Membrane Type 1 Matrix Metalloproteinase Proteolytic

MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE PROTEOLYTIC

ACTIVITY IN INITIAL ADHESIVE AND INVASIVE EVENTS OF OVARIAN

CANCER METASTASIS

A Dissertation

presented to

the Faculty of the Graduate School

at the University of Missouri

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

______

LANA BRUNEY

Dr. M. Sharon Stack, Dissertation Advisor

DECEMBER 2014

The undersigned, appointed by the Dean of the Graduate School, have examined the dissertation entitled

MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE PROTEOLYTIC

ACTIVITY IN INIITAL ADHESIVE AND INVASIVE EVENTS OF OVARIAN

CANCER METASTASIS

presented by Lana Bruney, a candidate for the degree of Doctor of Philosophy, and hereby certify that, in their opinion, it is worthy of acceptance.

______Dr. M. Sharon Stack

______Dr. Ronald Korthuis

______Dr. Kathy Timms

For my father, Bernard I. L. Bruney. Vous avez cru, même quand je ne l'ai pas.

ACKNOWLEDGEMENTS

I would like to express the deepest appreciation to my advisor, Dr. M. Sharon

Stack, one of the most inspiring women I know. Thank you for believing in me. I am extremely grateful for your patience, guidance, and insightful comments. I consider it an honor to have worked with you and think of you not only as a mentor, but as a friend. Thank you for helping me realize a dream.

I would also like to thank my committee members, Dr. Ronald Korthuis and Dr.

Kathy Timms. Thanks for serving on my committee, despite being so far away. I greatly appreciate your time, support, and suggestions.

I am indebted to my colleagues Dr. Yueying Liu, Jeffrey Johnson, Dr. Zonggao

Shi, Dr. Jing Yang and Dr. Yuliya Klymenko, and offer my sincere gratitude for all of your assistance, education, and time. Thanks so much for making the Stack

Lab a collaborative, open, and energetic environment.

Lastly, but certainly not least, I must thank the following people: My mother,

Cicely A. P. Bruney; Emilia Prosper; Matthew J. Rystrom; Mack Simon, Jr.; Telia

M. Bledson; Bernilia Bertrand; Dr. Natalie Downer; and Dr. Stancy Joseph.

Thank you for the support. Thanks for the encouragement. Thanks for listening.

But most of all, thanks for riding along on my journey.

TABLE OF CONTENTS

1. INTRODUCTION AND BACKGROUND ...... 1

I. INTRODUCTION TO OVARIAN CANCER ...... 1

Statistics...... 1

Etiology...... 2

Multi-step process of ovarian cancer spread...... 5

II. CELL ADHESION IN OVARIAN CANCER METASTASIS ...... 8

Cell adhesion receptors in ovarian cancer...... 8

Integrins...... 9

Cadherins...... 14

III. MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE ...... 18

Matrix metalloproteinases...... 18

Membrane type 1 matrix metalloproteinase...... 20

MT1-MMP structure...... 20

MT1-MMP in EOC...... 23

IV. INTEGRIN LINKED KINASE ...... 25

ILK structure...... 25

ILK controversy...... 27

ILK in EOC...... 29

iii

V. MUCIN 16/ CANCER ANTIGEN 125 ...... 31

Mucins...... 31

Mucin general structure...... 32

Mucins in EOC...... 33

MUC16...... 33

MUC16 structure...... 34

MUC16/CA-125 in EOC...... 35

VI. PROJECT RATIONALE ...... 39

2. ILK ALTERS MT1-MMP-DEPENDENT ACTIVITIES IN EARLY METASTATIC

OVARIAN CANCER EVENTS ...... 47

I. RATIONALE ...... 47

II. RESULTS ...... 49

ILK and MT1-MMP are co-expressed in human ovarian tumor tissues and

cells...... 49

siRNA knockdown of ILK expression in ovarian cancer cells...... 50

Silencing of ILK affects MT1-MMP regulated MCA formation...... 53

Down-regulation of ILK alters ovarian tumor adhesion and MT1-MMP

dependent invasion...... 55

siRNA-mediated knockdown of ILK does not affect ovarian tumor cell

adhesion to mesothelial tissue...... 57

III. DISCUSSION ...... 70

3. MT1-MMP SHEDDING OF MUC16/CA-125 MODULATES MESOTHELIAL

ADHESION AND INVASION ...... 78

I. RATIONALE ...... 78

II. RESULTS ...... 80

Inverse relationship between MT1-MMP and MUC16 expression in ovarian

cells and tissues...... 80

CA-125 is detected in the spent media of OVCA433-MT...... 81

Expression of MT1-MMP alters ovarian cancer cell:mesothelial cell

adhesion...... 82

MT1-MMP expression enhances ovarian cancer cell invasion through a live

mesothelial cell monolayer...... 83

MT1-MMP expression alters attachment of ovarian cancer cells to

mesothelial tissue...... 83

III. DISCUSSION ...... 92

4. OVERALL CONCLUSIONS ...... 99

5. MATERIALS AND METHODOLOGY ...... 105

I. MATERIALS ...... 105

Antibodies...... 105

Other materials...... 106

II. MODELS ...... 107

Cell lines and culture...... 107

Tumor Tissue Microarrays...... 110

III. EXPERIMENTAL METHODS ...... 111

Preparation of whole cell lysates ...... 111

Western blot analysis...... 111

Quantitative Real Time PCR (qPCR)...... 112

Immunohistochemistry...... 113

Immunofluorescence...... 114

FACS analysis...... 115

Enzyme-linked immunosorbent assay (ELISA)...... 115

Formation of multicellular aggregates (MCAs) ...... 116

Collagen I cell adhesion assay...... 117

Mesothelial cell adhesion assay...... 117

Collagen I invasion assay...... 118

Meso-mimetic invasion assay...... 118

Adhesion to a peritoneal explant...... 119

Wash protocol for adhesion assays...... 120

Statistical analyses...... 120

REFERENCES ...... 122

VITA ...... 144

vii

LIST OF FIGURES

Figure 1.1 MULTI-STEP PROCESS OF OVARIAN CANCER SPREAD ...... 41

Figure 1.2 INTEGRIN HETERODIMER ...... 42

Figure 1.3 CLASSICAL CADHERIN ...... 43

Figure 1.4 MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE ...... 44

Figure 1.5 INTEGRIN LINKED KINASE ...... 45

Figure 1.6 MUCIN16 ...... 46

Figure 2.1 EXPRESSION OF ILK IN OVARIAN ADENOCARCINOMAS AND

CELLS ...... 59

Figure 2.2 CO-LOCALIZATION OF ILK AND MT1-MMP IN OVARIAN CANCER

CELL LINE DOV13 ...... 61

Figure 2.3 siRNA DOWN-REGULATION OF ILK EXPRESSION IN OVARIAN

CANCER CELLS ...... 62

Figure 2.4 SILENCING OF ILK AFFECTS FORMATION OF MCAs ...... 64

Figure 2.5 SILENCING OF ILK ALTERS TUMOR CELL ADHESION AND

INVASION ...... 66

Figure 2.6 SILENCING OF ILK DOES NOT AFFECT CELL-TO-MESOTHELIAL

ADHESION IN AN EX VIVO PERITONEAL EXPLANT ...... 68

Figure 3.1 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN CANCER

TISSUES ...... 84

viii

Figure 3.2 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN CANCER

CELLS ...... 85

Figure 3.3 MT1-MMP OVEREXPRESSION IS ASSOCIATED WITH

DECREASED CELL SURFACE EXPRESSION OF MUC16 ...... 86

Figure 3.4 EXPRESSION OF MT1-MMP ALTERS TUMOR CELL ADHESION TO

AND INVASION THROUGH A LIVE MESOTHELIAL CELL MONOLAYER ...... 88

Figure 3.5 EXPRESSION OF MT1-MMP AFFECTS CELL-TO-MESOTHELIAL

ADHESION IN AN EX VIVO PERITONEAL EXPLANT ...... 90

LIST OF ABBREVIATIONS

3D Three-Dimensional

CA-125 Cancer Antigen 125

CMFDA 5-Chloromethylfluorescein Diacetate

DOV13 Ovarian cancer cell line

Ovarian cancer cell line DOV13 incubated with small DOV+QLT molecule ILK inhibitor QLT0267

Ovarian cancer cell line DOV13 transfected with ILK DOV-ILK-KD siRNA

ECM Extracellular Matrix

EOC Epithelial Ovarian Cancer

FBS Fetal Bovine Serum

GAPDH Glyceraldehyde-3-Phosphate Dehydrogenase

GM6001 (galardin/ilomastat), a non-specific MMP GM6001 inhibitor. Referred to as “GM” in figures

HPRT Hypoxanthine Ribosyltransferase

IHC Immunohistochemistry

ILK Integrin Linked Kinase

ILKAP Integrin Linked Kinase Associated Protein

LP9 Human mesothelial cell line

MCA Multicellular aggregate

MEM Minimum Essential Media

MMP Matrix Metalloproteinase

MT1-MMP Membrane-type I Matrix Metalloproteinase

MUC16 Mucin 16

Ovarian cancer cell line DOV13 transfected with non- NTC targeted control siRNA

Ovarian cancer cell line DOV13 transfected with non-

NTC+QLT targeted control siRNA incubated with small molecule ILK

inhibitor QLT0267

OSE Ovarian Surface Epithelium

OVCA Ovarian Cancer

OVCA433 Ovarian cancer cell line

Ovarian cancer cell line OVCA433 stably transfected with OVCA433-E240A MT1-MMP mutant E240A (catalytically inactive MT1-MMP)

Ovarian cancer cell line OVCA433 stably transfected with OVCA433-MT MT1-MMP

PINCH Particularly interesting new cysteine-histidine-rich protein

PBS Phosphate Buffered Saline

xii

Surveillance, Epidemiology, and End Results. A National

SEER Cancer Institute program that provides Americancancer

statistics

xiii

1. INTRODUCTION AND BACKGROUND

I. INTRODUCTION TO OVARIAN CANCER

Statistics.

Epithelial ovarian cancer (EOC) is one of the most common gynecologic

malignancies, generally developing in women over the age of forty (Jemal et al.

2011; Siegel et al. 2013). EOC symptoms are usually vague, and with no

sufficiently accurate screening test currently available, early diagnosis is often

difficult. When EOC are diagnosed prior to metastatic dissemination, the overall

5-year survival rate is 92%; however, nearly 85% of women with EOC are diagnosed with metastasis already present, dropping the survival rate to less

than 30% (American Cancer Society 2013; N Howlader et al. 2013). Relative survival varies by age; post-diagnosis, the 5-year survival rate of women aged sixty-five and older is nearly half that of women younger than sixty-five years of age (27% vs 56%, respectively) (Yancik 1993). With 14,030 deaths in 2013, EOC

kills more American women than any other gynecologic cancer (American

Cancer Society 2013). It is the fifth leading cause of overall cancer death among

American women, with 22,240 newly diagnosed cases expected in 2014

(American Cancer Society 2013).

Etiology.

The ovarian surface epithelium (OSE) is the modified pelvic mesothelium that

covers the ovary (McCluggage and Wilkinson 2005). This mesothelium is

comprised of a single layer of non-distinguished flat-to-cuboidal epithelial cells that are separated by a basement membrane from the underlying ovarian stroma

(Cannistra 2004; Gubbels et al. 2010). Despite this epithelial morphology, OSE cells exhibit some mesenchymal characteristics including: the expression of vimentin, and N-cadherin; the production of metalloproteases; and a distinct lack of E-cadherin expression (Auersperg et al. 1999; Strauss et al. 2011). This complex physiology sparked much speculation; the revelation that nearly 90% of human ovarian cancers are epithelial in origin initiated current dogma: EOC arise directly from the OSE. Over forty years ago, the available observational and epidemiological data were assessed to determine a biological initiator for EOC development. In a widely cited 1971 Lancet article, Fathalla postulated that the cyclical ovulation-induced damage and repair of the ovarian surface eventually results in a malignant transformation of the OSE, leading to tumor formation

(Fathalla 1971). This ‘incessant ovulation’ theory was based on data which

suggested a correlation between hormonal contraception and a decreased risk of

EOC, combined with the observation that nulliparity in the absence of hormonal contraception was associated with an increased risk of EOC (Fathalla 1971).

Studies focused on the development of peritoneal carcinomatosis by domestic hens forced to continually ovulate provided further support to suggest that this sort of repetitive damage to the OSE was instrumental in its malignant transformation (Bahr and Palmar 1989; Damjanov 1989). However, recently, the culmination of several morphological, molecular, and epidemiological observations has revealed information to question the current theory. These observations include: a mismatch of histologic and immunophenotypes in the

OSE versus those found within EOC; the pathologic similarity of EOC to

Mullerian epithelia (which includes epithelia of the fallopian tubes, endometrium, and endocervix); the morphological similarity of the three most common EOC tumor subtypes (serous, endometrioid, and mucinous) to other gynecologic carcinomas (fallopian tube, endometrium, and endocervix, respectively); the finding of precancerous lesions on sites other than the ovary or ovarian surface; and the post-surgical discovery of malignant tumors histologically identical to ovarian carcinomas in those who have had their ovaries removed for non- cancerous reasons (Bloss et al. 1993; Rodriguez and Dubeau 2001; Cheng et al.

2005; Dubeau 2008; Li et al. 2011). Therefore, alternate theories propose that many EOC originate outside the ovary, suggesting that epithelial cells derived

from the endometrium, the endocervix, or the fallopian tubes may play a role

(Crum et al. 2007; Hillier 2012; Kim et al. 2012b). Of the alternate theories proposed, the implication of fallopian tube fimbriae as a potential EOC origin site is arguably the most stongly supported (Kurman and Shih 2010; Roh et al. 2010;

Li et al. 2011; Kim et al. 2012b). The proximity of this fimbriated end to the ovarian surface may provide an opportunity for epithelia to detach and implant onto the ovarian surface, eventually leading to cellular transformation and EOC formation.

Complicating the determination of pathogenesis is the heterogeneity inherent in

EOC tumors. Kurman and Shih suggested nomenclature by which to categorize the various EOC tumor types into two broad groups, Type I or Type II, based upon similarities in morphologies and relative genetic disparity (Kurman and Shih

2010). Representing only about 25% of all EOC, Type I tumors include low-grade serous, low-grade endometrioid, clear cell, and mucinous carcinomas. These tumors are characterized by slow growth, a larger size while confined to the ovary, and distinct, specific mutations in a number of genes (including BRAF,

KRAS, EBB2, and PTEN) (Kurman and Shih 2010). Type II tumors, on the other hand, represent the majority of all EOC and are responsible for nearly 90% of all

EOC deaths (Kurman and Shih 2010). These aggressive tumors include high- grade serous, high-grade endometrioid, undifferentiated carcinomas, and

malignant mixed mesodermal tumors. Type II tumors are characterized by a high

frequency of TP53 mutations and, in contrast to Type I, are usually found in

areas other than the ovary.

Multi-step process of ovarian cancer spread.

Regardless of the site of origin, ovarian cancers undergo a unique mechanism of cancer spreading (Fig. 1.1). Once epithelial cells undergo neoplastic transformation, they detach and exfoliate from the primary tumor surface, disseminating into the peritoneal cavity (Fig. 1.1, A). These cells then exist as

single cells and cohesive multicellular aggregates (MCAs) which either remain

unattached within the peritoneal fluid (Fig. 1.1, B) or subsequently adhere to the

layer of mesothelial cells lining the peritoneal cavity (Fig. 1.1, C). After adhesion,

these cells induce mesothelial cell retraction and exposure of the underlying

extracellular matrix (ECM) to enable metastatic anchoring in the interstitial

collagen-rich sub-mesothelial ECM (Fig. 1.1, D).

As described above, the ability of tumor cells to successfully invade the sub-

mesothelial matrix precedes the implantation, formation, and anchoring of

secondary lesions. Mesothelial cell retraction is generally accepted as a major

part of this process, but the cellular and molecular mechanisms governing this

movement have yet to be elucidated. There is, however, evidence to suggest that

the integrity of the mesothelial cell monolayer is altered when tumor cells are

present. Electron micrographs of mesothelial tissue sections with peritoneal

metastases present revealed mesothelial cells as more rounded and separated

from each other, while those of normal peritoneal mesothelial cells (those without

metastases present) are shown as a flat sheet, so continuous and cohesive that

cell-cell boundaries are difficult to distinguish (Birbeck and Wheatley 1965; Witz et al. 1999). Additionally, micrographs of excised human peritoneum-associated

tumors have shown that the metastases are attached to connective tissue

directly beneath the mesothelial cell layer (sub-mesothelial ECM) and that

mesothelial cells are not present directly under the tumor mass (Witz et al. 1999).

In vitro assessment of interactions between ovarian MCAs and mesothelial cells

suggest that MCAs disrupt mesothelial cell-cell junctions and penetrate beneath mesothelial cells for high affinity binding to the sub-mesothelial matrix (Koga et al. 1980; Kiyasu et al. 1981; Niedbala et al. 1985). Recently, Iwanicki and colleagues have monitored this interaction (MCAs-mesothelial cell) in real time and have produced evidence to substantiate previous correlations between MCA attachment and mesothelial cell retraction (Iwanicki et al. 2011; Davidowitz et al.

2014). Iwanicki et al. further suggest that ovarian tumor spheroids, once attached and spread on the mesothelial monolayer, use integrin-dependent activation of myosin and traction force to physically displace mesothelial cells, clearing mesothelial cells from underneath the tumor spheroid to facilitate binding to the sub-mesothelial matrix (Iwanicki et al. 2011).

Metastases are most commonly found within the omentum, the peritoneum, the diaphragm, and bowel surfaces (Nguyen et al. 2009; Lengyel 2010). This multi- step process of cancer cell adhesion, migration, and anchoring eventually results in the death of the patient from the burden of metastatic disease (Skubitz 2002).

Although an early step in this process involves the adhesion of EOC cells to mesothelial cells and their associated sub-mesothelial ECM, few studies have focused on this interaction directly. This dissertation focuses on these processes, specifically the roles of membrane type 1 matrix metalloproteinase, integrin linked kinase, and mucin16, which will be reviewed in the following sections.

II. CELL ADHESION IN OVARIAN CANCER METASTASIS

Cell adhesion receptors in ovarian cancer.

Key molecules involved in homotypic and heterotypic cell-cell adhesion have

been identified and in many cases, these molecules or those involved in their

downstream signaling pathways have been recognized as potential therapeutic

or diagnostic molecular targets (Elmasri et al. 2009). In EOC metastasis,

multiple forms of cell-cell and cell-matrix adhesion are used to facilitate the

survival and dissemination of epithelial cells. When malignant epithelial cells are

shed into the peritoneum, released soluble adhesion molecules aid anoikis

resistance, homotypic cell-cell adhesion promotes the formation of MCAs, and

heterotypic cell-cell interactions bind these malignant cells to mesothelial cells

(Kim et al. 2012a). The molecules responsible for mediating cell adhesion are

typically multiprotein complexes consisting of three functional classes of proteins: cell adhesion receptors (typically transmembrane glycoproteins that mediate binding to the ECM or to receptors on other cells and determine the specificity of the cell-cell and cell-ECM receptor recognition and binding at the extracellular

surface (Gumbiner 1996)); ECM molecules (large glycoproteins which assemble

into complex structural arrays able to maintain interactions with multiple cell

surface receptors); and cytoplasmic plaque/peripheral membrane proteins (which

interface intracellularly with cell adhesion receptors, providing structural and

functional linkages between adhesion receptors and the actin microfilaments,

microtubules, and intermediate filaments of the cytoskeleton) (Gumbiner 1996).

Biochemical events within the cell regulate the functions of cell adhesion

complexes. Once thought to be static architectural entities, these complexes are

actually dynamic units able to facilitate signaling between the cell and the

external environment (Rosales et al. 1995).

Integrins.

In EOC metastasis, retraction of peritoneal mesothelial cells exposes the underlying collagen-rich ECM, resulting in integrin-mediated anchoring of metastatic cells. Cell adhesion occurs via the interaction of specific cell surface receptors with extracellular adhesive molecules (such as collagen, fibronectin, laminin, and vitronectin) which then trigger intracellular changes, resulting in altered cellular behavior (Akiyama 1996). The process of cell adhesion is initiated by a binding event and is often followed by cell flattening and spreading as cells adhere to the underlying matrix and form focal adhesions.

Arguably, the most well characterized cell adhesion receptors are integrins.

Integrins are a family of transmembrane glycoproteins which regulate many cell- cell and cell-matrix adhesive interactions. Integrins are comprised of a

functionally linked α and β subunit, the cytoplasmic domains of both which make important contributions to various aspects of overall integrin function (Fig. 1.2).

Each subunit consists of a large extracellular portion (~700-1100 amino acids), a

single membrane-spanning region, and a generally short (<80 amino acid)

cytoplasmic tail (Ruoslahti 1991; Hynes 1992; Monniaux et al. 2006;

Desgrosellier and Cheresh 2010; Aoudjit and Vuori 2012). αβ heterodimerization

and binding to ECM ligands induces dynamic conformational changes within the

domains (Monniaux et al. 2006; Gilcrease 2007). Both the α and β subunit cytoplasmic domains contribute to integrin functions including integrin- cytoskeletal interactions, cell motility, endocytosis, and signal transduction, through direct binding and activation of cytoplasmic proteins, most commonly focal adhesion kinase (FAK) (Burridge and Chrzanowska-Wodnicka 1996;

Yamada and Geiger 1997; Miyamoto et al. 1998). The β cytoplasmic domain in particular, has been noted for its role in facilitating cross-talk between different integrins (Blystone et al. 1995). In mammals, twenty-four distinct αβ heterodimers have been identified, with each α/β pairing providing specific ligand binding and the resulting ability to modulate a variety of intracellular signal transduction cascades (Hynes 1992; Rosales et al. 1995; Monniaux et al. 2006; Desgrosellier and Cheresh 2010).

In ovarian tissues, α2, α3, αv, β1 and β3 integrin subunits are highly expressed;

notably the expression of αv and β3 integrins in malignant EOC tumors is

significantly higher than that of other EOC tissue and there is evidence to

suggest that both are relevant to pelvic and abdominal diffusion and metastasis

of EOC cells (Davidson et al. 2003; Chen et al. 2009; Wang et al. 2011). Also,

expression of β1 has been directly correlated with shorter overall survival in

women with EOC (Cannistra 2004). Moreover, cDNA microarray studies have

shown that β1 integrin engagement also regulates expression of multiple gene

products that contribute to metastatic successes, specifically membrane type 1

matrix metalloproteinase (Casey et al. 2001; Barbolina et al. 2007). Integrin subunit αv has been found to regulate cell proliferation in EOC and β1 integrins

have been shown to mediate multiple adhesive events both in vitro and in vivo,

including: interactions between EOCs and the sub-mesothelial collagen rich

ECM; the direct binding of EOCs to the peritoneal mesothelial cells that line the

abdominal organs; and the cell-cell adhesion leading to MCA formation both in vitro and in vivo (Gardner et al. 1995; Li et al. 1999; Casey et al. 2001; Ellerbroek et al. 2001; Leung-Hagesteijn et al. 2001; Cruet-Hennequart et al. 2003; Durbin et al. 2009).

The relationship between integrin structure and function has been extensively scrutinized. These non-covalent, heterodimeric receptors physically interact with

ECM proteins (such as fibronectin, collagen, vitronectin, and laminin) and undergo conformational changes (clustering within the plasma membranes) within focal adhesions, contributing to integrin functions including integrin- cytoskeletal interactions, cell motility, endocytosis, and signal transduction

(Burridge and Chrzanowska-Wodnicka 1996; Van Nhieu et al. 1996; Yamada and Geiger 1997). This clustering within the plasma membrane connects extracellular signals to intracellular adaptor molecules (known as “outside-in”

signaling) to activate several cell signaling pathways, affecting the in vitro growth

morphology, survival, proliferation, and differentiation of both normal and

malignant cells (Davidson et al. 2003; Cabodi et al. 2010). Conversely, during

“inside-out” signaling, signals generated within the cytoplasmic domain can

induce conformational changes that alter ligand recognition (O’Toole et al. 1994;

Humphries 1996; Keely et al. 1998). Generally, integrin interaction and

association varies; while some integrins interact monogamously with a single

large ECM protein, most will actually bind several distinct ligands (Fu et al. 2012).

