The C G - r i c h regions cause low / housekeeping level expression o f P R L - 3 , whereas the T A T A box causes a high level o f P R L - 3 expression when the appropriate stimuli are present for induction. C G - r i c h regions are typically located i n the vicinity o f the transcription start site and are sufficient to direct transcription initiation i n T A T A - l e s s promoters. In addition, they usually display multiple transcription start sites and a lower level o f R N A synthesis than transcripts directed by T A T A promoters (Blake et al, 1990; L u et al, 1994; Pugh and Tjian, 1990).  79  W i t h respect to potential P R L - 3 transcript generation, spliced and unspliced human E S T s can be found within the full-length P R L - 3 promoter and appear to begin at the CpG-64 island (Figure 12). This suggests that the CpG-64 island can be a promoter o f P R L - 3 transcription. The 3-way regulatory potential was investigated between human, mouse, and rat to determine the level o f possible functional importance o f specific regions o f the P R L - 3 promoter. A broad peak was predicted that extended across the CpG-64 island (Figure 12), indicating that this promoter region can potentially regulate the transcription of the P R L - 3 gene. There are also two sharp peaks within the M i d d l e Range (-3656 / 2027) region. Regulatory potential analysis is based on the principle o f interspecies nucleotide identity correlating with functional importance. The Middle Range region exhibits minimal orthologue conservation and no putative T F B S s . Thus, these two sharp peaks in the M i d d l e Range regions may have little or no functional significance. Additionally, c/s-regulatory modules can only be found within the CpG-64 island and nowhere else i n the P R L - 3 promoter (Figure 12). Trans-activating factors mainly work i n a combinatorial fashion, or module, for promoter regulation. The lack o f any C R M s within the P R L - 3 promoter other than the CpG-64 island demonstrates the likely negligible transcriptional importance o f the rest o f the P R L - 3 promoter regions. This again validates the CpG-64 island as acting as a promoter for P R L - 3 gene expression.  In accordance with these bioinformatics analyses, luciferase reporter assays also indicate that the CpG-64 island has high promoter activity. It is typically the strongest transcriptional director among the various P R L - 3 promoter regions tested, especially in the more invasive / metastatic cell lines. In H E K 2 9 3 T cells, the CpG-64 island alone  80  induced a -131.5-fold increase in luciferase activity compared to the empty vector (Figure 5). Other promoter regions lacking the CpG-64 island induced no significant luciferase activity. The full-length promoter sequence containing the CpG-64 island also stimulated significant luciferase activity, whereas removal ofthe C p G - 6 4 island resulted in negligible induction o f luciferase expression. This suggests that the CpG-64 island indeed contains transcriptionally important regions whereas other P R L - 3 promoter sequences lack such activity, at least i n the cell lines and under the conditions examined.  In colorectal cancer cell lines (Figure 8), the CpG-64 island exhibited -2.3-fold higher promoter activity i n the metastatic cells (SW620) than i n the less invasive primary tumour cells (SW480). These results, obtained by conducting luciferase reporter assays i n a metastatic colorectal carcinoma cell line and a matching cell line derived from the primary tumour ofthe same patient, indicate that the C p G - 6 4 island can potentially be involved i n the metastatic transition o f colorectal carcinoma cells. This same CpG-64 island has high luciferase activity even i n the SW480 primary tumour cells, providing further evidence o f the transcriptional importance o f this promoter region. Moreover, the full-length P R L - 3 promoter containing the CpG-64 sequence is also more active i n the SW620 cells than in the SW480 cells, while this enhanced activity was lost upon deletion ofthe C p G - 6 4 region. Again, this supports a metastasis-associated promoter function o f this C G - r i c h region i n regulating P R L - 3 expression.  Consistent with the above findings in H E K 2 9 3 T and colorectal cancer cell lines, the CpG-64 island exhibited -8.7-fold higher promoter activity in the more invasive (DU145)  81 metastatic prostate cancer cells than i n the less invasive ( L N C a P ) metastatic prostate cancer cells (Figure 9). Interestingly, the CpG-64 island promoter activity was unusually low i n the L N C a P cells. Perhaps this is due to the androgen-dependent and less invasive nature o f L N C a P cells, while P R L - 3 has been associated with invasive properties. If the CpG-64 island is indeed the sole promoter o f P R L - 3 transcription, then it is reasonable that this promoter region would be less active i n a less invasive cell line. Since the C p G 64 island activity is higher i n the D U 1 4 5 cells than in the L N C a P cells, it would be expected that the same relationship is maintained when the CpG-64 is placed i n conjunction with other promoter regions (i.e. the full-length promoter). Surprisingly however, the full-length promoter stimulated ~3.6-fold higher luciferase expression in the less invasive L N C a P cells than in the more invasive D U 1 4 5 cells. The basis o f the inverse promoter activities o f the full-length and the CpG-64 P R L - 3 upstream sequences in L N C a P versus D U 1 4 5 cells is unclear. However, since both cell lines were derived from prostate tumour metastases, this might reflect activities regulated i n an unknown manner by the differential androgen-dependence or invasive abilities o f these cell lines. Lastly, since P R L - 3 is normally expressed in striated muscle, a normal skeletal muscle cell line (C2C12) was used to investigate P R L - 3 promoter activity. Surprisingly in C 2 C 1 2 cells, none o f the P R L - 3 promoter constructs induced any significant levels o f luciferase activity except the CpG-64 island. This again argues for the function o f the CpG-64 island as the major, i f not only, regulator o f normal P R L - 3 expression. To investigate which P R L - 3 promoter region(s) is operative i n malignant transformation, rhabdomyosarcoma ( R M S ) cell lines were used. A l l promoter regions (full-length and truncations) were comparatively more active in the R M S cell lines than in the C 2 C 1 2 cell  82  line. The CpG-64 island was ~6.5-fold more active as a promoter in the R D (embryonal R M S ) and ~3.9-fold more active in the R M S 13 (alveolar R M S ) cell lines compared to i n the C 2 C 1 2 cells (Figure 10). This again supports the promoter function o f the CpG-64 island in transcriptional control and also demonstrates its potential role i n up-regulating P R L - 3 expression in the malignant phenotype. Other promoter regions do not exhibit these drastic differences i n activity between normal and malignant muscle cell lines.  The luciferase reporter assay results i n various cell lines indicate that the CpG-64 island functions as the most important ( i f not the only) region ofthe P R L - 3 promoter. In the absence ofthe CpG-64 island, very little promoter activity is typically seen. This suggests that the CpG-64 island could possibly be the only P R L - 3 promoter. In malignant / metastatic transformation however, other promoter regions may participate in transcriptional regulation, as observed i n the prostate and muscle cancer cell lines.  4.1.2  CpG-64 island and cancer-associated c/s-elements  The CpG-64 island is potentially a promoter for P R L - 3 gene transcription. Thus, putative T F B S s that appear within the C p G - 6 4 island may exert significant transcriptional control on P R L - 3 expression. In addition, orthologue conservation (human / mouse) shows that this promoter region is highly conserved, further illustrating the functional importance o f the CpG-64 island. Although there were many cw-elements identified within the CpG-64 island ofthe P R L - 3 promoter (Figure 11 and Table 3), I have chosen four transcription factors (Max, n-myc, Mef-2, and Snail) for further discussion. These transcription factors  83  have demonstrated associations with muscle gene expression and / or the regulation o f gene expression in processes such as proliferation and cancer that are potentially relevant to P R L - 3 transcriptional regulation.  