UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The putative role of matrix metalloproteinase 13 and oncostatin M in the establishment of bone metastases Mancini, Stephanie Sarah Jane 2008

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

Item Metadata

Download

Media
24-ubc_2008_fall_mancini_stephanie_sarah_jane.pdf [ 3.95MB ]
Metadata
JSON: 24-1.0066946.json
JSON-LD: 24-1.0066946-ld.json
RDF/XML (Pretty): 24-1.0066946-rdf.xml
RDF/JSON: 24-1.0066946-rdf.json
Turtle: 24-1.0066946-turtle.txt
N-Triples: 24-1.0066946-rdf-ntriples.txt
Original Record: 24-1.0066946-source.json
Full Text
24-1.0066946-fulltext.txt
Citation
24-1.0066946.ris

Full Text

THE PUTATIVE ROLE OF MATRIX METALLOPROTE1NASE 13 AND ONCOSTAT1IST M IN THE ESTABLISHMENT OF BONE METASTASES by STEPHANIE SARAH JANE MANCINI B.Sc., The University of British Columbia, 2005 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Cell and Developmental Biology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) October 2008 © Stephanie Sarah Jane Mancini, 2008 ABSTRACT Breast cancer has a high propensity to metastasize to bone. While the genetic and epigenetic changes associated with metastatic breast cancer progression are being identified, the changes that drive metastatic progression are poorly understood. Proteases, and in particular matrix metalloproteinases (MMPs), have been shown to play a pivotal role in certain aspects of tumor metastasis by modifying the affected microenvironment. Bone matrix-depositing mouse MC3T3 osteoblasts were co-cultured with metastatic human MDA-MB-23 1 (MDA23 1) cells or the bone-homing MDA-MB 231-1 833/TR (1 833/TR) variant in an effort to identify novel, osteoclast-independent, changes to the tumor/bone microenvironment. Co-culture-induced changes in the complete “protease and inhibitor” expression profile in the osteoblasts and the tumor cells were then determined using targeted murine and human specific microarray chips (CLIP-CHIPTM).This analysis revealed an increase in the RNA expression of collagenase-3 (MMP 13) in the co-cultured osteoblasts that was confirmed by qPCR. Further, Western blotting indicated increased MIvIP13 protein secretion into the bone matrixltumor microenvironment by the co-cultured MC3T3 cells. The elevation in osteoblast-produced MMP 13 was observed when the co cultured tumor cells were in direct contact or separated by filters. Additionally, the elevation was also induced by conditioned medium derived from separate MDA23 1 or 1 833/TR cultures, which indicates that a soluble factor produced by the tumor cells is capable of inducing MMP 13. One soluble factor that appears to be produced by 1 833iTR cultures is oncostatin M. Oncostatin M is an interleukin-6 family cytokine that is known to upregulate MMP13 synthesis and secretion during chondrogenesis. Genome-wide Affymetrix® analysis revealed, and qPCR analysis confirmed, that oncostatin M receptor-specific subunit RNA was also significantly upregulated in co-cultured osteoblasts. Therefore, breast tumor cells may be capable of initiating protein degradative changes in the bone microenvironment that are independent of the much studied osteolytic degradation initiated by osteoclast activation. 11 TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES vi LIST OF FIGURES vu LIST OF ABBREVIATIONS viii ACKNOWLEDGEMENTS xi CHAPTER 1: INTRODUCTION 1 1.1 Metastatic breast cancer 1 1.2 Complications of bone metastases 1 1.3 Normal bone development 3 1.4 The vicious cycle between tumor and bone cells 1.5 Current and developing therapies 10 1.6 Matrix metalloproteinase 13 12 1.7 Oncostatin M 14 1.8 Objectives and hypothesis CHAPTER 2: MATERIALS AND METHODS 16 2.1 Tissue culture materials and methods 16 2.1.1 Cell culture 16 2.1.2 GFP expression in breast cancer cell lines 16 2.1.3 Fluorescence Automated Cell Sorting (FACS) 16 2.1.4 MC3T3 differentiation in 2- and 3-dimensions 17 2.1.5 Co-culture of MC3T3 osteoblasts and breast cancer cells in 2-and 3-dimensions 17 2.1.6 Collection and heat inactivation of conditioned medium 18 2.1.7 Transwell migration assay 18 2.1.8 LY294002 treatment 18 2.2 Cell and tissue staining materials and methods 19 2.2.1 Alkaline phopshatase and von Kossa staining 19 2.2.2 Gelfoam tissue staining 19 2.2.3 Hematoxylin and eosin Gelfoam staining 20 2.2.4 von Kossa Gelfoam staining 20 2.2.5 Immunohistochemistry 20 2.2.6 Immunocytochemistry and microscopy 21 2.3 RNA isolation and expression analysis materials and methods 22 2.3.1 TotalRNAlysis 22 2.3.2 cDNA archiving and quantitative PCR (qPCR) 22 2.3.3 CLIP-CHIPTM expression analysis 23 2.3.4 Affymetrix GeneChip® expression analysis 23 2.4 Protein isolation and analysis materials and methods 24 2.4.1 Whole cell protein lysis 24 2.4.2 Culture supernatant protein precipitation 24 2.4.3 Western blotting 25 111 CHAPTER 3: RESULTS .28 3.1 Differentiation of MC3T3 cells into an osteoblast-like state 28 3.2 Development of 2- and 3-dimensional tumor/osteoblast co-culture models 28 3.3 MC3T3 conditioned medium induces migration and actin reorganization in 1833/TR breast cancer cells 38 3.4 Focused CLIPCHIPTM microarray analysis of MC3T3 osteoblast gene expression in response to co-culturing with metastatic breast tumor cells 46 3.5 Matrix metalloproteinase 13 is induced in MC3T3 osteoblasts by tumor cells at both the transcript and protein levels 47 3.6 Genome-wide Affymetrix® analysis of MC3T3 cell gene expression in response to co-cultureing with bone metastatic 1833/TR breast tumor cells 56 3.7 Oncostatin M is a candidate molecule secreted by breast tumor cells that is capable of increasing the expression levels of both MMPI3 and OSMR in MC3T3 osteoblasts...68 3.8 Recombinant OSM and 1833/TR conditioned media-mediated upregulation of MMP13 and OSMR expression is not dependent upon P13K signaling cascade 75 CHAPTER 4: DISCUSSION 78 4.1 2- and 3-dimensional tumor/osteoblast co-culture models 78 4.2 Osteoblast produced factors are capable of increasing migration of tumor cells using a 2-dimensional co-culture model 79 4.3 A factor produced by breast tumor cells upregulates MMP13 expression and secretion in an osteoblastic microenvironment 80 4.4 Oncostatin M can upregulate MMP13 expression in osteoblasts 82 4.5 A potential second “vicious cycle” in bone metastasis 83 REFERENCES 87 APPENDICES 98 Appendix A: MMP-13 expression from MC3T3 cells after direct contact co-culture and FAC sorting 98 Appendix B: Anti-OSM blocking antibody experiment with recombinant OSM and 1833/TR conditioned medium 100 Appendix C: List of significant gene expression changes using CLIPCHIPTM microarray analysis of MC3T3 cells co-cultured with MDA-MB-231 breast tumors on filters 103 Appendix D: List of significant gene expression changes using CLIPCHIPTM microarray analysis of MC3T3 cells co-cultured with MDA-MB-231 breast tumor conditioned medium 105 Appendix E: List of significant gene expression changes using CLIPCHIPTM microarray analysis of MC3T3 cells co-cultured with MDA-MB-231-1833/TR breast tumor cells on filters 106 iv Appendix F: List of significant gene expression changes using CLIPCHIPTM microarray analysis of MC3T3 cells co-cultured with MDA-MB-231-1833/TR breast tumor conditioned medium 107 Appendix G: List of Gene Ontology analysis of Affymetrix® data — Conditioned medium versus serum free 108 Appendix H: List of Gene Ontology analysis of Affymetrix® data — Filter versus serum free 113 Appendix I: List of significant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Filter vs. serum-free 118 Appendix J: List of significant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Conditioned medium vs. serum-free 121 Appendix K: List of significant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Conditioned medium vs. filter 126 V LIST OF TABLES Table 1 — Primary antibodies for Western blotting 26 Table 2 — Secondary antibodies for Western blotting 27 Table 3 — CLIPCHIPTM microarray analysis for proteases and protease modifiers 48 Table 4 — Affymetrix® microarray analysis 59 Table 5 — Gene ontology analysis top 10 pathways — conditioned medium vs. serum-free 61 Table 6 — Gene ontology analysis top 10 pathways — filter vs. serum-free 62 vi LIST OF FIGURES Figure 1 — Differentiation of MC3T3 cells in 2-dimensional monolayers 29 Figure 2 — Differentiation of MC3T3 cells in 3-dimensional Gelfoam scaffolds 31 Figure 3 — Schematic of 2- and 3-dimensional culture models 34 Figure 4 — 2-dimensional co-culture with MDA23 1 cells in direct contact with differentiated MC3T3 monolayers 36 Figure 5 — 3-dimensional co-culture with MDA23 1 cells and differentiated MC3T3 cells in Gelfoam sponges 39 Figure 6 — 1 833/TR tumor cells migrate in response to MC3T3 osteoblast-conditioned medium 42 Figure 7 — Treatment with MC3T3-conditioned medium results in a reorganization of the actin cytoskeleton 44 Figure 8 — Validation of increased MIVIP 13 expression in MC3T3 cells upon exposure to breast tumor cells 49 Figure 9 — MMP13 expression is induced in MC3T3 cells in a time-dependent manner and is not enhanced by direct cell-cell interaction 52 Figure 10 — MMP 13 protein levels increase when co-cultured with MDA23 1 and 1 833/TR cells 54 Figure 11 — Affymetrix® analysis of MC3T3 cells co-cultured with 1 833/TR cells 57 Figure 12 — qPCR validation of increased mouse OSMR expression levels in MC3T3 osteoblasts upon exposure to 1 833/TR breast tumor cells 64 Figure 13 — Human OSM ligand and receptor expressed by 1833/TR tumor cells 66 Figure 14 — Heat-inactivation of 1 833/TR-conditioned media 69 Figure 15 — Recombinant oncostatin M increases MMP 13 and OSMR expression levels in MC3T3 osteoblast cells 71 Figure 16 — MIvIP 13 protein expression is increased by oncostatin M 73 Figure 17 — LY294002 treatment does not attenuate MIVIP 13 induction by OSM or 1 833/TR-conditioned medium 76 Figure 18—A potential pathway involving OSM and MIvIP13 in the enhancement of bone colonization by breast cancer cells 85 vii LIST OF ABBREVIATIONS 183 3/TR MDA-MB-23 1-1 833/TR cells 2D two dimensional 3D three dimensional Ab antibody ALP alkaline phosphatase aMEM minimum essential medium alpha modification AMP adenosine monophosphate ANT adenine nuclotide translocase AP- 1 activator protein 1 ATP adenosine triphosphate BCA bicinchoninic acid assay BMP bone morphogenetic protein BP bisphosphonate Bsp bone sialoprotein Ca2 divalent calcium cations CA9 carbonic anhydrase 9 eDNA copy deoxyribonucleic acid CM conditioned medium CNTF ciliary neurotrophic factor CO2 carbon dioxide COX-2 cyclooxygenase 2 CTGF connective tissue-derived growth factor CTR calcitonin receptor CXCL12 chemokine (C-X-C motif) ligand 12 CXCR4 chemokine (C-X-C motif) receptor 4 DABCO 1 ,4-diazabicyclo[2.2.2]octane DAPI 4’ 6-diamidino-2-phenylindole ddH2O distilled, deionized water DMF N,N-dimethylformamide DMSO dimethyl sulfoxide DNase deoxiribonuclease ECL enhanced chemiluminescence ECM extracellular matrix EDTA ethylenediaminetetraacetic acid EGF epidermal growth factor EtOH ethanol FACS fluorescence automated cell sorting FBS fetal bovine serum Fe immunoglobulin constant region viii FDR false discovery rate FGF fibroblast growth factor FPP farnesyl diphosphate GFP green fluorescence protein GGPP geranylgeranyl diphosphate GM-CSF granulocyte-macrophage colony-stimulating factor GO gene ontology GTPases guanine trinucleophosphatases HER2 human epidermal growth factor receptor 2 HI heat inactivated HRP horseradish peroxidase IFNy interferon y IGF insulin-like growth factor IL interleukin Jak janus kinase INK c-Jun N-terminal kinase LAP latency-associated peptide LIF leukemia inhibitory factor MAPK mitogen activated protein kinase MCP monocyte chemoattractant protein M-CSF macrophage colony-stimulating factor MDA23 1 MDA-MB-23 1 cells MIP-la macrophage inflammatory protein-la MMP matrix metalloproteinase MT 1 -MMP membrane type 1 matrix metalloproteinase NaC1 sodium chloride NBF neutral buffered formalin N-BP nitrogen-containing bisphosphonate OPG osteoprotegrin OPN osteopontin OSM oncostatin M OSMR oncostatin M receptor 1 Osx osterix PBS phosphate-buffered saline PDGF platelet-derived growth factor PFA paraformaldehyde PGE2 prostaglandin E2 P13K phosphomositide 3-kinase PKA protein kinase A PKB protein kinase B ix PPi inorganic pyrophosphate PTH parathyroid hormone PTHrP parathyroid hormone-related protein PVDF polyvinylidene fluoride qPCR quantitative polymerase chain reaction RANK receptor activator of NF-kappaB RANKL receptor activator of NF-kappaB ligand Rho ras homology RMA robust multichip average RNA ribonucleic acid Rnase ribonuclease R-Smad receptor Smad SDF- 1 stromal cell-derived factor 1 Smad SMA mothers against decapentaplegic SMP skim milk powder SPARC secreted protein, acidic, cysteine-rich STAT signal transducers and activator of transcription Sw3T3 Swiss 3T3 fibroblast cells T(3R TGF (3 receptor TCA trichloroacetic acid TGF(3 transforming growth factor (3 TGL thymidine kinase/GFP/luciferase TIMP tissue inhibitor of metalloproteinase TIVIA tissue microarray TNF a tumor necrosis factor a TNFR tumor necrosis factor receptor TRAF-6 TNFR-associated factor 6 TRAIL TNF-related apoptosis-inducing ligand TRAP tartrate resistant acid phosphatase Tyk tyrosine kinase VCAIVI- 1 vascular cell adhesion molecule 1 VEGF vascular endothelial growth factor VK von Kossa x ACKNOWLEDGEMENTS I would like to extend my thanks to all of the people who have helped me over the course of my graduate studies. First, I would like to thank Dr. Calvin Roskelley for giving me the opportunity to work in his lab. I would also like to thank the members of my advisory committee, Dr. Michael Underhill and Dr. Christopher Overall for their helpful suggestions through out my degree. I would also like to acknowledge the strong collaborative background of this project with Dr. Christopher Overall and his lab members. In particular, I would like to thank Dr. Charlotte Morrison for a wonderful collaboration through out the course of my studies. I would also like to extend special thanks to the current members of the Roskelley lab, Jane Cipollone, Pamela Austin, Spencer Freeman, Dr. Marcia Graves, and Dr. Sarah McLeod, who have all contributed throughful advice and support for this project. I would especially like to thank Jane Cipollone for her technical assistance with aspects of this project. Finally, I would like to thank my family, John, Alexandra, Kimberly, David, and Andrew 5, all of whom I am indebted to for their support during my graduate studies. xi CHAPTER 1: INTRODUCTION 1.1 Metastatic breast cancer Breast cancer is one of the most commonly diagnosed cancers in women, accounting for approximately 26% of the total new cancer cases, and for approximately 15% of all cancer related deaths (Jemal et al., 2007). At the time of diagnosis, approximately 60% of women present with localized disease, 31% with regional stage disease and 6% with distant metastases (Jemal et al., 2007). Ultimately, however, most breast cancer-related deaths are due to the emergence of distant metastatic disease rather than the expansion or re-emergence of the primary tumor (Mundy, 2002). Most solid tumor types tend to have a preference for metastasizing to certain organs and breast cancer tends to home to the bone, brain, liver and lungs (Coleman, 1997). Stephen Paget initially noticed this selective pattern in 1889 and hypothesized that “when a plant goes to seed, its seeds are carried in all directions; but they can only grow ftheyfall on congenial soil” (Paget, 1889) and he proposed that metastasis did not occur by chance but that the “seed” (i.e. the tumor cell) and the “soil” (i.e. the microenvironment at the distant, metastatic site) must be compatible (Reviewed by Ribatti et al., 2006). Essentially, under this paradigm, only those cells that gain a selective growth advantage due to factors present in the distant organ’s microenvironment would be able to establish a secondary tumor, or micrometastasis, in that tissue (Paget, 1889). In the case of breast cancers, the most common site of secondary tumor formation is bone (Coleman, 2006; Mundy, 1997; Mundy, 2002; Roodman, 2004). As a result, almost 70% of patients with advanced stage breast cancer have the presence of bone metastases as found by post-mortem examination (Coleman and Rubens, 1987). 1.2 Complications ofbone metastases Bone metastases occur most frequently in the red bone marrow of the proximal ends of long bones, the ribs, and the vertebral bodies (Mundy, 1997). The blood flow is high in the red bone marrow, which may help explain why metastases colonize this area of the bone 1 (Roodman, 2004). Osteolytic metastases (excessive bone destruction) are most common in breast cancer patients whereas osteoblastic metastases (excessive bone formation) are relatively rare (Blair et al., 2006; Mundy, 1997; Mundy, 2002; Roodman, 2004). In both types of bone lesions, however, there is increased osteoclast activity and bone resorption as measured by serum and urine markers (Seibel, 2005). Thus, it has been assumed that some degree of bone resorption must occur for tumor cells to successfully colonize and develop secondary site lesions in both osteolytic and osteoblastic metastatic lesions that originate in the breast (Nielsen et al., 1991; Wittrant et al., 2004b). Some of the most common clinical complications of metastatic bone lesions include pain, fractures, spinal cord or nerve compression syndromes, and hypercalcemia (Mundy, 1997; Mundy, 2002; Roodman, 2004). Commonly, the hypercalcemia is due to increased bone resorption and calcium release into the bloodstream that can be initiated by parathyroid hormone-related protein (PTHrP) secreted by the tumor cells which stimulates both osteoclast-mediated bone resorption and renal tubular calcium reabsorption back into the bloodstream (Burtis et al., 1990; Guise, 2000; Mundy, 1997). The majority of cancer patients have a dramatic decrease in their quality of life as a result of pain and discomfort created by the presence of metastases (Mantyh et al., 2002; Sabino and Mantyh, 2005). Pain due to bony metastases can be induced by either physical deformations or by chemical stimulation. Mechanical induction of pain is from fractures or micro-fractures, most frequently present in the proximal ends of long bones and in vertebral bodies (Mundy, 1997). Spinal cord or nerve compression syndromes can occur when extensive osteolytic destruction of vertebral bodies has occurred and the bone becomes fractured or deformed (Mundy, 1997). The bone formed in osteoblastic lesions is also susceptible to pathological fractures as the bone is weak due to its disorganized structure (Blair et al., 2006). These syndromes can also occur if the tumor growth directly impacts the spinal cord, or if there is excessive bone deposition from osteoblastic metastases and overgrowth into the vertebral canal (Mundy, 1997). Chemically-induced pain can occur as a result of both cancer cells and other tumor associated stromal cells secreting extra-cellular factors that stimulate or sensitize nociceptors (Reviewed by Sabino and Mantyh, 2005). Tumor-associated macrophages, neutrophils, and T-lymphocytes have all been shown to secrete factors such as 2 prostaglandins, tumor necrosis factor (TNF)-a, endothelins, interleukin (IL)- 1, IL-6, epidermal growth factor (EGF), transforming growth factor (TGF)-, and platelet-derived growth factor (PDGF) which have been shown to directly excite or sensitize primary afferent pain neurons (Reviewed by Sabino and Mantyh, 2005). These molecules are not only involved in the generation of pain, but they are also osteotropic and can act as microenvironmental modulators that facilitate tumor cell proliferation and increase osteoclastic activity, both of which can contribute to the preparation of the ‘soil’ that generates a tumor metastatic “niche” (see below). Tumor cells have also been shown to express cylcooxygenase (COX)-2 and COX-2 inhibitors attenuate bone pain and slow tumor growth and angiogenesis in a murine cancer model of bony metastasis (Gupta and Dubois, 2001). Acidic pH is also capable of stimulating pain, and the microenvironment surrounding breast tumor cell deposits in bone is typically of a low pH because activated osteoclasts generate free hydrogen ions at the bone resorptive interface (Reviewed by Katagiri and Takahashi, 2002; Sabino and Mantyh, 2005). Unfortunately, the current treatment options available for bone metastasis are not curative and thus are intended to ease the painful symptoms palliatively (Coleman, 2006; Mundy, 1997; Vakaet and Boterberg, 2004). Palliative care typically consists of localized radiotherapy, which is usually very effective in providing pain relief as a result of decreased pressure in the bone marrow due to cancer cell death and radiation-induced inflammatory cell death (Vakaet and Boterberg, 2004). Up to two thirds of patients will experience a significant decrease in pain with approximately half experiencing complete, long-term pain relief over the course of disease (Vakaet and Boterberg, 2004). 1.3 Normal bone development Developmentally, bone can be formed in two different ways, via intramembranous or endochondral ossification. One key difference between the two is that in intramembranous ossification osteoblast cells differentiate directly from mesenchymal precursors, whereas in endochondral ossification mesenchymal precursors first differentiate into chondrocytes that form a cartilaginous template, which is then remodeled into bone. 3 Regardless of how it is generated, adult bone is a very dynamic tissue that is regulated by highly regulated cycles of osteoclast-mediated matrix destruction and resorption as well as osteoblast-mediated matrix deposition and mineralization (Blair et al., 2006; Katagiri and Takahashi, 2002). These cycles are important for maintaining bone volume and integrity, for bone remodeling, and for calcium homeostasis. Osteoblasts, which share a common mesenchymal precursor with chondrocytes and adipocytes, secrete numerous extra-cellular matrix (ECM) proteins and regulate the formation of hydroxyapatite crystals to form a mature, mineralized bone matrix (Katagiri and Takahashi, 2002). Once osteoblastic fate has been determined, osteoblast differentiation is regulated and stimulated by a number of factors including fibroblast growth factors (FGFs) and bone morphogenetic proteins (BMPs), which are extra-cellular signaling molecules that belong to the TGFI3 super-family (Katagiri and Takahashi, 2002). RUNX2 (or Cbfal) and osterix (Osx or sp7) are two important transcription factors that drive the determination and differentiation of the osteoblast lineage (Gao et al., 2004; Komori et al., 1997). RIJNX2 is involved in the differentiation of both the chondrocyte and osteoblast fate, and Osx functions downstream of RUNX2 as an osteoblast-specific regulator (Gao et al., 2004; Komori et al., 1997). FGF-2 and BMP-2 are both able to stimulate Runx2 expression, the latter being mediated by the transcription factor D1x5 (Lee et al., 2003; Zhou et al., 2000). FGF-2, RUNX2, and BMP-2 have been shown to function in a regulatory cycle wherein FGF-2 influences osteoblast differentiation at an early stage and initiates Runx2 expression; Runx2 transcriptional activity then results in BMP-2 expression which further reinforces the cycle positively by upregulating Osx expression in a manner which is itself RTJNX2-dependent (Choi et al., 2005). FGFs, BMPs, IGF-I and TGFI3 induce the production of type I collagen, osteopontin, and osteonectin by osteoblasts, resulting in osteoid deposition during bone formation and turnover in a RUNX2-dependent manner (Zaidi, 2007). Osx further restricts the osteoblast lineage by upregulating the expression of bone-specific differentiation genes. Recently, it has been demonstrated that Osx can be induced by BMP-2 in Runx2-deficient mesenchymal cells through Msx2, a homeobox gene that promotes osteoblast differentiation (Matsubara et al., 2008; Nishio et al., 2006). IGF-I is also able to induce Osx expression independently of RUNX2 through p38, a component of the mitogen activated protein kinase (MAPK) signaling cascade (Celil et al., 2005). 