Additionally, some integrins recognize and associate with short peptide

sequences embedded within a larger macromolecule, while other integrins

interact with members of other adhesion receptor families (i.e. cadherins) (Fu et

al. 2012). In terms of simple cell adhesion, integrin expression is often redundant;

cells very often display many integrins capable of interacting with a specific ECM

protein (Fu et al. 2012). For example, the lateral and apical surfaces of ovarian

epithelial cells express both integrin α2β1 and α3β1, both of which interact with collagen I (Moser et al. 1996). Notably, it has also been suggested that expression of α2β1 integrin directly influences spheroid disaggregation and proteolysis in EOC (Schaefer et al. 2012).

Also associated with the engagement of β1 integrins is the extracellular lipid

signaling molecule, lysophosphatidic acid (LPA). LPA is constitutively produced

by peritoneal mesothelial cells and has been extensively studied in EOC (Rosanò

et al. 2006; Said et al. 2007b; Durbin et al. 2009). LPA activity has been shown to

promote prosurvival and proangiogenic signals, as well as cell proliferation and

migration (Fang et al. 2004; Huang et al. 2004; Chou et al. 2005; Symowicz et al.

2005; Said et al. 2007b; Liu et al. 2012b). Higher than normal levels have been

confirmed in the serum and ascites of EOC patients and some have suggested

use of LPA as a biomarker of EOC, but this remains controversial (Sutphen et al.

2004). Specifically, LPA has been shown to enhance the motility and

invasiveness of ovarian cancer cells and has a distinct role in the chemotactic

activity leading to shed ovarian cancer cells (Fishman et al. 2001; Ren et al.

2006; Said et al. 2007a). There is also evidence to suggest that the enhanced

migratory and invasiveness of EOC cells is attributed to augmented matrix

metalloproteinase (MMP) expression via LPA promotion of integrin activation (via

inside-out signaling) (Fishman et al. 2001; Said et al. 2007a). Fishman et al

noted increased β1 integrin expression when ovarian cancer cells were treated

with pathophysiological levels of LPA and further speculated that β1 integrin

clustering may regulate or promote expression of specific MMPs (Fishman et al.

2001).

Of the cell adhesion receptors, integrins are those most directly linked to the

processes underlying the unique mechanism of metastasis inherent to ovarian

cancer. Understanding the cell-cell and cell-matrix adhesive interactions

regulated by this family of receptors throughout the metastatic process (from

initial shedding events to intraperitoneal implantation) may be key to the development of pharmacological therapies designed to limit the extent of metastatic spread.

Cadherins.

In EOC, another important family of cell adhesion receptors is the cadherins.

These transmembrane glycoproteins mediate Ca2+-dependent intercellular

adhesion in specialized sites of cell-to-cell adhesion, known as adherens junctions (Elmasri et al. 2009). Cadherins are present in almost all solid tissues and play a fundamental role in normal development via the maintenance of proper cell-cell contacts (Huber et al. 1996; Skubitz 2002). There are several cadherin subfamilies (type I classical, type II atypical, desmosomal, flamingo,

ungrouped, etc); subtle differences between cadherins make each type uniquely suited for specific tissue and cell types (Chu et al. 2006). Type I classical cadherins (heretofore referred to simply as “cadherins”), including E- (epithelial),

N- (neural), and P- (placental), are homotypic with five Ca2+-dependent extracellular domains and a relatively short cytoplasmic domain (Fig. 1.3)

(Takeichi 1995).

Dynamic linkage of the cytoskeleton to adherens junctions in the plasma

membrane is required for cell adhesion; here, cadherins utilize the connection to

the cytoskeleton to exert their biological effects and control the specificity,

organization, and dynamics of cell-cell adhesions. As such, altered cadherin expression can often result in pathophysiological consequences including the progression of tumourigenesis, characterized by changes in cell differentiation, loss of cell-cell adhesion and enhanced cell migration (Zigler et al. 2010). Like integrins, cadherins have the ability to influence signaling pathways responsible for cell growth, survival and transcriptional activity. Cadherins also link to many of the same cytoskeletal elements as integrins. These commonalities suggest that

while cadherins and integrins are distinctive adhesive molecules in their own

right, they likely interact in an integrated network as opposed to separate

cascades, defined by the widely used term “adhesive crosstalk” (Weber et al.

2011).

Recently, Weber et al. have described three general modes of adhesive interaction for an integrin-cadherin integrated network: input-output signaling, convergent signaling, and lateral coupling (Weber et al. 2011). Input-output signaling describes adhesive events that modulate expression or functional activities of other adhesive molecules (Weber et al. 2011). These adhesive signals can potentially promote changes in the transcriptional or signaling effector activity that regulates cell-cell and cell-matrix adhesions, effect changes in membrane trafficking, modify cytoskeletal associations, and influence binding affinity (Weber et al. 2011). Convergent signaling utilizes communal downstream effector molecules (e.g., non-receptor tyrosine kinases, adaptor and scaffolding proteins, and small GTPases) of both integrin and cadherin adhesive events

(Weber et al. 2011). Integrin/cadherin convergent interaction also involves shared cytoskeletal networks, where cell-cell and cell-matrix adhesions are linked to shared structural elements, and overlapping macromolecular assemblies of actin, microtubules, and intermediate filaments serve as physical scaffolds to connect adhesion complexes (Weber et al. 2011). In this interaction, independently initiated adhesive signals converge to one or more shared effectors, effectively granting vincinal integrins and cadherins the ability to enact their effects synergistically (Weber et al. 2011). The third mode, lateral coupling, employs cytoplasmic adaptor or scaffolding proteins to promote lateral

associations of integrins and cadherins, irrespective of shared cytoskeletal structure, proximity, or cell-cell/cell-matrix adhesion (Weber et al. 2011).

However, the physiological significance of these lateral integrin-cadherin associations is as yet undetermined. While all three general modes are differentiated by their mechanisms, Weber et al. is careful to note that each adhesive interaction can become part of a complicated feedback loop where the different modes consort (i.e., the original adhesive signal is modulated by events activated downstream, thus initializing a separate integrin-cadherin adhesive interaction which may then continue the pattern of engagement/disengagement of adhesion molecules).

III. MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE

Matrix metalloproteinases.

Many of the extracellular signaling events responsible for the regulation of cell behavior occur at the cell membrane and are controlled by pericellular proteolysis. Matrix metalloproteinases (MMPs) are zinc-dependent extracellular matrix degrading proteases that facilitate proteolysis at the cell surface to directly influence cell behavior (Martin et al. 2000; Vu 2000; Löffek et al. 2011). MMPs have been implicated in normal physiological development as well as during disease processes. Normally, the proteolytic activity of MMPs is strictly controlled by endogenous inhibitors, but in pathological conditions, including oncogenic malignancies, arthritis, Alzheimer’s disease, and coronary artery disease, MMP activity is dysregulated, leading to excessive degradation of ECM components

(Vu 2000; Egeblad and Werb 2002; Freije et al. 2003). MMPs are up-regulated in almost every type of human oncogenic disease, exerting their pro-metastatic effects via cleavage of a diverse group of substrates, including ECM structural components, receptor tyrosine kinases, and other cell adhesion molecules

(Martin et al. 2000; Löffek et al. 2011; Vargová et al. 2012). Enhanced expression of specific MMPs is crucial to tissue remodeling and is thought to play

a major role in many cellular behaviors including: apoptosis, cell adhesion, proliferation, migration, invasion, metastasis, and tumor angiogenesis

(Kessenbrock et al. 2010; Pytliak et al. 2012; Vargová et al. 2012). These observations have prompted considerable interest in utilizing MMP inhibitors in large scale cancer clinical trials, but these have all met with limited efficacy and treatment failure, attributed to the broad spectrum nature of the compounds

(Lengyel 2010; Nakayama et al. 2012; Eiró 2013).

Nearly twenty-five MMPs have been described; these are numerically named and classified according to structure, with five secreted classes and three membrane- type (Vargová et al. 2012). Secreted MMPs localize to the cell surface through interactions with other molecules (integrins, proteoglycans, or various ECM proteins), while the membrane type MMPs are directly tethered to the cell membrane through covalent bonds. Whether secreted or membrane type, the basic structure of all MMPs is made up of five domains: a signal peptide, which directs MMPs to the plasma membrane; a zinc-containing catalytic domain; a prodomain that monopolizes the zinc active site, blocking the catalytic site from substrates; a hemopexin domain that mediates substrate interaction; and a hinge region, which links the catalytic and the hemopexin domain.

Membrane type 1 matrix metalloproteinase.

Of the membrane type MMPs, membrane type 1 matrix metalloproteinase (MT1-

MMP or MMP-14), a transmembrane proteinase that degrades interstitial collagen as well as a number of other substrates, has been studied the most extensively (Barbolina et al. 2007). MT1-MMP expression is important during development; MT1-MMP knockout mice develop dwarfism, display weakened bone formation, soft tissue degeneration, and inadequate collagen turnover

(Holmbeck et al. 1999). MT1-MMP is overexpressed in a number of human carcinomas, has been shown to enhance tumor cell growth, and has been implicated in epithelial to mesenchymal transition in prostate and squamous cell carcinoma cells (Doi et al. 2011; Yao et al. 2013). ECM degradation by MT1-

MMP activity has been directly associated with a number of pro-metastatic events, including proliferation, migration, chemoresistance and invasion in most cancer types (Egeblad and Werb 2002; Sounni et al. 2002; Doi et al. 2011).

MT1-MMP structure.

In addition to the five domains common to all MMPs, MT1-MMP also has a transmembrane domain, consisting of a stretch of hydrophobic amino acids that traverse the cell membrane, and a short (20 amino acid) cytoplasmic tail (Fig.

1.4, A) (Poincloux et al. 2009). MT1-MMP is activated intracellularly via a well

characterized process: the inactive zymogen is activated by pro-protein convertases, which cleave at the highly conserved furin cleavage site located between the propeptide and the catalytic domain (Barbolina and Stack 2008;

Vargová et al. 2012). Activated MT1-MMP is then transported to the cell surface, where it undergoes autocatalytic cleavage, forming soluble, inactive fragments

(Vargová et al. 2012). Localization to the plasma membrane allows MT1-MMP to modify its immediate pericellular environment, but the mechanism regulating this process has yet to be elucidated.

MT1-MMP activity at the cell surface is regulated through phosphorylation of its cytoplasmic tail, a required element for endocytosis, and numerous studies have demonstrated cytoplasmic tail involvement in intracellular signaling, intermolecular interactions, and cellular responses (Rozanov et al. 2004a; Wu et al. 2007; Moss et al. 2009b; Longuespée et al. 2012; Pytliak et al. 2012). Three potential phosphorylation sites have been identified: Threonine567 (Thr567),

Tyrosine573, and Serine577. Moss et al. have shown that MT1-MMP is phosphorylated on Thr567 and have further evidence to suggest that this modification regulates key characteristics of metastasis, including invasion and

3D growth (Fig. 1.4, B) (Moss et al. 2009b).

Though the role of MT1-MMP has been extensively studied throughout the cancer research community, the emphasis has nearly always been on assessment of functions prompted by catalytic activity. Prior to the late 2000s, there was little to no evidence available to support non-proteolytic functions of

MT1-MMP in cancer processes (though studies of MT1-MMP in immunological functions suggests that it may be able to regulate macrophage mobility independently of catalytic function, as well as associate with complement components in a receptor/ligand interaction) (Rozanov et al. 2004b; Sakamoto and Seiki 2009). There is now, however, evidence to suggest that MT1-MMP may be able to function in a non-proteolytic manner in cancer processes previously thought to be a consequence of catalytic engagement, specifically cell migration and invasion (Strongin 2010). Sounni et al. and D’Alessio el et al. have shown that MT1-MMP, when in complex with its catalytic-activity inhibitor, tissue inhibitor of metalloproteinase, induces the MEK/ERK signaling cascade to stimulate cell movement (D’Alessio et al. 2008; Sounni et al. 2010). More recently, MT1-MMP has been shown to promote mammary epithelial cell invasion of collagen type I through direct association of its transmembrane domain with β1 integrin (Mori et al. 2013). Clearly, the non-proteolytic activities of

MT1-MMP are just beginning to be elucidated. Understanding the mechanisms behind these activities and their potential additive effect to current proteolytic activity will be key for future therapeutics aimed at inhibiting MMPs.

MT1-MMP in EOC.

MT1-MMP is not detected in normal ovarian surface epithelium nor in benign

ovarian tumors, but is widely expressed in ovarian carcinomas of all histologic

types, with enhanced expression in metastases relative to primary tumors

(Sakata et al. 2000; Barbolina et al. 2007; Barbolina and Stack 2008). In EOC,

MT1-MMP expression has been shown to promote cell migration, extracellular

matrix invasion, angiogenesis, MCA formation, and expansive growth within 3D

collagen matrices; moreover, high expression of MT1-MMP has been linked with

decreased survival in EOC patients (Kamat et al. 2006; Barbolina et al. 2007;

Moss et al. 2009b; Kaimal et al. 2013).

MT1-MMP activity at the EOC cell surface is implicated in a number of pro-

metastatic events, including induction of cell-matrix detachment, promotion of

MCA formation, invasion of the collagen-rich sub-mesothelial matrix, and

metastatic shed (Ellerbroek et al. 2001; Moss et al. 2009a; Kaimal et al. 2013).

Moss et al. evaluated the potential effect of Thr567 phosphorylation on EOC cell

behavior (Moss et al. 2009b). To this end, MT1-MMP mutants at residue 567

were generated, replacing threonine with either glutamic acid (T567E; to mimic

constitutive phosphorylation) or alanine (T567A; to represent a phospho-

defective), transfected into an EOC cell line lacking endogenous MT1-MMP, and

assessed for MCA formation and 3D collagen growth (Moss et al. 2009b).

Constitutive phosphorylation at Thr567 resulted in larger (diameter) MCAs with multiple large invasive foci within the 3D collagen matrix, suggesting that the phosphorylation status of MT1-MMP cytoplasmic residue Thr567 regulates invasive growth within the sub-mesothelial matrix; however, the kinase that catalyzes this event has not been identified.

IV. INTEGRIN LINKED KINASE

Protein kinases play key roles in a variety of signal transduction pathways and

pathologic processes, where aberrant kinase activities have been found to

correlate with increases in cell proliferation and resistance to apoptosis (Kyriakis

2014). Integrin linked kinase (ILK), a serine/threonine protein kinase that is

expressed in virtually all mammalian cell types, mediates a number of cellular

functions by influencing intra- and extracellular processes. Required for embryonic development, ILK regulates integrin-mediated cell adhesion and provides a molecular scaffold for the assembly of signaling molecules, physically linking ECM proteins and growth factors via integrins and receptor tyrosine kinases to the actin cytoskeleton (McDonald et al. 2008; Widmaier et al. 2012).

Cellular processes facilitated by ILK activity include: integrin relocation to focal

adhesion sites, increased invasion of ECM, decreased cell-cell adhesion, and the

suppression of apoptosis and anoikis, (Hannigan et al. 2005; Widmaier et al.

2012).

ILK structure.

A yeast-two hybrid screen in the late nineties identified ILK, a protein capable of

interacting with the cytoplasmic domain of β1 and β3 integrins (Hannigan et al.

1996; Persad and Dedhar 2003). Structurally, ILK is 452 amino acids in length and is comprised of three conserved functional domains: the N-terminal/ankyrin repeat domain, a central pleckstrin homology-like domain, and the C-terminal kinase catalytic domain (Fig. 1.5) (Hannigan et al. 2005). The ankyrin repeats in the N-terminus are involved in protein-protein interactions, linking a range of

adaptor and signaling molecules, including PINCH (particularly interesting new

cysteine-histidine-rich protein) and ILK-associated protein (ILKAP), localizing the molecule to focal adhesions and regulating its function (Marotta et al. 2003;

Kovalevich et al. 2013). The pleckstrin homology domain participates in the regulation of the kinase activity, binding a lipid product of phosphatidylinositol 3- kinase PI3K, while the catalytic domain mediates ILK interaction via its integrin β cytoplasmic tail binding site (Dedhar 2000; Hannigan et al. 2005).

ILK activity is positively regulated in a PI3K-dependent manner by both cell-ECM interactions and growth factor receptors (Dedhar 2000; Cruet-Hennequart et al.

2003; Hannigan et al. 2005). Once activated, ILK is able to directly phosphorylate several key signaling molecules, including protein kinase B (Akt) at Ser473 and glycogen synthase kinase 3β (GSK3β), to affect cell survival, cell cycle, cell adhesion and ECM modification (Wu and Dedhar 2001). Phosphorylation of

GSK3β results in its inactivation, facilitating activation of several downstream proteins, including cyclin D1 and β-catenin, molecules key to cell cycle

progression and nuclear translocation (Dedhar 2000; Hannigan et al. 2005;

Maydan et al. 2010). Ser473 phosphorylation of Akt promotes cell survival;

interestingly, Akt can also phosphorylate GSK3β.

ILK activity is negatively regulated by tumor suppressor PTEN and ILK-

associated protein (ILKAP), a serine/threonine phosphatase (Leung-Hagesteijn

et al. 2001; Kumar et al. 2004). Inhibition of ILK has been shown to induce

apoptosis and cell cycle arrest, making ILK an attractive therapeutic target for

cancer treatment (Eke et al. 2009). Pharmacological inhibition of ILK in tumor

cells utilizes ATP competitive small molecule inhibitors. QLT0267, one such

molecule, has been shown to delay tumor growth and induce apoptosis in breast

cancer and human squamous cell carcinomas of the head and neck (Younes et

al. 2007; Kalra et al. 2009). In the same study, Kalra et al. also assessed the effect of QLT0267 when delivered in combination with a panel of common chemotherapies, revealing a synergistic interaction with docetaxel that significantly reduced tumor growth and extended survival in vivo compared to single agent controls (Kalra et al. 2009).

ILK controversy.

ILK function has been the subject of a great deal of controversy. Several researchers have debated ILK’s status as a bona fide kinase, suggesting that the molecule is actually a pseudokinase (Boudeau et al. 2006; Wickström et al. 2010;

Hannigan et al. 2011; Qin and Wu 2012; Ghatak et al. 2013). Much of this debate centers around the fact that the ILK catalytic domain contains atypical pseudoactive sites thought to be unable to catalyze phosphorylation. As a consequence, some have suggested suggesting that its functions are mediated exclusively through protein-protein interactions (Wickström et al. 2010; Hannigan et al. 2011). Functional analyses of catalytic domains have identified motifs critical for kinase activity, which includes an invariant lysine residue (necessary for ATP binding) and three highly conserved amino-acid triplets: HRD, DFG, and

APE (Hanks and Hunter 1995; Taylor et al. 2012; Kyriakis 2014). Of these, ILK only contains the invariant lysine and the APE triplet, substituting PRH and

SMAD, two non-conserved amino acid motifs, in place of the HRD and DFG triplets, respectively (Fig. 1.5) (Hannigan et al. 1996). Further support for

‘pseudo’ labeling, is the ability of ILK to mediate a number of biological events via

binding to both PINCH and actin-binding adaptor proteins α–parvin (actopaxin) or

β–parvin (affixin). The resulting ternary ILK-centric PINCH-ILK-parvin (PIP) complex is localized to cell-ECM adhesion sites, where it serves as a signaling mediator that transduces signals for downstream effectors to control cytoskeleton organization, spreading, motility, proliferation, and survival (Wu 2004; Qin and

Wu 2012; Honda et al. 2013; Kovalevich et al. 2013). These arguments, however, have proven faulty as similar deficits of canonical residues in catalytic domains have been described in several other kinases, all of which were, at one point, thought to be pseudokinases but have now been shown to possess protein kinase activity and are currently classified as atypical kinases (Manning et al.

2002; Taylor and Kornev 2010; Hannigan et al. 2011). Moreover, as the exact mechanism through which the PIP complex exerts its effects has yet to be elucidated, the assumption that its protein interactions exclude ILK serine/threonine kinase activity is fallacious (Wu 2004).

ILK in EOC.

EOC metastatic implantation is initiated via intraperitoneal adhesive events

resulting from cell-cell and cell-matrix interaction to mesothelial cells and tissues.

Numerous studies support a role for β1 integrin in these processes. As described

above, ILK is activated by β1 integrin adhesion, resulting in phosphorylation of

cytoplasmic substrates that regulate key cellular processes. Overexpression and

constitutive ILK activation promotes tumor formation in transgenic mice and

provokes oncogenic cell transformation into anchorage-independent, highly

migratory and invasive cells, a hallmark of EOC (McDonald et al. 2008; Widmaier

et al. 2012). In EOC, ILK expression is enhanced compared to benign ovarian

tumors and normal ovarian epithelium, and a direct relationship between ILK

expression and ovarian tumor grade has been shown (Ahmed et al. 2003). ILK may regulate EOC growth; there is evidence to suggest that silencing of ILK

(shRNA) in human ovarian cancer cell line SKOV3 induces up-regulation of pro- apoptotic bax expression and down-regulation of anti-apoptotic bcl-2 expression

(Liu et al. 2012a). Additionally, ILK gene silencing has been shown to suppress in vivo tumorigenesis of EOC cells; HO-8910 cells transfected with ILK antisense oligonucleotide and injected into nude mice revealed significantly delayed tumor formation and suppressed tumor growth, but the mechanism underlying these functions has yet to be elucidated (Li et al. 2013).

V. MUCIN 16/ CANCER ANTIGEN 125

Mucins.

Mucins are high molecular weight glycoproteins, generally found in the cell

membranes of human epithelial tissues that exist in relatively harsh environments

– environments that are exposed to a variety of microorganisms, toxins,

proteases, and microenvironmental changes, including pH, ionic concentration,

hydration, and oxygenation – where they play a key role in homeostatic maintenance (Kufe 2009; Zaretsky and Wreschner 2013). Nearly twenty mucin

genes have been identified. These are designated by numbers (MUC1, MUC2,

etc) and further classified into two groups: secreted/gel-forming and cell surface

(membrane-bound) (Kufe 2009). Secreted mucins form a mucous gel, creating a specific physical barrier to protect the epithelial cells lining the respiratory and gastrointestinal tracts, as well as those forming the ductal surfaces for organs, including the liver, breast, pancreas, and kidney. Membrane-bound mucins span

the cell membrane to conduct signals in response to external stimuli for a variety

of cellular responses, including proliferation, growth, differentiation, and

apoptosis (Kufe 2009; Zaretsky and Wreschner 2013). Dysregulation of mucin

production has been characterized in several adenocarcinomas, including

cancers of the lung, pancreas, colon, breast, and ovary (Niv 2000; Jeon et al.

2010; Rachagani et al. 2012).

Mucin general structure.

The general structure and biochemical composition of mucins provides protection

for the cell surface and serves to promote cell survival in the variable conditions

described above. Mucins are comprised of variable tandem-repeat structures,

sequences of amino acids that repeat. These structures are teeming with

prolines, threonines, and serines that are extensively glycosylated through O-

and N-linked oligosaccharides (Moniaux et al. 2001; Hattrup and Gendler 2008).

Molecular composition, tandem-repeat sequences, and higher order structures vary specifically depending on mucin type and contribute to specialized functions

(Chaturvedi et al. 2008). Transmembrane mucins MUC1, MUC4, and MUC16, for example, have a single membrane-spanning region, a cytoplasmic tail, and an extensive extracellular domain, which features a variable pattern of tandem- repeat domains, epidermal growth factor (EGF)-like domains, or sperm protein, enterokinase, and agrin (SEA) domains (Moniaux et al. 2001; Hollingsworth and

Swanson 2004; Hattrup and Gendler 2008). The arrangement of these domains form rod-like structures extending 200-2000 nm beyond the cell surface, giving these mucins their characteristic bottle-brush structure (Hattrup and Gendler

2008). All mucins consist of large protein cores, and the molecular weight of each

mucin is dependent upon the extent of glycosylation (Hattrup and Gendler 2008).

When highly expressed, transmembrane mucins have been shown to reduce drug efficacy, promote cellular growth, and protect cells from apoptosis; however, size, glycosylation, and projection from the cell surface can play important roles in specific function (Moniaux et al. 2001; Hollingsworth and Swanson 2004; Ren et al. 2004).

Mucins in EOC.

