Models and Mechanisms of Prostate Cancer Bone Metastases

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jessica Kelly Simmons

Graduate Program in Veterinary Biosciences

The Ohio State University

2014

Dissertation Committee:

Thomas Rosol, Adviser

Ramiro Toribio

Ahmad Shabsigh

Copyrighted by

Jessica Kelly Simmons

2014

Abstract

It is estimated that over 238,000 men were diagnosed with prostate cancer in 2013.

Prostate cancer death is typically secondary to metastatic disease. The most common site of metastasis is bone. These metastases are typically osteoblastic and result in increased morbidity and mortality through debilitating pain, pathologic fracture, spinal cord compression and paralysis. Therefore, the pathogenesis of prostate cancer bone metastasis represents a major clinical and research interest to investigate and develop therapeutic strategies. The purpose of this body of work was to develop and characterize a model of osteoblastic prostate cancer bone metastasis and to investigate the molecular pathogenesis of these metastasis.

In order to research prostate cancer, animal models that closely recapitulate the disease in man must be developed. To date, there is no complete animal model; instead, each model focuses on one or more phases of prostate cancer progression. The majority of the animal models of prostate cancer that develop bone metastases do not recapitulate the osteoblastic nature of metastases found in men. The first objective of this work was to characterize a naturally occurring model of osteoblastic prostate cancer metastasis in dogs. This model, the Probasco cell line, grew well in vitro and in vivo and formed dramatically osteoblastic metastases after intra-tibial or intra-cardiac injection. The

Probasco prostate cancer cell line will be a valuable model to investigate the mechanisms of prostate cancer pathogenesis and osteoblastic bone metastases.

The second objective was to investigate the molecular pathways important in the pathogenesis of prostate cancer bone metastases in relevant in vitro and in vivo models.

In prostate cancer cells expressing parathyroid hormone-related peptide (PTHrP), increased tumor growth and a shift to a more bone resorptive phenotype was evident in both in vitro and in vivo models. Since bone resorption is known to release factors that can stimulate tumor cell proliferation, the existence of a vicious cycle between prostate cancer growth and osteoclastic bone resorption may be playing a role in the increased tumor size.

The final objective was to expand upon our previous work detailing the significance of

Wnt signaling in prostate cancer bone metastasis. Previously, our lab established a canine prostate cancer cell line (Ace-1) that metastasized to bone and caused mixed osteoblastic and osteolytic lesions in nude mice. We had previously found that Ace-1 cells expressing

Dickkopf-1 (Dkk-1), an antagonize of canonical Wnt signaling and an enhancer of non- canonical Wnt/JNK signaling, resulted in increased tumor growth and number of metastases while decreasing new woven bone formation in the metastases. In this study, we confirmed enhanced Wnt/JNK signaling in vitro and how modulation of this pathway could alter gene expression and result in the increased cell growth and metastasis seen in vivo.

These studies support the role of PTHrP and Wnt signaling in the pathogenesis of prostate cancer bone metastasis. Additionally, a novel experimental model of osteoblastic prostate cancer bone metastasis was developed for the advancement of prostate cancer research.

iii

Dedication

This was a triumph,

and I could not have done it without my loving husband Erich.

Acknowledgements

I am very grateful to Dr. Thomas Rosol, my advisor, for his guidance and support throughout my graduate research training. I am also very grateful for Dr. Wessel Dirksen, who provided immense technical support and advice. The past and current members of the Rosol Lab all contributed to my success through advice, training, and collaboration.

My research is built upon the foundation of prostate cancer work done by Dr. Bruce

LeRoy and Dr. Nandu Thudi. I am very thankful for the undergraduate students who assisted me with these projects, including Ms. Keri Deininger, Mr. Josiah Pelot, and Mr.

James Feller.

I relied on numerous people within the Department, College and University, including

Ms. Shelly Haramia for slide scanning; Mr. Alan Flechtner, Ms. Anne Saulsbery and Ms.

Florinda Jaynes for slide preparation; Ms. Michelle Carlton for microCT support; Dr.

Matthew Allen and Ms. Sue Ringler for Faxitron and IVIS support, and all of the ULAR staff in Sisson Hall vivarium for facilitating my in vivo studies.

The Probasco cell line could not have been created without the efforts of Dr. Holly

Borghese and her colleges in the Biospecimen Repository.

My research and training were supported by the National Cancer Institute (award numbers T32-CA009338 and F32-CA165689).

Many thanks to my family for their love and support. They guided me through the hard times and celebrated with me in the good. Without Erich, none of this would not have been possible. Lastly, thank you Kaitlyn, because even though you are only a year old, you have impacted my life in ways I could never have imagined. Your smiles bring brightness and levity to every day.

Vita

2002...... Heritage High School

2006...... B.S.A. Animal Health, The University of Georgia

2009...... D.V.M, The University of Georgia

2009 to present ...... Veterinary Anatomic Pathology Resident Department of Veterinary Biosciences The Ohio State University

2010 to present ...... Post Doctoral Fellow Department of Veterinary Biosciences The Ohio State University

Publications

Stewart J. Wildebeest MCF in U.S. Cattle. SCWDS Briefs. 24 (1): 2-3. April 2008. (Published under maiden name of Stewart)

Stewart J. White Nose Syndrome in Bats. SCWDS Briefs. 24 (1):4-5. April 2008. (Published under maiden name of Stewart)

Simmons JK, McManamon R, Rech RR, Phillips AE, Howerth EW. Pathology in practice. Necrotizing pyogranulomatous meningoencephalitis with intralesional fungal hyphae, consistent with Cladophialophora bantiana. J Am Vet Med Assoc. 1;236(3):295- 7. Feb 2010. PMID: 20113240

Thudi NK, Shu ST, Martin CK, Lanigan LG, Nadella MV, Van Bokhoven A, Werbeck JL, Simmons JK, Murahari S, Kisseberth WC, Breen M, Williams C, Chen CS, McCauley LK, Keller ET, Rosol TJ. Development of a Brain Metastatic Canine Prostate

vii

Cancer Cell Line. The Prostate.71(12): 1251-62. September 2011. PMID: 21321976 PMCID: PMC3139788

Martin-Vaquero, Paula; da Costa, Ronaldo; Simmons, Jessica; Beamer, Gillian; Jäderlund, Karin; Oglesbee, Michael. A Novel Spongiform Leukoencephalopathy in Neonatal Border Terriers. JVIM. 26(2):402-6. March 2012.

Lisa Y. Wu, Jacqueline M. Johnson, Jessica K. Simmons, Desiree E. Mendes, Jonathan J. Geruntho, Tiancheng Liu, Wessel P. Dirksen, Thomas J. Rosol, William C. Davis, and Clifford E. Berkman. Biochemical characterization of prostate-specific membrane antigen from canine prostate carcinoma cells. The Prostate. 74(5):451-7. PMID: 24449207

Martin CK, Dirksen WP, Carlton MM, Lanigan LG, Pillai SP, Werbeck JL, Simmons JK, Hildreth BE 3rd, London CA, Toribio RE, Rosol TJ. Combined zoledronic acid and meloxicam reduced bone loss and tumour growth in an orthotopic mouse model of bone- invasive oral squamous cell carcinoma. Vet Comp Oncol. 2013 May 8. doi: 10.1111/vco.12037. [Epub ahead of print] PMID: 23651067

Fields of Study

Major Field: Veterinary Biosciences

viii

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgements ...... v

Vita ...... vii

List of Tables ...... xii

List of Figures ...... xiii

Chapter 1 Review of Animal Models of Prostate Cancer Bone Metastasis ...... 1

Abstract ...... 1

Introduction ...... 2

Canines in Prostate Cancer Research ...... 7

Human Prostate Cancer Xenografts in Immunocompromised Mice ...... 11

Prostate Cancer in Mice ...... 15

Mouse Prostate Reconstitution Model ...... 17

Prostate Cancer in Rats ...... 19 ix

Conclusions ...... 20

Tables ...... 22

Figures ...... 24

Chapter 2 Canine Prostate Cancer Cell Line (Probasco) Produces Osteoblastic Metastases in Vivo ...... 30

Abstract ...... 30

Introduction ...... 31

Materials and Methods ...... 33

Results ...... 40

Discussion ...... 45

Conclusions ...... 51

Tables ...... 52

Figures ...... 54

Chapter 3 PTHrP Expression Enhances Prostate Cancer Growth and Osteolysis in

Osteoblastic Bone Metastases ...... 64

Abstract ...... 64

Introduction ...... 65

Materials and Methods ...... 66

Results ...... 75

Discussion ...... 78

Conclusions ...... 84

Tables ...... 85

Figures ...... 87

Chapter 4 Dickkoft-1 (Dkk-1) Enhanced Non-canonical Wnt/JNK signaling in Prostate

Cancer Bone Metastases ...... 97

Abstract ...... 97

Introduction ...... 98

Materials and Methods ...... 100

Results ...... 105

Discussion ...... 106

Conclusions ...... 111

Tables ...... 112

Figures ...... 113

Bibliography ...... 118

List of Tables

Table 1.1 Canine Cell Lines that Metastasize to Bone ...... 22

Table 1.2 Human Cell and Tumor Lines used to Model Bone Metastasis in Mice ...... 23

Table 2.1 Antibodies used to immunohistochemically characterize the Probasco cell line.

...... 52

Table 2.2 The list of primers used for PCR and quantitative RT-PCR to characterize the

Probasco cell line...... 53

Table 3.1 Primers used for qRT-PCR with the Probasco-PTHrP and Probasco-Vector cell

lines...... 85

Table 3.2 Number and location of metastases following intracardiac injection of

Probasco-Vector and Probasco-PTHrP cells ...... 86

Table 4.1 Primers used for qRT-PCR to characterize Ace-1-Vector and Ace-1-Dkk-1

gene expression...... 112

xii

List of Figures

Figure 1.1 Different injection models of prostate cancer to investigate cancer progression

and metastasis...... 24

Figure 1.2 Effects of canine prostatic tissue implanted adjacent to murine calvaria in vivo.

...... 25

Figure 1.3 Intraprostatic Ace-1 tumor in an immunosuppressed dog...... 26

Figure 1.4 Nude mouse subcutaneously transplanted vertebral vossicle with canine Ace-1

tumor cells (T)...... 27

Figure 1.5 Intratibial Probasco tumor in a nude mouse...... 28

Figure 1.6 MatLyLu vertebral metastasis in a Copenhagen rat...... 29

Figure 2.1Photomicrographs of the primary prostatic carcinoma, the subcutaneous

xenograft, and the in vitro cell line (Probasco)...... 54

Figure 2.2 In vitro and in vivo growth patterns of Probasco cells...... 55

Figure 2.3Expression of select genes in the Probasco prostate cancer cells...... 56

Figure 2.4 µCT and radiographic analysis of a nude mouse tibia injected with Probasco

cells...... 57

Figure 2.5 Photomicrograph of H&E-stained tibia 6 weeks post-injection with Probasco

cells...... 58

xiii

Figure 2.6 Hybridization of canine autosomal centromeric BAC clones to an example

metaphase spread of Probasco...... 59

Figure 2.7 DNA copy number profiling of Probasco...... 60

Figure 2.8 DNA copy number profiling of Probasco recoded to human genome

coordinates and output as a human data ...... 62

Figure 3.1 In vitro mean cell diameter and actin immunocytochemistry of Probasco-

Vector and Probasco-PTHrP cells...... 87

Figure .3.2 In vitro and in vivo growth patterns of Probasco cells...... 88

Figure 3.3 Probasco-Vector and Probasco-PTHrP subcutaneous tumors ...... 89

Figure 3.4 Expression of genes and PTHrP secretion in the Probasco-Vector and

Probasco-PTHrP prostate cancer cells...... 91

Figure 3.5 Wound healing assay ...... 92

Figure 3.6 In vitro calvaria co-culture assay using 3.5mm neonatal murine calvaria discs

and Probasco-Vector and Probasco-PTHrP cell lines...... 93

Figure 3.7 Tibias from nude mice after injection of Probasco-Vector or Probasco-PTHrP

cells...... 94

Figure 3.8 Expression of specific mRNAs in the Probasco-PTHrP, Probasco-Vector, Leo,

Ace-1 and Ace-1-Dkk-1 canine prostate cancer cells...... 95

Figure 4.1 Secreted human Dkk-1 in culture media after 24 hours measured by ELISA.

...... 113

Figure 4.2 β-catenin immunohistochemical staining and quantification in Ace-1-Vector

and Ace-1-Dkk-1 cells ...... 114

xiv

Figure 4.3 AP-1 luciferase reporter activity of Ace-1-Vector and Ace-1-Dkk-1 cells with

no treatment, TGF-β, or SP600125...... 115

Figure 4.4 Comparison of cell proliferation and migration between Ace-1-Vector and

Ace-1-Dkk-1 cells. (n=3, repeated twice) ...... 116

Figure 4.5 Expression of select genes in Ace-1-Vector and Ace-1-Dkk-1 cells...... 117

Chapter 1 Review of Animal Models of Prostate Cancer Bone Metastasis

Abstract

Prostate cancer bone metastases are associated with a poor prognosis and are considered incurable. Insight into the formation and growth of prostate cancer bone metastasis is required for development of new imaging and therapeutic strategies to combat this devastating disease. Animal models are indispensable in investigating cancer pathogenesis and evaluating therapeutics. Multiple animal models of prostate cancer bone metastasis have been developed, but few effectively model prostatic neoplasms and osteoblastic bone metastases as they occur in men. This review discusses the animal models that have been developed to investigate prostate cancer bone metastasis, with a focus on canine models and also includes human xenograft and rodent models. Adult dogs spontaneously develop benign prostatic hyperplasia and prostate cancer with osteoblastic bone metastases. Large animal models, such as dogs, are needed to develop new molecular imaging tools and effective focal intraprostatic therapy. None of the

1 available models fully reflect the metastatic disease seen in men, although the various models have provided important insight into the metastatic process. As additional models are developed and knowledge from the different models is combined, the molecular mechanisms of prostate cancer bone metastasis can be deciphered and targeted for development of novel therapies and molecular diagnostic imaging.

Introduction

Prostate cancer is the most common newly diagnosed cancer in men in the United States and it is estimated to result in 10% of all cancer deaths in men in 2013 [1]. Between 70-

100% of patients that die due to prostate cancer have bone metastases [2, 3]. Skeletal metastases have a 5-year survival rate of 25% and median survival of 40 months [4]. The most common sites of bone metastasis in prostate cancer are pelvic bones, vertebral column, ribs, and long bones (primarily femur and humerus) [5, 6]. Bone metastases in prostate cancer are primarily osteoblastic, or bone-forming lesions. These metastases result in pain, fractures and nerve compressions that reduce quality of life and represent incurable disease to clinicians [7].

In vivo animal models are the key to understanding the pathogenesis of prostate cancer bone metastasis to develop better therapeutics. The perfect model would include a species with a prostate structurally and functionally similar to humans, have slowly developing

2 disease that progresses through hyperplasia, prostatic intra-epithelial neoplasia (PIN), locally invasive carcinoma and late-stage osteoblastic metastasis to the pelvis, vertebrae, ribs, and long bones [8]. No model exists at this time that perfectly recapitulates human disease; however, each available model has its applicability towards particular aspects of bone metastasis. As new models are developed that more fully encompass the spectrum of human prostate cancer bone metastasis, novel strategies for palliative and curative treatments can be developed.

In an effort to accurately replicate different stages of tumor progression and metastasis, multiple injection models have been developed. These include subcutaneous, orthotopic, intracardiac, tail vein, and bone (Figure 1.1). The injection models can be controversial because they fail to fully recapitulate human disease by eliminating the early stages of carcinogensis; however, they remain one of the most widely used investigative techniques of cancer until more complete models are developed.

Prostate Cancer in Dogs

The dog is the only species other than man to spontaneously develop benign prostatic hyperplasia (BPH) and prostate cancer at a significant incidence [9]. However, even though intact adult male dogs have a high incidence of BPH, they have a low incidence of spontaneous prostate cancer compared to men. One study reported the incidence of prostate cancer to be 0.2% in dogs and 12% in men [10]. Unlike the single-lobed

3 architecture of the human prostate, the canine prostate is bilobed and lacks the differential regions or zones that are characteristic of the human prostate gland (peripheral, transitional, and central zones) [11, 12]. All aged, intact male dogs will develop benign prostatic hyperplasia to a variable degree [11]. BPH can be severe enough to cause clinical disease, which includes dorsal compression of the colon and dyschezia. Dysuria is not a typical clinical problem in dogs with BPH compared to men. BPH does not appear to be a preneoplastic condition of the prostate in dogs, similar to men.

Prostatic intra-epithelial neoplasia (PIN) has been reported in dogs with or without invasive carcinoma, but this has not been confirmed by multiple investigations [13]. PIN may be a preneoplastic condition in dogs, but it is not known why PIN is not commonly recognized in dogs. It may be that the incidence of both PIN and prostate carcinoma are lower in dogs compared to men. It is common practice in some countries to neuter pet dogs. Interestingly, the incidence of prostate cancer is similar or greater in neutered dogs compared to intact dogs [14]. This suggests that prostate cancer in dogs may develop from androgen-independent stem cells that are present in the atrophied ducts of the prostate glands of neutered dogs [15]. An alternative theory is that non-testicular sources of androgen in the body may be sufficient to provide androgen activity on the prostate cells.

A striking difference between men and dogs is the role of androgens in prostate cancer pathogenesis. Androgens play a prominent role in the development and function of male accessory sex organs, and the androgen signaling pathway is considered essential to the

4 development of prostate cancer in men [16]. Androgen deprivation therapy is an established early therapy for androgen-dependent prostate cancer in men, and even after progression to androgen-resistance the androgen signaling pathway is still the primary target of many therapeutics [17]. In dogs the androgen receptor is present in normal prostatic tissue, is important for normal accessory sex organ maintenance and function, and necessary for the development of BPH in adult intact male dogs. In contrast to humans, most dogs with prostate cancer have absent expression of the androgen receptor

(AR) [18]. Since castration of dogs does not decrease the incidence of prostate cancer and androgen receptor expression is not present in canine prostate cancer, it is possible that the AR does not play a central role in the pathogenesis of canine prostate cancer.

Interestingly, secretory cells of the prostatic acini are androgen dependent; however, the prostatic ducts and urothelium of the prostatic urethra are maintained independent of androgen [19]. This suggests that canine prostate cancer may originate from prostatic ductular epithelium rather than the acini as occurs in men.

In men, immunohistochemical markers are often used to distinguish between prostatic and urothelial carcinoma. A prominent example is prostate specific antigen (PSA), which is found in prostatic epithelial cells, but not transitional cells of the urothelium [20]. PSA is not produced in dogs, but instead a related kallikrein enzyme, arginine esterase, is considered to be the main secretory product of the canine prostate [21]. Unfortunately, arginine esterase does not differentiate between prostatic or urothelial origin of canine prostate cancer [22]. Another marker used often in men is prostate specific membrane antigen (PSMA). In the past it has been reported that PSMA was not expressed in the 5 canine prostate; however, recent work has shown it to be present in at least 2 canine prostate cancer cell lines (Ace-1 and Leo) [23, 24]. Prostatic acid phosphatase has also been used in men as a prostate lineage specific marker; yet investigation of its use in canine prostate cancer has not been extensively pursued [25, 26]. In men, keratin 7 expression is also used for differentiating urothelial carcinoma from prostatic adenocarcinoma. Studies have shown that keratin 7 does not differentiate between these two tumors types in dogs, making it an unsuitable marker [22, 27]. At this time there is no definitive marker in dogs that can confidently differentiate between prostate tumors originating from the acinar secretory cells and the ductular epithelium or urothelium.

Prostate cancer affects older dogs (mean age of 10 years old) and the tumors can be heterogeneous. The most common diagnosis is prostatic adenocarcinoma, but as many as

53% of canine prostate carcinomas show mixed morphologic features [12, 28]. The most common pattern of growth is intra-alveolar, where the cells form papillary or cribiform projections of epithelium filling dilated ducts or alveoli [29]. The second pattern is acinar, in which 1-2 layers of cuboidal to columnar cells line acini (though in some cases, acini can be completely filled with a solid mass of cells) [29]. Unlike in men, grading is not routine. All canine prostate cancers are considered malignant with a high metastatic potential, and no survival difference has been found for the different histologic subtypes

[30]. At the time of necropsy, it has been reported that up to 80% of dogs with a primary prostate carcinoma will have gross metastases with skeletal metastases (predominantly in the axial skeleton) present in approximately 20% of cases [28]. A large proportion of

6 canine prostate cancer skeletal metastases are osteoblastic or mixed osteoblastic/osteolytic in nature, similar to men [31].

Canines in Prostate Cancer Research

Prostate tissue-bone interaction in vivo: Since a unique characteristic of prostate cancer bone metastasis is the ability to cause dramatic new bone formation, methods of investigating the tissue-specific mechanisms responsible for the osteoblastic phenotype are vital to understanding its pathogenesis. Dogs have proven to be an excellent model for these investigations. A novel approach was utilized to investigate the prostate-specific effects on bone [32]. Canine prostate, kidney, urinary bladder, spleen, and skeletal muscle were implanted adjacent to the calvaria of immunodeficient mice in vivo. Only the prostatic tissue resulted in significant rapid new bone formation (Figure 1.2). A possible mechanism for the new bone formation was discovered through in vitro studies co-culturing rat calvaria with homogenates of normal canine prostate. These experiments revealed that osteoblasts are activated by the canine prostatic tissue through an endothelin-dependent mechanism [33]. The importance of this work is highlighted by the fact that endothelin has become an important target in human prostate cancer metastases in the last five years [34-38]. This particular model has the advantage of providing an easily manipulated system for investigating prostate-specific mechanisms of new bone formation. Another benefit is the use of prostatic tissue rather than a cell line, which may more closely represent the complex microenvironment of metastases. Canine prostate, in 7 particular, is uniquely useful since it is known to produce bioactive factors that are also produced in the human prostate and are thought to be important in the pathogenesis of osteoblastic metastases [39, 40].

Canine Prostate Cancer Cell Lines: In order to model human prostate carcinoma bone metastasis, multiple canine cell lines have been created and utilized in vivo. There are six canine prostate cancer cell lines that have been reported: DPC-1 [41], CPA-1 [42], Ace-1

[43], Leo [44], CT-1258 [45] and Probasco. To date, only DPC-1, Ace-1, Leo, and

Probasco cells have been used to study bone metastasis (Table 1.1).

DPC-1: DPC-1 originated from an 11-year-old Doberman Pinscher with a poorly differentiated prostatic adenocarcinoma [41]. The DPC-1 cells are tumorigenic in both immunosuppressed mice (subcutaneous injection) and dogs (orthotopic injection). In studies where the cells were injected into the prostate glands of immunosuppressed dogs it was found that they metastasized to the pelvic bones in 2 out of 12 dogs, forming mixed osteoblastic/osteolytic metastases [41, 46].

Ace-1: The Ace-1 cell line was developed from a primary prostatic carcinoma (intra- alveolar type) of an 8-year-old male castrated Labrador Retriever. The cells are tumorigenic in immunocompromised mice and rats and form mixed osteoblastic/osteolytic metastases after intratibial, intracardiac, and intra-vossicle injection in nude mice [43, 47-49]. In bone, the cells initially induce the formation of new woven bone on metaphyseal trabecular bone and then on periosteal surfaces in areas with osteoclastic bone resorption of cortical bone. After intracardiac injection in mice, affected

8 bones included humeri, femurs, tibias, ribs, and lumbar vertebrae. The Ace-1 cells have also been injected orthotopically in immunosuppressed dogs to develop an experimental model of prostate cancer (Figure 1.3) [50]. The Ace-1 cells are versatile and have been used both for metastasis studies and also as a model of tumor-induced bone pain [51, 52].

The Ace-1 cells are very permissive to transfection and transduction. They have been stably transduced with yellow fluorescent protein (YFP) and luciferase (luc) to enable bioluminescent imaging for easier in vivo monitoring of metastasis and growth. Ace-1 cells have also been stably transfected with human Dickkopf-1 (Dkk-1) to determine the role of Wnt signaling in prostate cancer bone metastasis [53]. It was found that Dkk-1 upregulated non-canonical Wnt/JNK signaling and increased tumor growth and the number of metastases in vivo. Ace-1 cells have also been transfected with parathyroid hormone-related peptide (PTHrP) and it was found that PTHrP expression resulted in increased subcutaneous tumor growth and increased osteoclastogenesis, osteoblastogenesis, and angiogenesis in vossicles (neonatal murine vertebrae; Figure 1.4) transplanted in nude mice [48]. Ace-1 cells have proven to be a valuable model of mixed osteoblastic/osteolytic prostate cancer metastasis.

Leo: The Leo cell line originated from a primary prostate carcinoma from a 5-year-old mixed breed dog. Similar to Ace-1, these cells have been transduced with YFP/Luc to allow for in vivo bioluminescent imaging. After intra-cardiac injection into nude mice, the most common site of metastasis was the brain and spinal cord followed by metastases to long bones [44]. Bone metastases were characterized by loss of cortical and trabecular

9 bone. Brain metastasis in prostate cancer is rare, and there are few models available to study its pathogenesis. This unique cell line has the potential to contribute to this field.

Probasco: The Probasco cells came from the primary prostate carcinoma of a 10.5-year- old castrated mixed-breed dog that had been previously treated with palliative radiation therapy and metronomic chemotherapy (piroxicam, cyclophosphamide, toceranib phosphate, and chlorambucil). Like the Ace-1 cells, these cells have been transduced with

YFP/Luc and are very permissive to transduction or transfection. Following intracardiac injection, Probasco cells metastasized primarily to the appendicular skeleton, and both intratibial and intracardiac injections produced markedly osteoblastic tumors in bone

(Figure 1.5). Metastases had extensive new bone formation on both the endosteal and periosteal surfaces. In the early intra-tibial lesions (three weeks post-injection), the new woven bone was confined to the endosteal surfaces and marrow spaces. At six weeks post-injection the Probasco cells penetrated the cortical bone and induced marked periosteal new bone formation. The overall effect on the bone was markedly osteoblastic, with mild osteolysis. These cells have been transfected with PTHrP and it was found to increased osteolysis and tumor growth in vivo. The osteoblastic nature and ease of transfection/transduction make these cells useful for studying bone metastasis pathogenesis and osteoblastic metastases.

