A Dissertation Entitled the Role of Mdia2 in Adherens Junctions In

Home , MDia1, Plakoglobin, PLEKHA7, SMAP1, Stathmin

ADissertation entitled

The Role of mDia2 in Adherens Junctions in Epithelial Ovarian Cancers

by Yuqi Zhang

Submitted to the Graduate Faculty as partial fulﬁllment of the requirements for the Doctor of Philosophy Degree in Biomedical Sciences

Kathryn Eisenmann, Ph.D., Committee Chair

Rafael Garcia-Mata, Ph.D., Committee Member

Randall Ruch, Ph.D., Committee Member

Ivana de la Serna, Ph.D., Committee Member

Eda Yildirim-Ayan, Ph.D., Committee Member

Dr. Cyndee Gruden, Dean College of Graduate Studies

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of The Role of mDia2 in Adherens Junctions in Epithelial Ovarian Cancers by Yuqi Zhang

Submitted to the Graduate Faculty as partial fulﬁllment of the requirements for the Doctor of Philosophy Degree in Biomedical Sciences The University of Toledo February 2019

Epithelial ovarian cancer (EOC) cells disseminate within the peritoneal cavity, in part, via the peritoneal ﬂuid as single cells, clusters, or spheroids. Initial single cell egress from a tumor can involve disruption of cell-cell adhesions as cells are shed from the primary tumor into the peritoneum. In epithelial cells, Adherens Junc- tions (AJs) are characterized by homotypic linkage of E-cadherins on the plasma membranes of adjacent cells. AJs are anchored to the intracellular actin cytoskeletal network through a complex involving E-cadherin, p120 catenin, β-catenin, and

αE-catenin. However, the speciﬁc players involved in the interaction between the junctional E-cadherin complex and the underlying F-actin network remains unclear.

Recent evidence indicates that mammalian Diaphanous-related (mDia) formins plays a key role in epithelial cell AJ formation and maintenance through generation of linear actin filaments. We previously demonstrated that loss of the formin mDia2 was specifically associated with invasive single cell egress from EOC spheroids through disruption of junctional F-actin. In this work, we show that mDia2 has a role at AJs in ovarian cancer (OVCA) 429 cells and human embryonic kidney (HEK) 293 cells through its association with αE-catenin and β-catenin. mDia2 depletion in EOC cells leads to reduction in actin polymerization and disruption of cell-cell junctions with decreased interaction between β-catenin and E-cadherin. In summary, our findings indicate an essential role for mDia2 in AJ formation and stability in EOC cells. These iii effects are likely achieved through its interactions with and regulation of α-andβ- catenin. Our findings support a novel mechanism for EOC dissemination that should be considered in development of targeted therapy against this deadly disease.

iv Acknowledgments

IwouldliketothankmyadvisorDr.KathrynEisenmannforgivingmetheopportu- nity to work with her and her mentorship over the past several years. Her guidance and assistance were essential to the successful completion of the experiments pre- sented in this work. I also thank my committee members for their helpful insights. I am also grateful for the members of the Eisenmann lab and all my friends and teach- ers in the Cancer Biology department for supplying me with the technical toolbox to execute these experiments. Last but not least, I thank my parents for believing in me all these years and lending me the strength to get to where I am today.

v Contents

Abstract iii

Acknowledgments v

Contents vi

List of Figures xi

List of Abbreviations xiii

1 An Introduction to Epithelial Ovarian Cancer, Adherens Junctions,

and mDia Formins 1

1.1 An Introduction to Epithelial Ovarian Cancer (EOC) and its treatment 1

1.1.1 The epidemiology of EOC ...... 1

1.1.2 Symptoms and Diagnosis ...... 3

1.1.3 Prognosis and Treatment ...... 4

1.1.3.1 TreatmentofchemoresistantEOC ...... 5

1.2 Molecular mechanisms of EOC development and metastasis ...... 6

1.2.1 Molecular characteristics of EOC ...... 6

1.2.1.1 Low-grade serous ovarian cancer ...... 6

1.2.1.2 High-gradeserousovariancancer ...... 7

1.2.2 Origins of EOC ...... 9

1.2.3 Mechanisms of EOC metastasis ...... 11

1.2.3.1 HematogenousDissemination ...... 11

vi 1.2.3.2 Passive Dissemination ...... 12

1.2.3.3 RoleofAscitesinEOCMetastasis ...... 13

1.2.3.4 EOC seeding within the abdominal cavity ...... 14

1.2.3.5 EMT in EOC metastasis ...... 14

1.2.3.6 Potential therapeutic targets in EMT ...... 16

1.2.3.7 EMT and MET in EOC metastasis and chemoresistance 17

1.2.3.8 PartialEMT/METinEOC ...... 18

1.2.3.9 EMTandchemoresistance ...... 19

1.3 EMT and the Adherens Junction (AJ) ...... 20

1.3.1 AJformationandstabilization: anoverview ...... 20

1.3.2 The cytoskeleton in AJ stability and cell motility ...... 23

1.3.2.1 The cytoskeleton is necessary for AJ formation and

stability ...... 23

1.3.2.2 The actin cytoskeleton ...... 24

1.3.2.3 The microtubule cytoskeleton ...... 30

1.3.2.4 Intermediate ﬁlaments ...... 34

1.3.3 MechanismsofJunctionDestabilization...... 34

1.3.4 Cell motility is achieved through regulation of the cytoskeleton 35

1.3.4.1 Overview of membrane protrusions ...... 36

1.3.5 AJ regulators: cadherins, catenins, and vinculin ...... 38

1.3.5.1 Cadherins ...... 38

1.3.5.2 α-catenins ...... 39

1.3.5.3 β-catenin ...... 39

1.3.5.4 p120 catenin ...... 40

1.3.5.5 Vinculin...... 40

1.4 Formin proteins: an introduction ...... 41

1.4.1 Anoverviewofformins...... 41

vii 1.4.2 Structure of mDia formins ...... 41

1.4.3 LocalizationandfunctionofmDiaformins ...... 43

1.4.3.1 Formin function determines their intracellular local-

ization ...... 43

1.4.3.2 ForminsinAJs ...... 45

1.4.3.3 Integrative functions of mDia formins...... 47

1.4.3.4 Regulation by Rho-GTPases ...... 48

1.4.3.5 Dia-interacting protein (DIP) and the cell cycle in

mDia2 regulation ...... 50

1.5 Formins in development and disease ...... 50

1.5.1 Forminsindevelopment ...... 50

1.5.2 Forminsinimmunityandhomeostasis ...... 51

1.5.3 Formins in cancer ...... 52

1.5.4 Formins as therapeutic targets ...... 55

1.5.4.1 Formin inhibition ...... 55

1.5.4.2 Formin activation ...... 57

1.6 SummaryofPastFindingsandGapsintheKnowledge ...... 58

1.7 Hypothesis...... 59

2 Results 61

2.1 Introduction ...... 61

2.2 Results ...... 65

2.2.1 mDia2 is essential for junction integrity in spheroids...... 65

2.2.2 RoleofmDia2inAJformation ...... 67

2.2.3 mDia2 interacts with β-andα-catenin but not E-cadherin . . 69

2.2.4 mDia2 co-precipitates with β-andα-catenin in HEK293 cells 72

2.2.5 mDia2 aﬀects junctional stability ...... 73

viii 2.2.6 mDia2 expression aﬀects interactions between junctional proteins 76

2.2.7 Actin disruption does not inhibit interactions between mDia2

and α-andβ-catenin ...... 78

2.3 Discussion ...... 80

2.4 Conclusions ...... 84

2.5 Methods ...... 85

2.5.1 Cell lines and reagents ...... 85

2.5.2 Western blotting ...... 85

2.5.3 Transfection and knockdown ...... 86

2.5.4 Immunoprecipitation ...... 86

2.5.5 ImmunoﬂuorescenceandImageAnalysis ...... 87

2.5.6 InSituProximityLigationAssay(PLA) ...... 88

2.5.7 HangingDropAssay ...... 89

2.5.8 CalciumSwitchAssay ...... 90

2.5.9 Statistics ...... 90

2.6 Acknowledgements ...... 90

2.7 Funding ...... 91

3 Discussion 92

3.1 mDia2 interacts with both α-andβ-catenin ...... 92

3.2 Biomechanics of EOC ...... 93

3.3 Limitations ...... 93

3.4 Future Directions ...... 94

3.5 Summary of ﬁndings ...... 94

References 96

A Investigation of the respiratory diaphragm as a key site of EOC

ix invasionandmetastasis 149

A.1 Introduction...... 149

A.2 Methods: Development of an ex vivo diaphragm stretch model . . . . 150

A.3 Results...... 151

x List of Figures

1-1 Two models for anchorage of the E-cadherin/catenin complex to F-actin. 22

1-2 Formins mediate linear actin polymerization...... 26

1-3 Arp2/3/mediatesbranchedactinpolymerization...... 27

1-4 mDia formin domains...... 42

2-1 Analysis of mDia2 in a functional cell-cell adhesion assay...... 66

2-2 Quantiﬁcation of functional cell-cell adhesion assay ...... 67

2-3 Eﬀect of mDia2 on adherens junction formation...... 68

2-4 mDia2 interacts with β-catenin and αE-catenin but not E-cadherin. . . . 70

2-5 mDia2 interacts with β-catenin and αE-catenin but not E-cadherin. . . . 71

2-6 mDia2 does not interact with E-cadherin...... 72

2-7 mDia2 co-precipitates with αE- and β-catenininHEK293cells...... 73

2-8 mDia2 expression aﬀects junctional stability...... 74

2-9 Two models for anchorage of the E-cadherin/catenin complex to F-actin. 75

2-10 mDia2 expression aﬀects interactions between junctional proteins. . . . . 76

2-11 Quantiﬁcation of interactions between junctional proteins...... 77

2-12 Actin disruption does not inhibit interactions between mDia2 and α or

βcatenin...... 79

2-13 Quantiﬁcation of interactions between mDia2 and α or βcatenin...... 80

A-1 Spheroid invasion at 48 hours on stretched diaphragm explant...... 151

A-2 Spheroid invasion at 48 hours on unstretched diaphragm explant. . . . . 152

xi A-3 Normalized areas of GFP-SKOV3 invasion on stretched and unstretched

diaphragms...... 153

A-4 GFP-SKOV3 spheroids demonstrate collagen alignment and egress from

the spheroid as single cells and clusters...... 153

A-5 X-Z projections of GFP-SKOV3 spheroids seeded on stretched and un-

stretched diaphragm explants...... 154

xii List of Abbreviations

Arp ...... actin-related proteins AJ ...... adherens junction BD ...... basic domain BMP ...... bone morphogenetic protein CC ...... coiled coil DAD ...... diaphanous auto-regulatory domain DD ...... dimerization domain DID ...... diaphanous inhibitory domain DIP ...... Dia-interacting protein DRFs ...... Diaphanous-related formins EGFR ...... Epidermal Growth Factor Receptor EOC ...... epithelial ovarian cancer ESPs ...... early serous proliferations EMT ...... epithelial-to-mesenchymal FA ...... focal adhesion FSH ...... follicle-stimulating hormone FH ...... formin homology GBM ...... glioblastoma GAP ...... GTPase-activating protein GDI ...... guanine nucleotide displacement inhibitor GEF ...... guanine nucleotide exchange factors GFP ...... Green Flourescent Protein IL ...... interleukins IP ...... intraperitoneal IV ...... intravenous HGF ...... Hepatocyte Growth Factor HGSOC ...... high grade serous ovarian cancers HDAC ...... histone deacetylase LGSOC ...... low grade serous ovarian cancers LH ...... luteinizing hormone LPA ...... lysophosphatidic acid mDia ...... mammalian Diaphanous

xiii MAPK ...... mitogen-activated protein kinase MMP ...... matrix metalloproteinase MET ...... mesenchymal-epithelial-transition N-WASp ...... neuronal Wiskott-Aldrich Syndrome protein STIC ...... serous tubal intraepithelial carcinoma OSE ...... ovarian surface epithelium Rho ...... Ras homolog ROCK ...... RHO-associated Coiled-coil Containing Protein Kinase SRF ...... serum response factor SMIFH2 ...... small molecule inhibitor of the FH2 domain IMM ...... small molecule mDia2 intramimic TGF-β ...... transforminggrowthfactor-β uPA ...... urokinase plasminogen activator VEGF ...... Vascular Endothelial Growth Factor WAVE ...... WASP/WASP family verprolin-homologous protein ZEB1 ...... zinc ﬁnger E-box-binding homeobox ZA ...... zonula adherens

xiv Chapter 1

An Introduction to Epithelial

Ovarian Cancer, Adherens

Junctions, and mDia Formins

1.1 An Introduction to Epithelial Ovarian Cancer

(EOC) and its treatment

1.1.1 The epidemiology of EOC

Ovarian cancer is the deadliest gynecological malignancy, with 14,070 women estimated to die from the disease in the United States in 2018, according to the American

Cancer Society. Worldwide, 239,000 new cases of ovarian cancer and 152,000 related deaths occur annually (Reid, Permuth et al. 2017). There are four types of ovarian cancer classiﬁed by their cell of origin, with epithelial ovarian cancer (EOC) making up more than 90% of diagnosed cases (McCluggage 2011, Prat 2012, Rojas, Hirshﬁeld et al. 2016). EOC can further be divided by histologic subtype (histotypes), from most common to least, into high-grade serous (70%), endometrioid (10%), clear cell

(10%), low-grade serous (<5%), and mucinous (3%) (McCluggage 2011, Prat 2012,

1 Reid, Permuth et al. 2017). EOC is most commonly diagnosed in postmenopausal women, with the highest incidence rates (11.4 per 100,000 and 6.0 per 100,000) observed in Eastern and Central Europe, respectively (Doubeni, Doubeni et al. 2016,

Reid, Permuth et al. 2017). The lowest rates are seen in Asia and Africa ( 3per ≤ 100,000) (Reid, Permuth et al. 2017).

Known risk factors of ovarian cancer include familial genetic syndromes, nullipar- ity, and obesity. The greatest risk factors for ovarian cancer are the familial genetic syndromes, which involve inherited mutations in tumor suppressor genes and account for an estimated 10-12% of ovarian cancer cases (Pruthi, Gostout et al. 2010, Jelo- vac and Armstrong 2011, Doubeni, Doubeni et al. 2016). These include hereditary breast and ovarian cancer syndrome (BRCA1 and BRCA2 ), hereditary nonpolypo- sis colorectal cancer (MLH1, MLH3, MSH2, MSH6, TGFBR2, PMS1,andPMS2 ),

MUTYH-associated polyposis (MUTYH ), Peutz-Jeghers syndrome (STK11 ), and

PTEN hamartoma tumor syndrome (PTEN ) (Pruthi, Gostout et al. 2010, Hunn and Rodriguez 2012, Doubeni, Doubeni et al. 2016). Parity reduces ovarian cancer risk by 30-60% compared to rates in nulliparous women (Casagrande, Louie et al.

1979, Whittemore, Wu et al. 1988, Wu, Whittemore et al. 1988, Reid, Permuth et al. 2017). Meanwhile, obesity is associated with an increased risk of ovarian cancer, especially in pre-menopausal women (Doubeni, Doubeni et al. 2016, Reid, Permuth et al. 2017). In post-menopausal women, greater hip circumference, but not BMI, is associated with increased risk according to the Nurses Health Study (Kotsopoulos,

Baer et al. 2010, Reid, Permuth et al. 2017).

Protective factors against ovarian cancer include use of oral contraceptives (OCPs), salpingectomy, tubal ligation, and exercise. For every 5 years of OCP use, there is a

20% reduced risk of ovarian cancer (Collaborative Group on Epidemiological Studies of Ovarian, Beral et al. 2008, Reid, Permuth et al. 2017). Over the last 5 decades, it is estimated that OCP use has prevented 200,000 ovarian cancer cases (Reid, Per-

2 muth et al. 2017). Salpingectomy (removal of one or both fallopian tubes) and tubal ligation both reduce ovarian cancer risk, with tubal ligation especially reducing risk of endometrioid and clear cell EOC (Doubeni, Doubeni et al. 2016, Wentzensen, Poole et al. 2016, Reid, Permuth et al. 2017). Exercise may protect against ovarian cancer by reducing adipose tissue (and associated estrogen) as well as decreasing chronic inﬂammation (McTiernan 2008, Reid, Permuth et al. 2017).

1.1.2 Symptoms and Diagnosis

The majority of EOC cases (60%) are diagnosed in the late stages due to lack of symptoms early in the disease (Doubeni, Doubeni et al. 2016). Symptoms prior to diagnosis are often non-speciﬁc and include back pain, fatigue, abdominal pain or bloating, constipation, and urinary symptoms (Cannistra 2004). Some may experience paraneoplastic syndromes such as subacute cerebellar degeneration, seborrheic keratosis, and Trousseu syndrome, a condition characterized by migratory venous thromboses (Doubeni, Doubeni et al. 2016). The diagnosis can be made by assess- ing patient history (accounting for personal/family history of cancer), and perform- ing physical examination (potentially revealing pelvic/abdominal masses), transvagi- nal ultrasonography (to visualize the ovaries, diﬀerentiate between cystic and solid masses, and detect ascites), and laboratory testing (Doubeni, Doubeni et al. 2016).

A complete blood count (CBC), liver function tests (LFTs), and calcium are usually assessed (Doubeni, Doubeni et al. 2016). Cancer Antigen (CA)-125 is often used as a biomarker for ovarian cancer, but is only elevated in 50% of patients in the early stages of EOC. Its speciﬁcity is limited as both endometriosis and ﬁbroids can also result in

CA-125 elevation (Cannistra 2004, Clarke-Pearson 2009). Another biomarker, human epididymis protein 4 (HE4), is currently being used to detect disease recurrence and monitor progression after diagnosis (Doubeni, Doubeni et al. 2016). A sensitive and speciﬁc biomarker for early-stage EOC, which would have signiﬁcant implications for

3 disease prognosis, yet remains to be discovered.

1.1.3 Prognosis and Treatment

Survival rates for patients with EOC vary dramatically depending on the stage at diagnosis. According to the International Federation of Gynecology and Obstetrics

(FIGO), Stage I is classiﬁed as the tumor limited to one or both ovaries; Stage II is deﬁned as when the disease extends into the pelvic tissue (e.g., uterus and/or fallopian tubes) with or without malignant cells in peritoneal washings or ascites;

Stage III is characterized by metastasis outside the pelvis and/or regional lymph node involvement; and Stage IV is characterized by metastases to distant organs

(e.g,. liver, spleen, etc.) with or without pleural eﬀusion (Doubeni, Doubeni et al.

2016). Only 15% of cases are diagnosed as Stage I EOC, which is associated with a 92% 5-year survival rate (Reid, Permuth et al. 2017). Meanwhile, Stages III and

IV are associated with 39-59% and 17-28% 5-year survival rates, respectively (Chen,

Ruiz et al. 2003, Jelovac and Armstrong 2011, Allemani, Weir et al. 2015). In one study conducted in Korea of 22,880 women diagnosed with EOC in 1995-1999 and

2010-2014, 5-year relative survival rates only improved by less than 7% between the two time periods (Lee, Kim et al. 2018), reﬂecting the limitations in current available treatments.