4.1.2A  Max  The M y c / M a x / M a d network comprises a group o f transcription factors whose distinct interactions result in gene-specific transcriptional activation or repression. A great deal o f research indicates that this protein network plays roles in cell proliferation, differentiation, and death. Both M y c and M a d form heterodimers with M a x . These D N A bound heterodimers recruit coactivator or corepressor complexes that generate alterations in chromatin structure, which in turn modulate transcription (Grandori et al, 2000). In their biochemical behavior, the M a d proteins are similar to the M y c family members in that they homodimerize poorly but form specific heterodimers with M a x which can be detected in vitro and in vivo. Furthermore, M a d - M a x dimers recognize the same C A C G T G E-box sequence as do M y c - M a x dimers (Ayer et al, 1993; Zervos et al, 1993; H u r l i n e r a / . , 1995b). M y c proteins are prominent oncogenic transcription factors. O n the other hand, the correlation between M a d gene expression and differentiation has been confirmed by biochemical studies using cells that can be induced to undergo terminal differentiation in vitro. The increase in M a d protein levels is accompanied by a switch from M y c - M a x complexes to M a d - M a x complexes following induction o f differentiation (Ayer and Eisenman, 1993; Hurlin et al, 1995a). Thus, the M y c - M a x complex is responsible for cellular proliferation, whereas the M a d - M a x complex is responsible for  84  cellular differentiation. There are two M a x binding sites within the CpG-64 island o f the P R L - 3 promoter (Figure 11 and Table 3). Stable expression o f heterologous P R L - 1 , -2, or -3 increases cell growth rates (Diamond et al, 1994; Cates et al, 1996; Matter et al, 2001). Thus, it is plausible that the M y c - M a x complex binds to the C p G - 6 4 island and activates transcription o f P R L - 3 , and this in turn enhances proliferation. Moreover, it is also interesting to note that n-myc binding sequences are in proximity to M a x binding sequences within the CpG-64 island. This suggests a possible relationship between M y c M a x interaction and P R L - 3 expression in proliferating cells. O n the other hand, the M a d M a x complex may also bind to these M a x sites. Mef-2 binding sequences are also found within the C p G - 6 4 island, and Mef-2 is an activator and marker for muscle differentiation (Chen et al, 2001; Molkentin etal, 1995; Naidu et al, 1995; Black and Olson, 1998). Since P R L - 3 is naturally expressed in mature striated muscle, then it is plausible that the M a d - M a x complex can also bind to these sites and, together with Mef-2, activate P R L - 3 expression i n differentiating muscle cells.  4.1.2B  N-myc  The M y c family proteins function i n the integrated coregulation o f both proliferative and apoptotic signal transduction pathways (Cole and M c M a h o n , 1999). More specifically, the n-myc protein appears to act as a transcriptional regulator and is thought to govern the transcription o f critical genes involved i n mitogenesis and multidrug resistance. A close correlation between the expression o f n-myc and the multidrug resistance-associated protein gene has been reported (Norris et al, 1996; Haber et al, 1999). The childhood  85  tumour neuroblastoma is characterized by n-myc amplification i n aggressive and highly proliferative tumours that occur in a subset o f patients ( K i m and Carroll, 2004). N-myc gene amplification is found in -20-30% o f all neuroblastoma cases (Misawa et al, 2000). Direct evidence that n-myc contributes to tumourigenesis has been obtained from transgenic mice over-expressing n-myc, that showed a high incidence o f neuroblastoma (Weiss et al, 1997). Chronic upregulation o f n-myc i n the S H - E P neuroblastoma cell line accelerates progression into S-phase early after mitogenic stimulation o f quiescent cells (Lutz et al., 1996) and enhances the malignant phenotype o f neuroblastoma cells (Schweigerer et al, 1990). Since the n-myc protein is a well known transcription factor and is one o f the most powerful prognostic factors in predicting poor survival for neuroblastoma patients (Seeger et al, 1985), n-myc may also be involved i n P R L - 3 expression as there are four n-myc binding sequences within the CpG-64 island o f the P R L - 3 promoter. The n-myc binding sequence is the second most abundant c/s-element within the CpG-64 island. Thus, n-myc may be one o f perhaps several trans-activating factors that cause P R L - 3 over-expression i n highly metastatic and aggressive cancers.  4.1.2C  Mef-2  The capability to escape the differentiation program and undergo deregulated proliferation is a hallmark o f tumour cells. This may be achieved by several mechanisms, such as disrupting cell growth regulatory pathways and inactivating functional proteins that stimulate the differentiation program (Puri et al, 2000). Rhabdomyosarcoma ( R M S ) is a malignant tumour arising from skeletal muscle. It is the most common soft tissue  86  sarcoma in children under the age o f 15, accounting for 5-8% o f all cases o f childhood cancers (Miller et al, 1995; Pappo et al, 1995), and is the third most common extracranial solid tumour after neuroblastoma and W i l m ' s Tumour. R M S arises from muscle precursor cells that fail to complete the differentiation program despite the expression o f muscle regulatory proteins (Dias et al, 1992; Tapscott et al, 1993). R M S can be divided into two main categories, embryonal ( E R M S ) and alveolar ( A R M S ) . E R M S mainly occurs in young children and accounts for 60% o f all R M S cases. Histologically, malignant spindle and round cells that contain skeletal muscle crossstriations are located i n specific sites; such as the head/neck, genitourinary tract, and orbit. A R M S mainly occurs in adolescents as primary tumours o f the extremities or trunk. Histologically, fibrovascular septa form alveolar-like spaces filled with monomorphous malignant cells (Pappo et al, 1995; Pappo and Shapiro, 1997; Maurer et al, 1977). Mef-2 is a transcription factor involved in skeletal muscle differentiation and activation o f muscle-specific gene expression (Chen et al, 2001; Molkentin et al, 1995; Naidu et al, 1995; Black and Olson, 1998). Mef-2 localizes to the nucleus during muscle differentiation. Trafficking o f Mef-2 into the nucleus o f A R M S cells is impaired and thus, the transcriptional activity o f myogenic factors is impaired i n A R M S cells. Moreover, Mef-2 function is significantly compromised i n A R M S cells (Chen et al., 2001). Mef-2 is expressed i n A R M S and E R M S , but accumulates i n the cytoplasm o f A R M S cells. In contrast, E R M S cells accumulate large amounts of Mef-2 i n the nucleus upon differentiation stimuli. Inappropriate dysfunctional trafficking o f Mef-2 in A R M S cells relative to E R M S cells may correlate with the poor prognosis for individuals with A R M S (Chen et al, 2001). Since P R L - 3 is naturally expressed i n striated muscle and  87  Mef-2 binding sites (identified v i a bioinformatics analysis o f putative T F B S ) were found within the CpG-64 island, it is possible that P R L - 3 is also implicated i n malignant muscle transformation. C R M s analysis indicates that Mef-2 has the highest probability o f cooperative binding within the CpG-64 island o f the P R L - 3 promoter, thus P R L - 3 may be involved in myogenesis as well. Furthermore, Mef-2 binding sites are only found within this CpG-64 island. Luciferase reporter assays carried out i n normal and malignant muscle cell lines (Figure 10) demonstrated that the P R L - 3 promoter is i n general more active in the R M S cell lines than i n the normal muscle cell line (C2C12), suggesting a relationship between myocytic malignant transformation and P R L - 3 promoter activity. Although not statistically different (borderline), the luciferase activity induced by the CpG-64 island is higher i n the R D cells ( E R M S ) than i n the R M S 13 cells ( A R M S ) . The slight increase i n the CpG-64 island promoter activity may be attributed to the fact that Mef-2 is functional and localizes more i n the nucleus in E R M S cells than i n A R M S cells. In addition to other trans-activating factors contributing to the CpG-64 promoter activity, Mef-2 may play an additive role i n E R M S ; hence, the observed increase i n promoter activity o f the CpG-64 island in R D ( E R M S ) cells compared to R M S 13 ( A R M S ) cells.  4.1.2D  Snail  Epithelial-mesenchymal transition ( E M T ) is the process whereby epithelial cells acquire fibroblast-like properties and show reduced intercellular adhesion and increased motility. It is typically associated with the loss o f E-cadherin expression (Thiery, 2002; and Nieto, 2002). In cancer, down-regulation o f E-cadherin is a key step towards the invasive /  88 metastatic phase o f carcinoma, and dominant transcriptional repression is largely responsible for the loss o f E-cadherin expression (Behrens et al, 1991; Birchmeier et al, 1991, Hajra et al, 1999; J i et al, 1997; Giroldi et al, 1997). Snail is a zinc-finger transcription factor known to repress E-cadherin expression and induce E M T (Batlle et al, 2000; Guaita et al, 2002; Cano et al, 2000). Snail is expressed i n the invasive cells o f tumours induced in the skin o f mice (Cano et al, 2000), biopsies from patients with ductal breast carcinomas (Cheng et al, 2001; Blanco et al, 2002), gastric cancer (Rosivatz et al, 2002), and hepatocellular carcinoma (Sugimachi et al, 2003). Snail appears as an early marker o f the malignant phenotype and has implications as a prognostic factor (Blanco et al, 2002). The E-box sequence that binds Snail is the most abundant c/s-element within the CpG-64 island o f the P R L - 3 promoter. Since both P R L 3 and Snail have been implicated in cancerous metastasis and invasion, P R L - 3 and Snail may have a functional relationship. Nevertheless, since Snail itself is a key regulator o f E M T , it likely has many other targets, i n addition to E-cadherin. I have demonstrated a possible relationship between endogenous active Snail accumulation and P R L - 3 expression in invasive breast carcinoma cells (Figure 13). A mechanism whereby Snail induces P R L - 3 expression would be consistent with the known actions / effects o f these proteins in E M T .  