Canine Cell Lines in Immunocompromised Dogs: Dogs provide a useful large animal model for prostate cancer because they share several anatomic and physiologic similarities with men. The larger size of dogs also makes them helpful to investigate new

10 molecular diagnostics for imaging and surgical procedures that cannot be performed readily in mice. However, the genetic heterogeneity of dogs as well as their cost can make them a challenging research model [54]. Canine cell lines have been injected orthotopically in immunosuppressed dogs to investigate their growth and metastasis in vivo and response to histotripsy. The Ace-1 cells formed prostate neoplasms in all injected dogs and metastases to the lungs and regional lymph nodes occurred in 4/10 dogs. The tumors could be imaged by transrectal ultrasound and PET [55]. It was shown that histotripsy was effective in causing necrosis of the tumor tissue [50]. The DPC-1 cells have also been successfully implanted in dogs using a CT-guided transperitoneal approach with subsequent metastasis to lymph nodes (9/12), lungs (5/12) and bones

(mixed osteoblastic/osteolytic metastasis) in 2/12 dogs [46].

Human Prostate Cancer Xenografts in Immunocompromised Mice

Much research on prostate cancer bone metastasis is performed with human prostate cancer xenografts. The most commonly used cell lines are PC3 (derived from a bone metastasis), DU145 (derived from a brain metastasis), and LNCaP (derived from a lymph node metastasis). The PC3 and LNCaP cell lines and their sublines have been widely used to study bone metastasis, as well as several less commonly used cell lines. Of the human prostate cancer cell lines available PC3, PC3M, LAPC-4, LAPC-9, LNCaP,

LNCaPc4-2, MDA PCa 2b, Wish-pc2, CL-1, CWR22R and ARCaP have been shown to produce bone metastases in vivo (Table 1.2). 11

PC3 and PC3M: PC3 cells are androgen insensitive and form osteolytic metastases after orthotopic, intra-cardiac and intratibial injection [56-58]. After orthotopic or intra-cardiac injection, sites of bone metastasis include the skull, ribs, pelvis, femur and tibia, mandible, and occasionally the brain and spinal cord. PC3 cells will also form osteolytic metastases in human fetal and adult bones implanted subcutaneously after tail vein injection of neoplastic cells [59, 60] and when cells are injected adjacent to adult human bone implanted subcutaneously in immunocompromised mice [56]. This cell line has been transduced with green fluorescent protein (GFP) and luciferase (luc), allowing for more efficient imaging of metastases. A variant of PC3 which has a higher incidence of metastasis, PC3M, also forms osteolytic metastases (frequently within the mandible) after orthotopic and intracardiac injection [61, 62]. Mandibular metastases in prostate cancer

(and cancer in general) is an uncommon, but devastating event [63].

LAPC-4 and LAPC-9: Both LAPC-4 and LAPC-9 were derived from a femoral metastasis and are androgen dependent [64]. LAPC-4 produced mixed osteoblastic/osteolytic bone tumors when injected either orthotopically or adjacent to subcutaneously implanted fragments of adult human bone [65, 66]. LAPC-9, a more recently developed cell line, has been shown to produce predominantly osteoblastic tumors after intratibial injection [56]. Since osteoblastic tumors are unusual in human cell lines, the LAPC-9 model should prove to be a valuable tool to investigate the pathogenesis of osteoblastic metastasis. For both cell lines, severely immunocompromised mice (such as NOD/SCID) must be used to allow for tumor growth in vivo. 12

LNCaP and sublines: The LNCap cell line and two of its sublines, LNCaP c4-2 and CL-

1, have been used to study prostate cancer bone metastasis in vivo. LNCaP is androgen dependent and has been shown to form mixed osteoblastic/osteolytic tumors in subcutaneously implanted human adult bone after either direct injection into human bone or tail vein injection in immunocompromised mice [60, 67]. Interestingly, LNCaP does not metastasize to murine bone after orthotopic, intracardiac, or tail vein injection [61].

The LNCaP cell line has been transfected with luciferase for bioluminescent imaging [68,

69]. The subline LNCaP c4-2 is androgen independent and has been reported to metastasize to vertebrae and form mixed bone metastases after subcutaneous, orthotopic, and intracardiac injection [57, 70]. The CL-1 subline is also androgen independent and has been reported to directly invade and induce lysis of the lumbar vertebral column after orthotopic injection. Metastases to bone after orthotopic injection have been reported using green fluorescent protein (GFP) expression, although these were not confirmed histologically [58, 71]. A potential disadvantage of LNCaP cells is that they can be challenging to work with in vivo since they are slow growing and require the use of severely immunocompromised mice (such as NOD/SCID) and/or the use of the basement membrane product, Matrigel™ (Corning, Tewksbury, MA), with the cell injections.

MDA PCa 2b: The MDA PCa 2b cells were derived from a prostate cancer vertebral bone metastasis and demonstrate androgen-sensitive growth [54, 72]. The MDA PCA 2b line forms predominantly osteoblastic tumors when injected into mouse long bones

(femurs and tibias) or when directly injected into subcutaneously implanted human adult

13 bone [73, 74]. This cell line has been used extensively to investigate androgen signaling and progression to androgen independence [75-78].

CWR22R: The CWR22R cells were derived from a primary prostatic carcinoma and are androgen independent [79]. Transfection of the CWR22R cell line with lacZ allowed for visualization of micro-metastases after subcutaneous injection; however, histology was not used to confirm the metastases and no bone pathology was reported [80].

ARCaP: ARCaP cells were derived from the ascites fluid of a man diagnosed with metastatic prostate carcinoma. Mixed osteoblastic/osteolytic metastases were reported when the cells were injected orthotopically [81].The cell line is unique in that they are characterized by an androgen-repressed phenotype (growth is suppressed when exposed to androgens).

LuCaP 23.1 and LuCaP35: Both LuCaP 23.1 and 35 are serially transplantable, prostate specific antigen-producing tumor lines that do not grow in vitro and were developed from prostate cancer lymph node metastases [82, 83]. Both are androgen sensitive; however,

LuCaP 35 can transition to androgen independence [82]. When injected intra-tibially,

LuCaP 35 forms osteolytic metastases, while LuCaP 23.1 produces predominantly osteoblastic lesions [84]. After orthotopic implantation there was PSA mRNA detected in the bone marrow by real-time polymerase chain reaction (RT-PCR), suggesting micrometastases were present [85]. A challenge with the tumor lines is the lack of in vitro capacity for propagation and hence an inability to perform in vitro assays to investigate molecular mechanisms of pathogenesis.

Wish-pc2: The Wish-pc2 tumor line was developed from a poorly differentiated prostate carcinoma. These cells are androgen independent and considered to be prostatic neuroendocrine cells and represent a model for small cell carcinoma of the prostate.

Intraosseous (femur or tibia) injection of the cells into mice results in primarily osteolytic tumors [86].

Prostate Cancer in Mice

The prostate gland of the mouse is significantly different compared to men and dogs.

The anatomy of the murine prostate consistsof 4 grossly and histologically distinct lobes

(ventral, dorsal, lateral, and anterior) unlike the single lobe of the human prostate. The ventral lobe of the mouse prostate has no human counterpart, while the dorsal and lateral lobes are analogous to the peripheral zone of the human prostate and the anterior lobe is similar to the central zone in men [87]. Wild-type mice rarely develop spontaneous prostate cancer; therefore, transgenic mice have been developed to study particular mutations or transgenes in prostate cancer pathogenesis and treatment. A potential limitation in the use of transgenic mice is that few mutations can be studied concurrently, which does not reflect spontaneous invasive and metastatic prostate cancer. However, transgenic models do permit investigations on the interactions between neoplastic cells and the surrounding stroma and the role of stroma in cancer progression. Another challenge to developing a useful transgenic mouse model for prostate cancer research is that very few develop bone metastases. Most mouse models of prostate cancer do not 15 progress beyond the stage of invasive carcinoma [88]. This is unfortunate, since most morbidity and complications in prostate cancer are due to advanced or metastatic disease, which cannot be treated surgically. However, transgenic models can be useful to study chemoprevention of prostate cancer [89-91]. Only the TRAMP, LPB-Tag/PB-hepsin,

CR2-T-Ag, and 12T-7f transgenic models have been reported to develop metastasis to bone, albeit infrequently.

TRAMP: Bone metastasis has been infrequently reported in transgenic adenocarcinoma of the mouse prostate (TRAMP) mice and only in the FBV background strain.

Uncommon skeletal metastasis (one study reported 5/20 mice) has been reported to occur at approximately 23 weeks of age in the [TRAMP X FBV] F1strain, although histology revealed only minimal osteolytic/osteoblastic response to the presence of the neoplastic cells [92-95].

LPB-Tag/PB-hepsin: Femoral metastases were found in a small percentage of LPB-

Tag/PB-hepsin mice by 21 weeks; however, there was no associated osteolysis or new bone formation [96]. It is important to note that only femoral bones were examined.

CR2-T-Ag: Cryptdin-2/sv40 T (CR2-T-Ag) transgenic mice are considered a model of neuroendocrine prostate carcinoma and are reported to occasionally develop bone metastases by 24 weeks, though associated bone pathology was not described [97].

Similarly to the LPB-Tag/PB-hepsin study, only femurs were examined for metastases.

12T-7f: 12T-7f/MT-DNIIR and 12T-7f mice are two additional models of neuroendocrine prostate carcinoma and both rarely develop bone metastases, although no associated bone pathology has been described [98]. Only lumbar spine was evaluated.

There are other transgenic mice that are reported to have bone metastases; however, this was based on positive PCR from bone marrow or from areas of bone remodeling without histologic confirmation of neoplastic cells. These include the prostate-specific PTEN knockout mouse [99] and the Gγ/T-15stains [100]. These mice were not included in this review as their usefulness in studying the pathogenesis of bone metastases is limited by the lack of histologically confirmed disease.

Mouse Prostate Reconstitution Model

The basic concept of the mouse prostate reconstitution model is to combine prostate cells

(neoplastic or non-neoplastic) with urogenital sinus mesenchyme cells (rat or mouse) and implantation of the cells beneath the renal capsule of an immunodeficient mouse [31,

101]. This model has been used to study the differentiation of putative prostate cancer stem cells, the significance of carcinogenic genes that cannot be investigated with whole- animal knockouts, as well as epithelial and mesenchymal interactions in prostate cancer

[102-104]. One model that developed frequent micrometastases to bone is mouse prostate tissue over-expressing both ras and myc co-implanted with p53 knockout mouse urogenital sinus cells [105]. While the mouse prostate reconstitution model provides a

17 fascinating avenue to further explore the genetic pathogenesis of prostate cancer progression and metastasis, its use is limited by the complexity and high technical skill required to isolate and implant the tissue.

End-Stage Mouse Models of Tumor/Bone Interactions

The ability to address specific questions regarding how tumors interact with the bone microenvironment is critical to understand the cellular and molecular mechanisms of prostate cancer skeletal morbidity. Despite the fact that bone homing and the prostate cancer metastatic cascade in mice is quite different from dogs and humans, tissue reconstitution models can address specific mechanistic questions. The intratibial injection model has the advantage of direct placement of tumor cells into bone [106]. A disadvantage is that the trauma associated with injecting the tumor cells represents an atypical metastatic dissemination with wound healing consequences that are dramatic in bone. In addition, intra-tibial injection may force some of the tumor cells directly into bone marrow sinusoids and mimic an intravenous injection. Another novel model termed the „vossicle model‟ involves co-implantation of tumor cells and neonatal vertebral bodies [106]. This relatively simple model allows combinations of tumor cells or explants with murine bone. Isolated vertebral bones can be isolated from different types of genetically modified mice and/or luciferase-expressing mice in order to determine the impact of altering a gene in bone alone without systemic impact and to monitor tumor and/or bone growth over time.

Prostate Cancer in Rats

Similar to mice, the incidence of spontaneous prostate cancer in rats is rare in most strains. Some strains of rats; however, have a notable incidence of prostate cancer. As many as 30% and 16% of aged Lobund-Wistar and ACI/Seg rats, respectively, develop prostate carcinoma, although bone metastases do not occur [107]. To increase prostate tumorigenicity in Lobund-Wistar rats, intravenous injection of N-methyl-N-nitrosourea

(MNU) with or without subcutaneous implantation of testosterone proprionate pellets has been shown to produce more rapid and consistent development of prostate cancer with metastasis to the lung and lymph nodes; however, bone metastases were not reported

[108]. Most prostate cancer bone metastasis research performed is limited to two rat models: the MATLyLu subline and the PA-III cell line.

MATLyLu: This subline was developed from an original prostate carcinoma from a 22- month-old inbred Copenhagen male rat and the designation MATLyLu is an abbreviation for Metastatic Anaplastic Tumor Metastasizing to Lymph node and Lungs. This subline is considered a poorly differentiated carcinoma, is androgen insensitive and has been used to produce osteoclastic or, with variant R3327, osteoblastic bone metastases in

Copenhagen rats after tail vein, intracardiac, and intraosseous injection [109-113]. The

„osteoblastic‟ bone metastases represented osteolytic tumors with secondary periosteal bone proliferation due cortical bone erosion. Blomme et al have reported that the bone

19 metastases that occur with MATLyLu cells are predominantly osteolytic (Figure 1.6)

[112].

PA-III: The PA-III cell line was derived from a spontaneous prostate carcinoma from a

Lobund-Wistar rat. Similar to MayLyLu, this line is considered androgen insensitive and has been used to produce mixed osteolytic/osteoblastic lesions when implanted over the calvaria or scapula after periosteal disruption [107, 114].

Conclusions

There are a considerable number of prostate cancer models available for scientific studies although relatively few of them can be used to consistently model osteoblastic bone metastases as occurs in men. Since bone metastasis represents end-stage disease, it is important that useful animal models are available to investigate novel prevention and treatment strategies. Specific animal models should be carefully chosen for individual experimental designs so that they have the pathogenic characteristics to enable them to test hypotheses that are relevant to human prostate cancer. The canine and human osteoblastic prostate cancer lines have the greatest utility for in vivo experiments; however, no model fully recapitulates the disease as it occurs in men. The characterization of novel large animal models, such as experimental dogs with orthotopic implantation of canine prostate cancer cells lines, will be helpful to develop and assess new molecular imaging tools and effective intraprostatic therapies. It is hoped that

20 additional animal models are continually developed and critically evaluated as compared to human prostate cancer so the tools are available to effectively target prostate cancer to prevent, slow, or even reverse metastatic disease.

Tables

Cell line Models Notes DCP-1 Subcutaneous xenograft in mice

Orthotopic allograft in dogs Mixed osteoblastic/osteolytic metastases to pelvic bones in allograft model

Ace-1 Subcutaneous, intratibial, Mixed osteoblastic/osteolytic tumors in intratibial, intracardiac, and intra-vossicle intracardiac and intra-vossicle models. Metastasis to long xenografts in mice bones, ribs and vertebrae in intracardiac model

Orthotopic allograft in dogs Metastasis to lymph nodes and bone in allograft model

Leo Subcutaneous, intratibial and Osteolytic tumors in intratibial and intracardiac models. intracardiac xenografts in mice Metastasis primarily to brain and spinal cord, but also long bones in intracardiac model

Probasco Subcutaneous, intratibial and Osteoblastic tumors in intratibial and intracardiac models. intracardiac xenografts in mice Metastasis primarily to long bones

Table 1.1 Canine Cell Lines that Metastasize to Bone

Cell line Models Notes

PC3 Orthotopic, intratibial, and Osteolytic tumors intracardiac injection Metastasizes to skull, ribs, pelvis, femur and tibia Human fetal and adult bone with after orthotopic or intracardiac injection tail vein injection PC3M Orthotopic and intracardiac Osteolytic tumors with metastasis primarily to injection mandible LAPC-4 Orthotopic injection Mixed osteoblastic/osteolytic tumors Human adult bone with adjacent injection LAPC-9 Intratibial injection Osteoblastic tumors LNCAP Human adult bone with intra-bone Mixed osteoblastic/osteolytic tumors or tail vain injection LNCaP c4- Subcutaneous, orthotopic, and Vertebral osteolytic metastases 2 intracardiac injection LNCaP Orthotopic injection Osteolytic tumors CL-1 Direct invasion and metastasis to skull, rib, pelvis, femur and tibia MDA PCa Intrafemoral injection Osteoblastic tumors 2b Injected human adult bone CWR22R Subcutaneous injection LacZ-positive bone micro-metastases Metastases not histologically confirmed ARCaP Orthotopic injection Mixed osteoblastic/osteolytic tumors LuCap 35 Intratibial injection Osteolytic tumors in intratibial injections Orthotopic injection Bone marrow PSA positive sites after orthotopic injection (not histologically confirmed) LuCaP Intratibial injection Osteoblastic tumors 23.1 Wish-pc2 Intrafemoral or intratibial injection Osteolytic tumors Table 1.2 Human Cell and Tumor Lines used to Model Bone Metastasis in Mice

Figures

Figure 1.1 Different injection models of prostate cancer to investigate cancer progression and metastasis.

The metastatic cascade involves the local invasion of tumor cells from the prostate, intravasation into a blood or lymph vessel, extravasation out of the circulation, and survival and growth at a secondary site

(distant metastasis). At this time, metastasis models are not able to recapitulate this process from cancer initiation to metastatic disease. Therefore, different injection models are used to investigate this complex process. Prostatic (orthotopic) injections model invasion, extravasation, and distant metastasis.

Intravascular injections (intracardiac, tail vein) model extravasation and metastasis. Tail vein injections tend to favor lung metastases, whereas intracardiac injection of cancer cells into the left ventricle permits cancer cells to localize to any tissue of the body depending on its inherent metastatic phenotype. Finally, intratibial injections model the ability to grow in and modify the bone microenvironment. Together, the different injection models reveal important information about the pathogenesis of the metastatic cascade.

Figure 1.2 Effects of canine prostatic tissue implanted adjacent to murine calvaria in vivo.

(A) Decalcified calvarium with prostate tissue implanted between the skin and calvarium (above and out of the plane of the image). Note pre-existing calvaria (C) and marked periosteal thickening caused by dramatic new bone formation (arrowheads) [H&E]. (B) Undecalcified calcein-labeled prostate-implanted calvaria. Calcein is a green fluorescent calcium-binding dye that is taken up at sites of active bone mineralization. Note calcein fluorescence in mineralized periosteal new bone (arrowheads). Pre-existing calvaria (C) had minimal ﬂuorescence [Unstained].

Figure 1.3 Intraprostatic Ace-1 tumor in an immunosuppressed dog.

Ace-1 cells formed poorly demarcated masses (right side of image) that invaded into prostatic glands and ducts. The cells were polygonal to mildly spindle-shaped (epithelial-mesenchymal transition) and formed nests and papillary projections, frequently filling acini. Larger papillary projections were associated with coagulation necrosis. There is lymphoplasmacytic inflammation and desmoplasia surrounded affected acini and between the border of the invasive neoplastic cells into the normal prostatic tissue to the left. The left side of the image is composed of numerous variably sized, hypercellular glands lined by tall columnar to cuboidal epithelium and occasional prominent intra-luminal papillary projections (benign prostatic hyperplasia). [H&E]

Figure 1.4 Nude mouse subcutaneously transplanted vertebral vossicle with canine Ace-1 tumor cells (T).

The vossicle has cortical and trabecular bone with normal bone marrow (BM) hematopoietic cells (darkly stained cells). Note the new trabecular bone formation (arrowhead) induced by the Ace-1 tumor cells that surrounds the tumor cells. The image on the right is a magnification of the hyperplastic trabecular bone

(arrowheads) induced by the Ace-1 tumor cells (T). [H&E]

Figure 1.5 Intratibial Probasco tumor in a nude mouse.

(A) Micro-computed tomography (µCT) and of a nude mouse tibia injected with Probasco cells (right) and control tibia (left) at 3 weeks. Tibias in the Probasco-injected group had dramatic new bone formation signified by the increased radio-opacity in the diaphyseal region compared to the control bone. (B)

Photomicrograph of intratibial tumor. Densely packed large polygonal neoplastic cells formed sheets and irregular glands with surrounding large irregular trabeculae of induced woven bone. Numerous polygonal to cuboidal osteoblasts line the trabecular bone and in some areas there is apposition of the prostate cancer cells to the new woven bone surfaces. Mitotic figures are frequent. [H&E] 28

Figure 1.6 MatLyLu vertebral metastasis in a Copenhagen rat.

There is marked replacement of hematopoietic cells with anaplastic carcinoma cells with dramatic loss of the medullary bone. The surfaces of the remaining trabecular bone are scalloped and eroded and they have severely eroded endosteal surfaces due to osteoclastic bone resorption [H&E]. Inset: Numerous multinucleated osteoclasts line the scalloped surface of an eroded trabeculae [H&E].

Chapter 2 Canine Prostate Cancer Cell Line (Probasco) Produces Osteoblastic Metastases

in Vivo

Abstract

In 2012, over 240,000 men were diagnosed with prostate cancer and over 28,000 died from the disease. Animal models of prostate cancer are vital to understanding its pathogenesis and developing therapeutics. Canine models in particular are useful due to their similarities to late-stage, castration-resistant human disease with osteoblastic bone metastases. This study established and characterized a novel canine prostate cancer cell line that will contribute to the understanding of prostate cancer pathogenesis.

A novel cell line (Probasco) was derived from a mixed breed dog that had spontaneous prostate cancer. Cell proliferation and motility were analyzed in vitro. Tumor growth in vivo was studied by subcutaneous, intratibial, and intracardiac injection of Probasco cells into nude mice. Tumors were evaluated by bioluminescent imaging, Faxitron

30 radiography, µCT, and histology. RT-PCR and genome-wide DNA copy number profiling were used to characterize the cell line.

The Probasco cells grew in vitro (over 75 passages) and were tumorigenic in nude mice.

Probasco cells expressed high levels of BMP2, CDH1, MYO1, RUNX2, and SMAD5 and low levels of CXCL12 and SLUG mRNA. Following intracardiac injection, Probasco cells metastasized primarily to the appendicular skeleton, and both intratibial and intracardiac injections produced osteoblastic tumors in bone. Comparative genomic hybridization demonstrated numerous DNA copy number aberrations throughout the genome, including large losses and gains in multiple chromosomes.

The Probasco prostate cancer cell line will be a valuable model to investigate the mechanisms of prostate cancer pathogenesis and osteoblastic bone metastases.

Introduction

The American Cancer Society estimates that approximately 1,660,290 Americans will be diagnosed with cancer in 2013. Approximately 238,590 of those cases will be prostate cancer [115]. Prostate cancer affects 1 in 6 men in the western world, and 17% of affected men will die as a sequelae to metastatic disease [28]. The most common site of metastasis is bone, and approximately 90% of patients with metastatic disease will develop bone metastases that are predominantly osteoblastic or bone-forming [116, 117].

Even though the pathogenesis of prostate cancer has been studied for over 70 years, few

31 experimental models of osteoblastic skeletal metastasis have been developed [118].

Animal models of prostate cancer that mimic disease in men are vital for translational research and developing new diagnostics and therapeutics to improve patient quality of life.

To understand the pathogenesis of prostate cancer, animal models must reflect one or more stages of the human disease. This includes slow growth, increased incidence with age, prostatic intraepithelial neoplasia (PIN), initial androgen sensitivity followed by castration resistance, and development of predominantly osteoblastic bone metastases in late-stage disease. At this time, there is no animal model that encompasses all of these features; however, dogs are the best available model for spontaneous disease. Aside from humans, dogs are the only animal to naturally develop prostate cancer; however, the incidence is lower than in men [14]. Dogs develop prostate cancer late in life and the tumors often develop in the face of previous castration, indicating castration resistance

[55]. By the time prostate cancer is diagnosed in dogs, as many as 80% will have metastases and, similar to men, the metastases are often in bone and are typically osteoblastic [30]. For this reason, canine prostate cancer is a relevant model of late-stage human disease.

There are few canine prostate cancer cell lines used in research [41-45]. We report the formation and characterization of a novel canine prostate adenocarcinoma cell line originating spontaneously from a mixed-breed dog. The Probasco cells have unique morphologic features that include robust growth in vitro, slow but consistent growth in

32 vivo and dramatic osteoblastic bone metastases in nude mice. These characteristics make the Probasco cells a valuable resource in prostate cancer research.

Materials and Methods

Establishment of canine prostate carcinoma cell line (Probasco)

A section of tissue was aseptically removed from a primary prostate carcinoma of a 10.5- year-old castrated mixed-breed dog that had been previously treated with palliative radiation therapy and metronomic chemotherapy (piroxicam, cyclophosphamide, toceranib phosphate, and chlorambucil). The clinical disease was approximately nine months, and at the time of tumor sampling there was a suspected right hind limb metastasis that was not confirmed due to elective euthanasia. The tumor specimen was rinsed three times with sterile Dulbecco's phosphate-buffered saline (DPBS) and minced to approximately 1mm3 fragments with scalpels. The minced tumor was then cultured in

DMEM/F12 with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a

25cm2 flask. Medium was replaced every four days. Differential trypsinization with

0.25% trypsin ethylenediamine-tetraacetic acid (EDTA) was used to remove stromal cells, which released from the culture flasks earlier than the epithelial cells. Detachment of the cells was monitored every 2 min to remove the stromal cells. Once this population was removed, the remaining epithelial cells were collected and replated. The epithelial cells were allowed to become 70% confluent before passaging.