Current standard of care treatment for stage III and IV EOC involves cytoreductive surgery and adjuvant platinum/taxane duplet combinations such as carboplatin plus paclitaxel (McGuire, Hoskins et al. 1996, Vasey, Jayson et al. 2004). The route of administration of chemotherapy depends on whether the disease has been optimally

(<1 cm residual disease) or sub-optimally reduced by surgery. While intravenous

(IV) treatment or a combination of IV and intraperitoneal (IP) medication (IV/IP chemotherapy) can be administered for optimal surgical reduction, only IV treatment is recommended for sub-optimal reductions (Armstrong, Bundy et al. 2006). This is

4 due to the limitations of drug penetration into larger tumors. For optimally reduced

EOC, the most commonly used IV/IP chemotherapy regimen consists of 6 cycles of

IV paclitaxel (135 mg/m2)onday1,IPcisplatin(100mg/m2)onday2,andIP paclitaxel (60 mg/m2) on day 8 (Armstrong, Bundy et al. 2006). Clinical trials exploring dose modiﬁcations to this regimen have reported urinary tract infections, abdominal pain, and hyperglycemia (Dizon, Sill et al. 2011). IV/IP therapy has shown to improve disease-free survival and reduce risk of dying compared to standard IV therapy in a meta-analysis of 9 clinical trials with 2100 women (Jaaback,

Johnson et al. 2016). However, IV/IP therapy is associated with increased toxicity resulting in therapy discontinuation (Armstrong, Bundy et al. 2006).

1.1.3.1 Treatment of chemoresistant EOC

While the majority of EOC patients experience complete clinical remission after chemotherapy, about 75% of cases will recur (Romero and Bast 2012). Mainte- nance therapy has not improved survival outcomes, and thus only serial monitoring with physical examination, with CT scans as needed, and CA-125 studies are recommended (Mei, Chen et al. 2010). Treatment of relapsed EOC depends on the platinum-free interval (PFI). Patients with PFI of at least 6 months are considered to have platinum-sensitive disease while those with shorter PFI have platinum-resistant disease. Treatment of these patients attempts to maximize quality of life and control disease, and may include single or combination therapy with other taxane/platinum drugs, endocrine therapy, angiogenesis inhibition (bevacizumab) and the poly-ADP ribose polymerase (PARP) inhibitors (Hanker, Loibl et al. 2012). Currently PARP inhibitors rucaparib and olaparib are approved for patients with recurrent EOC, a known germline mutation in the BRCA genes, and disease progression despite 2-3 prior lines of treatment (Domchek, Aghajanian et al. 2016, Drew, Ledermann et al.

2016). Phase II clinical trials have found improved progression-free survival rates in

5 patients treated with combined olaparib and paclitaxel compared to standard carboplatin/paclitaxel chemotherapy, albeit with higher toxicities such as neutropenia

(Oza, Cibula et al. 2015). FDA approval of rucaparib was spurred by a study showing

65% complete or partial response to rucaparib monotherapy amongst patients with germline BRCA-mutant and platinum-sensitive HGSOC (Parkes and Kennedy 2016).

Many phase III trials are underway to assess the eﬃcacies of this new class of drugs.

1.2 Molecular mechanisms of EOC development

and metastasis

1.2.1 Molecular characteristics of EOC

FIGO classifies all serous ovarian carcinomas into low grade (Grade 1), intermediate grade (Grade 2), and high grade (Grade 3) tumors. Combining this with current molecular profiles and clinical features, the newest classification system from MD

Anderson Cancer Center separates ovarian tumors into Types I and II type tumors.

Interestingly, Type I includes low-grade serous ovarian cancers (LGSOC) comprising

5-10% of serous ovarian cancers, as well as mucinous, endometrioid, and clear cell carcionomas (Kaldawy, Segev et al. 2016). Meanwhile, Type II includes high grade serous ovarian cancers (HGSOC), as well as carcinosarcomas and undeﬁned carcinomas of the ovary (Koshiyama, Matsumura et al. 2014). While treatments are similar for both types, these types carry dramatically diﬀerent prognostic implications.

1.2.1.1 Low-grade serous ovarian cancer

Low grade serous ovarian cancer is most commonly associated with disruptions in the mitogen-activated protein kinase (MAPK) pathway, with activating mutations in the serine/threonine kinase BRAF and GTPase KRAS (Rojas, Hirshﬁeld et al.

6 2016, Roy and Cowden Dahl 2018). MAPK pathways are characterized by a series of phosphorylation steps that transmit growth signals into the nucleus to regulate transcription, chromatin remodeling, and protein synthesis. BRAF and KRAS mutations have been detected in both low-grade serous tumors and their presumptive precursor lesions, the serous borderline tumors. According to 2 separate and complementary studies by MD Anderson and Memorial Sloan Kettering, BRAF mutations were detected in 2% and 5% of low-grade cancers and 30% and 45% of borderline tumors, respectively (Wong, Tsang et al. 2010, Grisham, Iyer et al. 2013, Kaldawy, Segev et al. 2016). Meanwhile, KRAS mutation was found in 19% and 16%, respectively, of low-grade serous tumors with similar rates in borderline tumors (Kaldawy, Segev et al. 2016). The discrepancy in mutation rates of BRAF between low-grade serous and borderline tumors suggests that certain BRAF mutations may help protect against progression from borderline to low-grade serous tumors (Kaldawy, Segev et al. 2016).

Interestingly, BRAF mutations are also associated with decreased rates of recurrence among patients with borderline tumors (Tsang, Deavers et al. 2013). Additionally, patients with low-grade serous tumors with mutations in BRAF or KRAS have been observed to have better overall survival rates compared to wild type BRAF and KRAS

(median survival time of 106.7 months compared to 66.8 months, respectively) (Ger- shenson, Sun et al. 2015). Other mutations and pathways implicated in low-grade serous tumors include PTEN, PI3K, ARID1A-related chromatin remodeling, and the

Wnt/β-catenin pathway (Roy and Cowden Dahl 2018). Finally, the absence of NRAS mutations in borderline but presence in low-grade serous tumors suggests its role as adriverofLGSOC(Kaldawy,Segevetal.2016).

1.2.1.2 High-grade serous ovarian cancer

HGSOC, which makes up 75% of Type II ovarian cancers, results in 90% of deaths from ovarian cancer (Kurman and Shih Ie 2016). It is associated with high mitotic

7 activity, with disruptions in p53 and BRCA1/2, though mutations in ERRB2 and

AKT may also be involved (Rojas, Hirshﬁeld et al. 2016, Roy and Cowden Dahl

2018). CCNE1, which encodes the cell cycle regulator cyclin E1, was also frequently amplified in HGSOC (Cancer Genome Atlas Research 2011). Although all HGSOC are associated with a high frequency of p53 mutations, a subset of tumors is characterized by a closer morphological similarity to endometrioid and transitional cell carcinomas. These have been designated as the solid, pseudo-endometrioid, transitional cell-like (SET) variant and are associated with higher lymphocyte infiltration and mitotic index, but better clinical outcomes presumably due to greater chemosen- sitivity (Soslow, Han et al. 2012, Howitt, Hanamornroongruang et al. 2015, Kurman and Shih Ie 2016). Rates of BRCA1 mutations are significantly higher in this subset compared to other HGSOC.

Aside from their differences in molecular profile and prognoses, LGS and HGS ovarian cancers vary in their resistance to chemotherapy. Both are treated with surgical debulking (cytoreduction) followed by a platinum/taxane combination (Ger- shenson 2013, Kaldawy, Segev et al. 2016). However, HGS ovarian cancer is significantly more chemosensitive, with up to 80% of patients being disease-free after primary treatment compared to 52% in one study (Gershenson, Sun et al. 2006,

Ozols 2006). This discrepancy in chemoresistance has been attributed to longer cell cycles in LGSOC consistent with their indolent behavior (Schmeler and Gershen- son 2008). In contrast, the eventual chemoresistance of HGSOC may be attributed to both chemotherapy-driven selection of resistant subclones and acquisition of new mutations such as inactivating mutations in RB1, NF1, RAD51, and PTEN (Rojas,

Hirshﬁeld et al. 2016). Despite their chemoresistance, when no residual disease is achieved after optimal debulking, patients with LGSOC have longer overall survival than those with HGSOC (85% vs. 61% 5-year survival, respectively) (Grabowski,

Harter et al. 2016). In contrast, the survival rates for patients with LGSOC and HG-

8 SOC are similar when, after surgical debulking, there is residual disease larger than

1.0 cm2, suggesting that while the former may not respond well to chemotherapy, it is still a less aggressive disease (Ardighieri, Zeppernick et al. 2014).

In accordance with their varying molecular proﬁles, targeted therapies against both HGSOC and LGSOC are being developed. Clinical trials investigating targeted therapy with MEK1/2 inhibitors, alone or in combinations with other chemotherapy in the treatment of recurrent or progressive LGSOC are underway. A phase II trial of Selumetinib, one such MEK1/2 inhibitor, demonstrated a response rate of 15% for recurrent LGSOC with a median progression-free survival of 11 months (Farley,

Brady et al. 2013).

1.2.2 Origins of EOC

The origins of EOC are divided by the histological type and genomic proﬁle.

Amongst the non-serous EOC tumors, endometrioid carcinoma resembles endome- trial epithelium, mucinous resembles endocervical epithelium, and clear cell resembles glycogen-rich vaginal epithelium (Kim, Park et al. 2018). While LGSOC, mucinous, endometrioid, and clear cell carcinomas are thought to arise from endometriosis or fallopian tube-related serous borderline ovarian tumors, HGSOC and undiﬀerentiated carcinomas are thought to arise from fallopian tube epithelium (Rojas, Hirshﬁeld et al.

2016, Reid, Permuth et al. 2017). These precursor lesions are known as serous tubal intraepithelial carcinomas (STICs) and were ﬁrst discovered in women at high risk for ovarian cancer or with germline BRCA mutations who underwent risk-reducing salpingo-oophorectomy (RRSO) (Kurman and Shih Ie 2016). Instead of being present in the ovaries, these lesions were found only in the fallopian tubes. STICs were then found in 50-60% of patients with HGSOC ((Kindelberger, Lee et al. 2007, Przybycin,

Kurman et al. 2010, Kurman and Shih Ie 2016). Indeed, the risk of developing HG-

SOC is signiﬁcantly reduced in women who had underwent salpingectomy compared

9 to those who had not, again pointing to the fallopian tubes as the origin of HGSOC

(McAlpine, Hanley et al. 2014, Falconer, Yin et al. 2015).

However, not all HGSOC is believed to arise from STICs. Speciﬁcally, the SET variant is believed to arise from not just STICs, but other precursor lesions origi- nating from the fallopian tubes or elsewhere, potentially including ovarian surface epithelium (OSE) and ovarian cortical inclusion cysts-structures formed by invagi- nation of OSE into the ovarian stroma after ovulation (Kurman and Shih Ie 2016).

Indeed, inactivation of Rb1, p53, and BRCA1 in mouse OSE was shown to induce development of HGSOC, albeit often with low metastatic potential compared to the highly aggressive human HGSOC (Szabova, Yin et al. 2012). Alternatively, a new theory suggests that the absence of STICs in some patients with HGSOC are due to early serous proliferations (ESPs) that may lie undetected in the fallopian tubes until widespread metastases are found on the ovaries and within the abdominal cavity in general (Soong, Kolin et al. 2018). Yet, until further evidence surfaces, the origins of EOC remain unclear and the potential for HGSOC to arise from the ovaries still cannot be ruled out.

Regardless of whether EOC arises from the fallopian tubes or the ovaries themselves, how does ovarian cancer arise from normal epithelium? The incessant ovulation hypothesis and gonadotropin hypothesis have both been proposed to describe how EOC arises. The former draws from the increased prevalence of EOC with age and protection with parity, postulating that increased number of lifetime ovulatory cycles (along with the associated higher frequency of epithelial repair) increases likeli- hood of spontaneous mutations (Reid, Permuth et al. 2017). The latter proposes that increased exposure to hormones like luteinizing hormone (LH) and follicle-stimulating hormone (FSH) increases the risk of EOC (Reid, Permuth et al. 2017). Yet, little is known or proven still regarding the process by which EOC develops.

We do know that signiﬁcant inter- and intra-tumoral heterogeneity has been ob-

10 served in patients with EOC (Rojas, Hirshﬁeld et al. 2016). One study using in vitro and in vivo mouse models has shown that secondary lesions may arise mainly from detachment of multi-cellular aggregates from the primary tumor, which seed various secondary sites in the peritoneal cavity while maintaining their original phenotype

(Al Habyan, Kalos et al. 2018). Interestingly, inter-tumoral heterogeneity may not arise from mixing of aggregates from diﬀerent primary tumors. In mice with separate

GFP and mCherry-expressing primary tumors, metastatic lesions were single colored while mice with mosaic GFP/mCherry primary tumors had mosaic secondary tumors

(Al Habyan, Kalos et al. 2018). In both primary tumors and secondary clusters, a signiﬁcant number of EOC cells retained their proliferative capacities (Al Habyan,

Kalos et al. 2018). Together, it appears likely that rather than aggregating within the abdominal cavity, spheroids collectively detach and form secondary metastases that retain the molecular characteristics of the primary tumor.

1.2.3 Mechanisms of EOC metastasis

1.2.3.1 Hematogenous Dissemination

EOC is unique amongst carcinomas in that metastasis occurs in two separate processes: by passive dissemination within the abdominal cavity and by the more familiar hematogenous spread. The latter is a relatively more recent finding, when studies found that instead of randomly seeding all organs within the peritoneal cavity as may be expected with passive dissemination, there was a significantly greater number of metastatic lesions found in the omentum- a vascularized double layer of peritoneum hanging from the stomach that acts to restrict intra-abdominal infections and inflammation (Feki, Berardi et al. 2009). This, combined with studies showing circulating tumor cells in ovarian cancer patients, suggests a hematogenous route of dissemination, which was further confirmed by Pradeep et al., who showed

11 that diﬀerent gene expression proﬁles amongst SKOV3 ovarian cancer cells led to a predilection for hematogenous spread to the omentum (Phillips, Velasco et al. 2012,

Pradeep, Kim et al. 2014). More speciﬁcally, this subset of SKOV3 cells, termed

SKOV3-OM3 demonstrated greater expression of ERBB3 and NRG1 along with several epithelial-to-mesenchymal (EMT) markers including vimentin, MMP3, COL1A,

HMGA2, and HGF, and concomitant decreased expression of E-cadherin (Pradeep,

Kim et al. 2014).

1.2.3.2 Passive Dissemination

While hematogenous metastasis does occur, EOCs primary route of dissemination is passive, via transcoelomic spread within the abdominal cavity (Pradeep, Kim et al. 2014, Yeung, Leung et al. 2015). Cancer cells are shed from the primary tumor as single cells, clusters, or spheroids into the peritoneal cavity where they are carried with the accumulated peritoneal ﬂuid to various secondary sites such as the liver and diaphragm (as reviewed in Yeung et al., 2015). In patients, the sizes of these spheroids are highly variable, and in mouse models they are found in ascites as clusters ranging from less than 20 cells to over 100 cells (Allen, Porter et al. 1987, Al Habyan, Kalos et al. 2018). While both single cells and clusters can be shed into the abdominal cavity, evidence suggests that multi-cellular aggregates (spheroids) may be more resistant to anoikis, the process of programmed cell death in anchorage-dependent cells (Al

Habyan, Kalos et al. 2018). Speciﬁcally, over-expression of TrkB receptors and the

Hepatocyte Growth Factor (HGF) receptor MET are believed to inhibit anoikis as ovarian cancer cells ﬂoat in the peritoneal ﬂuid before seeding onto a secondary site

(Yu, Liu et al. 2008, Guadamillas, Cerezo et al. 2011).

12 1.2.3.3 Role of Ascites in EOC Metastasis

The characteristic accumulation of peritoneal fluid of metastatic EOC is known as ascites. Peritoneal fluid is continuously drained by the lymphatic system, which are enriched in the peritoneal lining of the abdominal cavity. In EOC, there is ob- struction of that lymphatic system mainly attributed to tumor burden. There is also increased neovascularization and increased vascular permeability promoted by vascular endothelial growth factor (VEGF) secretion from tumor cells that all contribute to fluid accumulation (as reviewed in Klymenko et al., 2017). Aside from the cancer cells, the ascites contains a significant number of white blood cells, along with high levels of lactate dehydrogenase and the mitogen lysophosphatidic acid (LPA) (as reviewed in Yeung et al., 2015). LPA is a highly bioactive lipid that induces matrix metalloproteinase (MMP) and urokinase plasminogen activator (uPA) secretion to promote EOC invasion (Pustilnik, Estrella et al. 1999, Fishman, Liu et al. 2001).

Another important component in ascites is the chemokine CXCL12, which binds to

CXCR4 receptors on EOC cells to increase expression of β1-integrin, also promoting cell migration (Balkwill 2004).

Aside from carrying EOC cells and inflammatory pro-invasive factors throughout the abdominal cavity, one study suggests that the laminar flow of ascites itself can further contribute to metastasis through induction of epithelial-mesenchymal transition (EMT). Rizvi et al. showed that shear stress increased vimentin expression and inhibited E-cadherin expression due to increased EGFR signaling (Rizvi, Gurkan et al. 2013). Accumulation of fluid in the abdominal cavity is also inherently associated with increased intraperitoneal pressure (IAP). Indeed, another study also demonstrated that under physiologically-relevant compression forces of around 22 mmHg, there was increased SNAI1 expression in EOC cells and aggregates (Klymenko, Kim et al. 2017). Overall, this suggests that biomechanical forces, and not just molecular interactions act to promote EOC progression.

13 1.2.3.4 EOC seeding within the abdominal cavity

Once EOC cells establish contact with the layer of epithelial cells known as mesothelium that line the abdominal cavity and organs, they adhere to and invade past the mesothelium and underlying basement membrane of collagen (types I and

IV), ﬁbronectin, and laminin (Witz, Montoya-Rodriguez et al. 2001). Various factors aﬀect EOC seeding. Fibronectin secretion from cancer-associated mesothelial cells and increased E-cadherin expression by EOC cells both promote cancer cell colo- nization of various secondary sites of metastasis (Imai, Horiuchi et al. 2004, Kenny,

Chiang et al. 2014). Chronic inﬂammation and scarring associated with aging promotes development of EOC at the ovaries as well as other sites of inﬂammation, such as the peritoneal lining along surgical wounds that release cytokines and chemokines

(Jia, Nagaoka et al. 2018). Cancer cells also preferentially home to milky spots, which consist of macrophages and lymphocytes, on the diaphragm and omentum

(Shimotsuma, Shields et al. 1993). Interactions between the cancer cells and white blood cells are believed to trigger intra-abdominal immune reactions and further cancer invasion through lymphatic tracts (Feki, Berardi et al. 2009). One consequence of intra-abdominal inﬂammation is the increased formation of ascites as mesothelial, immune, and stromal cells secrete interleukins (IL)-1, -6, and -8, promoting tumor angiogenesis and VEGF secretion (as reviewed in Yeung et al., 2015). Thus, the vis- cious cycle of inﬂammation, ascites development, and tumor dissemination propitiates itself.

1.2.3.5 EMT in EOC metastasis

Regardless of the route of dissemination, detachment of individual cells from a primary or secondary tumor involves disruptions in cell-cell junctions and increased cell motility, both of which are part of the hallmarks of EMT. Indeed in most carcinomas, tumor invasion and metastasis involves EMT, the process whereby cells acquire 14 genetic and epigenetic changes that promotes a more motile and invasive phenotype

(Klymenko, Kim et al. 2017). Inducers of EMT include various growth factors (e.g., epidermal growth factor (EGF), hepatocyte growth factor (HGF), and transforming growth factor-β (TGF-β)), hormones, and signaling pathways including the Wnt/β- catenin pathway and Notch-mediated transcription (Thiery, Acloque et al. 2009,

Klymenko, Kim et al. 2017). The process itself is prevalent in both physiologic and pathologic events, including embryogenesis, wound healing, tissue ﬁbrosis, and cancer

(as reviewed in Thiery et al., 2009).