Mucins are produced by all epithelial cells, though variations in the type, amount,

and expression is dependent upon cell and tissue type. The metastatic process

of epithelial cancers commonly exploits transmembrane mucins for their pro-

growth and pro-survival functions. The OSE expresses a mixed epithelo-

mesenchymal phenotype and is the only portion of the ovary known to express

mucins, with MUC1 being the only mucin found at a detectable level (Chauhan et

al. 2009). Overexpression of MUC1, MUC4, and MUC16 has been observed in

EOC, with benign and borderline carcinomas expressing less mucins than

malignant ovarian tumors (Giuntoli et al. 1998; Chauhan et al. 2006).

MUC16.

Similar to other transmembrane mucins, MUC16 has been extensively studied in

various epithelial tissues; reported functional roles include barrier protection of

the ocular surface, modulation of blastocyst binding to the uterine epithelium, and facilitation of T-cell apoptosis (Dogru et al. 2008; Zaretsky and Wreschner 2013;

Dharmaraj et al. 2014). Dysregulated expression of MUC16 has been implicated in a number of disease processes, including metastases of the pancreas, breast, and ovary (Nagata et al. 2007; Thériault et al. 2011; Lakshmanan et al. 2012). A large portion of MUC16 investigations have focused on its function in EOC metastasis, where its expression has been well documented; this is discussed in detail later in this chapter and is directly investigated in Chapter 3.

MUC16 structure.

Though acknowledged as a high molecular weight glycoprotein in the late eighties, the actual structure of MUC16 was not able to be confirmed and successfully cloned until the early 2000s (Davis et al. 1986; Lloyd et al. 1997; Yin and Lloyd 2001). To date, MUC16 is the largest mucin identified (Weiland et al.

2012). Similar to other transmembrane mucins, the MUC16 ectodomain is rich in serine, threonine, and proline, and heavily glycosylated with both O- and N-linked oligosaccharides (Weiland et al. 2012). Unique to MUC16 is a high leucine content, an extremely lengthy tandem repeat domain containing nearly three times as many amino acids as other transmembrane mucins, and multiple SEA modules within the protein sequence (other transmembrane mucins have a single SEA domain) (Fig. 1.6) (Moniaux et al. 2001; Hattrup and Gendler 2008;

Gipson et al. 2014). These repetitive sites, along with the extracellular portion of the 284-amino acid carboxy-terminal domain, have been suggested as potential sites for proteolytic cleavage (Hattrup and Gendler 2008; Weiland et al. 2012).

MUC16 also contains a short (35-amino acid) cytoplasmic tail with several

potential phosphorylation sites (Bouanene and Miled 2010). Phosphorylation of

the cytoplasmic tail has been reported in oncogenic processes and may be

responsible for the release of soluble proteolytic fragments of MUC16 into the

extracellular space, however, the exact interacting partner for the cleaved moiety

and phosphorylation site on the parent molecule have yet to be elucidated.

(Rachagani et al. 2009; Gipson et al. 2014). Additionally, given the proteolytic

activity of other molecules in the immediate pericelluar environment, proteolysis

of a membrane bound molecule could modulate release of soluble MUC16 via an

unknown mechanism.

MUC16/CA-125 in EOC.

MUC16 is highly expressed on the ovarian tumor cell surface but not on the

normal OSE. It has been shown to interact directly with mesothelin, a protein

expressed on the mesothelial cells that line the peritoneal cavity (Gubbels et al.

2006). MUC16 has also been shown to associate with receptors expressed on

the surface of natural killer cells, rendering these cells inactive and thus enabling

EOC cells to evade the immune response (Patankar et al. 2005). These

interactions may contribute to tumor cell growth as well as the adhesion and peritoneal spread characteristic of EOC metastasis (Scholler and Urban 2007;

Thériault et al. 2011). In addition to pro-metastatic implications, expression of

MUC16 may facilitate EOC resistance to drug treatment. When down-regulated in the MUC16-expressing ovarian tumor cell line OVCAR-3, knockdown cells were shown to be more sensitive to treatments with chemotherapeutics cisplatin and doxorubicin; conversely, overexpression in MUC16-negative ovarian tumor cell line SKOV3 increased resistance to these drugs (Boivin et al. 2009). Despite this evidence, no correlation between chemotherapy resistance and MUC16 expression in EOC patients has been reported.

Clinically, MUC16 is referred to as cancer antigen 125 (CA-125). In an attempt to develop antibodies reactive to ovarian cancer, Bast et al. developed ovarian cancer 125 (OC125), a murine monoclonal antibody that is able to react with a number of EOC cell lines and cryopreserved tumor tissue while remaining unreactive with nonmalignant ovarian tissues (Bast et al. 1981). OC125 was found to bind to a specific antigen expressed by nearly 80% of EOC; this antigen was defined as CA-125. A radioimmunoassay for CA-125 detection in serum and body fluids quickly followed, revealing an accumulation of CA-125 in EOC patients and demonstrating a significant correlation between antigen expression levels and the regression, stability or progression of EOC (Bast et al. 1983).

Since this discovery, CA-125 has been extensively investigated for use as a biomarker of EOC and has been implemented for use not only as a screening test for early EOC detection, but also as an assessment of therapeutic response to EOC treatment.

While considered to be the standard molecular marker of EOC malignancy for the past three decades, the use of CA-125 as a biomarker is not ideal. It is generally accepted that a serum concentration of CA-125 greater than 35U/ml is indicative of potential EOC malignancies, but elevated serum concentrations have also been noted in patients with non-gynecological cancers as well as in individuals with non-cancerous conditions such as menstruation and pregnancy

(Rosen et al. 2005; Chen et al. 2013b). Moreover, nearly a quarter of all EOC patients have normal levels of CA-125 when their malignancy is diagnosed

(Gupta and Lis 2009). Despite its poor as a screening test, CA-125 is still the most reliable source to gauge therapeutic response (Schmidt 2011). Current studies have recommended that CA-125 be measured repeatedly over time, with a general assumption that reduced levels indicate a favorable response to treatment (Gupta et al. 2010; Menczer 2013). Though there has been significant effort towards the development of novel biomarkers applicable to the diagnosis and management of EOC, to date, none have proven effective for use as an

EOC screening test and further studies are needed to elucidate an appropriate marker.

VI. PROJECT RATIONALE

Despite their epithelial morphology, OSE cells display mesenchymal

characteristics, including the production of mesenchymal metalloproteases and

cadherins and the lack of epithelial cadherins (Auersperg et al. 1999; Strauss et

al. 2011). This phenotypic plasticity further complicates identification of the

specific molecular events occurring during EOC metastatic progression, which

has historically been presumed to arise as a consequence of malignant

transformation and proliferation of epithelial cells from the OSE (Fathalla 1971).

Several recent studies have provided an alternative hypothesis, suggesting that

the initiating epithelial cells involved in formation of high-grade serous ovarian

cancer are derived from non-ovarian sources, such as the endocervix and fallopian tubes (Dubeau 2008; Berns and Bowtell 2012; Hillier 2012; Kim et al.

2012b). Regardless of the site of origin, EOC metastasis involves detachment of

epithelial cells from the primary tumor and dissemination into the peritoneal

cavity as single cells and MCAs, which then adhere to and anchor in the

mesothelial cell monolayer that lines the peritoneal cavity, to subsequently

proliferate within the interstitial collagen-rich sub-mesothelial matrix and establish secondary lesions (Hudson et al. 2008; Barbolina et al. 2009). Elucidating the

early molecular mechanisms involved in this metastatic process, specifically the adhesion of EOC cells to mesothelial cells and penetration of the associated sub- mesothelial extracellular matrix, is essential to the development of future therapeutic agents.

As stated earlier, MT1-MMP activity has been directly implicated in both the invasion of the sub-mesothelial collagen I matrix, and in the shedding of metastatic MCAs, but the molecular mechanisms behind these events are not completely understood (Moss et al. 2009a; Doi et al. 2011). Considering the well established role of MT1-MMP in the EOC metastatic process, identification of the molecules contributing to these pro-metastatic phenotypes is critical to future understanding of EOC metastatic spread.

The experiments detailed within this body of work were designed to investigate the initial adhesive and invasive events of ovarian cancer metastasis, as associated with MT1-MMP proteolytic activity. Specifically, the in vitro relationship between MT1-MMP and a potential phosphorylator, ILK, on adhesion and invasion is assessed in Chapter 2 while the effect of MT1-MMP activity on ovarian tumor cell ectodomain shedding is evaluated in Chapter 3.

1.1 MULTI-STEP PROCESS OF OVARIAN CANCER SPREAD

Figure 1.1 MULTI-STEP PROCESS OF OVARIAN CANCER SPREAD

Figure 1.1: Multi-step Process of Ovarian Cancer Spread. (A) Primary ovarian tumor formed by malignant transformation and proliferation of epithelial cells. (B) Cells are shed into the peritoneal cavity as both single cells and MCAs. (C) MCAs adhere to and disaggregate on human mesothelial cell monolayers, forming invasive foci. (D) Cells invade the collagen-rich sub-mesothelial matrix where they proliferate to form secondary lesions.

1.2 INTEGRIN HETERODIMER

Figure 2 1.2 INTEGRIN HETERODIMER

Figure 1.2: Integrin Heterodimer. Basic structure of an integrin at the cell surface, displaying α and β subunits.

1.3 CLASSICAL CADHERIN

Figure 31.3 CLASSICAL CADHERIN

Figure 1.3: Classical Cadherin. Basic structure of a type I classical cadherin at the cell surface.

1.4 MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE

Figure 41.4 MEMBRANE TYPE 1 MATRIX METALLOPROTEINASE

Figure 1.4: Membrane Type 1 Matrix Metalloproteinase. (A) Domain structure of MT1-MMP. MT1-MMP has a transmembrane domain, consisting of a stretch of hydrophobic amino acids that traverse the cell membrane and a 20-amino acid long cytoplasmic tail. (B) Schematic showing MT1-MMP cytoplasmic tail sequence and Thr567 phosphorylation site.

1.5 INTEGRIN LINKED KINASE

Figure 51.5 INTEGRIN LINKED KINASE

1.5: Integrin Linked Kinase. Domain structure of ILK. ILK is comprised of three highly conserved functional domains: an N-terminal ankyrin repeat domain, a central pleckstrin homology-like domain (blue), and the C-terminal kinase catalytic domain. Boxes highlight three amino acid triplets: APE, one third of the trifecta of highly conserved motifs common to most kinase catalytic domains and critical to kinase activity (others are HRD and DFG); and PRH and SMAD, non- conserved motifs unique to the ILK catalytic domain that align with the conserved HRD and DFG triplets found in typical kinase domains. Also highlighted (red) is an invariable lysine residue required for ATP binding.

1.6 MUCIN16

Figure 61.6 MUCIN16

1.6: Mucin16. Diagram of MUC16 molecular domains. MUC16 has a 35-amino acid long cytoplasmic tail, a transmembrane domain and a highly glycosylated extracellular domain laced with tandem repeat and SEA domains.

2. ILK ALTERS MT1-MMP-DEPENDENT ACTIVITIES IN

EARLY METASTATIC OVARIAN CANCER EVENTS

I. RATIONALE

As discussed earlier, the EOC metastatic process involves epithelial cells that detach from the primary tumor and are then shed into the peritoneal cavity as single cells and MCAs, which adhere intraperitoneally, and undergo localized invasion into the interstitial collagen-rich sub-mesothelial matrix, where they proliferate to anchor secondary lesions. The exact mechanism that controls the transition from detached cells to peritoneally anchored metastatic lesions is still unknown. β1 integrin activation is a key event in ovarian carcinoma metastatic dissemination and regulates expression of several gene products involved in metastasis, including MT1-MMP (Ellerbroek et al. 2001). Previous studies on the function of MT1-MMP cytoplasmic tail phosphorylation have uncovered a role for

MT1-MMP in MCA formation and invasion (Moss et al. 2009a). ILK, a β1 integrin cytoplasmic domain interacting molecule, is activated by integrin-mediated cell

adhesion (Lin et al. 2007). ILK activation has been shown to regulate several biological processes that suppress anoikis, a key event in ovarian cancer

metastasis. ILK-associated protein (ILKAP) negatively regulates ILK through

direct association, forming cytoplasmic complexes (Kumar et al. 2004).

Proteomic analysis of MT1-MMP interacting proteins have identified ILKAP as a

binding partner, which suggests that ILK may interact with MT1-MMP in the

metastatic process.

In this chapter, the role of ILK activity in early events of ovarian cancer

intraperitoneal metastasis is assessed. The experiments described within

investigate the hypothesis that ILK associates with MT1-MMP and regulates cell

behavior associated with early events in EOC metastasis.

II. RESULTS

ILK and MT1-MMP are co-expressed in human ovarian tumor tissues and cells.

Expression of ILK directly correlates with the progression of ovarian cancer, with enhanced expression found in high-grade human tumors compared to weak staining found in low-grade human tumors (Ahmed et al. 2003). To assess co- expression of ILK and MT1-MMP in human ovarian tumor tissues, serial sections of a tumor tissue microarray consisting of 72 cores of adenocarcinoma were subjected to IHC analysis for either ILK (Fig. 2.1, A-C) or MT1-MMP (Fig. 2.1, D-

F); representative examples are shown. From this TMA, a panel of 17 cores stained positively for ILK were also assessed for expression and localization of

MT1-MMP. Of these ILK positive cores, 52.9% also stained positively for MT1-

MMP. Real time quantitative PCR analysis for ILK and MT1-MMP expression in widely used EOC cells revealed expression of both ILK and MT1-MMP, with EOC cell line DOV13 revealing high endogenous MT1-MMP (Fig. 2.1, G).

Immunofluorescent studies revealed co-localization of ILK and MT1-MMP in

DOV13 grown atop a collagen type I matrix (Fig. 2.2).

siRNA knockdown of ILK expression in ovarian cancer cells.

To examine ILK contribution to metastatic events modulated by MT1-MMP, the original strategy for this project necessitated the overexpression and specific knockdown of ILK expression in EOC cell line DOV13, which expresses high endogenous MT1-MMP (Fig 2.1, G). Multiple attempts using a variety of methods

were used to overexpress ILK in MT1-MMP-expressing cell lines. While working closely with two post-doctoral employees in the laboratory as well as colleagues at the University of Notre Dame, several transfection methods were pursued to no avail. These are briefly described below, in the order they were attempted.

Initially, liposomal transfection using pcDNA 3.1 mammalian expression vector was attempted. The numerous attempts to optimize lipid-mediated delivery

included employment of different transfection reagents (including FuGENE6,

FuGENE HD, Lipofectamine 2000, and X-tremeGENE) at varying concentrations,

for time periods ranging from 12-72 hours. Modifications in cell confluence, cell

plating, and timing of transfection application were also explored. Spinfection

(where cells plated in transfection cocktail were centrifuged for a short period of

time) and serial transfections (where cells that have been incubated with

transfection cocktail for a given time period (12-72 hours) were subjected to a

cycle in which the transfection cocktail was removed, the cells were washed with

warmed antibiotic-free media, and fresh transfection cocktail was applied) were also included in optimization experiments.

Once lipid based transfection attempts were exhausted, electroporation, a transfection technology based on the momentary creation of small pores in cell membranes by applying an electrical pulse, was assessed for suitability.

Nucleofection™, a type of electroporation designed specifically for use in mammalian cells considered difficult or impossible to transfect (primary cells, for example) was utilized. Similar to traditional electroporation, the Nucleofector™ technology used electrical pulses to enable nucleic acid substrate delivery through the nuclear membrane and into the nucleus of the target cell. Cells were subjected to electrical pulses through various programs pre-designed by

Amaxa™ Nucleofector™ technology proprietary protocols. Cell conditions

including density, incubation medium, and passage number were also varied in

accordance with Amaxa™ programs for experimental optimization.

As a final attempt, retroviral production was pursued. Production consisted of the

cloning of the transgene into an infections plasmid via in vivo recombination in

bacteria, the rescue and propagation of the vector in complementary cells, and

the purification of the vector for introduction into the desired cells. While the

production process is simple to follow, according to Principles of Retroviral

Vector Design, efficient gene integration depends on a number of viral elements including: the promoter in the viral genome; the viral packaging signal (to direct incorporation of vector RNA); the signals prompting reverse transcription (a transfer RNA-binding site and polypurine tract for initiation of first- and second- strand DNA synthesis, and a repeated region at both ends of viral RNA required for transfer of DNA synthesis between templates; and short, partially inverted repeats located at the termini of the long terminal repeats required for integration

(Coffin et al. 1997). Successful incorporation of these elements, as well as ensuring a high titer for the resulting virus, can require a significant time investment. Over a year was devoted to the design and creation of an ILK retroviral system, but the overexpression of ILK in MT1-MMP expressing cell lines was unsuccessful.

To continue the investigation, an alternative strategy was devised, shifting the focus onto silencing rather than overexpression. As such, knockdown of endogenous ILK using siRNA (expression) and small molecule inhibitor QLT0267

(activity) was completed and the effects assessed.

Real time quantitative PCR analysis of ILK revealed significant suppression of

ILK in DOV13 cells transfected with ILK siRNA (DOV-ILK-KD) compared to non- targeted siRNA control (NTC) (24 hours, p=0.050), with maximal suppression achieved at 48 hours post transfection (p=0.050) (Fig. 2.3, A). Down-regulation was further confirmed via western blot analysis (as described in the Experimental

Methods section of the Materials and Methodology chapter) of cells cultured on a collagen I matrix (Fig. 2.3, B). Visual comparison of DOV-ILK-KD vs NTC suggest reduced expression of pAktSer473 when ILK is silenced (Fig. 2.3, B).

Densitometric analyses of three independent siRNA-treated cell cultures showed

that ILK kinase activity was decreased by an average of 82.4% (data not shown).

Comparatively, in experiments where DOV13 cells were treated with 25µM

QLT0267 (DOV+QLT), a small molecule inhibitor targeting ILK, a trend towards

decreased expression of pAktSer473 was visually observed, but the reductions

were not as pronounced as that of the siRNA-treated, nor were they statistically

calculated (Fig. 2.3, C). It should be noted that non-transfected DOV13 and NTC

behaved identically throughout these studies (statistically non-significant

difference), and were thusly utilized interchangeably in some aspects.

Silencing of ILK affects MT1-MMP regulated MCA formation.

Once shed from the primary tumor, free-floating ovarian cancer cells can form

MCAs. MCA formation has been shown to be associated with MT1-MMP (Moss

et al. 2009a). To assess the effect of ILK suppression on this phenomena, MCAs were generated using a modification of the hanging drop method (as described in the Experimental Methods section of Materials and Methodology), where a small quantity of DOV-ILK-KD cell suspension was gently pipetted onto the underside of a flat tissue culture lid, suspended over sterile liquid, and incubated. At 12 hours, both NTC and DOV-ILK-KD cells formed MCAs (Fig. 2.4, A-D); however, while NTC formed smooth, round and compact MCAs (Fig 2.4, A, B), DOV-ILK-

KD MCAs remained rough and irregularly shaped (Fig. 2.4, C, D). This observation remained for the lifespan of the MCA (up to 96 hours, data not shown).

MCA size/shape is correlated with MT1-MMP activity; Moss et al. have evidence to suggest that phosphorylation of MT1-MMP results in MCAs that are larger, ellipsoid in shape, and that tend to form long projections, possibly for invasive growth, while MCAs formed when MT1-MMP cannot be phosphorylated, are more spherical and lack invasive foci (Moss et al. 2009a). In order to make a quantitative comparison of MCA formation during ILK suppression, the length and the width of each MCA was measured using ImageJ (Fig 2.4, E), and the averaged ratio (L/W) was calculated (Fig 2.4, F). Length/width ratios confer shape; a ratio equal to 1 indicates a spherical shape, while a ratio greater than

one is an ellipsoidal shape. NTC formed MCAs that were spherical (1.05mm),

while DOV-ILK-KD formed MCAs that were more ellipsoidal (1.22mm).

Down-regulation of ILK alters ovarian tumor adhesion and MT1-MMP dependent invasion.

ILK activity is stimulated by adhesion to the extracellular matrix (Attwell et al.

2003). ILK has been shown to regulate integrin-mediated cell adhesion, cell migration, cytoskeletal reorganization, and cell-ECM interactions, while enzymatic activity of MT1-MMP has been well established as the prime determinant of a collagen-invasive phenotype in ovarian carcinomas (Ellerbroek et al. 1999; Dedhar 2000; Fishman et al. 2001; Kumar et al. 2004; Barbolina et al. 2007; Lin et al. 2007; Barbolina and Stack 2008; Eke et al. 2009; Moss et al.

2009a). Collagen type I, the most abundant matrix molecule in the sub- mesothelial stroma, is the preferred substrate for ovarian cell attachment during peritoneal metastasis (Moser et al. 1996). As such, the role of ILK in the

attachment and invasion of ovarian cancer cells to collagen type I was assessed.

Assays utilized DOV13, DOV+QLT, NTC, and DOV-ILK-KD.

To evaluate adhesion to collagen, cell lines were separately incubated atop a plated 10µg/ml collagen type I matrix for 30 minutes, washed to remove non- adherent cells, and adherent cells were enumerated. Control cells (DOV13 and

NTC) adhered robustly in a manner similar to each other (Fig. 2.5, A). Adhesion was significantly reduced in both DOV-ILK-KD and DOV+QLT compared to each respective control (DOV-ILK-KD vs NTC and DOV+QLT vs DOV13, p=0.050 for both) (Fig. 2.5, A).

For assessment of collagen invasive potential, cells were incubated for 24 hours in a transwell chamber containing a collagen type I-coated porous membrane.

Cells migrated to the underside of the collagen coated membrane were considered invasive and were quantified as described in the Experimental

Method section of Materials and Methodology. Silencing of ILK dramatically reduced invasive ability compared to NTC control (DOV-ILK-KD vs NTC, p=0.050) (Fig. 2.5, B). ILK inhibitor treated cells also displayed reduced invasion when compared to DOV13 control (DOV+QLT vs DOV13, p=0.050) (Fig. 2.5, B).

The effect of ILK silencing on the attachment to and invasion of a live mesothelial matrix was also assessed, as described in the Experimental Methods section of

Materials and Methodology. Briefly, to evaluate attachment, a mesothelial matrix was created, where human mesothelial cells (LP9) were plated atop a collagen type I matrix and grown to confluency. Fluorescently labeled EOC cells were then seeded atop this mesothelial monolayer, incubated in a 1:1 mix of LP9 complete

growth media and DOV13 complete growth media for 30 minutes, and washed to remove non-adherent cells. Adherent cells were enumerated as described in the

Experimental Methods section of Materials and Methodology. Similar to results

shown on a collagen type I matrix, knockdown of ILK significantly reduced

ovarian cancer cell adhesion to a confluent mesothelial monolayer (DOV-ILK-KD

vs NTC, p=0.050) (Fig. 2.5, C). A live meso-mimetic culture, wherein an 8µm porous membrane seated in the upper compartment of a transwell insert hosts a scaled 3D collagen type I matrix, upon which an LP9 monolayer is grown to confluence, was employed to assess invasion. For these studies, this 3D culture was plated with the indicated ovarian cancer cells and incubated in a 1:1 mix of complete growth media for 48 hours (as described in the Models section of

Materials and Methodology). Migrated cells on the underside of the pore filter were fixed stained and enumerated (as described in the Experimental Methods section of Materials and Methodology), revealing reduced invasion of both DOV-

ILK-KD and NTC+QLT (DOV-ILK-KD vs NTC and NTC+QLT vs NTC, p=0.050 for both) (Fig 2.5, D).

siRNA-mediated knockdown of ILK does not affect ovarian tumor cell adhesion to mesothelial tissue.

An early stage of metastasis involves adhesion of free-floating EOC cells to the

mesothelium, the single layer of flat cells covering the peritoneal cavity. This

heterotypic cell-cell interaction serves as the initial event in MT1-MMP-facilitated intraperitoneal metastasis; as such, the effect of ILK silencing on ovarian cancer cell attachment to live mesothelial tissue was assessed. In this experiment, peritoneal tissues from female c57bl/6 mice were excised, uniformly trimmed, and immobilized to a silicone coated dish before being seeded with fluorescently labeled ovarian cancer cells. After a 2 hour incubation, tissues are washed robustly, removed from the silicone bed and mounted onto a glass coverslip for analysis. Adhesion was quantified by fluorescence microscopy. Attachment of

DOV-ILK-KD was not statistically different from control (NTC) (p=0.275) (Fig 2.6,

C). Similarly, attachment of inhibitor treated cells did not differ significantly from control (p=0.513) (Fig 2.6, C).