Lentiviral luciferase transduction

The Probasco cells were grown in a 6-well culture plate with DMEM/F12, 10% FBS and

1% penicillin/streptomycin. Once the cells had reached approximately 95% confluence, the media was removed and replaced with 500µL of luciferase-containing virus (VC

2192 ConcpLuc (VSV-G)), 1.5mL of culture medium, and 1.6µL of polybrene stock

(8µg/mL). YFP-Luc lentiviral particles were produced by transient co-transfection of

293T cells with 10 mg packaging plasmid pCMVDR8.2, 2 mg envelope plasmid pMD.G

(obtained from Dr. Kathleen Boris-Lawrie, Ohio State Univ.) and 10 mg transfer plasmid pHIVSIN-YFP-Luc using calcium phosphate (Sigma–Aldrich, St. Louis, MO) as previously described [119]. Following centrifugation at 2700 rpm for 60 minutes at 30°C, the plate was placed in a cell culture incubator at 37°C and 5% CO2 for 24 hours, after which the virus-containing medium was aspirated and replaced with fresh culture media without virus. After 48 hours the cells were typsinized into a 25cm2 flask and cultured as described above.

Subcutaneous, intratibial, and intracardiac injection of Probasco cells into nude mice

All animal experimental procedures were approved by The Ohio State University

Institutional Laboratory Animal Care and Use Committee.

Subcutaneous: Five nude mice were anesthetized in an induction chamber with 3% isoflurane. The dorsal thoracic skin was scrubbed with alcohol-soaked gauze to remove superficial contamination. A suspension of 500,000 luciferase/YFP-transduced Probasco cells (Probasco-Luc) in 0.25mL of sterile DPBS was injected subcutaneously between the

34 shoulder blades using a 25-gauge needle. Tumor volume was measured weekly using calipers. Measurements were taken in three dimensions and tumor volume was calculated using the formula length x width x height x 1/2. Tumor growth was also monitored weekly using bioluminescent imaging.

Intratibial: Fifteen nude mice were anesthetized in an induction chamber with 3% isoflurane and maintained with 2.5% isoflurane. The right rear limb was scrubbed with alcohol-soaked gauze. The leg was held so the knee joint was at a 90° angle and a 27- gauge needle was introduced through the patellar ligament and into the tibial marrow space through the articular cartilage. Using a Hamilton syringe, 0.025mL of sterile DPBS containing a suspension of 200,000 Probasco-Luc cells was introduced into the marrow cavity. Tumor growth was monitored weekly using bioluminescent imaging.

Intracardiac: Eight nude mice were induced at 3% isoflurane gas anesthesia and maintained at 2.5%. The mice were placed in dorsal recumbency and their front limbs were placed perpendicular to their thorax and restrained to the procedure table with surgical tape. The ventral thorax was scrubbed with alcohol. A 1mL syringe was loaded with 0.1mL of DPBS containing a suspension of 300,000 cells. The syringe with a 25- gauge needle which was introduced into the thoracic cavity through the third intercostals space, approximate 1mm to the left and lateral from the sternum. Once a pulsatile jet of blood was present in the hub of the needle, the cell suspension was slowly injected into the left ventricle of the heart for 30 sec. Tumor growth was monitored weekly using bioluminescent imaging.

Bioluminescent imaging

Mice were anesthetized in an induction chamber with 3% isoflurane and maintained with

2% isoflurane. A 1ml syringe was used to inject 0.15ml of sterile DPBS containing

4.5mg D-Luciferin (Caliper Life Sciences) into the peritoneal cavity. Bioluminescent in vivo imaging was performed using the IVIS 100 (Caliper Life Sciences, Hopkinton, MA) and photon signal intensity was quantified using LivingImage software version 2.50

(Caliper Life Sciences). Imaging was performed every 2 minutes until peak photon signal was achieved (approximately 10 minutes post-injection). The IVIS 100 was set to 1 minute exposures with small binning.

In vitro growth rate

Probasco-Luc cells from passage 25 were plated in a six-well cell culture plate in triplicate at a density of 200,000 cells per well. Cells were cultured in DMEM/F12 with

10% FBS and 1% penicillin/streptomycin and incubated in 37°C and 5% CO2. The cells were collected using 0.25% trypsin at 24, 48, and 72 hours after the initial plating and counted with an automated cell counter (Nexcelom Bioscience, Lawrence, MA) using trypan-blue dye exclusion to differentiate between live and dead cells. The doubling time was calculated using the formula: (t2- t1) x log(n2)/log(n2/n1), where n is the cell number at time points (t).

Histopathology and immunohistochemistry

Complete necropsies were performed on the mice, subcutaneous tumors were weighed, and tissues were fixed in 10% neutral-buffered formalin at 4°C for 24 hr. Bones were decalcified in 10% EDTA (pH 7.4) for two weeks at 48°C and embedded in paraffin. The

36 specimens were sectioned (5 μm) and were either stained with hematoxylin and eosin

(H&E) or evaluated by immunohistochemistry against cytokeratins AE1/AE3, 5/6, 7, and

20 (Table 2.1).

RNA extraction and reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted from Probasco cells at passages 4, 5, and 6 and Ace-1 canine prostate carcinoma cells at passages 22, 23 and 24 using the Absolutely RNA Miniprep

Kit (Stratagene, Cat. No. 400800) and the manufacturer‟s protocol for adherent tissue culture cells. Total RNA (2.5 μg) was reverse transcribed using the Superscript II First

Strand cDNA synthesis kit (Invitrogen) and RT-PCR was performed for the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as well as the following genes related to bone microenvironment and prostate cancer progression and metastasis: bone morphogenic protein-2 (BMP2), cadherin-1 (CDH1), chemokine ligand

12(CXCL12), folate hydrolase 1(FOLH1), myoferlin (MYOF), runt-related transcription factor 2 (RUNX2), snail homolog 2 (SLUG), parathyroid hormone-related peptide

(PTHrP), androgen receptor (AR)and Mothers against decapentaplegic homolog 5

(SMAD5) using canine specific primers (Table 2.2).

DNA extraction and polymerase chain reaction (PCR)

Total DNA was extracted from Probasco cells at passage 6 using the DNeasy Blood &

Tissue Kit (Qiagen, Cat. No. 69581) and the manufacturer‟s protocol for adherent tissue culture cells. PCR was performed for the androgen receptor using canine specific primers

(Table 2.2).

Radiography

Radiography was performed postmortem on the mice that received intratibial and intracardiac injections. Formalin-fixed tissues were placed centrally on a Faxitron laboratory radiography system LX-60 (Faxitron X-ray Corp., Wheeling, IL) imaging platform and high resolution radiographs were taken at 25KV and 5 second exposures.

Micro-computed tomography (µCT)

The tibias were removed postmortem from mice injected intratibially and imaged using microcomputed tomography (µCT) (Siemens, Inveon Preclinical CT scanner, Siemens

AG, Munich, Germany). Images were acquired in 400 projections over 360 degrees at

100 kVp, 200 MA, 900 millisecond exposure, Bin 2, and a medium-high system magnification with a pixel width of 19 μm. Image data were reconstructed using Cobra software (Exxim, Pleasanton CA) and analyzed using 3D analysis software (Inveon

Research Workplace 3-Dimensional Image Software, Siemens PreClinical).

Segmentation thresholds were kept constant for all bones. Five mice were euthanized at three weeks post-injection to characterize early lesions and other mice were euthanized at six weeks (late lesions).

Analysis of genomic imbalances by comparative genomic hybridization (CGH) and fluorescent in situ hybridization (FISH)

Metaphase chromosome preparations and genomic DNA were obtained simultaneously from the same flask of cells. The cells in the flask were grown to 80% confluence and exposed to 50µg/ml of Karyomax (Invitrogen, Carlsbad, CA, USA) for four hours prior to routine recovery from the flask with conventional trypsin-EDTA. Pelleted cells were

38 divided equally into two tubes, one used to isolate genomic DNA for array comparative genomic hybridization (aCGH) and the second harvested for chromosome preparations by routine procedures [120, 121].

Chromosome preparations were isolated following conventional procedures of hypotonic treatment and fixation in 3:1 methanol:glacial acetic acid. To identify the position of the centromeres of the chromosomes, two BAC clones (330E21 and 326K03) from the

CHORI-82 BAC library, which have been shown previously to hybridize to all autosomal centromeres of the canine genome, were both labeled with SpectrumGreen-dUTP and hybridized to chromosome preparations of this cell line according to routine procedures

[122]. Chromosome preparations were counterstained with 80 ng/ml 4',6-diaminidino-2- phenylindole (DAPI) and mounted in antifade solution (Vectashield, Vector

Laboratories). Images were acquired with a fluorescence microscope (Axioplan 2ie,

Zeiss) equipped with a DAPI filter set and a cooled CCD camera (CoolSnapHQ,

Photometrics, Tuscon, AZ) both driven by dedicated software (SmartCapture 3, Digital

Scientific, Cambridge, U.K.).

Comparative genomic hybridization analysis: CGH was performed using a commercial

~180,000-feature microarray comprising repeat-masked ~60-mer oligonucleotides distributed at approximately 13kb intervals throughout the dog genome sequence assembly (canFam version 2.0, May 2005, [123]) (Agilent Technologies, Santa Clara,

CA). All steps were performed as described previously [121]. Briefly, DNA isolated from the cell line was labeled with cyanine-3dNTP and hybridized against a reference DNA sample (comprising equimolar quantities of constitutional DNA from 10 clinical healthy

39 male dogs) labeled with Cyanine-5dNTP. Following hybridization and washes the array image file was processed using Feature Extraction version 10.10 (Agilent Technologies) and imported into Nexus Copy Number version 7 (BioDiscovery, El Segundo, CA). DNA copy number calling was performed using the FASST2 segmentation algorithm in Nexus

Copy Number, using at least three consecutive probes with log2tumor: reference values of

≥0.201 or ≤-0.234 to indicate copy number gain or loss. Genomic recoding of the canine genome coordinate to represent the corresponding human genome locations was performed as described previously [124]. The resulting „humanized‟ data were output simultaneously to facilitate direct comparison with CGH data already available for human prostate specimens.

Results

Histopathologic characterization of the primary carcinoma and subcutaneous xenograft

The primary prostate carcinoma was a moderately differentiated alveolar-type prostatic carcinoma based on the light microscopic morphology. The carcinoma was composed of numerous ducts filled with polygonal epithelial cells arranged in nests and anastomosing cords (Figure 2.1A). The cancer cells contained moderate homogenous eosinophilic cytoplasm and a single, round to oval nucleus with finely stippled chromatin and 1-2

40 nucleoli. Desmoplasia was present throughout the tumor, as well as multifocal areas of necrosis and lymphocytic inflammation.

Probasco xenografts were grossly visible in the subcutis of all mice between 2-4 weeks after implantation. Mice were euthanized seven weeks after implantation, and tumors ranged in size from 170mm3 to 430mm3.The histologic appearance of the xenograft was similar to the primary carcinoma and was comprised of polygonal epithelial cells forming nests and anastomosing cords in a fine fibrovascular stroma with moderate to severe multifocal coagulation necrosis (Figure 2.1B). Pseudo-cyst formation was common in the tumor secondary to necrosis, and the neoplastic cells lined cavities containing sloughed cells. Occasionally, small foci of mineralization and ectopic bone formation were present.

Subcutaneous tumors were not highly invasive and were easily dissectible from the subcutaneous space.

Immunohistochemistry

The Probasco cells were positive for pan-cytokeratins AE1/AE3 and 5/6 and negative for cytokeratins 7 and 20 (data not shown).

In vitro and in vivo growth

The Probasco cells grew in an anchorage-dependent manner in vitro and formed monolayer sheets with polygonal cells exhibiting a tightly packed cobblestone growth pattern (Figure 2.1C). Probasco cells grew exponentially in vitro with a doubling time of approximately 18 hours under standard culture conditions (Figure 2.2A). In vivo, the doubling time of the viable cells in subcutaneous xenografts was approximately 2.5

41 weeks as estimated by BLI (Figure 2.2B). The subcutaneous xenograft volume doubled approximately every 1.4 weeks (Figure 2.2C). qRT-PCR and PCR of Probasco in vitro

The Probasco cells expressed high CDH1 (1700x), MYO1 (250x), FOLH1 (130x),

SMAD5 (100x), RUNX2 (50x), BMP2 (50x), and SLUG (10x) relative to GAPDH

(Figure 2.3A). There was little to no expression of CXCL12. Probasco cells expressed significantly greater BMP2, CDH1, MYOF, RUNX2, and SMAD5 mRNA compared to a second canine prostate carcinoma cell line, Ace-1 (Figure 2.3B). The expression of

CXCL12, PTHrP and SLUG were significantly lower in Probasco cells, and the expression of FOLH1 was similar between the two cell lines. Neither cell line expressed androgen receptor mRNA, though the gene was present by PCR (data not shown).

Intratibial Injections of Probasco in nude mice

Fourteen of 15 mice developed successful xenografts in the tibias. Radiographs and µCT imaging demonstrated extensive new bone formation on both the endosteal and periosteal surfaces of the tibias (Figures 2.4A and 2.4B). The new bone formed a radio-opaque starburst pattern of woven bone emanating from the periosteal surface, while the endosteal new bone presented as sclerosis of the marrow cavity with loss of distinctive trabeculae. In the early lesions (three weeks post-injection), the new woven bone was confined to the endosteal surfaces and marrow spaces. At six weeks post-injection the

Probasco cells penetrated the cortical bone and induced marked periosteal new bone formation. The overall effect on the bone was markedly osteoblastic, with mild osteolysis. Histology of the tibia demonstrated large multifocal to coalescing islands of

42 woven bone lined by both numerous cuboidal osteoblasts and polygonal neoplastic cells

(Figure 2.5). Large intratibial tumors formed pseudocystic cavities that contained necrotic cell and proteinaceous (eosinophilic) fluid.

Intracardiac injections (IC) and in vivo bioluminescent imaging (BLI) of Probasco in nude mice

Six of eight mice developed bone metastases (ranging from 1-5metastases per mouse) after IC injection with Probasco cells. At 5 minutes after injection of Probasco cells and

IP luciferin administration, the bioluminescence was diffusely distributed throughout the body, with the highest signal within the thoracic cavity. Two mice had bioluminescence within the forelimbs at day seven. By two weeks, six of the eight mice had bioluminescence within in the appendicular skeleton. The intensity of the BLI increased throughout weeks 3 and 4, reflecting increased tumor burden. At the end of week 4, bioluminescence demonstrated metastases in the femurs of two mice, tibias in six mice, radius/ulnar metastases in five mice, and vertebrae in four mice. One mouse with vertebral tumor growth had bilateral hind limb paralysis at the time of euthanasia.

Faxitron radiography demonstrated numerous radio-opaque radial spicules of periosteal new bone formation and medullary sclerosis in both the metaphysis and diaphysis of long bones, with minimal osteolysis.

Analysis of genomic imbalances by CGH and FISH

Analysis of >50 DAPI stained metaphase preparations indicated that while the cell line was not clonal, all of the cells evaluated were hyper diploid, containing between 88 and

108 chromosomes, with a mean of 92. Hybridization of BAC cones targeting the

43 centromeres of all canine autosomes [125] revealed that while most of the chromosomes were single armed, approximately 20% were bi-armed, indicative of centric fusions

(Figure 2.6). Centromeres were not detected on either of two grossly different bi-armed chromosomes, suggesting that these two aberrant structures may contain centromeres from the sex chromosomes, even though neither resembled a „normal‟ canine X or Y chromosome.

The genome wide profile of DNA isolated from passage four of the Probasco line revealed numerous copy number aberrations, representing whole (or almost whole) chromosomes (e.g., gain of dog chromosome (CFA) 4, 11, 20, 24 and 26 and loss of

CFA 7, 19, 27, 33, 34 and 36), large chromosome segments (e.g. gain of large regions of

CFA 9, 13, 14, 18 and loss regions of CFA 1, 3, 12, 13, 16, 22, 31 and Xq) (Figure 2.7).

In addition there were numerous focal copy number changes involving gains of small regions of CFA 2, 13, 20, 34 and 36, and losses of small regions of CFA 8 and 28 (Figure

2.7). The complex pattern of adjacent DNA copy number gain and loss along the CFA 13 and Xq is suggestive that these chromosomes may be involved in complex translocations events, which would merit further investigation. Recoding of the canine data to output the profile in the context of the human genome is presented in Figure 2.8.

Discussion

The creation of cell lines and animal models that closely recapitulate prostate cancer progression in men is essential for understanding prostate cancer pathogenesis and the development of novel therapeutics [21, 22]. Dog are unique in the animal kingdom because they develop spontaneous prostate cancer with an increasing incidence with age and frequently develop osteoblastic bone metastases [8]. In this study, we describe the development and characterization of a novel canine prostate cancer cell line (Probasco).

There are six canine prostate cancer cell lines that have been reported (DPC-1, CPA-1,

Ace-1, Leo, CT-1258 and Probasco), each with unique characteristics that contribute to prostate cancer research. The Leo cell line produces osteolytic bone metastases and, more importantly, consistently metastasizes to the brain and spinal cord after intracardiac injection [9]. The Ace-1 cell line demonstrates marked epithelial-mesenchymal transition and produces mixed osteoblastic and osteolytic metastases [10]. The Ace-1 cells have been used to develop an experimental model of prostate cancer in immunosuppressed dogs to investigate molecular imaging, focal therapy of prostate cancer, and metastasis to lymph nodes, lungs and bone [23]. The DPC-1 cell line is tumorigenic in both mice and dogs and produces undifferentiated prostate carcinoma xenografts [12, 24]. In immunosuppressed dogs, DPC-1 metastasizes to lung, lymph node, and produced mixed osteoblastic and osteolytic bone metastases [24]. The CPA-1 cell line, while not extensively characterized, is tumorigenic in nude mice and produces well differentiated prostate carcinoma xenografts [13]. The CT-1258 cell line produces tumors when

45 injected subcutaneously or intraperitoneally in immunodeficient mice and does not metastasize [11]. Of all the canine prostate cancer cell lines, the Probasco cell line is the only one to produce predominantly osteoblastic bone metastases. As men with prostate cancer bone metastases most commonly develop osteoblastic lesions, so the Probasco cells represent a uniquely valuable tool for evaluating this particular aspect of prostate cancer pathogenesis.

There are few animal models of prostate cancer bone metastasis that closely mimic human disease. Most currently available engineered mouse models of prostate cancer do not develop bone metastases, and xenograft models of prostate cancer bone metastasis often result in the formation of mixed or osteolytic metastases [25]. The two most commonly used human lines that stimulate new bone formation include the human

LNCap subline C4-2B and the LuCaP 23.1 tumor line; however, both of these form mixed metastases as they induce significant bone destruction as well as formation [26,

27]. A less commonly used human cell line, the MDA PCA 2b cells, forms predominantly osteoblastic lesions when injected into mouse long bones, however it takes significantly longer (60 days) than the Probasco cells (21 days) for osseous lesions to become evident [28]. It is expected that ability of the Probasco cells to mimic the osteoblastic nature of human prostate cancer bone metastases will further the understanding of the pathogenesis of osteoblastic metastases in prostate cancer, creating the potential for novel therapeutics and imaging modalities.

The pathogenesis of bone metastases has been intensely investigated over the last decade, but there is still much to understand. Cell lines that actively metastasize and grow in bone

46 are vital to advancing our understanding of this untreatable disease. Comparing cell lines that produce osteoblastic metastases to cell lines that produce mixed or osteolytic lesions may help identify important pathways in the molecular pathogenesis of osteoblastic prostate cancer bone metastases. As previously mentioned, the Ace-1 canine prostate cancer cell line (developed by our lab) metastasizes to bone and forms mixed osteoblastic/osteolytic lesions [10]. This is in contrast to the osteoblastic lesions induced by the Probasco cells. When comparing gene expression between these two cell lines, several genes were identified that may provide insight into the reasons for their different metastatic phenotypes. One such gene is BMP2. Interestingly, Probasco has a much higher expression of BMP2 compared to Ace-1. Bone morphogenic proteins are growth factors of the transforming growth factor-beta (TGF-β) superfamily that transduce signals through SMADs and have a critical role in bone formation [29]. BMP signaling activates

SMAD1/5/8, resulting in upregulation of transcription factor and coactivators that are important in osteoblast differentiation (i.e., Osterix and RUNX2) and new bone formation

[30, 31]. The elevated expression of BMP2 may play a role in the osteoblastic nature of

Probasco bone metastases. Another gene is CXCL12, a chemokine ligand that regulates proteases such as MMP9. The decreased expression of CXCL12 in the Probasco cells may in part account for the reduced osteolysis in bone metastases compared to the Ace-1 cells [32].

Osteomimicry has been implemented as an important feature in bone metastasis for over a decade, and has been found to be a feature of prostate cancer bone metastasis in men

[33]. The theory of osteomimicry proposes that neoplastic cells express bone matrix

47 proteins or osteoblast-related factors to enhance their survival and proliferation in bone.

One such factor is the osteoblast transcription factor, RUNX2 [34]. RUNX2 targets pathways such as Wnt, Src, BMP, and TGF-β and is considered a key regulator of metastasis-related genes in human prostate cancer [31]. Elevated RUNX2 expression has been documented in advanced human prostate cancer compared to early lesions [35]. The elevated expression of RUNX2 in the Probasco cells supports its use as a model of late- stage human prostate cancer and potentially provides a model to target osteomimicry in the development of therapeutics.

Both Probasco and Ace-1 have similar levels of expression of FOLH1. This gene, also known as prostate-specific membrane antigen, has been under intense investigation in recent years for its use as a target for prostate cancer imaging by positron emission tomography (PET), single photon emission computed tomography (SPECT), and transrectal gamma imaging (TRGI) [36-39]. FOLH1 is overexpressed in most human prostate cancers, and a similar overexpression in canine prostatic carcinoma cell lines makes them a useful tool in validating such imaging studies.

The role of parathyroid hormone-related peptide in prostate cancer is not clearly defined.

In other tumors, such as breast and lung cancer, PTHrP plays significant role osteolytic metastases. In these tumors, PTHrP is thought to be one of the main drivers of the

„vicious cycle‟ from its ability to enhance osteoclastogenesis and bone lyses [40]. The function of PTHrP in prostate cancer is less clear, though the predominantly osteoblastic nature of metastases suggests it plays a different role in the progression of bone metastases. One study has found that PTHrP overexpression in the Ace-1 cell line is

48 linked to increased osteoblastogenesis, osteoclastogenesis, and angiogenesis as well as increased bone turnover [41]. Interestingly, reports of PTHrP expression in prostate cancer bone metastases are conflicting, with several reports claiming its presence in

100% of human prostate cancer bone metastases and others reporting a less universal expression [42, 43]. The Probasco cell line may be useful as a model of bone metastases that have little to no expression of PTHrP, perhaps highlighting the importance of other pathways in disease progression.

Epithelial-mesenchymal transition (EMT) is a well known characteristic of invasive tumors. EMT is often characterized by loss of expression of cell-cell adhesion molecules like E-cadherin and increased expression of mesenchymal proteins, such as vimentin

[44]. The relatively high expression of CDH1 in Probasco cells compared to Ace-1 suggests that it has not undergone extensive EMT. This is supported by its polygonal cell shape both in culture and in vivo. The lack of prominent EMT may explain why the subcutaneous Probasco xenografts are minimally invasive and easy to dissect from the mouse when compared to Ace-1 xenografts, which are reported to invade into the underlying epaxial musculature, ribs and vertebrae [10].

Androgen receptor signaling is known to be a strong driver of prostate cancer progression in men; however, this does not seem to be the case in dogs. The lack of expression of the androgen receptor (AR) by Probasco is consistent with previous data stating that AR is not present in most canine prostate carcinomas [45]. Both the lack of AR expression and the frequent development of prostate cancer in neutered dogs suggest prostate cancer in dogs rapidly progresses to androgen independence. It is interesting to note that prostate

49 basal and stem cells are androgen independent and are present in the prostate gland of neutered dogs [6]. Investigations into the pathogenesis of prostate cancer in dogs will be valuable in developing novel therapeutics for the disease in dogs and men.

The genomic determinants for aggressiveness of prostate cancer are still poorly understood. Copy number variants in the genome are known to modulate cell function and disease. In recent years, several copy number variants have been shown to be associated with aggressive prostate cancer and poorer prognosis, including deletions in

1p36.13, 3p25, 3p14.2, 14q32.33, 15p11, and 20p13 and duplications in 12p13, 16q23.3, and 17p13.3[46, 47]. Array CGH analysis revealed that the Probasco cell line has numerous copy number variants including deletions in 1p36.13 and duplications in

17p13.3. These findings highlights the usefulness of the Probasco cell line as a model of human disease, as even on the genomic level there are similarities between the tumors in the two species.

Previous studies of high resolution aCGH of human prostate specimens revealed numerous genome copy number aberrations associated with high-risk disease [48, 49].

Among these aberrations the most frequent DNA copy number aberrations involve HSA

8, with gains of 8q (generally involving 8q21.3-24.3 (containing c-MYC) and deletions of 8p (generally involving 8p21). Recoded CGH data for Probasco revealed that humanized data did not reveal copy number changes in the regions that are evolutionarily conserved with HSA 8p, but did reveal complex copy number changes in the region of

8q23-24. The region of HSA 1 at 1q42.12-q42.3 demonstrated DNA copy gain in a third of human prostate cases [48]. It is interesting to note that not only is this region of the

50 genome also gained in Probasco, but that it is flanked by adjacent regions of DNA copy number loss (Figure 2.8), perhaps suggestive a region that has retained a critical association with the cancer across millions of years of speciation. Further consideration of such conserved aberrations, shared between human and dog prostate neoplasms, may provide a means to reduce the regions of significance in the human genome as we move towards better refining key genetic pathways and improved treatments.

Conclusions

Bone metastases are a devastatingly common complication of late state prostate cancer.

We have developed a canine prostate carcinoma cell line and xenograft model that preferentially metastasizes to bone and results in marked new bone formation. This model system will be vital in the investigations of the pathogenesis of osteoblastic metastases in prostate cancer.