Loss of cell-cell junctions (both tight junctions and adherens junctions) occurs early in EMT and involves the degradation and/or re-localization of cell-cell junctional proteins including E-cadherin and β-catenin (Lamouille, Xu et al. 2014). Meanwhile, loss of both cell junctions and contact with the basement membrane leads to loss of apical-basal polarity (Lamouille, Xu et al. 2014). For increased directional motility, actin-rich membrane protrusions, including lamellipodia and ﬁlopodia, promote ECM degradation while actin stress ﬁbers increase for greater cell contractility (Lamouille,

Xu et al. 2014).

In the context of EOC, EMT involves the increased expression of mesenchymal markers such as vimentin and N-cadherin, as well as epigenetic changes leading to

E-cadherin silencing (as reviewed in Klymenko et al., 2017). The latter comes about through zinc ﬁnger E-box-binding homeobox (ZEB1)-mediated recruitment of DNA methyltransferase 1 (DNMT1) the CDH1 promoter, and both SNAI1 and ZEB1- mediated recruitment of histone deacetylases (HDACs) (Peinado, Ballestar et al.

2004, Aghdassi, Sendler et al. 2012, Fukagawa, Ishii et al. 2015). Interestingly, certain histone methylations (e.g., H3K27me3, H3K9me2, H3K9me3, etc.) found in EOC are also found in breast cancer, suggesting a potential shared pathway for disease progression (as reviewed in Klymenko et al., 2017).

Many EMT-inducing factors are found in the ascites of EOC patients. Lysophos-

15 phatidic acid (LPA) in EOC patients plays an important role in induction of EMT through uPA-mediated cleavage of the extracellular domain of E-cadherin and conse- quent disruption of cell-cell junctions (Gil, Lee et al. 2008). At the same time, LPA has been shown to increase expression of matrix metalloproteinase (MMP)-9, increase

MMP-2 activity, promote β-catenin translocation to the nucleus, and induce actin reorganization with further eﬀects on cell adhesion and migration (Do, Symowicz et al.

2007, Liu, Burkhalter et al. 2012, Burkhalter, Westfall et al. 2015). Aside from LPA,

IL-6 has been shown to increase MMP-2 and -9 activity to promote EOC cell invasion

(So, Min et al. 2015). Additionally, expression of Wnt5a, another protein enriched in ascites, is associated with higher metastatic potential (Qi, Sun et al. 2014).

EMT is not only important in the earlier steps of EOC dissemination, but also in the later steps of seeding and invasion. Loss of E-cadherin helps active the Epidermal

Growth Factor Receptor (EGFR)/Focal Adhesion Kinase (FAK)/ERK-MAPK pathway, which leads to increased expression in α5-integrin and adhesion to mesothelial cells at various secondary sites (Sawada, Mitra et al. 2008, Vergara, Merlot et al.

2010). Sawada et al. go on to propose that the signiﬁcance of α5-integrin expression lies in the selection advantage conferred upon EOC cells when they can adhere to the

ﬁbronectin found in both the mesothelial cells (which quickly undergoes Fas/FasL- mediated apoptosis) and the submesothelial basement membrane (Sawada, Mitra et al. 2008).

1.2.3.6 Potential therapeutic targets in EMT

Understanding the underlying mechanisms of EMT in the context of EOC is important due to the potential of EMT inhibition in preventing EOC progression.

Inhibitors of the TGF-β pathway include small-molecule inhibitors of the kinase domains, immune modulation, as well as antisense-compound inhibitors (Fischer-

Colbrie, Witt et al. 1997). Vergara et al. showed that TGF-β treatment of SKOV3

16 cells leads to increased stress ﬁber formation and membrane protrusions (Vergara,

Merlot et al. 2010). The EGFR tyrosine kinase inhibitors geﬁtnib, erlotinib, and lapatinib have also been investigated while EGF stimulation has shown potential for prevention of epithelial inclusion cyst formation (Yingling, Blanchard et al. 2004,

Rivera, Vega-Villegas et al. 2008). HGF promotes cell scattering in EOC through the PI3K/AKT pathway and EMT through activation of p70S6K in EOC cell lines

(Pon, Zhou et al. 2008). Phosphorylation of p70S6K by paclitaxel has been shown to suppress its activity (Le, Hittelman et al. 2003). The vasoconstrictor peptide endothelin-1 (ET-1), which acts through activation of ETA receptor, is signiﬁcantly upregulated in EOC cells. ETA expression is associated with reduced E-cadherin expression and increased N-cadherin expression, but may be inhibited by atrasentan

(Spinella, Rosano et al. 2004, Rosano, Spinella et al. 2005). Finally, bone morphogenetic protein (BMP) 4 increases expression of transcriptional regulators Snail and

Slug while decreasing E-cadherin expression and activating Rho, Rac1, and Cdc42 small GTPases (Theriault, Shepherd et al. 2007). Inhibitors of BMP signaling prevent ALK1/2/3/6-mediated phosphorylation of Smad1/5/8 and have shown promise in the treatment of various diseases including prostate cancer (Lee, Cheng et al. 2011,

Sanvitale, Kerr et al. 2013).

1.2.3.7 EMT and MET in EOC metastasis and chemoresistance

Unlike in other carcinomas, the role of EMT in EOC progression and chemoresistance is not clear-cut. Indeed, results from 15 EOC cell lines and 70 ovarian carcinoma tissues showed increased expression of the microRNA (miR)-200 family members and decreased expression of ZEB1 and ZEB2 compared to normal ovarian surface epithelium (OSE) (Bendoraite, Knouf et al. 2010). This is surprising considering that these alterations are associated with mesenchymal-epithelial-transition

(MET). In light of this, it is interesting to note that normal ovarian surface epithe-

17 lium (OSE) on the ovaries does not express E-cadherin, but the OSE lining deep clefts, inclusion cysts, and ovarian tumors does (Sundfeldt, Piontkewitz et al. 1997).

When E-cadherin is introduced into the former in vitro, it has signiﬁcant eﬀects on adherens junctions formation and cell migration. Wu et al. showed in vitro that along with decreased β-catenin-dependent LEF-1 activity and increased spheroid formation, forced E-cadherin expression abrogated mesenchymal migration (Wu, Cipollone et al.

2008). This suggests that while E-cadherin may have a role in the early stages of

EOC dissemination, such as in the maintenance of anoikis-resistant spheroids within the abdominal cavity, its expression ultimately prevents mesenchymal invasion at the secondary sites.

In literature, EMT strongly correlates with EOC development and progression.

For example, expression of miR-9, which upregulates N-cadherin and vimentin while reducing E-cadherin expression is signiﬁcantly higher in ovarian cancer tissue compared to OSE (Zhou, Xu et al. 2017). Downregulation of E-cadherin is also associated with higher tumor grade and lower overall survival (Cho, Choi et al. 2006). Increased expression of transcriptional repressors of E-cadherin, such as Snail, is also associated with poor prognosis in EOC (Vergara, Merlot et al. 2010).

1.2.3.8 Partial EMT/MET in EOC

Across cancer types, malignant cells often do not fully transition from an epithelial to a mesenchymal phenotype, but instead adopt a stable intermediate phenotype which confers a number of beneﬁts including resistance to radiation and chemotherapy often seen in cells with the epithelial phenotypes (as reviewed in (Klymenko, Kim et al. 2017). This intermediate phenotype also demonstrates increased stemness, with increased expression of cancer stem cell markers, spheroid-forming and self- renewing capacity, and tumor-initiating potential (Kurrey, Jalgaonkar et al. 2009,

Santisteban, Reiman et al. 2009). Thus, it is not surprising that the line between

18 epithelial and mesenchymal phenotypes may sometimes be blurred in EOC. Cells with an intermediate (hybrid) mesenchymal phenotype with increased N-cadherin and vimentin expression are more resistant to anoikis than epithelial cells (Thiery, Acloque et al. 2009, Davidson, Holth et al. 2015). Higher vimentin expression is associated with increased chemoresistance to both platinum and paclitaxel (Kajiyama, Shibata et al. 2007, Haslehurst, Koti et al. 2012).

The intermediate phenotype can exist through self-renewal and proliferation, or through diﬀerentiation shift from either the epithelial or mesenchymal phenotype due to factors in the ascites (as reviewed in Klymenko et al., 2017). Certain EOC cell lines including OVCAR3 and OvCa432 can form clones of cells expressing both E-cadherin and N-cadherin, which have the potential to form spheroids with both epithelial and mesenchymal cells (Klymenko, Kim et al. 2017). EOC cells in serous eﬀusions can also undergo partial MET through reacquisition of E-cadherin (Davidson, Holth et al. 2015).

1.2.3.9 EMT and chemoresistance

Chemoresistance has traditionally been associated with EMT in various carcinomas and induction of MET has been the goal of several therapeutic approaches

(Shibue and Weinberg 2017). Surprisingly, in ovarian cancer, EMT is often, but not always, associated with increased chemoresistance. Induction of EMT in OVCA429 cells (a serous adenocarcinoma cell line) by activation of Notch3 signaling led to resistance to carboplatin (Gupta, Xu et al. 2013). Indeed, resistance to cisplatin, carboplatin, and/or paclitaxel has been observed with EMT induction (as reviewed by

Klymenko et al., 2017). However, another study of 46 ovarian cancer cell lines demonstrated greater cisplatin resistance via activation of NF-Kβ in epithelial cell types

(Miow, Tan et al. 2015). Overexpression of the EMT suppressor miR-200c/miR-141 induced even greater resistance to carboplatin than downregulation of members of

19 the miRNA-200 family.

What do clinical findings suggest regarding chemoresistance? In patients, spheroids with cells expressing more mesenchymal markers were found in ascites of those who had not previously received chemotherapy while spheroids with cells expressing epithelial markers were found in patients with recurrent or chemoresistant disease (Lat- ifi, Luwor et al. 2012). Even single cells from the latter group demonstrated greater resistance to cisplatin, indicating that chemoresistance could not entirely be explained by the compact structure of the spheroids. Yet, despite the fact that the conflicting evidence for and against the association between EMT and increased chemoresistance in EOC, evidence still implicates the mesenchymal phenotype in aggressive metastatic invasion (Klymenko, Kim et al. 2017). Therefore, efforts to prevent EMT continue to be the focus of anti-metastatic therapies.

1.3 EMT and the Adherens Junction (AJ)

1.3.1 AJ formation and stabilization: an overview

In addition to the loss of apical-basal polarity and changes in cell shape, EMT involves the loss of epithelial cell-cell junctions, which include tight junctions, adherens junctions (AJs), desmosomes, and scattered gap junctions (Lamouille, Xu et al. 2014). The AJs consist of both zonula adherens (ZA) and lateral junctions. The former comprises of transmembrane cadherin clusters that are more densely packed compared to the latter (Wu, Kanchanawong et al. 2015). Junction formation consists of 3 steps: (1) formation of cell-cell contacts through membrane protrusions, which lead to cadherin clustering, followed by (2) homophilic cadherin interactions between cells and ﬁnally, (3) contact expansion through actin reorganization (Cavey and Lecuit 2009).

The initiation of cell-cell contacts occurs through contact between lamellipodia

20 and membrane ruﬄes of neighboring cells (Vasioukhin, Bauer et al. 2000). Cadherin clusters form at punctate sites of contact (Kametani and Takeichi 2007), becom- ing immobilized at the plasma membranes, likely through their anchorage to actin

(Adams, Chen et al. 1998, Sako, Nagafuchi et al. 1998, Iino, Koyama et al. 2001,

Lambert, Choquet et al. 2002). Recruitment of more AJ proteins occurs over minutes as the puncta expand (Yonemura, Itoh et al. 1995, Adams, Chen et al. 1998).

Upon contact initiation, actin reorganization occurs as radial actin bundles connect cadherins to concentric actin rings under plasma membranes. The bundles are then replaced by the perijunctional actin belt and thicker actin arcs (Hirokawa, Keller et al. 1983). A number of actin regulators, including Rac1, Cdc42, Abl kinase, Arp2/3,

Cortactin, N-WASP, Formin-1, and Ena/VASP have been implicated in this process

(Vasioukhin, Bauer et al. 2000, Nakagawa, Fukata et al. 2001, Noren, Niessen et al.

2001, Kovacs, Ali et al. 2002, Kobielak, Pasolli et al. 2004).

Expansion of AJs relies on actin polymerization in membrane protrusions and actomyosin tension at contact sites. Speciﬁcally, Arp2/3-mediated branched actin polymerization in lamellipodia generates force to allow for contact expansion (Kovacs,

Ali et al. 2002, Helwani, Kovacs et al. 2004). At the same time, activated myosin

II and Rho at actin arcs contract at the edges of the cell, essentially generating an outward pulling force at the AJ (Krendel and Bonder 1999, Yamada and Nelson

2007). Actin bundles along the contact break and are pulled to the contact edges

(actin bundle snapping) (Krendel and Bonder 1999).

Actin-dependent stabilization of AJs depends on immobilization of E-cadherin clusters both by local actin ﬁlaments with low turnover rates and by the surrounding contractile actomyosin network (Cavey, Rauzi et al. 2008). Other regulators include

Eplin, an actin stabilizer recruited by junctional complexes and small GTPases (Abe and Takeichi 2008). Strength of the AJs is directly dependent upon anchorage of

E-cadherin to the actin cytoskeleton, not on levels of E-cadherin at the contact site

21 (Angres, Barth et al. 1996). However, destabilization of AJs occurs through cleavage of E-cadherins at the plasma membrane followed by their degradation (Lamouille, Xu et al. 2014).

Figure 1-1: (A) The traditional model indicates direct linkage between E- cadherin/β-catenin to F-actin via α-catenin. (B) An alternative model indicates cycling of α-catenin between E-cadherin/β- catenin and a perijunctional pool associating with F-actin. (Used with permission from Pokutta et al., Biochemical Society Trans- actions, 2008)

Aside from cadherins and actin, the other main players of AJs are the catenins, namely, α-catenin, β-catenin, and p120 catenin. The traditional model of AJ structure points to a direct linkage between the E-cadherin/β-catenin complex to F-actin via α-catenin (Fig. 1-1). However, some more recent studies revealed that binding

22 to β-catenin may reduce the aﬃnity of α-catenin for F-actin (Drees, Pokutta et al.

2005, Yamada, Pokutta et al. 2005). Thus, an alternative model has arisen, propos- ing that α-catenin cycles between two intracellular pools- one junctional, where it associates with the cadherin complex-and another cytosolic or perijunctional, where it associates with the underlying F-actin (Fig. 1-1) (Pokutta, Drees et al. 2008). The two α-catenin pools may allow for α-catenin to act as a molecular switch- turning oﬀ

Arp2/3-mediated branched actin-polymerization to promote linear actin polymerization, such as that mediated by the formin proteins (Gates and Peifer 2005, Pokutta,

Drees et al. 2008). This model is further complicated by studies that show a role for mechanical force in regulation of interactions between the actin and catenin binding. Indeed, the exact role of cortical F-actin in the AJ and its regulation by various catenins and actin-associated proteins (such as formins) is still a subject undergoing active investigation.

1.3.2 The cytoskeleton in AJ stability and cell motility

1.3.2.1 The cytoskeleton is necessary for AJ formation and stability

There are three types of cell-cell junctions: anchoring junctions, tight junctions, and gap junctions. Tight junctions contribute to polarity of epithelial cells, and regulate paracellular transport of water and solutes. Gap junctions allow for transport of small molecules and ions between cells to allow for electrical coupling across cells.

Anchoring junctions include AJs and desmosomes, and are necessary for the physical linkage between cells in tissue. The AJs provide mechanical support in morphogenesis, embryogenesis, tissue regeneration, and wound repair (Yonemura 2011). Previously, we discussed that the strength of AJs is dependent on anchorage of cadherins to the actin cytoskeleton, and that both local stable actin ﬁlaments and contractile actomyosin help to immobilize E-cadherin clusters at the junctions. To better understand

23 the role of the cytoskeleton in cell-cell junction stability, we ﬁrst look at the dynamic actin arrangements in epithelial cells during the process of AJ formation.

1.3.2.2 The actin cytoskeleton

Actin is one of the most abundant proteins found in nearly all eukaryotic cells.

With a molecular weight of 42 kDa, globular actin (G-actin) is the monomer upon which filamentous actin (F-actin) is polymerized. The roles of actin (also known as microfilaments or thin filaments) are as diverse as they are critical for normal cellular functioning. They include cellular motility, mitosis and cytokinesis, vesicle transport, cell signaling, gene expression, and maintenance of cell shape and cell-cell junctions. (Lodish 2008) To accommodate for these functions, there are 3 isoforms of actin in vertebrate cells. These are (1) α-actins, which are components of the muscular contractile apparatus; (2) β-actin, found in circular bundles, ventral stress

ﬁbers, cell-cell junctions and mitotic rings; and (3) -actin, found in the dorsal stress

ﬁbers, lamellae, lamellipodia cell cortex (Sun et al., 2015). Each G-actin molecule is linked to an ATP or ADP complexed with Mg2+ in the ATPase fold. In vitro, polymerization is induced by addition of cations (Mg2+, K+, or Na+) (Galinska-

Rakoczy, Wawro et al. 2009). F-actin consists of two winding strands of G-actin.

Actin is distinguished from other cytoskeleton types in electron micrographs by the binding of the actin-speciﬁc motor protein myosin. Myosin binding, with its unique orientation when bound to F-actin, also reveals a structural polarity in actin ﬁlaments.

(Moore, Huxley et al. 1970)

Growing actin ﬁlaments have pointed (-) ends and barbed (+) ends, where barbed ends are the sites of faster G-actin recruitment (Firat-Karalar, Hsiue et al. 2011).

In vitro, actin assembly occurs by nucleation, followed by elongation, and ﬁnally, maintenance of the steady state. In the nucleation phase, three G-actin subunit comes together to form a stable oligomer, or seed. Each binding reaction is powered by ATP

24 hydrolysis with the production of ADP and phosphate (Pi). In the elongation phase,

filament length increases as G-actin is added to both ends. As the number of F-actin increases while G-actin decreases, equilibrium is reached when the length of steady- state filament no longer changes even as G-actin continues to bind and dissociate from the ends. In the absence of actin nucleators, the rate of actin polymerization depends on the concentration of ATP-bound G-actin and addition to the (+) end is almost 10 times faster than to the (-) end. (Lodish 2008) Between the critical concentrations (0.12 M and 0.6 M) at which association and dissociation equilibrate, the continued net addition of ATP-G-actin at one end and net loss of ADP-G-actin at the other end of the filament is known as treadmilling (Neuhaus, Wanger et al.

1983, Lodish 2008). subsubsubsectionRegulation of the actin cytoskeleton While polymerization of G-actin into F-actin can occur spontaneously, various actin-binding proteins significantly accelerate and regulate the process. Profilin binds to G-actin, promoting the dissociation of ADP and replacement by ATP, and can bind other proteins with proline-rich residues while it is bound to actin. When G-actin is bound to profilin, it can only be added to the (+) end (Pollard, Blanchoin et al. 2000,

Pollard 2016). On the (-) end, severing proteins like coﬁlin destabilizes the ﬁlament to generate more free (-) ends (Pollard 2016). There are also capping proteins which helps to maintain the pool of actin monomers and limit the number of growing barbed ends (Pollard 2016). As they limit the length of actin, including that of cortical actin

ﬁlaments, these proteins (e.g., CAPZB) regulate cortex thickness and cell surface tension (Chugh, Clark et al. 2017). Meanwhile, tropomodulin binds to the (-) end and tropomyosin, preventing dissociation and helping to stabilize the actin ﬁlament

(Rao, Madasu et al. 2014).