2.1 EXPRESSION OF ILK IN OVARIAN ADENOCARCINOMAS

AND CELLS

Figure 72.1 EXPRESSION OF ILK IN OVARIAN ADENOCARCINOMAS AND CELLS

Figure 2.1: Expression of ILK in Ovarian Adenocarcinomas and Cells. (A-G) Microarrayed cores of ovarian adenocarcinoma were subjected to

immunohistochemical analyses for ILK (A-C) or MT1-MMP (D-F) as described in Materials and Methodology. 20X magnification. Arrows highlight positive staining for the indicated protein. Negative control tissues were processed as described in Materials and Methodology, without exposure to biotinylated secondary antibody. (G) Quantitative real time PCR analysis of ILK and MT1-MMP in ovarian cancer cells. The comparative CT method was used to determine average relative quantitation, as described in Experimental Methods section of Materials and Methodology. Results represent the mean of a minimum of three independent experiments.

2.2 CO-LOCALIZATION OF ILK AND MT1-MMP IN OVARIAN

CANCER CELL LINE DOV13

Figure 82.2 CO-LOCALIZATION OF ILK AND MT1-MMP IN OVARIAN CANCER CELL LINE DOV13

Figure 2.2: Co-localization of ILK and MT1-MMP in Ovarian Cancer Cell Line DOV13. (A-E) Subconfluent monolayers of DOV13 cells were seeded onto a 10µg/ml collagen type I matrix and processed for immunofluorescence. Images are representative of results obtained. The merged image shown (yellow, C) demonstrates co-localization of ILK (red, A) and MT1-MMP (green, B). (D-E) Secondary antibody controls for ILK (mouse, D) and MT1-MMP (rabbit, E).

2.3 siRNA DOWN-REGULATION OF ILK EXPRESSION IN OVARIAN CANCER CELLS

Figure 92.3 siRNA DOWN-REGULATION OF ILK EXPRESSION IN OVARIAN CANCER CELLS

Figure 2.3: SiRNA Down-regulation of ILK Expression in Ovarian Cancer Cells. (A) Quantitative real time PCR analysis of ILK levels as described in

Experimental Methods section of Materials and Methodology. Results are expressed relative to non-targeting control siRNA (NTC) and represent the mean of a minimum of three independent experiments. A Mann-Whitney U test was employed to determine statistical significance between NTC and DOV-ILK-KD within each time point as indicated. Non-transfected DOV13 and NTC behaved identically within each time point; a Mann-Whitney U test revealed no significant difference between non-transfected DOV13 and NTC within each time point (0hr, p=0.479; 24hr, p=0.487; 48hr, p=0.487). (B-C) Western blot analyses as described in Experimental Methods section of Materials and Methodology for expression of ILK downstream mediator pAktSer473. (B) siRNA treated DOV13 (DOV-ILK-KD) (as described in the Models section of Materials and Methodology), compared to NTC. (C) DOV13 cells were treated with 25µM of QLT0267 (DOV+QLT) or comparative volume of DMSO alone (Control), as described in the Models section of Materials and Methodology. Cell lysates were collected and processed for western blot analysis.

2.4 SILENCING OF ILK AFFECTS FORMATION OF MCAs

Figure 102.4 SILENCING OF ILK AFFECTS FORMATION OF MCAs

Figure 2.4: Silencing of ILK Affects Formation of MCAs. MCAs were generated using a modification of the hanging drop method (Kelm et al. 2003) as described in the Experimental Methods section of Materials and Methodology. (A-D) Representative examples of MCA formation at 12 hours for indicated cell lines. Scale bar, 400µm. (E) MCAs were measured for length (mm) across the

longest part of the structure and for width at exactly half of the length line, using ImageJ. (F) NTC and DOV-ILK-KD average L/W ratios ± standard error of the mean. Results are representative of five independent experiments. A Mann- Whitney U test was employed to determine a statistically significance difference between the two groups (NTC vs DOV-ILK-KD, p=0.009).

2.5 SILENCING OF ILK ALTERS TUMOR CELL ADHESION AND INVASION

Figure 112.5 SILENCING OF ILK ALTERS TUMOR CELL ADHESION AND INVASION

Figure 2.5: Silencing of ILK Alters Tumor Cell Adhesion and Invasion. (A) Ovarian cancer cells were allowed to adhere to a dish coated with collagen type I for 30min prior to washing (to remove unbound cells) and quantitation of adherent cells. Results are the percentage of total cells seeded and represent the mean of three independent experiments. Statistical significance was determined using Mann-Whitney U tests for indicated data groups; p-values are displayed accordingly (B) Tumor cells were seeded into a transwell apparatus

containing a collagen type I coated 8µm pore filter. After incubation for 24h, invading cells on the underside of the filter were stained and enumerated. Results are the ratio of migrated cells to total cells seeded, expressed as a percentage. Results represent the mean of three independent experiments. Statistical significance was determined using Mann-Whitney U tests for indicated data groups; p-values are displayed accordingly. (C) Ovarian cancer cell adhesion to a mesothelial cell monolayer. CMFDA-labeled cells were allowed to adhere to a confluent monolayer of LP9 mesothelial cells for 30min prior to washing. Adherent fluorescent cells were enumerated and are presented as a percentage of total cells seeded. Results represent the mean of three independent experiments. Statistical significance was determined using Mann- Whitney U tests for indicated data groups; p-values are displayed accordingly. (D) Ovarian cancer cell invasion through a live mesothelial cell monolayer. CMFDA-labeled tumor cells were incubated for 24h in a transwell chamber containing mesothelial cells plated on a collagen matrix cultured atop a porous membrane. Invaded cells were quantified as described in the Experimental Methods section of Methods and Methodology, and are expressed as a percentage of total cells seeded. Results represent the mean of four independent experiments. A Kruskal-Wallis Test was utilized to find a significant mean difference among all three groups (p=0.007). Mann-Whitney U tests for indicated data groups revealed significant reductions invasion (DOV-ILK-KD vs NTC and NTC+QLT vs NTC, p-values as indicated). Green bars – untransfected DOV13 (DOV13); red bars – DOV13 plus ILK small molecule inhibitor QLT0267 (DOV+QLT); blue bars – non targeting control siRNA (NTC); pink bars – DOV13 treated with ILK siRNA (DOV-ILK-KD).

2.6 SILENCING OF ILK DOES NOT AFFECT CELL-TO- MESOTHELIAL ADHESION IN AN EX VIVO PERITONEAL EXPLANT

Figure 122.6 SILENCING OF ILK DOES NOT AFFECT CELL-TO- MESOTHELIAL ADHESION IN AN EX VIVO PERITONEAL EXPLANT

Figure 2.6: Silencing of ILK Does Not Affect Cell-To-Mesothelial Adhesion in an ex vivo Peritoneal Explant. (A) Schematic of ex-vivo assay (as described in the Experimental Methods section of Materials and Methodology). Ovarian tumor cells were fluorescently labeled with CMFDA immediately prior to experimentation. An excised explant from the peritoneum of a female c57bl/6 mouse was pinned, mesothelium-side up, to a tissue culture dish containing an optically clear silastic resin. Tumor cells were allowed to adhere to mesothelial tissue for 2h prior to washing (to remove unattached cells). (B) Representative

images of fluorescently labeled ovarian tumor cells attached to explant. (C) Quantitation of adhesion to murine peritoneal tissue explant, as described in the Experimental Methods section of Methods and Methodology. Results are the mean result of three independent experiments, expressed as the relative number of cells per area. A Kruskal-Wallis test found no significant mean difference among the three groups (p=0.587). Mann-Whitney U tests for comparisons between groups (DOV-ILK-KD vs NTC and NTC+QLT vs NTC) also failed to illuminate any significant difference (p=0.275 and p=0.513, respectively). Blue – non targeting control siRNA (NTC); red – DOV13 plus ILK small molecule inhibitor QLT0267 (DOV+QLT); pink – DOV13 treated with ILK siRNA (DOV-ILK- KD).

III. DISCUSSION

Understanding of the molecular intricacies in the early events in ovarian cancer

metastasis, namely homotypic and heterotypic cell-cell adhesion and invasion, is key to creating successful future therapeutics. MT1-MMP, an extracellular matrix degrading protease localized to the cell surface, has been implicated in invasion and metastasis of a number of human tumor cells (Tsunezuka et al. 1996; Kim et al. 2007; Poincloux et al. 2009; Yao et al. 2013). MT1-MMP has not been detected normal OSE nor in benign ovarian tumors, but is widely expressed in ovarian carcinomas, where its enzymatic activity is key in promoting the migration to and invasion of collagen matrices (Afzal et al. 1998; Ellerbroek et al.

1999; Sakata et al. 2000; Wu et al. 2007). There is evidence to suggest that

MT1-MMP activity is regulated through phosphorylation of Thr567 on its cytoplasmic tail, but the kinase that catalyzes this event has not been identified

(Radeva 1997; Scott and Olson 2007).

Expression and activity of ILK, a serine/threonine kinase activated by integrin- mediated cell adhesion, is increased in several different cancer types (Hannigan

et al. 2005). Overexpression and constitutive ILK activation promotes tumor formation in transgenic mice and also provokes oncogenic cell transformation into anchorage-independent, highly migratory and invasive cells (Li et al. 2013).

In ovarian cancer, studies have shown up-regulated ILK expression compared to

benign tumors and normal ovarian epithelium; additionally, a direct relationship

between ILK expression and ovarian tumor grade has been shown (Ahmed et al.

2003). EOC metastatic implantation is initiated via intraperitoneal adhesive

events resulting from cell-cell and cell-matrix interaction to mesothelial cells and

tissues. β1 integrin-mediated adhesion has been implicated as a key event in this process and ILK is activated by β1 integrin adhesion, resulting in phosphorylation of cytoplasmic substrates that regulate key cellular processes. This chapter investigates the effect of ILK activity on MT1-MMP mediated events.

Initial experiments to assess expression of ILK in ovarian cancer tumor tissue

indicated moderate to strong expression. The staining of tumor tissue microarray

serial sections for either MT1-MMP or ILK revealed similar areas of expression,

with positive MT1-MMP immunoreactivity in nearly half of all tissues staining

positively for ILK expression. Evaluation of ovarian cancer cells for mRNA

expression of ILK and MT1-MMP indicated both ILK and MT1-MMP expression in

the majority of samples. Ovarian cancer cell line DOV13 revealed high

expression of both proteins; this was further confirmed with fluorescent co-

localization studies. Together, these results suggest that ILK and MT1-MMP may

have some interaction in EOC processes.

It has been previously demonstrated that MT1-MMP functions to promote

formation of MCAs as well as invasion of interstitial collagen-rich sub-mesothelial

matrix, an event secondary to an initial cell-cell adhesive interaction. As described previously, attempts to overexpress ILK failed. Considering the myriad of methods used to introduce ILK which each failed to produce viable cells, one may surmise that ILK overexpression is toxic to DOV13 cells. Therefore, the

contribution of ILK to MT1-MMP mediated processes, including MCA formation,

cell attachment, and collagen invasion, was investigated via targeted siRNA

knockdown of ILK. However, use of siRNA is not without controversy. It must be

noted that transcripts having less than 100% complementarity with an siRNA can

be at risk for ineffective or non-specific siRNA knockdown, a phenomenon

termed “off-targeting”. There is a growing body of evidence which suggests that

non-specific effects can be induced by siRNAs at the mRNA level (off-target gene modulation) and at the translational level (off-target protein regulation) in a concentration dependent manner (Saxena et al. 2003; Persengiev et al. 2004;

Fedorov et al. 2006; Naito and Ui-Tei 2013). Use of these non-specific siRNAs can result in observation of siRNA-specific effects in functional assays, prompting false-positive conclusions with respect to the role of the target gene. However,

observation of the same phenotype with multiple independent siRNA transfections should increase the confidence with which any observed phenotype can be ascribed to the silencing of the target gene (Jackson and Linsley 2004).

Jackson et al. recommend two options to mitigate off-target effects: pooling

siRNAs for the same gene and siRNA redundancy (Jackson and Linsley 2004).

Recent suggestions to reduce unintended effects include: employment of new,

experimentally validated methods for siRNA design (through updated

bioinformatics and recently built design algorithms); and utilization of an

alternative structure designed to maximize complementarity, coined asymmetric

shorter-duplex RNA (asiRNA) (Wang et al. 2009; Chang et al. 2013; Naito and

Ui-Tei 2013). Moreover, it has been suggested that strict use of minimally

effective doses of siRNA can reduce off-targeting (Tschuch et al. 2008; Caffrey et

al. 2011).

To adjust for non-specific effects resulting from siRNA mediated mechanisms,

studies used non-targeted siRNA transfected cells as comparative controls.

These non-targeted siRNA transfectants were shown to behave identically to

non-transfected controls; the two cell lines were therefore used interchangeably

in portions of this study. Additionally, siRNA doses were optimized prior to

experimentation to determine the lowest concentration of RNA duplex needed to

achieve the ILK suppression and the best pool of available duplexes for use.

Knockdown of ILK in ovarian cancer cell line DOV13 reduced adhesion to collagen and to a mesothelial monolayer, suggesting that ILK is involved in ovarian cancer cell-ECM and cell-cell attachment. Interaction of specific cell surface receptors with extracellular adhesive molecules like collagen, triggers transduction of signals to control cellular behavior (Akiyama 1996). The mechanism by which ILK regulates the dynamic rearrangement of cell-matrix adhesions has yet to be elucidated, and correlations between ILK expression and cell adhesion have shown varied responses, with reports of both increased and decreased adhesive ability in response to ILK knockdown (Attwell et al. 2003;

Kim et al. 2007). Since ILK activity is stimulated by adhesion to the extracellular matrix, it could be speculated that ovarian tumor cells make an initial, yet tenuous cell-ECM connection and require ILK activity for further anchoring. Loss of ILK expression could delay this cell-ECM interaction, requiring engagement of other molecules for adhesion to occur over time.

Investigations of the role of ILK in ovarian cancer cell to mesothelial cell adhesion lends further support to this postulate, with DOV-ILK-KD displaying altered adhesion to the LP9 mesothelial monolayer when assayed for 30 minutes.

Adhesion of DOV-ILK-KD to live mesothelial tissue, however, did not significantly differ from controls. It can be theorized that the ex vivo mesothelial monolayer is

so tightly woven together that early ovarian cancer cell-ECM connections cannot be made, effectively negating any potential adhesive benefit held from having normal ILK expression. Additionally, since the explant has been removed from its normal environment within the peritoneal cavity, it is reasonable to assume that there could be other soluble molecules in vivo that are no longer available to aid adhesive interactions once removed from the live environment. It would be interesting to evaluate invasion of this explant, where ovarian cancer cells would eventually come into contact with the sub-mesothelial collagen matrix and the true effect of ILK knockdown assessed.

To support ILK knockdown studies, pharmacological inhibition of ILK activity was attempted with use of small-molecule ILK inhibitor QLT0267. QLT0267 is a widely used ATP analog that competes with endogenous ATP for the ILK kinase domain ATP-binding site (McDonald et al. 2008). It has been shown to inhibit ILK kinase activity with a half maximal inhibitory concentration (IC50) of between 2-

5µM QLT0267, depending on cell type (Troussard et al. 2006). Here, QLT0267 was utilized at 25µM and displayed a minor effect on collagen adhesion but had no significant effect on mesothelial cell adhesion.

It has been suggested that while potent and selective, inhibitory molecules such as QLT0267 lack complete specificity and may result in off-target effects

(Bantscheff et al. 2007; Muranyi et al. 2009). This “off-targeting” may account for the behavior differences between inhibitor treated cells and siRNA transfected revealed in these studies. Additionally, it is feasible to speculate that inhibitor potency and/or effectiveness may have been reduced due to a number of factors, including storage for any length of time at less than optimal conditions (moisture, light, etc), and multiple exposures to extreme temperature fluctuations

(freeze/thaw cycles).

ILK is important for ovarian cancer cell invasion; knockdown of ILK reduced

DOV13 invasion of collagen I compared to controls. Additionally, a reduction in invasion of 3D meso-mimetic cultures (comprised of mesothelial cells overlaying a 3D collagen matrix) assayed over 48 hours was observed. Enzymatic activity of

MT1-MMP has been well established as the prime determinant of a collagen-

invasive phenotype in ovarian carcinomas (Ellerbroek et al. 1999; Barbolina et al.

2007; Sodek et al. 2007; Hudson et al. 2009). Thus, it is reasonable to speculate

that the abrogated invasion observed during ILK silencing is associated with

dysregulated activity of MT1-MMP.

In summary, these data revealed early evidence to implicate ILK expression in activation of MT1-MMP. siRNA mediated knockdown of ILK reduced/altered several MT1-MMP mediated pro-metastatic events, specifically: ovarian cancer cell adhesion to collagen; initial heterotypic cell-cell adhesion to mesothelial cells; invasion through meso-mimetic cultures of 3D collagen matrices; and MCA formation. Moreover, while ILK expression did not affect long term heterotypic cell-cell adhesion, the decreased invasive potential of siRNA transfected cells through meso-mimetic cultures, an event specifically catalyzed by MT1-MMP activity, strongly suggests a correlation between ILK expression and MT1-MMP phosphorylation.

3. MT1-MMP SHEDDING OF MUC16/CA-125 MODULATES

MESOTHELIAL ADHESION AND INVASION

I. RATIONALE

As discussed, the expression and activity of transmembrane protease MT1-MMP has been implicated in a number of pro-metastatic events including proliferation, adhesion, invasion and metastasis (Egeblad and Werb 2002; Sounni et al. 2002;

Doi et al. 2011). There is also, however, evidence to suggest that MT1-MMP activity at the cell surface can induce cell-matrix detachment, as well as promote the formation and shedding of metastatic MCAs in EOC (Moss et al. 2009a). The overexpression of MT1-MMP in EOC tumors relative to the normal ovary, with enhanced expression in metastases relative to primary tumors (Barbolina et al.,

2007), suggests that it may catalyze events that contribute to metastatic success.

MUC16, a cell surface bound glycoprotein, is highly expressed on the ovarian tumor cell surface (Yin et al. 2002; Kui Wong et al. 2003). An uncharacterized

proteolytic event catalyzes shedding of MUC16 from the tumor cell surface,

whereupon the shed ectodomain can be measured in peritoneal fluid and

circulating blood (Vergote et al. 1992; Goodell et al. 2009). This shed form is

detected in the serum of EOC patients as CA-125 and is considered to be the standard molecular marker of EOC malignancy (Bast et al. 1983; Scholler and

Urban 2007; Moore et al. 2010). Reported biological functions of MUC16 include the facilitation of tumor cell-to-mesothelial cell adhesive interaction on peritoneal surfaces, and modulation of tumor cell growth (Rump et al. 2004; Patankar et al.

2005; Gubbels et al. 2006; Thériault et al. 2011). Moreover, MUC16 has been shown to bind strongly to mesothelin, a protein present on peritoneal mesothelial cells (Rump et al. 2004; Gubbels et al. 2006; Scholler et al. 2007). This interaction is thought to facilitate the initial adhesion and subsequent implantation and peritoneal spread that characterizes EOC metastasis.

In this chapter, the relationship between MT1-MMP and MUC16 as it pertains to cellular events that mimic initial EOC metastatic events is assessed. The experiments described within test the hypothesis that MT1-MMP activity may trigger the MUC16/CA-125 ectodomain shedding thought to precede heterotypic adhesive events between ovarian cancer tumor cells and the mesothelium.

II. RESULTS

Inverse relationship between MT1-MMP and MUC16 expression in ovarian cells and tissues.

To assess the potential correlation between MUC16 and MT1-MMP, expression of both antigens was evaluated in human ovarian tumors as well as in ovarian cancer cells. Serial sections of a tumor tissue microarray consisting of various histological types of malignant ovarian cancer were subjected to immunohistochemical analysis for either MT1-MMP (Fig. 3.1, A-C) or MUC16

(Fig. 3.1, D-F). An inverse relationship between MT1-MMP and MUC16 expression is observed, wherein over half (56%) of all tissues with higher level

MT1-MMP expression (Fig. 3.1, C) exhibit low MUC16 (Fig. 3.1, F), while tissues with elevated MUC16 (Fig. 3.1, D) do not stain positively for MT1-MMP (Fig. 3.1,

A). Similar results were observed in cultured ovarian cancer cells (Fig. 3.2).

Using OVCAR3, an ovarian cancer cell line known to express high levels of

MUC16 (Yin and Lloyd 2001), no surface expression of MT1-MMP was observed

(Fig. 3.2, A-B). Similar results were obtained with parental OVCA433 cells (Fig.

3.2, C-D). However, when OVCA433 cells were engineered to overexpress

catalytically active MT1-MMP (designated OVCA433-MT), surface staining for

MUC16 was lost (Fig. 3.2, E-F). Expression of a catalytically inactive mutant of

MT1-MMP (E240A substitution), which represents a functional loss of MT1-MMP, restored MUC16 expression (Fig. 3.2, G-H). Real-time PCR analysis confirmed presence of MUC16 across OVCA samples (Fig. 3.3, A). Results were further confirmed using flow cytometry, wherein OVCA433-MT cells exhibited diminished surface MUC16 relative to parental OVCA433 (Fig. 3.3, B). MUC16 surface expression was restored in cells cultured with the broad spectrum MMP inhibitor

GM6001 or in cells transfected with the MT1-MMP inactive mutant OVCA433-

MT-E240A. Staining for MUC1, another ovarian tumor cell surface bound mucin that is released into the circulation and is employed as a biomarker was also assessed (Devine et al. 1992). No significant change in expression was found in

OVCA433-MT cells compared to control (data not shown).

CA-125 is detected in the spent media of OVCA433-MT.

As the data above demonstrate an inverse relationship between MT1-MMP and

MUC16, suggesting that MT1-MMP may participate in MUC16 ectodomain

shedding, conditioned medium from MT1-MMP-expressing cells were analyzed for the presence of shed MUC16. Parental OVCA433 cells, OVCA433-MT, and

OVCA433-E240A cells were each cultured for 48 hours, after which the conditioned media were assessed for human CA-125 using a commercially

available ELISA. While quantifiable CA-125 was found in all samples (Fig. 3.3,

C), the concentration of CA-125 found in conditioned medium from OVCA433-MT cells was significantly increased compared to controls. Shed CA-125 levels were decreased in cells cultured with broad spectrum MMP inhibitor GM6001 (25µM) and were below control levels in OVCA433-MT-E240A cells.

Expression of MT1-MMP alters ovarian cancer cell:mesothelial cell adhesion.

Peritoneal mesothelial cells express high levels of mesothelin and MUC16 has

been shown to interact strongly with mesothelin, with binding occurring via the N-

linked glycoproteins of cell surface bound MUC16 (Gubbels et al. 2006). To

evaluate whether MT1-MMP-catalyzed MUC16 ectodomain shedding alters the ability to bind mesothelin, a heterotypic cell adhesion assay was used. In this assay, fluorescently labeled ovarian cancer cells were incubated with confluent monolayers of adherent human peritoneal mesothelial cells (LP9) plated atop a collagen type I matrix, washed, and enumerated (Fig. 3.4). While parental

OVCA433 cells adhered avidly to mesothelial cells (Fig. 3.4, B), mesothelial adhesion was significantly reduced in OVCA433-MT cells. Mesothelial adhesion was partially restored in cells expressing catalytically inactive MT1-MMP-E240A.

Treatment of cells with GM6001 (galardin, GM), a non-specific MMP inhibitor, also partially restored mesothelial adhesion.

MT1-MMP expression enhances ovarian cancer cell invasion through a live mesothelial cell monolayer.

It has previously been shown that expression of MT1-MMP enhances the ability

of ovarian cancer cells to invade collagen type I gels (Ellerbroek et al. 2001;

Moss et al. 2009b). To evaluate the role of MT1-MMP in penetration of a more

physiologically relevant cell-matrix context, the ability to invade 3D meso-mimetic

cultures cultured atop porous membranes in a transwell was evaluated (Fig. 3.4,

C; Lengyel et al. 2013). Expression of MT1-MMP significantly enhanced meso-

mimetic invasion (Fig. 3.4, D), while cells expressing the catalytically inactive

MT1-MMP-E240A were less invasive than control OVCA433 cells.