Parker et al (2004) identified Snail as being induced in breast tumour vasculature. Surprisingly in the same study, P R L - 3 was also found to be highly induced in breast tumour vasculature (endothelium). Breast cancer is one o f the most studied cancers in relation to neovascularization and has provided a paradigm for the role o f angiogenesis i n  89  cancer (Leek, 2001). There is a lack o f molecular markers that allow accurate prediction o f responses to specific therapies, or more precisely determine whether a tumour is likely to metastasize to distant organs (Dexter et al, 1978; K l e i n et al, 2002). Enhanced angiogenesis is associated with an increased risk o f metastasis and poor prognosis in breast cancer (Linderholm et al, 2000; Olewniczak et al, 2002). Snail was found to be expressed in both endothelium and epithelium o f invasive breast cancer (Parker et al, 2004). Expression o f Snail i n tumour epithelium was previously reported (Fujita et al, 2003; Blanco et al, 2002), but its expression i n breast cancer endothelium is a novel finding. P R L - 3 expression was absent i n normal mammary epithelium and showed little expression in invasive breast epithelium, but was highly expressed in invasive breast endothelium. This is the first suggestion o f a proportional relationship between Snail and P R L - 3 expression.  A n inverse correlation o f Snail and E-cadherin expression has been reported i n many types o f human cancers, including squamous cell carcinoma. Yokoyama et al. (2003) compared three E-cadherin-negative squamous cell carcinoma cell lines with strong Snail expression to three other squamous cell carcinoma cell lines that express E-cadherin with weak Snail expression, and found that the cell lines with strong Snail expression were more invasive and had a higher ability to express matrix metalloproteinase-2 ( M M P - 2 ) . M M P s are a group o f matrix degrading enzymes which are highly expressed i n invasive cancer cells. Luciferase reporter assays revealed that Snail induces M M P - 2 promoter activity (Yokoyama et al, 2003). In addition, the M M P family is known to play a key role i n the invasion o f various human cancers including hepatocellular carcinoma  90  (Nebeshima et al, 2002). Reverse-Transcriptase P C R showed that M M P - 1 , -2, and -7 expressions were strongly upregulated by Snail (Miyoshi et al, 2004). The downregulation o f E-cadherin and upregulation o f M M P s by Snail further facilitates the E M T process. However, the precise mechanism o f how Snail promotes cancer invasion still remains unknown.  Prostate cancer is the most common malignancy detected i n men in Western countries (Gronberg, 2003; Amanatullah et al, 2000). The progression from androgen-dependence to independence (which occurs in patients treated with endocrine therapies) is associated with a higher invasive potential and malignant phenotype (Bonaccorsi et al, 2000; Cinar et al, 2001). This can be explained, at least in part, by the loss o f androgen control o f genes involved in limiting invasion (Baldi et al, 2003). Recent evidence indicates that androgen-sensitive prostate cancer cells have a less malignant phenotype characterized by reduced migration and invasion (Bonaccorsi et al, 2004). Moreover, upon E G F treatment, androgen-independent prostate cancer cells become more migratory (Bonaccorsi et al, 2004). However, transfection o f the androgen receptor ( A R ) c D N A into invasive androgen-independent cells confers a less malignant phenotype by interfering with E G F R autophosphorylation and signaling that leads to invasion in response to E G F (Bonaccorsi et al, 2004). Thus, androgens may disrupt EGF-mediated signaling pathways that lead to invasive properties. Interestingly, prominent Snail m R N A expression was evident after E G F treatment o f D U I 45 cells, the highly invasive androgen-independent metastatic prostate cancer cell line ( L u et al, 2003). A s shown in Figure 9, it is evident that the usually highly active CpG-64 island has minimal activity in  91  the androgen-dependent L N C a P cells, yet exhibits very high promoter activity in the androgen-independent D U 1 4 5 cells. If E G F stimulation does increase Snail expression, and i f Snail is involved in the positive transcriptional upregulation o f P R L - 3 , then P R L - 3 expression may be a downstream effect o f E G F signaling. Although this proposed mechanism has not yet been experimentally validated in this study, it leaves room for further work. Figure 9 indicates that the CpG-64 island induces higher luciferase activity in the cell line without A R (DU145) than i n the cell line with A R ( L N C a P ) . This suggests that the androgen receptor may be disrupting signaling pathways leading to P R L 3 transcriptional activity. This inverse relationship between A R and P R L - 3 is consistent with the proposed model as the loss o f A R and the over-expression o f P R L - 3 has been both attributed to increased invasiveness, migration, and metastatic capacity.  Saha et al. (2001) were the first to characterize P R L - 3 as being an important component o f colorectal cancer metastasis. Although we have shown that increased P R L - 3 promoter activity is evident in metastatic and invasive cell lines, the molecular actions o f P R L - 3 that can induce metastasis and invasion are still not clear. Although many gene changes have been attributed to cancer development, very few have been associated with metastasis. P R L - 3 is potentially a valuable marker to predict or detect metastasis and / or aggressive cancers.  92  4.2  POOR PROGNOSIS A N D F U N C T I O N A L L O C A L I Z A T I O N OF PRL-3  •PRL-3 expression is observed in clinical colorectal, uveal melanic, breast, and certain pediatric tumour samples. The localization o f the P R L - 3 I H C staining pattern is associated with the differentiation status ofthe cells. More specifically, P R L - 3 tends to be localized in the cell nuclei o f poorly differentiated pediatric tumours and localized in the cell cytoplasm o f well differentiated pediatric tumours (neuroblastoma and rhabdomyosarcoma). In addition, high levels o f P R L - 3 i n patient tumour samples are correlated with an unfavorable prognosis, as seen in uveal melanoma and breast carcinoma. These findings demonstrate the clinical significance o f P R L - 3 as a tumour aggressiveness / metastasis marker. However, further clinical studies are still needed to confirm the vital role o f P R L - 3 i n the metastatic process. Once the clinical role o f P R L - 3 has been validated, therapeutic development and testing can take place and in so doing, patients may then have a chance o f reducing the aggressiveness o f their disease and the risk o f relapse.  4.2.1  Colorectal C a r c i n o m a  To confirm previous observations o f P R L - 3 over-expression i n colorectal metastatic samples and as a proof-of-principle experiment testing the utility o f the P R L - 3 antibody in detecting endogenous P R L - 3 i n tumour sections, a tissue microarray ( T M A ) containing normal, primary, and matching metastatic tissues was purchased from Imgenex. Our  93  findings were consistent with previous observations, confirming that P R L - 3 is indeed over-expressed in colorectal metastases, and validating our method o f investigation.  It is interesting to note that the I H C staining o f P R L - 3 was only detected i n the cell cytoplasm from our investigation o f the colorectal cancer T M A . Virtually no nuclear I H C P R L - 3 staining was observed. Similarly, the I H C data from others (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004) also showed that P R L - 3 staining was predominantly cytoplasmic i n the colorectal cancer samples analyzed. Thus, virtually all I H C staining done by myself and others demonstrates predominant cytoplasmic localization o f P R L - 3 . The site o f action o f P R L - 3 therefore appears to be i n the cytoplasm, more specifically, at the plasma membrane, at least with respect to its role in tumour progression in metastatic colorectal cancer.  4.2.2  Uveal Melanoma  Isochromosome 8q was associated with a poor prognosis for uveal melanoma patients due to death by metastatic disease (White et al, 1998). Moreover, the gain o f 8q was significantly associated with poor overall survival (Aalto et al, 2001). The P R L - 3 gene is situated on chromosome 8q24.3, and P R L - 3 over-expression has been associated with metastatic disease (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004; M i s k a d et al, 2004; W u et al, 2004). P R L - 3 expression in uveal melanoma was explored in order to elucidate the potential o f P R L - 3 as a prognostic marker for uveal melanoma.  94  A T M A containing 184 cases was immunohistochemically stained for P R L - 3 . Weak P R L - 3 expression was detected in majority (-60%) o f cases, and only - 1 3 % had strong staining. Notably, the localization o f P R L - 3 staining was observed to vary from cytoplasmic to nuclear. Prenylation status determines P R L - 3 localization, with the prenylated phosphatase localized at the plasma membrane and the unprenylated P R L - 3 found in the nucleus (Zeng et al, 2000). This has led to the suggestion that P R L - 3 substrates and / or function may be localization-dependent. What these are, and how they relate to tumour progression and metastasis, are still unknown. A specific correlation between P R L - 3 localization and various histological parameters in this T M A has not yet been determined. However, although not statistically significant, a prognostic trend between P R L - 3 nuclear expression and clinical outcome is apparent. Cases with no or weak nuclear P R L - 3 I H C staining were grouped as 'negative' and cases with strong nuclear P R L - 3 I H C staining were termed 'positive'. A s seen i n the Kaplan-Meier survival curve for uveal melanoma (Figure 19), patients with positively staining samples have a poorer survival outcome than patients with negative staining samples. This reinforces the clinical ramifications o f P R L - 3 as being a key component o f an aggressive malignant phenotype and as a prognostic marker. To determine whether the subcellular localization o f P R L - 3 was linked to prognostic outcome, a survival curve was plotted for cytoplasmic P R L - 3 staining (data not shown). However, this revealed less statistical difference between the negative and positive cytoplasmic staining samples, although the survival linkage between positive P R L - 3 expression and poor outcome remains the same.  95  This suggests that P R L - 3 may also have a nuclear role in the progression o f certain melanomas.  