4 Knockout ofRunx2, however, results in the complete absence of bone formation. Therefore, while RUNX2-independent pathways contribute to, they alone are not sufficient for complete osteoblastogenesis (Matsubara et al., 2008). Osteoclasts are derived from the monocyte/macrophage lineage and their main role is bone destruction and resorption (Katagiri and Takahashi, 2002). Osteoclasts adhere to the bone matrix and create a sequestered local environment so that degradation only occurs locally at the site of osteoclast-matrix interaction (Chambers, 2000). This process is initiated when osteoclasts first dissolve the mineral crystallites in the bone matrix by decreasing the localized extracellular pH, which is followed by the secretion of cysteine proteinase, and matrix metalloproteinase (MMP) proteolytic enzymes that then digest the matrix proteins (Chambers, 2000; Everts et al., 2006). Cathepsin K is a cysteine proteinase that is very important in the degradation of bone matrix in both calvaria and long bones (Everts et al., 2006). MIVIP9 is the most common gelatinase expressed by osteoclasts that helps degrade collagen in the bone matrix (Wittrant et al., 2004b). Following this initial extracellular digestion, matrix protein fragments are subsequently endocytosed and either further digested intracellularly or secreted basolaterally by the osteoclasts (Chambers, 2000). Differentiated osteoclasts are characterized by being multinucleated with a ruffled border, as well as by expression of calcitonin receptor (CTR), tartrate resistant acid phosphatase (TRAP), and avf33-integrin (Katagiri and Takahashi, 2002; Takahashi et al., 1999; Wittrant et al., 2004b). Importantly, osteoblast cells are largely responsible for the differentiation of osteoclasts, and in vivo this is stimulated via macrophage colony- stimulating factor (M-CSF) and receptor-activator of nuclear factor (NF)-KB (RANK) signaling pathways (Katagiri and Takahashi, 2002; Wittrant et al., 2004a). RANK ligand (RANKL) is expressed as a membrane-bound protein on the surface of osteoblast cells (Lacey et al., 1998), and upon binding with RANK receptors on osteoclast precursors, it induces their differentiation and activation (Nakagawa et al., 1998). Stimulation with osteotropic factors such as 1 ,25-dihydroxyvitamin D3, PTH, prostaglandin E2 (PGE2), IL-i, and IL-li stimulates osteoblast cells to express more surface RANKL and further propel osteoclast differentiation (Katagiri and Takahashi, 2002). 5 After the binding of RANKL, the RANK receptor recruits the TNFR-associated factor (TRAF)-6 in the sub-plasmalemmal cytoplasm which then initiates three divergent downstream signaling pathways (Reviewed by Hofbauer and Heufelder, 2001). One pathway involves the c-Jun N-terminal kinase (INK) which activates c-fos/c-jun while another involves the activation of NF-KB. The third pathway is initiated by a c-src-mediated activation of A1Ct/PKB, which mediates cytoskeletal reorganization and anti-apoptotic signaling that contributes to the osteoclastic differentiation that is initiated by the first two pathways. Osteoprotegrin (OPG) is a soluble decoy molecule for RANKL that acts as a negative regulator of bone resorption by binding to RANKL and preventing it from binding to the RANK receptor on osteoclast precursors (Katagiri and Takahashi, 2002; Simonet et al., 1997; Tsuda et al., 1997). Many different cell types express OPG, including bone- marrow stromal cells, and it is the ratio between OPG and RANKL that determines the activation level of osteoclast cells through the RANK-signaling pathway during normal bone development and remodeling (Blair et al., 2006; Wittrant et al., 2004a). It is this tightly regulated balance between ostoblast matrix deposition and osteoclast matrix degradation during normal bone remodeling that tumor cells take advantage of when establishing metastases, both osteoblastic and osteolytic, in the bone microenvironment. 1.4 The vicious cycle between tumor and bone cells The bone microenvironment is enriched with a large milieu of growth factors that includes TGFI3, insulin-like growth factors (IGF)-I and IGF-II, FGF, PDGF and BMPs (Reviewed by (Roodman, 2004). These growth factors are deposited along with ECM proteins into the bone matrix by osteoblasts, and they are released from the ECM by osteoclast-mediated bone resorption (Katagiri and Takahashi, 2002). During normal bone remodeling, growth factor release from the bone matrix stimulates osteoblast growth and matrix production, forming new bone matrix (Zaidi, 2007). Tumor cell production of osteoclast activating factors such as PTHrP, IL-i, IL-6, IL-8, IL-il, TNFa, macrophage inflammatory protein-i -a (MIP- 1 a) and RANKL all increase osteoclast-mediated resorption which feeds forward to further increase growth factor release from the bone 6 matrix into the metastatic microenvironment (Reviewed by Guise et al., 2006; Guise and Mundy, 1998). Of these factors, TGF is very abundant and can play a role in the formation of both osteoblastic and osteolytic bone metastases. While the majority of bone metastases formed from breast cancer are osteolytic, many patients have mixed lesions with both osteolytic and osteoblastic activity. In fact 15- 20% of breast cancer cases are predominantly osteoblastic, and the formation of new bone occurs in response to osteoclast-mediated bone resoption (Roodman, 2004). Upon bone resorption and TGFI3 release from the bone ECM, osteoblasts increase collagen I synthesis and deposition, forming new bone (Zaidi, 2007). This growth factor stimulated bone formation must outweigh the rate of osteoclast-mediated bone degradation in order for an osteoblastic lesion to form. In the majority of breast cancer cases, however, the arrival of tumor cells results in more bone destruction than bone formation, resulting in osteolytic lesions. TGFI3 can stimulate PTHIP production by the tumor cells, defining it as an important functional mediator in osteolytic bone metastasis (Yin et a!., 1999). Specifically, TGF, PTHrP, RANK, RANKL and OPG signaling have all been found to be important components in the development of osteolytic bone metastases. Therefore, a so-called “vicious cycle” develops between tumor and bone cells that results in an abnormal, persistant, and ultimately pathologic, osteolytic bone destruction that is most frequently seen in breast cancer (Guise, 2000; Wittrant et al., 2004a; Yin et al., 1999; Yoneda et al., 2000). TGF3 keeps the vicious cycle moving forward by initiating a signaling cascade in tumor cells via a heterodimeric cell surface receptor composed of type I and type II receptor subunits, TRI and TRII (Reviewed by Lindemann et al., 2001). TGFI3 binds to the extracellular domain of TRII and Tj3RI is then recruited and activated via phosphorylation from TI3RJI. This leads to the recruitment, phosphorylation and activation of receptor (R) Smad proteins, either Smad2 or Smad3. The R-Smad then associates with a common mediator, Smad4, and the complex translocates to the nucleus where it can act as a transcription factor. For example, Smad3, in cooperation with an Ets transcription factor, binds to the AGAC box present in the PTHrP P3 promoter region and activates it leading to the gene’s increased expression (Lindemann et al., 2001). The importance of TGFI3-induced expression of PTHrP in osteolytic bony metastases was clearly demonstrated by expressing a series of receptor mutants in MDA 7 MB-23 1 (MDA23 1) breast cancer cells that were injected into mice by intracardiac injection to generate bone metastases (Yin et al., 1999). Specifically, Yin et a!. generated a dominant negative TGFI3 type II receptor that has no cytoplasmic tail and therefore no downstream signaling, resulting in less PTHrP production. Expression of this dominant negative mutant in the tumor cells decreases bone destruction, decreases the number of tumor-associated osteoclasts, and it increases survival. This effect was specific to the function of PHTrP because both forced expression of PTHrP or constitutively active T(3R1 in the tumor cells reversed the effect (Yin et al., 1999). Similarly, a decrease in bone tumor volume and number of tumor-associated osteoclasts occurs with the addition of a PTHrP blocking antibody (Guise et al., 2006). PTHrP is expressed in 50-60% of human primary breast tumors, and it is present at higher levels in metastatic cells that colonize the bone as compared to tumor cells at the primary and non-bone metastatic sites (Reviewed by Guise, 2000). In addition to TGFI3, PTHrP production in tumor cells can also be stimulated by other growth factors present in the bone microenvironment, including FGF1, FGF2, IGF-l, IGF-2, BMP-2, and PDGF (Yin et al., 1999). Interestingly, MCF-7 breast cancer cells do not produce PTHrP and they do not normally cause osteoblast-dependent osteolytic bone destruction unless PTHrP is force expressed (Swarthout et al., 2002). PTHrP signaling in osteoblast cells activates a number of transcription factors including RLTNX2, which stimulates RANKL and M-CSF expression but inhibits OPG expression (Hofbauer and Heufelder, 2001). Therefore, the increased formation and activation of osteoclasts is induced by RANKL and M-CSF that is produced by osteoblasts as a result of PTHrP signaling from the tumor cells (Swarthout et al., 2002). As suggested by the experimental data described above, an important factor in the development of osteolytic bone metastases is the ratio of RANKE to the decoy ligand OPG. This is further supported by clinical data. Specifically, in patients with osteolytic tumors, the RANKL/OPG ratio is often increased when compared to healthy tissue (Wittrant et al., 2004a). Importantly, RANKL is highly expressed in a subset of patients diagnosed with primary adenocarcinomas in the breast, prostate, and lung, many of whom later develop highly osteolytic disease. This suggests that tumor cell expression of RANKL can also enhance osteoclast activation directly (Huang et al., 2002). 8 Although the RANK-RANKE pathway has been a major research focus as a potential therapeutic target, other molecules and pathways are being uncovered that are involved in metastasis to bone and thus could be potential targets for therapy. Recently, Kang et al. isolated a subpopulation ofMDA231 cells, named 1833/TR cells, which are highly metastatic to bone (Kang et al., 2003). Genome-wide expression analysis revealed a number of genes that were associated with the increased bone metastatic ability, as compared with the parental MDA23 1 cells. These changes in gene expression were stable and were not altered by in vivo or in vitro culturing (Kang et al., 2003). Four of the most highly over-expressed genes included IL-li which activates osteoclast differentiation, connective tissue-derived growth factor (CTGF) which is an angiogenic growth factor, CXCR4 which is a bone-homing chemokine receptor, and MJvIP1 which cleaves collagen in osteolysis (Kang et al., 2003). Kang et al. assessed the effects of forced over-expression of these genes both individually and combinatorially, including the over-expression of osteopontin (OPN), which is over-expressed in highly metastatic cells (Kang et al., 2003). The over-expression of IL-li alone did not significantly increase the bone metastatic ability of the parental MDA23 1 cells, but over-expression of both IL-li and OPN did. Both CTGF and CXCR4 also failed to increase bone metastasis significantly when over-expressed individually. However, over-expression of either CTGF or CXCR4 with IL-li and OPN dramatically increased bone metastasis and they showed aggressiveness very close to that seen by the 1 833/TR bone metastatic subpopulation. Over-expression of MIvIP 1 individually or with IL-li and OPN was able to increase bone metastasis (Kang et al., 2003). Although stable over-expression of these genes is able to increase bone metastasis, the bone microenvironment is still able to influence their expression through TGF(3 signaling. Both IL-il and CTGF expression is increased in response to TGF3 signaling in a Smad-dependent manner (Kang et al., 2003). Thus, the use of large scale analyses to identify functionally important gene expression changes initiated by cancer cell-bone interactions may facilitate the identification of targetable molecules for new, metastastic specific therapies. Analyses of this kind also highlight the complex nature of the disease, which indicates that the blocking of more than one molecule and/or pathway will likely be required to generate effective therapeutic response. 9 1.5 Current and developing therapies The primary approach to clinical treatment of bone metastasis is to block the osteoclast activity, preventing the release of more growth factors, and hopefully disrupting the vicious cycle. This is currently done using bisphosphonates (BPs), which are specific inhibitors of osteoclast activity (Yoneda et a!., 2000). BPs, which are able to specifically target bone due to their high affinity for divalent cations including Ca2 (Coxon et al., 2006), are also widely used to treat other bone disorders including Paget’s disease, hypercalcinemia, and postmenopausal osteoporosis (Coxon et al., 2006; Fliesch, 1991; Gong et a!., 2003; Yoneda et a!., 2000). BP molecules are synthetic analogues of inorganic pyrophosphate (PPi; Reviewed by Coxon et al., 2006). There are two types of BP molecules, simple BPs that are the most structurally similar to PPi, and nitrogen-containing BPs (N-BPs) that have bulkier side chains. The simple BPs include clodronate and etidronate and they are metabolized by the osteoclasts into methylene-containing ATP analogues. These analogues accumulate in the cytosol and interfere with ATP-dependent enzymes, including adenine nucleotide translocase (ANT) which when inhibited can lead to mitochondrial membrane breakdown and cytochrome C release, triggering apoptosis (Reviewed by Coxon et a!., 2006). In contrast, the N-BPs, which include aledronate, ibandronate, risedronate, and zoledronate, are not metabolized. They act by inhibiting farnesyl diphosphate (FPP) synthase, depleting the cells of FFP and geranylgeranyl diphosphate (GGPP) isoprenoid lipids. This inhibition of prenylation results in disruption of smallGTPase signaling in the Ras, Rho and Rab families, which is required for the functioning of activated osteoclasts. (Reviewed by Coxon et al., 2006). In an animal model of breast cancer metastasis to bone, treatment with the bisphosphonate molecules risedronate, ibandronate, and zoledronate inhibits the development and progression of osteolytic bone metastases formed by MDA23 1 breast cancer cells (Hughes et a!., 1995; Sasaki et a!., 1995; Yoneda et al., 1997). This inhibition was driven by an increased apoptosis in both osteoclasts and breast cancer cells at the site of the bone metastasis (Hughes et a!., 1995). In clinical trials, bisphosphonate treatment is effective at increasing the time of bone lesion progression-free survival, but many patients will still develop bone metastases, causing skeletal-related complications, including pain 10 and bone fractures (Lipton, 2007). Fortunately, bisphosphonates have been very helpful symptomatically, particularly in terms of dealing with patient pain and discomfort, significantly increasing the patients’ quality of life (Devitt and McLachlan, 2008; Lipton, 2007). While these treatments are useful for short-term pain management, they have not proven curative. So far, none of the bisphosphonate treatments used in clinical trials increase the mean survival time of patients with metastatic disease (Devitt and McLachlan, 2008; Lipton, 2007; Major et al., 2005). Therefore, the development other therapeutic strategies are being actively developed. Because the RANKL/OPG ratio has the ability to regulate osteoclast activation, considerable research effort has been directed towards effectively targeting these molecules as a therapeutic strategy (Wittrant et al., 2004a). Current strategies include using soluble inhibitors of the RANK-signaling pathway (Reviewed by Wittrant et al., 2004a). In mouse models, the RANKL decoy OPG has been effective in the treatment and prevention of hypercalcemia, inhibition of osteoclastogenesis, tumor growth, osteolysis, skeletal destruction, decrease in tumor burden, increase in survival, and decrease in bone cancer pain (Reviewed by Wittrant et al., 2004a). In human patients, a single dose of recombinant OPG resulted in suppression of bone resorption and was well tolerated by the patients (Wittrant et al., 2004a). There are, however, potential drawbacks to OPG; it can bind to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) and therefore could potentially act as a protective compound for some cancer cells (Wittrant et al., 2004a). However, the use of RANK-Fc could potentially avoid this problem from arising because it cannot interfere with TRAIL-mediated cancer cell apoptosis (Wittrant et al., 2004a). RANK-Fc is a recombinant RANKL antagonist formed by fusing the Fc portion of human immunoglobulin G1 (hlgGi) to the extracellular domain of RANK (Sordillo and Pearse, 2003). RANK-Fc is active in vitro and is able to inhibit osteoclast formation (Sordillo and Pearse, 2003). In mouse and rat studies, RANK-Fc has been demonstrated to be a therapeutically active antagonist to RANKL signaling, it also inhibits tumor-induced osteolysis and decreases tumor burden specifically in osteolytic metastases (Sordillo and Pearse, 2003). Currently, a humanized monoclonal antibody against RANKL, which should mimic the action of RANK-Fc, is being tested in early stage clinical trials (Lipton et al., 2007). Combinatorial treatment using RANK-Fc and MIP- 1 a neutralizing antibodies 11 resulted in inhibition of the development and progression of experimental osteolytic lesions, as well as a decrease in overall tumor burden (Wittrant et a!., 2004a). Taken together, results from these studies indicate that direct, targeted therapy is possible. Thus, ongoing research into the identification of new molecules involved in the establishment of metastases should open up new avenues for therapeutic strategies. Two new potential therapeutic targets in bone metastasis that have been identified and studied in this thesis are matrix metalloproteinase 13, which is produced by osteoblasts, and oncostatin M, which is produced by breast tumor cells. Interestingly, and potentially importantly, these two molecules appear to be involved in a cancer cell/bone-induced pathway that is RANK ligand-independent. 1.6 Matrix metaioproteinase 13 Collagenase 3, or matrix metalloproteinase (IVIMP) 13, is a member of the IvilviP family that can cleave type II collagen and the large proteoglycan aggrecan, which are both prominent components of the bone matrix (Page-McCaw et al., 2007). During endochondral bone formation, MIVIP 13 is expressed by hypertrophic chondrocytes of the growth plate and by osteoblasts. Whole animal knockouts indicate that the protease is required for efficient ossification and bone formation at the growth plate of long bones as well as the proper formation and remodeling of bone trabeculae (Inada et a!., 2004; Stickens et al., 2004). An increase in MMP 13 expression in early invasive primary ductal breast carcinoma tissue has been correlated with decreased survival (Gonzalez et a!., 2007; Vizoso et a!., 2007). While the majority of the samples were positive for MMP 13 expression (Gonzalez et a!., 2007), expression in the tumor cells themselves was not associated with poor prognosis (Vizoso et a!., 2007). Instead, high levels of MMP 13 in either stromal fibroblasts or surrounding inflammatory mononuclear cells was associated with the presence of an infiltrating tumor edge, decreased relapse-free survival and increased the rate of the formation of distant metastases (Vizoso et al., 2007). Experimentally, mouse xenograft models in which MDA23 1 breast tumor cells were implanted orthotopically into the mammary fat pad induce increased expression of MMP 13 in the fat pad stroma (Lafleur et a!., 2005). Therefore, breast tumor cells can influence stroma! MMP 13 production in the 12 primary tumor site microenvironment. However, it was not previously known if breast tumor cells induce stromal MIvIP 13 expression in the distant site bone microenvironment. This is an issue I address directly in this thesis. Like other members of the MIVIP family, MIVIP 13 is released in an inactive state and can be activated by, and itself can activate, other MMPs. Specifically, MMP13 can be activated by MMP2, MIVIP3 or MIVIP14 (membrane type MT1-MMP) and active MMP13 itself can then cleave and activate IVIIVIP2 and 1v11v1P9 (Knauper et al., 1996; Leeman et al., 2002). Aside from matrix degradation and remodeling, ]VIIvIP13 cleaves and modifies the function of a number of growth factors and cytokines. For example, like MMP2 and MIVIP9, IvilviP 13 can release latent TGF from the ECM and activate it by cleaving off the latency-associated peptide (LAP), thereby effectively increasing both its bioavailability and activity (Dangelo et al., 2001). MIvIP13 can also activate CTGF, VEGF, and bFGF (Mott and Werb, 2004; Nagase et aL, 2006; Sounni and Noel, 2005) while it can inactivate monocyte chemoattractant protein (MCP)-3 and stromal cell-derived factor (SDF)- 1 (McQuibban et al., 2001; McQuibban et al., 2002; Nagase et al., 2006; Sounni and Noel, 2005). MMP 13 expression can be induced by a number of different signaling pathways in a number of different cell types. In osteoblast cells, MIVIP 13 induction occurs in an AP- 1- and Cbfal/RUNX2-dependent manner. Therefore, its expression increases during osteoblastic differentiation and bone mineralization (Winchester et al., 2000). PTH has also been shown to increase MMP 13 levels through a protein kinase A (PKA) and Cbfal/RTJNX2-dependent signaling pathway in osteoblasts (Quinn et al., 1990; Selvamurugan et al., 2000). In chondrocytes, IL-i and TNF-a both induce MIvIP13 expression via MAPK- and AP-i dependent pathways (Ahmad et al., 2007b; Liacini et al., 2003). Interestingly, the IL-i- mediated induction can be blocked by IFNy due to a STAT 1-dependent sequestration of the cyclic AMP response element-binding protein (CBP)/p3 00 transcription coactivator that binds to and directly activates the MIvIP 13 promoter (Ahmad et al., 2007a). Conversely, the cytokine oncostatin M (OSM), which is produced by breast tumor cells (see below), acts synergistically with IL-i to induce MIVIP 13 expression in chondrocytes (Litherland et al., 2008). 