MT1-MMP expression alters attachment of ovarian cancer cells to mesothelial tissue.

Tumor cell interaction with intact peritoneal mesothelium was also assessed

using an ex-vivo assay to examine fluorescent ovarian cancer cell attachment to

live mesothelial tissue. In this experiment, cells are incubated with immobilized

peritoneal explants and adhesion is quantified by scanning electron microscopy

or by fluorescence microscopy (Fig. 3.5, A-F). Relative to control OVCA433 cells, attachment of OVCA433-MT cells to the peritoneal explant was decreased.

When compared to OVCA433-MT, attachment of cells expressing the catalytically inactive MT1-MMP-E240A was significantly increased (Fig. 3.5, G).

3.1 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN

CANCER TISSUES

Figure 133.1 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN CANCER TISSUES

Figure 3.1: Expression of MT1-MMP and MUC16 in Ovarian Cancer Tissues. Microarrayed cores of ovarian adenocarcinoma were subjected to immunohistochemical analyses for MT1-MMP (A-C) or MUC16 (D-F) as described. Examples of variation in staining intensity are shown.

3.2 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN

CANCER CELLS

Figure 143.2 EXPRESSION OF MT1-MMP AND MUC16 IN OVARIAN CANCER CELLS

Figure 3.2: Expression of MT1-MMP and MUC16 in Ovarian Cancer Cells. Subconfluent monolayers of OVCA433, OVCA433-MT, OVCA433-E240A or OVCAR3 cells (as indicated) were processed for immunofluorescence. Staining for MT1-MMP (green – A, C, E, G) or MUC16 (red – B, D, F, H) as indicated.

3.3 MT1-MMP OVEREXPRESSION IS ASSOCIATED WITH

DECREASED CELL SURFACE EXPRESSION OF MUC16

Figure 153.3 MT1-MMP OVEREXPRESSION IS ASSOCIATED WITH DECREASED CELL SURFACE EXPRESSION OF MUC16

Figure 3.3: MT1-MMP Overexpression is Associated with Decreased Cell Surface Expression of MUC16. (A) Quantitative real time PCR analysis of MUC16 levels. No significant differences in MUC16 expression levels are observed (results of a student’s t-test are displayed). Additional statistical testing: no significant mean difference among the three groups (Kruskal-Wallis, p=0.191)

and significantly reduced expression of MUC16 compared to control (Mann- Whitney, p=0.037). (B) Flow cytometric analysis of surface MUC16 (statistics as displayed). Additional statistical testing: reduced MUC16 expression in MT1- MMP-overexpressing cells compared to control (433-MT vs 433, Mann Whitney, p=0.004); rescued MUC16 expression in GM6001 treated cells compared to MT1-MMP-overexpressing cells (433-MT+GM vs 433-MT, Mann-Whitney, p=0.025); rescued MUC16 expression when catalytic function is lost (433-E/A vs 433-MT, Mann-Whitney, p=0.004). (C) Quantitation of shed MUC16 ectodomain as determined by ELISA in 48h conditioned culture media (statistics as displayed). Additional statistical testing: increased MUC16 found in conditioned media of MT1-MMP-overexpressing cells compared to control (433-MT vs 433, Mann-Whitney, p=0.050); treatment of MT1-MMP-overexpressors with GM6001 decreased the amount of MUC16 shed into the media (433-MT+GM vs 433-MT, Mann-Whitney, p=0.050); less MUC16 found in media when catalytic function is lost (433-E/A vs 433-MT, Mann-Whitney, p=0.050). Red bars – OVCA433; green bars – OVCA433-MT1-MMP; yellow bars – Ovca433-MT1-MMP cultured with broad spectrum inhibitor GM6001; blue bars – OVCA433-MT1-MMP-E240A mutant.

3.4 EXPRESSION OF MT1-MMP ALTERS TUMOR CELL

ADHESION TO AND INVASION THROUGH A LIVE MESOTHELIAL

CELL MONOLAYER

Figure 163.4 EXPRESSION OF MT1-MMP ALTERS TUMOR CELL ADHESION TO AND INVASION THROUGH A LIVE MESOTHELIAL CELL MONOLAYER

Figure 3.4: Expression of MT1-MMP Alters Tumor Cell Adhesion to and Invasion through a Live Mesothelial Cell Monolayer. (A) Schematic representation of adhesion assay, depicting representative image of ovarian tumor cells (green) attached to a confluent mesothelial cell (LP9) monolayer. Scale bar, 400µm. (B) CMFDA-labeled cells were allowed to adhere to a

confluent monolayer of LP9 mesothelial cells for 30min prior to washing (to remove unbound cells) and quantitation of adherent fluorescent cells. Results are quantified as a percentage of total cells seeded and results of a student’s t-test are displayed. Additional statistics: significantly reduced adhesion of 433-MT to mesothelial cell monolayer compared to control (433-MT vs 433, Mann-Whitney, p=0.050); treatment of MT1-overexpressing cells with GM6001 increased adhesion to the mesothelial cell monolayer compared to 433-MT alone (433- MT+GM vs 433-MT, Mann Whitney, p=0.050); increased adhesion in catalytically inactive cells (433-E/A vs 433-MT, Mann Whitney, p=0.050). (C) Schematic representation of meso-mimetic invasion assay. (D) CMFDA-labeled cells were seeded atop a mesomimetic culture comprised of collagen embedded fibroblasts overlaid with a layer of LP9 mesothelial cells in a Boyden chamber containing an 8µm pore filter. Following incubation for 48h, cells penetrating the meso-mimetic through the pores to the underside of the filter were enumerated. Results depict the ratio of migrated cells to total cells seeded and statistics are as displayed. Addtiional statistics: a significant mean difference among all three groups found (Kruskal Wallis, p=0.027); increased invasion of MT overexpressing cells compared to control (433-MT vs 433, Mann Whitney, p=0.050); invasion is reduced in catalytically dead cells compared to both 433 control (433-E/A vs 433, Mann Whitney p=0.050) AND MT1-MMP-overexpressors (433-E/A vs 433-MT, Mann-Whitney p=0.050). Red bars – OVCA433; green bars – OVCA433-MT1- MMP; blue bars – OVCA433-MT1-MMP-E240A mutant.

3.5 EXPRESSION OF MT1-MMP AFFECTS CELL-TO-

MESOTHELIAL ADHESION IN AN EX VIVO PERITONEAL

EXPLANT

Figure 173.5 EXPRESSION OF MT1-MMP AFFECTS CELL-TO- MESOTHELIAL ADHESION IN AN EX VIVO PERITONEAL EXPLANT

Figure 3.5: Expression of MT1-MMP Affects Cell-to-Mesothelial adhesion in an ex vivo Peritoneal Explant. (A-C) Depiction of assay. (A) Ovarian cancer cells were fluorescently labeled with CMFDA. (B) An excised explant from the

peritoneum of a female mouse was pinned, mesothelium-side up, to a tissue culture dish containing an optically clear silastic resin. Tumor cells were allowed to adhere for 2h prior to washing (to remove unlabeled cells). (C) Representative scanning electron micrograph showing tumor cells (round) adherent to mesothelial monolayer. (D-F) Representative images of fluorescently labeled ovarian tumor cells attached to explant. Scale bar, 400µm. (G) Quantitation of adhesion of OVCA433, OVCA433-MT, or OVCA433-EA to murine peritoneal tissue explant. Results are expressed as the relative number of cells per area and results of student’s t-test are as displayed. Additional statistics: a mean difference among all three groups was found (Kruskal Wallis, p=0.027); reduced adhesion of MT1-MMP-overexpressing cells to tissue compared to control (433- MT vs 433, Mann Whitney, p=0.050); increased adhesion of catalytically dead cells to tissue compared to both MT1-MMP-overexpressing cells (433-E/A vs 433-MT, Mann-Whitney, p=0.050) and control (433-E/A vs 433-MT, Mann- Whitney, p=0.050). Red bars – OVCA433; green bars – OVCA433-MT1-MMP; blue bars – OVCA433-MT1-MMP-E240A mutant.

III. DISCUSSION

The primary cellular target for ovarian cancer metastasis is the mesothelial cell, which covers the peritoneum lining the peritoneal cavity. Ovarian cancer cells dissociated from the primary tumor metastasize intra-peritoneally through

adhesion to and localized invasion of peritoneal mesothelium to anchor

secondary lesions. Over 70% of EOC are diagnosed with intra-peritoneal

metastasis, when 5-year survival rates are less than 30%. However, according to

the most recent Surveillance, Epidemiology, and End Results (SEER) database report, when EOC are diagnosed prior to metastatic dissemination, the survival rate dramatically increases to 92% (N Howlader et al. 2013). Numerous studies investigating serum biomarkers to screen women at risk for EOC have assessed several potential markers but none are considered to have sufficient sensitivity and specificity for effective population-level early detection (Cramer et al. 2011;

Husseinzadeh 2011; Mai et al. 2011).

MUC16/CA-125, a cell surface glycoprotein, is highly expressed on ovarian tumors whereupon it is shed from the tumor surface via a proteinase-dependent mechanism. MUC16/CA-125 in the peritoneal fluid ultimately reaches the blood serum, where it is detected as the CA-125 antigen (Bast et al. 1983). CA-125 has been the standard molecular marker of EOC malignancy for several decades due to its elevated serum levels in 80% of advanced stage EOC patients, but it is still considered to be an imperfect tool for early detection (Bast et al. 1983; Tuxen et al. 1995; Mai et al. 2011). MUC16 functions in EOC metastasis have been well described. The interaction between MUC16 and mesothelin, a protein present on the surface of mesothelial cells, has been extensively investigated and many studies have implicated CA-125:mesothelin binding in the early adhesive events of EOC metastasis (Rump et al. 2004; Gubbels et al. 2006; Kaneko et al. 2009;

Chen et al. 2013a). While several studies have reported that CA-125 shedding can be modulated by cell cycle functions (cells predominantly in S and G2-M

phase show reduced CA-125 shedding), tyrosine kinase inhibition, and

interferon-γ (Marth et al. 1989, 2007; Zeimet et al. 1996), the precise impetus for

MUC16 ectodomain shedding (and the corresponding increase in circulating CA-

125), however, remains to be elucidated. This work presents evidence to

implicate MT1-MMP in this process.

Initial experiments to evaluate MUC16 expression on the surface of metastatic ovarian cancer tumor tissue indicated moderate to strong expression along tumor

borders, irrespective of histotype. Interestingly, comparison with serial sections

stained for the cell surface proteinase MT1-MMP revealed an inverse correlation

between MT1-MMP and MUC16 immunoreactivity. These results were confirmed

using ovarian cancer cell lines. OVCAR3 cells, which express high MUC16

levels, do not exhibit surface expression of MT1-MMP. Furthermore, when

MUC16-expressing OVCA433 cells were engineered to overexpress MT1-MMP,

surface expression of MUC16 was lost while mRNA expression remained

unaffected. This was accompanied by enhanced soluble MUC16 ectodomain

antigen in the conditioned medium measured using a commercially available

ELISA assay (Experimental Methods section of Materials and Methodology).

Incubation of these cells with a broad spectrum MMP inhibitor or expression of a

catalytically inactive MT1-MMP mutant abrogated this effect. Together, these

results support the hypothesis that MT1-MMP catalyzes MUC16 ectodomain

shedding, releasing the CA-125 antigen from the cell surface.

It has been previously demonstrated that expression of MT1-MMP can function

to promote pro-metastatic behavior of EOC cells, including invasion of the

interstitial collagen-rich sub-mesothelial matrix and promotion of proliferation in

constrained 3D collagen gels (Barbolina et al. 2007; Moss et al. 2009b).

However, it may be predicted that sub-mesothelial collagen-invasive activity is

secondary to an initial heterotypic cell-cell adhesive event occurring between the ovarian cancer cell and the mesothelial cell. MUC16 has been shown to interact strongly with mesothelin, with cell to mesothelin binding occurring via the N- linked glycoproteins of cell surface bound MUC16 (Gubbels et al. 2006). Hence, the interaction between mesothelial cell membrane-bound mesothelin and tumor cell surface MUC16 may represent the inaugural adhesive event prior to invasion. Results showing decreased adhesion of cells expressing catalytically active MT1-MMP to 3D meso-mimetic cultures as well as to intact ex vivo peritoneal tissue explants reveal a loss of ovarian tumor cell:mesothelial cell adhesion, supporting the hypothesis of decreased cell surface MUC16 when

MT1-MMP is expressed. Further support lies in the significantly increased adhesion of catalytically dead MT1-MMP expressing cells to intact ex vivo peritoneal tissue over that of MUC16 expressing control cells. These data are highly suggestive of a role for MT1-MMP in potentiating MUC16 functionality in ovarian tumor cell attachment to mesothelin. Nevertheless, enhanced meso- mimetic invasion is observed in MT1-MMP-expressing OVCA433 cells, suggesting that though initial ovarian tumor cell:mesothelial cell interaction is dependent on MUC16 expression, additional molecules must be employed to engage the necessary heterotypic adhesion.

The role of integrins as mediators of ovarian cancer adhesion has been widely reported, with specific attention given to the interaction between β1 integrin and fibronectin, expressed on ovarian cancer and mesothelial cells, respectively

(Lessan et al. 1999; Strobel and Cannistra 1999; Casey et al. 2001; Ahmed et al.

2005; Desgrosellier and Cheresh 2010). Furthermore, our group has also demonstrated that ovarian cancer cells exhibit preferential adhesion to interstitial collagens in a β1 integrin-dependent process (Moser et al. 1996; Ellerbroek et al.

2001). Interestingly, additional results indicate that adhesion of OVCA433-MT cells (with low surface MUC16) to a mesothelial monolayer was inhibited by β1 integrin function-blocking antibodies while adhesion of OVCA433-MT-E240A

(high surface MUC16) remained unaffected (data not shown). Together, these data support the hypothesis that initial events in ovarian tumor cell:mesothelial cell adhesion are preferentially mediated by MUC16, transitioning to β1 integrins either in the absence of MUC16 or to stabilize adhesive contacts. Future studies aimed at assessing molecules with potential to interact with β1 integrin, as well as identifying the downstream effectors of those interactions, may be instrumental towards elucidating specific molecular mechanisms involved in ovarian tumor cell adhesion.

Moss, et al. have demonstrated that expression of MT1-MMP is enhanced in experimentally generated ovarian cancer MCAs relative to single cells (Moss et

al. 2009b). It has been reported that MCAs, or spheroids, isolated from human ovarian cancer ascites are less adhesive compared to individual ovarian carcinoma cells (Burleson et al. 2004b). Burleson and coworkers postulated that cells in spheroid form may favor homotypic interactions over heterotypic interactions with mesothelium, and further suggest that adhesion complexes between MCAs and mesothelium are of lower affinity than those of single cells and are thus more susceptible to disruption by shear forces present in ascites fluid (Burleson et al. 2004a). Therefore, it is interesting to speculate that the MCA

population in human ovarian cancer ascites represents a subpopulation of cells

with relatively higher MT1-MMP and lower MUC16 compared to free-floating single cells; however this remains to be investigated.

In summary, the data disclosed here support a model wherein acquisition of catalytically active MT1-MMP expression in ovarian cancer cells induces MUC16 ectodomain shedding, resulting in increased soluble MUC16. Cells with decreased cell surface expression of MUC16 exhibit reduced adhesion to meso- mimetic cultures and to intact peritoneal explants. Nevertheless, MT1-MMP-

expressing cells are more invasive through meso-mimetic cultures comprised of

mesothelial cells overlaying a 3D collagen matrix. These data support a model

wherein initial interactions of ovarian tumor cells with peritoneal mesothelium

may be facilitated by a lower affinity interaction between MUC16, which

protrudes extensively from the cell surface, and mesothelin. Proteolytic clearing of MUC16, catalyzed by MT1-MMP, may then expose integrins for high affinity cell binding to peritoneal tissues, thereby anchoring metastatic lesions for subsequent proliferation within the collagen-rich sub-mesothelial matrix.

4. OVERALL CONCLUSIONS

The body of work presented in this dissertation investigates the molecular

mechanisms surrounding MT1-MMP activity in early stages of ovarian cancer

metastasis, utilizing in vitro models of homotypic and heterotypic cell-cell

adhesion, a meso-mimetic invasion assay, and ex vivo tissue explants. MT1-

MMP expression induces MUC16/CA-125 ectodomain shedding, which may then expose integrins at the ovarian tumor cell surface for high affinity cell-cell and cell-ECM binding. β1 integrin-mediated cell adhesion activates ILK, which may catalyze phosphorylation of Thr567 on the MT1-MMP cytoplasmic tail, promoting pro-metastatic events, including strengthening of adhesive contacts, invasion of the collagen-rich sub-mesothelial matrix, and MCA formation.

The majority of ovarian cancers are epithelial in origin, arising, arguably, from the single layer of cells that cover the ovary or fallopian tube. Metastatic ovarian tumors arise once an epithelial cell transforms, inducing detachment from the primary tumor site. These shed cells travel throughout the peritoneal cavity, escaping anoikis to survive as single cells and MCAs, and metastasize intraperitoneally through adhesion to and invasion of the mesothelial cell layer

covering the peritoneum, the primary microenvironment for ovarian cancer metastasis. Mesothelial cells lie atop a collagen type I-rich ECM; subsequent to the initial attachment of ovarian cancer cells, proteolytic activity catalyzes migration through the mesothelial monolayer and promotes invasion of the sub- mesothelial matrix. MT1-MMP enzymatic activity has been shown to be critical to this process. Utilizing phospho-defective and phospho-mimetic mutants, (T567A and T567E, respectively) Moss et al. have shown that phosphorylation of Thr567 in the MT1-MMP cytoplasmic tail may promote metastasis. Expression of MT1-

MMP stimulates the growth of collagen invasive MCAs and induces rapid detachment kinetics, spontaneous release of confluent cell-cell adherent sheets, and enhanced matrix invasion (Moss et al. 2009a, 2009b). The kinase responsible for Thr567 phosphorylation, however, has yet to be identified.

ILK is co-expressed with MT1-MMP in ovarian carcinomas and assessment of siRNA mediated ILK knockdown has revealed: decreased ovarian cancer cell adhesion to and invasion of collagen type I; decreased attachment to a mesothelial cell monolayer; and reduced invasive ability of a meso-mimetic culture. Invasion of sub-mesothelial collagen matrices is a well characterized consequence of MT1-MMP expression (Ellerbroek et al. 2001; Fisher et al. 2006;

Barbolina et al. 2007; Moss et al. 2009b). Proteomic analysis of MT1-MMP has identified ILKAP, a negative regulator of ILK, as a binding partner, prompting

100

speculation that ILK may also interact with MT1-MMP (Stack, unpublished data).

The alteration of MT1-MMP-mediated pro-metastatic events during ILK down- regulation supports this theory.

MT1-MMP expression at the cell surface has also been shown to induce cell- matrix detachment (Moss et al. 2009a). MUC16/CA-125 is detected in the serum of EOC patients and is considered to be the standard molecular marker of EOC malignancy, but the proteolytic event catalyzing this shedding has yet to be elucidated.

An inverse correlation between MT1-MMP and MUC16 immunoreactivity was observed in human ovarian tumors and cells and when MUC16-expressing ovarian cancer cells were engineered to overexpress MT1-MMP, surface expression of MUC16 was lost, while cells expressing an inactive mutant retained surface MUC16. Cells expressing catalytically active MT1-MMP displayed decreased adhesion to meso-mimetic cultures and intact ex-vivo peritoneal explants, yet demonstrated enhanced meso-mimetic invasion, thus supporting a theory where active MT1-MMP expression in ovarian cancer cells induces MUC16/CA-125 ectodomain shedding, resulting in a less adhesive phenotype, but, curiously, a more invasive one as well. This may be explained by

101

the structure of MUC16, which protrudes extensively from the cell surface. Initial

interactions of ovarian tumor cells with the peritoneal mesothelium may be

facilitated by a lower affinity interaction between the extended MUC16 and

mesothelin. The MT1-MMP catalyzed proteolytic clearing of MUC16 may then expose integrins for high affinity cell binding to peritoneal tissues, thereby anchoring metastatic lesions for subsequent proliferation within the collagen-rich

sub-mesothelial matrix.

Collectively, published research and the findings of this project prompt

speculation of a potential mechanism of metastatic promotion delineated as

follows: MT1-MMP activation catalyzes shedding of MUC16, thus exposing

integrins at the cell surface to facilitate cell-matrix and cell-cell attachment; β1

integrins are engaged and ILK is activated; ILK expression mediates further MT1-

MMP activity, and the cycle is restarted. While the data obtained supports this

hypothesis, the molecular complexity characteristic of ovarian cancer metastatic

success necessitates signaling pathways with compensatory mechanisms. For

example, the effect of ILK knockdown on mesothelial adhesion, when assayed

for 30 minutes demonstrated reduced cell-cell adhesion; however, one may

surmise that there may be, perhaps, a temporal limit on adhesion abrogation,

and a consequent engagement of other signaling factors to moderate adhesion.

Moreover, it should be noted that adhesive success bore no effect on invasive

102

potential, further underscoring the role of MT1-MMP activity. An interesting next

step would evaluate the effect of ILK overexpression on MUC16 cell surface

expression; however, as technical issues have prevented assessment of ILK-

overexpressed cells thus far, it may be more prudent to examine the effects of

ILK silencing on MUC16 expression.

A recent publication on mesothelin knockout mice also opens up an intriguing avenue for further analysis of early adhesive events (Zhang et al. 2014). It is

interesting to speculate whether the employment of mesothelin-deficient murine peritoneal tissue for ex vivo study, as described in this dissertation, would hinder tumor:mesothelial cell adhesion. It was this author’s intention to pursue these

studies for inclusion in this dissertation, however, this route was not feasible.

Zhang and colleagues had recently frozen embryos and had to re-derive animals,

thus requiring a 7-10 month waiting period before a breeding pair could be

provided to the research facility and an additional 2-3 months to provide sufficient

knockout offspring for analysis (Harper Cancer Research Center, University of

Notre Dame, South Bend, Indiana).

The overarching goal of this dissertation was to investigate the role of MT1-MMP

in the early mechanisms of ovarian cancer metastasis. Collectively, the data

103

presented here provide an excellent basis for further investigation of the microenvironment of the mesothelium and the molecules controlling adhesion to and invasion of this primary site of ovarian cancer metastasis.

104

5. MATERIALS AND METHODOLOGY

I. MATERIALS

Antibodies.

Anti-ILK antibodies were purchased from Millipore (Billerica, MA) (for Western blot analysis), Abcam (Cambridge, MA) (for immunoprecipation), or Sigma-

Aldrich (St. Louis, MO) (for immunofluorescence). Phospho-ILK (Thr173), ILKAP, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies were purchased from Santa Cruz Biotechnology, Inc (Santa Cruz, CA). Antibodies directed against phospho-Akt(Ser473), total Akt, and phospho-Threonine (in the sequence phospho-Threonine-X-Arginine) were purchased from Cell Signaling

Technology (Beverly, MA). β1 integrin functional inhibitor P5D2 was purchased from Millipore (Billerica, MA). MT1-MMP antibodies were purchased from

Epitomics (Burlingame, CA). CA-125 antibody for IHC was purchased from Dako

(Carpinteria, CA); and MUC16 antibodies for IF and flow cytometry were purchased from Abcam (Cambridge, MA). Alexa Fluor-488, 594-, and 647-

105

conjugated secondary antibodies were purchased from Invitrogen (Carlsbad,

CA).

Other materials.

Rat tail collagen type I was purchased from BD Biosciences (San Diego, CA).

SYLGARD® 184 Silicone Elastomer Kit was purchased from Dow Corning

(Midland, Mi). Cell culture media and additives were purchased from Gibco

Invitrogen (Carlsbad, CA) unless otherwise stated. Geneticin® (G418) and

TaqMan materials (TaqMan Fast Universal PCR Master Mix and TaqMan Gene

Expression Assay Mix) were purchased from Life Technologies, Inc (Carlsbad,

CA). Galardin/Ilomastat (GM6001) was purchased from Millipore (Billerica, MA).