Tables

Antibody Species Clone Company Dilution CK AE1/AE3 Mouse M3515 Dako 1:100 Corporation, Carpinteria, CA CK5/6 Mouse D5/16 B4 Dako 1:50 Corporation, Carpinteria, CA CK7 Mouse OV-TL 12/30 Dako 1:150 Corporation, Carpinteria, CA CK20 Mouse Ks20.8 Dako 1:250 Corporation, Carpinteria, CA Table 2.1 Antibodies used to immunohistochemically characterize the Probasco cell line.

qRT-PCR Primers Gene Forward Primer (5‟-3‟) Reverse Primer (5‟-3‟) BMP2 TGCGCAGCTTCCACCACGAAGAA CAAAGGTTCCTGCATCTGTTCCCGA CDH1 GCTGCTGACCTGCAAGGCGA GGCCGGGGTATCGGGGACAT CXCL12 ACTGTGCCCTTCAGATCGTGGCA CCACCTGCGCCTCTCACATCTTG FOLH1 GCAGGGGACCCTCTCACACCTG CTCGGAAGACCAACAGCCTCTGTGA GAPDH CCCACTCTTCCACCTTCGAC AGCCAAATTCATTGTCATACCAGG MYOF TGCCCCCGAAAGGCTGGGAAT ACTCCGTGTGCCCTGCGTCT RUNX2 TGCCTCTGGCCTTCCACTCTCAG TGCATTCGTGGGTTGGAGAAGCG SLUG GGCAAGGCGTTTTCCAGACCCT GGGCAAGAAAAAGGCTTCTCCCCAG SMAD5 CCCAGCCAATGGATACAAGC ATGCATGAAAAGCCTCCCCA PTHrP AGCTCGGCCGCCGGCTCAA GGAAGAATCGTCGCCGTAAG AR CACTCGGCCACCTCAGGGATTGG AACATGGTCCCTGGCAGTCTCCA PCR Primers AR CCTTTCCAGGAGCTTCCCGCTG CAGCGGAGCGCAAGGACGTA Exon 1- 1 AR GACAGCGCCAAGGAGTTGTGTAAGG GACAGCGCCAAGGAGTTGTGTAAGG Exon 1- 2 AR TAGCCCATGAGAGTGGCTGGTTCC AGGACTGCCCCAGGCTCTAAGC Exon 2 AR TGGCCATTCTCTGTTCACTTCTGTC AGCTGGGTGATGGCCACATAGGA

Exon 3 AR GCTTAGGCAACCTTTCTCACAGGG AGTGGGCATTATACCAATCCCTCCC Exon 4 AR TGCTCCTCAGCATTAATGGCCCCT CCAATCCCCCATTGCTTCCCACC Exon 5 AR GAGGGCACAGGGACCTGGAAATTC AGGAGCTGGCTTCTCCCCGAT Exon 6 AR TGACTGAAGGACCCAGGCGTACA CACCACCCAACAAGAGGTGCCA Exon 7 AR TGCATGCGCAAACACAAACACGT TCCCCAAGGCACTGCAGAGTAGT Exon 8 Table 2.2 The list of primers used for PCR and quantitative RT-PCR to characterize the Probasco cell line.

Figures

Figure 2.1Photomicrographs of the primary prostatic carcinoma, the subcutaneous xenograft, and the in vitro cell line (Probasco).

(A) Photomicrograph of hematoxylin and eosin (H&E)-stained primary prostatic carcinoma. The polygonal neoplastic cells form nests and chords separated by fibrous connective tissue. (B) Photomicrograph of

H&E-stained subcutaneous Probasco xenografts. Tumors were multilobular and contained multifocal necrosis and foci of ectopic ossification. (C) Phase contrast microscopy of Probasco cells (passage 48). The polygonal cells have a tightly packed cobblestone growth pattern.

Figure 2.2 In vitro and in vivo growth patterns of Probasco cells.

(A) In vitro growth curve of Probasco cells. Data presented as mean with a standard deviation of 3 replicates. (B) Graph represents the average bioluminescence measured from subcutaneous xenografts at the indicated time points. Data presented as a mean with a standard deviation of 5 mice. (C) Graph represents the average tumor volume of the subcutaneous xenografts. Data presented as a mean with a standard deviation of 5 mice.

Figure 2.3Expression of select genes in the Probasco prostate cancer cells.

(A) Graph represents the relative expression of BMP2, CDH1, MYOF, RUNX2, SMAD5, CXCL12,

PTHrP, SLUG, FOLH1, and AR within the Probasco cell line. (B) Graph represents the relative expression of the above genes compared to another canine prostate carcinoma cell line, Ace-1.

Figure 2.4 µCT and radiographic analysis of a nude mouse tibia injected with Probasco cells.

(A) µCT of Probasco injected tibia and control tibia at 3 weeks. Tibias in the Probasco-injected group had increased radio-opacity in the diaphyseal region compared to the control bones. (B) Radiograph of a control, early (3 weeks) and late (3 week) Probasco injected tibia. The early radiographs show increased diaphyseal intramedullary radio-opacity, while the radiograph taken at 6 weeks demonstrate the marked periosteal reaction that occurs as neoplastic cells exit through the cortex.

Figure 2.5 Photomicrograph of H&E-stained tibia 6 weeks post-injection with Probasco cells.

Neoplastic cells and osteoblasts line new woven bone present within the medullary cavity of the diaphysis and along the endosteal surfaces.

Figure 2.6 Hybridization of canine autosomal centromeric BAC clones to an example metaphase spread of

Probasco.

In this example the chromosome number was 108. The chromosomes are shown with DAPI counterstain

(A) and after application of a linear DAPI filter to reveal chromosome bands (B). In both cases the spectrum green labeled BAC probe reveals the presence and location of the centromeres. In this example there are 86 single-armed and 22 bi-armed structures. The asterisk in A and B indicate two clearly bi-armed chromosomes that did not exhibit autosomal centromere hybridization. This suggests that these two structures had sex chromosome-derived centromeres.

Figure 2.7 DNA copy number profiling of Probasco.

(A) Whole genome copy number profile: X-axis shows the 180,000 probes on the CGH array as log2 ratio organized according to chromosomal order from the top of CFA 1 to bottom of the X chromosome. The Y chromosome is not represented since the canfam2 genome assembly is female. The y-axis shows the log2 ratio of the fluorescence intensity of Cy3-labeled Probasco DNA and Cy5-labeled „reference‟ DNA (y- axis). B) 460 band ideogram representation of the dog [120] with the FASST2 called aberrations from A indicated as copy number gain and loss by shading in blue and red, respectively. The data indicate the presence of numerous DNA copy aberrations involving gains and losses of small to large, contiguous regions of the canine genome as well a complex series of alternating copy number status of CFA13.

Figure 2.7

Figure 2.8 DNA copy number profiling of Probasco recoded to human genome coordinates and output as a human data

(A) Recoded whole genome copy number data of Probasco as a pseudo-human profile, with the X-axis shows the 180,000 probes on the CGH array as log2 ratio organized according to chromosomal order from the top of human chromosome (HSA) 1 to the bottom of the HSA X. The y-axis shows the log2 ratio of the fluorescence intensity of Cy3-labeled Probasco DNA and Cy5-labeled „reference‟ DNA (y-axis). B)

Human850 band ideogram representation of the data in A with the FASST2 called aberrations indicated as copy number gain and loss by shading in blue and red, respectively.

Figure 2.8

Chapter 3 PTHrP Expression Enhances Prostate Cancer Growth and Osteolysis in

Osteoblastic Bone Metastases

Abstract

Parathyroid hormone-related protein (PTHrP) acts as an autocrine, intracrine, paracrine, and endocrine hormone. PTHrP has many roles in normal development and homeostasis, but it is perhaps most widely known for its roles in cancer. PTHrP is produced by numerous tumor types that metastasize to bone, and its role in skeletal metastases has been studied in tumors such as breast and lung cancer. The role of PTHrP in prostate cancer is still largely unknown. In this study a canine prostate cancer cell line, Probasco, was stably transfected with full-length PTHrP (Probasco-PTHrP). Cell proliferation and motility were measured in vitro. Tumor growth in vivo was measured after subcutaneous, intratibial, and intracardiac injection of cells into nude mice. Tumors were evaluated by bioluminescent imaging, Faxitron high resolution radiography and histology. RT-PCR was used to characterize select gene expression. In vitro Probasco-PTHrP had increased

64 proliferation compared to Probasco-vector cells. Probasco-PTHrP cells produced larger subcutaneous and intra-tibial tumors in vivo. While there was a range in phenotype, the

Probasco-PTHrP intra-tibial tumors had greater bone lysis. Following intracardiac injection, Probasco-PTHrP cells metastasized primarily to bone, and there was no change in number or tumor size compared to Probasco-vector cells. In conclusion, PTHrP causes increased tumor growth and bone lysis in vivo in the Probasco model of prostate cancer bone metastasis.

Introduction

Prostate cancer is a common malignancy in men and often metastasizes to bone in later stages of disease. Bone metastases result in significant morbidity and mortality through bone pain, pathologic fractures and nerve compression. Investigating how the relationship between prostate cancer and the bone microenvironment results in metastasis is vital in order to develop more effective therapeutics.

Much research has been performed to answer the question of why certain cancers metastasize to bone compared to other sites. The “seed and soil” hypothesis is often used to describe this phenomenon, but the specific factors that enable bone to serve as a preferred environment for prostate cancer are still unclear. Clinical and animal studies have shown that prostate cancer bone metastases are often in sites of active remodeling

[126-129]. From these observations it can be hypothesized that the microenvironment of

65 remodeling bone produces chemotactic and/or growth factors that support the development of metastases.

Many factors produced by cancer can instigate bone resorption and remodeling, but perhaps one of the most widely known is PTHrP. PTHrP affects bone by binding to the parathyroid hormone receptor (PTHR1) and stimulating osteoclastogenesis and osteoclast activity through increased production of receptor activator of NF-kappa B ligand

(RANKL) and monocyte chemotactic factor 1 (MCF-1) by osteoblasts [130]. High expression of PTHrP has been found in numerous malignancies that form bone metastases, and some studies have shown an association between PTHrP expression in prostate cancer and increased tumor progression and metastasis [131, 132].

In this study, it was hypothesized that PTHrP produced by prostate cancer would result in increased tumor growth and metastasis by altering the bone microenvironment.

Materials and Methods

Cell lines and tissue culture

The Probasco canine prostate cancer cell line was derived from a primary prostate carcinoma of a 10.5-year-old castrated mixed-breed dog that had been previously treated with palliative radiation therapy and metronomic chemotherapy (piroxicam, cyclophosphamide, toceranib phosphate, and chlorambucil) [133]. The clinical disease was approximately nine months, and at the time of tumor sampling there was a suspected

66 right hind limb metastasis that was not confirmed due to elective euthanasia. The

Probasco cell line was maintained at 37°C and 5% CO2 in DMEM/F12 media containing

10% fetal bovine serum (FBS) and 1% penicillin-streptomycin.

Lentiviral luciferase transduction

The Probasco cells were grown in a 6-well culture plate with DMEM/F12, 10% FBS and

1% penicillin/streptomycin. Once the cells had reached approximately 95% confluence, the media was removed and replaced with 500µL of luciferase-containing virus (VC

2192 ConcpLuc (VSV-G); Cincinnati Children‟s Hospital Viral Vector Production

Facility, Cincinnati, OH), 1.5mL of culture medium, and 1.6µL of polybrene stock

(8µg/mL). YFP-Luc lentiviral particles were produced by transient co-transfection of

293T cells with 10 μg packaging plasmid pCMVΔR8.2, 2 mg envelope plasmid pMD.G

(both from Dr. Kathleen Boris-Lawrie, Ohio State Univ.) and 10 μg transfer plasmid pHIVSIN-YFP-Luc using calcium phosphate (Sigma–Aldrich, St. Louis, MO) as previously described [119]. Following centrifugation at 2700 rpm for 60 minutes at 30°C, the plate was placed in a cell culture incubator at 37°C and 5% CO2 for 24 hours, after which the virus-containing medium was aspirated and replaced with fresh culture medium without virus. After 48 hours the cells were typsinized into a 25cm2 flask and cultured as described above.

Transfection and selection

Full-length PTHrP (-36~+141) cDNA was amplified from total RNA extracted from

HARA cells (obtained from Dr. H. Iguchi, National Kyushu Cancer Center, Fukuoka,

Japan) by RT-PCR using Platinum Taq HiFi polymerase (Invitrogen) according to the

67 published nucleotide sequence of human PTHrP (GenBank accession number

NM_198965). The primers used were: 5‟-CTA TAG GCT AGC GAG ACG ATG CAG

CGG AGA-3‟ (forward) and 5‟-CTA TAG CTC GAG TCA ATG CCT CCG TGA ATC-

3‟ (reverse). The underlined sequences represent unique NheI and XhoI restriction sites engineered into the primers, and the PCR-amplified produce was cloned into the expression vector pcDNA3.1(+) which is under the control of the CMV promoter

(Invitrogen). The recombinant plasmid, pcDNA3.1-PTHrP, was prepared using QIAprep

Miniprep Kit (Qiagen) and the nucleotide sequences was confirmed by DNA sequencing.

Probasco cells were transfected with either the empty vector (Probasco-Vector) or the

PTHrP -containing vector (Probasco-PTHrP) using Lipofectamine 2000 (Life

Technologies) following the manufacturer‟s protocols. After transfection, resistant cells were selected using 800 µg/ml of G418.

RNA extraction and quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR)

Total RNA was extracted from Probasco-Vector and Probasco-PTHrP cells (passages 10,

11, and 12) as well as the canine prostate carcinoma cell lines Leo (passages 32, 33, and

34) [44], Ace-1 (passages 22, 23, and 24) [43] and Ace-1-Dkk-1 (passages 25, 26 and 27)

[53] using the Absolutely RNA Miniprep Kit (Stratagene, Cat. No. 400800) and the manufacturer‟s protocol for adherent tissue culture cells. Total RNA (2.5 μg) was reverse transcribed using the Superscript II First Strand cDNA synthesis kit (Invitrogen) and RT-

PCR was performed using the QuantiTect SYBR Green PCR Kit (Qiagen, Cat. No.

204145) and primers for the housekeeping gene glyceraldehyde 3-phosphate

68 dehydrogenase (GAPDH), as well as the following genes: parathyroid hormone-related protein (PTHrP), B cell leukemia 2A (BCL2A), methyl-binding domain (MBD), insulin- like growth factor binding protein 6 (IGFBP6), retinoblastoma 1 (RB1), uroplakin 1B

(UPK1B), chemokine (C-X-C motif) ligand 12 (CXCL12), bone morphogenic protein 2

(BMP2), cadherin-1 (CDH1), folate hydrolase 1 (FOLH1), JUN, myoferlin (MYOF), runt-related transcription factor 2 (RUNX2), SNAIL, snail homolog 2 (SLUG), parathyroid hormone-related peptide (PTHrP), Mothers against decapentaplegic homolog

5 (SMAD5) and TWIST using canine specific primers (Table 3.1). RT-PCR reactions were repeated twice and crossing points were averaged. Genes were selected based on results from analysis of the canine 2.0 genome array (Affymetrix, Santa Clara, CA,

USA).

Canine 2.0 Genome Array

Three Probasco-Vector and Probasco-PTHrP mRNA samples at passages 10, 11, and 12 were prepared and microarray analysis with GeneChip Canine 2.0 Genome Arrays

(Affymetrix, Santa Clara, CA, USA) was performed at OSU's Comprehensive Cancer

Center Microarray Shared Resource. R version 2.15.1 (GNU, Boston, MA) was used for statistical analyses of these microarray data. Bioconductor "affy" package was used to import all the cel files. Microarray data were normalized using the "robust multi-array average" (RMA) algorithm [134]. Differential expression analysis of gene expression level was performed using the "limma" (Linear Models for Microarray Analysis) package from the Bioconductor project [135]. LIMMA takes into account the correlation between replicates and uses the empirical Bayes approach which gives stable inference also for a

69 relatively small number of arrays. Groups were compared pairwise, and the resulting p- values were adjusted using the false discovery rate (FDR) method.

PTHrP secretion

Media collected from 80% confluent Probasco-Vector and Probasco-PTHrP at passage 21 was sent to the Michigan State University Diagnostic Center for Population and Animal

Health Endocrinology department for PTHrP quantification using the PTHrP IRMA kit

(DSL8100, Beckman Coulter, Pasadena, CA). The lowest detectable standard was 2.0 pmol/L.

Wound Healing Assay

Probasco-PTHrP and Probasco-Vector cells were grown to full confluency in six-well culture plates in triplicate. Media was removed and the cells were rinsed with sterile

Dulbecco's phosphate-buffered saline (DPBS). The DPBS was removed and a sterile 200

L pipet tip was used to scratch three separate wounds (each approximately 2 cm long) through the cells. The cells were rinsed with DPBS and cultured in 1.5mL of DMEM/F12 with 0.1% bovine serum albumin (BSA). Using phase contrast, images of the scratch were taken at 0, 6, 12, and 24 hours.

In vitro growth rate

Probasco-PTHrP and Probasco-Vector cells from passage 6 were plated in a six-well culture plate in triplicate at a density of 200,000 cells per well. Cells were cultured in

DMEM/F12 with 10% FBS and 1% penicillin/streptomycin and incubated in 37°C and

5% CO2. The cells were collected using 0.25% trypsin at 24, 48, and 72 hours after the initial plating and counted with an automated cell counter (Nexcelom Bioscience,

Lawrence, MA) using trypan-blue dye exclusion to differentiate between live and dead cells. The doubling time was calculated using the formula: (t2- t1) x log(n2)/log(n2/n1), where n is the cell number at time points (t).

Immunocytochemistry

Probasco-Vector and Probasco-PTHrP were grown on glass coverslips and were briefly fixed in a solution of 1% formaldehyde in pH 6.5 stabilization buffer (127 mM NaCl, 5 mM KCl, 1.1 mM NaH2PO4, 0.4 mM KH2PO4, 2 mM MgCl2, 5.5 mM glucose, 1 mM

EGTA, 20 mM Pipes), and then fixed and permeabilized for 1 hour in a solution of 2% formaldehyde, 0.2% Triton X-100, and 0.5% deoxycholate in the same stabilization buffer. Actin was identified using Alexa-labeled phalloidin (Invitrogen Corp., Carlsbad,

CA). Cells were visualized using a Zeiss 510 META laser scanning confocal microscope using the filter set Fset15 (excitation: 546/12, emission: 590 nm).

In vitro calvaria co-culture

Two-day-old FBV mouse pups were euthanized and discs of calvaria (2 per mouse) were collected using 3.5mm biopsy punches (Fisher Scientific, cat. No. 33-33). The calvarial discs were washed twice with sterile DPBS and incubated overnight at 37C and 5%CO2 in 24-well cell culture plates with 1.5mL of BGJb medium with 0.1% BSA and 1%

Normocin (InvivoGen, San Diego, CA). Probasco-PTHrP and Probasco-Vector cells from passage 22 were plated at 50,000 cells per well in 24-well culture plates containing

DMEM/F12 with 0.1% BSA and incubated overnight. After incubation, the media was aspirated from the cells and replaced with media that was 50% DMEM/F12 and 50%

BGJb with a total of 0.1% BSA. One calvaria disc was added to each well of cells and

71 cultured for 48 hours. After 48 hours, media was removed and total medium calcium was measured using the QuantiChrom Calcium Assay Kit (BioAssay Systems, Hayward,

CA). Calvaria were removed and fixed in 95% ethanol and 5% acetic acid for 30 minutes.

A Faxitron laboratory radiography system LX-60 (Faxitron X-ray Corp., Wheeling, IL) imaging platform was used to take high-resolution radiographs of the calvaria at 25KV and 5 second exposures. Calvaria mineralization was estimated by converting the radiographs into black and white bitmap images and counting white pixels and black pixels using Image-Pro Premier 9.1 (Media Cybernetics, Rockville, MD).

Subcutaneous, intratibial, and intracardiac injection of Probasco cells into nude mice

All animal experiments were approved by The Ohio State University Institutional

Laboratory Animal Care and Use Committee. Mice were purchased from Taconic

(Cambridge City, IN) and injections were performed in mice that were 4 weeks old.

Vector and Probasco-PTHrP cells in 0.25mL of sterile DPBS was injected subcutaneously between the shoulder blades using a 25-gauge needle. Tumor volume was measured weekly using calipers. Measurements were taken in three dimensions and tumor volume was calculated using the formula length x width x height x 1/2. Tumor growth was also monitored weekly using bioluminescent imaging for a total of 7 weeks.

Intratibial: Five nude mice per group were anesthetized in an induction chamber with

3% isoflurane and maintained with 2.5% isoflurane. The right rear limb was scrubbed with alcohol-soaked gauze. The leg was held so the knee joint was at a 90° angle and a

27-gauge needle was introduced through the patellar ligament and into the tibial marrow space through the articular cartilage. Using a Hamilton syringe, 0.025mL of sterile DPBS containing a suspension of 200,000 Probasco-Vector or Probasco-PTHrP cells was introduced into the marrow cavity. Tumor growth was monitored weekly using bioluminescent imaging for a total of 6 weeks.

Intracardiac: Ten nude mice per group were induced with 3% isoflurane gas anesthesia and maintained at 2.5%. The mice were placed in dorsal recumbency and their front limbs were placed perpendicular to their thorax and restrained to the procedure table with surgical tape. The ventral thorax was scrubbed with alcohol. A 1mL syringe was loaded with 0.1mL of DPBS containing a 300,000 cell suspension of either Probasco-Vector or

Probasco-PTHrP cells. The syringe with a 25-gauge needle was introduced into the thoracic cavity through the third intercostals space, approximate 1mm to the left and lateral from the sternum. Once a pulsatile jet of blood was present in the hub of the needle, the cell suspension was slowly injected into the left ventricle of the heart for 30 sec. Tumor growth was monitored weekly using bioluminescent imaging for a total of 4 weeks. Three mice from each group did not survive the initial injection, so the final group size was seven mice.

Bioluminescent imaging

Mice were anesthetized in an induction chamber with 3% isoflurane and maintained with

2% isoflurane. A 1ml syringe was used to inject 0.15ml of sterile DPBS containing

Histopathology and immunohistochemistry

Radiography

Statistics

Results are displayed as means ± standard deviation. Normalized gene expression data was analyzed using one-way ANOVA followed by Sidak‟s multiple comparisons test to perform pair-wise comparisons using GraphPad Prism version 6.03 (La Jolla, CA). The

74 data from the remaining in vitro experiments and all in vivo experiments were analyzed by comparing the Probasco-PTHrP group to the Probasco-Vector group using Student‟s t- test. Data with P values < 0.05 were considered statistically significant.

Results

In vitro and in vivo growth

Both Probasco-Vector and Probasco-PTHrP cells grew in an anchorage-dependent manner in vitro and formed monolayer sheets of tightly packed polygonal cells; however, the mean cell diameter of typsinized Probasco-PTHrP cells was significantly smaller than the Probasco-Vector cells (Figure 3.1A). Immunocytochemistry for actin did not reveal a difference in the organization of the actin cytoskeleton between the two cell lines (Figure

3.1B). Probasco-PTHrP cells grew exponentially in vitro with a doubling time of approximately 24 hours, while the doubling time of Probasco-Vector was approximately

20 hours (Figure 3.2A). The subcutaneous xenograft volume was similar in both groups.

Probasco-PTHrP and Probasco-Vector xenografts double in size approximately every 1.5 and 1.5 weeks, respectively (Figure 3.2B). The doubling time of the viable cells in subcutaneous xenografts was approximately 30% faster in Probasco-PTHrP tumors

(doubling time of 1.3 weeks) compared to Probasco-Vector tumors (doubling time of 1.8 weeks) as estimated by BLI (Figure 3.2C; p<0.05). Intratibial tumor growth of Probasco-

PTHrP cells was increased compared to Probasco-Vector cells; however, statistical

75 significance was not reached (week 5 p=0.07; Figure 3.2D). Histologically, Probasco-

Vector subcutaneous tumors were comprised of moderately sized polygonal epithelial cells forming nests and anastomosing cords in a fine fibrovascular stroma with moderate central coagulation necrosis. Large pseudo-cyst formation characterized by epithelial- lined tubular or acinar structures dilated with either proteinaceous eosinophilic fluid or necrotic cell debris was common in Probasco-Vector tumors. In contrast, Probasco-

PTHrP tumors were composed of small and moderately sized polygonal epithelial cells forming nests and cords; however, coagulation necrosis and pseudo-cyst formation was less common (Figure 3.3). qRT-PCR of Probasco-Vector and Probasco-PTHrP in vitro

Compared to Probasco-Vector cells, Probasco-PTHrP cells expressed higher levels of

BCL2A (150%; p<0.001) and CXCL12 (154%; p<0.001) mRNA, and decreased amounts of IGFBP6 (55%, p<0.001) and UPK1B (63%, p<0.01) mRNA. Both cell lines had similar levels of expression of MBD and RB1 mRNA (Figure 3.4A).

PTHrP Secretion

Probasco-PTHrP cells secreted 150 pmol/L of PTHrP into the cell culture media while

Probasco-Vector cells did not secrete detectable PTHrP (Figure 3.4B).

Wound healing assay

Probasco-PTHrP and Probasco-Vector cells migrated across the scratch at the same rate.

Most scratches were fully healed by 12 hours, and no scratches were apparent by 24 hours (Figure 3.5).

Calvaria co-culture

Probasco-PTHrP cells produced more calvaria bone lysis resulting in more calcium in the media compared to Probasco-Vector; however, the difference was not statistically significant (Figure 3.6A and 3.6B).

Intratibial (IT) injections in nude mice

All 10 mice developed successful xenografts in the tibias. Radiographs of Probasco-

PTHrP injected limbs showed mixed osteoblastic/osteolytic lesions, with extensive radio- opaque new bone formation on both the endosteal and periosteal surfaces as well as multifocal areas of osteolysis characterized by radiolucency. Severity of osteolytic lesions varied within the Probasco-PTHrP injected tibias ranging from mild localized foci of bone resorption to a widespread “moth-eaten” pattern evident on radiography; however, overall osteolysis was increased compared to the markedly osteoblastic nature of Probasco-Vector (Figure 3.7A). Histology of Probasco-PTHrP injected tibias demonstrated multifocal islands of woven bone surrounded by loosely arranged bands of stromal cells which were then surrounded by densely packed polygonal neoplastic cells

(Figure 3.7B). This was in dramatic contrast to direct apposition of neoplastic cells to new woven bone and osteoblasts and a complete lack of stromal reaction found in

Probasco-Vector injected tibias. Compared to Probasco-Vector injected mice, Probasco-

PTHrP injected limbs had increased osteoclast formation and osteoclastic bone resorption.