13 1.7 Oncostatin M OSM is an IL-6 family cytokine that was originally isolated as a soluble molecule produced by macrophages that inhibits the growth of melanoma cells (Hoffman et al., 1996; Zarling et al., 1986). Human OSM is a highly stable protein; it can resist denaturing from pH 2 to pH 11 and heating at 56°C for 1 hour (Zarling et al., 1986). OSM does, however, lose activity at 90°C, unlike TGFI3 which is still active after heating to this temperature (Zarling et al., 1986). The IL-6 family of cytokines also includes ciliary neurotrophic factor (CNTF), IL-6, IL-il, cardiotrophin- 1, and leukemia inhibitory factor (LIF) which all signal through heterodimeric cytokine receptors that use the gp 130 subunit as a common component (Auguste et al., 1997; Lindberg et al., 1998; Rose-John et al., 2006). Specificity for OSM recognition is mediated by an OSM receptor 3/gp 130 “Type II OSMR” dimer, although OSM can also bind to the LIF receptor 3/gp 130 “Type I OSMR” dimer (Auguste et al., 1997; Mosley et al., 1996; Rose-John et al., 2006). Signaling through both the Type I and Type II OSMR pathways involves Jakl, Jak2, and Tyk2 kinase activation and STAT3 phosphorylation downstream of cytokine binding. In addition, STAT5b can also be specifically phosphorylated downstream of the activated Type II OSMR signaling (Auguste et al., 1997). Importantly, expression of the OSMR subunit itself is increased by Jakl, Jak2, and Tyk2 induced signaling (Radtke et al., 2002). Aside from its ability to inhibit tumor cell growth, OSM appears to play a role in other cellular processes including the regulation of acute inflammation (Hurst et al., 2002), skin inflammation (Boniface et al., 2007), hematopoiesis (Tanaka et al., 2003), spermatogenesis (de Miguel et al., 1997), and osteoblast differentiation (Malaval et al., 2005). Regarding the latter effect, treatment of early rat calvarial cell cultures with OSM increases the number of cell colonies (ie. stimulates proliferation), but it inhibits bone nodule formation (Malaval et al., 2005). The latter inhibition is associated with decreased expression of osteocalcin and bone sialoprotein (Bsp) at early differentiative stages (Malaval et al., 2005). OSM may also promote tumor progression by altering tumor-stroma interactions. For example, breast cancer cells can induce high expression of OSM by tumor associated neutrophils through GM-CSF signaling and this OSM expression, in turn, results 14 in increased VEGF production by the tumor cells which also become more invasive (Queen et a!., 2005). The related IL-6 family members OSM and LIF are also able to induce the expression of IvIIVIP 13 and tissue inhibitor of metalloproteinase (TIMP)- 1, but not MMP2 and 1v11V1P9 in osteoblast cells (Varghese et a!., 1999) indicating that OSM and MIvIP13 could be a pathway involved in metastatic breast cancer colonization of the bone. 1.8 Objectives and hypothesis In this thesis, I wanted to study the interactions between breast tumor cells and osteoblasts under conditions that approximate a bone microenvironment. Thus, I initially co cultured the two cell types under conditions in which osteoblastic differentiation was stimulated. My first experimental objective was to use these culture methods to assess the morphological changes that occurred in the tumor cells upon exposure to a bone-like microenvironment. My second objective was to determine what changes in gene expression occurred in the bone cells upon co-culture with metastastic tumor cells. I used two different microarray platforms (CLIP-CHIPTMand Affimetrix®) to assess significant gene expression changes across the entire genome. My hypothesis was that the breast tumor/osteoblast interactions in a bone-like microenvironment would initiate changes in gene expression in both cell types that have functional consequences for breast cancer metastasis to bone. Focusing on the changes in osteoblasts, I found that breast tumor cells increased the expression of MMP 13 and OSMR, and that this effect was potentially mediated, at least in part, through tumor produced OSM. Therefore, breast tumor cells may be capable of initiating protein degradative changes in the bone microenvironment by acting directly on the osteoblastic cell population. 15 CHAPTER 2: MATERIALS AND METHODS 2.1 Tissue culture materials and methods 2.1.1 Cell culture MC3T3 cells were maintained in Minimum Essential Media, alpha modification (aMEM, Sigma, St. Louis, MO), supplemented with 10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA) and 50 tg!m1 gentamycin sulfate (Sigma). MDA-MB-23 1 (MDA231) and MDA-MB-231-1833/TR (1833/TR) cells were maintained in DMEM/F12 (Sigma) supplemented with 10% FBS (Invitrogen) and 50 tg/ml gentamycin sulfate (Sigma). All cells were cultured at 37°C in a 5% CO2 humidified incubator. 2.1.2 GFF expression in breast cancer cell lines MDA23 1 cells were stably infected with a GFP-lenti viral vector in the Overall Lab at the University of British Columbia. 183 3/TR cells were stably infected with a thymidine kinase/GFP/luciferase (TGL) lenti viral construct, prior to in vivo selection for reproducible bone homing ability in the Siegel Lab at McGill University (Kang et al., 2003). They were obtained indirectly thru Arun Seth at Sunnybrook and Women’s College Health Sciences Centre, University of Toronto. 2.1.3 Fluorescence Automated Cell Sorting (FACS) Lenti viral-infected cells were maintained under regular growth conditions with no antibiotic selection. Cells were prepared for FAC sorting by trypsinization and resuspension in lmL of FACS buffer (10% FBS, filter sterilized phosphate-buffered saline (PBS), pH 7.4). MDA23 1 and 1 833/TR cells were sorted on a FACS Aria cell sorter (BD Biosciences, Mississauga, ON) for GFP expression thereby enriching for cells with high GFP expression. Uninfected MDA23 1 cells were used as a negative control. 16 2.1.4 MC3T3 dfferentiation in 2- and 3-dimensions In 2-dimensional monolayer cultures, MC3T3s were seeded at a density of io cells per cm2 in 6 well plates in regular growth media. For 3-dimensional culture models, Gelfoam gelatin sponges (Pharmacia & Upjohn, Mississauga, ON) were cut by hand to form approximately 1cm x 0.5 cm x 0.7 cm cubes (0.35 cm3). They were then hydrated in sterile distilled, deionized H2O (ddH2O) for a minimum of 24 hrs. Before use, the water was replaced with serum-free aIVIEM base media for a minimum of 24 hrs and stored in sterile conditions at room temperature until use. MC3T3 cells were seeded at 106 cells per sponge in 50 t1 of regular growth media. Cells were allowed to adhere to sponges for 2 hrs at 37°C before the addition regular serum-containing growth media to cover the sponges. After 24 hrs, regular growth medium supplemented with 100 mlvi 3-glycerophosphate (Sigma) and 50 tM ascorbic acid (Sigma) was added to the cultures (differentiation media). Media was changed every two days. 2.1.5 Co-culture ofMC3T3 osteoblasts and breast cancer cells in 2- and 3- dimensions MC3T3 cells were differentiated for 7 days as described above before the addition of the breast cancer cells, both in 2D monolayer cultures and in 3D sponge cultures. In monolayer, the breast cancer cells were seeded at 1 0 cells per cm2 and were either seeded directly on top of the differentiated MC3T3 monolayer, or on top of a 6 well format 1 tm pore transwell filter (BD Biosciences) in differentiation media. Cells were allowed to adhere for 6 hrs and then both monolayer and filter conditions were rinsed twice with serum-free aMEM base media, followed by addition of serum-free aMEM base media for the remainder of the experiment. MC3T3 monolayers that were exposed to breast cancer cell conditioned medium were rinsed with serum-free aMEM base media prior to treatment with the conditioned medium for the remainder of the experiment (as described). In 3D sponge cultures, breast cancer cells were seeded at 106 cells per sponge in 50 l of differentiation media. Cells were allowed to adhere to sponges for 2 hrs at 37°C before the addition of differentiation media to cover the sponges. 17 2.1.6 Collection and heat inactivation ofconditioned medium Conditioned medium was generated by rinsing cells twice with serum-free cLIVIEM base media to remove serum and differentiation factrs, followed by adding serum-free aMEM base media to dishes for 24 hrs. After 24 hrs, the conditioned medium was collected and filtered through low-protein binding 0.2 m pore syringe filters (PALL Corporation, East Hills, NY) to remove any cells or cell debris. Conditioned medium was then used either immediately or aliquoted and stored at -20°C. Conditioned medium was heat- inactivated at either 65°C or 90°C for 20 mm. Media was cooled to 37°C before use in cell culture treatments. 2.1.7 Transwell migration assay 1 833/TR cells were trypsinized and resuspended in serum-free uMEM base media at iø cells per mL. 200 tl of cells were then seeded (20,000 cells per filter) in 24-well format, 8 m pore size filters (BD Biosciences). 500 pl of the appropriate medium was placed below the filters in the wells and the cells were allowed to migrate for approximately 6 hrs. After 6 hrs, the tops of the filters were swabbed using a cotton swab to remove any cells that did not migrate through the pores. Filters were then fixed in absolute methanol at -20°C for 15 mm. Filters were stained using a 0.5% crystal violet (Sigma) solution in 70% methanol for 15 mm at room temperature, rinsed twice with distilled H20 (dH2O) and allowed to air- dry before imaging and quantification. Quantification involved counting 5 fields at 100X magnification (approximately the whole filter) for each condition in triplicate and averaging them and calculating the standard deviations. 2.1.8 LY294002 treatment MC3T3 cell monolayers were differentiated for 7 days and then treated with either serum-free media, 10 ng/ml OSM, or freshly isolated 1 833/TR-conditioned medium with or without LY294002 (Sigma). Briefly, monolayers were rinsed twice with serum-free aMEM base media to remove serum and differentiation factors followed by pre-treatment with 5 iM LY294002 for 30 mm in serum-free aMEM. Monolayers were then rinsed twice with 18 serum-free aMEM base media to remove the pre-treatment LY294002, and then treated with 10 pM LY294002 or 1:5000 DMSO (vehicle control) added to either serum-free, 10 ng/ml, or 1 833/TR CM. Protein samples were collected at 10 mm for Western blotting and RNA samples were collected at 24 hrs for qPCR analysis (described below). 2.2 Cell and tissue staining materials and methods 2.2.1 Alkaline phopshatase and von Kossa staining Cells grown in monolayer were fixed in 10% neutral buffered formalin (NBF, Fisher Scientific, Fair Lawn, NJ) for 2 hrs, and then subsequently incubated with alkaline phosphatase substrate solution for one hour at room temperature. The substrate solution contained 0.4% N,N-dimethylformamide (DMF), 0.1 mg/mi naphthoi AS-MX (Sigma), 0.6 mg/ml Fast Red Violet LB Salt (Sigma), and 50% 0.1 M Tris (pH 8.3) in dH2O. Subsequent to ALP staining, monolayers were incubated with a 2.5% silver nitrate (Sigma) solution for 1 hr at room temperature with protection from light. Monolayers were then rinsed twice with dH2O and allowed to air dry before imaging. 2.2.2 Gelfoam tissue staining Formalin-fixed, paraffin-embedded Gelfoam samples were sectioned into 5 urn slices on Superfrost/plus slides (Fisher Scientific). Sections were dried onto slides overnight at 50°C and then cooled to room temperature before staining. The Gelfoam sections were de-paraffinized by incubating the slides in xylene (Fisher Scientific) for 3 X 5 mm. Gelfoam sections were re-hydrated in absolute ethanol for 2 X 3 mm, followed by one 3 mm wash in 95% ethanol and one 3 mm wash in 70% ethanol. Gelfoam sections were then rinsed in ddH2O for 5 mm. Following staining, the sections were dehydrated in 2 X 10 sec washes in 95% ethanol, one 10 sec wash in absolute ethanol and one 1 mm wash in absolute ethanol. Gelfoam sections were then washed for 2 X 3 sec in xylene followed by one final 3 mm xylene wash. Gelfoam sections were mounted using Permount (Fisher Scientific). 19 2.2.3 Hematoxylin and eosin Gelfoam staining Following de-paraffinization and re-hydration, Gelfoam sections were stained in Gills Hematoxylin Stain No. 2 (Fisher Scientific) for 3 mm, followed by a brief rinse in dH2O. Hematoxylin stain was acidified by 10 brief dips in an acid alcohol solution (1% 1 OM HCI in 70% EtOH) and neutralized with LiCO3 (Fisher Scientific) for 90 sec and rinsed in running tap water for 2 mm, followed by 2 mm in dH2O. Slides were then incubated with Eosin-Y (Fisher Scientific) for 2 mm followed by dehydration and mounting as described above. 2.2.4 von Kossa Gelfoam staining Following de-paraffinization and re-hydration, Gelfoam sections were incubated with 2.5% silver nitrate solution for 1 hr with protection from light. Sections were rinsed with running water for 1 mm, followed by dH2O and then counterstained briefly (2 dips) with Gills Hematoxylin Stain No. 2. Sections were rinsed in running tap water for 5 mm followed by dehydration and mounting as described above. 2.2.5 Immunohistochemistry Following de-paraffinization and re-hydration, antigen retrieval was performed using citrate buffer, pH 6.0 (18 ml of 0.1 M citrate acid in 82 ml of 0.1 M sodium citrate, up to 1 L with dH2O), pre-heated to 90-98°C in a Coplin jar incubated for 30 mm in a steamer. Slides were incubated in the citrate buffer-filled Coplin jar and steamed for 30 mm. The Coplin jar was then removed from the steamer and allowed to cool to room temperature for 20 mm. The slides were rinsed with PBS for 3 X 5 mm. Endogenous peroxide activity was blocked by incubating the sections with 3% hydrogen peroxide for 10 mm. The slides were rinsed 3 X 5 mm with PBS and excess buffer was blotted using a kimwipe. The primary antibody dilutions were prepared in serum-free protein block solution (DAKO EnVisionTM + System, Troy, MI), were applied to the tissue and incubated in a sealed immunochamber overnight at 4°C. To detect human 20 podocalyxin, mouse monoclonal anti-podocalyxin 3D3 antibody (Kershaw et al., 1997) was diluted to 1:1000. To detect human Ki67, rabbit monoclonal anti-Ki67 antibody (Lab Vision Corporation) was diluted to 1:200. After primary antibody incubation, tissue sections were washed with PBS, 3 X 5 mm. Secondary antibody conjugated to horseradish peroxidase (HRP)-labelled polymer (DAKO EnVisionTM + System) was then applied to the tissue and incubated in a sealed immunochamber for 30 mm at room temperature followd by 3 X 5 mm washes with PBS. The HRP signal was then developed using Nova RedTM reagents prepared according to the manufacturer’s protocol and applied to the tissue for approximately 3-5 mm. Tissue sections were then rinsed 2 X 5 mm with dH2O and counterstained with Mayer’s hematoxylin (Fisher Scientific). The slides were washed with cold tap water for 10 mm and dehydrated as described above. 2.2.6 Immunocytochemistry and microscopy For fluorescence based cell staining, cells attached to 18 mm circle glass coverslips (Fisher Scientific) were fixed at room temperature in 4% paraformaldehyde (PFA, Electron Microscopy Sciences, Hatfield, PA) in PBS for 15 mm. Coverslips were washed for 2 X 5 mm PBS with gentle shaking and extracted using a 0.1% Triton Tx-100 (Fisher Scientific) solution for 15 mm at room temperature. Coverslips were washed 2 X 5 mm in PBS with gentle shaking and subsequently incubated with 50 il of a 1:250 dilution of rhodamine phalloidin (Molecular Probes, Eugene, OR) in 1% BSA in PBS for 20 mm. Following 2 X 10 mm washes in PBS with gentle shaking, coverslips were incubated with 1 ml of a 1:10000 dilution of 4’ 6-diamidino-2-phenylindole (DAPI, Sigma) in 1% BSA in PBS for 90 sec. Coverslips were washed 3 X 10 mm with PBS while gently shaking and then mounted with DABCO (95% glycerol, 2.5% l,4-diazabicyclo[2.2.2]octane (DABCO, Sigma) in PBS) on glass slides. Imaging was performed using the 60X oil immersion objective of an Olympus FV 1000 confocal microscope. 21 2.3 RNA isolation and expression analysis materials and methods 2.3.1 Total RNA lysis All RNA samples were colleted using the QIAGEN RNeasy RNA kit (QIAGEN Inc., Mississauga, ON) plus the QlAshredders (QLkGEN Inc.) for initial homogenization of samples. Samples were collected as per manufacturer’s instructions including the optional DNase I treatment (QIAGEN Inc.). In the final step, the samples were eluted in 30 tl RNase-free H20 and then analyzed for concentration and quality on a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE) before use. 2.3.2 cDNA archiving and quantitative PCR (qPCR) RNA was analyzed using the Applied Biosystems two-step reaction method. For step one, RNA was converted to cDNA using the cDNA archiving kit (Applied Biosystems, Foster City, CA) as per manufacturers instructions, but with a total reaction volume of 20 tl. A total of 2 pg RNA was converted to eDNA in a 20 tl reaction to yield a 100 ng4tl cDNA sample. The cycling parameters were 10 mm at 25°C, 120 mm at 37°C and, 5 sec at 85°C. In step two, 20 ng of cDNA template was used in each reaction with the appropriate primers for quantitative PCR using the 7500 Fast Real-Time PCR System (Applied Biosystems). For each gene of interest, a standard curve was generated from a common pool of cDNA samples and was specific to each experiment. TaqMan® Ribosomal RNA Control Reagent (Applied Biosystems) was used as the 1 8s internal control using the 3’ and 5’ primers at 60 nM and the probe at 100 nM final concentrations. The primers for each gene of interest were developed TaqMan® Gene Expression Assays (Applied Biosystems) that came pre-mixed as a 20X stock solution of both 3’ and 5’ primers and probe. Expression analysis kits used were for mouse MMP13 (Mm00439491_ml), mouse MIvIP14 (Mm00485054_ml), mouse MMP2 (Mm00439508_ml), mouse BSP (Mm00492555_ml), mouse Osx (Mm00504574_ml), mouse OSM (Mni0l 193966_mi), mouse OSMR (Mm00495424_ml), mouse gpl3O (Mm00439668 ml), human OSM (HsOOl7l 165_mi), and human OSIVIR (Hs003 84278_mi). All samples and standards were run in duplicate with a total reaction 22 volume of 15 l. The cycling parameters were 2 mm at 50°C, 10 mm at 95°C for the initial set-up, followed by 40 cycles of 15 sec at 95°C and 1 mm at 60°C for the denaturing and extension/annealing. All qPCR results were graphed by first normalizing gene expression quantities to their corresponding 1 8s internal control quantity. Relative RNA levels were then compared back to the average relative quantity of the first time-point or serum-free control condition. All error bars represent standard deviation. 2.3.3 CLIP-CHIPTMexpression analysis Charlotte Morrison in the Overall Lab at the University of British Columbia performed all CLIPCHIPTM expression analysis. Reverse transcription, purification, concentration, fluorescent labeling, hybridization, scanning, and analysis were performed as previously described (Kappelhoff and Overall, 2007). 2.3.4 AfJj’metrix GeneChip® expression analysis Hybridization was to the GeneChip® Mouse Genome 430 plus 2.0 array (Affymetrix mc, Santa Clara, CA) using a 1 cycle labeling protocol. The Genome Sciences Centre at the University of British Columbia performed expression analysis as follows (Qian, 2008). Calculation of statistical parameters was done using the MAS 5 algorithm for quality control, including scale factors, average background, and percentage of number of genes present and met the current Affymetrix® guidelines. Probe level intensities were normalized using quantile normalization across arrays and gene expression intensity values were estimated using the Robust Multichip Average (RMA) algorithm. Differentially expressed genes were calculated by comparing the MC3T3 samples treated with 1 833/TR cells on filters or 1 833/TR conditioned media with the MC3T3 samples in serum-free media (baseline). The samples treated with 1 833/TR cells on filters and with 1 833/TR conditioned media were also compared to each other. The sample size was n3. To identify differentially expressed genes, group comparison analyses were performed using empirical Bayes moderated t-statistics, and adjusted p-values were calculated using the Benjamini and 23 Hochberg approach to account for type I error rate and control the false discovery rate (FDR). Differentially expressed genes were defined by adjusted p-values 0.05 and the fold difference in mean expression 12.0I. Gene Ontology (GO) term enrichment analyses were used to determine clusters of upregulated signaling pathways. Fisher’s exact tests were used to determine which genes in a particular GO category deviate from the expected expression values. 2.4 Protein isolation and analysis materials and methods 2.4.1 Whole cell protein lysis Thirty five mm dishes of confluent cells were rinsed three times with serum-free cLIVIEM base media and lysed in 50 il of RIPA lysis buffer (50 mlvi Tris pH 7.5, 150 mlvi NaC1, 5 mM EDTA, 5% Nonidet-P40, 1% sodium deoxycholate, and 0.1% sodium dodecyl sulphate, SDS) containing the protease inhibitors aprotinin, leupeptin, phenylmethyl sulfonyl fluoride, pepA, EDTA, sodium vanadate, and sodium fluoride. Dishes were scraped using a sterile cell lifter and transferred to 1.5 ml centrifuge tubes on ice. Cells were incubated in lysis buffer on ice for 10 mm and then centrifuged at 14,000 rpm at 4°C for 15 mm. Supernatants were collected and used as whole cell lysates. 2.4.2 Culture supernatantprotein precipitation Serum-free culture supernatants were collected after 24 hrs of co-culture and precipitated with 10% trichloroacetic acid (TCA, Fisher Scientific) on ice for 20 mm in 1.5 ml centrifuge tubes. Samples were then centrifuged at 14,000 rpm at 4°C for 20 mm. The protein pellet was then subsequently washed with 70% EtOH and resuspended in non denaturing sample buffer. 24 2.4.3 Western blotting For whole cell lysates, protein concentrations were determined using Pierce BCA protein assay kit (Pierce, Rockford, IL) according to the manufacturer’s instructions. 20 tg of whole cell lysates, or 20 !l of TCA precipitated supernatants were separated by 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to a polyvinylidene fluoride (PVDF) membrane (Biorad, Herculaes, CA). PVDF membranes were incubated with a blocking solution containing either 5% skim milk powder (SMP) when blotting for Akt and actin, 5% bovine serum albumin (BSA, Fisher Scientific) when blotting for pSer473-Akt, in Tris (pH 7.5) buffered saline-Tween 20 (TBS-T), or the Odyssey® blocking buffer (LI-COP Biosciences, Lincoln, NE) when blotting for MMP 13 and MIvIP2, for 1 hr at room temperature, followed by a brief wash with TBS-T. Membranes were incubated with primary antibody dilutions (prepared according to Table 1) overnight at 4°C. Unbound primary antibody was removed by washing the PVDF membranes 3 X 10 mm with TBS-T. To detect the primary antibody for pAkt, Akt and actin, the PVDF membranes were incubated with HRP-conjugated secondary antibody dilutions, and for detecting the primary antibodies for MIVIP13 and MMP2, PVDF membranes were incubated with Alexa Fluor 680-conjugated secondary antibodies (prepared according to Table 2) for lhr at room temperature. Enhanced chemiluminescence reagents (ECL, Millipore Corporation, Billerica, MA) were applied to detect the HRP signal following exposure of the membrane to Kodak X OMAT film. A Kodak X-OMAT 1000A film processor was used to develop the film. Infrared secondary antibodies were detected and scanned using the Odyssey® Infrared Imaging System (LI-COR Biosciences). 