The EDTA-free Complete Protease Inhibitor and Halt Phosphatase Inhibitor cocktails used in modified RIPA buffer were purchased from Roche Diagnostics

(Indianapolis, IN) and Thermo Scientific (Rockford, Il), respectively. Protein G

Sepharose beads used in immunoprecipitation studies were purchased from

Sigma-Aldrich (St. Louis, MO). Immobilon-P polyvinylidene membranes were purchased from Millipore (Billerica, MA). SuperSignal West Dura

Chemiluminescent Substrate was purchased from Thermo Scientific (Rockford,

IL). ILK inhibitor QLT0267 was a generous gift from Dr. Shoukat Dedhar (BC

Cancer Research Centre, Vancouver, B.C.). All chemicals were analytical grade.

106

II. MODELS

Cell lines and culture.

All cells were maintained in a humidified incubator under standard conditions

(37°C with 5% CO2). Ovarian cancer cell lines DOV13 and OVCA433 were

generously provided by Dr. Robert Bast, Jr. (M.D. Anderson Cancer Center,

Houston, TX). The DOV13 cell line was chosen for experiments described in

Chapter 2 due to its high levels of endogenous MT1-MMP. This enzymatic profile

made these cells ideally suited for characterizing the effect of ILK on MT1-MMP phosphorylation. In contrast, the OVCA433 cell line was chosen for experiments described in Chapter 3 due to its status as the parenteral cell line of mutants engineered to overexpress MT1-MMP, which provided an optimal model for assessing the relationship between MT1-MMP and MUC16. Both DOV13 and

OVCA433 cell lines were maintained in minimum essential media (MEM)

supplemented with 10% fetal bovine serum (FBS), 1mM sodium pyruvate, 1mM

non-essential amino acid, 50U/ml penicillin, and 50µg/ml streptomycin; DOV13

medium required an addition of 10µg/ml human recombinant insulin (Life

Technologies, Inc., Carlsbad, CA).

107

Knockdown of ILK was completed using the ILK-targeted TriFECTa™ Dicer-

Substrate RNAi kit (Integrated DNA Technologies, San Diego, CA) and

Lipofectamine® RNAiMAX transfection reagent (Life Technologies, Inc.,

Carlsbad, CA) according to manufacturers’ protocol. Briefly, a transfection cocktail (Opti-MEM, Lipofectamine® RNAiMAX and either ILK siRNA duplex or

included non-targeting control siRNA) was placed in a cell culture vessel and

incubated at room temperature for 5 minutes. Subconfluent DOV13 cells were

trypsinized, neutralized, suspended in antibiotic-free media, and diluted to

2.5x105 cells/ml. Upon incubation termination, 1ml of cell suspension was added

to transfection cocktail and the entire solution incubated for 24 hours. After initial

incubation, samples were again subjected to fresh transfection cocktail for an

additional 24 hours. After this final incubation, samples were assessed to confirm

knockdown and utilized immediately as described.

For pharmacologic inhibition, DOV13 cells were subjected to QLT0267 (dissolved

in DMSO) at the indicated doses in serum-free media and incubated for 24

hours. As a control, DOV13 cells were treated with a comparative volume of

DMSO solution alone. After incubation, cells were utilized as indicated by assay protocol.

108

The transfection and generation of the stable cell lines OVCA433-MT and

OVCA433-E240A has been previously described (Moss et al. 2009a).

Transfected cells were maintained in the parental OVCA433 media described

above, with the addition of selection antibiotic Geneticin® (700µg/ml). Cells were

routinely FACS-sorted every 3-5 passages with anti-MT1-MMP antibodies to insure continued cell surface expression using FACSAria III Cell Sorter (BD

Biosciences, San Jose, CA) and all experiments were performed with freshly sorted cell populations. Untransfected parental OVCA433 cells and OVCA433 cells transfected with empty vector were designated as “control”; both behaved identically in assays and were used interchangeably.

The ovarian adenocarcinoma cell line OVCAR3 was obtained from American

Type Culture Collection (ATCC, Manassas, VA). OVCAR3 expresses high levels of MUC16, thus this cell line was selected for use as a MUC16 positive control

for those experiments described in Chapter 3. OVCAR3 cells were maintained in

RPMI 1640, 20% FBS, 1mM sodium pyruvate, 1mM non-essential amino acid,

2mM L-glutamine, 50U/ml penicillin, and 50µg/ml streptomycin.

The human mesothelial cell line LP9 was obtained from Coriell Aging Cell

Repository (Coriell Institute, Camden, NJ). The LP9 cell line was maintained in

109

1:1 ratio Medium 199 and Ham’s F12, supplemented with 15% FBS, 2mM

glutamine, 1mM HEPES, 20mg/ml epidermal growth factor (R&D systems,

Minneapolis, MN), 0.4µg/ml hydrocortisone, and 50U/ml penicillin, and 50µg/ml

streptomycin.

Tumor Tissue Microarrays.

Immunohistochemical detection of antigen content in malignant ovarian tissue

was performed as described below by using commercially available serialized tissue microarrays (Biomax 110118) containing a variety of tissues, including

numerous adenocarcinomas and several clear cell carcinomas (US Biomax,

Rockville, MD).

110

III. EXPERIMENTAL METHODS

Preparation of whole cell lysates

Whole cell lysates from attached cultures were collected by the following process: the cell monolayer was washed twice with phosphate-buffered saline, treated with 600µl modified RIPA lysis buffer (150mM NaCl; 50mM Tris, pH 7.5;

20mM NaF; 10mM Na2P2O7; 5mM EDTA; 1% Triton X-100; 0.1% SDS) containing protease inhibitor and/or phosphatase inhibitor cocktails as appropriate, for 10 minutes at 4ºC, scraped for complete collection, and then passed through a 26⅝ gauge syringe 5 times. Protein concentration was measured using DC™ Protein Assay (Bio-Rad Laboratories, Inc., Hercules, CA).

Western blot analysis.

For western blot analysis of whole cell lysates, lysates were collected as described above and protein content standardized. Protein (20µg) was electrophoresed on 9% SDS-polyacrylamide gels and transferred to methanol- activated polyvinylidene membranes. After transfer, membranes were incubated with 3% BSA in TBST (150mM NaCl, 25mM Tris, 0.05% Tween 20) to block non- specific binding for 1 hour at room temperature. Primary antibodies were diluted

111

as indicated in 3% BSA/TBST and incubated overnight at 4ºC. After washing, the

membranes were incubated with horseradish peroxidase-conjugated secondary

antibodies for 1 hour at room temperature, and then visualized with enhanced

chemiluminescence using ImageQuant™ LAS4000 (GE Healthcare Life

Sciences, Pittsburgh, PA). For assessment of controls, membranes were

stripped of antibody by washing for 1 hour at room temperature in an optimized

stripping buffer (150mM NaCl; 100mM β-mercaptoethanol; 50mM Tris, pH 6.8;

1% SDS; 0.02% NaN3), and then re-probed as appropriate for the control

antibody. Samples were normalized to control antibody (GAPDH) and quantified

using Multigauge v.2 for densitometric analysis (FUJIFILM, Tokyo, Japan).

Quantitative Real Time PCR (qPCR).

RNA was extracted from 106 cells (as indicated) using RNeasy Mini Kit (Qiagen,

Valencia, CA) in accordance with the manufacturer’s instructions. cDNA was synthesized from 1-5µg of total RNA using RT2 First Strand Kit (Qiagen,

Valencia, CA). Real time PCR was performed on a StepOnePlus™ Real-Time

PCR System (Applied Biosystems, Foster City, CA). Detection of ILK was achieved using a specific Taqman Gene Expression Assay (Hs00199714_m1,

Applied Biosystems, Foster City, CA) and conditions for amplification were as recommended by Applied Biosystems: an initial denaturation for 20 seconds at

95ºC followed by 40 cycles of 95ºC for 1 second and 60ºC for 20 seconds.

112

Taqman assay for housekeeping gene hypoxanthine ribosyltransferase (HPRT,

Hs99999909_m1, Applied Biosystems, Foster City, CA) was used for

normalization. Conditions for amplification using the iTaq SYBR Green Supermix

(Bio-Rad Laboratories, Inc., Hercules, CA) as a fluorescent reporter were as

follows: an initial denaturation for 10 min at 95ºC followed by 40 cycles of 95ºC

for 15s and 60ºC for 30s. PCR primer specificity was determined via melting

curves, where products were heated at 95ºC, cooled to 65ºC, and then slowly

melted at 0.5ºC/s up to 95ºC. Primer sequences for MUC16 were as follows:

forward, 5’-TGC GGT GTC CTG GTG ACC-3’; reverse, 3’-CAC CGG CAA GTT

CCA GTC-5’. GAPDH was used as an internal control in each reaction (Applied

-(∆C sample - ∆C control) Biosystems, Foster City, CA). The comparative CT method (2 t t )

was used to determine average relative quantitation.

Immunohistochemistry.

TMA of ovarian cancer tissues (as indicated throughout the text) were de- parrafinized in xylene for 5 minutes (thrice) and soaked in absolute alcohol for 3

minutes. To inhibit endogenous peroxidase, slides were incubated for 30 minutes

in 3.3% H2O2in methanol. Antigen retrieval was accomplished by incubation in

heated sodium citrate (10mM) for 1 hour. Slides were processed using

VECTASTAIN Elite ABC kit (Vector Laboratories, Burlingame, CA). Nonspecific interactions were blocked by using normal horse serum for 30 minutes at room

113

temperature. Primary antibodies were diluted 1:20 in 0.01% phosphate buffered

saline (PBS), pH 7.4 and incubated overnight at 4°C in a humidified chamber.

Bound antibodies were detected by using a BioGenex IHC kit containing

biotinylated secondary antibody and streptavidin-conjugated HRP enzyme

coupled with 3,3’-diaminobenzidine chromagen solution in proprietary

formulations (BioGenex, San Ramon, CA). Tissues were counterstained with

hematoxylin, blued with saturated lithium carbonate solution, and digitally

photographed on Aperio ImageScope (Leica Biosystems, Buffalo Grove, IL).

Immunofluorescence.

Cells were subcultured on 22mm2 glass coverslips (coated as indicated), washed

twice with ice-cold PBS, and fixed with 4% paraformaldehyde in 0.12M sucrose

in PBS for 10 minutes at room temperature. If necessary, cells were washed and

then permeabilized with 0.1% Triton-X for 5 minutes. Cells were blocked with 5%

normal goat serum in PBS for 1 hour at room temperature, incubated with

primary antibody (1:100) in 1% normal goat serum in PBS for 1h at 37°C, rinsed

thrice for 5 minutes with PBS, and incubated with appropriate Alexa-Fluor

conjugated secondary antibody at a 1:500 dilution for 30 minutes at 37°C. After

washing, cells were allowed to dry, mounted with VECTASHIELD Mounting

Media with 4’, 6-diamidino-2-phenylindole (DAPI) (Vector laboratories,

114

Burlingame, CA), and visualized on an EVOS® FL digital inverted fluorescence

microscope (Advanced Microscopy Group, Mill Creek, WA).

FACS analysis.

Subconfluent cells were harvested with 20mM EDTA and washed twice with

PBS. (For studies utilizing inhibitor GM6001, subconfluent monolayers were

incubated in 25µg/ml GM6001 for 18-20 hours at 37ºC). Cells (2x105) recovered

in complete media for 30 minutes at 37°C, were centrifuged, washed twice, and

immunolabeled using indicated antibody for 30 minutes at 4°C. All antibodies were diluted 1:1000. After washing, cells were stained with secondary antibody conjugated to Alexa Fluor 647 (1:500) for 30 minutes at 4°C, washed twice, and resuspended in 0.5ml PBS for analysis on a Beckman Coulter FC500 Flow

Cytometer (Beckman Coulter, Inc, Indianapolis, IN). Data was analyzed using

FlowJo (TreeStar, Ashland, OR). Experiments were conducted in triplicate.

Results are expressed as an average of the percentage of fluorescent units found in each treatment compared to secondary stained controls.

Enzyme-linked immunosorbent assay (ELISA).

A commercially available sandwich ELISA kit (Sigma, St. Louis, MO) consisting

of 96-well plates coated with antibody specific for human CA-125 was employed

115

to assess protein concentrations of CA-125 present in the conditioned media of

freshly sorted cells plated for 48 hours (up to confluence). After incubation, media

was collected, centrifuged gently to remove cellular debris, and pipetted into

wells. Biotinylated detection antibody and HRP-conjugated streptavidin were

used for detection of bound CA-125. Samples were assayed in duplicate and

biological replicates of the experiment were performed. Absorbance readings

were made at 450nm and were acquired on a Molecular Devices Spectramax

microtiter plate reader (Sunnyvale, CA). CA-125 levels were determined by

interpolation from a standard curve.

Formation of multicellular aggregates (MCAs)

MCAs were generated using a modification of the hanging drop method as previously described (Kelm et al. 2003). Briefly, cells were harvested in complete

medium and diluted to a concentration of 1x105 cells/ml. From this, 20µl of cell

suspension was gently pipetted onto the underside of a 150mm tissue culture

plate lid. To avoid dehydration of the hanging droplet, 20ml of sterile PBS was

placed in the tissue culture dish immediately prior to inverting the droplet-

containing lid. MCA formation was monitored after incubation at 37ºC for 6-12

hours.

116

Collagen I cell adhesion assay.

Tissue culture wells were coated with 10µg/ml collagen type I in sodium

carbonate, pH 9.6 overnight at 4°C, washed with PBS, and air dried. Ovarian cancer cells were seeded as indicated, allowed to adhere at 37ºC for 30 minutes,

and washed for removal of non-adherent cells as described later in this section

(“Wash protocol for adhesion assays”). After washing, cells were fixed and

adherent cells were enumerated for analysis. Assays were performed in triplicate and five 20X fields/well were counted.

Mesothelial cell adhesion assay.

Tissue culture wells were coated with 10µg/ml collagen type I in sodium

carbonate, pH 9.6 overnight at 4°C, washed with PBS, and air dried. LP9 human

peritoneal mesothelial cells were seeded and grown for 48 hours to form a tightly

woven monolayer. Cancer cells were labeled with CellTracker™ Green, (5-

Chloromethylfluorescein Diacetate (CMFDA), Life Technologies Inc, Carlsbad,

CA) for 30 minutes at 37ºC. The live mesothelial monolayer was washed twice

with PBS, seeded with cells as indicated, and allowed to adhere for 30 minutes.

Wells were then washed for removal of non-adherent cells as described later in

this section (“Wash protocol for adhesion assays”) and fixed. Fluorescent cells

were enumerated. Assays were performed in triplicate and five fields/well were

counted.

117

Collagen I invasion assay.

An 8µm microporous membrane located within the upper compartment of a transwell insert (BD Biosciences, San Diego, CA) was coated with 10µg/ml collagen type I in sodium carbonate, pH 9.6 overnight at 4°C, washed with PBS, and air dried. Cells were seeded atop the filter and the apparatus incubated at

37ºC for 24 hours. After incubation, migrated cells passing through the 8µm pore filter were fixed, stained with Diff-Quik (Fisher Scientific, Pittsburgh, PA) and enumerated. All experiments were completed in triplicate and five fields/well were counted.

Meso-mimetic invasion assay.

A 3D collagen type I matrix was prepared as previously described (Moss et al.

2009b) and plated atop an 8µm microporous membrane within a transwell insert

(BD Biosciences, San Diego, CA). The matrix was then overlaid with LP9 human peritoneal mesothelial cells (Lengyel et al. 2013). The LP9 were allowed to grow

to confluence, forming a tight monolayer. CellTracker™ Green-labelled

OVCA433 cells were seeded atop the live monolayer and the co-culture was incubated at 37°C for 48 hours in a 1:1 ratio of complete media for each cell type.

Migrated cells passing through the 3-dimensional culture and the 8µm pore filter

were fixed, stained with Diff-Quik (Fisher Scientific, Pittsburgh, PA) and

118

enumerated. All experiments were completed in triplicate and five fields/well were

counted.

Adhesion to a peritoneal explant.

Peritoneal tissues from female c57bl/6 mice between 22 and 26 weeks of age

were used for assessments of cellular adhesion to mesothelial tissues. Tissues

were excised, rinsed twice, and submerged in PBS. While submerged, tissues

were uniformly trimmed to 2.0 x 2.0mm square, using tweezers to steady tissue

against a ruler and surgical scissors to cut. Once cut, tissues were gently placed

upon and pinned to a SYLGARD® 184 silicone coated dish (here, the

SYLGARD® 184 silicone coat serves as a cushioned, nonreactive surface that

aids stationary positioning and pinning of sample tissue). Throughout surgical

removal and preparation, extreme care was taken to ensure tissues were not

stretched; this included handling tissues with only one surgical apparatus at a time (other than when being cut) and refraining from tissue adjustment after placement upon the SYLGARD® 184 silicone coated dish. Cells to be assayed were labeled with CellTracker™ Green, trypsinized, and diluted to 2x105 cells/ml.

2.5ml of cell suspension was added to tissue and the assay was incubated for 2 hours at 37ºC. After incubation, the tissue was washed for removal of non-

adherent cells as described later in this section (“Wash protocol for adhesion

assays”), removed from the silicone bed, and mounted onto a glass coverslip for

119

imaging (EVOS® FL) and enumeration. Only cells on the mesothelial surface

were enumerated (cells also adhere to exposed collagen on the “wounded” cut

edges of the tissue block). Alternatively, tissues were processed for scanning

electron microscopy, as described previously (Barbolina et al. 2013). Briefly,

post-assay tissues were washed as described later in this section (“Wash

protocol for adhesion assays”), fixed, dehydrated, and dried. Once dried,

samples were placed on carbon stubs, subjected to Flash-Dry™ silver paint,

coated with platinum, and examined using a Magellan 400 scanning electron

microscope (FEI, Hillsboro, OR).

Wash protocol for adhesion assays.

Incubation media from each well was removed by suction pipette and wells were placed on a laboratory shaker. Room temperature PBS (2.5ml) was pipetted onto the wall of each well (to keep the integrity of the sample intact) and each apparatus was shaken at a high velocity (level 5 on shaker dial) for 5 minutes.

This procedure of removal, PBS application, and shaking was repeated 5 times for each sample. Once completed, samples were readied for additional steps, as detailed in each respective assay description.

Statistical analyses.

120

Statistical significance is defined as p<0.05 and was calculated employing a

variety of statistical tests (Student’s t-test, Mann-Whitney U Test, Kruskal-Wallis)

as indicated using SigmaPlot v.12 (Systat Software Inc., San Jose, CA).

Statistical tests were chosen based on the normality of the data set. Parametric

tests (i.e. Student’s t-test) were used to compare the means of two independent

samples; however, since much of the data within these studies could not be

completely described by two parameters (mean and standard deviation), non-

parametric tests were utilized. The Kruskal-Wallis test, a non-parametric version

of ANOVA, was used in instances where there were equal sample sizes in all

groups and the comparatives had one nominal variable and one measurement

variable. This test was employed to examine whether the mean ranks of the

measurement variable were the same in all groups. Measurement observations were converted to their ranks in the overall data set; when scores received tied ranks, a correction factor was used. The Mann-Whitney U test, the non-

parametric analogue to the Student’s t-test, was also used instances where the

comparative had one nominal variable and one measurement variable, however

this test was only utilized when comparing exactly two values. Statistical

significance is stated throughout the text (the statistical test used and the

resulting p-value).

121

REFERENCES

Afzal, S., Lalani, E.N., Poulsom, R., Stubbs, a, Rowlinson, G., Sato, H., et al. (1998). MT1-MMP and MMP-2 mRNA expression in human ovarian tumors: possible implications for the role of desmoplastic fibroblasts. Hum. Pathol. 29: 155–65.

Ahmed, N., Riley, C., Oliva, K., Stutt, E., Rice, G.E., and Quinn, M. a (2003). Integrin-linked kinase expression increases with ovarian tumour grade and is sustained by peritoneal tumour fluid. J. Pathol. 201: 229–37.

Ahmed, N., Riley, C., Rice, G., and Quinn, M. (2005). Role of integrin receptors for fibronectin, collagen and laminin in the regulation of ovarian carcinoma functions in response to a matrix microenvironment. Clin. Exp. Metastasis. 22: 391–402.

Akiyama, S.K. (1996). Integrins in cell adhesion and signaling. Hum. Cell. 9: 181–6.

American Cancer Society (2013). Cancer facts & figures 2013 (Atlanta).

Aoudjit, F., and Vuori, K. (2012). Integrin signaling in cancer cell survival and chemoresistance. Chemother. Res. Pract. 2012: 283181.

Attwell, S., Mills, J., Troussard, A., Wu, C., and Dedhar, S. (2003). Integration of Cell Attachment , Cytoskeletal Localization , and Signaling by Integrin-linked Kinase ( ILK ), CH-ILKBP , and the Tumor Suppressor PTEN. 14: 4813–4825.

Auersperg, N., Pan, J., Grove, B.D., Peterson, T., Fisher, J., Maines-Bandiera, S., et al. (1999). E-cadherin induces mesenchymal-to-epithelial transition in human ovarian surface epithelium. Proc. Natl. Acad. Sci. U. S. A. 96: 6249–54.

Bahr, J.M., and Palmar, S. (1989). The influence of aging on ovarian function. CRC Crit. Rev. Poult. Biol. 2: 103–110.

122

Bantscheff, M., Eberhard, D., Abraham, Y., Bastuck, S., Boesche, M., Hobson, S., et al. (2007). Quantitative chemical proteomics reveals mechanisms of action of clinical ABL kinase inhibitors. Nat. Biotechnol. 25: 1035–44.

Barbolina, M. V, Adley, B.P., Ariztia, E. V, Liu, Y., and Stack, M.S. (2007). Microenvironmental regulation of membrane type 1 matrix metalloproteinase activity in ovarian carcinoma cells via collagen-induced EGR1 expression. J. Biol. Chem. 282: 4924–31.

Barbolina, M. V, Liu, Y., Gurler, H., Kim, M., Kajdacsy-Balla, A. a, Rooper, L., et al. (2013). Matrix rigidity activates Wnt signaling through down-regulation of Dickkopf-1 protein. J. Biol. Chem. 288: 141–51.

Barbolina, M. V, Moss, N.M., Westfall, S.D., Liu, Y., Burkhalter, R.J., Marga, F., et al. (2009). Microenvironmental regulation of ovarian cancer metastasis. Cancer Treat. Res. 149: 319–34.

Barbolina, M. V, and Stack, M.S. (2008). Membrane type 1-matrix metalloproteinase: substrate diversity in pericellular proteolysis. Semin. Cell Dev. Biol. 19: 24–33.

Bast, R.C., Feeney, M., Lazarus, H., Nadler, L.M., Colvin, R.B., and Knapp, R.C. (1981). Reactivity of a monoclonal antibody with human ovarian carcinoma. J. Clin. Invest. 68: 1331–7.

Bast, R.C., Klug, T.L., St John, E., Jenison, E., Niloff, J.M., Lazarus, H., et al. (1983). A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N. Engl. J. Med. 309: 883–7.

Berns, E.M.J.J., and Bowtell, D.D. (2012). The changing view of high-grade serous ovarian cancer. Cancer Res. 72: 2701–4.

Birbeck, M.S., and Wheatley, D.N. (1965). An electron microscopic study of the invasion of ascites tumor cells into the abdominal wall. Cancer Res. 25: 490–7.

Bloss, J.D., Liao, S.Y., Buller, R.E., Manetta, A., Berman, M.L., McMeekin, S., et al. (1993). Extraovarian peritoneal serous papillary carcinoma: a case-control retrospective comparison to papillary adenocarcinoma of the ovary. Gynecol. Oncol. 50: 347–51.

Blystone, S.D., Lindberg, F.P., LaFlamme, S.E., and Brown, E.J. (1995). Integrin beta 3 cytoplasmic tail is necessary and sufficient for regulation of alpha 5 beta 1

123

phagocytosis by alpha v beta 3 and integrin-associated protein. J. Cell Biol. 130: 745–54.

Boivin, M., Lane, D., Piché, A., and Rancourt, C. (2009). CA125 (MUC16) tumor antigen selectively modulates the sensitivity of ovarian cancer cells to genotoxic drug-induced apoptosis. Gynecol. Oncol. 115: 407–13.