Intracardiac injection (IC) and in vivo bioluminescent imaging (BLI) in nude mice

Incidence of bone metastasis was measured by BLI. Six of seven mice injected with

Probasco-PTHrP cells developed bone metastases, ranging from 1-4 metastases per mouse. Similarly, six of seven mice injected with Probasco-Vector developed bone metastases, ranging from 1-5 metastases per mouse (Table 3.2). The locations of bone metastases were similar, with limbs being the most common site followed by mandible and tail. qRT-PCR of Probasco-Vector, Probasco-PTHrP, Ace-1, Ace-1-Dkk-1 and Leo in vitro

Compared to three additional canine prostate carcinoma cell lines (Ace-1, Ace-1-Dkk-1, and Leo), Probasco-Vector and Probasco-PTHrP cells expressed significantly more

SMAD5 (140%; p<0.05) and RUNX2 (approximately 130%; p<0.05) mRNA and significantly less SNAIL mRNA (10-33%; p<0.05) (Figure 3.7A and 3.7B). Probasco-

Vector and Probasco-PTHrP cell expression of BMP2, SLUG, TWIST and MYOF mRNA was most similar to Leo cells while expression of CDH1 and JUN mRNA was more similar to the Ace-1 and Ace-1-Dkk-1 cells.

Discussion

PTHrP was first discovered in 1987 as a factor that induced hypercalcemia of malignancy

[136]. Since then, PTHrP has been shown to play a significant role in the bone metastasis

78 of many neoplasms, such as breast and lung cancer [137, 138]. PTHrP can facilitate bone metastases by autocrine, paracrine, and intracrine signaling in neoplastic cells to promote proliferation and cell survival. PTHrP has three functional domains: an N-terminal, midregion and C-terminal domain [139, 140]. The N-terminal domain binds and activates the

PTH/PTHrP receptor (PTHR1) in either an autocrine, paracrine or endocrine manner

[141]. The mid-region contains the nuclear localization sequence (NLS) and is vital to intracrine signaling and the regulation of proliferation. The C-terminal domain

(osteostatin) has an anabolic effect in bone and inhibits osteoclastic bone resorption, although its effect in cancer is not well characterized [142, 143]. In bone, PTHrP signaling stimulates osteoblasts to produces RANKL, a factor that enhances osteoclast formation and maturation. As the population of mature osteoclasts expands, bone resorption is enhanced and calcium and growth factors such as transforming growth factor-beta (TGF-β) and fibroblast growth factors (FGFs) are released [144]. The calcium can then further stimulate PTHrP secretion from the neoplastic cells resulting in a

“vicious cycle” of positive feedback [145]. Similarly, TGF-β can stimulate PTHrP production, thus amplifying the cycle [146].

Normal prostatic tissue expresses very low levels of PTHrP and its expression is localized to the epithelial and neuroendocrine cells [147]. PTHrP expression is greater in well to poorly differentiated prostate cancer and prostatic intraepithelial hyperplasia

(PIN) (87-100% of cells) compared to benign disturbances of growth such as prostatic hyperplasia (33% of cells) [148, 149]. In many human prostate cancer cell lines, PTHrP expression enhances tumor growth in bone and prevents apoptosis [150].

In this study, the effects of PTHrP in prostate cancer bone metastasis were investigated using the canine prostate cancer cell line, Probasco, stably transfected with full-length human PTHrP. The Probasco cell line was very useful in this regard due to its markedly osteoblastic lesions in bone, similar to what is seen in men [116]. Intratibial tumors produced by Probasco cells produce marked endosteal and periosteal new bone formation with minimal lysis radiographically and histologically. The Probasco cells do not express detectable PTHrP mRNA or secrete detectable PTHrP protein. The Probasco cell line is very permissive to transfection, and transfection of full-length human PTHrP DNA resulted in high levels of mRNA expression and protein secretion.

The Probasco-PTHrP cells had increased proliferation in vivo as evidenced by increased subcutaneous and intratibial tumor BLI and increased subcutaneous tumor volume compared to the vector controls. This may be explained, in part, by the altered gene expression in the PTHrP-expressing cells. Probasco-PTHrP cells expressed significantly higher levels of BCL2A and significantly lower levels of IGFBP6. BCL2A is not expressed in normal prostate, but is often expressed in prostate and other cancers [151]. It promotes prostate cancer cell survival through the inhibition of apoptosis and has also been implicated in the development of androgen-independent prostate cancer due to its increased expression in the advanced stages of disease [152, 153]. Further investigation will be needed to determine if BCL2A enhances resistance to apoptosis in the Probasco-

PTHrP cells.

The role of IGFBP6 in prostate cancer has not been widely studied. The function of

IGFBP6 is predominantly to bind insulin-like growth factor II and inhibit its function by

80 blocking receptor binding [154, 155]. In this way the ultimate action of IGFBP6 is to reduce cellular proliferation. The combined actions of increased BCL2A and decreased

IGFBP6 could contribute to the larger tumors present in the Probasco-PTHrP injected mice by enhancing cell proliferation and survival.

In this model of prostate cancer bone metastasis there was significant bone production in the Probasco-PTHrP tumors in addition to increased osteolysis. While PTHrP is normally considered an inducer of osteolysis, it can have dramatic anabolic effects as well. PTHrP

1-36 increases bone density when administered by daily injection to women with osteoporosis [156, 157]. PTHrP 107-139 also has been shown to be anabolic in bone when administered in vivo or in vitro by increasing osteoblastogenesis and bone mineralization and decreasing osteoclast formation [143]. Further investigation into the mechanism of new bone formation by the Probasco cell line is needed.

In the Ace-1 model of canine prostate carcinoma that produces mixed osteoblastic and osteolytic bone metastase, Ace-1 cells transfected with rat full-length PTHrP induced significantly greater new bone formation and angiogenesis in subcutaneous bone vossicles in vivo [48]. It was found that tumor-produced PTHrP enhanced proliferation of bone marrow stromal cells between the tumor and bone surfaces. In our investigation we also found dramatic proliferation of stromal cells between the Probasco cells and tibial bone surfaces that was only present in the Probasco-PTHrP tumors. Angiogenesis was not characterized in this model; however, it would be interesting to verify a similar increase in vascularity in the Probasco-PTHrP bone tumors. Several factors are believed to play a role in increasing bone marrow stromal cells, including platelet-derived growth

81 factor (PDGF) and transforming growth factor-beta (TGF-β) [158, 159]. TGF-β is abundant in bone matrix and is released in response to factors that stimulate osteoclastic bone resorption [160]. Through osteolysis induced by PTHrP, it is possible that the subsequently released TGF-β is stimulating stromal cell proliferation. This could explain why the stromal proliferation is not apparent in the Probasco-Vector tumors since bone resorption is minimal. TGF-β also is able to block matrix degradation by reducing collagenase-like protease synthesis while enhancing protease inhibitor expression, which may play a role in reducing osteolysis and creating the mixed osteoblastic/osteolytic metastases in the Probasco-PTHrP injected limbs [161].

In vitro, Probasco-PTHrP cells have a significantly smaller mean cell diameter than the

Probasco-Vector cells. The affect of PTHrP on cell size has not been widely studied, but one study has reported a reduction in lung cancer cell size with PTHrP expression [162].

In that study, it was postulated that the reduction in cell size was secondary to a more indolent growth pattern. For Probasco-PTHrP cells this is less likely to be the case since the cells had a similar growth rate in vitro as the vector controls. In contrast, PTHrP has been reported to increase cell volume in certain non-neoplastic tissues (such as cardiomyocytes and trophoblasts) [163, 164].

We did not find a difference between incidence of metastasis and PTHrP expression, unlike other studies [48]. This was not surprising because the in vitro wound healing assay did not show a difference in migration between these two cell lines. While it is thought that PTHrP enhances cancer growth and progression, there is conflicting evidence with respect to its prognostic significance in prostate cancer. While PTHrP

82 expression is found in most metastatic prostate tumors, it has also been shown to be expressed in 33% of benign hyperplasia and 87% of well-differentiated tumors [148,

165]. As PTHrP has the ability to act as an autocrine, intracrine, paracrine, or endocrine hormone, its downstream effects can be varied depending on numerous factors [166]. In this particular model system it may be that the balance of these effects, while stimulating tumor growth, does not enhance metastasis.

Our laboratory has developed 5 canine prostate carcinoma cell lines: Ace-1 [43], Ace-1-

Dkk-1 [53], Leo [44], Probasco [133] and Probasco-PTHrP. Ace-1 cells are characterized as metastasizing primarily to bone after intracardiac injection and forming mixed osteoblastic/osteolytic metastases. Ace-1-Dkk-1 cells are stably transfected with human

Dkk-1 and produce bone tumors that are osteolytic with no evidence of new bone formation. Leo cells are unique in that they metastasize primarily to brain after intracardiac injection; in bone the Leo cells form osteolytic metastases. The Probasco cell line is the most osteoblastic of the canine prostate carcinoma cell lines while the

Probasco-PTHrP cells form metastases more similar to the Ace-1 cell line. In the gene expression comparisons between the cell lines, Probasco-Vector and Probasco-PTHrP expressed significantly more SMAD5 and RUNX2 mRNA and significantly less SNAIL mRNA than the other three cell lines. SMAD5 and RUNX2 are important in osteoblast differentiation and perhaps contribute to the osteoblastic nature of the Probasco-Vector and Probasco-PTHrP metastases [167, 168]. SNAIL is associated with epithelial- mesenchymal transition (EMT), and the lower expression of SNAIL mRNA in the two

Probasco cell lines is not surprising because they have the least evidence of EMT histologically compared to the other canine prostate carcinoma cell lines [169].

Conclusions

This study demonstrated that prostate cancer-derived PTHrP enhanced osteoclastic bone resorption and prostate cancer cell proliferation. The effects on cell proliferation were possibly due to enhanced expression of oncogenes and reduced expression of tumor suppressor genes, such as BCL2A and IGFBP6. Targeting factors such as PTHrP that mediate the interactions between prostate cancer and the bone microenvironment may provide novel therapeutics in the treatment of prostate cancer bone metastasis. Lastly, since there was no difference in the incidence of metastasis after intracardiac injection between Probasco-Vector and Probasco-PTHrP cells, this suggests that PTHrP is not always necessary to confer that ability of cancers to metastasize and grow in bone.

Tables

Gene Forward Primer (5’-3’) Reverse Primer (5’-3’) GAPDH CCCACTCTTCCACCTTCGAC AGCCAAATTCATTGTCATACCAGG PTHrP AGCTCGGCCGCCGGCTCAA GGAAGAATCGTCGCCGTAAG BCL2A GACAAAACGGAGGCTGGGAA CCTCTGGGCCAGTCAGTAGT MBD TGTAGACGAGGACGAGCTACC CTTCTGGTTGTTGCCCCCAA IGFBP6 GTCCAAGATGCTGAAATGGGC TAGGCACGTAGAGGGTGTGA RB1 TGCCAACGCCAACAAAAATGA TGGTTTAGGAGGGTTGCTGC UPK1B GACAGACTCATGCTCCAGGAC ATCACACAGCACTGACGAGG CXCL12 ACTGTGCCCTTCAGATCGTGGCA CCACCTGCGCCTCTCACATCTTG BMP2 TGCGCAGCTTCCACCACGAAGAA CAAAGGTTCCTGCATCTGTTCCCGA CDH1 GCTGCTGACCTGCAAGGCGA GGCCGGGGTATCGGGGACAT CXCL12 ACTGTGCCCTTCAGATCGTGGCA CCACCTGCGCCTCTCACATCTTG FOLH1 GCAGGGGACCCTCTCACACCTG CTCGGAAGACCAACAGCCTCTGTGA GAPDH CCCACTCTTCCACCTTCGAC AGCCAAATTCATTGTCATACCAGG MYOF TGCCCCCGAAAGGCTGGGAAT ACTCCGTGTGCCCTGCGTCT RUNX2 TGCCTCTGGCCTTCCACTCTCAG TGCATTCGTGGGTTGGAGAAGCG SLUG GGCAAGGCGTTTTCCAGACCCT GGGCAAGAAAAAGGCTTCTCCCCAG JUN CGACATGGAGTCGCAAGAGA GCTTCCTTTTCCGGCACTTG SNAIL ACACGCTGCCCTGCGTCTGT GGTCTGCAGGTGGGCCCGAA SMAD5 CCCAGCCAATGGATACAAGC ATGCATGAAAAGCCTCCCCA PTHrP AGCTCGGCCGCCGGCTCAA GGAAGAATCGTCGCCGTAAG TWIST GGCAGGGCCGGAGACCTAGATG TCCACGGGCCTGTCTCGCTT Table 3.1 Primers used for qRT-PCR with the Probasco-PTHrP and Probasco-Vector cell lines.

Table 3.2 Metastasis following intracardiac injection of Probasco-Vector and Probasco-PTHrP cells

Cell line Proportion of mice Number of Average number of Metastasis that developed metastases per metastases per locations metastases mouse mouse

Probasco- 6/7 1-4 2.7 Limbs, Vector mandible, tail and digits

Probasco- 6/7 1-5 3 PTHrP

Table 3.2 Number and location of metastases following intracardiac injection of Probasco-Vector and

Probasco-PTHrP cells

Figures

A 16

n o

r 14 *

i m 13

12 r P to r c H e T -V -P o c sc as a b b ro ro P P B

Figure 3.1 In vitro mean cell diameter and actin immunocytochemistry of Probasco-Vector and Probasco-

PTHrP cells.

(A) Probasco-PTHrP cells were approximately 1 micron smaller in diameter than Probasco-Vector cells in culture (n=3; Student‟s t-test; *, p<0.05). (B) There was no difference in the arrangement of the actin cytoskeleton between the Probasco-Vector (left) and Probasco-PTHrP (right) cells. The diameter of the cytoskeleton was larger in the Probasco-Vector cells.

A In Vitro Growth Curve B Subcutaneous tumors 3000000 Probasco-Vector 1200 Probasco-Vector Probasco-PTHrP

1000 Probasco-PTHrP

l l

e 2000000 800

/ m

s 600

l m e

C 1000000 400 200 0 0 1 2 3 1 2 3 4 5 6 7 Day Week

C D Intratibial tumors )

Subcutaneous tumors )

5 = 2.0109 Probasco-Vector SQ = 9

n 2.010

n Probasco-Vector IT

(

Probasco-PTHrP SQ c

* c Probasco-PTHrP IT e

9 e 9 S

1.510 S 1.510

/ s

* s

n o 9 o

t 9

1.010 * t 1.010 o

* o

8 8 e 5.010 e 5.010

g *

g a

* a

e v

v 0 A A 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Week Week

Figure .3.2 In vitro and in vivo growth patterns of Probasco cells.

(A) In vitro growth curve of Probasco-Vector and Probasco-PTHrP cells. There was no significant difference in cell proliferation. (n=3; Student‟s t-test, p=0.07) (B) Graph represents the average tumor volume of the subcutaneous xenografts. There was no significant difference in tumor volume between the two groups (n=5). (C) Graph represents the average bioluminescence measured from subcutaneous xenografts at the indicated time points. Probasco-PTHrP tumors had significantly greater BLI at weeks 2-7.

(n=5; Student‟s t-test, *, p<0.05) (D) Graph represents the average bioluminescence measured from intratibial xenografts at the indicated time points. There was no significant difference in tumor BLI between the two groups; however, the P value at week 5 was 0.07 (n=5, Student‟s t-test). 88

Figure 3.3 Probasco-Vector and Probasco-PTHrP subcutaneous tumors

Probasco-Vector subcutaneous tumors (top) were comprised of moderately sized polygonal epithelial cells forming nests, anastomosing cords and sheets with moderate central coagulation necrosis (N). Pseudo-cyst (C) formation characterized by epithelial- lined tubular or acinar structures dilated with either proteinaceous eosinophilic fluid or necrotic cell debris was common. In contrast, Probasco-PTHrP subcutaneous (bottom) were composed of small to moderately sized polygonal epithelial cells forming nests, cords and sheets; however, coagulation necrosis and pseudo-cyst formation was less common. Insets show high power magnification (40X) of cell populations in respective tumors.

Figure 3.3

C N

Probasco-Vector

Probasco-PTHrP90

n 4

) o

i Probasco-Vector H

s * * D

s Probasco-PTHrP

e P

r 3

d 2

e 1 m

v *

r t

o *

( e

R 0 A D 6 1 B 2 2 B P B 1 1 L B R K L C M F P C B G X I U C

) 200

l o

m 150

(

P r

H 100

d e

t 50

c e

S 0 r P to r c H e T -V -P o o c c s s a a b b ro ro P P

Figure 3.4 Expression of genes and PTHrP secretion in the Probasco-Vector and Probasco-PTHrP prostate cancer cells.

(A) Probasco-PTHrP cells expressed significantly higher BCL2A (149%, p<0.001) and CXCL12 (153%, p<0.001) and significantly lower IGFBP6 (60%, p<0.001), UPK1B (63%, p<0.01) mRNA compared to

Probasco-Vector cells. MBD and RB1 mRNA expression was similar between the two cells. (B) Probasco-

PTHrP secreted 150 pmol/L in vitro while Probasco-Vector did not secrete detectable PTHrP (level of detectability is 0.2 pmol/L).

50 h

t Probasco-Vector

h a

r 40 Probasco-PTHrP

f a

o 30

o i

c o

r 10

P s I 0 0 10 20 30 Hours

Figure 3.5 Wound healing assay

Graph represents the average amounts of time for Probasco-Vector and Probasco-PTHrP cells to migrate to close a scratch made to confluent cells in vitro. There was no difference in cell migration between the two groups (n=3; Student‟s t-test)

A B

8 Probasc-Vector 80 Probasco-Vector )

Probasco-PTHrP n Probasco-PTHrP

/ t

a g

6 z

(

n i

4 i 75

a t

i 2

R M 0 70 2 r P y to r a c H D e T -V -P o o c sc s a a b b ro ro P P

Figure 3.6 In vitro calvaria co-culture assay using 3.5mm neonatal murine calvaria discs and Probasco-

Vector and Probasco-PTHrP cell lines.

(A) Amount of calcium present in the media after 2 days of culture. There was no significant difference in calcium released into the media by calvaria lysis (n=6; Student‟s t-test, p=0.3) (B) Relative calvaria mineralization after two days of culture. There was no significant difference in calvaria mineralization between the two groups. (n=6; Student‟s t-test, p=0.4)

T S OB T

b OB L

Figure 3.7 Tibias from nude mice after injection of Probasco-Vector or Probasco-PTHrP cells.

(A) Faxitron radiograph of Probasco-Vector (upper left) and Probasco-PTHrP (upper right) injected limbs and the contralateral, non-injected limbs (lower right and left). The Probasco-Vector injected tibia had marked intramedullary and periosteal new bone formation. In contrast, the Probasco-PTHrP injected tibia had numerous radiolucent foci in the intramedullary and periosteal new bone representing osteolysis. (B)

Photomicrograph of Probasco-Vector (right) and Probasco-PTHrP (left) tumor-bone interface (40x). In the

Probasco-Vector tumors, polygonal neoplastic cells (T) and osteoblasts (OB) lined new woven bone (b) present within the medullary cavity of the diaphysis and along the endosteal surfaces. In contrast, Probasco-

PTHrP tumors had multiple layers of spindle-shaped stromal cells (S) separating the smaller polygonal neoplastic cells (T) from the bone and osteoblasts and multinucleated osteoclasts (OC). Osteolysis (L) is evident by the scalloped bone surface deep to osteoclasts.

Figure 3.8 Expression of specific mRNAs in the Probasco-PTHrP, Probasco-Vector, Leo, Ace-1 and Ace-

1-Dkk-1 canine prostate cancer cells.

Probasco-Vector and Probasco-PTHrP cells expressed significantly greater SMAD5 (B; 140%; p<0.05) and

RUNX2 (B; 130%; p<0.05) mRNA and significantly less SNAIL mRNA (A; 10-33%; p<0.05) compared to the Ace-1, Ace-1-Dkk-1, and Leo canine prostate carcinoma cell lines. Probasco-Vector and Probasco-

PTHrP cell expression of BMP2 (A), SLUG (A), TWIST (A) and MYOF (B) mRNA was most similar to

Leo cells while expression of CDH1 (B) and JUN (B) mRNA was more similar to Ace-1 and Ace-1-Dkk-1 cells. (n=3, repeated twice; *, p<0.05, two way ANOVA)

Figure 3.8

Probasco-PTHrP Probasco-Vector Leo Ace-1 * Ace-1-Dkk-1

A *

n )

o 1000 i

H * s

D * s 100 * * e * P * * r * *

A *

p 10 * *

o t

N e

R 0.1

e 0.01

t o

a 0.001

e (

R 0.0001

2 1 G IL T P H U A IS M L L B O S N W F S T

B n 3

)

o i

H *

p x G 2 * * *

E *

o * * * *

t * * *

A * * *

* d

N *

i l

m 1

i r

e (

R 0 1 2 N F P 2 5 1 r X D H L U O H D J Y T N A C C M U M X P R C S 96

Chapter 4 Dickkoft-1 (Dkk-1) Enhanced Non-canonical Wnt/JNK signaling in Prostate

Cancer Bone Metastases

Abstract

The molecular mechanisms by which prostate cancer cells metastasize and grow in bone are not fully understood; however, we hypothesized that the Wnt signaling pathways plays an important role. To investigate the contribution of the Wnt signaling pathways in prostate cancer bone metastases, we over-expressed the Wnt/JNK pathway agonist,

DKK-1, in the mixed osteoblastic and osteolytic Ace-1 prostate cancer cells. Previous work had shown that DKK-1 expression increased the number and lytic nature of bone metastases in vivo. This study focused on elucidating how enhanced Wnt/JNK signaling could alter metastasis and the bone microenvironment. Ace-1 cells stably expressing human DKK-1 or empty vector were cultured in vitro. Wnt/JNK signaling was investigated by AP-1 reporter activity and qRT-PCR. Treatment with a non-canonical

Wnt/JNK agonist and antagonist were performed and the changes in reporter activity, cell proliferation and migration were investigated. In this study it was shown that DKK-1 97 significantly increased non-canonical Wnt/JNK signaling. Subsequent gene expression alterations included a dramatic decrease in the mRNA expression of genes important for osteoblast maturation. Treatment with a Wnt/JNK agonist enhanced tumor cell proliferation and migration, and antagonist treatment reduced proliferation and migration.

The present study showed that DKK-1 was a potent activator of non-canonical Wnt/JNK signaling and provides a possible mechanism by which DKK-1 expression inhibits bone growth and enhances tumorigenesis in prostate cancer metastases.

Introduction

One in six men will be diagnosed with prostate cancer in the western world, and of those men 1 in 6 will die of metastatic disease [28]. The most common site of metastasis is bone [2]. Metastasis to bone is considered incurable and is associated with significant pain and morbidity. Studies on prostate cancer metastases have shown the importance of interactions with the bone microenvironment for growth and development, but much remains unknown [11].

An important pathway in the regulation of the bone microenvironment is the canonical

Wnt signaling pathway. The canonical pathway ultimately leads to the stabilization of ß- catenin [170]. Normal levels of ß-catenin promote cell adhesion and control cell shape, and up-regulation affects cell proliferation and differentiation through activation of transcription factors and alterations in gene transcription [171]. In bone, canonical Wnt

98 signaling promotes osteoblast differentiation and, indirectly, new bone formation [172]. It has been proposed in prostate cancer that tumor-derived Wnts stimulate osteoblast differentiation and the formation of osteoblastic metastases in the bone through activation of the canonical Wnt pathway in a paracrine fashion [173].

Wnt signaling is not limited solely to the canonical Wnt/β-catenin pathway, but can also function though several “non-canonical” pathways. The non-canonical Wnt pathways, while less understood, are involved in cell polarity, cell adhesion, and other cell functions through regulation of pathways such as phospholipase C and c-Jun N-terminal kinases

(JNK) [171, 174]. Dkk-1 (Dickkopf-1) is known to be both a potent inhibitor of canonical Wnt signaling and an up-regulator of the non-canonical Wnt/JNK pathway

[175-177]. Research on prostate carcinoma cell lines has shown that Dkk-1 can result in changes in metastatic phenotype and altered tumor growth [178, 179].

In bone, the ability of Dkk-1 to inhibit canonical Wnt signaling results in reduced osteoblastic bone formation and secondary osteolysis [180]. A consequence of osteolysis is the release of transforming growth factor-ß (TGF-ß), one of the most abundant cytokines in the bone matrix [181]. TGF-ß is an inducer of epithelial-mesenchymal transition (EMT) during development and its overexpression in cancer suggests a similar role in neoplasia [182, 183]. EMT in cancer is associated with tumor migration, invasion, and metastasis. TGF-ß enables downstream signaling by binding to cell surface receptors which leads to the activation of Smads. Smads are transcription factors that are located within the cytoplasm and regulate transcription of TGF-ß target genes. Smads alone have a low affinity for DNA and need to interact with cofactors for effective DNA binding and

99 transcription. One such cofactor is Activator protein 1 (AP-1), a product of the non- canonical Wnt/JNK pathway [184]. Through this signaling pathway, TGF-β indirectly enhances Wnt/JNK signaling [185]. The AP-1/Smad 3 complex promotes gene transcription of vimentin, endothelin-1 (ET-1), and c-jun [182, 184]. Increased expression of vimentin and decreased expression of E-cadherin is one of the hallmarks of EMT

[186]. Interestingly, the combined role of bone-derived TGF-ß and non-canonical Wnt signaling has never been investigated in prostate cancer bone metastases.

Using a canine model of prostate cancer, the Ace-1 cell line, previous work in our lab has shown that increased Dkk-1 expression enhanced non-canonical Wnt/JNK signaling as well as increased tumor growth and the number of metastases after intracardiac injection

[53]. The goal of this study was to elucidate the molecular role of Dkk-1 in prostate cancer growth and metastasis.