25 Table 1 — Primary antibodies for Western blotting Primary antibody Company Dilution Dilution Buffer MIVIP- 13 Millipore Corporation, 1:400 Odyssey® Mouse anti-MIVIP- 13 Billerica, MA blocking buffer MIVIP-2 Overall Lab, UBC 1:400 Odyssey® Rabbit anti-MMP-2 Vancouver, BC blocking buffer pSer473 Akt Cell Signaling Technologies, Inc 1: 1000 1% BSA in TBS Rabbit anti-pS er473 Akt Boston, MA Akt Cell Signaling Technologies, Inc 1:1500 5% BSA in TBS Rabbit anti-Akt Boston, MA Actin Sigma, 1:500 1% BSA in TBS-T Mouse anti-actin St. Louis, MO + NaAzide 26 Table 2 — Secondary antibodies for Western blotting Secondary antibody Company Dilution Dilution Buffer Goat anti-mouse IgG Jackson Immunoresearch, 1:20000 1% BSA in TBS HRP West Grove, PA Goat anti-rabbit IgG Jackson Immunoresearch, 1:20000 1% BSA in TBS HRP West Grove, PA Goat anti-mouse IgG Molecular Probes, 1:10000 Odyssey® blocking Alexa Fluor 680 Eugene, OR buffer Goat anti-rabbit IgG Molecular Probes, 1:10000 Odyssey® blocking Alexa Fluor 680 Eugene, OR buffer 27 CHAPTER 3: RESULTS 3.1 Differentiation ofMC3T3 cells into an osteoblast-like state The immortal MC3T3 cell line, derived from mouse calvaria, consists of pre osteoblastic fibroblasts that differentiate into osteoblasts in medium containing (3- glycerophosphate and ascorbic acid. The latter differentiation factors are required for the MC3T3 cells to produce, deposit, and mineralize a bone-like matrix. Additionally, I seeded MC3T3 cells at high-density (1 O cells per cm2) in 2-dimensional (2D) monolayers to ensure efficient and relatively uniform matrix deposition and nodule formation across the cultures (Figure 1). After 7 days of culture, the MC3T3 cells differentiated into an early osteoblast like state which was indicated by their increased alkaline phosphatase activity (Figure 1A), as well as increased Osterix (Osx) and bone sialoprotein (Bsp) expression RNA in both the control and differentiation treatment conditions (Figure 1 B). Mineralization began robustly at approximately 21 days, specifically in the differentiation media conditions, as shown by von Kossa staining (Figure 1 A). MC3T3 cells also differentiated when they were cultured in 3-dimensions (3D) on the collagen spicules present within Gelfoam gelatin sponge scaffolds (Figure 2). Similar to 2D culture, MC3T3 in 3D culture cells expressed Osx and Bsp in both control and differentiation medium (Figure 2B) while von Kossa staining indicated that they mineralized the scaffold matrix specifically in differentiation media (Figure 2A). 3.2 Development of 2- and 3-dimensional tumor/osteoblast co-culture models An overarching goal of this research was to determine if interactions between breast tumor cells and osteoblasts, independent of osteoclasts, lead to microenvironmental changes that could, potentially, promote bone metastasis. Therefore, I set out to co-culture these cell types in two different ways: in 2D culture to carry out gene expression and biochemical analyses; and in 3D culture to make it possible to subsequently carry out functional analyses 28 Figure 1 — Differentiation of MC3T3 cells in 2-dimensional monolayers MC3T3 cells were seeded at a density of i0 cells per cm2 and cultured in either oMEM with 10% FBS and gentamycin (Control) or additionally supplemented with 1- glycerophosphate and ascorbic acid (Differentiation) for a total of 28 days. (A) Monolayers were fixed and stained for alkaline phosphatase activity (red) and mineralization (black) at days 0, 7, 14, 21, and 28. Note that monolayers treated with either control or differentiation media both showed an increase in alkaline phophatase activity over time, but mineralization only occurred in cultures supplemented with differentiation factors. (B) RNA was isolated at days 0, 7, 14, 21, and 28 and used for qPCR analysis for the differentiative markers mouse Osterix (Osx) and Bone Sialoprotein (Bsp). Note that both Osx and Bsp expression levels were markedly increased after 7 days of culturing in control and differentiation media indicating that the monolayer cultures were in an early osteoblast state. RNA was analyzed in duplicate with n2 samples per condition, and the above graph represents one trial of three individual experiments. 29 A Control Differentiation B Control Differentiation Mouse Osterix (Osx) O5I i1[11[ Day 0 Day 7 Day 14 Day 21 Day 28 35 43 Sialoprotein (Is IL Day 0 Day 7 Day 14 Day 21 Day 28 00 C” 30 Figure 2 — Differentiation of MC3T3 cells in 3-dimensional Gelfoam scaffolds MC3T3 cells were seeded at 106 cells per sponge and were cultured in either control or differentiation media for up to 7 days. (A) 3D cultures were formalin fixed on days 0, 4, and 7 for histological analysis. Formalin-fixed, paraffm embedded Gelfoam sponges were sectioned and stained using hematoxylin and eosin (H&E) staining and von Kossa (VK) staining for mineralization. 3D cultures treated with differentiation media exhibited mineralization at days 4 and 7 (arrowheads), whereas the control cultures exhibited no mineralization. Scale bar — 100 tm. (B) RNA was isolated from Gelfoam sponges on days 0, 4, and 7 and used for qPCR analysis of Osx and Bsp expression. Both Osx and Bsp expression increased after 4 days, and continued to increase after 7 days in both control and differentiated cultures indicating that the MC3T3s were in an early osteoblast state. Bsp levels were markedly increased in the cultures treated with differentiation media. RNA was analyzed in duplicate with n2 samples per condition, and the above graph represents one trial of three individual experiments. 31 A Control H&E VK Differentiation H&E VK Mouse Osterix (Osx) RI Ii DayO Day4 Day7 B 14 12 10 8 : Control • Differentiation Mouse Bone Sialoprotein (Bsp) DayO Day4 Day7 32 under conditions that would more closely mimic the in vivo situation. The methodological design of these two co-culture systems is shown schematically in Figure 3. For 2D co-culturing, MC3T3 cells were differentiated as described above for 7 days prior to interaction with tumor cells. Both MDA-MB-23 1 (MDA23 1) and MDA-MB-23 1- 1833/TR (1833/TR) tumor cells were used in these co-culture studies. MDA231 cells are a metastatic human breast cancer cell line and 1 833/TR cells are a sub-population of MDA23 1 cells selected in vivo for their propensity to metastasize specifically to bone (Kang et al., 2003). 2D co-cultures were maintained in three subtly different ways (Figure 3A): with the two cell types in direct contact (Figure 3A left branch) to test for changes produced by cell-cell interaction and/or soluble paracrine factors produced within the co-culture (ie. both two-way “direct” and “soluble” signaling); with the two cell types separated by filters (Figure 3A middle branch) to test for changes caused by the production of soluble factors within the co-culture (ie. two-way soluble signaling only); and by swapping conditioned medium (Figure 3A right branch) between cell types cultured in separate dishes to test for changes in one cell type caused by factors produced by the other cell type (ie. one-way soluble signaling only). For direct contact co-cultures, GFP-labeled MDA23 1 or 1 833/TR tumor cells were seeded directly on top of 7-day differentiated MC3T3 osteoblastic monolayers (Figure 3A left). The tumor cells adhered to the osteoblastic monolayers within 6 hrs (Figure 4A) and they were identified with certainty using epifluorescent microscopy (Figure 4B) to show the presence of their GFP that is stably expressed in these cells (See materials and methods for details). To isolate cell-type specific RNA for microarray analysis after direct co-culture, the overlaid monolayers were trypsinized and FAC sorted based on GFP expression to separate the tumor population (GFP-positive) from the MC3T3 population (GFP-negative). Alternatively, intact co-cultures were lysed and qPCR analysis with species-specific primers was carried out (see below). For across-filter 2D co-cultures (Figure 3A middle), MC3T3 cells were differentiated for 7 days in the bottom of tissue culture dishes and then either MDA23 1 or 1 833/TR tumor cells were seeded on top of 1 tm pore size filters placed in the same dish. These filters allowed for exchange of media and soluble molecules between the two cell types (ie. two-way indirect, paracrine signaling) in the absence of direct cell contact. 33 Figure 3 — Schematic of 2- and 3-dimensional culture models (A) Schematic representation of the 2-dimensional culture model. Naïve MC3T3 pre osteoblasts were seeded at high density and cultured with differentiation supplements for 7 days. After 7 days, MC3T3 cells were then co-cultured in one of three ways: with tumor cells either in direct contact (Direct Contact — two-way “direct” and paracrine signaling), on top of 1 tm pore filters (Filter — two-way paracrine signaling), or by swapping conditioned medium from the opposite cell type (Conditioned Medium — one-way paracrine signaling). (B) Schematic representation of the 3-dimensional culture model. Naïve MC3T3 cells were seeded at high density onto hydrated 3D Gelfoam scaffolds and then cultured with differentiation supplements for 7 days. After 7 days, 106 tumor cells were added to the cultures and allowed to interact for 24 hrs prior to formalin fixation or RNA isolation. 34 A 2D co-culture schematic I___________________________ 7 days differentiation I I I I Conditioned Medium MC3T3 I Tumor Cells I uI 35 Direct Contact Filter B 3D co-culture schematic MC3T3 Tumor Cells Figure 4 — 2-dimensional co-culture with MDA23 1 cells in direct contact with differentiated MC3T3 monolayers MC3T3 cells were plated at 1 O cells per cm2 and differentiated for 7 days in differentiation medium. MDA23 1 cells were then plated directly on top of the MC3T3 monolayers at 1 O cells per cm2. (A) Phase microscopy images (lOx magnification) showing confluent, differentiated MC3T3 cells alone (left image), and co-cultured with MDA23 1 cells in direct contact for 6 hrs in differentiation media (pre-wash, middle image), and following washing to remove serum (post-wash, right image). (B) GFP expression of MDA23 1 cells was visualized using epifluorescent microscopy (20x magnification) after 24 hrs of co-culture. MC3T3 cells did not express GFP (left image) and therefore GFP-positive MDA23 1 cells were clearly identified alone (middle image) and when in direct co-culture with MC3T3 cells (right image). 36 AMC3T3 MC3T3 + MDA23I Pre-wash MC3T3 + MDA23 1 Post-wash B MC3T3 MDA23I MC3T3 + MDA23 1 •i .J.__. ‘I.. .. . c7rJ. -$4 .. . - 4 ....,‘ — .4 s__ .: :: r i:’ !4 i .4 V4. •.3t’.2 37 Expression analysis was then carried out using RNA isolated directly from the separate populations on the bottom of the dish and on top of the filter insert (see below). For 2D conditioned media (CM) experiments (Figure 3A right), CM was collected from 7-day differentiated MC3T3 monolayers and MDA23 1 or 1 833/TR cells that were seeded at the same high density as in the filter and direct co-cultures (1 cells per cm2). This CM was then added to the other cell type and changes in gene expression due to one- way paracrine signaling were assessed using RNA isolated directly from the cells maintained in the separate dishes (see below). For 3D co-culture experiments, MC3T3 cells were allowed to attach to the spicules of Gelfoam gelatin sponges and then cultured for 7 days prior to the loading of either MDA23 1 or 1 833/TR tumor cells into the sponges (Figure 3B). Alternatively, tumor cells alone were loaded directly into the sponges. 24 hrs after tumor cell loading, the cultures were formalin-fixed followed by paraffin embedding, sectioning and staining. Immunostaining for endogenous podocalyxin, a cell surface glycoprotein marker that is highly expressed in MDA23 1 cells (Somasiri et al., 2004), was used to distinguish tumor cells from MC3T3 osteoblasts in these 3D cultures (Figure 5). Both MC3T3 osteoblasts and MDA23 1 tumor cells were able to survive and proliferate when maintained either individually or together in the 3D scaffolds as demonstrated by Ki67 staining (Figure 5). Therefore, this novel culture system can be used in the future to assess the functional effects of tumor and osteoblastic cells on each other in 3D culture. 3.3 MC3T3 conditioned medium induces migration and actin reorganization in 1833/TR breast cancer cells Fibroblast cells are a major component of the stromal microenvironment and since fibroblast-conditioned medium (Clvi) can induce tumor cell migration, I chose to assess whether the CM generated by MC3T3 osteoblasts, which share a number of characteristics with stromal fibroblasts, could induce the migration of 1 833/TR tumor cells. Furthermore, the ability of the MC3T3 CM to induce phenotypic changes in the tumor cells would suggest that these co-culture methods could be used to study the interactions between tumor cells and cells of the bone microenvironment. 38 Figure 5 — 3-dimensional co-culture with MDA231 cells and differentiated MC3T3 cells in Gelfoam sponges Both MC3T3 and MDA23 1 cells were seeded in 3D sponge scaffolds at 106 cells per sponge either alone or in co-culture. MC3T3 cells were differentiated in Gelfoam sponges for 7 days, where as MDA23 1 cells were cultured in sponges for only 24 hrs prior to fixation. For co-cultured sponges, 106 MDA23 1 cells were added to pre-differentiated MC3T3 3D cultures for 24 hrs. 3D sponges were formalin-fixed and embedded in paraffin for sectioning and staining. Staining for Ki67, a marker of proliferation, showed that both MC3T3 (top panel, left) and MDA23 1 (top panel, middle) cells that were cultured alone in 3D sponges were able to proliferate. Note that the co-cultures of MC3T3 and MDA23 1 cells were also positive for Ki67 (top panel, right, solid arrows) indicating that both cell types were proliferating when co-cultured in the sponges. Podocalyxin staining was used to identify the two different cell populations. MC3T3 cells were negative for podocalyxin expression (bottom panel, left), where as IvIDA23 1 cells stained highly positive for podocalyxin (bottom panel, middle). Staining co-cultures for podocalyxin expression showed the presence of both cell populations as unstained MC3T3 cells (hatched arrow) and positively stained MDA23 1 cells (solid arrow). Scale bar — 100 tm. 39 M C3 T3 + M D A 23 1 M D A 23 1O nl y M C3 T3 O nl y C) I C 1 833/TR cells migrated in response to CM derived from both MC3T3 cells alone and MC3T3-1833/TR co-cultures (Figure 6A), indicating that there are, indeed, soluble molecules present in the CM that can induce migration. There was no discernible difference between MC3T3 only CM and either contact or filter co-culture CM. Therefore, a one-way indirect paracrine signaling pathway from osteoblasts appears to be sufficient to initiate breast tumor cell migration. To determine if the paracrine signal that induced tumor cell migration was generated by a factor, or factors, produced specifically by MC3T3 osteoblasts, conditioned medium from Swiss 3T3 (Sw3T3) non-osteoblast fibroblasts was used for comparison (Figure 6B). The Sw3T3 CM was able to induce 1833/TR tumor cell migration at a similar level to the MC3T3 CM positive control. However, the MC3T3 osteoblast CM was considerably more resistant to heat-inactivation than was the Sw3T3 CM. Therefore, it is likely that there are different, heat stable factors produced by osteoblasts that are not produced by fibroblasts (Figure 6B). I also examined the actin cytoskeleton of the 1 8331TR tumor cells upon exposure to CM as another phenotypic readout to assess if these 2D co-culture methods were suitable for identifying tumor-bone interactions. Most notably the MC3T3 CM resulted in the formation of f-actin-rich lamellapodia-like structures at the leading edges of 1 8331TR cells (Figure 7A). This type of actin reorganization was consistent with the increased migratory response elicited by the MC3T3 CM seen in Figure 6. This tentative conclusion was further supported by the observation that heat-inactivation of the MC3T3 osteoblast CM ameliorated the actin re-organization and leading edge ruffling. Treatment with Sw3T3 fibroblast CM caused some reorganization of the actin cytoskeleton but many fewer cells exhibited as marked an increase in lamellapodia-like structures (Figure 7B). Heat- inactivation of Sw3T3 CM resulted in an increase in the presence of stable actin stress fibers in the tumor cells. Additionally, there was an apparent retraction of the cell membranes, which in combination with increased stable stress fibers may explain why there was less response in the migration assay. Taken together, the changes observed using these two phenotypic endpoints demonstrated that the 2D co-culture methods used here should lead to gene expression alterations in each cell type. 41 Figure 6 — 1833/TR tumor cells migrate in response to MC3T3 osteoblast-conditioned medium (A) Conditioned medium was derived from MC3T3 differentiated monolayers, and MC3T3- 1 833/TR co-cultures in both direct contact and across filter conditions. 1 833/TR cells were placed on top of 8 pm pore filters and the different types of CM were placed in the chamber below. 1833/TR cells migrated in response to both contact and filter MC3T3-1833/TR CM at similar levels to MC3T3 CM. (B) Conditioned medium was derived from MC3T3 differentiated monolayers and Sw3T3 fibroblasts, and subjected to heat-inactivation at 65°C for 20 mill prior to use in the migration assay. 1833/TR cells were placed on top of 8 tm pore filters and the different types of CM were placed in the chamber below. 1 833/TR cells migrated in response to MC3T3 CM and their migration was markedly decreased by heat inactivation. 1 833/TR cells also migrated in response to Sw3T3 fibroblast derived CM, but this migration response was not as robust in comparison to the MC3T3 CM positive control media, and was severely attenuated by heat-inactivation. Migration was quantified with n=3 samples per condition, and the above graph represents one trial of two individual experiments. 42 A 400 C 250 200 150 - 100 L) 50 0 400 C Cl) 250 200 150 • 100 L) 50 0 B :zzJzzzzz:::zzzzzzzz::zzz: Filter •Conditioned Medium DHeat Inactivated CM MC3T3 CM Sw3T3 CMMC3T3 CM Contact MC3T3-1833/TR CM 43 Figure 7 — Treatment with MC3T3-conditioned medium reu1ts in a reorganization of the actin cytoskeleton 1 833/TR cells were treated with conditioned medium for 6 hrs, and then fixed and stained for f-actin. (A) 1 833/TR cells were treated with MC3T3 derived conditioned medium (MC3T3 CM) or MC3T3 CM that was heat-inactivated for 20 mm at 65°C prior to use. 1 833/TR cells treated with MC3T3 CM reorganized their actin cytoskeletons and have increased f-actin-rich lamellapodia-like structures (arrowheads), compared to serum-free (SF), indicative of increased motility. Heat-inactivation of MC3T3-conditioned medium attenuated this effect. (B) 1 833/TR cells were treated with Sw3T3 fibroblast conditioned medium (Sw3T3 CM) or Sw3T3 CM that was heat-inactivated for 20 mm at 65°C prior to use. Cell were fixed after 6 hrs of treatment and stained for actin. Sw3T3-conditioned medium caused some reorganization and but the increase in number of lamellapodia-like structures was less than that seen with the MC3T3-conditioned medium. Heat inactivation of Sw3T3 conditioned medium resulted in an increase in stable actin stress fibers and retraction of the cell membranes. lx zoom scale bar —20 !Im; 2x zoom scale bar — 10 tm. 44 Alx zoom 2x zoom Serum Free MC3T3 CM HI-MC3T3 CM B lx zoom 2x zoom Sw3T3 CM HI-Sw3T3 CM 45 3.4 Focused CLIP-CHIP1M microarray analysis ofMC3T3 osteoblast gene expression in response to co-culturing with metastatic breast tumor cells Proteases and protease inhibitors are important molecules for the formation and maintenance of bone (Page-McCaw et al., 2007). They are also implicated in cancer progression and metastasis (Gonzalez et al., 2007; Martin and Matrisian, 2007; Vizoso et al., 2007). Therefore, my initial microarray analyses were done using the CLIP-CHIPTM microarray that was developed by the Overall Lab at the University of British Columbia to assess RNA levels of the great majority of all the human and murine proteases and many of their modulators (Kappelhoff and Overall, 2007; Overall and Dean, 2006). The first analyses were performed on RNA collected from the MC3T3 osteoblasts that were co cultured in direct contact with either MD23 1 or 1 833/TR cell in an effort to cast a wide net for any and all changes caused by two-way direct and indirect/paracrine signaling effects. These first experiments were viewed as an initial screening tool and statistical significance could not be reached due to the high variability between experiments (data not shown). Initially, I attempted to FAC sort cells before isolating RNA from the two different cell populations after co-culturing. However, the time for processing during FAC sorting led to problems with RNA stability and specificity regardless of the presence or absence of co culturing (Appendix A). Because of this technical limitation, all subsequent CLIP-CHIPTM microarray analyses were carried out on RNA collected from MC3T3 cells co-cultured with MDA231 or 1833/TR cells on filters, or with 1v1DA231- or 1833/TR-CM such that individual cell types could be collected and lysed directly. Therefore, gene expression changes caused by direct cell-cell contact were not assessed in the CLIPCHIPTM screens. This technical limitation did not, however, extend to analysis by qPCR as use of species specific primers allowed for differential expression analysis within a mixed RNA population. Therefore, this direct culture method was utilized in qPCR-based experiments where direct co-culture monolayers were lysed intact. Regardless of the method of co culture, all gene expression levels were compared back to baseline levels in MC3T3 monolayers kept in serum-free medium for the duration of the experiments and expressed as fold change. 46 CLIPCHIPTMbased analysis revealed that a number of genes were significantly up- (red) or down- (green) regulated in MC3T3 cells when tumor cells on filters or tumor cell CM were present (summary in Table 3; for complete data set, see Appendices C-F). The most significantly upregulated gene in both filter and CM conditions for MDA23 1 and 1 833/TR co-cultures was collagenase 3, also known as matrix metalloproteinase (IvilviP)- 13. Co-cultured MC3T3 cells also expressed increased levels of osteonectin!SPARC, an extracellular calcium binding protein involved in the initiation and promotion of matrix mineralization that is highly expressed in primary tumor associated stromal cells (Framson and Sage, 2004). Expression of the genes that code for the complement components C3 and C ira were also upregulated; both are components of the innate immune response (Lacroix et al., 2001; Sahu and Lambris, 2001). Very recently, another group has also determined that osteoblasts express inflammatory cytokines in response to treatment with MDA23 1 tumor cell conditioned media (Kinder et al., 2008). The expression of legumain, a lysosomal endopeptidase highly expressed in colon, prostate and breast tumors and associated with poor prognosis (Murthy et a!., 2005), was also upregulated. Interestingly, the expression of survivin was actually decreased in the co cultured osteoblasts. Survivin inactivates caspases to prevent programmed cell death and is highly expressed in cancer, including colon, lung, breast, brain and melanoma (Sah et al., 2006). This decrease in survivin was coupled with an increase in the expression of the pro apoptotic caspases 4 and 12, suggesting that the breast tumor cells may actually be triggering an apoptotic response in the MC3T3 osteoblast cells. 3.5 Matrix metaioproteinase 13 is induced in MC3T3 osteoblasts by tumor cells at both the transcript andprotein levels There was a highly significant upregulation of MIVIP 13 gene expression in all four co-culture conditions assessed by CLIP-CHIPTM analysis (Table 3). MMP2 and MIVIP14 were two proteases whose expression levels did not change significantly and thus, they were used as negative controls for validation by qPCR using species-specific primers (Figure 8). qPCR confirmed that MIvlPl3 was markedly upregulated in MC3T3 cells when they were co-cultured with either MDA23 1 and 1 833/TR cells on filters or with their CM. MMP2 and 47 Table 3 — CLIP_CHIPTM microarray analysis for proteases and protease modifiers MC3T3 + 1833/TR I MC3T3 + MDA231 Differentially Expressed mRNA ICondjtjoned Conditioned in MC3T3s Filter Medium Filter collagenase 3 (IV1Iv1P- 13) complement component 3 PHEX endopeptidase aptoglobin-1 )roteasome catalytic subunit 3i caspase-4/1 1 carboxypeptidase X2 roteasome catalytic subunit 2i complement component Clra osteoblast serine protease cytosol alanyl aminopeptidase ransferrin receptor 2 protein caspase- 12 sparc/osteonectin, testican-2 legumain survivin nesoderm-specific transcript Note: red = upregulated, green = down-regulated. Above represents a summary of significant hits with p <0.05 and 12.0I fold changes in expression. Charlotte Morrison in the Overall Lab at the University of British Columbia performed all CLIP-CHIPTM analyses which were done in duplicate with n=3 for 1833/TR filter and CM, and n=2 for MDA231 filter and CM. For complete data set see Appendices C-F. 48 Figure 8 — Validation of increased MMP13 expression in MC3T3 cells upon exposure to breast tumor cells cDNA samples used for CLIPCHIPTM analysis were also used for qPCR for mouse MMP 13 expression to confirm the increase in expression. Expression levels for mouse MIVIP2 and MMP14 were used as non-changing controls. Note that qPCR analysis confirmed that IvllvIPl3 expression was markedly upregulated in MC3T3 cells when they were co-cultured with either MDA23 1 or 1 833/TR breast tumor cells across filters or when treated with their conditioned medium. In contrast, neither IvllvlP2 nor MMP 14 expression levels changed under any treatment conditions. All relative expression levels were normalized back to 1 8s expression internal control for each sample and replicate, and then compared to the serum-free baseline expression level. RNA was analyzed in duplicate with n=3 samples per condition used for the CLIPCHIPTM analysis. 