P R L - 3 expression has been implicated in mouse B 1 6 melanoma invasiveness and metastatic potential ( W u et al, 2004). Heterologous expression o f P R L - 3 i n melanoma cells with low invasive properties resulted in altered extracellular matrix adhesion and up-regulated integrin-mediated cell spreading efficiency ( W u et al, 2004). This study also confirmed that P R L - 3 facilitated lung and liver metastasis o f the melanoma cells with intrinsically low metastatic ability in an experimental metastasis model i n mice, consistent with the observed accelerated proliferation ( M T T assay) and growth rate o f the cells both in vitro and in vivo. Since P R L - 3 is involved i n metastatic progression i n mouse melanoma models, and is implicated as a prognostic indicator o f poor clinical survival, then P R L - 3 may also be involved i n melanoma progression as well.  4.2.3  Breast C a r c i n o m a  There have been a number o f publications recently regarding the role o f the Snail family o f zinc-finger transcription factors in tumourigenesis. The relationship between Snail and E M T is clear, and Snail promotes the invasive capacity ofthe epithelium i n breast cancer (Parker et al, 2004). The transition from epithelial to mesenchymal cells promotes enhanced cellular migratory and invasive properties; hence such a transition may contribute to tumour progression. Since Snail has been shown to be over-expressed i n invasive breast epithelium and endothelium (Fujita et al, 2003; Blanco et al, 2002;  96  Parker et al, 2004), and P R L - 3 expression increased under conditions that permitted accumulation o f Snail (Figure 13), the expression o f P R L - 3 i n breast cancer was explored to elucidate the potential o f P R L - 3 as a prognostic marker for breast cancer.  The staining localization o f P R L - 3 was again observed to be quite variable i n the 495 cases o f breast carcinoma examined. Although the majority o f the breast carcinoma cases on the T M A scored positive for nuclear P R L - 3 expression, a significant portion exhibited strictly or concomitant cytoplasmic P R L - 3 staining. Others have reported that P R L - 3 is not typically expressed i n normal breast epithelial tissue, but when it is expressed, P R L - 3 mainly resides i n the nucleus (Miskad et al, 2004). The significance o f P R L - 3 localization in breast cancer has yet to be determined. O f the 495 breast carcinoma cases examined, over 50% o f the tissues had undetectable cytoplasmic P R L - 3 expression. About a quarter o f the cases had weak cytoplasmic P R L - 3 expression, and only a fraction (<3%) o f the cases displayed strong cytoplasmic P R L - 3 staining. Interestingly however, the patient samples that expressed high levels o f cytoplasmic P R L - 3 had a noticeable and statistically significant reduction in survival outcome (Figure 21). Patients with no or low cytoplasmic P R L - 3 expression tended to do better than those with high P R L - 3 expression. This further exemplifies the clinical relevance o f P R L - 3 as a biomarker for cancerous progression and as a prognostic indicator. Since P R L - 3 is not naturally found i n breast epithelia, the presence o f P R L - 3 in breast cancer may also indicate that P R L - 3 could be involved in tumourigenesis and / or progression o f breast carcinoma.  97  B y the same token, nuclear P R L - 3 expression was also evaluated relative to patient survival. The majority ofthe cases on this breast cancer T M A had strong nuclear P R L - 3 staining (Table 8). However, when survival outcomes o f patients with or without nuclear P R L - 3 expression were compared, there were no differences between the groups (Figure 22). This may suggest that nuclear P R L - 3 has no role i n disease progression i n breast cancer, and that cytoplasmic (prenylated) P R L - 3 does have an affect on the clinical picture o f breast cancer patients.  Her-2 is a well known receptor tyrosine kinase that is activated by a paracrine secretion mechanism and has a role in both early proliferative and late invasive / metastatic changes in breast cancer. Over-expression o f the Her-2 receptor has been associated with poor prognosis (Winters et al, 2003) in breast cancer. Heterodimerization o f Her-2 with other members o f its family o f receptor kinases have been shown to either activate the M A P kinase or PI3-kinase pathways (Nahta et al, 2004). Her-2 heterodimers are very stable, and hence can stimulate prolonged signaling. L i k e Her-2, prenylated P R L - 3 is localized at the plasma membrane and can activate various growth, angiogenic, anti-apoptotic, or migratory signaling pathways. Nonetheless, cytoplasmic P R L - 3 is associated with poor clinical survival for breast cancer patients, and signifies a functional localization o f PRL-3.  98  4.2.4  Pediatric Tumours: Neuroblastoma and Rhabdomyosarcoma  The functional localization o f the P R L s is an issue o f interest and not much is known o f their possibly regulatable localization and cytoplasmic versus nuclear substrates and functions. However, P R L - 1 has been found to be localized i n the nucleus during proliferation o f liver cells and during differentiation o f intestinal cells (Diamond et al, 1996; K o n g et al, 2000).  In general, poorly differentiated neuroblastomas and rhabdomyosarcomas are more prognostically unfavorable than the well differentiated tumours. I H C staining o f P R L - 3 in neuroblastoma and rhabdomyosarcoma tissues revealed that the less differentiated tumours have predominantly nuclear P R L - 3 , whereas the well differentiated tumours display mainly cytoplasmic P R L - 3 (Figure 23). Although the functional localization o f P R L - 3 i n pediatric tumours has yet to be determined, a relationship between tissue differentiation and P R L - 3 localization can be inferred. It is not clear whether P R L - 3 expression / localization causes differentiation or i f differentiation causes specific P R L - 3 expression / localization, but P R L - 3 localization is associated with differentiation status of the tissue. A statistical correlation between P R L - 3 I H C localization and tissue differentiation is seen in neuroblastoma (Table 11), but not in rhabdomyosarcoma (Table 12). More clinical samples need to be analyzed before verifying whether the possible trend in the observations is due to chance or a real trend.  99  It is also interesting to note that P R L - 3 and Mef-2 localization appears to have an inverse correlation. Mef-2 accumulates i n the nucleus o f well differentiated (embryonal) R M S and i n the cytoplasm o f poorly differentiated (alveolar) R M S (Chen et al, 2001). Conversely, P R L - 3 staining is localized in the cytoplasm o f well differentiated (embryonal) R M S and i n the nucleus o f poorly differentiated (alveolar) R M S . This is quite fascinating and proposes a potential role P R L - 3 may have on cellular differentiation. However, a specific relationship between P R L - 3 and Mef-2 localization has not been determined.  100  CHAPTER 5  SUMMARY  5.1  CONCLUSIONS  Metastasis is the leading cause o f death i n cancer patients and involves a complex, multistep process including detachment o f tumour cells from the primary cancer, invasion o f surrounding tissue, entry into the circulatory system, reinvasion, and proliferation at a distant secondary site (Boyd, 1996; Bashyam, 2002). The process o f metastasis is highly complex, and is the result o f accumulating genetic mutations that cause changes i n different places and times, rather than just the alteration o f a single gene. P R L - 3 may act as a key player to initiate and maintain tumour cell growth i n the distant organs by transforming the tumour cell characteristics and subsequently enhancing their metastatic ability.  Here, I demonstrated that the P R L - 3 promoter has differential transcriptional activities i n various cell lines exhibiting assorted levels o f invasiveness, metastatic state, malignancy, and cellular differentiation. More specifically, reporter assays confirm that the CpG-64 island is an important region for P R L - 3 transcriptional regulation. The CpG-64 island is more active in colorectal metastatic cells than i n primary tumour cells, more active in invasive prostate androgen-independent tumour cells than i n androgen-dependent tumour cells, and more active in malignant muscle cells than in normal muscle cells.  101  Bioinformatics analyses revealed that this CpG-64 island has multiple cancer-associated cz's-elements, high orthologue conservation, potential transcript generation, detectable regulatory potential, and high cz's-regulatory module binding probability. Other regions in the P R L - 3 promoter have some, but not all, o f these above characteristics. Together, our findings suggest that the CpG-64 island is a candidate alternative promoter region that is differentially utilized i n various stages o f cancerous development.  