1.3.2.2.1.1. Multiple classes of actin nucleators

Actin nucleation is the rate-limiting step of F-actin formation. To accelerate the process, there are multiple classes of actin nucleators. Here, we discuss the formin

25 family and the Arp2/3 complex. While the former is responsible for linear actin nucleation and polymerization, the latter is responsible for branched actin networks, both of which are essential for normal cellular processes (Alberts 2001, Bershadsky

2004). Linear actin is found in stress ﬁbers and contractile rings during cytokinesis while branched actin is found in the leading edge lamellae of migrating cells.

1.3.2.2.1.1.1. Formin-dependent actin polymerization

Figure 1-2: Two formin proteins interact through FH2 domains, which processively nucleate (A) and elongate F-actin ﬁlaments (B). (Used with permission from Bershadsky, Trends in Cell Bio, 2004.)

As previously discussed, formins nucleate long linear actin ﬁlaments. Formins form homodimers with the two formin homology (FH)-2 domains combining into adoughnut-shapedcomplexthatisabletobindtwoactinsubunitsfromthetwo intertwined helical strands at the (+) end (Figure 1-2 a). It then processively elon- gates the ﬁlament by alternating between the strands, allowing for addition of new

ATP-G-actin to one strand while maintaining its attachment to the other (Alberts 26 2001, Bershadsky 2004). Throughout this process, the proline-rich FH1 domain of the formin helps to recruit proﬁlin-ATP-actin by binding to proﬁlin (Figure 1-2b).

At the same time, formin binding at the (+) end prevents binding of actin by capping protein (Bershadsky 2004). Formin-dependent actin nucleation is regulated by

Rho-GTPases, which will be discussed in detail in following sections.

1.3.2.2.1.1.2. Arp2/3-dependent actin polymerization

Figure 1-3: The Arp2/3 complex consists of Arp2 and Arp3 (orange spheres) as well as other associated subunits (a). The complex binds the side of an existing ﬁlament, promoting branched actin polymerization at the barbed (+) end (b). (Used with permission from Bershadsky, Trends in Cell Bio, 2004.)

The Arp2/3 complex, which is found in all eukaryotes, consists of 7 subunits.

Two of these are actin-related proteins (Arp). Unlike the formins, Arp2/3 requires a pre-formed actin ﬁlament as well as a regulatory protein such as Wiskott-Aldrich syndrome protein (WASp) in order to be active (Lodish 2008). Upon its activation,

Arp2 and Arp3 changes conformation to resemble the (+) end of the actin ﬁlament.

27 However, since it is attached to the side of the ﬁlament (Fig. 1-3), new G-actin is added to Arp2/3 at an angle of 70 degrees from the old ﬁlament (Bershadsky

2004, Pollard 2016). Regulation of Arp2/3 occurs via activation or inhibition of its regulatory proteins. For example, Cdc42, a Rho-GTPase, binds to and activates

WASp, allowing it to not only bind and activate Arp2/3 but also bind new G-actin

(Rohatgi, Ma et al. 1999).

1.3.2.2.2. Toxins and drugs that aﬀect actin dynamics

Various natural toxins have been harnessed for investigation of actin structure and dynamics. For the former, the fungal toxin phalloidin irreversibly stabilizes actin, and, when ﬂuorescently labeled, allows for visualization of actin ﬁlaments in vitro. Another sponge toxin, jasplakinolide, also stabilizes actin dimers. Cytochalasin D binds to the

(+) end of F-actin preventing addition of G-actin to that end. Notably, neither F- actin stabilization with jasplakinolide nor depolymerization with cytochalasin changes the amount of cytosolic α-catenin monomers and homodimers (Yamada, Pokutta et al. 2005). Finally, latrunculin binds and sequesters G-actin, preventing actin polymerization. Both cytochalasin D and latrunculin lead to actin disassembly in live cells (Wakatsuki, Okatani et al. 2000).

1.3.2.2.3. Branched and unbranched actin participate in AJ stabilization

In epithelial cells, radial actin filaments such as that found in finger-like membrane protrusions called filopodia are found at sites of cadherin clustering, but some are gradually replaced by the circumferential actin belt as the junctions mature (as reviewed in Harris and Tepass 2010). The new AJs are supported by both newly polymerized actin near the cadherin clusters and pre-existing cortical actin (Va- sioukhin, Bauer et al. 2000, Ivanov, Hunt et al. 2005, Kishikawa, Suzuki et al.

2008). The proteins responsible for new actin polymerization include the Arp2/3 complex, formins, and members of the Ena/VASP family (e.g., MENA) (Harris and

Tepass 2010). Arp2/3 is known to co-immunoprecipitate with E-cadherin and its in-

28 hibition signiﬁcantly aﬀects cell-cell junctions formation, as well as F-actin turnover in mature AJs (Verma, Shewan et al. 2004, Kovacs, Verma et al. 2011). The formin, mDia1 also localizes to AJs in a RhoA-dependent manner (Carramusa, Ballestrem et al. 2007). mDia proteins are also known to be required for junction maintenance (Sa- hai and Marshall 2002, Carramusa, Ballestrem et al. 2007, Homem and Peifer 2008) but the mechanism of its recruitment to the AJ is unclear. Magie et al. propose that α-catenin and p120 catenin directly recruit it via recruitment of Rho1 (Magie,

Meyer et al. 1999). During the process of AJ formation, levels of branched actin poly- merizers (e.g., Arp2/3, and cortactin) decrease as cadherins begin to cluster just as

α-catenin localizes to AJs (Helwani, Kovacs et al. 2004, Verma, Shewan et al. 2004,

Perez-Moreno and Fuchs 2006, Yamada and Nelson 2007). Unbranched actin poly- merizers such as Formin-1 and Ena/VASP are concurrently recruited (Vasioukhin,

Bauer et al. 2000, Kobielak, Pasolli et al. 2004).

1.3.2.2.3.1 Rho-GTPases are key orchestrators of actin dynamics

The recruitment of actin nucleators and regulatory proteins is largely associated with Rho-GTPase signaling. The Rho-GTPases (i.e., Rho, Rac, Cdc42) are commonly referred to as molecular switches, which are active when bound to GTP and inactive when bound to GDP (Haga and Ridley 2016). In the cytosol, they remain in an inactive state bound to guanine nucleotide displacement inhibitor (GDI). Various signaling pathways lead to the translocation of the Rho-GTPases to the membrane along with their activation (through exchange of GDP for GTP) by guanine nucleotide exchange factors (GEFs) (Haga and Ridley 2016). Rho-GTPases are inactivated by

GTPase-activating proteins (GAPs). In cancer, expression or activity of GEFs and

GAPs are often disrupted, with eﬀects on cell migration and junctions (Ellenbroek and Collard 2007).

Rac and Rho are speciﬁcally implicated in AJ formation while Cdc42 is necessary for AJ maintenance (Braga 2002, Yap and Kovacs 2003). Once activated, Rac recruits

29 the WASp family protein verprolin homologue 2 (WAVE2), which activates Arp2/3 with the help of WAVE1 (Yamazaki, Oikawa et al. 2007). Rac is especially important in the initial steps of AJ formation as it promotes exploratory membrane protrusions

(Harris and Tepass 2010). On the other hand, Rho is more concentrated in the contractile actin bundles of more mature AJs (Yamada and Nelson 2007). Early in

AJ formation, Rho is inhibited by Rac and the kinase Abelson (Sander, ten Klooster et al. 1999, Mertens, Rygiel et al. 2005). Rho activation is speculated to be achieved indirectly by inhibition of Rac, which can occur through partitioning defective 3 homologue (PAR3)-dependent inhibition of T lymphoma invasion and metastasis- inducing protein 1 (TIAM1) (Harris and Tepass 2010). Rho activates Rho kinase

(ROCK) to induce actomyosin contractility necessary for membrane alignment and junction maturation (Delanoe-Ayari, Al Kurdi et al. 2004, Yamada and Nelson 2007).

1.3.2.3 The microtubule cytoskeleton

Microtubules are also an important component of AJs. Therefore we devote the section below to an overview of their structure and function.

1.3.2.2.3.1 Rho-GTPases are key orchestrators of actin dynamics

Compared to the actin cytoskeleton, the microtubule network is more rigid. Like actin, though, they can exist as single strands (of up to 20 µm) or bundles, as seen in motile structures like cilia and ﬂagella. Like actin, they also have intrinsic polarity and speciﬁc motor proteins that bind to them, kinesins and dyneins. In addition to organelle transport, microtubules are most well-recognized for their role in mitosis, wherein they form the signature mitotic spindle that separates chromosomes to opposite ends of the cell. (Lodish 2008).

Dimers of α-andβ-tubulin make up 13 protoﬁlaments, which associate with one another to form a tubule about 25 nm in diameter. Each subunit can bind a GTP molecule, but only the GTP bound to β-tubulin can be hydrolyzed. Polarity is

30 derived from the alternation between α-andβ-tubulin subunits down the length of the tubulin, leaving an α-tubulin on one end and β-tubulin on the other (Akhmanova and Steinmetz 2015). The latter is designated as the (+) end due to its faster growth rate. Most microtubule structures in the cell consist of singlet microtubules, but in cilia and flagella, a 13-protofilament tubule combines with 10-protofilament tubules to form doublet or triplet microtubules (Lodish 2008).

1.3.2.3.2. Microtubule regulation and dynamics

Just as there is a seed from which actin is elongated, all microtubules begin as microtubule-organizing centers (MTOCs), also known as centrosomes in interphase cells or spindle poles in mitotic cells. Centrosomes themselves consist of a pair of centrioles, which are made of triplet microtubules. Proteins surrounding the centrioles, such as the γ-tubulin ring complex (γ-TURC) are responsible for adding new tubulin dimers to the centrosomes. Microtubules also undergo treadmilling like actin, but uniquely experience dynamic instability wherein individual microtubules can undergo catastrophe followed by shrinkage that can be reversed with rescue (Akhmanova and

Steinmetz 2015). The length of microtubules is determined by the balance between catastrophes and rescues, as well as by the presence of anchors such as organelles, which can bind to the (+) end and prevent disassembly.

Similar to actin, tubulin can assemble in the absence of its associated proteins, the microtubule-associated proteins (MAPs). However, stabilization of microtubules in cells requires MAPs, which increase microtubule growth rates (Desai and Mitchi- son 1997). One of these is the family of tau proteins, which includes tau, MAP2, and MAP4. MAPs are regulated through phosphorylation by microtubule aﬃnity- regulating kinase (MAPK/Par-1) and cyclin-dependent kinases (CDKs). Other proteins promote microtubule disassembly and include the kinesin-13 family, Oncoprotein

(OP) 18/stathmin, and katanin (Cassimeris 2002, Howard and Hyman 2003). Finally, it is worth noting that some proteins, including certain members of the formin family,

31 regulate both actin and microtubule dynamics, indicating their central role in regulation and coordination between elements of the cytoskeleton (Bartolini, Moseley et al. 2008).

1.3.2.3.2.1. mDia formins regulate microtubule dynamics

Independent of their ability to polymerize actin, mDia formins also stabilize microtubules, preventing their disassembly. Speciﬁcally, the FH1 and FH2 domains of mDia2 can bind to microtubules in cells, reducing both assembly and disassembly rates (Bartolini, Moseley et al. 2008). Interestingly, mDia1 was shown to promote microtubule stabilization through regulation of glycogen synthase kinase-3 β (GSK3β) via protein kinase Cs (PKCs). Separately, an mDia2-APC/EB1 pathway has been proposed to regulate indirect microtubule stabilization (as reviewed in DeWard and

Alberts 2008). De-tyrosinated -tubulin within polarized microtubules stabilized by the Rho/mDia axis (also known as Glu MTs) contribute to cell polarity by helping to promote transport of vesicles and actin regulators to the leading edge (Lin, Gunder- sen et al. 2002). This is important in cell migration as Glu MTs form in the leading edge of cells before migration and orient in the direction of the migration (Gundersen and Bulinski 1988).

1.3.2.3.3. Drugs that aﬀect microtubule dynamics

Due to their essential role in mitosis, the microtubules have been the target of many therapeutic agents. Colchicine is one of the best-known agents and acts by sequestering tubulin dimers, preventing microtubule assembly (Dalbeth, Lauterio et al. 2014). Colchicine treatment results in eventual loss of all microtubules in the cell except for the centrosomes. It is a drug known for its use in the treatment of acute gout, an inﬂammatory joint disease (Dalbeth, Lauterio et al. 2014). Nocodazole, another inhibitor of microtubule polymerization is often used in the research setting and also acts by dimer sequestration (Florian and Mitchison 2016).

Anti-microtubule chemotherapy involves either microtubule stabilization or desta-

32 bilization (by inhibition of polymerization). The former includes the class of drugs known as taxanes (which includes paclitaxel). These drugs are known to prevent mitosis, and are thus widely used in chemotherapy for various cancers including ovarian, breast, lung, bladder, and esophageal carcinomas (Zhang, Yang et al. 2014). In contrast, another spindle poison, the vinca alkaloids, prevent microtubule polymerization. One of the most well-known of these, vincristine, is used in the treatment of acute lymphocytic leukemia, acute myeloid leukemia, Hodgkins disease, neuroblas- toma, and small cell lung cancer (Jordan 2002, 2003).

1.3.2.3.4. Role of Microtubules at the AJ

In the cell, microtubules are necessary for the transport of various AJ proteins

(Harris and Tepass 2010). Physically, the growing microtubule (+) ends typically face the junctions, while the (-) ends are oriented toward the cell center (Harris and Tepass

2010). AJs are believed to stabilize microtubules by anchoring them and inhibiting dynamic instability, a process discussed in the previous section on microtubules

(Waterman-Storer, Salmon et al. 2000). The (+) ends also recruit plus end-tracking proteins (+TIPs) such as cytoplasmic linker protein 170 (CLIP170), which localize with microtubules in cadherin-catenin clusters (Ligon, Karki et al. 2001, Stehbens,

Paterson et al. 2006, Ligon and Holzbaur 2007). Dynein, the microtubule-speciﬁc motor protein, also associates with β-catenin and is believed to pull microtubules to the cell cortex (Karki, Ligon et al. 2002, Harris and Tepass 2010). Even without AJs,

α-catenin expression in cell membranes is able to aﬀect microtubule polymerization

(Shtutman, Chausovsky et al. 2008). Another catenin, p120, binds PLEKHA7, which interacts with NEZHA, a protein that associates with microtubule (-) ends. Thus AJ localization of PLEKHA7 results in recruitment of NEZHA-bound microtubule to the

AJ. Both PLEKHA7 and NEZHA have been shown to be necessary for AJ stability, again supporting the important role of microtubules at the AJ (Meng, Mushika et al.

2008, Harris and Tepass 2010).

33 1.3.2.4 Intermediate ﬁlaments

The third and remaining component of the cytoskeleton consist of intermediate

filaments. These proteins consist of at least five families of proteins encoded by 70 genes that are known for their tensile strength (Herrmann and Aebi 2000, Fletcher and Mullins 2010). All intermediate filament proteins contain a core α-helical rod of 310 residues (Lodish 2008). They are separated into five classes based on their varying C- and N-terminals. To assemble into filaments, the proteins dimerize and form protofilaments. A single filament consists of 16 protofilaments. Unlike actin and microtubules, intermediate filaments do not have intrinsic polarity and do not have any known associated nucleating proteins. (Lodish 2008). Keratins, desmin, vimentin, neurofilaments, and lamins are several examples of this third type of cytoskeleton network (Chung, Rotty et al. 2013). In accordance with their role as key structural components of the cell, intermediate filaments span the cytoplasm of cells and the inner nuclear envelope. They are also much more stable, with slower turnover rates compared to microtubules and actin (Herrmann and Aebi 2000, Lodish 2008).

1.3.3 Mechanisms of Junction Destabilization

Junction stability results from there being greater stabilizing vs. destabilizing forces at cell-cell contacts. The latter involves membrane protrusive forces while the former involves junctional Myo-II-dependent tension, which aligns actin ﬁlaments parallel to the cell membranes, preventing protrusion formation (Gloushankova, Alieva et al. 1997, Sahai and Marshall 2002, Zhang, Betson et al. 2005).

Cadherin endocytosis is an essential step in AJ destabilization. Although the exact mechanism for cadherin translocation from the junctions to clathrin-coated pits is unclear, loss of p120 catenin was shown to help trigger the process. The catenin has been speculated to inhibit endocytic machinery (including the ligase hakai) near

34 the cadherin juxtamembrane domain (Fujita, Krause et al. 2002, Xiao, Garner et al. 2005, Ishiyama, Lee et al. 2010). Rho-GTPases also regulate endocytosis of AJ proteins. Speciﬁcally, Cdc42 acts through its eﬀectors, Cdc42-interacting protein 4

(CIP4), to promote Dynamin-dependent vesicle scission D. melanogaster (Leibfried,

Fricke et al. 2008). Also, Georgiou et al. showed that together with Par6 and aPKC,

Cdc42 regulates Arp2/3-mediated actin polymerization to regulate endocytosis and

AJ stability (Georgiou, Marinari et al. 2008). Additionally, Cdc42 aﬀects traﬃcking of Crumbs, another regulator of endocytosis of AJ proteins (Harris and Tepass 2008).

Meanwhile, Rho has also been shown to help maintain AJ in D. melanogaster pigment epithelial cells by inhibiting AJ endocytosis (Warner and Longmore 2009). Finally,

ADP-ribosylation factor 6 (ARF6) has been shown to promote dynamin-dependent

AJ endocytosis while the ARF6-GAP, SMAP1 prevents endocytosis (D’Souza-Schorey

2005, Kon, Tanabe et al. 2008).

1.3.4 Cell motility is achieved through regulation of the cy-

toskeleton

The same factors that induce endocytosis of AJ proteins also lead to increased cell motility through EMT. Cell motility relies on significant rearrangements of the cytoskeleton to create membrane protrusions and contractile force, which propels movement. While microtubules generate force for movement and intermediate filaments switch from cytokeratin to vimentin during EMT, actin polymerization and reorganization occur to generate membrane protrusions such as lamellipodia, filopodia, invadopodia, and membrane blebs (Ridley 2011).

35 1.3.4.1 Overview of membrane protrusions

Lamellipodia are actin-rich structures extending 3-5 µmfromtheleadingedge

(also known as the lamella) of migrating cells (Gupton, Eisenmann et al. 2007). They do not contain any microtubules, but rather contain branched actin along with actin nucleators like Arp2/3 (Ridley 2011). A ready supply of linear F-actin is supplied, in part, by the mDia formins in the lamellae (Gupton, Eisenmann et al. 2007).

Filopodia are distinguished from other membrane protrusions by their slender

ﬁnger-like morphology (Arjonen, Kaukonen et al. 2011). Like lamellipodia, their formation is driven by actin nucleators, namely, the mDia formins and WASp (Sarmiento,

Wang et al. 2008, Yang and Svitkina 2011). Other proteins such as the actin-bundling fascin and Rho GTPases (Cdc42, RhoF, and Rif) are also important in ﬁlopodia formation and stability (Ridley 2011).

In cancer cells, invadopodia are a specialized type of membrane protrusion that uses the same actin machinery as lamellipodia and ﬁlopodia but can also degrade extracellular matrix (ECM) (Yamaguchi 2012). In addition to actin nucleators and

MMPs, these structures also contain microtubules, which serve to provide structural support as well as transport of proteases (Ridley 2011).

Cell migration is an intricately coordinated process involving: (1) extension of lamellipodia from the leading edge; (2) adhesion of lamellipodia to the ECM; (3) forward propulsion of the cell body upon contraction in the cell rear; and (4) detachment of the trailing end with endocytic recycling of the cell membrane and integrins

(Lodish 2008). Rac activates the Arp2/3 complex to form lamellipodia, while Cdc42 and its eﬀector Par-6 is required for orientation and polarity of the cell as it migrates

(Etienne-Manneville, Manneville et al. 2005). WASp and Arp2/3 both are activated by Cdc42 to induce ﬁlopodia formation (Welch and Mullins 2002, Svitkina, Bulanova et al. 2003). Contraction in the cell rear is mediated by Rho, which activates the myosin II-dependent contractile actin machinery (Kodama, Lechler et al. 2004).