49 MsMMPI3 • MsMMP2 D MsMMP14 D Filter Co-culture Conditioned Medium 14 14 __________________________________________________________ __________________________________________________________ MC3T3 MC3T3 MC3T3 MC3T3 MC3T3 MC3T3 (Serum Free) (MDA231) (1833/TR) (Serum Free) (MDA23I) (1833/TR) 50 MIVIP14 levels did not change as assessed by qPCR, consistent with the CLIP-CHIPTM microarray data. By CLIP-CHIPTManalysis, the fold expression increase of MIvIP 13 was highest in the CM treatments (Table 3), which suggests that there may be an inhibitory feedback loop that is only present in the two-way paracrine signaling filter co-culture method. In order to determine if direct contact (ie. two-way direct and paracrine signaling) altered MIvIP 13 induction, it was important to do qPCR analysis with species-specific primers to avoid FAC sorting of direct contact co-cultures. I also wanted to determine if the IvilviP 13 induction caused by the presence of tumor cells or tumor cell CM was an “all-or-none” effect (ie. maximal induction immediately), or if the level of induction changed over time. I therefore chose to do a time course by qPCR analysis using all three co-culture methods (Figure 9). The induction of MIvIP 13 in MC3T3 osteoblasts by co-culturing with both MDA23 1 and 1 833/TR cells increased over time (Figure 9). The kinetics of the MMP 13 induction in the direct co-cultures were similar to those seen with both the filter co-culture and CM treatments, particularly with the 1 833/TR cells, indicating that direct contact did not enhance this induction. It is therefore likely that a soluble molecule inherently produced by the tumor cells induces MIVIP 13 at the transcript level. For the above MMP 13 transcript induction to be functionally relevant, an increase in MIvIP 13 protein would have to occur, which could then act on the microenvironment. To test for this, secreted MMP 13 protein was examined by Western blotting of the concentrated co-culture supernatants (Figure 10). Supernatants were collected after co-culturing of the MC3T3 osteoblasts with either the tumor cells or tumor cell CM for 24 hours; the supernatants therefore contained all soluble molecules produced by both osteoblast and tumor cells. MMP 13 protein production by MC3T3 osteoblasts was increased in the presence of either MDA23 1 or 1 833/TR tumor cells or their CM, but was predominantly in the pro-MIvlPl3 form as shown by Western blotting for mouse MMP13 (Figure 10). This strongly suggests that breast tumor cells are capable of inducing the production of factors in the bone microenvironment that could lead to the remodeling of that microenvironment in an osteoclast-independent fashion. MIvIP 13 has many substrates, including type II collagen, an important ECM component in endochondral bone formation. In addition, MMP13 plays a functional role in 51 Figure 9 — MMP13 expression is induced in MC3T3 cells in a time-dependent manner and is not enhanced by direct cell-cell interaction MC3T3 cells were co-cultured with 1v1DA23 1 or 183 3/TR cells using direct contact, filter and conditioned medium co-culture methods for up to 48hrs. RNA was collected at 0, 24, and 48 hrs and used for qPCR analysis. MMP 13 expression was markedly increased in all three types of culture conditions with both tumor cell types indicating that the factor responsible for MMP 13 induction is likely to be a soluble molecule present in the conditioned medium and not a cell surface molecule. MMP 13 induction generally increased with time in all three co-culture conditions where as IvllvlP2 and MIvIP 14 expression levels remained the same. RNA was analyzed in duplicate with n2 samples per condition, and the above graph represents one trial of three individual experiments for all conditions except MDA23 1 CM which was one trial of two individual experiments. 52 Ms MMP13 Ms MMP2 Q Ms MMP14 D 4-. z 01 1 833/TR Contact Co-culture 14 . 12 4) 10 c’ < 8 6 2 0 -- Oh 24h 48h 1,1 1 833/TR Filter Co-culture 12 T 14 12 in -- MDA23 1 Contact Co-culture 14 ---- 12 •---- 1O .---- 8 •---- 6 •---- 4 .--.. 2 •---- 0 Oh 24h 48h MDA231 Filter Co-culture 14 1aIi.I Ij=j Oh 24h 48h 4-. 01 1:0 _______________ Oh 24h 48h 1 833/TR Conditioned Media IA MDA23 1 Conditioned Media 01 1.-I. •---- 12 •---- 1O ----- 8 •--•- 6 ---- ‘i-i In IIr 8 6 4 2 0 Oh 24h -:::I’Fi:Iri 48h Oh 24h 48h 53 Figure 10 — MMP13 protein levels increase when co-cultured with MDA231 and 1833/TR cells Mouse MC3T3 cells were co-cultured with (A) human MDA23 1 and (B) 1 833/TR cells for 24 hrs and the culture supernatants were collected and TCA precipitated for mouse MIVIP 13 protein detection by Western blotting. Note that both MDA23 1 and 1 833/TR cells induce increased levels of IvIMP 13 protein when co-cultured in direct contact or across filters, as well as through their conditioned medium indicating that the molecule responsible was likely to be a soluble molecule. The osteoblast-produced MIv1P1 3 in response to tumor cells or their conditioned medium was the pro-MIV1P13 (inactive) form. MMP2 was used as a control blot where protein levels did not change in any of the tumor cell treatments. The blots shown here are a representative trial of three individual experiments. 54 - r M D A 23 1 CM M D A 23 1 M C3 T3 Co nt ac t Fi lte r CM Co nt ro l (D [1 r[ 83 3/T R CM I 8 33 /T R 00 - U I U I the normal maintenance and remodeling of bony trabeculae (Page-McCaw et al., 2007; Stickens et al., 2004). In breast cancer patients with an increased propensity to develop distant metastases, IvIIV1P1 3 expression is upregulated in primary tumor-associated stromal cells (Gonzalez et al., 2007; Vizoso et al., 2007). Therefore, identifying potential factors produced by breast tumor cells that induce IVIMP 13 expression in osteoblasts became the next focus of this research. 3.6 Genome-wide Affrmetrix® analysis of MC3 T3 cell gene expression in response to co-cultureing with bone metastatic 1833/TR breast tumor cells In an effort to determine molecules that may be involved in the tumor cell-mediated MMP 13 upregulation in osteoblasts, I chose to use genome-wide Affymetrix® microarray analysis. Only the 1 833/TR cells were used for this analysis because of their repeatable pattern of bone metastasis. Because of this I reasoned that they should cause gene expression changes in the MC3T3 osteoblasts that were more specific to bone homing tumors than the parental MDA23 1 cell population that metastasizes to sites other than just the bone. Affymetrix® microarray analysis was therefore preformed on RNA collected from the MC3T3 osteoblasts that were co-cultured with either 1 833/TR cells on filters, or with 1 833/TR CM for 24 hours to assess two-way and one-way paracrine signaling effects, respectively. The analyses were performed on RNA isolated from osteoblasts so that the data obtained from these experiments could potentially reveal what molecules are involved in MMP 13 upregulation (ie. what pathway was involved), as well as show other interesting, non-protease proteins whose expression was altered by the presence of tumor cells. All gene expression levels were compared back to baseline levels in MC3T3 monolayers kept in serum-free medium for the duration of the experiment and expressed as fold change. Affymetrix®-based analysis revealed that a number of genes were significantly up (red) or down- (green) regulated in MC3T3 cells when 1 833/TR cells or CM were present (Figure 11; summary in Table 4, for complete data set, see Appendices I-K). Some of the interesting genes that were upregulated in the co-cultured MC3T3 osteoblasts and that could potentially be involved in enhancing the establishment of bone metastases were chemokine 56 Figure 11 — Affymetrix® analysis of MC3T3 cells co-cultured with 1833/TR cells MC3T3 cells were co-cultured for 24 hrs with 1 833/TR cells on filters or with 1 833/TR conditioned medium. RNA samples were collected in triplicate and used for Affymetrix® analysis on the GeneChip® Mouse Genome 430 plus 2.0 array. Differentially expressed genes were defined by adjusted p-values 0.05 and the fold difference in mean expression 12.0I. (A) Changes in gene expression levels in MC3T3 cells co-cultured with 1 833/TR cells on filters as compared to serum-free baseline expression levels are plotted as up- regulated (red) and down-regulated (green) circles. There were 84 up-regulated and 21 down-regulated significant changes in gene expression when the MC3T3 cells were co cultured with 1 833/TR cells on filters. (B) Changes in gene expression levels in MC3T3 cells treated with 1 833/TR conditioned medium as compared to serum-free baseline expression levels are plotted as up-regulated (red) and down-regulated (green) circles. There were 39 up-regulated and 117 down-regulated significant changes in gene expression when the MC3T3 cells were treated with 1 833/TR-conditioned medium. (C) Changes in gene expression levels in MC3T3 cells treated with 1 833/TR conditioned medium as compared to MC3T3 cells co-cultured with 1 833/TR cells on filters expression levels are plotted as up regulated (red) and down-regulated (green) circles. There were only 2 up-regulated and 19 down-regulated significant changes in gene expression between the MC3T3 cells treated with 1 833/TR conditioned medium and the MC3T3 cells co-cultured with the 1 8331TR cells on filters. Analysis was performed at the Genome Sciences Centre, Vancouver, BC. For complete data sets, see Appendices I-K. 57 A I 833/TR Filter vs. SF Baseline based on Adjusted p-value 00- l) z C- B I 833/TR CM vs. SF Baseline based on Adjusted p-value C I 833/TR CM vs. 1 833/TR Filter based on Adjusted p-value z > I 00 -6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6 Log2 Fold Change Log2 Fold Change -6 -4 -2 0 2 4 6 Log2 Fold Change 58 Table 4 — Affymetrix® microarray analysis Differentially Expressed mRNA in MC3T3 I CM vs. SF I Filter vs. SF lipocalin 2 aptoglobin serum amyloid A 3 oncostatin M receptor umor necrosis factor receptor superfamily, member 9 serine (or cysteine) peptidase inhibitor, dade A, member 31S. suppressor of cytokine signaling 3 embigin natrix metallopeptidase 13 carbonic anhydrase 9 B-cell leukemiallymphoma 3 chemokine (C-X-C motif) ligand 12 vascular cell adhesion molecule 1 rofilin 1 Note: red upregulated, green down-regulated. Above represents a summary of significant hits with p 0.05 and ?12.0I fold changes in expression. For complete data sets see Appendices I — K. 59 ligand 12 (CXCL 1 2ISDF- 1), carbonic anhydrase 9 (CA9), tumor necrosis factor receptor (TNFR), and vascular cell adhesion molecule 1 (VCAM- 1). SDF- 1 is particularly interesting because it has the ability to increase the metastatic properties of tumor cells in culture including migration and invasion (Kang et al., 2005) and also plays a role in the establishment of bone metastases with its receptor CXCR4 (Wang et al., 2006). The ability of tumor cells to increase the expression of SDF-1 in the bone microenvironment would be an advantage for colonization because it could increase the recruitment of new tumor cells, enhancing tumor establishment at the distant site. Tumor expression of CA9 is associated with decreased relapse free survival and is an independent prognostic marker in lymph node-positive women (Brennan et al., 2006). Radiotherapy is one of the palliative therapies given to patients with bone metastases and CA9, a hypoxia-induced protein that is involved in the buffering of C02,may confer resistance to radiotherapy (Brennan et al., 2006). Lipocalin is another independent predictor of poor outcome in breast cancer as it is associated with lymph node metastasis and decreased disease-free survival (Bauer et al., 2008). Functionally, lipocalin forms a complex with MIVIP9, protecting MMP9 from degradation (Fernandez et al., 2005). The upregulation of both TNFR and VCAM-1 may indicate that tumor cells can enhance their ability to adhere in the bone microenvironment. TNFcc is able to induce the expression of VCAM- 1 in endothelial cells and breast cancer cells can bind to VCAIVI- 1 (Nizamutdinova et al., 2008). While these changes in gene expression will be important for future studies of bone metastasis, the major purpose for the Affymetrix® analysis was to determine what signaling pathways or other proteins were involved in the upregulation of MIvIP 13 in the MC3T3 osteoblasts. For that I used gene ontology (GO) analysis of the Affymetrix® data. GO analysis assigns genes to functional categories on the assumption that genes involved in a particular active biological process are likely to have the same expression profile (Werner, 2008). This GO analysis technique was used to identifr pathways or proteins with similar expression profiles to MMP 13. GO analysis was done by comparing the IVIMP 13 expression profile in both the 1 833/TR CM and filter conditions versus MMP 13 expression profile in serum-free. GO clustered groups that had similar expression profiles to MMP 13 were ranked in order of statistical significance for both the CM (Table 5) and filter (Table 6) co-cultures. The oncostatin M receptor {3 (OSMR) activity expression profile was 60 Table 5 — Gene ontology analysis top 10 pathways — conditioned medium vs. serum- free Rank inTerm Annotated Significant Expected classic p-value DNA-dependent ATPase 81 9 0.36 1 1.OOE-10 activity structure-specific DNA 118 8 0.52 2 5.90E-08binding single-stranded DNA 83 7 0.37 3 9.1OE-08binding ribonucleoside-diphosphate 10 3 0.04 4 1.OOE-05 reductase activity oxidoreductase activity 10 3 0.04 5 1.OOE-05 oxidoreductase activity 11 3 0.05 6 1 .40E-05 kinase regulator activity 219 7 0.97 7 5.80E-05 oncostatin-M receptor 3 2 0.01 8 5.80E-05 activity nu DNA polymerase 3 2 0.01 9 5.80E-05 activity double-stranded DNA 50 4 0.22 10 7.30E-05binding Analysis shows the different GO categories (“Term” column) and describes how the observed number of genes per term (“Significant” column) deviated from the expected amount (“Expected” column). The “Annotated” column represents the number of genes defined under the given GO category. Ranking was based on p-value. For complete list see Appendix H. 61 Table 6— Gene ontology analysis top 10 pathways — filter vs. serum-free Rank inAnnotated Significant ExpectedTerm classic p-value oncostatin-M receptor 3 2 0.01 1 2.1OE-05 activity unspecific monooxygenase 37 3 0.1 2 0.00014 activity monooxygenase activity 199 5 0.53 3 0.0002 protein dimerization activity 660 8 1.77 4 0.0004 oxidoreductase activity 57 3 0.15 5 0.00049 heme binding 268 5 0.72 6 0.00079 tetrapyrrole binding 268 5 0.72 7 0.00079 protein kinase activator 23 2 0.06 8 0.00174 activity protein homodimerization 322 5 0.87 9 0.00179 activity kinase activator activity 27 2 0.07 10 0.0024 Analysis shows the different GO categories (“Term” column) and describes how the observed number of genes per term (“Significant” column) deviated from the expected amount (“Expected” column). The “Annotated” column represents the number of genes defined under the given GO category. Ranking was based on p-value. For complete list see Appendix G. 62 significantly correlated with the MMP 13 expression profile and was the top activity pathway found in the filter co-culture condition (Table 6). OSMR levels were increased in the genome-wide Affymetrix® analysis (Table 6), and OSIVIR was also in the top 10 signaling pathways for the CM GO analysis (Table 5) indicating that it was a potential candidate for being involved in MJVIP13 induction. The ligand oncostatin M (OSM) is a member of the IL-6 family of cytokines, all of which signal through heterodimeric receptors that use the gp 130 subunit as a common component (Auguste et al., 1997; Lindberg et al., 1998; Rose-John et al., 2006). The OSMRIgp 130 Type II heterodimer complex mediates specificity for OSM recognition, although OSM can also bind to the LIF receptor 3/gp 130 type I heterodimer complex (Auguste et al., 1997; Mosley et al., 1996; Rose-John et al., 2006). Validation of a subset of the gene expression changes putatively identified by the Affymetrix® and GO analysis was done by qPCR using species-specific primers for the MIvIPs, OSMR, gpl3O and the OSM ligand. These experiments confirmed that both MMP 13 and OSMR expression levels were increased in MC3T3 cells when they were co cultured with 1833/TR cells on filters or 1833/TR CM (Figure 12). In contrast, gpl3O, MMP 14 and MtVIP2 expression levels did not change (Figure 12), all of which was consistent with the original Affymetrix® analysis. Species-specific transcripts for the OSM ligand were not detectable in the mouse MC3T3 osteoblasts under any conditions (data not shown). However, species-specific transcripts for both OSM ligand and OSMR were present in human 1 833/TR human breast tumor cells, although the levels of both transcripts did not change in response to co-culturing (Figure 13). Thus, the receptor for OSM-specific signaling is upregulated in osteoblasts after exposure to breast tumor cells and the tumor cells themselves contain transcripts for the OSM ligand. Together, these findings suggest that a functional signaling pathway may be induced under conditions of co-culture. Since MIvIP 13 is involved in the regulation of matrix and matrix-bound cytokine release (Dangelo et al., 2001; McQuibban et al., 2001; Mott and Werb, 2004; Nagase et al., 2006), it is possible that tumor cell production of OSM and the subsequent upregulation of MMP 13 expression by osteoblasts may be a functional osteoclast-independent pathway that could enhance breast cancer colonization of bone. 63 Figure 12 — qPCR validation of increased mouse OSMR expression levels in MC3T3 osteoblasts upon exposure to 1833/TR breast tumor cells MC3T3 RNA used for Affymetrix® analysis was re-analyzed using species-specific qPCR primers for mouse MIVIP13, OSMR, gpl3O, MIVIP14 and MIVIP2. qPCR analysis showed that both MMP13 and OSMR expression levels were markedly increased when MC3T3 cells were co-cultured with 1 833/TR cells on filters, and when treated with 1 833/TR- conditioned medium. As expected the gpl3O, MMP14, and MIVIP2 expression levels remained unchanged which was consistent with the Affymetrix® analysis. The graph shown here represents the n3 samples used for Affymetrix® analysis. 64 • Ms MMP13 Li Ms OSMR I Ms gpl3O El Ms MMPI4 IL Ms MMP2 6 1’ ------ 1+ 1jTT IZ1 MC3T3 MC3T3 MC3T3 (Serum Free) (1833/TR Filter) (1833/TR CM) 65 Figure 13 — Human OSM ligand and receptor expressed by 1833/TR tumor cells RNA from 1 833/TR cells that were co-cultured with MC3T3 cells across filters was used for qPCR analysis of the human OSM and human OSMR expression levels in these cells. Both OSM and OSMR were expressed in the 1 833/TR cells, but their expression levels did not change when co-cultured with MC3T3 cells. RNA was analyzed in duplicate with n=3 samples per condition, and the above graph represents one individual experiment. 66 • Hu OSMR C Hu OSM 2 ‘ 1.2 1833/TR 1833/TR (Serum Free) (MC3T3 Filter) 67 3.7 Oncostatin M is a candidate molecule secreted by breast tumor cells that is capable of increasing the expression levels of both MMP13 and OSMR in MC3T3 osteoblasts OSM is a heat stable protein (Zarling et al., 1986). To determine whether 1833/TR breast tumor cell conditioned medium contains heat stable factors consistent with the presence of OSM, it was heated to 65°C. This did not attenuate its ability to induce MIVIP13 and OSMR expression when added to MC3T3 cells (Figure 14). In contrast, this induction capability was dramatically reduced when the conditioned medium was heated to 90°C, a temperature which has been shown previously to completely denature human OSM (Zarling et al., 1986). Interestingly, the attenuation was not complete at 90°C, suggesting that there may be other even more heat-stable factors present in the 1 833/TR CM involved in the induction of IvIMP13 and OSMR transcript levels in MC3T3 cells. Recombinant human OSM increased the production of MIVIP13 and OSMR transcripts in a dose-dependent manner when it was added directly to MC3T3 osteoblasts (Figure 15). There was a specificity to this induction that was similar to that produced by 1 8331TR CM, given that neither gp 130 nor MMP 14 expression levels changed in response to the recombinant OSM treatment. Furthermore, treatment with 10 ng/ml of recombinant OSM was also able to increase the production of the MIvIP 13 protein to levels similar to those seen when MC3T3 osteoblasts were treated with either the 1 833/TR or MDA23 1 CM (Figure 16). Taken together, these data suggest that recombinant OSM may be acting through the same pathway as the conditioned medium. Unfortunately, however, I have been unable to demonstrate definitively that OSM itself in the conditioned media contributed to the MIVIP 13 and OSMR induction as blocking antibody experiments were inconclusive (Appendix B). Furthermore, while the OSM RNA was present in 1833/TR cells (Figure 13), I was unable to directly demonstrate the presence of OSM in the 1 833/TR conditioned media; This may have been due to inefficient binding of the antibody used as I was able to immunoprecipitate only a small fraction of recombinant OSM when the antibody was in great molar excess (Appendix B). 68 Figure 14 — Heat-inactivation of 1833/TR-conditioned media MC3T3 cells were treated for 24 hrs with 1 833/TR-conditioned media that was heat- inactivated (HI-CM) for 20 mm at 65°C or 90°C. Treatment with 65°C HI-CM induced MIvIP13 and OSMR at similar levels to non-heated (37°C) 1833CM, where as 90°C HI-CM induced MIvIP 13 and OSMR at lower levels indicating that one or more of the factors involved in stimulation is denatured at 90°C. Mouse gp130 and MMP14 were used as negative controls, and neither treatment with 65°C or 90°C HI-CM altered their expression levels. RNA was analyzed in duplicate with n=2 samples per condition, and the above graph represents one trial of three individual experiments. 69 B Ms MMP 13 D Ms OSMR B Ms gp 130 D Ms MMP 14 5 ------ in-. - SF 37°C 65°C 90°C 1833/TR CM 70 Figure 15 — Recombinant oncostatin M increases MMP13 and OSMR expression levels in MC3T3 osteoblast cells MC3T3 cells were directly treated for 24 hrs with increasing concentrations of recombinant human OSM at 0, 0.5, 2, 5, and 10 ng/ml final concentration. 18331TR-conditioned medium (1833/TR CM) was used as a positive control. Both MIVIP13 and OSMR expression levels increased as OSM concentration increased showing that OSM is able to induce IV11VIP 13 and OSMR expression in a dose-dependent manner and at levels similar to 1 833/TR-conditioned medium. RNA was analyzed in duplicate with n=2 samples per condition, and the above graph represents one trial of three individual experiments. 