Investigating P R L - 3 expression in clinical tumour samples, I confirmed that P R L - 3 expression is observed i n colorectal tumours, and discovered that it is also expressed i n uveal melanic, breast, and certain pediatric tumours. The localization o f the P R L - 3 I H C staining pattern is associated with the differentiation status ofthe cells in some tumour types. P R L - 3 tended to have a nuclear localization in poorly differentiated pediatric tumours and was localized in the cytoplasm o f well differentiated pediatric tumours (neuroblastoma and rhabdomyosarcoma). Furthermore, high levels o f P R L - 3 in clinical tumour samples were correlated with an unfavorable prognosis, as seen in uveal melanoma and significantly so in breast carcinoma. These findings demonstrate the clinical relevance o f P R L - 3 in screening, monitoring, and prognosis o f cancer patients. Furthermore, they suggest that P R L - 3 could be a valid therapeutic target. Knowledge o f the substrates and functional localization o f P R L - 3 w i l l be invaluable to the development o f P R L - 3 directed inhibitors for anti-cancer treatment.  102  5.2  F U T U R E DIRECTIONS  W i t h the remaining time left in the lab (prior to September 2005), I propose to carry out the following investigations. Since specific growth factors have been shown to exacerbate cancer progression, I intend to analyze the P R L - 3 promoter plasmids in selected cancer cell lines under growth factor stimulation. IGF-1 stimulation has been shown to cause plasma membrane ripples, motile actin microspike formation, and cell separation i n non-motile lowly invasive breast cancer M C F - 7 cells (Guvakova et al, 2002). Thus, IGF-1 stimulation o f M C F - 7 cells may stimulate, through specific P R L - 3 promoter regions, the upregulated expression o f P R L - 3 that plays a role i n enhanced tumour cell invasive properties. O n another note, E G F stimulation o f D U 1 4 5 cells has been shown to upregulate Snail expression. E G F stimulation o f D U 1 4 5 cells could provide cellular conditions that allow testing o f whether Snail causes a positive transcriptional upregulation o f P R L - 3 . Moreover, IGF-1 has been shown to cause n-myc induction (Misawa et al, 2000) and increased motility (Meyer et al, 2001) in neuroblastoma cell lines. Thus, IGF-1 stimulation o f San-2 cells (neuroblastoma cell line with no n-myc gene amplification) may provide appropriate conditions to determine i f a specific region o f the P R L - 3 promoter is responsible for increased P R L - 3 expression that may be linked to neuroblastoma cell motility, and i f n-myc is responsible for this increased transcriptional activity. IGF-1 has also been shown to induce differentiation i n primitive muscle cells ( X u and W u , 2000). Therefore, IGF-1 stimulation in normal (C2C12) and malignant ( R D and R M S 13) muscle cells may shed light on the role o f P R L - 3 in muscle differentiation.  103  To determine whether the CpG-64 island is an authentic promoter of P R L - 3 , 1 w i l l use the Rapid Amplification o f c D N A Ends (5 - R A C E ) method. 5 ' - R A C E determines the coding sequence at the 5 '-end o f the transcript; hence, it identifies the first transcribed exon. O n a different note, all P R L - 3 promoter regions investigated using reporter assays, i n this study, contain non-methylated cytosines. Since cytosine-methylation is a prominent epigenetic transcriptional regulatory process, and genomic cytosines may or may not be methylated in the P R L - 3 promoter, I intend to investigate whether methylation is responsible for P R L - 3 transcriptional regulation i n tumourigenesis and metastasis by investigating the methylation status o f genomic D N A from various cancer cell lines.  In recent years, upregulated P R L - 3 expression has been correlated with metastasis and poor clinical outcome o f several adult cancers (Saha et al, 2001; Bardelli et al, 2003; Peng et al, 2004; Kato et al, 2004; M i s k a d et al, 2004; W u et al, 2004). N o similar studies have been conducted to determine i f P R L - 3 expression is a marker o f similar clinical trends i n pediatric malignancies. I w i l l construct a 168 case pediatric tumour tissue microarray consisting o f the most common types of pediatric malignancies with, where available, matching metastases. The sections from this array w i l l be I H C stained for P R L - 3 to evaluate P R L - 3 protein levels and localization. These sections w i l l also be probed for P R L - 3 m R N A levels using Chromagenic In-Situ Hybridization. The I H C and Chromagenic In-Situ Hybridization data collected on this array for P R L - 3 w i l l subsequently be correlated with clinical outcome in pediatric cases to infer prognostication.  104  REFERENCES  Aalto, Y . , L . Eriksson, S. Seregard, O. Larsson, and S. Knuutila. "Concomitant loss o f chromosome 3 and whole arm losses and gains o f chromosome 1, 6, or 8 i n metastasizing primary uveal melanoma." Investigative Ophthalmology & Visual Science, 2001, 42(2): 313-7. Amanatullah, D F , A T Reutens, B T Zafonte, M . F u , S. M a n i , and R G Pestell. "Cell-cycle dysregulation and the molecular mechanisms o f prostate cancer." Frontiers in Bioscience, 2000, 5: D372-90. Ayer, D E and R N Eisenman. " A switch from M y c : M a x to M a d : M a x heterocomplexes accompanies monocyte/macrophage differentiation." Genes and Development, 1993, 7: 2110-9. Ayer, D E , L . Kretzner, and R N Eisenman. " M a d : a heterodimer partner for M a x that antagonizes M y c transcriptional activity." C e l l , 1993, 72: 211-22. Baldi, E . , L . Bonaccorsi, and G . Forti. "Androgen receptor: good guy or bad guy in prostate cancer invasion?" Endocrinology, 2003, 144: 1653-5. Bardelli, A . , S. Saha, J A Sager, K E Romans, B . X i n , S D Markowitz, C . Lengauer, V E Velculescu, K W Kinzler, and B . Vogelstein. " P R L - 3 Expression in Metastatic Cancers." Clinical Cancer Research, 2003, 9: 5607-15. Bashyam, M D . "Understanding cancer metastasis: an urgent need for using differential gene expression analysis." Cancer, 2002, 94: 1821-29. Batlle, E . , E . Sancho, C . Franci, D . Dominguez, M . Monfar, J. Baulida, and A . Garcia De Herreros. "The transcription factor Snail is a repressor o f E-cadherin gene expression in epithelial tumour cells." Nature C e l l Biology. 2000, 2(2): 84-9. Behrens, J., W . Lowrick, L . Klein-Hitpass, and W . Birchmeier. "The E-cadherin promoter: functional analysis o f a G C - r i c h region and an epithelial cell-specific palindromic regulatory element." Proceedings o f the National Academy o f Sciences o f the United States o f America, 1991, 88(24): 11495-9. Birchmeier, W . , J. Behrens, K M Weidner, U H Frixen, and J. Schipper. "Dominant and recessive genes involved in tumour cell invasion." Current Opinion i n C e l l Biology, 1991,3(5): 832-40. Black, B L and E N Olson. "Transcriptional control o f muscle development by myocyte enhancer factor-2 ( M E F 2 ) proteins." Annual Review o f C e l l & Developmental Biology, 1998, 14: 167-96.  105  Blake, M C , R C Jambou, A G Swick, J W Kahn, and J C Azizkhan. "Transcriptional initiation is controlled by upstream G C - b o x interactions i n a T A T A - l e s s promoter." Molecular C e l l Biology, 1990, 10: 6632-41. Blanco, M J , G . Moreno-Bueno, D . Sarrio, A . Locascio, A . Cano, J. Palacios, and M A Nieto. "Correlation o f Snail expression with histological grade and lymph node status in breast carcinomas." Oncogene, 2002, 21: 3241-6. Blume-Jensen, P. and T. Hunter. "Oncogenic kinase signaling." Nature, 2001, 411: 35565. Bonaccorsi, L . , V . Carloni, M . Muratori, A . Salvadori, A . Giannini, M . Carini, M . Serio, G . Forti, and E . Baldi. "Androgen receptor expression i n prostate cancinoma cells suppresses alpha6beta4 integrin-mediated invasive phenotype." Endocrinology, 2000, 141:3172-82. Bonaccorsi, L . , S. Marchiani, M . Muratori, V . Carloni, G . Forti, and E . Baldi. "Signaling Mechanisms That Mediate Invasion i n Prostate Cancer Cells." Annals o f N e w Y o r k Academy o f Sciences, 2004, 1028: 283-8. Boyd, D . "Invasion and metastasis." Cancer & Metastasis Reviews, 1996, 15: 77-89. Brown-Shimer, S., K A Johnson, D E H i l l , and A M Bruskin. "Effect o f protein tyrosine phosphatase I B expression on transformation by the human neu oncogene." Cancer Research, 1992, 52: 478-82. Cano, A . , M A Perez-Moreno, I. Rodrigo, A . Locascio, M J Blanco, M G del Barrio, F. Portillo, and M A Nieto. "The transcription factor Snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression." Nature C e l l Biology, 2000, 2(2): 76-83. Carter, D A . "Expression o f a novel rat protein tyrosine phosphatase gene." Biochimica et Biophvsica Acta, 1998, 1442(2-3):405-8. Casey, P J , P A Solski, C J Der, and J E Buss. "p21Ras is modified by a farnesyl isoprenoid." Proceedings o f the National Academy o f Sciences o f the United States o f America, 1989, 86: 8323-7. Cates, C A , R L Michael, K R Stayrook, K A Harvey, Y D Burke, S K Randall, P L Crowell, and D N Crowell. "Prenylation o f oncogenic human P T P ( C A A X ) protein tyrosine phosphatases." Cancer Letters. 1996, 110(1-2): 49-55. Chambers, A F , A C Groom, and I C MacDonald. "Dissemination and growth o f cancer cells in metastatic sites." Nature Reviews Cancer, 2002, 2(8): 563-72.  