Bouanene, H., and Miled, A. (2010). Conflicting views on the molecular structure of the cancer antigen CA125/MUC16. Dis. Markers. 28: 385–94.

Boudeau, J., Miranda-Saavedra, D., Barton, G.J., and Alessi, D.R. (2006). Emerging roles of pseudokinases. Trends Cell Biol. 16: 443–52.

Burleson, K.M., Casey, R.C., Skubitz, K.M., Pambuccian, S.E., Oegema, T.R., and Skubitz, A.P.N. (2004a). Ovarian carcinoma ascites spheroids adhere to extracellular matrix components and mesothelial cell monolayers. Gynecol. Oncol. 93: 170–81.

Burleson, K.M., Hansen, L.K., and Skubitz, A.P.N. (2004b). Ovarian carcinoma spheroids disaggregate on type I collagen and invade live human mesothelial cell monolayers. Clin. Exp. Metastasis. 21: 685–97.

Burridge, K., and Chrzanowska-Wodnicka, M. (1996). Focal adhesions, contractility, and signaling. Annu. Rev. Cell Dev. Biol. 12: 463–518.

Cabodi, S., Pilar Camacho-Leal, M. del, Stefano, P. Di, and Defilippi, P. (2010). Integrin signalling adaptors: not only figurants in the cancer story. Nat. Rev. Cancer. 10: 858–70.

Caffrey, D.R., Zhao, J., Song, Z., Schaffer, M.E., Haney, S. a, Subramanian, R.R., et al. (2011). siRNA off-target effects can be reduced at concentrations that match their individual potency. PLoS One. 6: e21503.

Cannistra, S. a (2004). Cancer of the ovary. N. Engl. J. Med. 351: 2519–29.

Casey, R.C., Burleson, K.M., Skubitz, K.M., Pambuccian, S.E., Oegema, T.R., Ruff, L.E., et al. (2001). Beta 1-integrins regulate the formation and adhesion of ovarian carcinoma multicellular spheroids. Am. J. Pathol. 159: 2071–80.

Chang, C., Hong, S.W., Dua, P., Kim, S., and Lee, D. (2013). The design, preparation, and evaluation of asymmetric small interfering RNA for specific gene silencing in mammalian cells. Methods Mol. Biol. 942: 135–52.

124

Chaturvedi, P., Singh, A.P., and Batra, S.K. (2008). Structure, evolution, and biology of the MUC4 mucin. FASEB J. 22: 966–81.

Chauhan, S.C., Kumar, D., and Jaggi, M. (2009). Mucins in ovarian cancer diagnosis and therapy. J. Ovarian Res. 2: 21.

Chauhan, S.C., Singh, A.P., Ruiz, F., Johansson, S.L., Jain, M., Smith, L.M., et al. (2006). Aberrant expression of MUC4 in ovarian carcinoma: diagnostic significance alone and in combination with MUC1 and MUC16 (CA125). Mod. Pathol. 19: 1386–94.

Chen, J., Zhang, J., Zhao, Y., Li, J., and Fu, M. (2009). Integrin beta3 down- regulates invasive features of ovarian cancer cells in SKOV3 cell subclones. J. Cancer Res. Clin. Oncol. 135: 909–17.

Chen, S.-H., Hung, W.-C., Wang, P., Paul, C., and Konstantopoulos, K. (2013a). Mesothelin binding to CA125/MUC16 promotes pancreatic cancer cell motility and invasion via MMP-7 activation. Sci. Rep. 3: 1870.

Chen, X., Zhang, J., Cheng, W., Chang, D.Y., Huang, J., Wang, X., et al. (2013b). CA-125 level as a prognostic indicator in type I and type II epithelial ovarian cancer. Int. J. Gynecol. Cancer. 23: 815–22.

Cheng, W., Liu, J., Yoshida, H., Rosen, D., and Naora, H. (2005). Lineage infidelity of epithelial ovarian cancers is controlled by HOX genes that specify regional identity in the reproductive tract. Nat. Med. 11: 531–7.

Chou, C.-H., Wei, L.-H., Kuo, M.-L., Huang, Y.-J., Lai, K.-P., Chen, C.-A., et al. (2005). Up-regulation of interleukin-6 in human ovarian cancer cell via a Gi/PI3K- Akt/NF-kappaB pathway by lysophosphatidic acid, an ovarian cancer-activating factor. Carcinogenesis. 26: 45–52.

Chu, Y.-S., Eder, O., Thomas, W. a, Simcha, I., Pincet, F., Ben-Ze’ev, A., et al. (2006). Prototypical type I E-cadherin and type II cadherin-7 mediate very distinct adhesiveness through their extracellular domains. J. Biol. Chem. 281: 2901–10.

Coffin, J., Hughes, S., Varmus, H., and Editors (1997). Principles of Retroviral Vector Design. In Retroviruses, (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press),.

Cramer, D.W., Bast, R.C., Berg, C.D., Diamandis, E.P., Godwin, A.K., Hartge, P., et al. (2011). Ovarian cancer biomarker performance in prostate, lung, colorectal,

125

and ovarian cancer screening trial specimens. Cancer Prev. Res. (Phila). 4: 365– 74.

Cruet-Hennequart, S., Maubant, S., Luis, J., Gauduchon, P., Staedel, C., and Dedhar, S. (2003). alpha(v) integrins regulate cell proliferation through integrin- linked kinase (ILK) in ovarian cancer cells. Oncogene. 22: 1688–702.

Crum, C.P., Drapkin, R., Miron, A., Ince, T.A., Muto, M., Kindelberger, D.W., et al. (2007). The distal fallopian tube: a new model for pelvic serous carcinogenesis. Curr. Opin. Obstet. Gynecol. 19: 3–9.

D’Alessio, S., Ferrari, G., Cinnante, K., Scheerer, W., Galloway, A.C., Roses, D.F., et al. (2008). Tissue inhibitor of metalloproteinases-2 binding to membrane- type 1 matrix metalloproteinase induces MAPK activation and cell growth by a non-proteolytic mechanism. J. Biol. Chem. 283: 87–99.

Damjanov, I. (1989). Ovarian tumours in laboratory and domestic animals. Curr. Top. Pathol. 78: 1–10.

Davidowitz, R.A., Selfors, L.M., Iwanicki, M.P., Elias, K.M., Karst, A., Piao, H., et al. (2014). Mesenchymal gene program-expressing ovarian cancer spheroids exhibit enhanced mesothelial clearance. J. Clin. Invest. 124: 2611–25.

Davidson, B., Goldberg, I., Gotlieb, W.H., Kopolovic, J., Risberg, B., Ben-Baruch, G., et al. (2003). Coordinated expression of integrin subunits, matrix metalloproteinases (MMP), angiogenic genes and Ets transcription factors in advanced-stage ovarian carcinoma: a possible activation pathway? Cancer Metastasis Rev. 22: 103–15.

Davis, H.M., Zurawski, V.R., Bast, R.C., and Klug, T.L. (1986). Characterization of the CA 125 antigen associated with human epithelial ovarian carcinomas. Cancer Res. 46: 6143–8.

Dedhar, S. (2000). Cell-substrate interactions and signaling through ILK. Curr. Opin. Cell Biol. 12: 250–6.

Desgrosellier, J.S., and Cheresh, D. a (2010). Integrins in cancer: biological implications and therapeutic opportunities. Nat. Rev. Cancer. 10: 9–22.

Devine, P.L., McGuckin, M.A., and Ward, B.G. (1992). Circulating mucins as tumor markers in ovarian cancer (review). Anticancer Res. 12: 709–17.

126

Dharmaraj, N., Chapela, P.J., Morgado, M., Hawkins, S.M., Lessey, B. a, Young, S.L., et al. (2014). Expression of the transmembrane mucins, MUC1, MUC4 and MUC16, in normal endometrium and in endometriosis. Hum. Reprod. 29: 1730– 8.

Dogru, M., Matsumoto, Y., Okada, N., Igarashi, A., Fukagawa, K., Shimazaki, J., et al. (2008). Alterations of the ocular surface epithelial MUC16 and goblet cell MUC5AC in patients with atopic keratoconjunctivitis. Allergy. 63: 1324–34.

Doi, T., Maniwa, Y., Tanaka, Y., Tane, S., Hashimoto, S., Ohno, Y., et al. (2011). MT1-MMP plays an important role in an invasive activity of malignant pleural mesothelioma cell. Exp. Mol. Pathol. 90: 91–6.

Dubeau, L. (2008). The cell of origin of ovarian epithelial tumours. Lancet Oncol. 9: 1191–7.

Durbin, A.D., Hannigan, G.E., and Malkin, D. (2009). Oncogenic ILK, tumor suppression and all that JNK. Cell Cycle. 8: 4060–6.

Egeblad, M., and Werb, Z. (2002). New functions for the matrix metalloproteinases in cancer progression. Nat. Rev. Cancer. 2: 161–74.

Eiró, N. (2013). Clinical Relevance of Matrix Metalloproteases and their Inhibitors in Breast Cancer. J. Carcinog. Mutagen. S13:

Eke, I., Hehlgans, S., and Cordes, N. (2009). There’s something about ILK. Int. J. Radiat. Biol. 85: 929–36.

Ellerbroek, S.M., Fishman, D.A., Kearns, A.S., Bafetti, L.M., and Stack, M.S. (1999). Ovarian carcinoma regulation of matrix metalloproteinase-2 and membrane type 1 matrix metalloproteinase through beta1 integrin. Cancer Res. 59: 1635–41.

Ellerbroek, S.M., Wu, Y.I., Overall, C.M., and Stack, M.S. (2001). Functional interplay between type I collagen and cell surface matrix metalloproteinase activity. J. Biol. Chem. 276: 24833–42.

Elmasri, W.M., Casagrande, G., Hoskins, E., Kimm, D., and Kohn, E.C. (2009). Cell adhesion in ovarian cancer. Cancer Treat. Res. 149: 297–318.

127

Fang, X., Yu, S., Bast, R.C., Liu, S., Xu, H.-J., Hu, S.-X., et al. (2004). Mechanisms for lysophosphatidic acid-induced cytokine production in ovarian cancer cells. J. Biol. Chem. 279: 9653–61.

Fathalla, M.F. (1971). Incessant ovulation--a factor in ovarian neoplasia? Lancet. 2: 163.

Fedorov, Y., Anderson, E.M., Birmingham, A., Reynolds, A., Karpilow, J., Robinson, K., et al. (2006). Off-target effects by siRNA can induce toxic phenotype. RNA. 12: 1188–96.

Fisher, K.E., Pop, A., Koh, W., Anthis, N.J., Saunders, W.B., and Davis, G.E. (2006). Tumor cell invasion of collagen matrices requires coordinate lipid agonist- induced G-protein and membrane-type matrix metalloproteinase-1-dependent signaling. Mol. Cancer. 5: 69.

Fishman, D.A., Liu, Y., and Ellerbroek, S.M. (2001). Lysophosphatidic Acid Promotes Matrix Metalloproteinase ( MMP ) Activation and MMP-dependent Invasion in Ovarian Cancer Cells Lysophosphatidic Acid Promotes Matrix Metalloproteinase ( MMP ) Activation and MMP-dependent Invasion in Ovarian Cancer Cells 1. 3194–3199.

Freije, J.M.P., Balbín, M., Pendás, A.M., Sánchez, L.M., Puente, X.S., and López-Otín, C. (2003). Matrix metalloproteinases and tumor progression. Adv. Exp. Med. Biol. 532: 91–107.

Fu, G., Wang, W., and Luo, B.-H. (2012). Overview: structural biology of integrins. Methods Mol. Biol. 757: 81–99.

Gardner, M.J., Jones, L.M., Catterall, J.B., and Turner, G.A. (1995). Expression of cell adhesion molecules on ovarian tumour cell lines and mesothelial cells, in relation to ovarian cancer metastasis. Cancer Lett. 91: 229–34.

Ghatak, S., Morgner, J., and Wickström, S.A. (2013). ILK: a pseudokinase with a unique function in the integrin-actin linkage. Biochem. Soc. Trans. 41: 995–1001.

Gilcrease, M.Z. (2007). Integrin signaling in epithelial cells. Cancer Lett. 247: 1– 25.

Gipson, I.K., Spurr-Michaud, S., Tisdale, A., and Menon, B.B. (2014). Comparison of the Transmembrane Mucins MUC1 and MUC16 in Epithelial Barrier Function. PLoS One. 9: e100393.

128

Giuntoli, R.L., Rodriguez, G.C., Whitaker, R.S., Dodge, R., and Voynow, J.A. (1998). Mucin gene expression in ovarian cancers. Cancer Res. 58: 5546–50.

Goodell, C.A., Belisle, J. a, Gubbels, J.A., Migneault, M., Rancourt, C., Connor, J., et al. (2009). Characterization of the tumor marker muc16 (ca125) expressed by murine ovarian tumor cell lines and identification of a panel of cross-reactive monoclonal antibodies. J. Ovarian Res. 2: 8.

Gubbels, J. a a, Belisle, J., Onda, M., Rancourt, C., Migneault, M., Ho, M., et al. (2006). Mesothelin-MUC16 binding is a high affinity, N-glycan dependent interaction that facilitates peritoneal metastasis of ovarian tumors. Mol. Cancer. 5: 50.

Gubbels, J.A., Claussen, N., Kapur, A.K., Connor, J.P., and Patankar, M.S. (2010). The detection, treatment, and biology of epithelial ovarian cancer. J. Ovarian Res. 3: 8.

Gumbiner, B.M. (1996). Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell. 84: 345–57.

Gupta, D., Lammersfeld, C.A., Vashi, P.G., and Braun, D.P. (2010). Longitudinal monitoring of CA125 levels provides additional information about survival in ovarian cancer. J. Ovarian Res. 3: 22.

Gupta, D., and Lis, C.G. (2009). Role of CA125 in predicting ovarian cancer survival - a review of the epidemiological literature. J. Ovarian Res. 2: 13.

Hanks, S.K., and Hunter, T. (1995). Protein kinases 6. The eukaryotic protein kinase superfamily: kinase (catalytic) domain structure and classification. FASEB J. 9: 576–96.

Hannigan, G., Troussard, A.A., and Dedhar, S. (2005). Integrin-linked kinase: a cancer therapeutic target unique among its ILK. Nat. Rev. Cancer. 5: 51–63.

Hannigan, G.E., Leung-Hagesteijn, C., Fitz-Gibbon, L., Coppolino, M.G., Radeva, G., Filmus, J., et al. (1996). Regulation of cell adhesion and anchorage- dependent growth by a new beta 1-integrin-linked protein kinase. Nature. 379: 91–6.

Hannigan, G.E., McDonald, P.C., Walsh, M.P., and Dedhar, S. (2011). Integrin- linked kinase: not so “pseudo” after all. Oncogene. 30: 4375–85.

129

Hattrup, C.L., and Gendler, S.J. (2008). Structure and function of the cell surface (tethered) mucins. Annu. Rev. Physiol. 70: 431–57.

Hillier, S.G. (2012). Nonovarian origins of ovarian cancer. Proc. Natl. Acad. Sci. U. S. A. 109: 3608–9.

Hollingsworth, M. a, and Swanson, B.J. (2004). Mucins in cancer: protection and control of the cell surface. Nat. Rev. Cancer. 4: 45–60.

Holmbeck, K., Bianco, P., Caterina, J., Yamada, S., Kromer, M., Kuznetsov, S. a, et al. (1999). MT1-MMP-deficient mice develop dwarfism, osteopenia, arthritis, and connective tissue disease due to inadequate collagen turnover. Cell. 99: 81– 92.

Honda, S., Shirotani-Ikejima, H., Tadokoro, S., Tomiyama, Y., and Miyata, T. (2013). The integrin-linked kinase-PINCH-parvin complex supports integrin αIIbβ3 activation. PLoS One. 8: e85498.

Huang, M.-C., Lee, H.-Y., Yeh, C.-C., Kong, Y., Zaloudek, C.J., and Goetzl, E.J. (2004). Induction of protein growth factor systems in the ovaries of transgenic mice overexpressing human type 2 lysophosphatidic acid G protein-coupled receptor (LPA2). Oncogene. 23: 122–9.

Huber, O., Bierkamp, C., and Kemler, R. (1996). Cadherins and catenins in development. Curr. Opin. Cell Biol. 8: 685–91.

Hudson, L.G., Moss, N.M., and Stack, M.S. (2009). EGF-receptor regulation of matrix metalloproteinases in epithelial ovarian carcinoma. Future Oncol. 5: 323– 38.

Hudson, L.G., Zeineldin, R., and Stack, M.S. (2008). Phenotypic plasticity of neoplastic ovarian epithelium: unique cadherin profiles in tumor progression. Clin. Exp. Metastasis. 25: 643–55.

Humphries, M.J. (1996). Integrin activation: the link between ligand binding and signal transduction. Curr. Opin. Cell Biol. 8: 632–40.

Husseinzadeh, N. (2011). Status of tumor markers in epithelial ovarian cancer has there been any progress? A review. Gynecol. Oncol. 120: 152–7.

Hynes, R.O. (1992). Integrins: versatility, modulation, and signaling in cell adhesion. Cell. 69: 11–25.

130

Iwanicki, M.P., Davidowitz, R. a, Ng, M.R., Besser, A., Muranen, T., Merritt, M., et al. (2011). Ovarian cancer spheroids use myosin-generated force to clear the mesothelium. Cancer Discov. 1: 144–57.

Jackson, A.L., and Linsley, P.S. (2004). Noise amidst the silence: off-target effects of siRNAs? Trends Genet. 20: 521–4.

Jemal, A., Bray, F., and Ferlay, J. (2011). Global Cancer Statistics. 61: 69–90.

Jeon, J.M., Lee, H.W., Park, J.Y., Jung, H.R., Hwang, I., Kwon, S.Y., et al. (2010). Expression of MUC1 and MUC4 and Its Prognostic Significance in Non- Small Cell Lung Carcinoma. Korean J. Pathol. 44: 397.

Kaimal, R., Aljumaily, R., Tressel, S.L., Pradhan, R. V, Covic, L., Kuliopulos, A., et al. (2013). Selective blockade of matrix metalloprotease-14 with a monoclonal antibody abrogates invasion, angiogenesis, and tumor growth in ovarian cancer. Cancer Res. 73: 2457–67.

Kalra, J., Warburton, C., Fang, K., Edwards, L., Daynard, T., Waterhouse, D., et al. (2009). QLT0267, a small molecule inhibitor targeting integrin-linked kinase (ILK), and docetaxel can combine to produce synergistic interactions linked to enhanced cytotoxicity, reductions in P-AKT levels, altered F-actin architecture and improved treatment outc. Breast Cancer Res. 11: R25.

Kamat, A. a, Fletcher, M., Gruman, L.M., Mueller, P., Lopez, A., Landen, C.N., et al. (2006). The clinical relevance of stromal matrix metalloproteinase expression in ovarian cancer. Clin. Cancer Res. 12: 1707–14.

Kaneko, O., Gong, L., Zhang, J., Hansen, J.K., Hassan, R., Lee, B., et al. (2009). A binding domain on mesothelin for CA125/MUC16. J. Biol. Chem. 284: 3739– 49.

Keely, P., Parise, L., and Juliano, R. (1998). Integrins and GTPases in tumour cell growth, motility and invasion. Trends Cell Biol. 8: 101–6.

Kelm, J.M., Timmins, N.E., Brown, C.J., Fussenegger, M., and Nielsen, L.K. (2003). Method for generation of homogeneous multicellular tumor spheroids applicable to a wide variety of cell types. Biotechnol. Bioeng. 83: 173–80.

Kessenbrock, K., Plaks, V., and Werb, Z. (2010). Matrix metalloproteinases: regulators of the tumor microenvironment. Cell. 141: 52–67.

131

Kim, G., Davidson, B., Henning, R., Wang, J., Yu, M., Annunziata, C., et al. (2012a). Adhesion molecule protein signature in ovarian cancer effusions is prognostic of patient outcome. Cancer. 118: 1543–53.

Kim, J., Coffey, D.M., Creighton, C.J., Yu, Z., Hawkins, S.M., and Matzuk, M.M. (2012b). High-grade serous ovarian cancer arises from fallopian tube in a mouse model. Proc. Natl. Acad. Sci. U. S. A. 109: 3921–6.

Kim, Y., Lee, S., Ye, S., Lee, J.W., Yb, K., Sy, L., et al. (2007). Epigenetic regulation of integrin-linked kinase expression depending on adhesion of gastric carcinoma cells.

Kiyasu, Y., Kaneshima, S., and Koga, S. (1981). Morphogenesis of peritoneal metastasis in human gastric cancer. Cancer Res. 41: 1236–9.

Koga, S., Kudo, H., Kiyasu, Y., Kaneshima, S., Iitsuka, Y., Takeuchi, T., et al. (1980). A scanning electron microscopic study on the peritoneal implantation of ascites hepatoma AH100B cells in rats. Gann. 71: 8–13.

Kovalevich, J., Tracy, B., and Langford, D. (2013). PINCH: MOre than just an adaptor protein in cellular response. J. Cell Physiol. 226: 940–947.

Kufe, D.W. (2009). Mucins in cancer: function, prognosis and therapy. Nat. Rev. Cancer. 9: 874–85.

Kui Wong, N., Easton, R.L., Panico, M., Sutton-Smith, M., Morrison, J.C., Lattanzio, F. a, et al. (2003). Characterization of the oligosaccharides associated with the human ovarian tumor marker CA125. J. Biol. Chem. 278: 28619–34.

Kumar, A.S., Naruszewicz, I., Wang, P., Leung-Hagesteijn, C., and Hannigan, G.E. (2004). ILKAP regulates ILK signaling and inhibits anchorage-independent growth. Oncogene. 23: 3454–61.

Kurman, R., and Shih, le-M. (2010). The Origin and Pathogenesis of Epithelial Ovarian Cancer - A Proposed Unifying Theory. Am. J. Surg. Pathol. 34: 433–443.

Kyriakis, J.M. (2014). In the beginning, there was protein phosphorylation. J. Biol. Chem. 289: 9460–2.

Lakshmanan, I., Ponnusamy, M.P., Das, S., Chakraborty, S., Haridas, D., Mukhopadhyay, P., et al. (2012). MUC16 induced rapid G2/M transition via

132

interactions with JAK2 for increased proliferation and anti-apoptosis in breast cancer cells. Oncogene. 31: 805–17.

Lengyel, E. (2010). Ovarian cancer development and metastasis. Am. J. Pathol. 177: 1053–64.

Lengyel, E., Burdette, J.E., Kenny, H. a, Matei, D., Pilrose, J., Haluska, P., et al. (2013). Epithelial ovarian cancer experimental models. Oncogene. 1–15.

Lessan, K., Aguiar, D.J., Oegema, T., Siebenson, L., and Skubitz, A.P. (1999). CD44 and beta1 integrin mediate ovarian carcinoma cell adhesion to peritoneal mesothelial cells. Am. J. Pathol. 154: 1525–37.

Leung-Hagesteijn, C., Mahendra, A., Naruszewicz, I., and Hannigan, G.E. (2001). Modulation of integrin signal transduction by ILKAP, a protein phosphatase 2C associating with the integrin-linked kinase, ILK1. EMBO J. 20: 2160–70.

Li, F., Zhang, Y., and Wu, C. (1999). Integrin-linked kinase is localized to cell- matrix focal adhesions but not cell-cell adhesion sites and the focal adhesion localization of integrin-linked kinase is regulated by the PINCH-binding ANK repeats. J. Cell Sci. 112 ( Pt 2: 4589–99.

Li, J., Abushahin, N., Pang, S., Xiang, L., Chambers, S.K., Fadare, O., et al. (2011). Tubal origin of “ovarian” low-grade serous carcinoma. Mod. Pathol. 24: 1488–99.

Li, Q., Li, C., Zhang, Y.-Y., Chen, W., Lv, J.-L., Sun, J., et al. (2013). Silencing of integrin-linked kinase suppresses in vivo tumorigenesis of human ovarian carcinoma cells. Mol. Med. Rep. 7: 1050–4.

Lin, S.-W., Ke, F.-C., Hsiao, P.-W., Lee, P.-P., Lee, M.-T., and Hwang, J.-J. (2007). Critical involvement of ILK in TGFbeta1-stimulated invasion/migration of human ovarian cancer cells is associated with urokinase plasminogen activator system. Exp. Cell Res. 313: 602–13.