71 •MsMMP13 CMsOSMR •Msgpl3O MsMMP14 25 20 15 1o 5 0 0 0.5 2 5 10 1833/TR CM ng/ml OSM 72 Figure 16 — MMP13 protein expression is increased by oncostatin M MC3T3 cells were treated with MDA231 CM, 1833/TR CM and 10 ng/ml final concentration of recombinant human OSM for 24 hrs. Culture supernatants were then TCA precipitated and subjected to Western blotting for mouse MMP13. OSM treatment resulted in an increase in MIVIP 13 protein production at similar levels to both MDA23 1- and 1 833/TR-conditioned medium. 73 Se ru m Fr ee lO ng /m iO SM j 18 33 /T RC M I I Co nt ro l CD 3.8 Recombinant OSM- and 1833/TR conditioned media-mediated upregulation ofMMP13 and OSMR expression is not dependent upon P13K signaling cascade In chondrocytes, OSM synergizes with IL-i to increase MIvIP 13 expression in a PI3KJAkt-dependent manner (Litherland et al., 2008). Since osteoblasts and chondrocytes are derived from a common precursor I reasoned that OSM might use the same pathway to induce MIvIP 13 in the latter cell types. To test this, LY294002 was added at 10 itM to pharmacologically block P13K in both recombinant OSM- and 1 833/TR CM-treated cells. Western blotting for p5er473-Akt, a downstream readout, indicated that P13K signaling was blocked under these conditions (Figure 17A). Surprisingly, however, qPCR analysis revealed that IVIIvIP 13 and OSMR transcript levels actually increased under these conditions, independent of the presence of recombinant OSM- or tumor cell-conditioned media (Figure 17B). Thus, not only does it appear that P13K is not required for MIVIP13/OSMR induction in osteoblasts but, in fact, there may even be a PI3KJAkt-dependent inhibitory signaling pathway present in these cells. Importantly, neither gp 130 nor MMP 14 transcript levels were significantly altered upon inhibition of PI3K!Akt, indicating some specificity to the effect on IVIMP 13 and OSMR transcription. Taken together, these data indicate that the signaling events that lead to MIvIP13 and OSMR induction in osteoblasts by recombinant OSM- and 1833/TR-conditioned medium are P13K-independent and may differ from those that occur in chondrocytes. 75 Figure 17 — LY294002 treatment does not attenuate MMP13 induction by OSM or 1833/TR-conditioned medium MC3T3 differentiated monolayers were pre-treated with 5 M LY294002 in serum-free medium for 30 mm prior to stimulation. MC3T3 cells were then treated with either serum- free medium (SF), 10 ng/ml recombinant OSM (OSM), or 1833/TR conditioned medium (CM) with or without 10 tM LY294002 (DMSO used as vehicle control). (A) MC3T3 cells treated for 10 mm were lysed for Western blotting of pSer473-Akt, total-Akt and actin to demonstrate Akt inhibition. (B) MC3T3 cells were treated for 24 hrs and then lysed for RNA to use in qPCR analysis for mouse MIVIP13, OSMR, gpl3O, and MIVIP14 expression levels. LY294002 inhibition of Akt resulted in a marked increase in expression levels of both MMP13 and OSMR when added to serum-free medium, 10 ng/ml OSM, and 1833/TR- conditioned medium. MMP14 and gpl3O levels were unchanged upon LY294002 inhibition of Akt in all three conditions. RNA was analyzed in duplicate with n=2 samples per condition, and the above graphs represents one trial of three individual experiments. 76 ApSer473-Akt Akt Actin DMSO LY294002 SF OSM CM 5P (Y” - — — — — — 30 25 20 15 10 5 1 • DMSO []LY294002 Mouse MMP13 Mouse OSMR 3025 -I- 20 --. c --. 15z ______________________________________________ — : ___________________________________________________________ Serum Free OSM 1833CM El B 160 80 60 .40 20 0 1833CM Mouse gpl3O — I _-+i Serum Free OSM Mouse MMPI4 30 ‘ 25 -a 15 10 0 Serum Free OSM 1833CM — — - Serum Free OSM 1833CM 77 CHAPTER 4: DISCUSSION 4.1 2- and 3-dimensional tumor/osteoblast co-culture models In this research I used co-culture models to study the interactions that take place when metastatic breast tumor cells arrive in the bone microenvironment. I also identified specific interactions that may help contribute to the establishment of metastatic disease. My primary focus was the microenvironment, as it is becoming clearer that interactions between tumor cells and the distant site stroma play a significant role in metastatic progression. The MC3T3 cell line was chosen to model the bone microenvironment because of its ability to differentiate reproducibly into an osteoblast-like state in both 2- and 3-dimensional culture (Figures 1 and 2). Initial experiments with the GFP-tagged MDA23 1 breast carcinoma cell line facilitated optimization of the co-culture conditions. Specifically, these tumor cells were able to adhere to MC3T3 osteoblast monolayers with minimal cell loss when serum and differentiation factors were removed for the remainder of the experiment (Figure 4A). Another useful feature of this 2D co-culture method was that the two cell populations could be distinguished based on GFP tagging (Figure 4B), which facilitated the isolation of cell type-specific RNA for post co-culture gene expression analyses. The 3D co-culture methods have the potential to be useful to study functional tumor- bone interactions immunohistochemically when the osteoblasts lay down a mineralized matrix on collagenous spicules and the tumor cells cluster in nests in the free spaces in between the spicules (ie. approximating bone marrow niches). Under these conditions the two different cell populations could be distinguished by staining for distinct proteins such as podocalyxin, which is only expressed by the tumor cells (Figure 5) followed by dual staining with Ki67 for proliferation. In this way, different potential metastastic therapies could be tested in a histotypic culture model with relatively high throughput before undertaking more resource- and time-intensive mouse studies. Specifically, this 3D co culture model can be used to test the efficacy of potential therapeutic agents on the growth and survival of breast tumor cells within a recapitulated bone microenvironment. Metastatic breast lesions are predominantly osteolytic (Blair et al., 2006; Mundy, 1997; Mundy, 2002; Roodman, 2004). Therefore, most studies that have examined breast 78 tumor/bone microenvironmental interactions have focused on the activation state of osteoclasts. As a result, treatments (ie. bisphosphonates) have been developed to suppress osteoclast activity with some symptomatic but very limited curative success (Mundy, 1997; Yoneda et a!., 2000). I chose to focus, instead, on the tumor/osteoblast microenvironment to identify novel interactions that may be important early in the metastatic process, independent of osteoclast activity. However, for the tumor cells/osteoblast co-cultures to be more clinically relevant will ultimately require the addition of osteoclasts to make sure all of the components of the “vicious cycle” of bone metastases are present (Blair et al., 2006; Guise, 2000; Guise and Mundy, 1998; Guise et al., 1996). Such an approach would further facilitate the surrogate in vitro therapeutic agent testing described above, especially given that it will modify signals initiated by interactions between tumor cells and osteoblasts with functional consequences. This may be the case with MMP13 (see below). 4.2 Osteoblast produced factors are capable of increasing migration of tumor cells using a 2-dimensional co-culture model 2D co-culture and conditioned media approaches have been used previously by other research groups to study tumor-stromal interactions in breast cancer (Mercer and Mastro, 2005; Mercer et al., 2004; Queen et al., 2005). In those studies, serum was present during the co-culture or conditioned medium treatments, which could potentially have altered gene expression profiles and cytokine responsiveness. Therefore, in these experiments I eliminated serum, which was particularly important for assessing the migration and cytoskeletal changes in 1 833/TR tumor cells upon treatment with MC3T3 osteoblast conditioned medium. While the 1833/TR cells are metastatic, and intrinsically motile, exposure to MC3T3 conditioned medium resulted in the formation of f-actin rich lamellipodial-like protrusions (Figure 7) and increased their migration (Figure 6). One potential candidate molecule present in the co-culture microenvironment that could be responsible for enhancing this migratory activity is chemokine ligand 12 (CXCL 12/ SDF- 1) given that Affymetrix® analysis indicated that its expression in osteoblasts was increase in the co-cultures (Table 4). SDF- 1 is normally expressed by osteoblasts in order to retain hematopoetic stem cells in the bone marrow and it can induce the directed migration of 79 lymphocytes (Burger and Kipps, 2006; Pelletier et al., 2000; Sun et al., 2005). In carcinomas, SDF- 1 has the ability to increase the migratory and invasive metastatic properties of tumor cells that express its receptor, CXCR4 (Kang et al., 2005; Sun et al., 2005; Wang et al., 2006). Based on a human tissue microarray study, increased expression of CXCR4 correlates with higher tumor grade, poor outcome, HER2 amplification and hormone receptor negative status (Salvucci et al., 2006). Experimentally, HER2 increases the expression of CXCR4 expression in tumor cells which enhances their chemotaxis towards SDF-1 (Li et al., 2004) while CXCR4 blocking antibodies decrease tumor size and metastastic spread to bone in mouse models (Sun et al., 2005). Thus, to determine if SDF l/CXCR4 signaling is involved in the directed migration of 1 833/TR cells towards MC3T3 CM, future experiments should test the effect of SDF- 1 or CXCR4 blocking antibodies. 4.3 A factor produced by breast tumor cells upregulates MMP13 expression and secretion in an osteoblastic microenvironment As described above, a major focus of this project was to identify pathways and molecules in the tumor/bone microenvironment that are involved in the establishment of metastases. More specifically, I was looking for pathways that are initiated separately from the afready well-characterized TGF(3/PTHrP/RANK!RANKL pathway which requires the presence of osteoclasts and direct cell-cell contact (Lacey et al., 1998; Nakagawa et al., 1998). Therefore, I utilized an osteoclast-free co-culture system in which indirect cell-cell (ie. paracrine) interactions could be examined. Using this co-culture method with human tumor and mouse osteoblast lines also enabled quick expression analysis with two different microarray platforms and qPCR. The initial protease and protease inhibitor expression analysis using the CLIPCHIPTM microarray consistently showed [vIMP 13 was significantly upregulated in the mouse osteoblasts in the presence of either tumor cells or their CM (Table 3). Given the importance of MMP 13 in skeletal remodeling (Inada et al., 2004; Page McCaw et al., 2007; Stickens et al., 2004), I chose to follow up on this protease as the primary focus of the experimental portion of this thesis. I first determined if the level of MIvIP 13 induction in osteoblasts was an “all-or none” response (ie. maximal induction immediately) based on the presence of the tumor 80 cells or factors produced by the tumor cells (ie. paracrine molecules), or if the level of induction changed as the co-culture continued over time. Time course analysis showed that MMP 13 transcript expression increased over time and with similar kinetics in all three co culture conditions (Figure 9). Increased MN4P 13 expression was also detected by Western blotting on culture supernatants (Figure 10), but it was mostly the pro-MMP 13 form that was produced. Although MMP13 is activated by MMP2 and MMP14 (Knauper et aL, 1996), the levels of these two IvilviPs did not change in these co-culture experiments. It is possible that in a true in vivo setting, there would be active MIvIP2 or IvilviP 14 in the bone stroma. For example, osteoclast cells could play an important activator role because they secrete MMP 14 that could then activate the newly produced pro-MMP 13 (Dallas et al., 2002; Sato et al., 1997). While the classical role for all MMPs is matrix degradation, there are also other roles that they can play in the metastatic setting. MIVIP 13 can release and activate matrix bound growth factors like CTGF, VEGF, FGF2, and TGFI3 (Dangelo et al., 2001; Mott and Werb, 2004; Nagase et al., 2006; Sounni and Noel, 2005) and therefore any tumor cell with the ability to induce MMP13 activity in the bone microenvironment would have an advantage when trying to establish a micrometastasis. CTGF and VEGF are angiogenic growth factors (Hashimoto et aL, 2002; Kang et al., 2003; Kozlow and Guise, 2005). CTGF stimulates tumor growth and angiogenesis, as well as regulates the activity of VEGF by forming a CTGF-VEGF inactive complex (Hashimoto et al., 2002). VEGF is released from this inhibitory complex through cleavage by MMPs, including MMP13, which restores the angiogenic activity of VEGF (Hashimoto et al., 2002). VEGF is also involved in the regulation of tumor cell migration and invasion through its ability to upregulate CXCR4 expression, increasing chemotaxis to SDF-1 (Bachelder et al., 2002). FGF2 is involved in normal bone development and turnover, and can be released from ECM by MMP13 activity (Whitelock et al., 1996). Once released, soluble FGF2 induces MIVIP 13 expression through an AP- 1-dependent mechanism (Varghese et al., 2000), which could potentially contribute to an increase in MIvIP 13 expression in bone metastasis. TGF(3 is an abundant growth factor in bone matrix that enhances bone metastasis (Kozlow and Guise, 2005) through the induction of PTHrP in tumor driven osteolysis (Guise et al., 2006; Yin et al., 1999) and it can also stimulate the expression of CTGF (Kang et al., 2003). Given that MIvIP13 activates 81 TGFI3 by LAP cleavage (Dangelo et al., 2001) and that TGF(3 is clearly involved in osteolytic bone metastases (Guise et al., 2006; Yin et al., 1999), it is possible that tumor- induced IvilviP 13 upregulation in osteoblasts could promote osteolytic disease progression through TGFI3. 4.4 Oncostatin M can upregulate MMP13 expression in osteoblasts Many different pathways in osteoblasts induce MMP 13 expression. In normal bone development, MIvIP13 induction occurs through an AP-l and Cbfal (RUNX2)-dependent manner, and expression increases during differentiation and mineralization (Winchester et al., 2000). PTH also increases IvilviP 13 levels through a PKA and Cbfal -dependent signaling pathway (Quinn et al., 1990; Selvamurugan et al., 2000). GO analysis of the Affymetrix® data was used to find putative pathways or proteins involved in the upregulation of MMP 13. GO analysis requires that genes are categorized based on their biological function or pathway and comparisons are done assuming that genes in the same biological pathway are likely to have the same expression profile (Werner, 2008). Using these criteria OSMR activity significantly correlated with 1 833/TR cell-induced MMP 13 expression (Tables 5 and 6). Since previous research shows that OSM up-regulates IVIIVIP 13 expression (Litherland et al., 2008; Varghese et al., 1999), and the OSM ligand was produced by the 1 833/TR tumor cells at the transcript level (Figure 13), I chose to follow up on this pathway to determine if OSM was involved in the induction of MIVIP13 by 1833/TR tumor cells. When it was originally isolated, human OSM was shown to be a very stable protein that was resistant to heat up to 90°C, where it lost its biological effectiveness (Zarling et al., 1986). Thus, to determine whether or not the 1833/TR CM likely contained OSM, it was heat-inactivated at either 65°C or 90°C prior to use in culture. The ability of the conditioned medium to increase both MMP13 and OSMR expression was attenuated only when conditioned medium was heat-inactivated at 90°C (Figure 14), consistent with the idea that OSM is present. To test whether OSM alone could specifically increase the expression of MMP13, MC3T3 monolayers were treated with recombinant human OSM. Treatment with recombinant OSM increased the transcript levels of both MIvIP 13 and OSMR in a dose 82 dependent manner (Figure 15) and the 10 ng/ml OSM treatment increased the expression of the pro-MMP 13 peptide to similar levels as seen with both MDA23 1 and 183 3/TR CM (Figure 16). Of note, however, was that the expression levels of the gp 130 subunit of the OSM heterodimeric cytokine receptor were unaltered by both tumor CM and recombinant OSM treatment (Figures 12 and 15). In chondrocytes OSM induces IvilviP 13 expression synergistically with IL- 1 in a PI3KIAkt-dependent signaling cascade (Litherland et al., 2008). Since chondrocytes and osteoblasts are derived from a common precursor, it was possible that the MIvIP 13 upregulation seen in the MC3T3 osteoblast co-cultures with 1 833/TR cells was occuring through a PI3KJAkt-dependent signaling cascade. Therefore, to test this hypothesis I chose to pharmacologically inhibit P13K using LY294002. In contrast to the Litherland et al. experiments, addition of LY294002 failed to inhibit the transcriptional upregulation of either MMP13 or OSMR (Figure 17). Surprisingly, the addition of LY294002 actually resulted in a dramatic increase in MMP 13 and OSMR expression levels indicating that there was likely a PI3KJAkt-dependent signaling cascade that negatively regulates the transcription of both proteins. OSMR signaling involves activation of Jaki, Jak2 and Tyk2 kinase activity that lead to phosphorylation of STAT3 and STAT Sb downstream of cytokine binding (Auguste et al., 1997). Regulation of OSMR expression itself also involves Jakl, Jak2, and Tyk2, but not Jak3 (Radtke et al., 2002). It is therefore possible that the predominant signaling pathway that leads to the increased transcription of both of these molecules involves the Jak/STAT secondary messengers instead of PI3K/Akt. 4.5 A potential second “vicious cycle” in bone metastasis Using indirect co-culture methods has allowed me to elucidate a potential pathway involving MIvIP13 and OSM that is independent of the classical TGF3/PTHrP/RANKJ RANKL vicious cycle in bone metastasis. Despite this finding, it is possible that this OSM/MMP 13 pathway could act together with the TGF(3/PTHrP/RANK!RANKL pathway and enhance bone matrix degradation (Figure 18). In theory, breast tumor cells that produce both PTHrP and OSM could have a significant advantage in establishing micrometastases 83 compared to cells that express just one or the other. For example, PTFIrP could stimulate increased expression of RANKL and M-CSF by the osteoblast cells, thereby activating the osteoclast cells, which would then release proteases, including MIvIP 14 (Sato et al., 1997). This osteoclastic MMP 14 would then activate the increased pro-MIVIP 13 (Dallas et al., 2002; Sato et al., 1997) prouced by osteoblastic cells in response to the OSM from the tumor cells. This would then lead to bone matrix remodeling and further growth factor release from the bone matrix. Also relevant is the fact that OSM can stimulate SDF- 1 expression in stem cell populations (Lee et al., 2007). Therefore, OSM may be playing a role in the upregulation of SDF- 1 observed in the osteoblast population in my studies. SDF- 1 could then recruit more tumor cells to the bone, thereby enhancing its metastatic colonization. In this postulated model, there would be enhanced TGFI3 and other growth factor release, which could also stimulate the formation of new bone (Zaidi, 2007). Since osteoblastic metastases form after initial bone degradation (Roodman, 2004), it is also possible that OSM and MMP 13 could play a role in the establishment of both osteolytic and osteoblastic tumor colonization. Using paracrine-mediated assays, I have shown that there is a RANK-independent pathway involving OSM and MIvIP13. The combination of both RANK-dependent and RANK-idependent pathways could be what ultimately leads to the truly devastating and vicious cycle of bone metastases. Since current therapeutic strategies designed to treat bone metastasis by blocking osteolysis with either bisphophonates or anti-RANK treatment are not curative (Blair et al., 2006; Coleman, 2006; Sordillo and Pearse, 2003), it will be important to determine if blocking this adjacent OSMIMMP 13 pathway (Figure 18) will enhance the treatment of metastatic bone disease. Culture models such as the ones described in this thesis will provide an important platform for testing novel therapeutics and combinatorial therapy against bone metastases in a recapitulated bone microenvironment. 84 Figure 18— A potential pathway involving OSM and MMP13 in the enhancement of bone colonization by breast cancer cells Tumor cells (blue) produce both PTHrP and OSM. PTHrP release would stimulate the production of RANKL and M-CSF by osteoblast cells (grey), stimulating the maturation and activation of osteoclast cells (green), which would release protons and proteases, including MIvIP 14, to degrade the bone matrix. OSM would stimulate the production of pro-MIVIP 13 by osteoblasts, which would then be activated by the IvIIvlPl4 in the microenvironment. Active MMP 13 could then cleave and activate TGFI3, which would then result in increased expression of PTHrP by the tumor cells, propagating the vicious cycle. OSM would also stimulate the production of VCAM-1 and SDF-l. SDF-1 could then recruit more tumor cells and enhance the metastatic colonization of bone. Therapeutically, targeting this OSM pathway could be by preventing the binding of OSM to the OSMR expressed on the surface of the osteoblast cells. 85 Potential therapeutic target ,0”0hibition of OSM receptor binding ii ) Wc+ ç+ 0 0 PTIHIrP * • M-CSF RANKL ? RANK VCAM-1 • SDF-i Latent TG93 A TGF3 ‘) proMMPl3 MMPI3 ) MMPI4/MMP2 Proteinase + Proton ECM Growth Factors 86 REFERENCES Ahmad, R., Qureshi, H. Y., El Mabrouk, M., Sylvester, J., Ahmad, M., and Zafarullah, M. (2007a). Inhibition of interleukin 1-induced matrix metalloproteinase 13 expression in human chondrocytes by interferon gamma. Ann Rheum Dis 66, 782-789. Ahmad, R., Sylvester, J., and Zafarullah, M. (2007b). MyD88, IRAK1 and TRAF6 knockdown in human chondrocytes inhibits interleukin- 1-induced matrix metalloproteinase 13 gene expression and promoter activity by impairing MAP kinase activation. Cell Signal 19, 2549-2557. Auguste, P., Guillet, C., Fourcin, M., Olivier, C, Veziers, J., Pouplard-Barthelaix, A., and Gascan, H. (1997). Signaling of type II oncostatin M receptor. J Biol Chem 272, 15760- 15764. Bachelder, R. E., Wendt, M. A., and Mercurio, A. M. (2002). Vascular endothelial growth factor promotes breast carcinoma invasion in an autocrine manner by regulating the chemokine receptor CXCR4. Cancer Res 62, 7203-7206. Bauer, M., Eickhoff, J. C., Gould, M. N., Mundhenke, C., Maass, N., and Friedl, A. (2008). Neutrophil gelatinase-associated lipocalin (NGAL) is a predictor of poor prognosis in human primary breast cancer. Breast Cancer Res Treat 108, 389-397. Blair, J. M., Zhou, H., Seibel, M. J., and Dunstan, C. R. (2006). Mechanisms of disease: roles of OPG, RANKL and RANK in the pathophysiology of skeletal metastasis. Nat Clin Pract Oncol 3, 4 1-49. Boniface, K., Diveu, C., Morel, F., Pedretti, N., Froger, J., Rayon, E., Garcia, M., Venereau, E., Preisser, L., Guignouard, E., et al. (2007). Oncostatin M secreted by skin infiltrating T lymphocytes is a potent keratinocyte activator involved in skin inflammation. J Immunol 178, 4615-4622. Brennan, D. J., Jirstrom, K., Kronblad, A., Millikan, R. C., Landberg, G., Duffy, M. I., Ryden, L., Gallagher, W. M., and O’Brien, S. L. (2006). CA IX is an independent prognostic marker in premenopausal breast cancer patients with one to three positive lymph nodes and a putative marker of radiation resistance. Clin Cancer Res 12, 6421-6431. Burger, J. A., and Kipps, T. J. (2006). CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood 107, 176 1-1767. Burtis, W. J., Brady, T. G., Orloff J. J., Ersbak, J. B., Warrell, R. P., Jr., Olson, B. R., Wu, T. L., Mitnick, M. E., Broadus, A. E., and Stewart, A. F. (1990). Immunochemical characterization of circulating parathyroid hormone-related protein in patients with humoral hypercalcemia of cancer. N Engi J Med 322, 1106-1112. 