106  Chen, S L , M . Wang, B . Hosking, and G . Muscat. "Subcellular Localization ofthe Steroid Receptor Coactivators (SRCs) and M E F 2 i n M u s c l e and Rhabdomyosarcoma Cells." Molecular Endocrinology, 2001, 15(5): 783-96. Cheng, C W , P E W u , J C Y u , C S Huang, C T Y u e , C W W u , and C Y Shen. "Mechanisms of inactivation o f E-cadherin in breast carcinoma: Modification o f the two-hit hypothesis of tumour suppressor gene." Oncogene, 2001, 20: 3814-23. Cinar, B . , K S Koeneman, M . Edlund, G S Prins, H E Zhau, and L W Chung. "Androgen receptor mediates the reduced tumour growth, enhanced androgen responsiveness, and selected target gene transactivation in a human prostate cancer cell line." Cancer Research, 2001, 61(19): 7310-7. Cole, M D and S B M c M a h o n . "The M y c oncoprotein: a critical evaluation o f transactivation and target gene regulation." Oncogene, 1999, 18: 2916-24. Damjanov, I. Pathology, for the Health-Related Professions, 2  nd  edition. W B Saunders  Company, U S A , 2000. Davidson, E H . Genomic regulatory systems: development and evolution. Academic  Press, San Diego, 2001. Dennis, J L , J K Vass, E C W i t , W N Keith, and K A Oien. "Identification from public data o f molecular markers o f adenocarcinoma characteristic o f the site o f origin." Cancer Research, 2002, 62: 5999-6005. Dexter, D L , H M K o w a l s k i , B A Blazar, Z . Fligiel, R. Vogel, and G H Heppner. "Heterogeneity o f tumour cells from a single mouse mammory tumour." Cancer Research, 1978,38:3174-81. D i Cristofano, A . and P P Pandolfi. "The multiple roles o f P T E N i n tumour suppression." C e l l , 2000, 100: 387-90. Diamond, R H , D E Cressman, T M Laz, C S Abrams, and R. Taub. " P R L - 1 , a unique nuclear protein tyrosine phosphatase, affects cell growth." Molecular & Cellular Biology, 1994, 14(6): 3752-62. Diamond, R H , C . Peters, SP Jung, L E Greenbaum, B A Haber, D G Silberg, P G Traber, and R . Taub. "Expression o f P R L - 1 nuclear PTPase is associated with proliferation i n liver but with differentiation in intestine." American Journal o f Physiology, 1996,271(1 Pt 1): G121-9. Dias, P., D M Parham, D N Shapiro, SJ Tapscot, and P J Houghton. "Monoclonal antibodies to the myogenic regulatory protein M y o D l : Epitope mapping and diagnostic utility." Cancer Research. 1992, 52: 6431-9.  107  Egeblad, M . and Z . Werb. " N e w functions for the matrix metallo-proteinases in cancer progression." Nature Reviews Cancer, 2002, 2: 161-74. Ellenrieder, V . , G . Adler, and T M Gress. "Invasion and metastasis in pancreatic cancer." Annals o f Oncology. 1999, 10 Suppl 4: 46-50. Elnitski, L . , R C Hardison, J. L i , S. Yang, D . Kolbe, P. Eswara, M J O'Connor, S. Schwartz, W . M i l l e r , and F. Chiaromonte. "Distinguishing Regulatory D N A From Neutral Sites." Genome Research. 2003, 13: 64-72. Fidler, IJ. "Seed and soil revisited: contribution ofthe organ microenvironment to cancer metastasis." Surgical Oncology Clinics o f North America. 2001, 10(2): 257-69, v i i - v i i i i . Fischer, E H , H . Charbonneau, and N K Tonks. "Protein Tyrosine Phosphatases: A Diverse Family o f Intracellular and Transmembrane Enzymes." Science, 1991, 253(5018): 401-6. Fossa, S D . "Management o f hormone resistant prostate cancer." A c t a Urologica Belgica, 1994, 62(1): 73-6. Frazer, K A , L . Elnitski, D M Church, I. Dubchak, and R C Hardison. "Cross-species sequence comparisons: a review o f methods and available resources." Genome Research, 2003, 13: 1-12. Frith, M C , U . Hansen, and Z . Weng. "Detection o f cw-element clusters in higher eukaryotic D N A . " Bioinformatics, 2001, 17(10): 878-89. Fujita, N . , D L Jaye, M . Kajita, C . Geigerman, C S Moreno, and P A Wade. " M T A 3 , a M i 2 / N u R D complex subunit, regulates an invasive growth pathway i n breast cancer." C e l l , 2003, 113: 207-19. Gardiner-Garden, M . and M . Frommer. " C p G islands in vertebrate genomes." Journal o f Molecular Biology, 1987, 196(2): 261-82. Ghosh, A . , X . Wang, E . K l e i n , and W D Heston. "Novel role o f prostate-specific membrane antigen in suppressing prostate cancer invasiveness." Cancer Research, 2005, 65(3): 727-31. Giordano, T J , K A Shedden, D R Schwartz, R . K u i c k , J M Taylor, N . Lee, D E Misek, J K Greenson, S L Kardia, D G Beer, G . Rennert, K R Cho, S B Gruber, E R Fearon, and S. Hanash. "Organ-specific molecular classification o f primary lung, colon, and ovarian adenocarcinomas using gene expression profiles." American Journal o f Pathology, 2001, 159: 1231-8. -  108  Giroldi, L A , P P Bringuier, M . de Weijert, C . Jansen, A . van Bokhoven, and J A Schalken. "Role o f E-boxes in the repression o f E-cadherin expression." Biochemical & Biophysical Research Communications, 1997, 241(2): 453-8. Glomset, J A , M H Gelb, and C C Farnsworth. "Prenyl proteins i n eukaryotic cells: a new type o f membrane anchor." Trends in Biochemical Science, 1990, 15(4): 139-42. Grandori, C , S M Cowley, L P James, and R N Eisenman. "The M y c / M a x / M a d network and the transcriptional control o f cell behavior." [Review] Annual Review o f C e l l & Developmental Biology. 2000, 16: 653-99. Griffin, C A , P P Long, and A P Schachat. "Trisomy 6p i n an Ocular Melanoma." Cancer Genetics & Cytogenetics. 1998, 32(1): 129-32. Gronberg, H . "Prostate cancer epidemiology." Lancet, 2003, 361: 859-64. Guaita, S., I. Puig, C . Franci, M . Garrido, D . Dominguez, E . Batlle, E . Sancho, S. Dedhar, A G De Herreros, and J. Baulida. "Snail induction o f epithelial-to-mesenchymal transition i n tumour cells is accompanied by MTJC1 repression and Z E B 1 expression." Journal o f Biological Chemistry. 2002, 277(42): 39209-16. Gumucio, D L , H . Heilstedt-Williamson, T A Gary, S A Tarle, D A Shelton, D . Tagle, J. Slightom, M . Goodman, and F S Collins. "Phylogenetic footprinting reveals a nuclear protein which binds to silencer sequences in the human y and s globin genes." Molecular C e l l Biology, 1992, 12: 4919-29. Guvakova, M A , J C Adams, and D . Boettiger. "Functional role o f a-actinin, PI3-kinase and M E K 1 / 2 i n insulin-like growth factor I receptor kinase regulated motility o f human breast carcinoma cells." Journal o f C e l l Science, 2002, 115: 4149-65. Haber, M . , S B Bordow, J. Gilbert, J. Madafiglio, M . Kavallaris, G M Marshall, E B Mechetner, JP Fruehaufi L . Tee, S L Conn, H . Salwen, M L Schmidt, and M D Norris. "Altered expression o f the M Y C N oncogene modulates M R P gene expression and response to cytotoxic drug i n neuroblastoma cells." Oncogene, 1999, 18: 2777-82. Hajra, K M , X . J i , and E R Fearon. "Extinction o f E-cadherin expression i n breast cancer via a dominant repression pathway acting on proximal promoter elements." Oncogene, 1999, 18: 7274-9. Hancock, JF, A l Magee, J E Childs, and C J Marshall. " A l l Ras proteins are polyisoprenylated but only some are palmitoylated." C e l l , 1989,57: 1167-77. Hannenhalli, S. and S. Levy. "Promoter prediction in the human genome." Bioinformatics. 2001, 17(Suppl.l): S90-6.  109  Hardison, R C , J L Slightom, D L Gumucio, G . Goodman, N . Stojanovic, and W . Miller. "Locus control regions o f mammalian (3-globin gene clusters: Combining phylogenetic analyses and experimental results to gain functional insights." Gene, 1997b, 205: 73-94. Horsman, D E and V A White. "Cytogenetic Analysis o f Uveal Melanoma." Cancer, 1993,71(3): 811-819. Hsu, F D , T O Nielsen, A . Alkushi, B . Dupuis, D . Huntsman, C L L i u , M v d Rijn, and C B Gilks. "Tissue Microarrays A r e an Effective Quality Assurance T o o l for Diagnostic Immunohistochemistry." Modern Pathology, 2002, 15(12): 1374-1380. Hunter, T. and B M Sefton. "Transforming gene product o f Rous sarcoma virus phosphorylates tyrosine." Proceedings o f the National Academy o f Sciences o f the United States o f America, 1980, 77: 1311-5. Hunter, T. "Protein kinases and phosphatases: the yin and yang o f protein phosphorylation and signaling." Cell, 1995, 80(2): 225-36. Hunter, T. "Signaling-2000 and beyond." C e l l 2000, 100: 113-27. Hurlin, P J , K P Foley, D E Ayer, R N Eisenman, D . Hanahan, and J M Arbeit. "Regulation o f M y c and M a d during epidermal differentiation and HPV-associated tumourigenesis." Oncogene. 1995a, 11: 2487-501. Hurlin, P J , C . Queva, P J Koskinen, E . Steingrimsson, D E Ayer, N G Copeland, N A Jenkins, and R N Eisenman. "Mad3 and Mad4: novel Max-interacting transcriptional repressors that suppress c-myc-dependent transformation and are expressed during neural and epidermal differentiation." E M B O Journal, 1995b. 14(22): 5646-59. Jackson, J., J. Meisinger, S. Patel, Z C L i m , K . Vellody, R. Metz, and M R Young. "Protein phosphatase-2A associates with the cytoskeleton to maintain cell spreading and reduced motility o f nonmetastatic Lewis lung carcinoma cells: the loss o f this regulatory control i n metastatic cells." Invasion Metastasis, 1997, 17: 199-209. Ji, X . , A S Woodard, D L R i m m , and E R Fearon. "Transcriptional defects underlie loss o f E-cadherin expression i n breast cancer." C e l l Growth and Differentiation, 1997, 8: 7738. Jiang, W G . "E-cadherin and its associated protein catenins, cancer invasion and metastasis." British Journal o f Surgery. 1996, 83: 437-46. Jiao, W . , K . M i y a z a k i , and Y . Kitajima. "Exogenous expression o f E-cadherin in gallbladder carcinoma cell line G-415 restores its cellular polarity and differentiation." International Journal o f Oncology. 2001. 19: 1099-107.  