36 During EMT, actin polymerization along with its interactions with actin-binding proteins and myosin at the leading edge generates the force required for cell motility.

Overexpression of Arp2/3 is often seen in malignant tumors (Iwaya, Oikawa et al.

2007, Spence, Timpson et al. 2012, Monteiro, Rosse et al. 2013) and overexpression of cortactin is associated with metastasis (Han, Gambin et al. 2014, Helgeson,

Prendergast et al. 2014). Upon activation by Rho-GTPases, WASP/WASP family verprolin-homologous protein (WAVE) proteins activate Arp2/3 to promote branched actin polymerization (Sun, Fang et al. 2015). The Rho in filopodia (Rif)/mammalian diaphanous 2 (mDia2) axis has also specifically been shown to stimulate filopodia protrusions (as reviewed in Sun et al, 2015).

Using latrunculin, which prevents formation of new actin ﬁlaments, researchers have found that actin in the leading edges of migrating cells is highly dynamic with a turnover rate of 30-180 seconds while stress ﬁbers are relatively stable structures with turnover of 5-10 minutes. In cancer, certain pseudopodia-enriched proteins including

AHNAK, spetin-9 and S100A11 are upregulated, and their inhibition prevents cell migration and invasion (Shankar and Nabi 2015). Actin levels can also aﬀect intercellular junctions as reduction β-actin has been shown to result in its distribution in lamellar ruﬄes instead of cell junctions (Shagieva, Domnina et al. 2012, Sun, Fang et al. 2015).

Another type of cell migration termed amoeboid migration is also seen in cancer cells, which are known to have migratory plasticity. When proteolytic enzymes such as the MMPs are inhibited, cancer cells can undergo mesenchymal-amoeboid transition

(MAT) (Wolf, Mazo et al. 2003). In amoeboid migration, non-apoptotic membrane blebs protrude as cortical actin is disrupted and retract with actomyosin contraction.

This process is mediated by RhoA and its eﬀectors ROCK and the formin mDia2

(Eisenmann, Harris et al. 2007, Fackler and Grosse 2008, Wyse, Lei et al. 2012).

37 1.3.5 AJ regulators: cadherins, catenins, and vinculin

1.3.5.1 Cadherins

Cadherins are a super-family of cell-adhesion molecules (CAMs) found in AJs and desmosomes, with at least six subfamilies. One of these subfamilies, classical cadherins includes epithelial (E)-cadherin, neural (N)-cadherin, and placental

(P)-cadherin. Binding is calcium-dependent and primarily hemophilic, though heterophilic interactions can occur (Vasioukhin, Bauer et al. 2000, Prakasam, Maruthamuthu et al. 2006).

Classical cadherins contain a transmembrane domain, a C-terminal cytoplasmic domain, and ﬁve extracellular domains necessary for Ca2+ binding. The E-, N-, and P-cadherins have unique distal extracellular N-terminals. Via their cytoplasmic domains (catenin-binding domain and juxtamembrane domain), cadherins associate with catenins, including β-catenin, γ-catenin (also known as plakoglobin), and p120 catenins (Kourtidis, Lu et al. 2017). Cadherins can directly bind to actin cytoskeleton at the cytoplasmic tail and the F-actin network regulates localization of cadherin clusters (Hong, Troyanovsky et al. 2013, Wu, Kanchanawong et al. 2015). How- ever, binding to β-catenin is speciﬁcally required for AJ maturation (as reviewed in

Kourtidis et al. 2017). Phosphorylation of either E-cadherin or β-catenin aﬀects

β-catenin binding to the catenin-binding domain. Meanwhile E-cadherin binding to the lipid kinase Iγ PI phosphate kinase (PIPKIγ)promotesintracellularE-cadherin traﬃcking but binding to phosphatase PTPm helps maintain membrane localization

(Brady-Kalnay, Rimm et al. 1995, Ling, Bairstow et al. 2007).

In addition to endocytosis with subsequent degradation, E-cadherin levels are regulated at the transcriptional level by Snail and other zinc-ﬁnger family transcription factors implicated in EMT, including SIP1, Slug, and Twist. (Kourtidis, Lu et al.

2017) Additionally, certain microRNAs like the miR-200 family promote E-cadherin

38 expression, often by targeting EMT inducers such as ZEB1 and SIP1 (Gregory, Bert et al. 2008).

1.3.5.2 α-catenins

Three diﬀerent isoforms of α-catenin are expressed in epithelial cells (αE-catenin), neurons (αN-catenin), and testes (αT-catenin), with all having the ability to interact with cadherins and mediate cell-cell adhesion (Kobielak and Fuchs 2004). More speciﬁcally, in vivo studies have shown that αE-catenin is required for AJ formation in epithelial cells, and cancer cells without αE-catenin do not form mature AJs

(Shimoyama, Nagafuchi et al. 1992, Yonemura, Itoh et al. 1995, Adams and Nelson

1998). AJ disassembly involves dissociation of both actin and αE-catenin from the

E-cadherin/β-catenin complex either through tyrosine phosphorylation of β-catenin or interaction between IQGAP1 and β-catenin (as reviewed by Kobielak and Fuchs

2004). The mechanism by which αE-catenin promotes stable AJs is a subject of ongo- ing investigation. While the classical model of the AJ proposes a direct linkage of the

E-cadherin/β-catenin complex and the underlying F-actin via α-catenin, studies have shown that when associated with the complex αE-catenin actually has signiﬁcantly reduced aﬃnity for F-actin (Yamada, Pokutta et al. 2005). This, together with the observation that stretching of α-catenin exposes binding sites for actin-binding proteins such as vinculin, suggests that the mechanism of α-catenin-dependent junction stabilization is not so straightforward (Yonemura 2011).

1.3.5.3 β-catenin

In epithelial cells, β-catenin mostly localizes to junctional cadherin-catenin complexes (as reviewed in Kourtidis et al. 2017). In the absence of Wnt signaling,

β-catenin is phosphorylated and translocates to the cytoplasm where it associates with complexes formed by Axin, adenomatous polyposis coli (APC), and the ser-

39 ine/threonine kinase glycogen synthase 3β (GSK-3β)(Polakis2000).Itisthenubiq- uitinated and degraded by this complex. When Wnt binds its receptor Frizzled, the

Axin/APC/GSK-3β complex is inhibited and cytosolic β-catenin translocates to the nucleus. (MacDonald, Tamai et al. 2009, Clevers and Nusse 2012). It then activates the T-cell factor/lymphoid enhancer factor (Tcf/Lef) family of transcription factors and promotes expression of target genes, including Axin2 and LGR5 (Clevers and

Nusse 2012).

1.3.5.4 p120 catenin

One of the main functions of p120 catenins is to stabilize cadherins at the AJ

(Ireton, Davis et al. 2002, Kourtidis, Ngok et al. 2013). Interaction between p120 and E-cadherin is disrupted upon binding of the ligase Hakai to the JMD and results in

E-cadherin endocytosis (Fujita, Krause et al. 2002). Interestingly p120 also associates with microtubules and kinesins, which regulate their localization (Chen, Kojima et al. 2003, Yanagisawa, Kaverina et al. 2004). The catenin is involved in inside-out cadherin signaling, wherein its phosphorylation regulates cadherin function (Petrova,

Spano et al. 2012). Indeed, p120 stabilizes AJs by recruiting ROCK1 and actin to

AJs. Overexpression of p120 catenin leads to their colocalization with N-cadherin in intracellular vesicles and subsequent association with kinesin 1. Disruptions in these interactions prevent N-cadherin localization to AJs via microtubule-dependent traﬃcking (Chen, Kojima et al. 2003).

1.3.5.5 Vinculin

Vinculin localizes to both focal adhesions and AJs. Its binding to α-catenin is force-dependent (Miyake, Inoue et al. 2006). Interestingly, there is signiﬁcant sequence similarity before vinculin and α-catenin. Upon myosin II inhibition, vinculin translocates from the AJs (Yonemura 2011). Since it is an actin-binding protein, vin-

40 culin enhances mechano-sensing by E-cadherin and increases F-actin concentrations at the AJ in response to force (le Duc, Shi et al. 2010). Overall, the force-dependent interactions between vinculin and AJ components set a precedent for discovery of other force-dependent associations actin-binding proteins to the AJ and the catenins.

1.4 Formin proteins: an introduction

1.4.1 An overview of formins

There are 15 formin isoforms in mammals and 6 in Drosophila with molecular weights ranging from 120 to 200 kDa (Higgs and Peterson 2005, Chhabra and Higgs

2007, Chesarone, DuPage et al. 2010). These 15 mammalian formin isoforms, called

Diaphanous-related formins (DRFs) are divided into families, including but not limited to mammalian Diaphanous (mDia), disheveled-associated activator of morphogenesis (DAAM), formin (FMN), and formin homology domain-containing protein

(FHOD) families (Kuhn and Geyer 2014). All DRFs can dimerize at the FH2 domain, the domain responsible for linear actin polymerization (Chhabra and Higgs

2007). As the mDia proteins are the main subject of investigation in this work, we focus on the mDia formins below.

1.4.2 Structure of mDia formins

The gene DIAPH1 on chromosome 5q31.3 encodes mDia1 protein (also known as

DRF1, hDia1, Dia1, or Diaph1), while the gene DIAPH3 on chromosome 13q21.2 encodes mDia2 (Diaph2, Dia2, hDia2) (Katoh and Katoh 2004, Eisenmann, Dykema et al. 2009). The mDia formins are characterized by 7 major domains from N- to

C-terminus (see Figure 1-4): the GTPase binding domain (GBD), the diaphanous inhibitory domain (DID), the dimerization domain (DD), the coiled coil (CC) domain,

41 Figure 1-4: In the autoinhibited state, the DAD domain binds to DID, preventing actin polymerization by the FH2 domain. Binding by Rho-GTPases at the GBD activates the formin by preventing this association. (Used with permission from DeWard and Al- berts, Current Biology, 2008)

the formin homology (FH) 1 and FH2 domains, the diaphanous auto-regulatory domain (DAD). The proline-rich FH1 domain binds profilin, which helps to recruit actin monomers to the growing actin filament and significantly helps to increases the rate of polymerization by increasing the concentration of G-actin at the (+) end (Higgs

2005, Paul and Pollard 2009). The FH2 domain is the functional domain of formins in that it is essential for both actin nucleation and elongation (Pruyne, Evangelista et al. 2002, Kovar 2006). As mentioned previously, the FH2 domain processively binds to the growing (+) end of the actin ﬁlament, promoting polymerization both by catalyzing the addition of new G-actin and by preventing the binding of capping proteins (Harris, Rouiller et al. 2006, Baarlink, Brandt et al. 2010, Chesarone, Du-

Page et al. 2010). In mDia1 and mDia2, the mDia FH2 domain, together with the

DD, are responsible for formation of active mDia homodimers (Pruyne, Evangelista et al. 2002, Sagot, Rodal et al. 2002, Baarlink, Brandt et al. 2010). However, to homodimerize and bind to actin, the mDia formin must ﬁrst be released from its autoinhibited state. This is accomplished by the Rho-GTPase binding to the GBD

(Baarlink, Brandt et al. 2010, Chesarone, DuPage et al. 2010). In the autoinhibited 42 state, the DID at the N-terminus is folded onto and interacts with the diaphanous auto-regulatory domain (DAD) at the C-terminus (Alberts 2001, Li and Higgs 2005,

Wallar, Stropich et al. 2006). When mDia1 and mDia2 are inhibited, homodimer formation via their FH2 domains is inhibited (Copeland, Green et al. 2007). Other domains of the mDia formins include the basic domain (BD) and CC domain, the former of which mediates localization of mDia2 to the plasma membrane (Wallar,

Stropich et al. 2006). Located at the N-terminus, the BD associates with the membrane via electrostatic interactions (Gorelik, Yang et al. 2011). The CC domain is also involved in dimerization, and with the DD, enables membrane localization of the mDia formins (Copeland, Green et al. 2007).

1.4.3 Localization and function of mDia formins

1.4.3.1 Formin function determines their intracellular localization

Expression of most formins (including those in the Dia family) is similar in tissues with mesenchymal, epithelial, and neural origins (Krainer, Ouderkirk et al. 2013).

Just as actin is a crucial component of a variety of cellular structures and processes, the mDia formins are essential for the actin remodeling that enables cell adhesion, cytokinesis, cell polarity, and tissue morphogenesis. The diﬀerent isoforms of mDia formins diﬀer in their actin-polymerizing functions in the cell. The formin mDia1 or

DRF1 polymerizes actin ﬁlaments that underlie stress ﬁbers (Watanabe, Kato et al.

1999, Hotulainen and Lappalainen 2006), mechanotransduction, cell polarization, migration, axonal outgrowth, exocrine vesicle secretion in apical membrane (Kuhn and

Geyer 2014). Together with another formin called Fmnl3, mDia1 has also been implicated in maintaining monolayer integrity during collective cell migration (Rao and

Zaidel-Bar 2016). Both mDia1 and mDia2 speciﬁcally mediate endocytosis along with intracellular endosome traﬃcking, possibly by regulating formation of the endosome

43 actin coating, which prevents fusion (Fernandez-Borja, Janssen et al. 2005, Wallar,

Deward et al. 2007, Bartolini, Moseley et al. 2008). mDia2 also polymerizes F-actin

ﬁlaments that underlie ﬁlopodia (Peng, Wallar et al. 2003, Pellegrin and Mellor 2005,

Beli, Mascheroni et al. 2008), and is involved in endosome traﬃcking (Wallar, Deward et al. 2007). It is also involved in cytokinesis, nucleation of erythroblasts, and cell movements during gastrulation (Kuhn and Geyer 2014). Finally, mDia3 (or DRF2) polymerizes F-actin which that underlies stress ﬁbers (Yasuda, Oceguera-Yanez et al.

2004, Kuhn and Geyer 2014), and is also involved in endosome traﬃcking (Gasman,

Kalaidzidis et al. 2003, Kuhn and Geyer 2014).

Depending on their function, mDia formins can localize to the cell membrane, the perinuclear region, or the nucleus. To localize to the membrane, they interact with membrane-associated Rho GTPase, or directly via membrane binding motif (poly- blastic clusters towards the N-terminus) of mDia1 and mDia2 that form electrostatic interactions with membrane phospholipids (Ramalingam, Zhao et al. 2010, Gorelik,

Yang et al. 2011). Indeed, various post-translational modiﬁcations such as phosphorylation and myristoylation enable membrane localization of formins (as reviewed by Griksheit and Grosse 2016). Depending on the formin, membrane localization can also occur through FH3-IQGAP1 interactions (in mDia1) (Brandt, Marion et al.

2007), or FH1-SH3 interactions with scaﬀolding proteins with BAR domains (Frost,

Unger et al. 2009, Kuhn and Geyer 2014). Also at the cell periphery, as mentioned previously, formins are essential for actin polymerization underlying membrane protrusions. Speciﬁcally, mDia2 promotes lamellipodial protrusion through recruitment of Ena/VASP (Bogdan, Schultz et al. 2013). Compared to the FH2 domain of mDia1, the FH2 of mDia2 can also bundle F-actin to promote protrusive structures like ﬁlopodia ((Peng, Wallar et al. 2003, Schirenbeck, Bretschneider et al. 2005, Har- ris, Rouiller et al. 2006). In migrating cells, formins are necessary for formation of lamella and focal adhesions during cell migration (Gupton, Eisenmann et al. 2007).

44 They also polymerize actin in potentially invasive structures like lamellipodia and

ﬁlopodia (Yang, Czech et al. 2007, Goh and Ahmed 2012).

Both mDia1 and mDia2 also localize to the nucleus, with eﬀects on transcription

(Baarlink, Wang et al. 2013, Kuhn and Geyer 2014). One mechanism by which it does so is through the serum response factor (SRF). Actin polymerization leads to decreased G-actin, which is sensed by the SRF cofactor MAL, which binds G- actin. Reduced G-actin leads to nuclear accumulation of MAL, SRF activation, and transcription of genes encoding for cytoskeleton proteins such as actin, vinculin, and

β1-integrin, as well as early growth response-1 (Egr1), a protein involved in tumor suppression (Eisenmann, Dykema et al. 2009, Baarlink, Brandt et al. 2010).

1.4.3.2 Formins in AJs

Formin localization to AJs can occur upon Rho-GTPase activation. This is ex- empliﬁed by the localization of Dia1 (the human mDia1 homologue) to AJs of MCF7 epithelial cells upon activation by RhoA (Carramusa, Ballestrem et al. 2007). Sa- hai and Marshall also showed that mDia1, but not ROCK acts to stabilize junctions downstream of Rho (Sahai and Marshall 2002). Similarly, FMNL-2 localizes to AJs of Madin-Darby canine kidney (MDCK) cells when activated by Rac1 (Grikscheit,

Frank et al. 2015). Depletion of mDia formins in MDCK cells inhibits cell-cell junction formation (Xing, Wang et al. 2007). Formin-1 was also shown to localize to AJs of keratinocytes, and interact with α-catenin to stabilize junctions (Kobielak, Pasolli et al. 2004). FMNL3 is also necessary for actin polymerization and stabilization of

AJs in endothelial tissue (Phng, Gebala et al. 2015). More recently, Rao and Zaidel-

Bar showed that formin inhibition led to increased cell spreading with concomitant loss of lateral AJs (Rao and Zaidel-Bar 2016). In all of these cases, junctional localization of formins was associated with their roles in junction formation and maintaining epithelial integrity either in monolayers or 3D structures with central lumens.

45 One proposed mechanism for formin-mediated AJ stabilization is through the actin-dependent localization of AJ components such as E-cadherin and catenins to the junctions. E-cadherin is known to be immobilized at cell-cell junctions via binding interactions through the extracellular domain or interaction with actin at the cytoplasmic domain (Erami, Timpson et al. 2015). Rao and Zaidel-Bar showed that though depletion of Fmnl3 in Eph4 mammary cells resulted in an increase in G-actin at the expense of junctional F-actin levels, the reverse resulted from Fmnl3 overexpression (Rao and Zaidel-Bar 2016). As expected with the decrease in F-actin, depletion of Fmnl3 resulted in punctate AJs and decreased junctional E-cadherin levels.

Previously, Sahai and Marshall had also showed that mDia1 activity promotes association of α-catenin with the junctional E-cadherin/β-catenin complex in HEK293 cells

(Sahai and Marshall 2002). They also demonstrated that the actin-polymerizing function, and not the microtubule-organizing ability of mDia1 is speciﬁcally responsible for this.

Another proposed mechanism for formin-mediated AJ stabilization is through shifting of the balance between non-branched and branched actin polymerization at the junctions. Formin inhibition with the pan-formin small molecule inhibitor of the FH2 domain (SMIFH2) led not only to the reduction of AJ E-cadherin levels but also increased Arp2/3-mediated cell spreading (Rao and Zaidel-Bar 2016). This harkens to the observed interaction between α-catenin and formins, in which the former helps to recruit actin nucleators such as the latter to the AJ. Indeed, in the absence of α-catenin, Formin-1 could not localize to junctions in vivo (Kobielak, Pa- solli et al. 2004). Homodimers of α-catenin have also been observed in vitro to inhibit

Arp2/3-mediated branched actin polymerization in favor of linear actin polymerization by proteins such as the formins (Drees, Pokutta et al. 2005). Interestingly, while most cases of formin-mediated AJ stabilization involves junctional localization of the formin, some examples show that junctional localization may not be essential for this

46 function in some formins. Rao and Zaidel-Bar showed that while mDia1 is diﬀusely localized in Eph4 mammary epithelial cells, it is nevertheless essential for AJ stability, and over-expression of its DAD domain led to increased levels of F-actin and junctional E-cadherin (Rao and Zaidel-Bar 2016). Together, these ﬁndings indicate that formins can be regulated by α-catenin and work to indirectly promote AJ stability through actin polymerization.