87 Celil, A. B., Hollinger, J. 0., and Campbell, P. G. (2005). Osx transcriptional regulation is mediated by additional pathways to BMP2/Smad signaling. J Cell Biochem 95, 518-528. Chambers, T. J. (2000). Regulation of the differentiation and function of osteoclasts. J Pathol 192, 4-13. Choi, K. Y., Kim, H. J., Lee, M. H., Kwon, T. G., Nah, H. D., Furuichi, T., Komori, T., Nam, S. H., Kim, Y. J., and Ryoo, H. M. (2005). Runx2 regulates FGF2-induced Bmp2 expression during cranial bone development. Dev Dyn 233, 115-121. Coleman, R. E. (1997). Skeletal complications of malignancy. Cancer 80, 1588-1594. Coleman, R. E. (2006). Clinical features of metastatic bone disease and risk of skeletal morbidity. Clin Cancer Res 12, 6243s-6249s. Coleman, R. E., and Rubens, R. D. (1987). The clinical course of bone metastases from breast cancer. Br J Cancer 55, 61-66. Coxon, F. P., Thompson, K., and Rogers, M. J. (2006). Recent advances in understanding the mechanism of action of bisphosphonates. Cuff Opin Pharmacol 6, 307-312. Dallas, S. L., Rosser, J. L., Mundy, G. R., and Bonewald, L. F. (2002). Proteolysis of latent transforming growth factor-beta (TGF-beta )-bindmg protein- 1 by osteoclasts. A cellular mechanism for release of TGF-beta from bone matrix. J Biol Chem 277, 21352-21360. Dangelo, M., Sarment, D. P., Billings, P. C., and Pacifici, M. (2001). Activation of transforming growth factor beta in chondrocytes undergoing endochondral ossification. J Bone Miner Res 16, 2339-2347. de Miguel, M. P., de Boer-Brouwer, M., de Rooij, D. G., Paniagua, R., and van Dissel Emiliani, F. M. (1997). Ontogeny and localization of an oncostatin M-like protein in the rat testis: its possible role at the start of spermatogenesis. Cell Growth Differ 8, 611-6 18. Devitt, B., and McLachlan, S. A. (2008). Use of ibandronate in the prevention of skeletal events in metastatic breast cancer. Ther Clin Risk Manag 4, 453-458. Everts, V., Korper, W., Hoeben, K. A., Jansen, I. D., Bromme, D., Cleutjens, K. B., Heeneman, S., Peters, C., Reinheckel, T., Saftig, P., and Beertsen, W. (2006). Osteoclastic bone degradation and the role of different cysteine proteinases and matrix metalloproteinases: differences between calvaria and long bone. J Bone Miner Res 21, 1399-1408. Fernandez, C. A., Yan, L., Louis, G., Yang, J., Kutok, J. L., and Moses, M. A. (2005). The matrix metalloproteinase-9/neutrophil gelatinase-associated lipocalin complex plays a role in breast tumor growth and is present in the urine of breast cancer patients. Clin Cancer Res 11, 5390-5395. 88 Fliesch, H. (1991). Bisphophonates, pharmacology and use in the treatment of tumor- induced hypercalcemic and metastatic bone diseases. Drug 42, 919-942. Framson, P. E., and Sage, E. H. (2004). SPARC and tumor growth: where the seed meets the soil? J Cell Biochem 92, 679-690. Gao, Y., Jheon, A., Nourkeyhani, H., Kobayashi, H., and Ganss, B. (2004). Molecular cloning, structure, expression, and chromosomal localization of the human Osterix (SP7) gene. Gene 341, 101-110. Gong, L., Hoshi, K., Ejiri, S., Nakajima, T., Shingaki, S., and Ozawa, H. (2003). Bisphosphonate incadronate inhibits maturation of ectopic bone induced by recombinant human bone morphogenetic protein 2. J Bone Miner Metab 21, 5-11. Gonzalez, L. 0., Pidal, I., Junquera, S., Corte, M. D., Vazquez, J., Rodriguez, J. C., Lamelas, M. L., Merino, A. M., Garcia-Muniz, J. L., and Vizoso, F. J. (2007). Overexpression of matrix metalloproteinases and their inhibitors in mononuclear inflammatory cells in breast cancer correlates with metastasis-relapse. Br J Cancer 97, 957- 963. Guise, T. A. (2000). Molecular mechanisms of osteolytic bone metastases. Cancer 88, 2892- 2898. Guise, T. A., Mohammad, K. S., Clines, G., Stebbins, E. G., Wong, D. H., Higgins, L. S., Vessella, R., Corey, E., Padalecki, S., Suva, L., and Chirgwin, J. M. (2006). Basic mechanisms responsible for osteolytic and osteoblastic bone metastases. Clin Cancer Res 12, 6213s-6216s. Guise, T. A., and Mundy, G. R. (1998). Cancer and bone. Endocr Rev 19, 18-54. Guise, T. A., Yin, J. J., Taylor, S. D., Kumagai, Y., Dallas, M., Boyce, B. F., Yoneda, T., and Mundy, G. R. (1996). Evidence for a causal role of parathyroid hormone-related protein in the pathogenesis of human breast cancer-mediated osteolysis. J Clin Invest 98, 1544- 1549. Gupta, R. A., and Dubois, R. N. (2001). Colorectal cancer prevention and treatment by inhibition of cyclooxygenase-2. Nat Rev Cancer 1, 11-2 1. Hashimoto, G., Inoki, I., Fujii, Y., Aoki, T., Ikeda, E., and Okada, Y. (2002). Matrix metalloproteinases cleave connective tissue growth factor and reactivate angiogenic activity of vascular endothelial growth factor 165. J Biol Chem 277, 36288-36295. Hofbauer, L. C., and Heufelder, A. E. (2001). Role of receptor activator of nuclear factor kappaB ligand and osteoprotegerin in bone cell biology. J Mol Med 79, 243-253. Hoffman, R. C., Moy, F. J., Price, V., Richardson, J., Kaubisch, D., Frieden, E. A., Krakover, J. D., Castner, B. J., King, J., March, C. J., and Powers, R. (1996). Resonance assigmuents for Oncostatin M, a 24-kDa alpha-helical protein. J Biomol NMR 7, 273-282. 89 Huang, L., Cheng, Y. Y., Chow, L. T., Zheng, M. H., and Kumta, S. M. (2002). Tumour cells produce receptor activator of NF-kappaB ligand (RANKL) in skeletal metastases. J Clin Pathol 55, 877-878. Hughes, D. E., Wright, K. R., Uy, H. L., Sasaki, A., Yoneda, T., Roodman, G. D., Mundy, G. R., and Boyce, B. F. (1995). Bisphosphonates promote apoptosis in murine osteoclasts in vitro and in vivo. J Bone Miner Res 10, 1478-1487. Hurst, S. M., McLoughlin, R. M., Monslow, J., Owens, S., Morgan, L., Fuller, G. M., Topley, N., and Jones, S. A. (2002). Secretion of oncostatin M by infiltrating neutrophils: regulation of IL-6 and chemokine expression in human mesothelial cells. J Immunol 169, 5244-5251. Inada, M., Wang, Y., Byrne, M. H., Rahman, M. U., Miyaura, C., Lopez-Otin, C., and Krane, S. M. (2004). Critical roles for collagenase-3 (Mmp 13) in development of growth plate cartilage and in endochondral ossification. Proc Nat! Acad Sci U S A 101, 17 192- 17197. Jemal, A., Siegel, R., Ward, E., Murray, T., Xu, J., and Thun, M. J. (2007). Cancer statistics, 2007. CA Cancer J Clin 57, 43-66. Kang, H., Watkins, G., Parr, C., Douglas-Jones, A., Manse!, R. E., and Jiang, W. G. (2005). Stromal cell derived factor-i: its influence on invasiveness and migration of breast cancer cells in vitro, and its association with prognosis and survival in human breast cancer. Breast Cancer Res 7, R402-410. Kang, Y., Siegel, P. M., Shu, W., Drobnjak, M., Kakonen, S. M., Cordon-Cardo, C., Guise, T. A., and Massague, J. (2003). A multigenic program mediating breast cancer metastasis to bone. Cancer Cell 3, 53 7-549. Kappelhoff, R., and Overall, C. (2007). The CLIP-CHIP Oligonucleotide Microarray: Dedicated array for analysis of all protease, nonproteolytic homolog, and inhibitor gene transcripts in human and mouse, 49 edn: John Wiley & Sons, Inc.). Katagiri, T., and Takahashi, N. (2002). Regulatory mechanisms of osteoblast and osteoclast differentiation. Oral Dis 8, 147-159. Kershaw, D. B., Beck, S. G., Wharram, B. L, Wiggins, J. E., Goyal, M., Thomas, P. E., and Wiggins, R. C. (1997). Molecular cloning and characterization of human podocalyxin-like protein. Orthologous relationship to rabbit PCLP1 and rat podocalyxin. J Biol Chem 272, 15708-15714. Kinder, M., Chislock, E., Bussard, K. M., Shuman, L., and Mastro, A. M. (2008). Metastatic breast cancer induces an osteoblast inflammatory response. Exp Cell Res 314, 173-183. Knauper, V., Will, H., Lopez-Otin, C., Smith, B., Atkinson, S. J., Stanton, H., Hembry, R. M., and Murphy, G. (1996). Cellular mechanisms for human procollagenase-3 (MIVIP- 13) 90 activation. Evidence that MT 1 -MIVIP (MIvIP- 14) and gelatinase a (MMP-2) are able to generate active enzyme. J Biol Chem 271, 17124-17131. Komori, T., Yagi, H., Nomura, S., Yamaguchi, A., Sasaki, K., Deguchi, K., Shimizu, Y., Bronson, R. T., Gao, Y. H., Inada, M., et at. (1997). Targeted disruption of Cbfal results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89, 755- 764. Koziow, W., and Guise, T. A. (2005). Breast cancer metastasis to bone: mechanisms of osteolysis and implications for therapy. J Mammary Gland Biol Neoplasia 10, 169-180. Lacey, D. L., Timms, E., Tan, H. L., Kelley, M. J., Dunstan, C. R., Burgess, T., Elliott, R., Colombero, A., Elliott, G., Scully, S., et al. (1998). Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93, 165-176. Lacroix, M., Ebel, C., Kardos, J., Dobo, J., Gal, P., Zavodszky, P., Arlaud, G. J., and Thielens, N. M. (2001). Assembly and enzymatic properties of the catalytic domain of human complement protease Cir. J Biol Chem 276, 36233-36240. Lafleur, M. A., Drew, A. F., de Sousa, E. L., Buck, T., Bills, M., Walker, E. C., Williams, E. D., Waltham, M., and Thompson, E. W. (2005). Upregulation of matrix metalloproteinases (MMPs) in breast cancer xenografts: a major induction of stromal MIvIP 13. Tnt J Cancer 114, 544-5 54. Lee, M. H., Kim, Y. J., Kim, H. J., Park, H. D., Kang, A. R., Kyung, H. M., Sung, J. H., Wozney, J. M., and Ryoo, H. M. (2003). BMP-2-induced Runx2 expression is mediated by D1x5, and TGF-beta 1 opposes the BMP-2-induced osteoblast differentiation by suppression of D1x5 expression. J Biol Chem 278, 34387-34394. Lee, M. J., Song, H. Y., Kim, M. R., Sung, S. M., Jung, J. S., and Kim, J. H. (2007). Oncostatin M stimulates expression of stromal-derived factor-i in human mesenchymal stem cells. Tnt J Biochem Cell Biol 39, 650-659. Leeman, M. F., Curran, S., and Murray, G. I. (2002). The structure, regulation, and function of human matrix metalloproteinase-13. Crit Rev Biochem Mol Biol 37, 149-166. Li, Y. M., Pan, Y., Wei, Y., Cheng, X., Zhou, B. P., Tan, M., Zhou, X., Xia, W., Hortobagyi, G. N., Yu, D., and Hung, M. C. (2004). Upregulation of CXCR4 is essential for HER2-mediated tumor metastasis. Cancer Cell 6, 459-469. Liacini, A., Sylvester, J., Li, W. Q., Huang, W., Dehnade, F., Ahmad, M., and Zafarullah, M. (2003). Induction of matrix metalloproteinase- 13 gene expression by TNF-alpha is mediated by MAP kinases, AP- 1, and NF-kappaB transcription factors in articular chondrocytes. Exp Cell Res 288, 208-2 17. Lindberg, R. A., Juan, T. S., Weicher, A. A., Sun, Y., Cupples, R., Guthrie, B., and Fletcher, F. A. (1998). Cloning and characterization of a specific receptor for mouse oncostatin M. Mo! Cell Biol 18, 3357-3367. 91 Lindemann, R. K., Ballschmieter, P., Nordheim, A., and Dittmer, J. (2001). Transforming growth factor beta regulates parathyroid hormone-related protein expression in MDA-MB 231 breast cancer cells through a novel Smad/Ets synergism. J Biol Chem 276, 46661- 46670. Lipton, A. (2007). Treatment of bone metastases and bone pain with bisphosphonates. Support Cancer Ther 4, 92-100. Lipton, A., Steger, G. G., Figueroa, J., Alvarado, C., Solal-Celigny, P., Body, J. J., de Boer, R., Berardi, R., Gascon, P., Tonkin, K. S., et a!. (2007). Randomized active-controlled phase II study of denosumab efficacy and safety in patients with breast cancer-related bone metastases. J Clin Oncol 25, 443 1-4437. Litherland, G. J., Dixon, C., Lakey, R. L., Robson, T., Jones, D., Young, D. A., Cawston, T. E., and Rowan, A. D. (2008). Synergistic collagenase expression and cartilage collagenolysis are phosphatidylinositol 3-kinase/Aict signaling-dependent. J Biol Chem 283, 14221-14229. Major, P. P., Cook, R. J., Chen, B. L., and Zheng, M. (2005). Survival-adjusted multiple- event analysis for the evaluation of treatment effects of zoledronic Acid in patients with bone metastases from solid tumors. Support Cancer Ther 2, 234-240. Malaval, L., Liu, F., Vemallis, A. B., and Aubin, J. E. (2005). GP13O/OSMR is the only LIF/IL-6 family receptor complex to promote osteoblast differentiation of calvaria progenitors. J Cell Physiol 204, 585-593. Mantyh, P. W., Clohisy, D. R., Koltzenburg, M., and Hunt, S. P. (2002). Molecular mechanisms of cancer pain. Nat Rev Cancer 2, 201-209. Martin, M. D., and Matrisian, L. M. (2007). The other side of MiviPs: protective roles in tumor progression. Cancer Metastasis Rev 26, 7 17-724. Matsubara, T., Kida, K., Yamaguchi, A., Hata, K., Ichida, F., Meguro, H., Aburatani, H., Nishimura, R., and Yoneda, T. (2008). BMP2 regulates osterix through Msx2 and Runx2 during osteoblast differentiation. J Biol Chem. McQuibban, G. A., Butler, G. S., Gong, J. H., Bendall, L., Power, C., Clark-Lewis, I., and Overall, C. M. (2001). Matrix metalloproteinase activity inactivates the CXC chemokine stromal cell-derived factor-i. J Biol Chem 276, 43503-43508. McQuibban, G. A., Gong, J. H., Wong, J. P., Wallace, J. L., Clark-Lewis, I., and Overall, C. M. (2002). Matrix metalloproteinase processing of monocyte chemoattractant proteins generates CC chemokine receptor antagonists with anti-inflammatory properties in vivo. Blood 100, 1160-1167. Mercer, R. R., and Mastro, A. M. (2005). Cytokines secreted by bone-metastatic breast cancer cells alter the expression pattern of f-actin and reduce focal adhesion plaques in osteoblasts through P13K. Exp Cell Res 310, 270-281. 92 Mercer, R. R., Miyasaka, C., and Mastro, A. M. (2004). Metastatic breast cancer cells suppress osteoblast adhesion and differentiation. Clin Exp Metastasis 21, 427-435. Mosley, B., De Imus, C., Friend, D., Boiani, N., Thoma, B., Park, L. S., and Cosman, D. (1996). Dual oncostatin M (OSIV1) receptors. Cloning and characterization of an alternative signaling subunit conferring OSM-specific receptor activation. J Biol Chem 271, 32635- 32643. Mott, J. D., and Werb, Z. (2004). Regulation of matrix biology by matrix metalloproteinases. Curr Opin Cell Biol 16, 558-564. Mundy, G. R. (1997). Mechanisms of bone metastasis. Cancer 80, 1546-1556. Mundy, G. R. (2002). Metastasis to bone: causes, consequences and therapeutic opportunities. Nat Rev Cancer 2, 584-593. Murthy, R. V., Arbman, G., Gao, J., Roodman, G. D., and Sun, X. F. (2005). Legumain expression in relation to clinicopathologic and biological variables in colorectal cancer. Clin Cancer Res 11, 2293-2299. Nagase, H., Visse, R., and Murphy, G. (2006). Structure and function of matrix metalloproteinases and TIMPs. Cardiovasc Res 69, 562-573. Nakagawa, N., Kinosaki, M., Yamaguchi, K., Shima, N., Yasuda, H., Yano, K., Morinaga, T., and Higashio, K. (1998). RANK is the essential signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem Biophys Res Commun 253, 395-400. Nielsen, 0. 5., Munro, A. J., and Tannock, I. F. (1991). Bone metastases: pathophysiology and management policy. 3 Clin Oncol 9, 509-524. Nishio, Y., Dong, Y., Paris, M., O’Keefe, R. J., Schwarz, E. M., and Drissi, H. (2006). Runx2-mediated regulation of the zinc finger Osterix/5p7 gene. Gene 372, 62-70. Nizamutdinova, I. T., Lee, G. W., Lee, J. S., Cho, M. K., Son, K. H., Jeon, S. J., Kang, S. S., Kim, Y. S., Lee, J. H., Seo, H. G., et a!. (2008). Tanshinone I suppresses growth and invasion of human breast cancer cells, IvIDA-MB-23 1, through regulation of adhesion molecules. Carcinogenesis. Overall, C. M., and Dean, R. A. (2006). Degradomics: systems biology of the protease web. Pleiotropic roles of IvIMPs in cancer. Cancer Metastasis Rev 25, 69-75. Page-McCaw, A., Ewald, A. J., and Werb, Z. (2007). Matrix metalloproteinases and the regulation of tissue remodelling. Nat Rev Mol Cell Biol 8, 22 1-233. Paget, 5. (1889). The distribution of secondary growths in cancer of the breast. Lancet 1, 571-573. 93 Pelletier, A. J., van der Laan, L. J., Hildbrand, P., Siani, M. A., Thompson, D. A., Dawson, P. E., Torbett, B. E., and Salomon, D. R. (2000). Presentation of chemokine SDF-l alpha by fibronectin mediates directed migration of T cells. Blood 96, 2682-2690. Qian, H. (2008). In, p. Personal communications regarding Affymetrix analysis. Queen, M. M., Ryan, R. E., Holzer, R. G., Keller-Peck, C. R., and Jorcyk, C. L. (2005). Breast cancer cells stimulate neutrophils to produce oncostatin M: potential implications for tumor progression. Cancer Res 65, 8896-8904. Quinn, C. 0., Scott, D. K., Brinckerhoff, C. E., Matrisian, L. M., Jeffrey, J. J., and Partridge, N. C. (1990). Rat collagenase. Cloning, ammo acid sequence comparison, and parathyroid hormone regulation in osteoblastic cells. J Biol Chem 265, 22342-22347. Radtke, S., Hermanns, H. M., Haan, C., Schmitz-Van De Leur, H., Gascan, H., Heinrich, P. C., and Behrmann, I. (2002). Novel role of Janus kinase 1 in the regulation of oncostatin M receptor surface expression. J Biol Chem 277, 11297-11305. Ribatti, D., Mangialardi, G., and Vacca, A. (2006). Stephen Paget and the ‘seed and soil’ theory of metastatic dissemination. Clin Exp Med 6, 145-149. Roodman, G. D. (2004). Mechanisms of bone metastasis. N Engi J Med 350, 1655-1664. Rose-John, S., Scheller, J., Elson, G., and Jones, S. A. (2006). Interleukin-6 biology is coordinated by membrane-bound and soluble receptors: role in inflammation and cancer. J Leukoc Biol 80, 227-236. Sabino, M. A., and Mantyh, P. W. (2005). Pathophysiology of bone cancer pain. 3 Support Oncol 3, 15-24. Sah, N. K., Khan, Z., Khan, G. J., and Bisen, P. 5. (2006). Structural, functional and therapeutic biology of survivin. Cancer Lett 244, 164-171. Sahu, A., and Lambris, J. D. (2001). Structure and biology of complement protein C3, a connecting link between innate and acquired immunity. Immunol Rev 180, 35-48. Salvucci, 0., Bouchard, A., Baccarelli, A., Deschenes, J., Sauter, G., Simon, R., Bianchi, R., and Basik, M. (2006). The role of CXCR4 receptor expression in breast cancer: a large tissue microarray study. Breast Cancer Res Treat 97, 275-283. Sasaki, A., Boyce, B. F., Story, B., Wright, K. R., Chapman, M., Boyce, R., Mundy, G. R., and Yoneda, T. (1995). Bisphosphonate risedronate reduces metastatic human breast cancer burden in bone in nude mice. Cancer Res 55, 355 1-3557. Sato, T., del Carmen Ovejero, M., Hou, P., Heegaard, A. M., Kumegawa, M., Foged, N. T., and Delaisse, J. M. (1997). Identification of the membrane-type matrix metalloproteinase MT1-MIvIP in osteoclasts. J Cell Sci 110 (Pt 5), 589-596. 94 Seibel, M. J. (2005). Clinical use of markers of bone turnover in metastatic bone disease. Nat Clin Pract Oncol 2, 504-5 17. Selvamurugan, N., Pulumati, M. R., Tyson, D. R., and Partridge, N. C. (2000). Parathyroid hormone regulation of the rat collagenase-3 promoter by protein kinase A-dependent transactivation of core binding factor aiphal. J Biol Chem 275, 503 7-5042. Simonet, W. S., Lacey, D. L., Dunstan, C. R., Kelley, M., Chang, M. S., Luthy, R., Nguyen, H. Q., Wooden, S., Bennett, L., Boone, T., et a!. (1997). Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 89, 309-3 19. Somasiri, A., Nielsen, J. S., Makretsov, N., McCoy, M. L., Prentice, L., Gilks, C. B., Chia, S. K., Gelmon, K. A., Kershaw, D. B., Huntsman, D. G., et a!. (2004). Overexpression of the anti-adhesin podocalyxin is an independent predictor of breast cancer progression. Cancer Res 64, 5068-5073. Sordillo, E. M., and Pearse, R. N. (2003). RANK-Fc: A therapeutic antagonist for RANK-L in myeloma. Cancer 97, 802-8 12. Sounni, N. E., and Noel, A. (2005). Membrane type-matrix metalloproteinases and tumor progression. Biochimie 87, 329-342. Stickens, D., Behonick, D. J., Ortega, N., Heyer, B., Hartenstein, B., Yu, Y., Fosang, A. J., Schorpp-Kistner, M., Angel, P., and Werb, Z. (2004). Altered endochondral bone development in matrix metalloproteinase 13-deficient mice. Development 131, 5883-5895. Sun, Y. X., Schneider, A., Jung, Y., Wang, J., Dai, J., Cook, K., Osman, N. I., Koh-Paige, A. J., Shim, H., Pienta, K. J., et aL (2005). Skeletal localization and neutralization of the SDF- 1 (CXCL 1 2)/CXCR4 axis blocks prostate cancer metastasis and growth in osseous sites in vivo. J Bone Miner Res 20, 318-329. Swarthout, J. T., D’Alonzo, R. C., Selvamurugan, N., and Partridge, N. C. (2002). Parathyroid hormone-dependent signaling pathways regulating genes in bone cells. Gene 282, 1-17. Takahashi, N., Udagawa, N., and Suda, T. (1999). A new member of tumor necrosis factor ligand family, ODF/OPGL/TRANCE/RANKL, regulates osteoclast differentiation and function. Biochem Biophys Res Commun 256, 449-45 5. Tanaka, M., Hirabayashi, Y., Sekiguchi, T., Inoue, T., Katsuki, M., and Miyajima, A. (2003). Targeted disruption of oncostatin M receptor results in altered hematopoiesis. Blood 102, 3154-3162. Tsuda, E., Goto, M., Mochizuki, S., Yano, K., Kobayashi, F., Morinaga, T., and Higashio, K. (1997). Isolation of a novel cytokine from human fibroblasts that specifically inhibits osteoclastogenesis. Biochem Biophys Res Commun 234, 137-142. 95 Vakaet, L. A., and Boterberg, T. (2004). Pain control by ionizing radiation of bone metastasis. mt J Dev Biol 48, 599-606. Varghese, S., Rydziel, S., and Canalis, E. (2000). Basic fibroblast growth factor stimulates collagenase-3 promoter activity in osteoblasts through an activator protein-i-binding site. Endocrinology 141, 2185-2191. Varghese, S., Yu, K., and Canalis, E. (1999). Leukemia inhibitory factor and oncostatin M stimulate collagenase-3 expression in osteoblasts. Am J Physiol 276, E465-47 1. Vizoso, F. J., Gonzalez, L. 0., Corte, M. D., Rodriguez, J. C., Vazquez, J., Lamelas, M. L., Junquera, S., Merino, A. M., and Garcia-Muniz, J. L. (2007). Study of matrix metalloproteinases and their inhibitors in breast cancer. Br J Cancer 96, 903-911. Wang, J., Loberg, R., and Taichman, R. S. (2006). The pivotal role of CXCL12 (SDF 1)/CXCR4 axis in bone metastasis. Cancer Metastasis Rev 25, 573-5 87. Werner, T. (2008). Bioinformatics applications for pathway analysis of microarray data. Cuff Opin Biotechnol 19, 50-54. Whitelock, J. M., Murdoch, A. D., lozzo, R. V., and Underwood, P. A. (1996). The degradation of human endothelial cell-derived perlecan and release of bound basic fibroblast growth factor by stromelysin, collagenase, plasmin, and heparanases. J Biol Chem 271, 10079-10086. Winchester, S. K., Selvamurugan, N., D’Alonzo, R. C., and Partridge, N. C. (2000). Developmental regulation of collagenase-3 mRNA in normal, differentiating osteoblasts through the activator protein-i and the runt domain binding sites. J Biol Chem 275, 233 10- 233i8. Wittrant, Y., Theoleyre, S., Chipoy, C., Padrines, M., Blanchard, F., Heymann, D., and Redini, F. (2004a). RANKL/RANKJOPG: new therapeutic targets in bone tumours and associated osteolysis. Biochim Biophys Acta 1704, 49-5 7. Wittrant, Y., Theoleyre, S., Couillaud, S., Dunstan, C., Heymann, D., and Redini, F. (2004b). Relevance of an in vitro osteoclastogenesis system to study receptor activator of NP-kB ligand and osteoprotegerin biological activities. Exp Cell Res 293, 292-301. Yin, J. J., Selander, K., Chirgwin, J. M., Dallas, M., Grubbs, B. G., Wieser, R., Massague, J., Mundy, G. R., and Guise, T. A. (i999). TGF-beta signaling blockade inhibits PTHrP secretion by breast cancer cells and bone metastases development. J Clin Invest 103, 197- 206. Yoneda, T., Michigami, T., Yi, B., Williams, P. J., Niewolna, M., and Hiraga, T. (2000). Actions of bisphosphonate on bone metastasis in animal models of breast carcinoma. Cancer 88, 2979-2988. 