110  Karolchik, D . , R . Baertsch, M . Diekhans, TS Furey, A . Hinrichs, Y T L u , K M Roskin, M . Schwartz, C W Sugnet, D J Thomas, R J Weber, D . Haussler, and W J Kent. "The U C S C Genome Browser Database." Nucleic Acids Research, 2003, 31(1): 51-4. Kassis, J., D A Lauffenburger, T. Turner, and A . Wells. "Tumour invasion as dysregulated cell motility." Seminars i n Cancer Biology, 2001, 11(2): 105-17. Kato, H . , S. Semba, U A Miskad, Y . Seo, M . Kasuga, and H . Y o k o z a k i . " H i g h Expression o f P R L - 3 Promotes Cancer C e l l Motility and Liver Metastasis i n Human Colorectal Cancer: A Predictive Molecular Marker o f Metachronous Liver and Lung Metastases." Clinical Cancer Research, 2004, 10: 7318-28. K i m , M . and W L Carroll. "Autoregulation o f the N-myc Gene Is operative i n Neuroblastoma and Involves Histone Deacetylase 2." Cancer, 2004, 101(9): 2106-15. Kinzler, K W and B . Vogelstein. The Genetic Basis of Human Cancer. M c G r a w - H i l l  Press, Toronto, 1998. K l e i n , C A , T J Blankenstein, O. Schmidt-Kittler, M . Petronio, B . Polzer, N H Stoecklein, and G . Riethmuller. "Genetic heterogeneity o f single disseminated tumour cells i n minimal residual cancer." Lancet, 2002, 360: 683-9. Kolbe, D . , J. Taylor, L . Elnitski, P. Eswara, J. L i , W . M i l l e r , R. Hardison, and F. Chiaromonte. "Regulatory Potential Scores From Genome-Wide Three-Way Alignments o f Human, Mouse, and Rat." Genome Research, 2004, 14: 700-707. K o n g , W H , G P Swain, S X L i , and R H Diamond. " P R L - 1 PTPase expression is developmentally regulated with tissue-specific patterns i n epithelial tissues." American Journal o f Physiology - Gastrointestinal & Liver Physiology. 2000, 279(3): G613-21. Kouraklis, G . , T J Triche, R. Wesley, and M . Tsokos. " M y c oncogene expression and nude mouse tumourigenicity and metastasis formation are higher in alveolar than embryonal rhabdomyosarcoma cell lines." Pediatric Research, 1999, 45(4 Pt 1): 552-8. Leek, R D . "The prognostic role o f angiogenesis i n breast cancer." Anticancer Research, 2001,21:4325-31. Lenhard, B . , A . Sandelin, L . Mendoza, P. Engstrom, N . Jareborg, and W W Wasserman. "Identification o f conserved regulatory elements by comparative genome analysis." Journal o f Biology, 2003. 2(21:13. Levinson, A D , H . Opperman, H E Varmus, and J M Bishop. "The purified product o f the transforming gene o f avian sarcoma virus phosphorylates tyrosine." Journal o f Biological Chemistry, 1980, 255(24): 11973-80.  111  Linderholm, B . , B . Lindh, B . Tavelin, K . Grankvist, and R. Henriksson. "p53 and vascular endothelial-growth-factor ( V E G F ) expression predicts outcome i n 833 patients with primary breast carcinoma." International Journal o f Cancer, 2000, 89: 51-62. L u , J., W . Lee, C . Jiang, and E B Keller. "Start site selection by S p l in the T A T A - l e s s human Ha-ras promoter." Journal o f Biological Chemistry, 1994,269(7): 5391-402. L u , Z . , S. Ghosh, Z . Wang, and T. Hunter." Downregulation o f caveolin-1 function by E G F leads to the loss o f E-cadherin, increased transcriptional activity o f (3-catenin, and enhanced tumour cell invasion." Cancer C e l l , 2003, 4: 499-515. Lutz, W . , M . Stohr, J. Schurmann, A . Wenzel, A . Lohr, and M . Schwab. "Conditional expression o f N - m y c i n human neuroblastoma cells increases expression o f aprothymosin and ornithine decarboxylase and accelerates progression into S-phase early after mitogenic stimulation o f quiescent cells." Oncogene, 1996, 13: 803-12. Mancini, A . , A . K o c h , R. W i l m s , and T. Tamura. "The SH2-containing inositol 5phosphatase (SHIP)-l is implicated in the control o f cell-cell junction and induces dissocitation and dispersion o f M D C K cells." Oncogene, 2002, 21: 1477-84. Matter, W F , T. Estridge, C . Zhang, R. Belagaje, L . Stancato, J. D i x o n , B . Johnson, L . Bloem, T. Pickard, M . Donaghue, S. Acton, R. Jeyaseelan, V . Kadambi, and C J Vlahos. "Role o f P R L - 3 , a human muscle-specific tyrosine phosphatase, in angiotensin-]! signaling." Biochemical & Biophysical Research Communications, 2001, 283(5): 10618. Marshall, C J . "Protein Prenylation: A Mediator o f Protein-Protein Interactions." Science, 1993,259: 1865-6 Maurer, H M , T. M o o n , M . Donaldson, C . Fernandez, E A Gehan, D . Hammond, D M Hays, W . Lawrence Jr., W . Newton, A . Ragab, B . Raney, E H Soule, W W Sutow, and M . Tefft. "The intergroup rhabdomyosarcoma study: a preliminary report." Cancer, 1977, 40: 2015-2026. Meyer, G E , E . Shelden, B . K i m , and E L Feldman. "Insulin-like growth factor I stimulates motility in human neuroblastoma cells." Oncogene, 2001, 20: 7542-50. M i l l e r , R W , J L Y o u n g Jr., and B . Novakovic. "Childhood Cancer." Cancer. 1995, 75: 395-405. Misawa, A . , H . Hosoi, A . Arimoto, T. Shikata, S. A k i o k a , T. Matsumura, P J Houghton, and T. Sawada. "N-myc induction stimulated by Insulin-like Growth Factor I through Mitogen-activated Protein Kinase Signaling Pathway in Human Neuroblastoma Cells." Cancer Research. 2000, 60: 64-9.  112 M i s k a d , U A , S. Semba, H . Kato, and H . Y o k o z a k i . "Expression o f P R L - 3 Phosphatase in Human Gastric Carcinomas: Close Correlation with Invasion and Metastasis." Pathobiology, 2004, 71: 176-84. M i y o s h i , A . , Y . Kitajima, K . Sumi, K . Sato, A . Hagiwara, Y . Koga, and K . Miyazaki. "Snail and SIP1 increase cancer invasion by upregulating M M P family i n heptocellular carcinoma cells." British Journal o f Cancer, 2004, 90: 1265-73. M o h n , K L , T M L a z , J C Hsu, A E Melby, R. Bravo, and R . Taub. "The immediate-early growth response i n regenerating liver and insulin-stimulated H-35 cells: comparison with serum-stimulated 3T3 cells and identification o f 41 novel immediate-early genes." Molecular & Cellular Biology, 1991, 11(1): 381-90. Molkentin, J D , B L Black, JF Martin, and E N Olson. "Cooperative activation o f muscle gene expression by M E F 2 and myogenic b H L H proteins." C e l l , 1995, 83: 1125-36. Montagna, M . , O. Serova, B S Sylla, M G Mattei, and G M Lenoir. "Localization o f the human phosphotyrosine phosphatase-related genes (h-PRL-1) to chromosome bands Ip35-p34, 17ql2-q21, Ilq24-q25 and 12q24." Human Genetics, 1996, 98: 738-40. M u k a i , S., K . M i y a z a k i , and H . Yakushiji. "The role o f E-cadherin i n the differentiation o f gallbladder cancer cells." C e l l Tissue Research, 2001, 306: 117-28. M u r a , C , G L Gac, S. Jacolot, and C . Ferec. "Transcriptional regulation o f the human H F E gene indicates high liver expression and erythropoiesis coregulation." F A S E B Journal, 2004, 18(15): 1922-4. Nabeshima, K . , T. Inoue, Y . Shimano, and T. Sameshima. "Matrix metalloproteinases in tumour invasion: Role for cell migration." Pathology International, 2002, 52: 255-64. Nahta, R., M C Hung, and F J Esteva. "The HER-2-targeting antibodies trastuzumab and pertuzumab synergistically inhibit the survival o f breast cancer cells." Cancer Research, 2004, 64(7): 2343-6. Naidu, P S , D C Ludolph, R Q To, T J Hinterberger, and SF Konieczny. "Myogenin and M E F 2 function synergistically to activate the M R F 4 promoter during myogenesis." Molecular C e l l Biology. 1995, 15: 2707-18. Neel B G . "Role o f phosphatases in lymphocyte activation." Current Opinion i n Immunology, 1997, 9(3): 405-20. Neel, B G and N K Tonks. "Protein tyrosine phosphatase i n signal transduction." Current Opinion i n C e l l Biology. 1997, 9: 193-204.  113  Newling, D W . "The management o f hormone refractory prostate cancer." European Urology, 1996, 29 Suppl 2: 69-74. Nieto, M A . "The Snail superfamily o f zinc-finger transcription factors." Nature Reviews Molecular C e l l Biology, 2002, 3: 155-66. Norris, M D , S B Bordow, G M Marshall, PS Haber, S L Cohn, and M . Haber. "Expression ofthe gene for multidrug-resistence-associated protein and outcome i n patients with neuroblastoma." N e w England Journal o f Medicine, 1996, 334: 231-8. Olewniczak, S., M . Chosia, A . Kwas, A . K r a m , and W . Domagala. "Angiogenesis and some prognostic parameters o f invasive ductal breast carcinoma i n women." Polish Journal o f Pathology. 2002, 53(4): 183-8. Oliver, C J , R T Terry-Lorenzo, E . Elliott, W A Bloomer, S. L i , D L Brautigan, R J Colbran, and S. Shenolikar. "Targeting protein phosphatase 1 (PP1) to the actin cytoskeleton: the neurabin I/PP1 complex regulates cell morphology." Molecular C e l l Biology, 2002, 22: 4690-701. Osterlind, A . "Trends i n incidence o f ocular malignancy melanoma i n Denmark 19431982." International Journal o f Cancer. 1987,40: 161. Pappo, A S , D N Shapiro, W M Crist, and H M Maurer. "Biology and therapy o f pediatric rhabdomyosarcoma." Journal o f Clinical Oncology, 1995, 13: 2123-39. Pappo, A S and D N Shapiro. "Rhabdomyosarcoma: biology and therapy." Cancer Treatment Research, 1997, 92: 309-39. Parker, B S , P. Argani, B P Cook, H . Liangfeng, S D Chartrand, M . Zhang, S. Saha, A . Bardelli, Y . Jiang, T B St. Martin, M . Nacht, B A Teicher, K W Klinger, S. Sukumar, and S L Madden. "Alterations in Vascular Gene Expression i n Invasive Breast Carcinoma." Cancer Research, 2004, 64: 7857-66. Pathak, M K , D . Dhawan, D J Lindner, E C Borden, C . Farver, and T. Y i . "Pentamidine is an Inhibitor o f P R L Phosphatases with Anticancer Activity." Molecular Cancer Therapeutics, 2002, 1: 1255-64. Peng, L R , J Y N i n g , L . Meng, and C C Shou. "The association ofthe expression level o f protein tyrosine phosphatase P R L - 3 protein with liver metastasis and prognosis o f patients with colorectal cancer." Journal o f Cancer Research and Clinical Oncology, 2004, 130(9): 521-6. Ponniah, S., D Z Wang, K L L i m , and C J Pallen. "Targeted disruption o f the tyrosine phosphatase P T P a leads to constitutive downregulation o f the kinases Src and F y n . " Current Biology. 