Materials and Methods

Cell Culture

The canine Ace-1 prostate cancer cells were previously derived from a spontaneous dog prostate carcinoma by our laboratory [43]. Ace-1 cells stably expressing human Dkk-1

(Ace-1-Dkk-1) or empty vector (Ace-1-Vector) were previously created by transfection and selection with G418 (Sigma-Aldrich Co., St. Louis, MO) [53]. Ace-1 and Ace-1-

Dkk-1 cells were maintained at 37°C in Dulbecco‟s Modified Eagle‟s Medium/Ham‟s

100

Nutrient Mixture F12 (DMEM/F12) (Invitrogen, Carlsbad, CA) supplemented with 0.1% bovine serum albumin (BSA; Sigma Aldrich) in a 5% CO2-humidified chamber.

Dkk-1 ELISA

Ace-1-Vector and Ace-1-Dkk-1 cells were plated in 6-well plates in triplicate at a density of 500,000 cells per well and cultured in DMEM/F12 with 10% fetal bovine serum and

1% penicillin/streptomycin. After 24 hours, media was collected and Dkk-1 concentrations were measured using the DuoSet Human Dkk1 ELISA Kit as described previously [187] (R&D Systems, Minneapolis, MN, USA). The lowest standard of the assay was 62.5 pg/mL.

β-catenin immunohistochemistry

Ace-1-Vector and Ace-1-Dkk-1 were grown to 70% confluency, typsinized, and centrifuged at 1000RPM for 5 minutes. The cell pellet was then fixed in 10% neutral- buffered formalin for 24 hours and then embedded in paraffin. The specimens were sectioned (5 μm) and evaluated by immunohistochemistry for β-catenin (BD Pharmingen, cat no 610153). Image Pro Plus 9.0 was used to perform batch analysis and quantify total cell staining (reported as optical density, OD). The values from five fields of approximately 700 cells were averaged.

AP-1 reporter transfection

To investigate the activity of Wnt/JNK signaling, 10,000 Ace-1-Vector or Ace-1-Dkk-1 cells in 96-well plates in triplicate were transiently transfected with 0.1µg of either an inducible AP-1 reporter construct, a negative control, or a positive control purchased from SABiosciences (cat. No CCS-011L). Transient transfection was accomplished using

101

0.5µL Lipofectamine 2000 transfection reagent (Invitrogen) and Opti-MEM media (Life

Technologies) according to the manufacturer‟s instructions. Cells were cultured for 24 hours in DMEM/F12 and 0.1% BSA alone or with either 20µM of SP600125 (Santa Cruz

Biotechnology, Dallas, TX), a selective JNK inhibitor, or 10ng/mL TGF-β (Prospec, East

Brunswick, NK). AP-1 reporter activity was measured using a Dual-Luciferase Reporter

Assay System (Promega) using 20µL of lysate. The internal constitutively active renilla luciferase positive control was used to normalize for transfection efficiency. The negative control was used to normalize for background luciferase reporter activity.

Wound healing assay

Ace-1 and Ace-1-Dkk1 were grown to confluence in six-well cell culture plates in triplicate. Media was removed and the cells were rinsed with sterile Dulbecco's phosphate-buffered saline (DPBS). The DPBS was removed and a sterile 200 L pipet tip was used to scratch three separate wounds through the cells. The cells were rinsed with

DPBS and cultured in 1.5mL of DMEM/F12 with 0.1% bovine serum albumin (BSA) alone or with either 20µM of SP600125 or 10ng/mL TGF-β. Using phase contrast, images of the scratch were taken at 0, 6, 12, and 24 hours. The rate of wound closure was calculated using the slope of the line graph created from the time point data.

In vitro growth rate

Ace-1 and Ace-1-Dkk-1 cells were plated in a six-well cell culture plate in triplicate at a density of 200,000 cells per well. Cells were cultured in DMEM/F12 with 0.1% BSA and incubated at 37°C and 5% CO2. The cells were collected using 0.25% trypsin at 24, 48, and 72 hours after the initial plating and counted with an automated cell counter

102

(Nexcelom Bioscience, Lawrence, MA) using trypan-blue dye exclusion to differentiate between live and dead cells. The doubling time was calculated using the formula: (t2- t1) x log(n2)/log(n2/n1), where n is the cell number at time points (t).

RNA extraction and quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR)

Total RNA was extracted from Ace-1-Vector (passage 23, 24 and 25) and Ace-1-Dkk-1 cells (passages 36, 36, and 37) using the Absolutely RNA RT-PCR Miniprep Kit

(Stratagene, Cat. No. 400800) and the manufacturer‟s protocol for adherent tissue culture cells. Total RNA (0.5 μg) was reverse transcribed using the Superscript II First Strand cDNA synthesis kit (Invitrogen) and RT-PCR was performed in duplicate using the

QuantiTect SYBR Green PCR Kit (Qiagen, Cat. No. 204145) and primers for the housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as well as the following genes: bone morphogenic protein 2 (BMP2), folate hydrolase 1 (FOLH1),

FOS, phosphatase and tensin homolog 1 (PTEN1), runt-related transcription factor 2

(RUNX2) and TWIST using canine specific primers (Table 4.1). Genes were selected based on results from a Canine 2.0 Genome array.

Canine 2.0 Genome Array

Three Ace-1-Vector (passage 23, 24 and 25) and Ace-1-Dkk-1 (passages 36, 36, and 37) mRNA samples at passages 10, 11, and 12 were selected and microarray analysis with

GeneChip Canine 2.0 Genome Arrays (Affymetrix, Santa Clara, CA, USA) was performed at OSU's Comprehensive Cancer Center Microarray Shared Resource. R version 2.15.1 (GNU, Boston, MA) was used for statistical analyses of these microarray

103 data. Bioconductor "affy" package was used to import all the cel files. Microarray data were normalized using the "robust multi-array average" (RMA) algorithm [134].

Differential expression analysis of gene expression level was performed using the

"limma" (Linear Models for Microarray Analysis) package from the Bioconductor project

[135]. LIMMA takes into account the correlation between replicates and uses the empirical Bayes approach which gives stable inference also for a relatively small number of arrays. Groups were compared pairwise, and the resulting p-values were adjusted using the false discovery rate (FDR) method.

Statistics

Results are displayed as means ± standard deviation. Normalized gene expression, AP-1 reporter activity, proliferation and migration data was analyzed using two-way ANOVA followed by Sidak‟s multiple comparisons test to perform pair-wise comparisons using

GraphPad Prism version 6.03 (La Jolla, CA). The data from the Dkk-1 ELISA and β- catenin optical density were analyzed by comparing the Ace-1-Dkk-1 group to the Ace-1-

Vector group using Student‟s t-test. Data with P values < 0.05 were considered statistically significant.

104

Results

Dkk-1 secretion

Secretion of human Dkk-1 by Ace-1-Dkk-1 cells was confirmed by ELISA. Ace-1-

Vector had no detectable Dkk-1 secretion while Ace-1-Dkk-1 media contained approximately 4.8 ng/mL after 24 hours of culture (Figure 4.1).

β-catenin immunohistochemistry

For both cell lines, β-catenin immunostaining was primarily within the cytoplasm

(Figures 4.2A and 4.2B). Ace-1-Dkk-1 had significantly lower mean cell optical density of β-catenin immunostaining (Figure 4.2C, p=.011).

AP-1 Reporter activity

Ace-1-Dkk-1 cells had approximately twice the amount of AP-1 reporter activity compared to Ace-1-Vector cells. Treatment with the selective JNK inhibitor, SP600125, reduced AP-1 reporter activity of Ace-1-Dkk-1 cells by approximately 50% (p<0.0001), while treatment with TGF-β increased activity by approximately 24% (p<0.05). For the

Ace-1-Vector cells, treatment did not result in significant changes in reporter activity

(Figure 4.3).

Growth rate

Proliferation of Ace-1-Vector and Ace-1-Dkk-1 cells was similar in vitro (Figure 4.4A).

Treatment with TGF-β significantly reduced the doubling time in the Ace-1-Dkk-1 cells by approximately 20% (p<0.01). No change in proliferation was seen in Ace-1-Vector

105 cells after treatment with SP600125 or TGF-β and no change was seen in Ace-1-Dkk-1 cells after treatment with SP600125.

Wound healing assay

Ace-1-Vector and Ace-1-Dkk-1 cells had similar rates of wound healing in vitro (Figure

4.4B). The rates of migration increased for both Ace-1-Vector (13% increase; p<0.001) and Ace-1-Dkk-1 (36% increase; p<0.0001) after treatment with TGF-β. There was no change in the rates of migration for Ace-1-Vector cells after treatment with SP600125, while Ace-1-Dkk-1 migration was reduced by approximately 67% (p<0.05). qRT-PCR

Expression of BMP2 and FOLH1 were markedly decreased (>95%; p<0.001) in Ace-1-

Dkk-1 cells, while there were moderate decreases in FOS (60% reduction; p<0.001),

PTEN1 (50% reduction; p<0.01), and RUNX2 (65% reduction; p<0.001). TWIST expression was increased by approximately 58% (p<0.001) in Ace-1-Dkk-1 cells compared to Ace-1-Vector cells (Figure 4.5).

Discussion

There are 19 different Wnt proteins in humans, and the Wnts can induce activation of several signal transduction cascades depending on the cellular context [188]. The two broad categories of signaling are the canonical and non-canonical Wnt pathways.

Canonical Wnt signaling results in the stabilization and subsequent nuclear localization

106 of β-catenin where it can alter gene expression pertaining to stem cell self-renewal, cell proliferation, differentiation, migration and apoptosis [189]. Non-canonical Wnt signaling can be divided into several categories: Wnt/JNK, Wnt/Rock, and the

Wnt/calcium signaling pathways [190-192]. Previously we had shown that Dkk-1 expression in the canine prostate cancer cell line, Ace-1, increased tumor growth and bone metastasis and it was proposed that this was through activation of the non-canonical

Wnt/JNK signaling pathway [53]. In this manuscript we confirmed increased Wnt/JNK signaling in the Ace-1-Dkk-1 cell line and showed how modification of Wnt/JNK signaling resulted in downstream effects that contributed to enhanced tumor growth and metastasis.

The Wnt/JNK pathway begins with Wnt binding to a Frizzled receptor resulting in phosphorylation of disheveled (Dsh) and subsequent activation of JNK [193]. JNK then activates the AP-1 transcription factor which localizes to the nucleus to modify gene transcription [194, 195]. AP-1 binds to the motif 5‟-TGA[CG]TCA-3‟ and enhances or represses transcription of many of genes [196, 197]. We have confirmed increased basal

Wnt/JNK signaling in the Ace-1-Dkk-1 cells by showing increased activity of an inducible AP-1 reporter compared to Ace-1-Vector cells. Treatment of Ace-1-Dkk-1 cells with a selective JNK inhibitor, SP600125, markedly decreased activity of the AP-1 reporter, effectively knocking down Wnt/JNK activity. Less striking effects on AP-1 reporter activity, migration and proliferation were seen with treatment of the Ace-1-

Vector cells. This was likely due to its lower basal activity.

107

TGF-β is abundant in bone matrix and is released in response to osteolysis [160].

Interestingly, TGF-β has been reported to be an activator of JNK signaling [198]. The increased osteolytic phenotype of the Ace-1-Dkk-1 cells in vivo would result in greater

TGF-β release and enhanced Wnt/JNK signaling to induce the “vicious cycle” of bone resorption and tumor growth. Treatment of Ace-1-Dkk-1 cells with TGF-β in vitro markedly increased the AP-1 reporter activity, supporting this theory. This conclusion was further supported by the significantly increased proliferation and migration in the

Ace-1-Dkk-1 cells after treatment with TGF-β. In vivo, enhanced proliferation and migration could result in larger tumors and increased numbers of metastases, as was found in our previous study [53].

In this study, we found that Dkk-1 significantly altered gene expression in the Ace-1 cells. Genes with the greatest decrease in expression in the Ace-1-Dkk-1 cells included

BMP2, FOLH1, FOS, PTEN1, and RUNX2 while TWIST was shown to be moderately increased. All of these genes have shown to have importance in cancer progression and metastasis.

TWIST has an essential role in cancer metastasis through its ability to induce epithelial- mesenchymal transition, angiogenesis, and apoptotic resistance [199]. In vivo, Ace-1-

Dkk-1 cells were shown to have an increased rate of metastasis after intracardiac injection in nude mice [53]. Increased TWIST expression is commonly associated with enhanced metastasis and it is likely that it contributed to the increased bone metastasis by the Ace-1-Dkk-1 cells, although further studies will be necessary to determine the role of

TWIST in this model [200]. Recently, small RNA and small molecule TWIST inhibitors

108 have been developed for use as cancer therapeutics, potentially making the Ace-1-Dkk-1 cells a suitable model to investigate their effectiveness [199].

BMP2 and RUNX2 are two genes associated with osteoblast maturation and bone formation [201, 202]. BMP2 is considered a potent inducer of osteoblast differentiation through its effect on RUNX2 expression in osteoblasts [203]. RUNX2 is considered a master osteoblast transcription factor and regulates many genes that determine the osteoblast phenotype [204, 205]. In this study we found that Dkk-1 reduced the expression of these two genes in Ace-1-Dkk-1 cells. In mice it was found that the Ace-1-

Dkk-1 cells produced more lytic bone tumors compared to Ace-1-Vector cells, and this effect may in part have been due to the reduced expression of BMP2 and RUNX2 resulting in reduced osteoblastic differentiation and subsequent enhanced osteolysis.

PTEN1 is a tumor suppressor gene and its expression is commonly decreased in advanced prostate cancer. PTEN1 signaling regulates cell division and can promote apoptosis through its regulation of the phosphoinositide 3-kinase (PI3K) signaling pathway [206, 207]. Decreased PTEN1 expression leads to uncontrolled cell proliferation, decreased apoptosis, as well as enhanced angiogenesis, all of which may be playing a role in the enhanced metastasis demonstrated by Ace-1-DKK-1 cells [208].

In our model, the Ace-1-Dkk-1 cell line expressed lower levels of FOS mRNA compared to the Ace-1-Vector cells. FOS is generally considered a proto-oncogene, however it has a wide variety of functions including effects on differentiation, proliferation, apoptosis and immune response that can alter its role in metastasis [209]. In ovarian, mammary, lung and thyroid cancer, decreased FOS expression has been correlated with increased

109 tumor progression and metastasis and an adverse outcome [210-213]. One possible explanation is that FOS can enhance cell sensitivity to apoptosis, although the mechanism is poorly understood [214]. It is possible that the decreased expression of FOS in the Ace-

1-Dkk-1 cells could enhance cell survival, thereby promoting metastasis and tumor growth.

Interestingly, there was decreased expression of FOLH1 (prostate specific membrane antigen) in the Ace-1-Dkk-1 cells. Many studies have found a positive correlation between prostate cancer progression and FOLH1 expression [215, 216], and FOLH1 is currently being investigated as a target for prostate cancer imaging and therapeutics [217,

218]. In men, FOLH1 expression has been correlated with androgen independence [219].

Since androgen signaling is not as significant in canine prostate cancer, FOLH1 expression may not have the selective advantage as it does in human disease [220, 221].

Therefore the decrease in FOLH1 expression in the Ace-1-Dkk-1 cell line was not surprising.

Β-catenin stabilization and accumulation is one of the main results of canonical Wnt signaling [176]. In previous studies we found that Ace-1-Vector and Ace-1-Dkk-1 cells did not have significantly different amounts of total β-catenin and concluded that canonical Wnt signaling did not play a role in the altered metastatic phenotype between the two cell lines [53]. In this study, immunohistochemistry revealed that there was a significant difference in the amounts of total cell β-catenin between the two cell lines.

Dkk-1 in known to be an inhibitor of canonical Wnt signaling and promotes the destruction of β-catenin, therefore it is not surprising that the Ace-1-Dkk-1 cells have less

110 total β-catenin compared to the Ace-1-Vector cells [222]. Decreased canonical Wnt signaling is typically associated with decreased tumor progression and metastasis; however, these effects may be overshadowed by the effects of increased Wnt/JNK signaling in the Ace-1-Dkk-1 cells [223].

Conclusions

In previous studies we found that Dkk-1 expression in the canine prostate carcinoma cell line, Ace-1, resulted in increased tumor growth, incidence of bone metastasis, and tumor- associated bone lysis in vivo. This study provides evidence that Dkk-1 upregulates the non-canonical Wnt/JNK pathway resulting in downstream alterations in gene expression important in osteoblast stimulation, cell proliferation and EMT of tumor cells. We also found that Ace-1-Dkk-1 cells treated with TGF-β, a growth factor released from the bone during osteolysis, increased Wnt/JNK signaling, cellular proliferation and migration.

Targeting the Wnt/JNK signaling pathway may provide a novel therapeutic pathway to reduce prostate cancer proliferation and metastasis in dogs and men.

111

Tables

Gene Forward Primer (5’-3’) Reverse Primer (5’-3’) GAPDH CCCACTCTTCCACCTTCGAC AGCCAAATTCATTGTCATACCAGG BMP2 TGCGCAGCTTCCACCACGAAGAA CAAAGGTTCCTGCATCTGTTCCCGA FOLH1 GCAGGGGACCCTCTCACACCTG CTCGGAAGACCAACAGCCTCTGTGA FOS TGTGAAGACCATGACGGGAG CCTTTCCCTTCGGATTCTCCT PTEN1 AAAGCTGGAAAGGGACGAACTGGTG ACACATAGCGCCTCTGACTGGGAAT RUNX2 TGCCTCTGGCCTTCCACTCTCAG TGCATTCGTGGGTTGGAGAAGCG TWIST GGCAGGGCCGGAGACCTAGATG TCCACGGGCCTGTCTCGCTT Table 4.1 Primers used for qRT-PCR to characterize Ace-1-Vector and Ace-1-Dkk-1 gene expression.

112

Figures

6 )

l *

n (

e 2

e S 0 r 1 to - c k e k -D -V 1 -1 - e e c c A A

Figure 4.1 Secreted human Dkk-1 in culture media after 24 hours measured by ELISA.

There is no detectable Dkk-1 secretion by Ace-1 cells, while Ace-1-Dkk-1 cells secrete approximately 4.8 ng/mL. (*, p<0.0001, Student‟s t-test, n=3)

113

C y

t s

i 0.4

t m

i Ace-1-Vector

s h

n Ace-1-Dkk-1

c e

o 0.3

s *

t n

p 0.2

m l

0.1

t M

a 0.0 c - r 1

 o - ct k e k V -D - -1 -1 e ce c A A

Figure 4.2 β-catenin immunohistochemical staining and quantification in Ace-1-Vector and Ace-1-Dkk-1 cells

Photomicrograph of Ace-1-Vector (A) and Ace-1-Dkk-1 cells (B) showing predominantly intracytoplasmic

β-catenin immunohistochemical staining. (C) Quantification of β-catenin immunohistochemical staining over 5 fields of Ace-1-Vector and Ace-1-Dkk-1 paraffin-embedded cell pellets. Ace-1-Vector cells have a higher mean total cell OD compared to Ace-1-Dkk-1 (*, p<0.05; Student‟s t-test).

114

* 80

s * t

i * Ace-1-Vector n

U Ace-1-Dkk-1

e 60

e f

i 40

e v

i 20

l e

R 0 t  5 n - 2 e F 1 G 0 tm 0 a T 6 re P T S o N

Figure 4.3 AP-1 luciferase reporter activity of Ace-1-Vector and Ace-1-Dkk-1 cells with no treatment,

TGF-β, or SP600125.

Basal AP-1 reporter activity was markedly higher in the Ace-1-Dkk-1 cells compared to Ace-1-Vector cells

(n=3, p<0.0001, two way ANOVA), indicating a higher level of basal Wnt/JNK signaling. TGF-β treatment significantly increased Wnt/JNK signaling in only in the Ace-1-Dkk-1 cells (n=3, p<0.05, two way

ANOVA). Similarly, treatment with SP600125 resulted in a significant decrease in AP-1 reported activity in the Ace-1-Dkk-1cells (n=3, p<0.0001, two way ANOVA).

115

A 30

) Ace-1-Vector

s r

u Ace-1-Dkk-1 o

h (

e m

g * n

i 20

b u

D 15 t  6 n - 2 e F 1 G 0 tm 0 a T 6 re P T S o N

B 0.8 *

r Ace-1-Vector

i * a

p Ace-1-Dkk-1

e 0.7

R *

d n

u 0.6

o W

f o

0.5

e t

a R 0.4 t  5 n - 2 e F 1 G 0 tm 0 a T 6 B re P T S o N

Figure 4.4 Comparison of cell proliferation and migration between Ace-1-Vector and Ace-1-Dkk-1 cells.

(n=3, repeated twice)

(A) Ace-1-Vector and Ace-1-Dkk-1 doubling times after no treatment, TGF-β, or SP600126. Basal cell proliferation was not significantly different between the two cell lines; however, treatment with TGF-β significantly reduced the doubling time in the Ace-1-Dkk-1 cells (p<0.01, two way ANOVA). There was no change in Ace-1-Dkk-1 proliferation after SP600125 treatment. (B) basal rates of wound closure in Ace-

1-Vector and Ace-1-Dkk-1 cells were similar. Treatment with TGF-β significantly increased migration in both groups (p<0.001 for Ace-1-Vector and p<0.0001 for Ace-1-Dkk-1 cells) while SP600125 significantly decreased migration only in the Ace-1-Dkk-1 cells (p<0.05, two way ANOVA).

116

n 2.0

) o

i * Ace-1-Vector

s D

s Ace-1-Dkk-1

e P

r 1.5

d 1.0

z i

l *

a *

e 0.5 m

v *

N ( e * * R 0.0 2 1 S 1 2 T P H O N X S M L F E N I B O T U W F P R T

Figure 4.5 Expression of select genes in Ace-1-Vector and Ace-1-Dkk-1 cells.

The graph represents the relative expression of BMP2, FOLH1, FOS, PTEN1, RUNX2, and TWIST mRNA normalized to GAPDH in both cell lines. Expression of BMP2 and FOLH1 mRNA were markedly reduced in Ace-1-Dkk-1 cells, while expression of FOX, PTEN1, and RUNX2 mRNA were reduced by approximately 50-65%. TWIST expression was increased by approximately 58% in the Ace-1-Dkk-1 cells. n=3, repeated twice. *, p<0.01

117

Bibliography

1. Siegel R, Naishadham D, Jemal A: Cancer statistics, 2013. CA: a cancer journal for clinicians 2013, 63(1):11-30.

2. Coleman RE: Clinical features of metastatic bone disease and risk of skeletal morbidity. Clinical cancer research : an official journal of the American Association for Cancer Research 2006, 12(20 Pt 2):6243s-6249s.

3. Mehra R, Kumar-Sinha C, Shankar S, Lonigro RJ, Jing X, Philips NE, Siddiqui J, Han B, Cao X, Smith DC, Shah RB, Chinnaiyan AM, Pienta KJ: Characterization of bone metastases from rapid autopsies of prostate cancer patients. Clinical cancer research : an official journal of the American Association for Cancer Research 2011, 17(12):3924-3932.

4. Sturge J, Caley MP, Waxman J: Bone metastasis in prostate cancer: emerging therapeutic strategies. Nature reviews Clinical oncology 2011, 8(6):357-368.

5. Heidenreich A: Bisphosphonates in the management of metastatic prostate cancer. Oncology 2003, 65 Suppl 1:5-11.

6. Weilbaecher KN, Guise TA, McCauley LK: Cancer to bone: a fatal attraction. Nature reviews Cancer 2011, 11(6):411-425.

7. Morgans AK, Smith MR: Bone-targeted agents: preventing skeletal complications in prostate cancer. The Urologic clinics of North America 2012, 39(4):533-546.

8. Langdon SP: Animal modeling of cancer pathology and studying tumor response to therapy. Current drug targets 2012, 13(12):1535-1547.

118

9. Waters DJ, Hayden DW, Bell FW, Klausner JS, Qian J, Bostwick DG: Prostatic intraepithelial neoplasia in dogs with spontaneous prostate cancer. The Prostate 1997, 30(2):92-97.

10. Gamlem H, Nordstoga K, Glattre E: Canine neoplasia--introductory paper. APMIS Supplementum 2008(125):5-18.

11. Rosol TJ, Tannehill-Gregg SH, LeRoy BE, Mandl S, Contag CH: Animal models of bone metastasis. Cancer 2003, 97(3 Suppl):748-757.

12. Smith J: Canine prostatic disease: a review of anatomy, pathology, diagnosis, and treatment. Theriogenology 2008, 70(3):375-383.

13. Waters DJ, Bostwick DG: Prostatic intraepithelial neoplasia occurs spontaneously in the canine prostate. The Journal of urology 1997, 157(2):713-716.

14. Teske E, Naan EC, van Dijk EM, Van Garderen E, Schalken JA: Canine prostate carcinoma: epidemiological evidence of an increased risk in castrated dogs. Molecular and cellular endocrinology 2002, 197(1-2):251-255.

15. Moulay M, Liu W, Willenbrock S, Sterenczak KA, Carlson R, Ngezahayo A, Murua Escobar H, Nolte I: Evaluation of stem cell marker gene expression in canine prostate carcinoma- and prostate cyst-derived cell lines. Anticancer research 2013, 33(12):5421-5431.

16. Sharifi N: Mechanisms of androgen receptor activation in castration-resistant prostate cancer. Endocrinology 2013, 154(11):4010-4017.

17. Derleth CL, Yu EY: Targeted therapy in the treatment of castration-resistant prostate cancer. Oncology (Williston Park) 2013, 27(7):620-628.

18. Gallardo F, Mogas T, Baro T, Rabanal R, Morote J, Abal M, Reventos J, Lloreta J: Expression of androgen, oestrogen alpha and beta, and progesterone receptors in the canine prostate: differences between normal, inflamed, hyperplastic and neoplastic glands. Journal of comparative pathology 2007, 136(1):1-8.