1.4.3.3 Integrative functions of mDia formins

Given its ability to bind to and regulate both F-actin and microtubules, mDia formins are uniquely endowed with the ability to coordinate formation of complex cytoskeleton structures. Formin-coordinated parallel alignment of microtubules and

F-actin in some cells such as HeLa cells is important for changes in morphology

(Gasteier, Schroeder et al. 2005). Individual formins have been implicated in speciﬁc cellular processes. For example, mDia2 localizes to mitotic spindles of HeLa cells

(Kato, Watanabe et al. 2001). F-actin polymerized by mDia2 at the contractile ring is necessary for cytokinesis (Watanabe, Ando et al. 2008). Meanwhile, mDia1, which helps to mediate macrophage phagocytosis, interacts with and is localized to phagocytic cups by the microtubule-binding protein CLIP-170 (as reviewed in

Bartolini and Gundersen 2011). With regards to the AJ in particular, Sahai and

Marshall found that while microtubule disruption did not prevent junctional α-catenin localization, it did not result in a more dynamic cell periphery (Sahai and Marshall

2002). However, while actin disruption by itself did prevent α-catenins localization to the junctions, it did not have an eﬀect on cell border dynamics (Sahai and Marshall

2002). These data suggest separate and potentially complimentary eﬀects of actin and microtubules in AJ dynamics, and simultaneously supporting the role of formins

(which regulate both) in AJ formation and stabilization.

47 1.4.3.4 Regulation by Rho-GTPases

The various functions of formins are regulated and coordinated spatially and tem- porally by the family of Ras homolog (Rho) small proteins with molecular weights of around 21 kDa. Their distinct active and inactive states when bound to GTP or GDP allow the Rho-GTPases to act as a molecular switch. Cycling between GTP and GDP is controlled by guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs), with the former responsible for Rho activation via displacing the bound GDP in favor of GTP binding (as reviewed in (Porter, Papaioannou et al.

2016)). On the other hand, GAPs promote GTP hydrolysis and Rho-GTPase inactivation (Porter, Papaioannou et al. 2016). While activation of certain members of the Rho-GTPase family has been associated with tumorigenesis, activation of others such as Rac1 has been shown to inhibit cancer progression (as reviewed in Porter et al. 2016). As RhoA-D, Rac, and Cdc42 have all been shown to activate mDia2

(Lammers, Meyer et al. 2008), it is important understand their functions in the cell.

In mammals, there are 23 Rho-GTPases and their eﬀectors include, but are not limited to, mDia formins along with ROCK, Rhotekin, Rhophilin, WASp, PKN and

Citron-K (as reviewed in Thumkeo et al. 2013). It is well established that RhoA, which activates mDia formins, is important in promoting actin stress ﬁber formation. It is also important in cytokinesis, transcription, cell transformation, cell cycle progression, cell migration, cell-cell junctions, and vesicular traﬃcking among other key cellular processes (Hill, Wynne et al. 1995, Olson, Ashworth et al. 1995, Qiu,

Chen et al. 1995, Braga, Machesky et al. 1997, Ridley 2006). The importance of RhoA in maintaining AJs is seen in the loss of AJs and apical-basal polarity of neuroepithelium in mice with conditionally depleted of RhoA (Herzog, Loetscher et al. 2011, Katayama, Leslie et al. 2012). It is also worth noting that RhoA levels increase with maturation of ovarian cancer spheroids in vitro, suggesting that

RhoA coordinates AJ stability. The proposed mechanism is that higher levels of

48 RhoA promote ROCK-mediated invasive transitions while simultaneously suppress- ing mDia2-mediated junction-stabilization.

RhoA-C share 85% amino acid sequence similarity (Thumkeo, Watanabe et al.

2013), yet are responsible for diﬀering functions within the cell. For instance, RhoB is primarily localized to endosomes compared to RhoA and RhoC, which localize to the cytoplasm and plasma membrane (Adamson, Paterson et al. 1992, Zalcman, Clos- son et al. 1995). Indeed, RhoB together with mDia2, when localized to endosomes, are necessary for vesicle traﬃcking (Wallar, Deward et al. 2007). Interestingly, RhoB may have an anti-metastatic role as its depletion has been associated with increased cell migration (Vega, Colomba et al. 2012, Ridley 2013). Meanwhile, RhoC has been particularly implicated in cancer cell invasion and EMT (as reviewed in Thumkeo et al. 2013) and over-expression of RhoC induced ROCK-dependent membrane blebbing along with disruption of junctional localization of α-catenin (Sahai and Marshall

2002). Cell invasion can occur through the RhoC activation of the formin FMNL3

(Vega, Fruhwirth et al. 2011). In contrast, more recently, FMNL3 was found to be recruited to AJs downstream of Cdc42 and Src kinase, and its depletion led to dissociation of leader cells from migrating monolayers (Rao and Zaidel-Bar 2016). This indicates the complex and varying eﬀects of Rho-GTPases on their formin eﬀectors.

To promote cell migration and invasion, the Rho-GTPases speciﬁcally interact with a diverse group of downstream eﬀectors. For example both Cdc42 and Rif

GTPases regulate mDia2 to regulate ﬁlopodia formation (Peng, Wallar et al. 2003,

Baarlink, Brandt et al. 2010, Goh and Ahmed 2012). At the same time, Cdc42 also interacts with and regulates neuronal Wiskott-Aldrich Syndrome protein (N-

WASp) (Carlier, Ducruix et al. 1999), which in turn regulates Arp2/3 whose increased expression has also been linked to cancer metastasis (Ellenbroek and Collard 2007).

Other Rho-GTPase eﬀectors, such as ROCK, which promotes actomyosin contraction and rear end retraction during cell migration, have also been implicated in cancer cell

49 invasion (Manning, Parker et al. 2000, Rathinam, Berrier et al. 2011).

1.4.3.5 Dia-interacting protein (DIP) and the cell cycle in mDia2 regu-

lation

Aside from regulation by the Rho-GTPases, mDia2 is uniquely inactivated by

Dia-interacting protein (DIP), a protein involved in stress ﬁber formation, cell migration, vesicular traﬃcking, and focal adhesion (FA) assembly (Satoh and Tominaga

2001, Meng, Numazaki et al. 2004, Wyse, Lei et al. 2012). Though it binds both mDia1 and mDia2, DIP only inhibits mDia2 FH2-mediated F-actin polymerization and bundling, resulting in decreased ﬁlopodia formation and non-apoptotic membrane blebbing (Eisenmann, Harris et al. 2007). This has been attributed to disruption of cortical actin. Further studies implicate the mDia2:DIP axis in amoeboid morphological transitions in MDA-MB-231 breast cancer cells (Wyse, Lei et al. 2012). Once again, this highlights the role for mDia formins in cancer invasion and metastasis.

Levels of mDia2 are also regulated by ubiquitination followed by degradation such as during the process of mitosis (DeWard and Alberts 2009). Indeed, mDia2 expression is cell cycle-dependent, with stable expression through the S- and G2/M phases, and poly-ubiquitination following M phase and reduced levels in G0/G1 (DeWard and Alberts 2009). Thus the levels of mDia2, and not just its activity, are under tight temporal regulation.

1.5 Formins in development and disease

1.5.1 Formins in development

As actin polymerization is crucial for cell migration and cytokinesis, it is unsurprising that they are important in normal growth and development. For instance,

Formin 1 may be essential for limb development as evidenced by reduction of dig- 50 its in knockout mice (Zhou, Leder et al. 2009). Depletion of Formin 1 isoform IV

(Fmn1-IV) in mice resulted in mouse embryonic ﬁbroblasts (MEFs) with disrupted cell spreading and focal adhesion (FA) assembly (Dettenhofer, Zhou et al. 2008).

Interestingly, some of these mice had kidney aplasia (Wynshaw-Boris, Ryan et al.

1997). Hu et al. also showed that Formin 1 interacts with ﬁlamin B to coordinate chondrocyte proliferation and diﬀerentiation in the growth plate in mice (Hu, Lu et al.

2014). In mice, Daam1 is also important in heart morphogenesis while in Drosophila and Xenopus it contributes to trachea formation and gastrulation respectively (Li,

Hallett et al. 2011).

Formins may also have an eﬀect on fertility, hearing, and olfaction. Depletion of

Fmn2 in mice interferes with meiosis, the process responsible for generation of oocytes

(Leader, Lim et al. 2002). Interestingly, certain mutations in DIAPH2 (encoding a mutant mDia3) can result in premature ovarian failure while mutations in DIAPH1

(encoding a mutant mDia1) can result in autosomal dominant non-syndromal sen- sorineural hearing loss (Lynch, Lee et al. 1997, Bione, Sala et al. 1998). Depletion of mDia1 and mDia3 in mice also results in disruptions in migration of olfactory inhibitory interneurons (Shinohara, Thumkeo et al. 2012).

1.5.2 Formins in immunity and homeostasis

Formins are important for immunity due to their eﬀects on migration and adhesion of immune cells such as T cells and neutrophils. In mice, DIAPH1 deletion led to disruptions in T-cell development, proliferation, migration to lymphoid organs like the spleen and lymph nodes (Eisenmann, Harris et al. 2007, Sakata, Taniguchi et al. 2007). In mice, lymphoid cells require mDia1 for migration and chemotaxis

(Eisenmann, West et al. 2007). However, expression of the constitutively active form of mDia1 (with truncated GBD and partially truncated DID domains) resulted in actin accumulation in Jurkat T cells with decreased chemotaxis with TCR engage-

51 ment, again suggesting that mDia1-mediated cytoskeleton dynamics are essential for proper T cell function (Vicente-Manzanares, Rey et al. 2003). Similar to T-cells, neutrophil polarization and chemotaxis is dependent on mDia1 as part of Src family kinase and LARG/RhoA/ROCK signaling pathways (Zhang, Shehabeldin et al. 1999,

Peng, Wallar et al. 2003). Thus mDia formins play an essential role in the regulation of immunity.

In addition to immune cells, erythrocytes and platelets require formins for proper function. Previously we discussed the importance of mDia2 in cytokinesis. Thus, it is not hard to imagine that mDia2 is also essential for formation of the contractile actin ring that is required for the process of enucleation, where cells such as erythrocytes precursor (erythroid) cells extrude their nuclei (Ji, Jayapal et al. 2008). Meanwhile, thrombin-activated platelets also rely on RhoA activation of mDia1 and DAAM1 for actin remodeling and cell spreading (Higashi, Ikeda et al. 2008, Gao, Chen et al.

2009).

1.5.3 Formins in cancer

In addition to cell migration, certain formins are involved in cell proliferation.

In humans, 5q chromosomal deletion with concomitant deletion of mDia1, results in myelodyplastic syndrome in some patients (Peng, Kitchen et al. 2007, Eisenmann,

Dykema et al. 2009). In humans, myelodysplastic syndrome is considered a cancer characterized by decreased production of various types of cells found in the blood, including red and white blood cells, often resulting in anemia, recurrent infections, and increased risk of leukemia. Deletion of mDia1 in mice results in a myelodysplastic syndrome with impaired migration and adhesion of leukocytes (Peng, Kitchen et al.

2007, Tanizaki, Egawa et al. 2010). The idea that mDia1 regulates proliferation of hematopoietic stem cell is further supported by the fact that these mice also developed hypercellular bone marrow with increased spleen and bone marrow production of

52 myeloid and erythroid cells (Peng, Kitchen et al. 2007). Finally, an association between increased expression of FMNL1 and lymphoid malignancies such as chronic lymphocytic leukemia (CLL) has also been observed (Favaro, de Souza Medina et al.

2003).

Given their roles in junction stabilization, cell polarity and migration, various members have been implicated in tumorigenesis and disease progression of solid tumors. In adenocarcinoma cells depleted of N-WASp, RhoA-activated mDia1 mediates stress ﬁber and membrane protrusion formation, structures that promote motility

(Sarmiento, Wang et al. 2008). In glioma cells, mDia1 regulates FA assembly and migration (DeWard, Eisenmann et al. 2010). Dia1-mediated focal adhesion stabilization promotes invasive protrusions into collagen (Fessenden, Beckham et al. 2018).

Meanwhile, mDia2 has been implicated in invasion and metastasis of various cancers, including breast, cervical, and prostate cancers (DeWard, Eisenmann et al. 2010).

For instance, depletion of mDia2 in breast cancer cells resulted in reduction of formation of the pro-invasive actin-based membrane protrusions known as invadopodia, again implicating mDia2 in cancer progression (Lizarraga, Poincloux et al. 2009).

As previously mentioned, various studies have shown that the mDia formins are involved in endosome traﬃcking as well as non-apoptotic membrane blebbing (Wallar and Alberts 2003, Eisenmann, Harris et al. 2007, Wallar, Deward et al. 2007).

Thus it is unsurprising that studies also suggest their involvement in formation of microvesicles- 50-nm membrane vesicles extruded from the cell that contain signaling proteins and microRNAs (as reviewed in DeWard et al. 2010). Indeed, Di Vizio et al. observed formin-mediated microvesicle (oncosome) production, with enhanced cancer metastasis in EGF-induced prostate cancer cells (Di Vizio, Kim et al. 2009).

The mechanism for formin-mediated cancer cell migration and invasion is a topic of active investigation. For example, increased FMNL2 expression is associated with metastatic colorectal cancer cells and lymph node metastases (Li, Zhu et al. 2010).

53 However, there is no evidence to indicate whether a causal relationship exists between

FMNL2 expression and colorectal cancer progression. More recently, Kim et al. also found that mDia1 is upregulated in invasive breast cancer cells (Kim, Jung et al.

2016). Enhanced invasion proposed to occur through mDia1-mediated traﬃcking of membrane type 1 (MT1)-MMP to the plasma membrane via microtubules (Kim,

Jung et al. 2016). In addition, the authors found that in cells with reduced mDia1 expression, there was also reduced expression of certain stem cell markers like CD44 and CD133 (Kim, Jung et al. 2016). This suggests a role for formins in expression of cancer stem cell genes, with further implications for the eﬀects of formins in cancer progression.

Though many studies implicate formins in tumorigenesis and cancer metastasis with a focus on their role in migration and invasion, others show that they may also function as tumor suppressors. For example, mDia proteins associate with the known tumor suppressor adenomatous polyposis coli (APC), and certain members are hypothesized to act downstream of this protein to prevent tumorigenesis (Wen, Eng et al. 2004, Peng, Kitchen et al. 2007). Also, expression of interfering mDia1 (with truncated mDia1 FH2 domain) in Ras-transformed MEFs increased their tumorigenic capacity (Kamasani, Duhadaway et al. 2007). At the same time, mDia2 depletion in ovarian cancer spheroids resulted in amoeboid invasive transitions in vitro (Pettee,

Dvorak et al. 2014). Furthermore, triple negative breast cancer (TNBC) tissue isolated from patients demonstrate signiﬁcantly reduced DIAPH3 expression, which also correlated with increased metastasis (Jiang 2017). In accordance with this, the authors found that DIAPH3 over-expression suppressed migration and invasion of

MDA-MB-231 breast cancer cells. Loss of DIAPH3 is also associated with 20% of primary and 60% of metastatic human prostate tumors, along with breast cancer and hepatocarcinoma progression (Di Vizio, Kim et al. 2009, Wyse, Lei et al. 2012). Di

Vizio et al. showed that depletion of mDia2 in EGF-stimulated DU145 and LNCaP

54 prostate cancer cells resulted in increased membrane blebbing with increased motility and invasion (Di Vizio, Kim et al. 2009). Finally, loss of mDia2 has been associated with acquisition of the amoeboid phenotype and tumor progression in prostate, breast, and hepatocellular carcinomas (Hager, Morley et al. 2012). Further, mDia2 silencing in prostate cancer cells promoted shedding of extracellular vesicles (EV) containing factors that enhance cancer cell proliferation and suppress tumor-inﬁltrating immune cells (Kim, Morley et al. 2014).

While some studies found increased formin expression to correlate with increased invasive capacity, in certain contexts formin expression may be associated with less invasive phenotypes. For example, Rao and Zaidel-bar used semi-quantitative RT-PCR to analyze levels of FMNL3 in non-transformed, epithelial, mesenchymal, and intermediate/transitioning ovarian cancer cell lines. They observed the highest expression in non-transformed and epithelial ovarian cancer cells with minimal expression in mesenchymal cell types (Rao and Zaidel-Bar 2016). Thus, beyond tumorigenesis and migration, formin activity may have eﬀects on cancer cell morphology with implications for chemoresistance and metastatic potential.

1.5.4 Formins as therapeutic targets

1.5.4.1 Formin inhibition

Rizvi et al. ﬁrst identiﬁed the small molecule inhibitor of the FH2 domain

(SMIFH2), a 2-thiooxodihydropyrimidine-4,6-dione derivative that inhibits FH1FH2- mediated assembly of profiling-bound G-actin, most likely targeting the FH2 domain in formins (Rizvi, Neidt et al. 2009). It has been observed to specifically inhibit formin-mediated actin assembly and stress fiber formation without affecting spontaneous or Arp2/3-mediated actin polymerization. SMIFH2 exerts concentration- dependent effects on NIH 3T3 fibroblasts, which vary from growth inhibition and

55 cytokinesis deﬁcits over 4 days at 2.5 µM, to membrane blebbing at 10 µM. A549 lung carcinoma cells required higher dose (30 µM) of SMIFH2 for growth inhibition and induction of membrane blebbing. These results are consistent with the idea that mDia inhibition results in disrupted cortical actin accompanied by membrane blebbing and amoeboid motility (Eisenmann, Harris et al. 2007, Wyse, Lei et al.

2012). In spite of its induction of blebbing, Rizvi et al. showed signiﬁcantly reduced migration rates in SMIFH2-treated NIH 3T3 cells (Rizvi, Neidt et al. 2009).

Since the publication of the ﬁndings of Rizvi et al. in 2009, other studies have sug- gested that instead of inhibiting migration and invasion, formin inhibition may not be an ideal anti-invasive strategy for cancer cells. For instance, treatment of ES2 ovarian cancer cells with SMIFH2 led to formation of disorganized spheroids with loose cells in the periphery supporting the concept that formin inhibition has disruptive eﬀects on cell-cell junctions, with implications for cancer dissemination (Pettee, Dvorak et al.

2014). Arden et al. also showed that in glioblastoma (GBM) cells, mDia inhibition reduced persistent migration but did not disrupt random migration (Arden, Lavik et al.

2015). This eﬀect suggests a role for mDia formins in Rho-GTPase-mediated maintenance of cell polarity (Fukata, Nakagawa et al. 2003). Finally, SMIFH2 treatment was also observed to increase migration and prevent mitosis in U2OS osteosarcoma cells (Isogai, van der Kammen et al. 2015).

Discrepancies between the eﬀects of SMIFH2 on cell migration (as observed in

NIH 3T3 and U2OS cells) has been attributed to different concentrations of the reagent used, differences in formin expression amongst cell lines, and/or differences in experimental set up (Isogai, van der Kammen et al. 2015). Indeed, in the literature from 2009 to 2014, SMIFH2 concentrations ranging from 100 nM to 100 µMhavebeen used, with 10-30 µMconcentrationsbeingthemostcommonlyused(Isogai,vander

Kammen et al. 2015). Treatment times range from 490 seconds to 72 hours (Isogai, van der Kammen et al. 2015).