96 Yoneda, T., Sasaki, A., Dunstan, C., Williams, P. J., Bauss, F., De Clerck, Y. A., and Mundy, G. R. (1997). Inhibition of osteolytic bone metastasis of breast cancer by combined treatment with the bisphosphonate ibandronate and tissue inhibitor of the matrix metalloproteinase-2. J Clin Invest 99, 2509-2517. Zaidi, M. (2007). Skeletal remodeling in health and disease. Nat Med 13, 791-801. Zarling, J. M., Shoyab, M., Marquardt, H., Hanson, M. B., Lioubin, M. N., and Todaro, G. J. (1986). Oncostatin M: a growth regulator produced by differentiated histiocytic lymphoma cells. Proc Nati Acad Sci U S A 83, 9739-9743. Zhou, Y. X., Xu, X., Chen, L., Li, C., Brodie, S. G., and Deng, C. X. (2000). A Pro25OArg substitution in mouse Fgfrl causes increased expression of Cbfal and premature fusion of calvarial sutures. Hum Mol Genet 9, 200 1-2008. 97 APPENDICES Appendix A: MMP-13 expression from MC3T3 cells after direct contact co culture and FA C sorting MC3T3 cells were co-cultured with either MDA23 1 or 183 3/TR cells that expressed GFP for 24 hrs. Following co-culture, monolayers were trypsinized and resuspended for FAC sorting to collect population specific RNA based on GFP expression. MC3T3 cells kept in serum-free (SF) aMEM for 24 hrs were used as a control for expression level comparisons. MMP 13 expression levels were markedly lower in MC3T3 SF cells that were lysed after trypsinization and FAC sorting (Sus-Lyse), compared to the expression level from MC3T3 SF cells lysed intact in the tissue culture dish (PL-Lyse) indicating that the RNA collected after FAC sorting was not representative of the co-culture level. RNA collected from these FAC sorted populations was also used for .CLIPCHIPTM microarray analysis, but due to very high variability between replicates, no statistical data could be generated, suggesting that this direct contact co-culture method would not be a viable option for microarray analysis. 98 Mouse MMP13 1.2 1 0.8 0.6 0.4 0.2 0 PL-Lyse MC3T3-SF MC3T3-SF MC3T3 MC3T3 Sus-Lyse MDA23 1 1 833/TR 99 Appendix B: Anti-OSM blocking antibody experiment with recombinant OSM and 1833/TR conditioned medium Two independent blocking antibody experiments (A and B) were done using a mouse anti-human OSM antibody (R&D Systems Inc., Minneapolis, MN) to try and neutralize 1833/TR-produced human OSM present in the conditioned media. Both blocking antibody experiments were done by pre-incubating all medium with the indicated amount of antibody for 1 hour at 37°C prior to use in culture. 10 ng/ml of recombinant OSM was used for both experiments, and 1 833/TR-conditioned medium was collected immediately before use and kept at 37°C. (A) The ability for 183 3/TR-conditioned medium to increase MIVIP 13 expression appeared to be blocked by the presence of 10 g of blocking antibody, but the effect of the recombinant OSM was not. There was no effect on OSMR, gpl3O or MIVIP14 expression levels in the presence of blocking antibody. (B) Repetition of the same culture conditions with mouse IgG control antibodies showed that the presence of 20 tg of anti OSM antibody was able to block the effect of recombinant OSM, but not that of the 1 833/TR-conditioned medium. This was opposite to the result seen in the previous experiment shown in A, and was with twice the amount of antibody. Similarly to the previous experiment, the levels of OSMR, gp 130 and MMP 14 were not affected by either concentration of the anti-OSM antibody, nor the IgG control antibody. (C) Whole cell lysates from 1833/TR cells grown for either 24 or 48 hrs at i0 cells per cm2 were immunoprecipitated using protein-G sepharose beads (Sigma) and 5 tg of the same anti- human OSM antibody used in the blocking antibody experiments. 100 ng of recombinant OSM (rOSM) was used as a loading control for Western blotting, and 100 ng was also used for the immunoprecipitation positive control. The very small quantity of rOSM that was detected by immnoprecipitation showed that this particular antibody had low affinity for human OSM, and was likely the reason that both blocking antibody and immunoprecipitation experiments failed. 100 j NoAb • lOugIgG 2Oug lgG • lOug • 2Oug Mouse MMPI3 35 30 25 SF OSM 1833CM Mouse gpl3O 35 30 25 20 Is 10 5 0 ab 35 30 25 20 15 10 5 0 35 30 25 20 15 10 5 0 Mouse OSMR ________________ • SF OSM 1833CM I Oug Anti-OSM Mouse MMPI4 SF OSM 1833CM Mouse OSMR SF OSM 1833CM Mouse MMPI3 I14121086 4 2 0 14 12 10 8 6 4 2 0 j:::::::::::::j’jJ SF OSM 1833CM Mouse gpl3O ,-------__----------------I-- -- I - :t ara I 14 12 10 8 6 4 2 0 14 12 10 8 6 4 2 0 SF OSM 1833CM Mouse MMP14 aIa aj’ .àr SF OSM 1833CM SF OSM - 1833CM 101 CAnti-human OSM IP lOOng 1833/TRWCL lOOng Ab Beads rOSM 2” rC a -s 102 Appendix C: List ofsignificant gene expression changes using CLIP CHIPFMmicroarray analysis ofMC3T3 cells co-cultured with MDA-MB-231 breast tumors on filters (page 1 of2) Description ubiquitin C-terminal hydrolase 5 averageM -0.42 6.56 averageA I F’ collagenase3 2.95 11.14 complement component 3 2.77 6.27 PHEX endopeptidase 1.53 8.45 proteasome catalytic subunit 3i 1.46 8.41 caspase-4/11 1.30 6.69 proteasome catalytic subunit 2i 1.29 8.44 complement component Cira 1.19 10.98 cytosol alanyl aminopeptidase 1.10 9.28 transferrin receptor 2 protein 1.03 6.46 caspase-12 0.93 7.27 carboxypeptidase X2 0.89 6.27 legumain 0.86 8.29 tissue inhibitor of metalloprotease- 1 0.85 11.25 proteasome catalytic subunit ii 0.82 8.12 dihydropyrimidinase-rel. prot. 3 0.81 6.67 cathepsin L2 0.62 9.12 haptoglobin-1 0.62 6.28 complement component Clrb 0.61 7.28 ADAIVI19 0.59 5.97 Cl inhibitor 0.51 6.85 lysosomal carboxypeptidase A 0.51 9.63 testis senile protease 6 0.51 8.79 protease inhibitor 6b 0.47 7.49 elafin precursor 0.45 9.42 adipocyte-enh. binding protein 1 0.42 7.72 protease inhibitor 9e 0.40 6.70 granzyme B 0.39 7.55 nidogen 0.37 9.04 protease inhibitor 9d 0.36 6.31 cystatinT -0.30 7.87 proteasome alpha 3 subunit -0.36 9.29 FACE-2/RCE1 -0.36 8.25 glandular kallikrein mK24 -0.38 6.34 USP14 -0.39 6.55 aminoacylase -0.41 6.77 calpain 6 -0.41 10.09 103 Appendix C: List ofsignificant gene expression changes using CLIP CHIPFM microarray analysis ofMC3T3 cells co-cultured with MDA-MB-231 breast tumors on filters (page 2 of2) proteasome alpha 4 subunit -0.44 9.10 TESP1 -0.46 5.12 jammin-like protease 2 -0.46 8.29 aminopeptidase-like 1 -0.48 7.36 Padl-homolog -0.49 8.78 proliferation-association protein 1 -0.51 10.18 colligin!CBP1 -0.58 8.89 proteasome beta 3 subunit -0.58 11.09 protease inhibitor 2 -0.58 5.36 plasminogen -0.59 5.04 CGI-77 -0.62 8.10 prolyl oligopeptidase -0.67 7.48 testis serine protease 5 -0.67 5.43 survivin -0.71 8.89 ADAMTS1 -0.80 6.52 ubiguitin C-terminal hydrolase 3 -0.81 7.34 mesoderm-specific transcript -1.09 Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.0I fold changes in expression. 8.24 104 Appendix D: List ofsignificant gene expression changes using CLIP CHIPMmicroarray analysis ofMC3T3 cells co-cultured with MDA-MB-231 breast tumor conditioned medium Description averageM averageA I Fold Chan collagenase3(MIvIP-13) 3.81 11.80 Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.OI fold changes in expression. 105 averaeM brain serine proteinase 2 -0.77 Description Appendix E: List ofsignificant gene expression changes using CLIP CHIPFMmicroarray analysis ofMC3T3 cells co-cultured with MDA-MB 231-1833/FR breast tumor cells on filters __________ __________ averaeA collagenase 3 2.57 12.36 caspase-12 1.03 8.07 PHEX endopeptidase 1.02 9.73 proteasome catalytic subunit 2i 1.00 7.73 carboxypeptidase X2 0.98 6.56 proteasome catalytic subunit 3i 0.94 7.94 caspase-4/11 0.85 7.10 legumain 0.82 8.62 ADAMTS5/11 0.82 6.21 cIAP2 0.72 5.89 proteasome catalytic subunit ii 0.64 8.18 complement component Clra 0.63 10.50 haptoglobin-l 0.63 6.16 transferrin receptor 2 protein 0.61 6.35 umbelical vein proteinase 0.59 8.54 complement Cl r-homolog 0.58 5.90 stromelysin 3 0.58 7.57 abhydrolase dom. containing 4 0.57 8.66 procollagen C-proteinase 0.52 6.35 protease inhibitor 9d 0.51 6.14 follistatin-like 1 0.40 9.77 TRAF-binding protein domain 0.39 8.10 ubiguitin C-terminal hydrolase 5 -0.43 7.47 WAP four-disulfide core 6-like 1 -0.47 4.92 calpain 6 -0.48 10.95 USP16 -0.54 7.41 CGI-77 -0.56 8.69 survivin -0.56 9.74 ADAMTS18 -0.59 5.27 Padi-homolog -0.63 8.68 DUB2a-like2 -0.72 4.19 Note: red = upregulated, green = down-regulated. Above represents all sigr p 0.05 and 12.0I fold changes in expression. 5.63 106 Appendix F: List ofsignificant gene expression changes using CLIP CHIPEM microarray analysis ofMC3T3 cells co-cultured with MDA-MB 231-1833/TR breast tumor conditioned medium Description averaeM collagenase 3 3.89 12.8 haptoglobin-1 1.54 7.1 complement component 3 1.44 6 carboxypeptidase X2 1.3 7.7 osteoblast serine protease 1.21 9.5 caspase-4/ll 1 7.2 sparc/osteonectin, testican-2 0.96 7.2 neprilysin 0.81 8.2 legumain 0.77 8.9 complement component Clra 0.72 10.7 caspase-12 0.71 7.2 transferrin receptor 2 protein 0.66 6.6 angiotensinogen!AGT 0.62 8.4 ADAM23 0.6 6.4 proteinase nexin 1/GDN 0.59 11.4 complement component Clrb 0.58 7.6 ADAM9 0.51 8.2 umbelical vein proteinase 0.5 8.4 cystatin ElM 0.47 7.1 tissue inhibitor of metalloprotease- 1 0.45 1 1.5 aminopeptidase PILS 0.45 7 complement Clr-homolog 0.43 6.5 cIAPl 0.37 7.1 mitochondrial ribosomal protein L4 (Mrpl4) -0.38 8.6 carboxypeptidase E -0.52 13.3 MIG3O-like protease inhibitor -0.71 8.3 survivin -0.97 9.6 Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.0I fold changes in expression. averaeA I F’’ 107 Appendix G: List ofGene Ontology analysis ofAffymetrix® data — Conditioned medium versus serum free (page 1 of5) Rank in Term Annotated Significant Expected classic p-value DNA-dependent 1 .00E- ATPase activity 81 9 0.36 1 10 structure-specific DNA 5.90E- binding 118 8 0.52 2 08 single-stranded DNA 9.1 OE binding 83 7 0.37 3 08 ribonucleoside diphosphate reductase 1 .OOE act... 10 3 0.04 4 05 oxidoreductase activity, 1 .OOE acting on CH2 g... 10 3 0.04 5 05 oxidoreductase activity, 1 .40E- acting on CH2 g... 11 3 0.05 6 05 5.80E- kinase regulator activity 219 7 0.97 7 05 oncostatin-M receptor 5. 80E- activity 3 2 0.01 8 05 flu DNA polymerase 5.80E- activity 3 2 0.01 9 05 double-stranded DNA 7.30E- binding 50 4 0.22 10 05 insulin-like growth factor receptor bind... 24 3 0.11 11 0.00016 DNA replication origin binding 5 2 0.02 12 0.00019 cyclin-dependent protein kinase regulato... 65 4 0.29 13 0.0002 telomerase activity 6 2 0.03 14 0.00029 growth factor activity 287 7 1.27 15 0.00031 ATPase activity 749 11 3.32 16 0.00053 ATPase activity, coupled 645 10 2.86 17 0.00063 vascular endothelial growth factor recep... 12 2 0.05 18 0.00125 DNA bending activity 12 2 0.05 19 0.00125 protein kinase regulator activity 187 5 0.83 20 0.00151 108 Appendix G: List ofGene Ontology analysis ofAffymetrix® data — Conditioned medium versus serum free (pajie 2 of5) nucleoside triphosphatase activity 1271 14 5.64 21 0.00159 enzyme inhibitor activity 402 7 1.78 22 0.00218 pyrophosphatase activity 1329 14 5.9 23 0.0024 cytoskeletal protein binding 1044 12 4.63 24 0.00244 hydrolase activity, acting on acid anhyd... 1335 14 5.92 25 0.0025 nucleotidase activity 17 2 0.08 26 0.00254 hydrolase activity, acting on acid anhyd... 1338 14 5.94 27 0.00255 actin binding 686 9 3.04 28 0.00362 actin filament binding 73 3 0.32 29 0.00424 RNA-directed DNA polymerase activity 22 2 0.1 30 0.00426 11 -beta-hydroxysteroid dehydrogenase act... 1 1 0 31 0.00444 dCMP deaminase activity 1 1 0 32 0.00444 lactosylceramide 4- alpha-galactosyltrans... 1 1 0 33 0.00444 protein binding, bridging 76 3 0.34 34 0.00474 cytokine activity 364 6 1.62 35 0.00582 DNA binding 4518 31 20.05 36 0.00799 thymidine kinase activity 2 1 0.01 37 0.00885 3’(2’),5’-bisphosphate nucleotidase acti... 2 1 0.01 38 0.00885 hemoglobin binding 2 1 0.01 39 0.00885 uridine nucleosidase activity 2 1 0.01 40 0.00885 receptor binding 1136 11 5.04 41 0.01256 1 -phosphatidylinositol 4-kinase activity 3 1 0.01 42 0.01325 uracil DNA N glycosylase activity 3 1 0.01 43 0.01325 rRNA primary transcript binding 3 1 0.01 44 0.0 1325 tau protein binding 3 1 0.01 45 0.01325 109 Appendix G: List of Gene Ontology analysis ofAffvmetrix® data — Conditioned medium versus serum free (pare 3 of5) calcium-dependent protein binding 3 1 0.01 46 0.01325 tRNA-pseudouridine synthase activity 43 2 0.19 47 0.01567 DNA helicase activity 44 2 0.2 48 0.0 1637 interstitial collagenase activity 4 1 0.02 49 0.01763 protein kinase C inhibitor activity 4 1 0.02 50 0.0 1763 flap endonuclease activity 4 1 0.02 51 0.01763 pseudouridine synthase activity 49 2 0.22 52 0.02006 GPI anchor binding 234 4 1.04 53 0.02065 interleukin-6 receptor binding 5 1 0.02 54 0.02199 S100 beta binding 5 1 0.02 55 0.02199 protein binding 12850 69 57.02 56 0.02363 transcription termination factor activit... 6 1 0.03 57 0.02633 RNA polymerase I transcription terminati... 6 1 0.03 58 0.02633 prothoracicotrophic hormone activity 6 1 0.03 59 0.02633 deoxynucleoside kinase activity 6 1 0.03 60 0.02633 binding 23719 115 105.25 61 0.02667 nucleotidyltransferase activity 261 4 1.16 62 0.02929 ErbB-3 class receptor binding 7 1 0.03 63 0.03065 kinase inhibitor activity 63 2 0.28 64 0.0320 1 collagenase activity 8 1 0.04 65 0.03496 caspase inhibitor activity 8 1 0.04 66 0.03496 insulin binding 8 1 0.04 67 0.03496 DNA-directed DNA polymerase activity 67 2 0.3 68 0.03582 chemokine activity 67 2 0.3 69 0.03582 chemokine receptor binding 68 2 0.3 70 0.0368 110 Appendix G: List ofGene Ontology analysis ofAffymetrix® data — Conditioned medium versus serum free (page 4 of5) RNA binding 1518 12 6.74 71 0.03788 phosphatidate phosphatase activity 9 1 0.04 72 0.03924 peptide hormone binding 9 1 0.04 73 0.03924 lithium ion binding 9 1 0.04 74 0.03924 intramolecular transferase activity 76 2 0.34 75 0.04501 enzyme regulator activity 1561 12 6.93 76 0.04521 transforming growth factor beta receptor... 12 1 0.05 77 0.05198 adenyl nucleotide binding 3478 22 15.43 78 0.05534 protein homodimerization activity 322 4 1.43 79 0.05566 inositol or phosphatidylinositol kinase... 88 2 0.39 80 0.05843 inositol or phosphatidylinositol phospha... 92 2 0.41 81 0.06317 1 -acylglycerol-3 - phosphate 0- acyltransfe... 15 1 0.07 82 0.06454 5’-nucleotidase activity 15 1 0.07 83 0.06454 DNA N-glycosylase activity 15 1 0.07 84 0.06454 phosphoinositide 3- kinase regulator acti... 15 1 0.07 85 0.06454 G-protein-coupled receptor binding 94 2 0.42 86 0.06559 identical protein binding 640 6 2.84 87 0.06617 chromatin binding 343 4 1.52 88 0.06695 protein disulfide isomerase activity 16 1 0.07 89 0.0687 intramolecular oxidoreductase activity, ... 16 1 0.07 90 0.0687 hydrolase activity 4582 27 20.33 91 0.07249 111 Appendix G: List ofGene Ontology analysis ofAffymetrix® data — Conditioned medium versus serum free (page 5 of5) endodeoxyribonuclease activity 17 1 0.08 92 0.07283 protein phosphatase binding 17 1 0.08 93 0.07283 nucleic acid binding 7311 40 32.44 94 0.0798 1 intramolecular oxidoreductase activity, ... 19 1 0.08 95 0.08105 phosphoinositide binding 373 4 1.66 96 0.08497 protein kinase C binding 20 1 0.09 97 0.08513 metal ion transporter activity 110 2 0.49 98 0.086 FAD binding 111 2 0.49 99 0.08733 acylglycerol 0- acyltransferase activity 22 1 0.1 100 0.09323 112 Appendix H: List ofGene Ontology analysis ofAffymetrix® data — Filter versus serumfree (page 1 of5) Rank inAnnotated Significant Expected classic p-valueTerm oncostatin-M receptor 2.1 OE3 2 0.01 1 activity 05 unspecific monooxygenase 37 3 0.1 2 0.00014 activity monooxygenase 199 5 0.53 3 0.0002 activity protein dimerization 660 8 1.77 4 0.0004 activity oxidoreductase activity, 57 3 0.15 5 0.00049 acting on paire... heme binding 268 5 0.72 6 0.00079 tetrapyrrole binding 268 5 0.72 7 0.00079 protein kinase activator 23 2 0.06 8 0.00174 activity protein homodimerization 322 5 0.87 9 0.00 179 activity kinase activator activity 27 2 0.07 10 0.0024 amine N methyltransferase 1 1 0 11 0.00269 activity 3-aipha-hydroxysteroid 1 1 0 12 0.00269dehydrogenase (B-... oxidoreductase activity 1631 11 4.38 13 0.00422 oxidoreductase activity, 247 4 0.66 14 0.00448 acting on paire... iron ion binding 594 6 1.6 15 0.00535 1 ,4-alpha-glucan branching enzyme 2 1 0.01 16 0.00537 activi... glycogen (starch) 2 1 0.01 17 0.00537 synthase activity hemoglobin binding 2 1 0.01 18 0.00537 myosin II binding 2 1 0.01 19 0.00537 deaminase activity 42 2 0.11 20 0.00573 hydrolase activity, 44 2 0.12 21 0.00628 acting on carbon-nit... 113 Appendix H: List ofGene Ontology analysis ofAfJ’metrix® data — Filter versus serum free (‘pa re 2 of5) AMP deaminase 4 1 0.01 22 0.01071 activity extracellular matrix 4 1 0.01 23 0.01071 constituent conferr... cytidine deaminase 5 1 0.01 24 0.01337 activity nitric-oxide synthase 5 1 0.01 25 0.01337 activity alcohol dehydrogenase 5 1 0.01 26 0.01337(NADP+) activity chemokine activity 67 2 0.18 27 0.0141 chemokine receptor 68 2 0.18 28 0.0145binding telomerase activity 6 1 0.02 29 0.0 1602 aipha-amylase activity 6 1 0.02 30 0.0 1602 amylase activity 7 1 0.02 31 0.01867 GTP-Rho binding 7 1 0.02 32 0.0 1867 RNA binding 1518 9 4.08 33 0.02061 single-stranded DNA 83 2 0.22 34 0.02111binding nucleoside binding 8 1 0.02 35 0.0213 1 phosphatidate 9 1 0.02 36 0.02394phosphatase activity aldehyde dehydrogenase 10 1 0.03 37 0.02656 [NAD(P)+] activit... G-protein-coupled 94 2 0.25 38 0.0266 receptor binding 3-cliloroallyl aldehyde 11 1 0.03 39 0.02918dehydrogenase act... phosphatidylinositol 11 1 0.03 40 0.02918transporter activit... identical protein 640 5 1.72 41 0.02919binding transforming growth 12 1 0.03 42 0.03179factor beta receptor... FAD binding 111 2 0.3 43 0.03609 actin binding 686 5 1.84 44 0.03763 aldehyde dehydrogenase (NAD) 15 1 0.04 45 0.03958 activity 5’-nucleotidase activity 15 1 0.04 46 0.03958 114 Appendix H: List ofGene Ontology analysis ofAffymetrix® data — Filter versus serumfree (pa ‘e 3 of5) hematopoietinlinterfero 117 2 0.31 47 0.03971 n-class (D200-dom... structure-specific DNA 118 2 0.32 48 0.04032binding extracellular matrix 120 2 0.32 49 0.04156 structural constitu... growth factor activity 287 3 0.77 50 0.0424 nucleotidase activity 17 1 0.05 51 0.04474 UDP glucosyltransferase 17 1 0.05 52 0.04474 activity FMN binding 19 1 0.05 53 0.04987 glucosyltransferase 19 1 0.05 54 0.04987 activity aldo-keto reductase 20 1 0.05 55 0.05243 activity RNA-directed DNA 22 1 0.06 56 0.05752polymerase activity transferase activity, 327 3 0.88 57 0.05 825transferring hexos... hydrolase activity, 156 2 0.42 58 0.06619 acting on carbon-nit... steroid dehydrogenase 28 1 0.08 59 0.07263 activity myosin binding 30 1 0.08 60 0.07762 oxidoreductase activity, 35 1 0.09 61 0.08996 acting on paire... oxidoreductase activity, 187 2 0.5 62 0.09036 acting on the C... protein kinase regulator 187 2 0.5 63 0.09036 activity polypeptide N acetylgalactosaminyltra 36 1 0.1 64 0.09241 nsf... UDP glycosyltransferase 194 2 0.52 65 0.096 13 activity acid phosphatase 38 1 0.1 66 0.09729 activity carbonate dehydratase 39 1 0.1 67 0.09972 activity NADP binding 40 1 0.11 68 0.10214 115 Appendix H: List of Gene Ontology analysis ofAffymetrix® data — Filter versus serum free (pa Fe 4 of5) oxidoreductase activity, 204 2 0.55 69 0.10455 acting on CH-OH... lipid phosphatase 41 1 0.11 70 0.10456 activity tRNA-pseudouridine 43 1 0.12 71 0. 10937 synthase activity RNA splicing factor 44 1 0.12 72 0.11177 activity, transester... kinase regulator 219 2 0.59 73 0.11754 activity pseudouridine synthase 49 1 0.13 74 0.12366 activity phospholipid 50 1 0.13 75 0.12602transporter activity coenzyme binding 235 2 0.63 76 0.13181 phospholipaseA2 53 1 0.14 77 0.13306 activity extracellular matrix 55 1 0.15 78 0.13773 structural constitu... insulin-like growth 56 1 0.15 79 0. 14005factor binding Rho GTPase binding 57 1 0.15 80 0.14236 transferase activity, 489 3 1.31 81 0.14475transferring glyco... protein heterodimerization 252 2 0.68 82 0.14739 activity acetylgalactosaminyltra 60 1 0.16 83 0.14928 nsferase activity cytoskeletal protein 1044 5 2.81 84 0.14984binding oxidoreductase activity, 61 1 0.16 85 0.15157 acting on the a... chymotrypsin activity 62 1 0.17 86 0.15385 cofactor binding 267 2 0.72 87 0.16143 transcriptional 518 3 1.39 88 0.16324 repressor activity trypsin activity 73 1 0.2 89 0.17859 actin filament binding 73 1 0.2 90 0.17859 manganese ion binding 286 2 0.77 91 0.17955 oxidoreductase activity, 1 0.2 92 0.18301 acting on the a... 116 Appendix H: List of Gene Ontology analysis ofAffymetrix® data — Filter versus serum free (pa re 5 of5) intramolecular 76 1 0.2 93 0.18522transferase activity protein binding, 76 1 0.2 94 0.18522bridging phosphoric monoester 572 3 1.54 95 0.19939hydrolase activity isomerase activity 324 2 0.87 96 0.2 166 DNA binding 4518 15 12.15 97 0.22682 protein binding 12850 38 34.54 98 0.2539 cytokine activity 364 2 0.98 99 0.25627 metal ion transporter 110 1 0.3 100 0.25668 activity 117 Appendix I: List ofsignificant gene expression changes in MC3 T3 osteoblasts by Affymetrix® analysis — Filter vs. serum-free (page 1 of3 118 Appendix I: List ofsignificant gene expression changes in MC3T3 osteoblasts by A l’metrix® analysis — Filter vs. se 119 Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.0I fold changes in expression. Appendix I: List ofsignificant gene expression changes in MC3T3 osteoblasts by A ‘ vmetrix® analysis — Filter vs. serum-free (1- -- ‘3 of3) I 120 LAppendix J: List ofsignificant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Conditioned medium vs. serum-free (page 1 of5) 121 Appendix J: List ofsignificant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Conditioned medium vs. serum-free 122 Appendix J: List ofsignificant gene expression changes in MC3T3 osteoblasts by Afmetrix® analysis — Conditioned medium vs. serum-free (--- 3 of.5 123 Appendix J: List ofAffymetrix® analysis — Conditioned medium vs. serum ‘ree (page 4 of5) Ras association (RaibS/AF-6) domain family 2 thyroid hormone receptor interactor 13 LSM12 homolog (S. cerevisiae) trans-acting transcription factor 3 cell division cycle 27 homolog (S. cerevisiae) miniebromosome maintenance deficient 2 mitotin (S. cerevisiae S100 protein, beta polypeptide, neural ribonucleotide reductase M2 denticleless homolog (Drosophila) zinc finger protein 207 early growth response 3 minichromosome maintenance deficient 5, cell division cycle 46 (S. cerevisiae) cell division cycle associated 8 E2F transcription factor 7/11 similar to E2F transcription factor 7 high mobility group box 2 kinesin family member 2C similar to developmentally regulated RNA-binding protein 1 dyskeratosis congenita 1, dyskerin homolog (human) minichromosome maintenance deficient 7 (S. cerevisiae) muucbromosome mamtenance deficient 6 (MIS5 homolo S nombe) (S cerevisiae) phosphatidyli.nositol 4-kinase, catalytic, beta polypeptide dyskeratosis ccngenita 1, dyskerin honiolog (human) Transcribed locus, weakly similar to XP_001068222.1 similar to arylsulfatase E precursor [Rattus norveicusj Transcribed locus ribonucleotide reductase M2 structural maintenance of chromosomes 2 minichromosome maintenance deficient 2 mitotin (S. cerevisiae) RAP2B, member of RAS oncogene family profilin 1 papillary renal cell carcinoma (translocation-associated) polymerase (DNA directed), eta (RAD 30 related) TAP binding protein WD repeat domain 1 germ cell-specific gene 2 centrosomal protein 55 Wilms’ tumour 1-associating protein high mobility group box 2 KH-type splicing regulatory protein RJKEN cDNA 2600005003 gene dCMP deaminase retinoblastoma binding protein 4 124 Appendix J: List ofAffymetrix® analysis — Conditioned medium vs. serum- I7fl[gjyaflJ Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.0I fold changes in expression. 125 Appendix K: List ofsignificant gene expression changes in MC3T3 osteoblasts by Affymetrix® analysis — Conditioned medium vs. filter Note: red = upregulated, green = down-regulated. Above represents all significant hits with p 0.05 and 12.0I fold changes in expression. 126

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0066946/manifest

Comment

Related Items