119

19. Leav I, Schelling KH, Adams JY, Merk FB, Alroy J: Role of canine basal cells in postnatal prostatic development, induction of hyperplasia, and sex hormone- stimulated growth; and the ductal origin of carcinoma. The Prostate 2001, 48(3):210- 224.

20. Genega EM, Hutchinson B, Reuter VE, Gaudin PB: Immunophenotype of high- grade prostatic adenocarcinoma and urothelial carcinoma. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc 2000, 13(11):1186-1191.

21. Gobello C, Castex G, Corrada Y: Serum and seminal markers in the diagnosis of disorders of the genital tract of the dog: a mini-review. Theriogenology 2002, 57(4):1285-1291.

22. LeRoy BE, Nadella MV, Toribio RE, Leav I, Rosol TJ: Canine prostate carcinomas express markers of urothelial and prostatic differentiation. Veterinary pathology 2004, 41(2):131-140.

23. Aggarwal S, Ricklis RM, Williams SA, Denmeade SR: Comparative study of PSMA expression in the prostate of mouse, dog, monkey, and human. The Prostate 2006, 66(9):903-910.

24. Wu LY, Johnson JM, Simmons JK, Mendes DE, Geruntho JJ, Liu T, Dirksen WP, Rosol TJ, Davis WC, Berkman CE: Biochemical characterization of prostate-specific membrane antigen from canine prostate carcinoma cells. The Prostate 2014, 74(5):451-457.

25. Paner GP, Luthringer DJ, Amin MB: Best practice in diagnostic immunohistochemistry: prostate carcinoma and its mimics in needle core biopsies. Archives of pathology & laboratory medicine 2008, 132(9):1388-1396.

26. McEntee M, Isaacs W, Smith C: Adenocarcinoma of the canine prostate: immunohistochemical examination for secretory antigens. The Prostate 1987, 11(2):163-170.

27. LeRoy B, Painter A, Sheppard H, Popiolek L, Samuel-Foo M, Andacht TM: Protein expression profiling of normal and neoplastic canine prostate and bladder tissue. Veterinary and comparative oncology 2007, 5(2):119-130. 120

28. Chen Y, Scher HI: Prostate cancer in 2011: Hitting old targets better and identifying new targets. Nature reviews Clinical oncology 2012, 9(2):70-72.

29. Meuten DJ: Tumors in Domestic Animals, 4th edn: Wiley-Blackwell; 2002.

30. Cornell KK, Bostwick DG, Cooley DM, Hall G, Harvey HJ, Hendrick MJ, Pauli BU, Render JA, Stoica G, Sweet DC, Waters DJ: Clinical and pathologic aspects of spontaneous canine prostate carcinoma: a retrospective analysis of 76 cases. The Prostate 2000, 45(2):173-183.

31. Valkenburg KC, Williams BO: Mouse models of prostate cancer. Prostate cancer 2011, 2011:895238.

32. LeRoy BE, Bahnson RR, Rosol TJ: New bone formation in nude mouse calvaria induced by canine prostate tissue. Molecular and cellular endocrinology 2002, 197(1-2):257-263.

33. LeRoy BE, Sellers RS, Rosol TJ: Canine prostate stimulates osteoblast function using the endothelin receptors. The Prostate 2004, 59(2):148-156.

34. Heidegger I, Massoner P, Eder IE, Pircher A, Pichler R, Aigner F, Bektic J, Horninger W, Klocker H: Novel therapeutic approaches for the treatment of castration-resistant prostate cancer. The Journal of steroid biochemistry and molecular biology 2013, 138:248-256.

35. Rove KO, Crawford ED: Evolution of treatment options for patients with CRPC and bone metastases: bone-targeted agents that go beyond palliation of symptoms to improve overall survival. Oncology (Williston Park) 2011, 25(14):1362- 1370, 1375-1381, 1387.

36. Pinto A, Merino M, Zamora P, Redondo A, Castelo B, Espinosa E: Targeting the endothelin axis in prostate carcinoma. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2012, 33(2):421-426.

37. Roh M, Abdulkadir SA: Targeting the endothelin receptor in prostate cancer bone metastasis: Back to the mouse? Cancer biology & therapy 2010, 9(8):615-617.

121

38. Shao N, Wang Y, Jiang WY, Qiao D, Zhang SG, Wu Y, Zhang XX, Wang JL, Ding Y, Feng NH: Immunotherapy and endothelin receptor antagonists for treatment of castration-resistant prostate cancer. International journal of cancer Journal international du cancer 2013, 133(7):1743-1750.

39. Ishizaka K, Azuma H, Matsubara O, Kitahara S, Oshima H: Production of endothelin by canine prostatic epithelial cells and its stimulatory effects on their growth. Journal of andrology 1999, 20(4):529-536.

40. Blomme EA, Sugimoto Y, McCauley LK, Lin YC, Capen CC, Rosol TJ: Stromal and epithelial cells of the canine prostate express parathyroid hormone-related protein, but not the PTH/PTHrP receptor. The Prostate 1998, 36(2):110-120.

41. Anidjar M, Villette JM, Devauchelle P, Delisle F, Cotard JP, Billotey C, Cochand-Priollet B, Copin H, Barnoux M, Triballeau S, Rain JD, Fiet J, Teillac P, Berthon P, Cussenot O: In vivo model mimicking natural history of dog prostate cancer using DPC-1, a new canine prostate carcinoma cell line. The Prostate 2001, 46(1):2-10.

42. Eaton CL, Pierrepoint CG: Growth of a spontaneous canine prostatic adenocarcinoma in vivo and in vitro: isolation and characterization of a neoplastic prostatic epithelial cell line, CPA 1. The Prostate 1988, 12(2):129-143.

43. LeRoy BE, Thudi NK, Nadella MV, Toribio RE, Tannehill-Gregg SH, van Bokhoven A, Davis D, Corn S, Rosol TJ: New bone formation and osteolysis by a metastatic, highly invasive canine prostate carcinoma xenograft. The Prostate 2006, 66(11):1213-1222.

44. Thudi NK, Shu ST, Martin CK, Lanigan LG, Nadella MV, Van Bokhoven A, Werbeck JL, Simmons JK, Murahari S, Kisseberth WC, Breen M, Williams C, Chen CS, McCauley LK, Keller ET, Rosol TJ: Development of a brain metastatic canine prostate cancer cell line. The Prostate 2011, 71(12):1251-1263.

45. Fork MA, Murua Escobar H, Soller JT, Sterenczak KA, Willenbrock S, Winkler S, Dorsch M, Reimann-Berg N, Hedrich HJ, Bullerdiek J, Nolte I: Establishing an in vivo model of canine prostate carcinoma using the new cell line CT1258. BMC cancer 2008, 8:240.

122

46. Anidjar M, Scarlata E, Cury FL, Rocha J, Hamel L, Luz M, Chevalier S: Refining the orthotopic dog prostate cancer (DPC)-1 model to better bridge the gap between rodents and men. The Prostate 2012, 72(7):752-761.

47. Li X, Liao J, Park SI, Koh AJ, Sadler WD, Pienta KJ, Rosol TJ, McCauley LK: Drugs which inhibit osteoclast function suppress tumor growth through calcium reduction in bone. Bone 2011, 48(6):1354-1361.

48. Liao J, Li X, Koh AJ, Berry JE, Thudi N, Rosol TJ, Pienta KJ, McCauley LK: Tumor expressed PTHrP facilitates prostate cancer-induced osteoblastic lesions. International journal of cancer Journal international du cancer 2008, 123(10):2267- 2278.

49. Wise-Milestone L, Akens MK, Rosol TJ, Hojjat SP, Grynpas MD, Whyne CM: Evaluating the effects of mixed osteolytic/osteoblastic metastasis on vertebral bone quality in a new rat model. Journal of orthopaedic research : official publication of the Orthopaedic Research Society 2012, 30(5):817-823.

50. Schade GR, Keller J, Ives K, Cheng X, Rosol TJ, Keller E, Roberts WW: Histotripsy focal ablation of implanted prostate tumor in an ACE-1 canine cancer model. The Journal of urology 2012, 188(5):1957-1964.

51. Halvorson KG, Kubota K, Sevcik MA, Lindsay TH, Sotillo JE, Ghilardi JR, Rosol TJ, Boustany L, Shelton DL, Mantyh PW: A blocking antibody to nerve growth factor attenuates skeletal pain induced by prostate tumor cells growing in bone. Cancer research 2005, 65(20):9426-9435.

52. Halvorson KG, Sevcik MA, Ghilardi JR, Rosol TJ, Mantyh PW: Similarities and differences in tumor growth, skeletal remodeling and pain in an osteolytic and osteoblastic model of bone cancer. The Clinical journal of pain 2006, 22(7):587-600.

53. Thudi NK, Martin CK, Murahari S, Shu ST, Lanigan LG, Werbeck JL, Keller ET, McCauley LK, Pinzone JJ, Rosol TJ: Dickkopf-1 (DKK-1) stimulated prostate cancer growth and metastasis and inhibited bone formation in osteoblastic bone metastases. The Prostate 2011, 71(6):615-625.

54. Navone NM, Logothetis CJ, von Eschenbach AC, Troncoso P: Model systems of prostate cancer: uses and limitations. Cancer metastasis reviews 1998, 17(4):361-371. 123

55. Keller JM, Schade GR, Ives K, Cheng X, Rosol TJ, Piert M, Siddiqui J, Roberts WW, Keller ET: A novel canine model for prostate cancer. The Prostate 2013, 73(9):952-959.

56. Lee YP, Schwarz EM, Davies M, Jo M, Gates J, Zhang X, Wu J, Lieberman JR: Use of zoledronate to treat osteoblastic versus osteolytic lesions in a severe- combined-immunodeficient mouse model. Cancer research 2002, 62(19):5564-5570.

57. Wu TT, Sikes RA, Cui Q, Thalmann GN, Kao C, Murphy CF, Yang H, Zhau HE, Balian G, Chung LW: Establishing human prostate cancer cell xenografts in bone: induction of osteoblastic reaction by prostate-specific antigen-producing tumors in athymic and SCID/bg mice using LNCaP and lineage-derived metastatic sublines. International journal of cancer Journal international du cancer 1998, 77(6):887-894.

58. Yang M, Jiang P, Sun FX, Hasegawa S, Baranov E, Chishima T, Shimada H, Moossa AR, Hoffman RM: A fluorescent orthotopic bone metastasis model of human prostate cancer. Cancer research 1999, 59(4):781-786.

59. Nemeth JA, Harb JF, Barroso U, Jr., He Z, Grignon DJ, Cher ML: Severe combined immunodeficient-hu model of human prostate cancer metastasis to human bone. Cancer research 1999, 59(8):1987-1993.

60. Yonou H, Yokose T, Kamijo T, Kanomata N, Hasebe T, Nagai K, Hatano T, Ogawa Y, Ochiai A: Establishment of a novel species- and tissue-specific metastasis model of human prostate cancer in humanized non-obese diabetic/severe combined immunodeficient mice engrafted with human adult lung and bone. Cancer research 2001, 61(5):2177-2182.

61. Pettaway CA, Pathak S, Greene G, Ramirez E, Wilson MR, Killion JJ, Fidler IJ: Selection of highly metastatic variants of different human prostatic carcinomas using orthotopic implantation in nude mice. Clinical cancer research : an official journal of the American Association for Cancer Research 1996, 2(9):1627-1636.

62. Kozlowski JM, Fidler IJ, Campbell D, Xu ZL, Kaighn ME, Hart IR: Metastatic behavior of human tumor cell lines grown in the nude mouse. Cancer research 1984, 44(8):3522-3529.

124

63. Parkins GE, Klufio GO: Prostate cancer metastasis to the mandible: case report. East African medical journal 2009, 86(5):251-252.

64. Craft N, Chhor C, Tran C, Belldegrun A, DeKernion J, Witte ON, Said J, Reiter RE, Sawyers CL: Evidence for clonal outgrowth of androgen-independent prostate cancer cells from androgen-dependent tumors through a two-step process. Cancer research 1999, 59(19):5030-5036.

65. Davies MR, Lee YP, Lee C, Zhang X, Afar DE, Lieberman JR: Use of a SCID mouse model to select for a more aggressive strain of prostate cancer. Anticancer research 2003, 23(3B):2245-2252.

66. Tsingotjidou AS, Zotalis G, Jackson KR, Sawyers C, Puzas JE, Hicks DG, Reiter R, Lieberman JR: Development of an animal model for prostate cancer cell metastasis to adult human bone. Anticancer research 2001, 21(2A):971-978.

67. Yonou H, Kanomata N, Goya M, Kamijo T, Yokose T, Hasebe T, Nagai K, Hatano T, Ogawa Y, Ochiai A: Osteoprotegerin/osteoclastogenesis inhibitory factor decreases human prostate cancer burden in human adult bone implanted into nonobese diabetic/severe combined immunodeficient mice. Cancer research 2003, 63(9):2096-2102.

68. Dias SJ, Li K, Rimando AM, Dhar S, Mizuno CS, Penman AD, Levenson AS: Trimethoxy-resveratrol and piceatannol administered orally suppress and inhibit tumor formation and growth in prostate cancer xenografts. The Prostate 2013, 73(11):1135-1146.

69. Scatena CD, Hepner MA, Oei YA, Dusich JM, Yu SF, Purchio T, Contag PR, Jenkins DE: Imaging of bioluminescent LNCaP-luc-M6 tumors: a new animal model for the study of metastatic human prostate cancer. The Prostate 2004, 59(3):292-303.

70. Thalmann GN, Anezinis PE, Chang SM, Zhau HE, Kim EE, Hopwood VL, Pathak S, von Eschenbach AC, Chung LW: Androgen-independent cancer progression and bone metastasis in the LNCaP model of human prostate cancer. Cancer research 1994, 54(10):2577-2581.

71. Tso CL, McBride WH, Sun J, Patel B, Tsui KH, Paik SH, Gitlitz B, Caliliw R, van Ophoven A, Wu L, deKernion J, Belldegrun A: Androgen deprivation induces 125 selective outgrowth of aggressive hormone-refractory prostate cancer clones expressing distinct cellular and molecular properties not present in parental androgen-dependent cancer cells. Cancer journal 2000, 6(4):220-233.

72. Navone NM, Olive M, Ozen M, Davis R, Troncoso P, Tu SM, Johnston D, Pollack A, Pathak S, von Eschenbach AC, Logothetis CJ: Establishment of two human prostate cancer cell lines derived from a single bone metastasis. Clinical cancer research : an official journal of the American Association for Cancer Research 1997, 3(12 Pt 1):2493-2500.

73. Yang J, Fizazi K, Peleg S, Sikes CR, Raymond AK, Jamal N, Hu M, Olive M, Martinez LA, Wood CG, Logothetis CJ, Karsenty G, Navone NM: Prostate cancer cells induce osteoblast differentiation through a Cbfa1-dependent pathway. Cancer research 2001, 61(14):5652-5659.

74. Goya M, Miyamoto S, Nagai K, Ohki Y, Nakamura K, Shitara K, Maeda H, Sangai T, Kodama K, Endoh Y, Ishii G, Hasebe T, Yonou H, Hatano T, Ogawa Y, Ochiai A: Growth inhibition of human prostate cancer cells in human adult bone implanted into nonobese diabetic/severe combined immunodeficient mice by a ligand-specific antibody to human insulin-like growth factors. Cancer research 2004, 64(17):6252- 6258.

75. Zhao XY, Malloy PJ, Krishnan AV, Swami S, Navone NM, Peehl DM, Feldman D: Glucocorticoids can promote androgen-independent growth of prostate cancer cells through a mutated androgen receptor. Nature medicine 2000, 6(6):703-706.

76. Hara T, Nakamura K, Araki H, Kusaka M, Yamaoka M: Enhanced androgen receptor signaling correlates with the androgen-refractory growth in a newly established MDA PCa 2b-hr human prostate cancer cell subline. Cancer research 2003, 63(17):5622-5628.

77. Zhao XY, Boyle B, Krishnan AV, Navone NM, Peehl DM, Feldman D: Two mutations identified in the androgen receptor of the new human prostate cancer cell line MDA PCa 2a. The Journal of urology 1999, 162(6):2192-2199.

78. Liu XH, Kirschenbaum A, Yao S, Liu G, Aaronson SA, Levine AC: Androgen- induced Wnt signaling in preosteoblasts promotes the growth of MDA-PCa-2b human prostate cancer cells. Cancer research 2007, 67(12):5747-5753.

126

79. Pretlow TG, Wolman SR, Micale MA, Pelley RJ, Kursh ED, Resnick MI, Bodner DR, Jacobberger JW, Delmoro CM, Giaconia JM, et al.: Xenografts of primary human prostatic carcinoma. Journal of the National Cancer Institute 1993, 85(5):394-398.

80. Holleran JL, Miller CJ, Culp LA: Tracking micrometastasis to multiple organs with lacZ-tagged CWR22R prostate carcinoma cells. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society 2000, 48(5):643-651.

81. Zhau HY, Chang SM, Chen BQ, Wang Y, Zhang H, Kao C, Sang QA, Pathak SJ, Chung LW: Androgen-repressed phenotype in human prostate cancer. Proceedings of the National Academy of Sciences of the United States of America 1996, 93(26):15152- 15157.

82. Corey E, Quinn JE, Buhler KR, Nelson PS, Macoska JA, True LD, Vessella RL: LuCaP 35: a new model of prostate cancer progression to androgen independence. The Prostate 2003, 55(4):239-246.

83. Ellis WJ, Vessella RL, Buhler KR, Bladou F, True LD, Bigler SA, Curtis D, Lange PH: Characterization of a novel androgen-sensitive, prostate-specific antigen- producing prostatic carcinoma xenograft: LuCaP 23. Clinical cancer research : an official journal of the American Association for Cancer Research 1996, 2(6):1039-1048.

84. Corey E, Quinn JE, Bladou F, Brown LG, Roudier MP, Brown JM, Buhler KR, Vessella RL: Establishment and characterization of osseous prostate cancer models: intra-tibial injection of human prostate cancer cells. The Prostate 2002, 52(1):20-33.

85. Corey E, Quinn JE, Vessella RL: A novel method of generating prostate cancer metastases from orthotopic implants. The Prostate 2003, 56(2):110-114.

86. Pinthus JH, Waks T, Schindler DG, Harmelin A, Said JW, Belldegrun A, Ramon J, Eshhar Z: WISH-PC2: a unique xenograft model of human prostatic small cell carcinoma. Cancer research 2000, 60(23):6563-6567.

87. Lee CH, Akin-Olugbade O, Kirschenbaum A: Overview of prostate anatomy, histology, and pathology. Endocrinology and metabolism clinics of North America 2011, 40(3):565-575, viii-ix.

127

88. Kasper S: Survey of genetically engineered mouse models for prostate cancer: analyzing the molecular basis of prostate cancer development, progression, and metastasis. Journal of cellular biochemistry 2005, 94(2):279-297.

89. Berman-Booty LD, Sargeant AM, Rosol TJ, Rengel RC, Clinton SK, Chen CS, Kulp SK: A review of the existing grading schemes and a proposal for a modified grading scheme for prostatic lesions in TRAMP mice. Toxicologic pathology 2012, 40(1):5-17.

90. Berman-Booty LD, Chu PC, Thomas-Ahner JM, Bolon B, Wang D, Yang T, Clinton SK, Kulp SK, Chen CS: Suppression of prostate epithelial proliferation and intraprostatic progrowth signaling in transgenic mice by a new energy restriction- mimetic agent. Cancer prevention research 2013, 6(3):232-241.

91. Sargeant AM, Klein RD, Rengel RC, Clinton SK, Kulp SK, Kashida Y, Yamaguchi M, Wang X, Chen CS: Chemopreventive and bioenergetic signaling effects of PDK1/Akt pathway inhibition in a transgenic mouse model of prostate cancer. Toxicologic pathology 2007, 35(4):549-561.

92. Gingrich JR, Barrios RJ, Morton RA, Boyce BF, DeMayo FJ, Finegold MJ, Angelopoulou R, Rosen JM, Greenberg NM: Metastatic prostate cancer in a transgenic mouse. Cancer research 1996, 56(18):4096-4102.

93. Hensley PJ, Kyprianou N: Modeling prostate cancer in mice: limitations and opportunities. Journal of andrology 2012, 33(2):133-144.

94. Roy-Burman P, Wu H, Powell WC, Hagenkord J, Cohen MB: Genetically defined mouse models that mimic natural aspects of human prostate cancer development. Endocrine-related cancer 2004, 11(2):225-254.

95. Gupta S, Hastak K, Ahmad N, Lewin JS, Mukhtar H: Inhibition of prostate carcinogenesis in TRAMP mice by oral infusion of green tea polyphenols. Proceedings of the National Academy of Sciences of the United States of America 2001, 98(18):10350-10355.

96. Klezovitch O, Chevillet J, Mirosevich J, Roberts RL, Matusik RJ, Vasioukhin V: Hepsin promotes prostate cancer progression and metastasis. Cancer cell 2004, 6(2):185-195. 128

97. Garabedian EM, Humphrey PA, Gordon JI: A transgenic mouse model of metastatic prostate cancer originating from neuroendocrine cells. Proceedings of the National Academy of Sciences of the United States of America 1998, 95(26):15382- 15387.

98. Tu WH, Thomas TZ, Masumori N, Bhowmick NA, Gorska AE, Shyr Y, Kasper S, Case T, Roberts RL, Shappell SB, Moses HL, Matusik RJ: The loss of TGF-beta signaling promotes prostate cancer metastasis. Neoplasia 2003, 5(3):267-277.

99. Wang S, Gao J, Lei Q, Rozengurt N, Pritchard C, Jiao J, Thomas GV, Li G, Roy- Burman P, Nelson PS, Liu X, Wu H: Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer cell 2003, 4(3):209-221.

100. Perez-Stable C, Altman NH, Mehta PP, Deftos LJ, Roos BA: Prostate cancer progression, metastasis, and gene expression in transgenic mice. Cancer research 1997, 57(5):900-906.

101. Cai Y, Kregel S, Vander Griend DJ: Formation of human prostate epithelium using tissue recombination of rodent urogenital sinus mesenchyme and human stem cells. Journal of visualized experiments : JoVE 2013(76).

102. Buttyan R: The mouse prostate reconstitution model of prostate diseases. The Prostate 1997, 33(3):164-165.

103. Thompson TC, Timme TL, Park SH, Yang G, Ren C: Mouse prostate reconstitution model system: A series of in vivo and in vitro models for benign and malignant prostatic disease. The Prostate 2000, 43(4):248-254.

104. Cunha GR, Hayward SW, Wang YZ: Role of stroma in carcinogenesis of the prostate. Differentiation; research in biological diversity 2002, 70(9-10):473-485.

105. Thompson TC, Park SH, Timme TL, Ren C, Eastham JA, Donehower LA, Bradley A, Kadmon D, Yang G: Loss of p53 function leads to metastasis in ras+myc- initiated mouse prostate cancer. Oncogene 1995, 10(5):869-879.

129

106. Park SI, Kim SJ, McCauley LK, Gallick GE: Pre-clinical mouse models of human prostate cancer and their utility in drug discovery. Current protocols in pharmacology / editorial board, SJ Enna 2010, Chapter 14:Unit 14 15.

107. Blouin S, Basle MF, Chappard D: Rat models of bone metastases. Clinical & experimental metastasis 2005, 22(8):605-614.

108. Pollard M, Suckow MA: Hormone-refractory prostate cancer in the Lobund- Wistar rat. Experimental biology and medicine 2005, 230(8):520-526.

109. Liepe K, Geidel H, Haase M, Hakenberg OW, Runge R, Kotzerke J: New model for the induction of osteoblastic bone metastases in rat. Anticancer research 2005, 25(2A):1067-1073.

110. Geldof AA, Rao BR: Prostatic tumor (R3327) skeletal metastasis. The Prostate 1990, 16(4):279-290.

111. Haq M, Goltzman D, Tremblay G, Brodt P: Rat prostate adenocarcinoma cells disseminate to bone and adhere preferentially to bone marrow-derived endothelial cells. Cancer research 1992, 52(17):4613-4619.

112. Blomme EA, Dougherty KM, Pienta KJ, Capen CC, Rosol TJ, McCauley LK: Skeletal metastasis of prostate adenocarcinoma in rats: morphometric analysis and role of parathyroid hormone-related protein. The Prostate 1999, 39(3):187-197.

113. Tennant TR, Kim H, Sokoloff M, Rinker-Schaeffer CW: The Dunning model. The Prostate 2000, 43(4):295-302.

114. Koutsilieris M: PA-III rat prostate adenocarcinoma cells (review). In vivo 1992, 6(2):199-203.

115. Society AC: Cancer Facts and Figures 2013. In. Atlanta: American Cancer Society; 2013.

116. Larson SR, Zhang X, Dumpit R, Coleman I, Lakely B, Roudier M, Higano CS, True LD, Lange PH, Montgomery B, Corey E, Nelson PS, Vessella RL, Morrissey C:

130

Characterization of osteoblastic and osteolytic proteins in prostate cancer bone metastases. The Prostate 2013, 73(9):932-940.

117. Rajpar S, Fizazi K: Bone targeted therapies in metastatic castration-resistant prostate cancer. Cancer journal 2013, 19(1):66-70.

118. Alva A, Hussain M: The changing natural history of metastatic prostate cancer. Cancer journal 2013, 19(1):19-24.

119. Tannehill-Gregg SH, Levine AL, Nadella MV, Iguchi H, Rosol TJ: The effect of zoledronic acid and osteoprotegerin on growth of human lung cancer in the tibias of nude mice. Clinical & experimental metastasis 2006, 23(1):19-31.

120. Breen M, Bullerdiek J, Langford CF: The DAPI banded karyotype of the domestic dog (Canis familiaris) generated using chromosome-specific paint probes. Chromosome Research 1999, 7(5):401-406.

121. Thomas R, Seiser EL, Motsinger-Reif A, Borst L, Valli VE, Kelley K, Suter SE, Argyle D, Burgess K, Bell J, Lindblad-Toh K, Modiano JF, Breen M: Refining tumor- associated aneuploidy through 'genomic recoding' of recurrent DNA copy number aberrations in 150 canine non-Hodgkin lymphomas. Leuk Lymphoma 2011, 52(7):1321-1335.