56 Besides their effects on cell migration, formin inhibition has effects on cell-cell junctions, p53 expression, and microtubule stabilization. The first of these has also been exploited as potential method to enhance penetration of standard chemotherapy into 3D tumor structures. For example, it has been observed that SMIFH2 treatment potentiates the anti-viability effects of cisplatin in spheroids formed from ES2 cells

(Ziske, Pettee et al. 2016). Meanwhile, Isogai et al. found that at certain concentrations, SMIFH2 can inhibit p53 transcriptional activity separately from its effects on mDia2, and the fact that expression of p53 did not correlate mDia2 expression suggests independent regulation (Isogai, van der Kammen et al. 2015). However, as actin depolymerization is known to affect p53 expression and activity and formins are prominent actin nucleators, mDia inhibition may yet also have an indirect effect on this key tumor suppressor (Rubtsova, Kondratov et al. 1998). Finally, SMIFH2 treatment of Arabidopsis cells decreased microtubule density (Rosero, Zarsky et al. 2013).

This is unsurprising given the role of formins in microtubule stabilization. These eﬀects are all important to bear in mind when considering SMIFH2 as a therapeutic agent.

1.5.4.2 Formin activation

Compared to formin inhibition, fewer studies have explored the eﬀects of formin agonism on cells. Two types of agents have been studied: (1) the small molecule mDia2 intramimics IMM-01 and IMM-02, and (2) the mDia2 DAD peptides. The

IMMs activate mDia formins through disruption of the DID-DAD auto-inhibitory interaction (Lash, Wallar et al. 2013). Treatment of ﬁbroblasts and colon cancer cells with these compounds result in microtubule stabilization, cell-cycle arrest, and apoptosis (Lash 2013). In mice, mDia activation has been shown to reduce colon cancer growth and proliferation (Lash, Wallar et al. 2013). Also, treatment of glioblastoma

(GBM) cells with formin activators abrogated both directional and random migration

57 in two-dimensional (2D) models and successfully inhibited GBM invasion in spheroid assays and in ex vivo living rat brain slice invasion assays (Arden, Lavik et al. 2015).

The eﬀect of IMM treatment on microtubules is unclear. While Arden et al. observed that IMMs did not stabilize microtubules in cultured GBM cells, Lash observed microtubule stabilization in mouse embryonic ﬁbroblasts (NIH3T3 cells) and metastatic colon adenocarcinoma cells (SW450 cells) (Lash, Wallar et al. 2013, Arden,

Lavik et al. 2015). As in formin inhibition, discrepancies may be due to diﬀerences in the cell lines used as well as experimental set ups. Further studies are needed to determine the eﬀects of IMMs and mDia2-DADs on microtubule dynamics.

Currently, a major advantage of formin agonism (via either IMMs or mDia2-

DADs) over formin inhibition lies in their significantly reduced in vivo toxicity compared to SMIFH2 at functional concentrations. Indeed LeCorgne et al. observed significant morphological defects in developing zebrafish treated with 5-10 µM SMIFH2, ≥ while treatment with IMM and mDia2-DADs at 10-20 and 50 µM, respectively, had minimal toxic effects (LeCorgne, Tudosie et al. 2018). Lash et al. also showed that

IMMs decreased tumor burden in mouse models of hepatocellular carcinoma with limited toxicity (Lash, Wallar et al. 2013). These ﬁndings along with their inhibitory eﬀects on cell growth and motility make formin agonism an attractive anti-metastasis strategy.

1.6 Summary of Past Findings and Gaps in the

Knowledge

EOC is unique amongst cancers in that its main route of dissemination is transcoelomic.

Cancer cells are shed from the primary tumor as single cells or clusters into the peritoneal cavity where they are carried with the accumulated peritoneal ﬂuid (ascites) to various secondary sites such as the liver and diaphragm. Current standard-of-care

58 therapies like cisplatin and paclitaxel target cell proliferation with limited success, especially in chemoresistant disease. The unique route of EOC dissemination has not been exploited in these therapies, which can partly be attributed to the lack of understanding of the molecular interactions that preserve the structural integrity of primary tumors and the spheroids they shed.

Previous published work from our lab revealed that EOC spheroids are held together by cell-cell junctions lined by F-actin and, peripherally, mDia2. Adherens junctions are a type of junctions speciﬁc to epithelial cells including EOCs. AJs, which consist of homotypic linkages between cadherins on neighboring cells, provide mechanical stabilization for EOC spheroids. While it is known that the cadherins and catenins are the hallmark proteins of AJs, the proteins that assemble the F-actin network associated with these junctions is yet unclear. The mDia formins are essential actin nucleators known to polymerize various actin-based structures in the cell, including membrane protrusions and stress ﬁbers. Certain members of the formin family have already been known to localize to AJs various cell types, often upon activation by Rho-GTPases. Some interact with AJ proteins such as α-catenin. Yet, the role of mDia formins in AJ formation and stabilization in EOC is unclear.

1.7 Hypothesis

The hypothesis of this work is that mDia2 is important for AJ formation and stability in EOC. Just as it was observed to colocalize with junctional actin in EOC spheroids, mDia2 is expected to be an essential nucleator of junctional actin as well as stress fibers. mDia2 is also expected to interact with the catenins, affecting junctional stability by regulating their localization. Using cluster dissociation assays, immunofluorescence microscopy, and immunoprecipitation assays, we investigate the role of mDia2 AJ formation and stabilization. These experiments show that mDia2

59 interacts with select junctional players and is essential for the mechanical integrity of the AJ in EOC cells.

60 Chapter 2

Results

mDia2 formin selectively interacts with catenins and not E-cadherin to

regulate Adherens Junction formation

1 1, Yuqi Zhang and Kathryn M. Eisenmann ∗ 1Department of Cancer Biology, University of Toledo Health Science Center, 3000

Arlington Ave, Toledo, OH 43614 USA.

([email protected]; [email protected])

*corresponding author

Submitted - BMC- Molecular and Cell Biology

2.1 Introduction

Ovarian cancer is the deadliest gynecological malignancy, with 14,070 women estimated to die from the disease in the United States in 2018, according to the American

Cancer Society. There are at least four types of ovarian cancer classiﬁed by their cell type of origin, with epithelial ovarian cancer (EOC) representing 90% of diagnosed cases (Rojas, Hirshﬁeld et al. 2016). Most patients are diagnosed in the late stages, when peritoneal dissemination has already occurred (Karst and Drapkin 2010), and

75% of patients develop relapsing disease after undergoing current standard of care 61 treatment with cytoreductive surgery and adjuvant platinum/taxane duplet combinations (Romero and Bast 2012).

While hematogenous metastasis does occur, EOCs primary route of dissemination is transcoelomic (Pradeep, Kim et al. 2014, Yeung, Leung et al. 2015). Cancer cells are shed from the primary tumor as single cells or clusters into the peritoneal cavity where they are carried with the accumulated peritoneal ﬂuid (ascites) to secondary sites such as the liver and diaphragm (as reviewed in (Yeung, Leung et al.

2015)). Since cancer cell detachment from the primary tumor initiates metastasis, it is important to understand the molecular mechanisms involved in this process.

A type of cell-cell junction present in epithelial cells is the Adherens Junction (AJ), which typically consists of E-cadherin and p120, α-, and β-catenin (Gumbiner 2005,

Yamada, Pokutta et al. 2005, Yonemura, Wada et al. 2010). AJ formation is calcium- dependent. In the presence of calcium, the extracellular domains of E-cadherin change conformation to allow for homotypic linkages between cells whose adjacent membranes are 20 nm apart (Vasioukhin, Bauer et al. 2000, Yonemura, Wada et al. 2010).

AJ stability is dependent on AJ protein anchorage to the underlying cortical F- actin cytoskeleton and continuous F-actin polymerization (Hong, Troyanovsky et al.

2013, Grikscheit and Grosse 2016). While the classical AJ indicates direct linkage between the E-cadherin/β-catenin complex to F-actin via α-catenin (which is herein interchangeably referred to as epithelial α-catenin or αE-catenin) , binding to β- catenin may reduce the aﬃnity of α-catenin for F-actin (Drees, Pokutta et al. 2005,

Yamada, Pokutta et al. 2005). A recent alternative model proposes that α-catenin cycles between two intracellular pools- one junctional, where it associates with the cadherin complex-and another cytosolic or perijunctional that associates with the underlying F-actin (Pokutta, Drees et al. 2008). The two α-catenin pools may allow for α-catenin to act as a molecular switch- turning oﬀArp2/3-mediated branched actin-polymerization to promote linear actin polymerization (Gates and Peifer 2005,

62 Pokutta, Drees et al. 2008). The exact role of cortical F-actin in the adherens junction and its regulation by various catenins and actin-associated proteins is still a subject undergoing active investigation.

Various actin-associating and/or polymerizing proteins are known to aﬀect AJ stability and are regulated by the catenins. Among these are various members of the formin family of proteins, which act as downstream eﬀectors of Rho-GTPases.

Upon binding of a Rho-GTPase to the GTPase-binding domain (GBD) of the formin, the formin is released from its autoinhibited state, promoting non-branched F-actin polymerization, and, in some cases, bundling (Firat-Karalar and Welch 2011). Formin family members such as formin-1, Diaphanous-related formin 1 (mDia1), and Fmnl3 localize to and/or strengthen the AJ by increasing E-cadherin junctional localization while decreasing its mobility within the plasma membrane (Carramusa, Ballestrem et al. 2007, Grikscheit and Grosse 2016, Rao and Zaidel-Bar 2016). Formin localization to AJs was regulated through Rho-signaling and post-translational modiﬁcations such as phosphorylation and myristoylation (Takeya, Taniguchi et al. 2008, Chesarone,

DuPage et al. 2010, Cheng, Zhang et al. 2011, Kuhn and Geyer 2014, Rao and

Zaidel-Bar 2016). Formins may also be mechanosensitive, responding to external forces. mDia1-mediated F-actin polymerization increased upon pN force application to actin ﬁlaments (Jegou, Carlier et al. 2013).

Both formin inhibition and activation were associated with increased cellular invasion. For example, mDia formins were required for formation of invadopodia and invasion by MDA-MB-231 breast adenocarcinoma cells (Lizarraga, Poincloux et al.

2009). Indeed, both mDia1 and mDia2 localize to ﬁlopodial tips in various mammalian cell lines including tumor cells (Peng, Wallar et al. 2003, Higashida, Miyoshi et al. 2004, Pellegrin and Mellor 2005, Yang, Czech et al. 2007, Sarmiento, Wang et al. 2008). Previously, mDia1 depletion inhibited Src accumulation at focal adhesions, with eﬀects on adhesion stability as well as cellular migration (Yamana, Arakawa et

63 al. 2006, DeWard, Eisenmann et al. 2010). Furthermore, mDia2 speciﬁcally regulated epithelial cell migration by localizing to the lamella of migrating epithelial cells to polymerize and maintain cortical actin (Gupton, Eisenmann et al. 2007). How- ever, suppression of both intrinsic and direction migration of U87 glioblastoma cells from spheroids in an ex vivo model was seen upon mDia agonism (Arden, Lavik et al.

2015). In accordance with these ﬁndings, deletion of the DIAPH3 locus (encoding mDia2) was associated with metastatic disease through its regulation of amoeboid migration in prostate cancer (Di Vizio, Kim et al. 2009, DeWard, Eisenmann et al. 2010). Indeed, mDia2 was shown to be important for maintaining membrane integrity, as its inhibition by Dia-interacting protein (DIP) led to membrane blebbing and amoeboid motility (Eisenmann, Harris et al. 2007). In the context of ovarian diseases, disruption of other formins (e.g., mDia3) was associated with eﬀects on ovarian development and premature ovarian failure (Castrillon and Wasserman 1994, Bione,

Sala et al. 1998, DeWard, Eisenmann et al. 2010). Together, these ﬁndings indicate a key role for mDia in metastatic processes. However, the mechanism(s) behind invasive transitions in EOC and how it relates to mDia activity and/or localization remains unclear.

In ovarian cancer, decreased E-cadherin expression is associated with invasive peritoneal seeding and is more commonly observed in borderline ovarian tumors and carcinomas compared to benign tumors (Vergara, Merlot et al. 2010). Previously, mDia2 depletion promoted ovarian cancer spheroid invasion by driving single cell invasive egress from the spheroid (Pettee, Dvorak et al. 2014). The exact mechanisms of action remained elusive, however. Here, we show that mDia2 is important for

AJ formation and stability in EOC monolayers. Depletion of mDia2 leads to loss of junctional continuity and decreased resistance to mechanical shearing. We observe interactions between mDia2 and α-andβ-catenins. Interestingly, mDia2 does not interact with either E- or N-cadherin, indicating that it may not directly interact

64 with the classical junctional cadherin complex. Finally, we show that the interactions between mDia2 and the catenins may not be dependent upon an intact F-actin network. Collectively, these data indicate a critical role for mDia formins in regulating

EOC AJs during invasive transitions.

2.2 Results

2.2.1 mDia2 is essential for junction integrity in spheroids.

AkeystepinEOCdisseminationisthesheddingofcancercellsfromtheprimary tumor or secondary metastases into the peritoneal cavity, a process that depends on the disruption of cell-cell junctions. To investigate the role of mDia2 in EOC junction integrity, stable mDia2 shRNA OVCA429 cells co-expressing GFP were generated, along with control shRNA GFP-expressing OVCA429 cells. Knockdown (KD) was conﬁrmed with Western blotting (Fig. 2-1A). E-cadherin expression was unchanged upon mDia2 depletion.

To determine the effect of mDia2 on junctional strength, a hanging drop assay was used to measure cellular resistance to mechanical shear forces (Kim, Islam et al. 2000, Ehrlich, Hansen et al. 2002). Single-cell suspensions of OVCA429 GFP- expressing mDia2 KD or control cells were seeded as droplets and cultured for 30 minutes to 4 hours (Fig. 2-1B). Initially at 30 minutes and 1 hour, control cells formed larger and more loosely-packed spheroids compared to mDia2 KD cells. This is consistent with previous findings indicating that mDia2 depletion in ovarian cancer cells led to increased spheroid compaction (Pettee, Dvorak et al. 2014). However, by 2 hours, mDia2 KD and control cells formed similar sized spheroids (Fig. 2-1C). At the specified time points, mechanical trituration as applied to the spheroids formed. Cell clusters were then enumerated as either >20 cell, 11-20 cells or <11 cells. OVCA429 mDia2-depleted spheroids were less resistant to shear forces than control cells, as the

65 Figure 2-1: Analysis of mDia2 in a functional cell-cell adhesion assay. A. Western blotting demonstrates knockdown of mDia2 in OVCA429-mDia2 KD cells compared to control knockdown OVCA429 cells. B. OVCA429-mDia2 KD and OVCA429-control cells were plated as single droplets (5000 cells/drop) in hanging drop cultures in RPMI media and triturated at speciﬁed time points. Representative ﬁelds at 0.5, 1, 2, and 4 hours before and after trituration are shown. Scale bar = 1000 µm. C. Cluster diameters of OVCA429-mDia2 KD and OVCA429-control cells over 4 hours are shown. D. Images of clusters of OVCA429- mDia2 KD and OVCA429-control cells after trituration. Scale bar = 1000 µm. Experiment was performed 3 times in triplicates. Arepresentativeexperimentofthreeisshown.

numbers of clusters >20 cells were reduced to 9% from 16%, respectively, in 4hrs (Fig

2-1D, 2-2A-B). This corresponded to decreases in cell clusters of 11-20 cells of 26% of

OVCA429-control cells compared to 9% of OVCA429-mDia2 KD at 2 hours. These results suggest that mDia2 may have a role in stabilization of cell-cell junctions in

EOC clusters and resistance to shear forces.

66 Figure 2-2: Quantiﬁcation of functional cell-cell adhesion assay. A-B. Graphs show percentage of cells (OVCA429-control and OVCA429- mDia2 KD cells) in aggregates of 0-10 cells, 11-20 cells, and >20 cells at speciﬁed time points. For each time point, >200 cells were scored.

2.2.2 Role of mDia2 in AJ formation

In AJ formation, the linkage between cadherins on adjacent cells is dependent on calcium (Vasioukhin, Bauer et al. 2000). To determine the eﬀects of mDia2 on AJ formation, a calcium switch assay combined with immunoﬂuorescence (IF) was used to visualize E-cadherin and F-actin in mDia2 KD and control cells. Both mDia2 KD cells and control cells formed AJs when cultured in calcium-containing medium (Fig.

2-3A). Upon culturing in calcium-free media, AJs in both the mDia2 KD and control cells were abolished, but by 4 hours of calcium repletion with calcium-containing media, AJs were beginning to form in both cells (Fig.2-3A). Junctional continuity was then quantiﬁed, where a junction was determined to be continuous if the longest

E-cadherin-positive region was at least 50% the length between two vertices of a cell-cell junction (Carramusa, Ballestrem et al. 2007). The number of continuous

E-cadherin-positive junctions was signiﬁcantly greater for control cells (59% of total junctions) compared to mDia2 KD cells (19% of total junctions) (Fig. 2-3B). These

67 data indicate that mDia2 may not only have an eﬀect in junction integrity, but also in the formation of the AJ in EOC cells.

Figure 2-3: Effect of mDia2 on adherens junction formation. A. OVCA429- mDia2 KD and OVCA429-control cells were plated at 200,000 cells per 35 mm well and grown to 70% confluence in RPMI medium. Cells were then cultured in calcium-free RPMI for 16 hours followed by regular RPMI and processed and imaged at specified times. For control, cells were in fresh RPMI for 16 hours and again for 4 hours. Cells were stained for E-cadherin and F-actin. Representative fields at 0 and 4 hours after calcium repletion and control cells are shown. Scale bar = 50 µm. B. Quantification of continuous and discontinuous junctions per E- cadherin stain, in OVCA429-mDia2 KD and OVCA429-control cells from 0 to 4 hours of calcium repletion. A representative experiment of two is shown. *p<0.05. Error bars denote SEM.

68 2.2.3 mDia2 interacts with β- and α-catenin but not E-cadherin

As downstream eﬀectors of Rho-GTPases, various formin family members including FMNL2, Formin-1, and mDia1 localized to cadherin-based cell-cell junctions and interact with junctional proteins in epithelial cells (Sahai and Marshall 2002, Ko- bielak, Pasolli et al. 2004, Carramusa, Ballestrem et al. 2007, Eisenmann, Harris et al. 2007, Dettenhofer, Zhou et al. 2008, Homem and Peifer 2008, Grikscheit,

Frank et al. 2015, Grikscheit and Grosse 2016). Given its eﬀect on AJ formation and stability, we next asked whether endogenous mDia2 interacts with junctional proteins in OVCA429 monolayers. OVCA429 cells express mDia2, mDia1, E-cadherin,

N-cadherin, β-catenin, and α-catenin (Fig. 2-4A-C). Interaction between mDia2 with both α-andβ-catenin was detected by co-immunoprecipitation (IP). Previously,

Rac1-activated FMNL2 was shown to bind to α-catenin and E-cadherin in MCF10A cells, while Formin-1 was shown to bind to α-catenin in keratinocytes (Kobielak, Pa- solli et al. 2004, Grikscheit, Frank et al. 2015, Grikscheit and Grosse 2016). In our system, mDia2 co-precipitates with both α-andβ-catenins, yet does not interact with either E- or N-cadherin, suggesting that mDia2 may be interacting with a cytosolic, rather than the membrane-associated junctional pool of catenins (Fig. 2-4A). mDia1 does not interact with β-catenin (Fig. 2-4B), indicating a formin-speciﬁcity to this interaction. This is consistent with previous ﬁndings suggesting that mDia2, and not mDia1, is involved in junctional stability in ovarian cancer spheroids (Pettee, Dvorak et al. 2014). Neither mDia1 nor mDia2 precipitated with the cadherins (Fig. 2-4A,

C) and Proximity Ligation Assays (PLA) conﬁrmed a lack of mDia2 and E-cadherin interaction in cells (Fig. 2-5). A positive PLA signal indicates when two proteins are within 40 nm of each other. Collectively, these data indicate that neither mDia1, nor mDia2 associate with AJ-associated E-cadherin complexes.