1999, 9: 535-8.  114  Portera, C A Jr., R S Berman, and L M Ellis. "Molecular determinants o f colon cancer metastasis." Surgical Oncology, 1998, 7(3-4): 183-95. Price, L S and J G Collard. "Regulation o f the cytoskeleton by Rho-family GTPases: implications for tumour cell invasion." Seminars i n Cancer Biology, 2001, 11(2): 16773. Pugh, B F and R. Tjian. "Mechanism o f transcriptional activation by S p l : evidence for coactivators." C e l l , 1990, 61: 1187-97. Puri, P L , Z . W u , P. Zhang, L D Wood, K S Bhakta, J. Han, J R Feramisco, M . Karin, and J. Wang. "Induction o f terminal differentiation by constitutive activation o f p38 M A P kinase in human rhabdomyosarcoma cells." Genes and Development, 2000, 14: 574-84. Ridley, A . "Molecular switches i n metastasis." Nature, 2000, 406: 466-7. Roach P J . "Multisite and hierarchal protein phosphorylation." Journal o f Biological Chemistry, 1991, 266(22): 14139-42. Rosivatz, E . , I. Becker, K . Specht, E . Fricke, B . Luber, R. Busch, H . Hofler, and K F Becker. "Differential expression ofthe epithelial-mesenchymal transition regulators Snail, S I P l and Twist in gastric cancer." America Journal o f Pathology, 2002, 161: 1881-91. Saha, S., A . Bardelli, P. Buckhaults, V E Velculescu, C . Rago, B . St. Croix, K E Romans, M A Choti, C . Lengauer, K W Kinzler, and B . Vogelstein. " A Phosphatase Associated with Metastasis o f Colorectal Cancer." Science, 2001, 294: 1343-6. Schafer, W R , R. K i m , R. Sterne, J. Thorner, S H K i m , and J. Rine. "Genetic and Pharmacological Suppression o f Oncogenic Mutations in Ras Genes o f Yeast and Humans." Science, 1989, 245: 379-85. Schweigerer, L . , S. Breit, A . Wenzel, K . Tsunamoto, R. Ludwig, and W . Schwab. "Augmented M Y C N expression advances the malignant phenotype o f human neuroblastoma cells: evidence for induction o f autocrine growth factor activity." Cancer Research. 1990, 50: 4411-6. Seabra, M C . "Membrane association and targeting o f prenylated Ras-like GTPases." (review) Cellular Signalling. 1998, 10(3): 167-72. Seeger, R C , G M Brodeur, H . Sather, A . Daltofi, S E Siegel, K Y Wong, and D . Hammond. "Association o f multiple copies o f N-myc oncogene with rapid progression o f neuroblastomas." N e w England Journal o f Medicine, 1985,334: 1111-6.  115  Sefton, B M , T. Hunter, E H B a l l , and SJ Singer. " V i n c u l i n : a cytoskeletal target o f the transforming protein o f Rous sarcoma virus." C e l l , 1981,24: 165-74. S i , X . , Q. Zeng, C H N g , W . Hong, and C J Pallen. "Interaction o f farnesylated P R L - 2 , a protein-tyro sine phosphatase, with the beta-subunit o f geranylgeranyltransferase JJ." Journal o f Biological Chemistry, 2001, 276(35): 32875-82. Sinensky, M . "Functional aspects o f polyisoprenoid protein substituents: roles in proteinprotein interaction and trafficking." Biochimica et Biophysica Acta, 2000, 1529(1-3): 203-9. Sisley, K . , I G Rennie, D W Cottam, A M Potter, C W Potter, and R C Rees. "Cytogenetic findings in six posterior uveal melanomas: involvement o f chromosomes 3, 6, and 8." Genes, Chromosomes and Cancer, 1990, 2: 205-9. Stearns, M E and T. McGarvey. "Prostate cancer: therapeutic, diagnostic, and basic studies." Laboratory Investigation, 1992, 67(5): 540-52. Streuli, M . "Protein tyrosine phosphatases i n signaling." Current Opinion in C e l l Biology, 1996, 8(2): 182-8. Su, J., M . Muranjian, and J. Sap. "Receptor protein tyrosine phosphatase alpha activates Src-family kinases and controls integrin-mediated responses in fibroblasts." Current Biology, 1999, 9(10): 505-11. Sugimachi, K . , S. Tanaka, T. Kameyama, K . Taguchi, S. Aishima, M . Shimada, K . Sugimachi, and M . Tsuneyoshi. "Transcriptional repressor Snail and progression o f human hepatocellular carcinoma." Clinical Cancer Research, 2003,130: 4057-71. Sun, H . and N K Tonks. "The coordinated action o f protein tyrosine phosphatases and kinases in cell signaling." Trends in Biochemical Science. 1994, 19(11): 480-5. Susan, A . and B U Schumacher. Metastasis Research Protocols, vol. L Humana Press, N e w Jersey, 2001. Tan, Y H . " Y i n and yang o f phosphorylation in cytokine signaling." Science, 1993, 262: 376-7. Tanaka, M . , Y . Kitajima, G . Edakuni, S. Sato, and K . Miyazaki. "Abnormal expression o f E-cadherin and p-catenin may be a molecular marker o f submucosal invasion and lymph node metastasis in early gastric cancer." British Journal o f Surgery, 2002, 89: 23644. Tapscott, SJ, M J Thayer, and H . Weintraub. "Deficiency in Rhabdomyosarcoma o f a factor required for M y o D activity and myogenesis." Science, 1993, 259: 1450-3.  116  Thiery, JP. "Epithelial-mesenchymal transition in tumour progression." Nature Reviews Cancer, 2002, 2: 442-54. Tonks, N K and B G Neel. "From form to function: signaling by protein tyrosine phosphatases." CeU, 1996, 87: 365-8. Ureta-Vidal, A . , L . Ettwiller, and E . Birney. "Comparative genomics: genome-wide analysis in metazoan eukaryotes." Nature Reviews Genetics, 2003, 4(4): 251-62. Varedi, M . , A . Ghahary, P G Scott, and E E Tredget. "Cytoskeleton regulates expression o f genes for transforming growth factor-beta 1 and extracellular matrix proteins in dermal fibroblasts." Journal o f C e l l Physiology, 1997, 172: 192-9. Vega, S., A V Morales, O H Ocana, F. Valdes, I. Fabregat, and M A Nieto. "Snail blocks the cell cycle and confers resistance to cell death." Genes and Development, 2004, 18: 1131-43. Velculescu, V E , S L Madden, L . Zhang, A E Lash, J. Y u , C . Rago, A . L a i , 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, et al. "Analysis o f human transcriptomes." Nature Genetics, 1999, 23: 387-8. Wang, Q., D I R Holmes, S M Powell, Q L L u , and J. Waxman. "Analysis o f stromalepithelial interactions i n prostate cancer identifies P T P C A A X 2 as a potential oncogene." Cancer Letters. 2002, 175: 63-9. Wasserman, W W and A . Sandelin. "Applied Bioinformatics for the Identification o f Regulatory Elements." Nature Reviews Genetics. 2004, 5: 276-287. Weiss, W A , K . Aldage, G . Mohapatra, B G Feuerstein, and J M Bishop. "Targeted expression o f M Y C N causes neuroblastoma in transgenic mice." E M B O Journal, 1997, 16(11): 2985-95. Weiss, L . "Metastasis o f cancer: a conceptual history from antiquity to the 1990s." Cancer & Metastasis Reviews. 2000, 19(3-4): 193-383. White, V A , J D Chambers, P D Courtright, W Y Chang, and D E Horsman. "Correlation o f Cytogenetic Abnormalities with the Outcome o f Patients with Uveal Melanoma." American Cancer Society. 1998, 83(2): 354-9. Winters, Z E , R D Leek, M J Bradburn, C J Norbury, and A L Harris. "Cytoplasmic p21 W A F 1 / C I P 1 expression is correlated with H E R - 2 / neu in breast cancer and is an independent predictor o f prognosis." Breast Cancer Research. 2003, 5(6): R242-9.  117  W u , X P , H . Zeng, X M Zhang, Y . Zhao, H B Sha, X M Ge, M Y Zhang, X . Gao, and Q. X u . "Phosphatase o f Regenerating Liver-3 Promotes Motility and Metastasis o f Mouse Melanoma Cells." American Journal o f Pathology, 2004 164(6): 2039-54. X u , Q. and Z . W u . "The Insulin-Growth Factor-Phosphatidylinositol 3-Kinase-Akt Signaling Pathway Regulates Myogenin Expression in Normal Myogenic Cells but Not in Rhabdomyosarcoma-derived R D cells." Journal o f Biological Chemistry, 2000, 275(47): 36750-7. Yokoyama, K . , N . Kamata, R. Fujimoto, S. Tsutsumi, M . Tomonari, M . Taki, H . Hosokawa, and M . Nagayama. "Increased invasion and matrix metalloproteinase-2 expression by Snail-induced mesenchymal transition i n squamous cell carcinoma." International Journal o f Oncology, 2003, 22(4): 891-8. Zeng, Q., W . Hong, and Y H Tan. "Mouse P R L - 2 and P R L - 3 , two potentially prenylated protein tyrosine phosphatases homologous to P R L - 1 . " Biochemical & Biophysical Research Communications. 1998, 244(2): 421-7. Zeng, Q, X . S i , H . Horstmann, Y . X u , W . Hong, and C J Pallen. "Prenylation-dependent association o f protein tyrosine phosphatase P R L - 1 , -2, and -3 with the plasma membrane and the early endosome." Journal o f Biological Chemistry. 2000, 275(28): 21444-52. Zeng, Q., J M Dong, K . Guo, J. L i , H X Tan, V . K o h , C J Pallen, E . Manser, and W . Hong. " P R L - 3 and P R L - 1 promote cell migration, invasion, and metastasis." Cancer Research, 2003, 63: 2716-22. Zervos, A S , J. Gyuris, and R. Brent. " M x i l , a protein that specifically intereacts with M a x to bind M y c - M a x recognition sites." C e l l , 1993, 72: 223-32. Zhang, F L and P J Casey. "Protein prenylation: molecular mechanisms and functional consequences." Annual Review o f Biochemistry, 1996, 65: 241-69. Zhang, L . , W . Zhou, V E Velculescu, S E Kern, R H Hruban, S R Hamilton, B . Vogelstein, and K W Kinzler. "Gene expression profiles in normal and cancer cells." Science, 1997, 276: 1268-72. Zhao, Z . , C C Lee, D G Monckton, A . Yazdani, M I Coolbaugh, X . L i , J. Bailey, Y . Shen, and C T Caskey. "Characterization and genomic mapping o f genes and pseudogenes o f a new human protein tyrosine phosphatase." Genomics, 1996,35: 172-81. Zheng, X M , Y . Wang, and C J Pallen. " C e l l transformation and activation o f pp60-src by overexpression o f a protein tyrosine phosphatase." Nature, 1992, 359: 336-9.  118  Zhou, B P , J. Deng, W . X i a , J. X u , Y M L i , M . Gunduz, and M C Hung. "Dual regulation o f Snail by G S K - 3 (3-mediated phosphorylation i n control o f epithelial-mesenchymal transition." Nature C e l l Biology. 2004, 6(10): 931-40.  

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.831.1-0092194/manifest

Comment

Related Items