122. Breen M, Hitte C, Lorentzen TD, Thomas R, Cadieu E, Sabacan L, Scott A, Evanno G, Parker HG, Kirkness EF, Hudson R, Guyon R, Mahairas GG, Gelfenbeyn B, Fraser CM, Andre C, Galibert F, Ostrander EA: An integrated 4249 marker FISH/RH map of the canine genome. BMC Genomics 2004, 5(1):65.

123. Lindblad-Toh K, Wade CM, Mikkelsen TS, Karlsson EK, Jaffe DB, Kamal M, Clamp M, Chang JL, Kulbokas EJ, 3rd, Zody MC, Mauceli E, Xie X, Breen M, Wayne RK, Ostrander EA, Ponting CP, Galibert F, Smith DR, DeJong PJ, Kirkness E, Alvarez P, Biagi T, Brockman W, Butler J, Chin CW, Cook A, Cuff J, Daly MJ, DeCaprio D, Gnerre S, Grabherr M, Kellis M, Kleber M, Bardeleben C, Goodstadt L, Heger A, Hitte C, Kim L, Koepfli KP, Parker HG, Pollinger JP, Searle SM, Sutter NB, Thomas R, Webber C, Baldwin J, Abebe A, Abouelleil A, Aftuck L, Ait-Zahra M, Aldredge T, Allen N, An P, Anderson S, Antoine C, Arachchi H, Aslam A, Ayotte L, Bachantsang P, Barry A, Bayul T, Benamara M, Berlin A, Bessette D, Blitshteyn B, Bloom T, Blye J, Boguslavskiy L, Bonnet C, Boukhgalter B, Brown A, Cahill P, Calixte N, Camarata J, Cheshatsang Y, Chu J, Citroen M, Collymore A, Cooke P, Dawoe T, Daza R, Decktor K, 131

DeGray S, Dhargay N, Dooley K, Dorje P, Dorjee K, Dorris L, Duffey N, Dupes A, Egbiremolen O, Elong R, Falk J, Farina A, Faro S, Ferguson D, Ferreira P, Fisher S, FitzGerald M, Foley K, Foley C, Franke A, Friedrich D, Gage D, Garber M, Gearin G, Giannoukos G, Goode T, Goyette A, Graham J, Grandbois E, Gyaltsen K, Hafez N, Hagopian D, Hagos B, Hall J, Healy C, Hegarty R, Honan T, Horn A, Houde N, Hughes L, Hunnicutt L, Husby M, Jester B, Jones C, Kamat A, Kanga B, Kells C, Khazanovich D, Kieu AC, Kisner P, Kumar M, Lance K, Landers T, Lara M, Lee W, Leger JP, Lennon N, Leuper L, LeVine S, Liu J, Liu X, Lokyitsang Y, Lokyitsang T, Lui A, Macdonald J, Major J, Marabella R, Maru K, Matthews C, McDonough S, Mehta T, Meldrim J, Melnikov A, Meneus L, Mihalev A, Mihova T, Miller K, Mittelman R, Mlenga V, Mulrain L, Munson G, Navidi A, Naylor J, Nguyen T, Nguyen N, Nguyen C, Nicol R, Norbu N, Norbu C, Novod N, Nyima T, Olandt P, O'Neill B, O'Neill K, Osman S, Oyono L, Patti C, Perrin D, Phunkhang P, Pierre F, Priest M, Rachupka A, Raghuraman S, Rameau R, Ray V, Raymond C, Rege F, Rise C, Rogers J, Rogov P, Sahalie J, Settipalli S, Sharpe T, Shea T, Sheehan M, Sherpa N, Shi J, Shih D, Sloan J, Smith C, Sparrow T, Stalker J, Stange-Thomann N, Stavropoulos S, Stone C, Stone S, Sykes S, Tchuinga P, Tenzing P, Tesfaye S, Thoulutsang D, Thoulutsang Y, Topham K, Topping I, Tsamla T, Vassiliev H, Venkataraman V, Vo A, Wangchuk T, Wangdi T, Weiand M, Wilkinson J, Wilson A, Yadav S, Yang S, Yang X, Young G, Yu Q, Zainoun J, Zembek L, Zimmer A, Lander ES: Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature 2005, 438(7069):803-819.

124. Thomas R, Seiser EL, Motsinger-Reif A, Borst L, Valli VE, Kelley K, Suter SE, Argyle D, Burgess K, Bell J, Lindblad-Toh K, Modiano JF, Breen M: Refining tumor- associated aneuploidy through 'genomic recoding' of recurrent DNA copy number aberrations in 150 canine non-Hodgkin lymphomas. Leukemia & lymphoma 2011, 52(7):1321-1335.

125. Thomas R, Duke SE, Karlsson EK, Evans A, Ellis P, Lindblad-Toh K, Langford CF, Breen M: A genome assembly-integrated dog 1 Mb BAC microarray: a cytogenetic resource for canine cancer studies and comparative genomic analysis. Cytogenetic and genome research 2008, 122(2):110-121.

126. Imbriaco M, Larson SM, Yeung HW, Mawlawi OR, Erdi Y, Venkatraman ES, Scher HI: A new parameter for measuring metastatic bone involvement by prostate cancer: the Bone Scan Index. Clinical cancer research : an official journal of the American Association for Cancer Research 1998, 4(7):1765-1772.

127. Roudier MP, True LD, Higano CS, Vesselle H, Ellis W, Lange P, Vessella RL: Phenotypic heterogeneity of end-stage prostate carcinoma metastatic to bone. Human pathology 2003, 34(7):646-653. 132

128. Schneider A, Kalikin LM, Mattos AC, Keller ET, Allen MJ, Pienta KJ, McCauley LK: Bone turnover mediates preferential localization of prostate cancer in the skeleton. Endocrinology 2005, 146(4):1727-1736.

129. Miwa S, Mizokami A, Keller ET, Taichman R, Zhang J, Namiki M: The bisphosphonate YM529 inhibits osteolytic and osteoblastic changes and CXCR-4- induced invasion in prostate cancer. Cancer research 2005, 65(19):8818-8825.

130. Liao J, McCauley LK: Skeletal metastasis: Established and emerging roles of parathyroid hormone related protein (PTHrP). Cancer metastasis reviews 2006, 25(4):559-571.

131. Deftos LJ, Barken I, Burton DW, Hoffman RM, Geller J: Direct evidence that PTHrP expression promotes prostate cancer progression in bone. Biochemical and biophysical research communications 2005, 327(2):468-472.

132. Guise TA: The vicious cycle of bone metastases. Journal of musculoskeletal & neuronal interactions 2002, 2(6):570-572.

133. Simmons JK, Dirksen WP, Hildreth BE, 3rd, Dorr C, Williams C, Thomas R, Breen M, Toribio RE, Rosol TJ: Canine prostate cancer cell line (Probasco) produces osteoblastic metastases in vivo. The Prostate 2014.

134. Bolstad BM, Irizarry RA, Astrand M, Speed TP: A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 2003, 19(2):185-193.

135. Smyth GK: Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Statistical applications in genetics and molecular biology 2004, 3:Article3.

136. Suva LJ, Winslow GA, Wettenhall RE, Hammonds RG, Moseley JM, Diefenbach-Jagger H, Rodda CP, Kemp BE, Rodriguez H, Chen EY, et al.: A parathyroid hormone-related protein implicated in malignant hypercalcemia: cloning and expression. Science 1987, 237(4817):893-896.

133

137. Mundy GR: Metastasis to bone: causes, consequences and therapeutic opportunities. Nature reviews Cancer 2002, 2(8):584-593.

138. Deng X, Tannehill-Gregg SH, Nadella MV, He G, Levine A, Cao Y, Rosol TJ: Parathyroid hormone-related protein and ezrin are up-regulated in human lung cancer bone metastases. Clinical & experimental metastasis 2007, 24(2):107-119.

139. Hildreth BE, 3rd, Werbeck JL, Thudi NK, Deng X, Rosol TJ, Toribio RE: PTHrP 1-141 and 1-86 increase in vitro bone formation. The Journal of surgical research 2010, 162(2):e9-17.

140. Toribio RE, Brown HA, Novince CM, Marlow B, Hernon K, Lanigan LG, Hildreth BE, 3rd, Werbeck JL, Shu ST, Lorch G, Carlton M, Foley J, Boyaka P, McCauley LK, Rosol TJ: The midregion, nuclear localization sequence, and C terminus of PTHrP regulate skeletal development, hematopoiesis, and survival in mice. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 2010, 24(6):1947-1957.

141. Wysolmerski JJ: Parathyroid hormone-related protein: an update. The Journal of clinical endocrinology and metabolism 2012, 97(9):2947-2956.

142. Martin TJ, Moseley JM, Gillespie MT: Parathyroid hormone-related protein: biochemistry and molecular biology. Critical reviews in biochemistry and molecular biology 1991, 26(3-4):377-395.

143. de Castro LF, Lozano D, Portal-Nunez S, Maycas M, De la Fuente M, Caeiro JR, Esbrit P: Comparison of the skeletal effects induced by daily administration of PTHrP (1-36) and PTHrP (107-139) to ovariectomized mice. Journal of cellular physiology 2012, 227(4):1752-1760.

144. Roodman GD: Mechanisms of bone metastasis. The New England journal of medicine 2004, 350(16):1655-1664.

145. Sanders JL, Chattopadhyay N, Kifor O, Yamaguchi T, Brown EM: Ca(2+)- sensing receptor expression and PTHrP secretion in PC-3 human prostate cancer cells. American journal of physiology Endocrinology and metabolism 2001, 281(6):E1267-1274.

134

146. Kiriyama T, Gillespie MT, Glatz JA, Fukumoto S, Moseley JM, Martin TJ: Transforming growth factor beta stimulation of parathyroid hormone-related protein (PTHrP): a paracrine regulator? Molecular and cellular endocrinology 1993, 92(1):55-62.

147. Iwamura M, Wu G, Abrahamsson PA, di Sant'Agnese PA, Cockett AT, Deftos LJ: Parathyroid hormone-related protein is expressed by prostatic neuroendocrine cells. Urology 1994, 43(5):667-674.

148. Asadi F, Farraj M, Sharifi R, Malakouti S, Antar S, Kukreja S: Enhanced expression of parathyroid hormone-related protein in prostate cancer as compared with benign prostatic hyperplasia. Human pathology 1996, 27(12):1319-1323.

149. Bryden AA, Hoyland JA, Freemont AJ, Clarke NW, George NJ: Parathyroid hormone related peptide and receptor expression in paired primary prostate cancer and bone metastases. British journal of cancer 2002, 86(3):322-325.

150. Dougherty KM, Blomme EA, Koh AJ, Henderson JE, Pienta KJ, Rosol TJ, McCauley LK: Parathyroid hormone-related protein as a growth regulator of prostate carcinoma. Cancer research 1999, 59(23):6015-6022.

151. Vaux DL, Cory S, Adams JM: Bcl-2 gene promotes haemopoietic cell survival and cooperates with c-myc to immortalize pre-B cells. Nature 1988, 335(6189):440- 442.

152. Catz SD, Johnson JL: BCL-2 in prostate cancer: a minireview. Apoptosis : an international journal on programmed cell death 2003, 8(1):29-37.

153. McDonnell TJ, Troncoso P, Brisbay SM, Logothetis C, Chung LW, Hsieh JT, Tu SM, Campbell ML: Expression of the protooncogene bcl-2 in the prostate and its association with emergence of androgen-independent prostate cancer. Cancer research 1992, 52(24):6940-6944.

154. Martin JL, Coverley JA, Baxter RC: Regulation of immunoreactive insulin-like growth factor binding protein-6 in normal and transformed human fibroblasts. The Journal of biological chemistry 1994, 269(15):11470-11477.

135

155. Roghani M, Hossenlopp P, Lepage P, Balland A, Binoux M: Isolation from human cerebrospinal fluid of a new insulin-like growth factor-binding protein with a selective affinity for IGF-II. FEBS letters 1989, 255(2):253-258.

156. Horwitz MJ, Tedesco MB, Gundberg C, Garcia-Ocana A, Stewart AF: Short- term, high-dose parathyroid hormone-related protein as a skeletal anabolic agent for the treatment of postmenopausal osteoporosis. The Journal of clinical endocrinology and metabolism 2003, 88(2):569-575.

157. Augustine M, Horwitz MJ: Parathyroid hormone and parathyroid hormone- related protein analogs as therapies for osteoporosis. Current osteoporosis reports 2013, 11(4):400-406.

158. McCarthy DM: Annotation. Fibrosis of the bone marrow: content and causes. British journal of haematology 1985, 59(1):1-7.

159. Shehata M, Schwarzmeier JD, Hilgarth M, Hubmann R, Duechler M, Gisslinger H: TGF-beta1 induces bone marrow reticulin fibrosis in hairy cell leukemia. The Journal of clinical investigation 2004, 113(5):676-685.

160. Sethi N, Dai X, Winter CG, Kang Y: Tumor-derived JAGGED1 promotes osteolytic bone metastasis of breast cancer by engaging notch signaling in bone cells. Cancer cell 2011, 19(2):192-205.

161. Le Bousse-Kerdiles MC, Martyre MC, Samson M: Cellular and molecular mechanisms underlying bone marrow and liver fibrosis: a review. European cytokine network 2008, 19(2):69-80.

162. Hastings RH, Montgrain PR, Quintana R, Rascon Y, Deftos LJ, Healy E: Cell cycle actions of parathyroid hormone-related protein in non-small cell lung carcinoma. American journal of physiology Lung cellular and molecular physiology 2009, 297(4):L578-585.

163. Schluter KD, Weber M, Piper HM: Effects of PTH-rP(107-111) and PTH-rP(7- 34) on adult cardiomyocytes. Journal of molecular and cellular cardiology 1997, 29(11):3057-3065.

136

164. El-Hashash AH, Kimber SJ: PTHrP induces changes in cell cytoskeleton and E-cadherin and regulates Eph/Ephrin kinases and RhoGTPases in murine secondary trophoblast cells. Developmental biology 2006, 290(1):13-31.

165. Iwamura M, di Sant'Agnese PA, Wu G, Benning CM, Cockett AT, Deftos LJ, Abrahamsson PA: Immunohistochemical localization of parathyroid hormone- related protein in human prostate cancer. Cancer research 1993, 53(8):1724-1726.

166. Soki FN, Park SI, McCauley LK: The multifaceted actions of PTHrP in skeletal metastasis. Future oncology 2012, 8(7):803-817.

167. Yang G, Yuan G, Li X, Liu P, Chen Z, Fan M: BMP-2 Induction of Dlx3 Expression Is Mediated by p38/Smad5 Signaling Pathway in Osteoblastic MC3T3- E1 Cells. Journal of cellular physiology 2014, 229(7):943-954.

168. James AW: Review of Signaling Pathways Governing MSC Osteogenic and Adipogenic Differentiation. Scientifica 2013, 2013:684736.

169. Smith BN, Odero-Marah VA: The role of Snail in prostate cancer. Cell adhesion & migration 2012, 6(5):433-441.

170. Morin PJ: beta-catenin signaling and cancer. BioEssays : news and reviews in molecular, cellular and developmental biology 1999, 21(12):1021-1030.

171. Widelitz R: Wnt signaling through canonical and non-canonical pathways: recent progress. Growth factors 2005, 23(2):111-116.

172. Rossini M, Gatti D, Adami S: Involvement of WNT/beta-catenin signaling in the treatment of osteoporosis. Calcified tissue international 2013, 93(2):121-132.

173. Hall CL, Keller ET: The role of Wnts in bone metastases. Cancer metastasis reviews 2006, 25(4):551-558.

174. McNeill H, Woodgett JR: When pathways collide: collaboration and connivance among signalling proteins in development. Nature reviews Molecular cell biology 2010, 11(6):404-413.

137

175. Hall CL, Daignault SD, Shah RB, Pienta KJ, Keller ET: Dickkopf-1 expression increases early in prostate cancer development and decreases during progression from primary tumor to metastasis. The Prostate 2008, 68(13):1396-1404.

176. Pinzone JJ, Hall BM, Thudi NK, Vonau M, Qiang YW, Rosol TJ, Shaughnessy JD, Jr.: The role of Dickkopf-1 in bone development, homeostasis, and disease. Blood 2009, 113(3):517-525.

177. Rao TP, Kuhl M: An updated overview on Wnt signaling pathways: a prelude for more. Circulation research 2010, 106(12):1798-1806.

178. Hall CL, Bafico A, Dai J, Aaronson SA, Keller ET: Prostate cancer cells promote osteoblastic bone metastases through Wnts. Cancer research 2005, 65(17):7554-7560.

179. Hall CL, Kang S, MacDougald OA, Keller ET: Role of Wnts in prostate cancer bone metastases. Journal of cellular biochemistry 2006, 97(4):661-672.

180. Daoussis D, Andonopoulos AP: The emerging role of Dickkopf-1 in bone biology: is it the main switch controlling bone and joint remodeling? Seminars in arthritis and rheumatism 2011, 41(2):170-177.

181. Bismar H, Kloppinger T, Schuster EM, Balbach S, Diel I, Ziegler R, Pfeilschifter J: Transforming growth factor beta (TGF-beta) levels in the conditioned media of human bone cells: relationship to donor age, bone volume, and concentration of TGF-beta in human bone matrix in vivo. Bone 1999, 24(6):565-569.

182. Fuxe J, Vincent T, Garcia de Herreros A: Transcriptional crosstalk between TGF-beta and stem cell pathways in tumor cell invasion: role of EMT promoting Smad complexes. Cell cycle 2010, 9(12):2363-2374.

183. Massague J: TGFbeta in Cancer. Cell 2008, 134(2):215-230.

184. Shaulian E: AP-1--The Jun proteins: Oncogenes or tumor suppressors in disguise? Cellular signalling 2010, 22(6):894-899.

138

185. Derynck R, Zhang YE: Smad-dependent and Smad-independent pathways in TGF-beta family signalling. Nature 2003, 425(6958):577-584.

186. Verrecchia F, Vindevoghel L, Lechleider RJ, Uitto J, Roberts AB, Mauviel A: Smad3/AP-1 interactions control transcriptional responses to TGF-beta in a promoter-specific manner. Oncogene 2001, 20(26):3332-3340.

187. Thudi NK, Martin CK, Nadella MV, Fernandez SA, Werbeck JL, Pinzone JJ, Rosol TJ: Zoledronic acid decreased osteolysis but not bone metastasis in a nude mouse model of canine prostate cancer with mixed bone lesions. The Prostate 2008, 68(10):1116-1125.

188. Eaton S: Release and trafficking of lipid-linked morphogens. Current opinion in genetics & development 2006, 16(1):17-22.

189. Zeng X, Tamai K, Doble B, Li S, Huang H, Habas R, Okamura H, Woodgett J, He X: A dual-kinase mechanism for Wnt co-receptor phosphorylation and activation. Nature 2005, 438(7069):873-877.

190. Bejsovec A: Wnt pathway activation: new relations and locations. Cell 2005, 120(1):11-14.

191. Clevers H: Wnt/beta-catenin signaling in development and disease. Cell 2006, 127(3):469-480.

192. Veeman MT, Axelrod JD, Moon RT: A second canon. Functions and mechanisms of beta-catenin-independent Wnt signaling. Developmental cell 2003, 5(3):367-377.

193. Li F, Chong ZZ, Maiese K: Vital elements of the Wnt-Frizzled signaling pathway in the nervous system. Current neurovascular research 2005, 2(4):331-340.

194. Goldstein B, Takeshita H, Mizumoto K, Sawa H: Wnt signals can function as positional cues in establishing cell polarity. Developmental cell 2006, 10(3):391-396.

195. Katoh M: WNT/PCP signaling pathway and human cancer (review). Oncology reports 2005, 14(6):1583-1588. 139

196. Qing J, Zhang Y, Derynck R: Structural and functional characterization of the transforming growth factor-beta -induced Smad3/c-Jun transcriptional cooperativity. The Journal of biological chemistry 2000, 275(49):38802-38812.

197. Vries RG, Prudenziati M, Zwartjes C, Verlaan M, Kalkhoven E, Zantema A: A specific lysine in c-Jun is required for transcriptional repression by E1A and is acetylated by p300. The EMBO journal 2001, 20(21):6095-6103.

198. Liu Q, Zhang Y, Mao H, Chen W, Luo N, Zhou Q, Chen W, Yu X: A crosstalk between the Smad and JNK signaling in the TGF-beta-induced epithelial- mesenchymal transition in rat peritoneal mesothelial cells. PloS one 2012, 7(2):e32009.

199. Khan MA, Chen HC, Zhang D, Fu J: Twist: a molecular target in cancer therapeutics. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2013, 34(5):2497-2506.

200. Wu CY, Hung JJ, Wu KJ: Linkage between Twist1 and Bmi1: molecular mechanism of cancer metastasis/stemness and clinical implications. Clinical and experimental pharmacology & physiology 2012, 39(8):668-673.

201. Carroll SH, Ravid K: Differentiation of mesenchymal stem cells to osteoblasts and chondrocytes: a focus on adenosine receptors. Expert reviews in molecular medicine 2013, 15:e1.

202. Kamiya N: The role of BMPs in bone anabolism and their potential targets SOST and DKK1. Current molecular pharmacology 2012, 5(2):153-163.

203. Zaidi M: Skeletal remodeling in health and disease. Nature medicine 2007, 13(7):791-801.

204. Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G: Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 1997, 89(5):747-754.

205. Zhao Z, Zhao M, Xiao G, Franceschi RT: Gene transfer of the Runx2 transcription factor enhances osteogenic activity of bone marrow stromal cells in

140 vitro and in vivo. Molecular therapy : the journal of the American Society of Gene Therapy 2005, 12(2):247-253.

206. Paramio JM, Navarro M, Segrelles C, Gomez-Casero E, Jorcano JL: PTEN tumour suppressor is linked to the cell cycle control through the retinoblastoma protein. Oncogene 1999, 18(52):7462-7468.

207. Radu A, Neubauer V, Akagi T, Hanafusa H, Georgescu MM: PTEN induces cell cycle arrest by decreasing the level and nuclear localization of cyclin D1. Molecular and cellular biology 2003, 23(17):6139-6149.

208. Carnero A, Blanco-Aparicio C, Renner O, Link W, Leal JF: The PTEN/PI3K/AKT signalling pathway in cancer, therapeutic implications. Current cancer drug targets 2008, 8(3):187-198.

209. Karin M, Liu Z, Zandi E: AP-1 function and regulation. Current opinion in cell biology 1997, 9(2):240-246.

210. Mahner S, Baasch C, Schwarz J, Hein S, Wolber L, Janicke F, Milde-Langosch K: C-Fos expression is a molecular predictor of progression and survival in epithelial ovarian carcinoma. British journal of cancer 2008, 99(8):1269-1275.

211. Kustikova O, Kramerov D, Grigorian M, Berezin V, Bock E, Lukanidin E, Tulchinsky E: Fra-1 induces morphological transformation and increases in vitro invasiveness and motility of epithelioid adenocarcinoma cells. Molecular and cellular biology 1998, 18(12):7095-7105.

212. Levin WJ, Press MF, Gaynor RB, Sukhatme VP, Boone TC, Reissmann PT, Figlin RA, Holmes EC, Souza LM, Slamon DJ: Expression patterns of immediate early transcription factors in human non-small cell lung cancer. The Lung Cancer Study Group. Oncogene 1995, 11(7):1261-1269.

213. Jin SP, Kim JH, Kim MA, Yang HK, Lee HE, Lee HS, Kim WH: Prognostic significance of loss of c-fos protein in gastric carcinoma. Pathology oncology research : POR 2007, 13(4):284-289.

141

214. Fleischmann A, Jochum W, Eferl R, Witowsky J, Wagner EF: Rhabdomyosarcoma development in mice lacking Trp53 and Fos: tumor suppression by the Fos protooncogene. Cancer cell 2003, 4(6):477-482.

215. Gasinska A, Luczynska E, Wilk W, Cichocka A: Differences in the expression of telomerase and prostate-specific membrane antigen in non-advanced prostatic cancer. Folia histochemica et cytobiologica / Polish Academy of Sciences, Polish Histochemical and Cytochemical Society 2013, 51(1):66-72.

216. Kasperzyk JL, Finn SP, Flavin R, Fiorentino M, Lis R, Hendrickson WK, Clinton SK, Sesso HD, Giovannucci EL, Stampfer MJ, Loda M, Mucci LA: Prostate-specific membrane antigen protein expression in tumor tissue and risk of lethal prostate cancer. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2013, 22(12):2354-2363.

217. Osborne JR, Akhtar NH, Vallabhajosula S, Anand A, Deh K, Tagawa ST: Prostate-specific membrane antigen-based imaging. Urologic oncology 2013, 31(2):144-154.

218. Olson WC, Heston WD, Rajasekaran AK: Clinical trials of cancer therapies targeting prostate-specific membrane antigen. Reviews on recent clinical trials 2007, 2(3):182-190.

219. Wright GL, Jr., Grob BM, Haley C, Grossman K, Newhall K, Petrylak D, Troyer J, Konchuba A, Schellhammer PF, Moriarty R: Upregulation of prostate-specific membrane antigen after androgen-deprivation therapy. Urology 1996, 48(2):326- 334.

220. Perner S, Hofer MD, Kim R, Shah RB, Li H, Moller P, Hautmann RE, Gschwend JE, Kuefer R, Rubin MA: Prostate-specific membrane antigen expression as a predictor of prostate cancer progression. Human pathology 2007, 38(5):696-701.

221. Leroy BE, Northrup N: Prostate cancer in dogs: comparative and clinical aspects. Veterinary journal 2009, 180(2):149-162.

222. Robinson DR, Zylstra CR, Williams BO: Wnt signaling and prostate cancer. Current drug targets 2008, 9(7):571-580. 142

223. Rabbani SA, Arakelian A, Farookhi R: LRP5 knockdown: effect on prostate cancer invasion growth and skeletal metastasis in vitro and in vivo. Cancer medicine 2013, 2(5):625-635.

143