Liu, Q., Xiao, L., Yuan, D., Shi, X., and Li, P. (2012a). Silencing of the integrin- linked kinase gene induces the apoptosis in ovarian carcinoma. J. Recept. Signal Transduct. Res. 32: 120–7.

133

Liu, Y., Burkhalter, R., Symowicz, J., Chaffin, K., Ellerbroek, S., and Stack, M.S. (2012b). Lysophosphatidic Acid disrupts junctional integrity and epithelial cohesion in ovarian cancer cells. J. Oncol. 2012: 501492.

Lloyd, K.O., Yin, B.W., and Kudryashov, V. (1997). Isolation and characterization of ovarian cancer antigen CA 125 using a new monoclonal antibody (VK-8): identification as a mucin-type molecule. Int. J. Cancer. 71: 842–50.

Löffek, S., Schilling, O., and Franzke, C.-W. (2011). Biological role of matrix metalloproteinases: a critical balance. Eur. Respir. J. 38: 191–208.

Longuespée, R., Boyon, C., Desmons, A., Vinatier, D., Leblanc, E., Farré, I., et al. (2012). Ovarian cancer molecular pathology. Cancer Metastasis Rev. 31: 713–32.

Mai, P.L., Wentzensen, N., and Greene, M.H. (2011). Challenges related to developing serum-based biomarkers for early ovarian cancer detection. Cancer Prev. Res. (Phila). 4: 303–6.

Manning, G., Whyte, D.B., Martinez, R., Hunter, T., and Sudarsanam, S. (2002). The protein kinase complement of the human genome. Science. 298: 1912–34.

Marotta, A., Parhar, K., Owen, D., Dedhar, S., and Salh, B. (2003). Characterisation of integrin-linked kinase signalling in sporadic human colon cancer. Br. J. Cancer. 88: 1755–62.

Marth, C., Egle, D., Auer, D., Rössler, J., Zeimet, A.G., Vergote, I., et al. (2007). Modulation of CA-125 tumor marker shedding in ovarian cancer cells by erlotinib or cetuximab. Gynecol. Oncol. 105: 716–21.

Marth, C., Fuith, L.C., Böck, G., and Dapunt, O. (1989). Modulation of Ovarian Carcinoma Tumor Marker CA-125 by γ Modulation of Ovarian Carcinoma Tumor Marker CA-125 by 7-Interferon. 6538–6542.

Martin, J., Yung, S., Robson, R.L., Steadman, R., and Davies, M. (2000). Production and regulation of matrix metalloproteinases and their inhibitors by human peritoneal mesothelial cells. Perit. Dial. Int. 20: 524–33.

Maydan, M., McDonald, P.C., Sanghera, J., Yan, J., Rallis, C., Pinchin, S., et al. (2010). Integrin-linked kinase is a functional Mn2+-dependent protein kinase that regulates glycogen synthase kinase-3β (GSK-3beta) phosphorylation. PLoS One. 5: e12356.

134

McCluggage, W.G., and Wilkinson, N. (2005). Metastatic neoplasms involving the ovary: a review with an emphasis on morphological and immunohistochemical features. Histopathology. 47: 231–47.

McDonald, P.C., Fielding, A.B., and Dedhar, S. (2008). Integrin-linked kinase-- essential roles in physiology and cancer biology. J. Cell Sci. 121: 3121–32.

Menczer, J. (2013). Andrology & Gynecology : Current Research The Significance of Normal Pretreatment Levels of Ca125 ( < 35 U / Ml ) in Epithelial Ovarian Carcinoma. 22: 9–11.

Miyamoto, S., Katz, B.Z., Lafrenie, R.M., and Yamada, K.M. (1998). Fibronectin and integrins in cell adhesion, signaling, and morphogenesis. Ann. N. Y. Acad. Sci. 857: 119–29.

Moniaux, N., Escande, F., Porchet, N., Aubert, J.P., and Batra, S.K. (2001). Structural organization and classification of the human mucin genes. Front. Biosci. 6: D1192–206.

Monniaux, D., Huet-Calderwood, C., Bellego, F. Le, Fabre, S., Monget, P., and Calderwood, D.A. (2006). Integrins in the ovary. Semin. Reprod. Med. 24: 251– 61.

Moore, R.G., MacLaughlan, S., and Bast, R.C. (2010). Current state of biomarker development for clinical application in epithelial ovarian cancer. Gynecol. Oncol. 116: 240–5.

Mori, H., Lo, A.T., Inman, J.L., Alcaraz, J., Ghajar, C.M., Mott, J.D., et al. (2013). Transmembrane/cytoplasmic, rather than catalytic, domains of Mmp14 signal to MAPK activation and mammary branching morphogenesis via binding to integrin β1. Development. 140: 343–52.

Moser, T.L., Pizzo, S. V, Bafetti, L.M., Man, D.A.F., and Sharon, M. (1996). Evidence for preferential adhesion of ovarian epithelial carcinoma cells to type I collagen mediate by the 012Pl TO TYPE I COLLAGEN MEDIATED BY THE 012Pl. 701: 695–701.

Moss, N.M., Barbolina, M. V, Liu, Y., Sun, L., Munshi, H.G., and Stack, M.S. (2009a). Ovarian cancer cell detachment and multicellular aggregate formation are regulated by membrane type 1 matrix metalloproteinase: a potential role in I.p. metastatic dissemination. Cancer Res. 69: 7121–9.

135

Moss, N.M., Wu, Y.I., Liu, Y., Munshi, H.G., and Stack, M.S. (2009b). Modulation of the membrane type 1 matrix metalloproteinase cytoplasmic tail enhances tumor cell invasion and proliferation in three-dimensional collagen matrices. J. Biol. Chem. 284: 19791–9.

Muranyi, A.L., Dedhar, S., and Hogge, D.E. (2009). Combined inhibition of integrin linked kinase and FMS-like tyrosine kinase 3 is cytotoxic to acute myeloid leukemia progenitor cells. Exp. Hematol. 37: 450–60.

N Howlader, Noone, A., Krapcho, M., Garshell, J., Neyman, N., Altekruse, S., et al. (2013). SEER Cancer Statistics Review (CSR) 1975-2010 (Bethesda, MD).

Nagata, K., Horinouchi, M., Saitou, M., Higashi, M., Nomoto, M., Goto, M., et al. (2007). Mucin expression profile in pancreatic cancer and the precursor lesions. J. Hepatobiliary. Pancreat. Surg. 14: 243–54.

Naito, Y., and Ui-Tei, K. (2013). Designing functional siRNA with reduced off- target effects. Methods Mol. Biol. 942: 57–68.

Nakayama, K., Nakayama, N., Katagiri, H., and Miyazaki, K. (2012). Mechanisms of ovarian cancer metastasis: biochemical pathways. Int. J. Mol. Sci. 13: 11705– 17.

Nguyen, D.X., Bos, P.D., and Massagué, J. (2009). Metastasis: from dissemination to organ-specific colonization. Nat. Rev. Cancer. 9: 274–84.

Nhieu, G.T. Van, Krukonis, E.S., Reszka, A.A., Horwitz, A.F., and Isberg, R.R. (1996). Mutations in the cytoplasmic domain of the integrin beta1 chain indicate a role for endocytosis factors in bacterial internalization. J. Biol. Chem. 271: 7665– 72.

Niedbala, M.J., Crickard, K., and Bernacki, R.J. (1985). Interactions of human ovarian tumor cells with human mesothelial cells grown on extracellular matrix. An in vitro model system for studying tumor cell adhesion and invasion. Exp. Cell Res. 160: 499–513.

Niv, Y. (2000). Mucin and Colorectal Cancer. 775–777.

O’Toole, T.E., Katagiri, Y., Faull, R.J., Peter, K., Tamura, R., Quaranta, V., et al. (1994). Integrin cytoplasmic domains mediate inside-out signal transduction. J. Cell Biol. 124: 1047–59.

136

Patankar, M.S., Jing, Y., Morrison, J.C., Belisle, J. a, Lattanzio, F. a, Deng, Y., et al. (2005). Potent suppression of natural killer cell response mediated by the ovarian tumor marker CA125. Gynecol. Oncol. 99: 704–13.

Persad, S., and Dedhar, S. (2003). The role of integrin-linked kinase (ILK) in cancer progression. Cancer Metastasis Rev. 22: 375–84.

Persengiev, S.P., Zhu, X., and Green, M.R. (2004). Nonspecific, concentration- dependent stimulation and repression of mammalian gene expression by small interfering RNAs (siRNAs). RNA. 10: 12–8.

Poincloux, R., Lizárraga, F., and Chavrier, P. (2009). Matrix invasion by tumour cells: a focus on MT1-MMP trafficking to invadopodia. J. Cell Sci. 122: 3015–24.

Pytliak, M., Vargová, V., and Mechírová, V. (2012). Matrix metalloproteinases and their role in oncogenesis: a review. Onkologie. 35: 49–53.

Qin, J., and Wu, C. (2012). ILK: a pseudokinase in the center stage of cell-matrix adhesion and signaling. Curr. Opin. Cell Biol. 24: 607–13.

Rachagani, S., Torres, M.P., Kumar, S., Haridas, D., Baine, M., Macha, M.A., et al. (2012). Mucin (Muc) expression during pancreatic cancer progression in spontaneous mouse model: potential implications for diagnosis and therapy. J. Hematol. Oncol. 5: 68.

Rachagani, S., Torres, M.P., Moniaux, N., and Batra, S.K. (2009). Current status of mucins in the diagnosis and therapy of cancer. Biofactors. 35: 509–27.

Radeva, G. (1997). Overexpression of the Integrin-linked Kinase Promotes Anchorage-independent Cell Cycle Progression. J. Biol. Chem. 272: 13937– 13944.

Ren, J., Agata, N., Chen, D., Li, Y., Yu, W., Huang, L., et al. (2004). Human MUC1 carcinoma-associated protein confers resistance to genotoxic anticancer agents. Cancer Cell. 5: 163–75.

Ren, J., Xiao, Y., Singh, L.S., Zhao, X., Zhao, Z., Feng, L., et al. (2006). Lysophosphatidic acid is constitutively produced by human peritoneal mesothelial cells and enhances adhesion, migration, and invasion of ovarian cancer cells. Cancer Res. 66: 3006–14.

137

Rodriguez, M., and Dubeau, L. (2001). Ovarian tumor development: insights from ovarian embryogenesis. Eur. J. Gynaecol. Oncol. 22: 175–83.

Roh, M.H., Yassin, Y., Miron, A., Mehra, K.K., Mehrad, M., Monte, N.M., et al. (2010). High-grade fimbrial-ovarian carcinomas are unified by altered p53, PTEN and PAX2 expression. Mod. Pathol. 23: 1316–24.

Rosales, C., O’Brien, V., Kornberg, L., and Juliano, R. (1995). Signal transduction by cell adhesion receptors. Biochim. Biophys. Acta. 1242: 77–98.

Rosanò, L., Spinella, F., Castro, V. Di, Dedhar, S., Nicotra, M.R., Natali, P.G., et al. (2006). Integrin-linked kinase functions as a downstream mediator of endothelin-1 to promote invasive behavior in ovarian carcinoma. Mol. Cancer Ther. 5: 833–42.

Rosen, D.G., Wang, L., Atkinson, J.N., Yu, Y., Lu, K.H., Diamandis, E.P., et al. (2005). Potential markers that complement expression of CA125 in epithelial ovarian cancer. Gynecol. Oncol. 99: 267–77.

Rozanov, D. V, Savinov, A.Y., Golubkov, V.S., Postnova, T.I., Remacle, A., Tomlinson, S., et al. (2004a). Cellular membrane type-1 matrix metalloproteinase (MT1-MMP) cleaves C3b, an essential component of the complement system. J. Biol. Chem. 279: 46551–7.

Rozanov, D. V, Sikora, S., Godzik, A., Postnova, T.I., Golubkov, V., Savinov, A., et al. (2004b). Non-proteolytic, receptor/ligand interactions associate cellular membrane type-1 matrix metalloproteinase with the complement component C1q. J. Biol. Chem. 279: 50321–8.

Rump, A., Morikawa, Y., Tanaka, M., Minami, S., Umesaki, N., Takeuchi, M., et al. (2004). Binding of ovarian cancer antigen CA125/MUC16 to mesothelin mediates cell adhesion. J. Biol. Chem. 279: 9190–8.

Ruoslahti, E. (1991). Integrins. J. Clin. Invest. 87: 1–5.

Said, N. a., Najwer, I., Socha, M.J., Fulton, D.J., Mok, S.C., and Motamed, K. (2007a). SPARC Inhibits LPA-Mediated Mesothelial—Ovarian Cancer Cell Crosstalk. Neoplasia. 9: 23–35.

Said, N., Socha, M.J., Olearczyk, J.J., Elmarakby, A. a, Imig, J.D., and Motamed, K. (2007b). Normalization of the ovarian cancer microenvironment by SPARC. Mol. Cancer Res. 5: 1015–30.

138

Sakamoto, T., and Seiki, M. (2009). Cytoplasmic tail of MT1-MMP regulates macrophage motility independently from its protease activity. Genes Cells. 14: 617–26.

Sakata, K., Shigemasa, K., Nagai, N., and Ohama, K. (2000). Expression of matrix metalloproteinases (MMP-2, MMP-9, MT1-MMP) and their inhibitors (TIMP-1, TIMP-2) in common epithelial tumors of the ovary. Int. J. Oncol. 17: 673–81.

Saxena, S., Jónsson, Z.O., and Dutta, A. (2003). Small RNAs with imperfect match to endogenous mRNA repress translation. Implications for off-target activity of small inhibitory RNA in mammalian cells. J. Biol. Chem. 278: 44312–9.

Schaefer, A., Nethe, M., and Hordijk, P.L. (2012). Ubiquitin links to cytoskeletal dynamics, cell adhesion and migration. Biochem. J. 442: 13–25.

Schmidt, C. (2011). CA-125: a biomarker put to the test. J. Natl. Cancer Inst. 103: 1290–1.

Scholler, N., Garvik, B., Hayden-Ledbetter, M., Kline, T., and Urban, N. (2007). Development of a CA125-mesothelin cell adhesion assay as a screening tool for biologics discovery. Cancer Lett. 247: 130–6.

Scholler, N., and Urban, N. (2007). CA125 in ovarian cancer. Biomark. Med. 1: 513–23.

Scott, R.W., and Olson, M.F. (2007). LIM kinases: function, regulation and association with human disease. J. Mol. Med. (Berl). 85: 555–68.

Siegel, R., Naishadham, D., and Jemal, A. (2013). Cancer statistics, 2013. CA. Cancer J. Clin. 63: 11–30.

Skubitz, A.P.N. (2002). Adhesion molecules. Cancer Treat. Res. 107: 305–29.

Sodek, K.L., Ringuette, M.J., and Brown, T.J. (2007). MT1-MMP is the critical determinant of matrix degradation and invasion by ovarian cancer cells. Br. J. Cancer. 97: 358–67.

Sounni, N.E., Devy, L., Hajitou, A., Frankenne, F., Munaut, C., Gilles, C., et al. (2002). MT1-MMP expression promotes tumor growth and angiogenesis through an up-regulation of vascular endothelial growth factor expression. FASEB J. 16: 555–64.

139

Sounni, N.E., Rozanov, D. V, Remacle, A.G., Golubkov, V.S., Noel, A., and Strongin, A.Y. (2010). Timp-2 binding with cellular MT1-MMP stimulates invasion-promoting MEK/ERK signaling in cancer cells. Int. J. Cancer. 126: 1067–78.

Strauss, R., Li, Z.-Y., Liu, Y., Beyer, I., Persson, J., Sova, P., et al. (2011). Analysis of epithelial and mesenchymal markers in ovarian cancer reveals phenotypic heterogeneity and plasticity. PLoS One. 6: e16186.

Strobel, T., and Cannistra, S. a (1999). Beta1-integrins partly mediate binding of ovarian cancer cells to peritoneal mesothelium in vitro. Gynecol. Oncol. 73: 362– 7.

Strongin, A.Y. (2010). Proteolytic and non-proteolytic roles of membrane type-1 matrix metalloproteinase in malignancy. Biochim. Biophys. Acta. 1803: 133–41.

Sutphen, R., Xu, Y., Wilbanks, G.D., Fiorica, J., Grendys, E.C., LaPolla, J.P., et al. (2004). Lysophospholipids are potential biomarkers of ovarian cancer. Cancer Epidemiol. Biomarkers Prev. 13: 1185–91.

Symowicz, J., Adley, B.P., Woo, M.M.M., Auersperg, N., Hudson, L.G., and Stack, M.S. (2005). Cyclooxygenase-2 Functions as a Downstream Mediator of Lysophosphatidic Acid to Promote Aggressive Behavior in Ovarian Carcinoma Cells Cyclooxygenase-2 Functions as a Downstream Mediator of Lysophosphatidic Acid to Promote Aggressive Behavior in Ovarian C. 2234– 2242.

Takeichi, M. (1995). Morphogenetic roles of classic cadherins. Curr. Opin. Cell Biol. 7: 619–27.

Taylor, S.S., Keshwani, M.M., Steichen, J.M., and Kornev, A.P. (2012). Evolution of the eukaryotic protein kinases as dynamic molecular switches. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 367: 2517–28.

Taylor, S.S., and Kornev, A.P. (2010). Yet another “active” pseudokinase, Erb3. Proc. Natl. Acad. Sci. U. S. A. 107: 8047–8.

Thériault, C., Pinard, M., Comamala, M., Migneault, M., Beaudin, J., Matte, I., et al. (2011). MUC16 (CA125) regulates epithelial ovarian cancer cell growth, tumorigenesis and metastasis. Gynecol. Oncol. 121: 434–43.

140

Troussard, A. a, McDonald, P.C., Wederell, E.D., Mawji, N.M., Filipenko, N.R., Gelmon, K. a, et al. (2006). Preferential dependence of breast cancer cells versus normal cells on integrin-linked kinase for protein kinase B/Akt activation and cell survival. Cancer Res. 66: 393–403.

Tschuch, C., Schulz, A., Pscherer, A., Werft, W., Benner, A., Hotz-Wagenblatt, A., et al. (2008). Off-target effects of siRNA specific for GFP. BMC Mol. Biol. 9: 60.

Tsunezuka, Y., Kinoh, H., and Takino, T. (1996). Expression of Membrane-type Matrix Metalloproteinase 1 ( MT1-MMP ) in Tumor Cells Enhances Pulmonary Metastasis in an Experimental Metastasis Assay Expression of Membrane-type Matrix Metalloproteinase 1 ( MT1-MMP ) in Tumor Cells Enhances Pulmonary Metast. 1: 5678–5683.

Tuxen, M.K., Sölétormos, G., and Dombernowsky, P. (1995). Tumor markers in the management of patients with ovarian cancer. Cancer Treat. Rev. 21: 215–45.

Vargová, V., Pytliak, M., and Mechírová, V. (2012). Matrix metalloproteinases. EXS. 103: 1–33.

Vergote, I.B., Onsrud, M., Børmer, O.P., Sert, B.M., and Moen, M. (1992). CA125 in peritoneal fluid of ovarian cancer patients. Gynecol. Oncol. 44: 161–5.

Vu, T.H. (2000). Matrix metalloproteinases: effectors of development and normal physiology. Genes Dev. 14: 2123–2133.

Wang, X., Wang, X., Varma, R.K., Beauchamp, L., Magdaleno, S., and Sendera, T.J. (2009). Selection of hyperfunctional siRNAs with improved potency and specificity. Nucleic Acids Res. 37: e152.

Wang, Y., Liu, J., Lin, B., Wang, C., Li, Q., Liu, S., et al. (2011). Study on the Expression and Clinical Significances of Lewis y Antigen and Integrin αv, β3 in Epithelial Ovarian Tumors. Int. J. Mol. Sci. 12: 3409–21.

Weber, G.F., Bjerke, M.A., and DeSimone, D.W. (2011). Integrins and cadherins join forces to form adhesive networks. J. Cell Sci. 124: 1183–93.

Weiland, F., Martin, K., Oehler, M.K., and Hoffmann, P. (2012). Deciphering the Molecular Nature of Ovarian Cancer Biomarker CA125. Int. J. Mol. Sci. 13: 10568–82.

141

Wickström, S. a, Lange, A., Montanez, E., and Fässler, R. (2010). The ILK/PINCH/parvin complex: the kinase is dead, long live the pseudokinase! EMBO J. 29: 281–91.

Widmaier, M., Rognoni, E., Radovanac, K., Azimifar, S.B., and Fässler, R. (2012). Integrin-linked kinase at a glance. J. Cell Sci. 125: 1839–43.

Witz, C.A., Monotoya-Rodriguez, I.A., and Schenken, R.S. (1999). Whole explants of peritoneum and endometrium: a novel model of the early endometriosis lesion. Fertil. Steril. 71: 56–60.

Wu, C. (2004). The PINCH-ILK-parvin complexes: assembly, functions and regulation. Biochim. Biophys. Acta. 1692: 55–62.

Wu, C., and Dedhar, S. (2001). Integrin-linked kinase (ILK) and its interactors: a new paradigm for the coupling of extracellular matrix to actin cytoskeleton and signaling complexes. J. Cell Biol. 155: 505–10.

Wu, Y.I., Munshi, H.G., Snipas, S.J., Salvesen, G.S., Fridman, R., and Stack, M.S. (2007). Activation-coupled membrane-type 1 matrix metalloproteinase membrane trafficking. Biochem. J. 407: 171–7.

Yamada, K.M., and Geiger, B. (1997). Molecular interactions in cell adhesion complexes. Curr. Opin. Cell Biol. 9: 76–85.

Yancik, R. (1993). Ovarian cancer. Age contrasts in incidence, histology, disease stage at diagnosis, and mortality. Cancer. 71: 517–23.

Yao, G., He, P., Chen, L., Hu, X., Gu, F., and Ye, C. (2013). MT1-MMP in breast cancer: induction of VEGF-C correlates with metastasis and poor prognosis. Cancer Cell Int. 13: 98.

Yin, B.W., and Lloyd, K.O. (2001). Molecular cloning of the CA125 ovarian cancer antigen: identification as a new mucin, MUC16. J. Biol. Chem. 276: 27371–5.

Yin, B.W.T., Dnistrian, A., and Lloyd, K.O. (2002). Ovarian cancer antigen CA125 is encoded by the MUC16 mucin gene. Int. J. Cancer. 98: 737–40.

Younes, M.N., Yigitbasi, O.G., Yazici, Y.D., Jasser, S. a, Bucana, C.D., El- Naggar, A.K., et al. (2007). Effects of the integrin-linked kinase inhibitor QLT0267

142

on squamous cell carcinoma of the head and neck. Arch. Otolaryngol. Head. Neck Surg. 133: 15–23.

Zaretsky, J.Z., and Wreschner, D.H. (2013). Mucins – Potential Regulators of Cell Functions: Gel-Forming and Soluble Mucins (Tel Aviv: BENTHAM SCIENCE PUBLISHERS).

Zeimet, a G., Marth, C., Offner, F. a, Obrist, P., Uhl-Steidl, M., Feichtinger, H., et al. (1996). Human peritoneal mesothelial cells are more potent than ovarian cancer cells in producing tumor marker CA-125. Gynecol. Oncol. 62: 384–9.

Zhang, J., Bera, T.K., Liu, W., Du, X., Alewine, C., Hassan, R., et al. (2014). Megakaryocytic potentiating factor and mature mesothelin stimulate the growth of a lung cancer cell line in the peritoneal cavity of mice. PLoS One. 9: e104388.

Zigler, M., Dobroff, A.S., and Bar-Eli, M. (2010). Cell adhesion: implication in tumor progression. Minerva Med. 101: 149–62.

143

VITA

Lana Y. N. Bruney was born in Houston, Texas. Lana is a first generation

American. She spent the first few years of her life in her family’s home country,

Dominica, then returned to the United States. Lana has earned a Bachelor of

Science in Chemistry and a Masters in Health Administration. During her academic career, Lana has competed for and successfully obtained numerous national awards, writing awards, honors certifications, and fellowships. She is currently part of the prestigious Presidential Management Fellows Program and has started a career in government.

144