To conﬁrm interactions between mDia2 and α-andβ-catenins and to localize

69 these interactions within cells, we performed PLA in conjunction with E-cadherin and F-actin IF. PLA interaction between a known direct interaction pair of α-and

β-catenin robustly detected both cytosolic and junctional interactions, while PLA reactions with mDia2 antibody alone revealed no signal (Fig. 2-5A). mDia2 and

α-andβ-catenin interactions were observed and were mostly localized not to the junctional area as demarcated by E-cadherin, but to the non-junctional area (Fig.

2-5A, B).

Figure 2-4: mDia2 interacts with β-catenin and αE-catenin but not E- cadherin. A. Immunoprecipitation for mDia2 in OVCA429 cells followed by immunoblotting for mDia2, E-cadherin, β-catenin, and αE-catenin. B. Immunoprecipitation in OVCA429 cells for β-catenin (β-cat) followed by immunoblotting for mDia2, mDia1, E-cadherin, N-cadherin, β-catenin and αE-catenin. C. Immuno- precipitation in OVCA429 cells for E-cadherin or N-cadherin followed by immunoblotting for mDia1, mDia2, β-catenin, αE- catenin, E-cadherin and N-cadherin. A-B were repeated thrice and representative experiments shown..

70 Figure 2-5: mDia2 interacts with β-catenin and αE-catenin but not E- cadherin. D. Proximity ligation assay (PLA) using OVCA429 cells shows close (<40 nm) interaction between mDia2 and αE- catenin and between mDia2 and β-catenin. Cells were ﬁxed and incubated with primary antibodies against the indicated PLA pairs and E-cadherin. F-actin was stained with phalloidin. Scale bar = 50 µm. E. Quantiﬁcation of junctional mDia2/catenin interactions in D. *p<0.01 relative to αE-catenin/β-catenin PLA pair.

The lack of interaction between mDia2 and E-cadherin was further conﬁrmed by a PLA using antibodies against mDia2 and E-cadherin (Fig. 2-6). Taken together, these results indicate that mDia2 interacts with the α-andβ-catenins in a spatially distinct manner than that of junctional E-cadherin.

71 Figure 2-6: mDia2 does not interact with E-cadherin. OVCA429 cells were treated with PLA probes targeting mDia2 only or mDia2 and E-cadherin. F-actin is stained with phalloidin and nuclei with DAPI. )

2.2.4 mDia2 co-precipitates with β- and α-catenin in HEK293

cells

To validate the interactions between mDia2 and α-andβ-catenins, we exogenously expressed mDia2 and α-orβ-catenin in HEK293 cells. IP for GFP revealed GFP- mDia2 interaction with HA-α-catenin (Fig. 2-7A). Next, we transfected HEK293 cells with Flag-mDia2 and GFP-β-catenin. IP for GFP showed GFP-β-catenin inter- acted with Flag-mDia2 (Fig. 2-7B). These results conﬁrm our previous ﬁndings in

OVCA429 cells, conﬁrming interactions between mDia2 and β-andα-catenin.

72 Figure 2-7: mDia2 co-precipitates with αE- and β-catenin in HEK293 cells. A. Cells were transfected with GFP-mDia2 and HA αcatenin or GFP and HA empty vectors. IP for GFP was followed by immunoblotting for mDia2, αE-catenin, and GFP. The experiment was repeated thrice. B. Cells are transfected with Flag-mDia2 and GFP-βcatenin or Flag and GFP empty vectors. IP for GFP was followed by immunoblotting for Flag and GFP.

2.2.5 mDia2 aﬀects junctional stability

To assess the requirement for mDia2 on junctional protein expression and localization, we visualized junctional markers E-cadherin and β-catenin in mDia2- and control-depleted OVCA429 cells. While total E-cadherin levels were not aﬀected by mDia2 depletion relative to control cells (Fig. 2-1A), mDia2 knockdown cells ex- hibited clustered and discontinuous localization of E-cadherin and β-catenin at the junctional region compared to the linear junctional staining of both proteins in control cells (Fig. 2-8A).

This effect was similar to the clustered staining of E-cadherin and β-catenin seen upon treatment with untransfected OVCA429 cells with a small molecule inhibitor of 73 Figure 2-8: mDia2 expression affects junctional stability. A. OVCA429- mDia2 KD and OVCA429-control cells were plated, grown to confluence, processed and stained for E-cadherin and β-catenin. Coverslips were imaged with fluorescence microscopy at 60x mag- nification. Boxes denote regions of interest (ROI). Scale bars =50µm. B. Graphs show ratios of continuous to discontinuous adherens junctions as measured by E-cadherin staining. Three fields and at least 50 junctions per condition were quantified. C. Phalloidin was used to stain for F-actin in OVCA429- mDia2 KD and OVCA429-control cells. Scale bars = 50 µm. D. Actin filaments per cell for OVCA429-mDia2 KD and OVCA429- control cells are shown. Five separate fields and at least 70 cells per condition were quantified. The experiment was repeated thrice.*p<0.05.

the formin homology 2 (FH2) domain (SMIFH2), a broad-spectrum formin inhibitor

(Fig. 2-9A) (Rizvi, Neidt et al. 2009). The number of continuous junctions was quantiﬁed. The proportion of continuous junctions was signiﬁcantly higher for control cells compared to mDia2 knockdown cells. In control cells, 76% of junctions between cells were continuous, compared to 22% of mDia2 knockdown cell junctions (Fig. 2-

8B). This indicates that mDia2 is essential for junctional localization of E-cadherin and β-catenin, and is consistent with our previous ﬁndings suggesting loss of junction strength upon mDia2 depletion.

74 To evaluate F-actin levels underlying AJs in mDia2- and control-depleted cells, we stained OVCA429s with phalloidin. We observed a marked decrease in F-actin staining in the mDia2-depleted cells relative to control cells (Fig. 2-8C). Actin filaments were quantified using Filaquant software, which showed a significant reduction in F-actin in mDia2 knockdown cells relative to control cells (Fig. 2-8D), just as formin inhibition with SMIFH2 also reduced filament levels (Fig. 2-9B). These results suggest that mDia2 depletion leads to disruption in junctional localization of

β-catenin and E-cadherin with concurrent reduction in F-actin.

Figure 2-9: Formin inhibition disrupts AJs and decreases F-actin filaments. OVCA429 cells were treated with 40 µM SMIFH2 or DMSO for 8 hours. A. Junctions are stained for β-catenin and E-cadherin. Rectangles mark regions of interest (ROIs). Scale bars = 50 µm. B. F-actin was stained with phalloidin and detected filaments shown in Filaquant software analysis. C. A significant reduction in F-actin is observed in SMIFH2-treated cells. Five random fields and at least 58 cells were analyzed per condition. *p<0.05. Values denote mean SEM.

75 2.2.6 mDia2 expression aﬀects interactions between junc-

tional proteins

To determine whether a causal relationship exists between F-actin reduction and

AJ disruption, we investigated interactions between junctional proteins upon mDia2 depletion. AJs are characterized by stable interactions between E-cadherin and catenins (reviewed in (Cavey and Lecuit 2009)). We therefore evaluated the interactions between E-cadherin and β-catenin in mDia2 knockdown and control OVCA429s.

We used PLA probes to visualize E-cadherin and β-catenin interaction (Figure 2-

10A), as well as α-catenin and β-catenin (Figure 2-10B), in conjunction with phalloidin staining. We observed a decrease in β-catenin/E-cadherin interaction in mDia2 knockdown cells with notably more disorganized F-actin (Fig. 2-10A).

Figure 2-10: mDia2 expression aﬀects interactions between junctional proteins. A. Representative images of OVCA429-mDia2 KD and OVCA429-control cells labeled to visualize F-actin and β- catenin/E-cadherin interactions by PLA. B. Representative images of OVCA429-mDia2 KD and OVCA429-control cells labeled to visualize F-actin and α-catenin/β-catenin interactions by PLA.

76 Quantiﬁcation revealed a signiﬁcant reduction in β-catenin/E-cadherin interaction in mDia2 knockdown cells compared to control cells (Fig. 2-11A). This indicates that mDia2-mediated junction stability may, in part, be attributed to an F-actin- dependent stabilization of the β-catenin/E-cadherin complex.

As β-catenin is also associated with both junctional and cytosolic α-catenin, we investigated whether mDia2 depletion impacts α-andβ-catenin interactions using

PLA probes to visualize α-andβ-catenin interactions in conjunction with F-actin.

Interestingly there was a signiﬁcant increase in α-orβ-catenin interactions in mDia2 knockdown cells compared to control cells (Fig. 2-10B, 2-11B). These results suggest that mDia2 promotes E-cadherin/β-catenin interactions while preventing α-/β- catenin interactions. Whether one interaction occurs at the expense of the other is unclear as both α-andβ-catenin and their interacting complex have multiple eﬀects on the cell, including AJ stabilization and transcriptional regulation.

Figure 2-11: Quantification of interactions between junctional proteins. A. Quantification of PLA detecting β-catenin/E-cadherin interactions (n=60, 101). B. Quantification of PLA detecting α- catenin/β-catenin interactions. Six fields were analyzed per condition (A, B). *p<0.05, **p<0.001. Scale bars = 50 µm. All values denote mean +/- SEM.

77 2.2.7 Actin disruption does not inhibit interactions between

mDia2 and α- and β-catenin

We previously showed a reduction in levels and organization of F-actin in mDia2 knockdown OVCA429 cells (Fig. 2-8C-D). This was in response to global suppression of mDia2-dependent F-actin dynamics. We therefore evaluated whether F-actin was necessary to facilitate interactions between mDia2 and α-orβ-catenin or between β and α-catenin. We used cytochalasin D (CytoD) to globally inhibit actin polymerization. PLA was used to visualize interactions between these protein pairs in CytoD- or DMSO-treated OVCA429s. By 30 minutes, CytoD-treated cells had visible and widespread disruption in the cytoskeletal network accompanied by AJ disruption as visualized by E-cadherin staining (Fig. 2-12A-C). Cortical actin underlying the junctions was clustered and discontinuous in CytoD- treated cells compared to the continuous linear junctional staining seen in control cells. Interestingly, there was comparable PLA signal for the α-/β-catenin and β-catenin/mDia2 PLA pairs between treatment and control cells (Fig. 2-12A-C), suggesting that the interaction between mDia2 and β-catenin or α-andβ-catenin do not occur through mutual binding to F-actin. This is consistent with previous studies, which pointed to an α-catenin pool that heterodimerizes with β-catenin and concurrently displays decreased affinity for F-actin (Gates and Peifer 2005, Pokutta, Drees et al. 2008). This pool of α- catenin should be unaffected by disturbances in the actin cytoskeleton. Interestingly, asmallyetsignificantincreaseinmDia2/α-catenin interactions was observed upon

CytoD treatment and global F-actin polymerization defects (Fig. 2-12B, 2-13). These data collectively suggest that mDia2 can interact with α-catenin in the absence of organized F-actin networks. Indeed, previously Formin-1 was shown to interact with

α-catenin in a puriﬁed system in absence of F-actin (Kobielak, Pasolli et al. 2004).

However the primary αcatenin-binding sequence of Formin-1 is not signiﬁcantly ho-

78 mologous to that of mDia formins, suggesting an alternate mechanism. Furthermore, suppression of polymerized F-actin may underlie conformational changes in α-catenin to enhance its localized interaction with mDia2 away from the AJ (Drees, Pokutta et al. 2005, Yamada, Pokutta et al. 2005, Desai, Sarpal et al. 2013). Together, this suggests that actin disruption does not prevent interactions between mDia2 and the catenins, but instead may alter the nature and location of these interactions, such that they are no longer contribute to junctional stabilization.

Figure 2-12: Actin disruption does not inhibit Interactions between mDia2 and α or βcatenin. A-C. Representative images of OVCA429- mDia2 KD and OVCA429-control cells labeled to visualize α- catenin/β-catenin, α-catenin/mDia2, or mDia2/β-catenin by PLA are shown.

79 Figure 2-13: Quantiﬁcations of PLA detecting α-catenin/β-catenin, α- catenin/mDia2, or mDia2/β-catenin interactions. Ten ﬁelds analyzed per condition.

2.3 Discussion

The poor prognosis associated with EOC can largely be attributed to its diagnosis in the late stages of the disease when cancer cells have already disseminated within the peritoneal cavity. Previously, depletion of mDia2, but not mDia1, was associated with single cell invasion from ovarian cancer spheroids (Pettee, Dvorak et al. 2014).

Here, we show for the first time that mDia2 is essential for maintenance of cell-cell junction strength in EOC spheroids and AJ formation. This may be attributed to interactions between mDia2 and AJ proteins, specifically α-andβ-catenin. Interest- ingly, mDia2 does not appear to interact with the catenins at the junctions, indicating that its localization to the AJ is not necessary for its role in junction maintenance. As expected, mDia2 depletion resulted in a significant reduction in actin filament levels and marked disorganization in cytoskeleton architecture. However, disruption of the

F-actin network did not prevent interactions between α-andβ-catenin or between mDia2 and β-catenin. These data suggest a key role for mDia2 in AJ formation and stability in EOC cells which may not be entirely dependent on its actin-polymerizing activity. In this study, we did not assess the roles of targeting mDia2-directed micro-

80 tubule stabilization in AJ function, as mDia formins were shown to strengthen AJs in the absence of microtubules (Carramusa, Ballestrem et al. 2007).

Epithelial cells, including EOC cells, develop cadherin-based AJs, via a three-step process which involves initiation of the cell-cell contact, expansion of the contact interface, and ﬁnally, stabilization of the contact to form multi-cellular structures including epithelial monolayers and three-dimensional (3D) spheroids (Zhang, Bet- son et al. 2005, Cavey and Lecuit 2009, Baum and Georgiou 2011, Grikscheit and

Grosse 2016). During cell migration, actin polymerization generates dynamic plasma- membrane protrusions, which include ﬁnger-like ﬁlopodia characterized by parallel

F-actin bundles extending beyond the leading edge of lamellipodia (Wallar and Al- berts 2003, Baum and Georgiou 2011). Contact initiation occurs when junctional proteins at the ﬁlopodia tips form homophilic linkages with proteins in the neighboring cell (Raich, Agbunag et al. 1999, Vasioukhin, Bauer et al. 2000, Grikscheit and

Grosse 2016). Cadherins then cluster at the newly formed contact and induce actin- remodeling to form a stable junction (Zhang, Betson et al. 2005, Baum and Georgiou

2011, Grikscheit and Grosse 2016). Our ﬁndings that mDia2 plays a role in junction formation are consistent with role for other formins including Formin-1 and mDia1 in junction stabilization in keratinocytes and MCF7 breast cancer cells (Kobielak,

Pasolli et al. 2004, Carramusa, Ballestrem et al. 2007, Grikscheit and Grosse 2016).

This, compounded with the ﬁnding that mDia2 is a predominant formin regulating ovarian cancer spheroid organization (Pettee, Dvorak et al. 2014) supports a key role for mDia2 in AJ formation and stabilization in EOC.

Stability of AJs is thought to depend upon anchorage of the junctional complex of

E-cadherin and catenins to the underlying F-actin cytoskeleton (Hong, Troyanovsky et al. 2013). Traditional models have implicated α-catenin as the link between the E- cadherin/β-catenin junctional complex and the underlying F-actin (Gates and Peifer

2005, Pokutta, Drees et al. 2008). Yet, binding to β-catenin decreased the aﬃnity of

81 α-catenin for F-actin, so an alternative model was proposed wherein α-catenin cycles between a junctional pool bound to the E-cadherin/β-catenin complex and a perijunctional pool bound to F-actin along with various actin-binding proteins such as the formins (Drees, Pokutta et al. 2005, Yamada, Pokutta et al. 2005, Pokutta, Drees et al. 2008). In the present study, mDia2 associated with both α-andβ-catenins, but neither with E- nor N-cadherin. Furthermore, these mDia2-catenin interactions occurred predominantly in the non-junctional region. This suggests that mDia2 either associates with α-andβ-catenin separately as a duplex or together as a cytosolic or nuclear triplex. We cannot rule out the possibility of a triple complex of E-cadherin,

β-catenin, and mDia2, although our co-IPs failed to support this notion. At the same time, there remains the possibility that under certain speciﬁc conditions, mDia2 may act as the link between the E-cadherin complex and α-catenin bound to F-actin.

Formins are known to maintain junctional stability through both cortical actin polymerization and junctional contractility regulation. Indeed, mDia1 regulates junctional tension by reorganizing actin into actomyosin bundles at the AJ in Caco-2 colon epithelial cells (Acharya, Wu et al. 2017). This is signiﬁcant given that α-catenin binding to both F-actin and the actin-binding protein vinculin are force-dependent due to conformational changes in α-catenin that occur with force application (Yone- mura, Wada et al. 2010, Yonemura 2017). As mDia2 is important for junctional stability and resistance to shear force in EOC cells (Figure 2-1), it is reasonable to surmise that mDia2 stabilizes junctions by providing contractile force at the AJ. This, in turn, would enhance anchorage of the junctional E-cadherin/β-catenin complex to the F-actin network via α-catenin.

What if the actin network itself is disrupted? While inhibition of actin polymerization with CytoD led to AJ disruption, it did not reduce non-junctional mDia2 and β-catenin interactions (Fig. 2-12C, 2-13). Meanwhile, F-actin disruption slightly increased interactions between mDia2 and α-catenin (Fig. 2-12B, 2-13). There-

82 fore, although AJ stability involves actin-dependent localization of junctional proteins (Erami, Timpson et al. 2015), mDia2 does not require F-actin to associate with either α-orβ-catenin in the non-junctional region. As β-catenin is not known to bind to F-actin, actin depolymerization is not expected to aﬀect its interaction with mDia2. While α-catenin homodimers bind to F-actin in the cytosol, α-catenins major stoichiometric binding partner in the cytosol is β-catenin (Drees, Pokutta et al.

2005, McCrea and Gottardi 2016). The slight increase in α-catenin interaction with mDia2 could potentially be attributed to a change in α-catenin conformation that occurs upon actin depolymerization, just as α-catenin conformation is regulated by actomyosin contractile force (Acharya, Wu et al. 2017).

We show here that mDia2 depletion signiﬁcantly reduces global F-actin levels

(Fig. 2-8D). Decreased F-actin could potentially decrease the number of α-catenin homodimers that preferentially binds to it, concomitantly increasing the pool of α/β- catenin heterodimers. Indeed, we observed that mDia2 depletion is associated with an increase in α/β-catenin interaction. We also observed a concurrent reduction in

β-catenin/E-cadherin interaction. This is consistent with the concept that stability of the junctional β-catenin/E-cadherin complex is dependent on its anchorage to F- actin. Together, these ﬁndings support the notion of mDia2s indirect role in junction stabilization through F-actin polymerization and bundling.

Our ﬁndings are consistent with recent publications that propose a role for formins in regulating the epithelial mesenchymal transition (EMT). Both formin inhibition with SMIFH2 and depletion of mDia1 and mDia2 prevented TGF-β-induced EMT in lung, mammary, and renal epithelial cells (Rana, Aloisio et al. 2018). Others demonstrated the role of formins including FHOD1 and FMNL2 in the morphological changes associated with EMT (Li, Zhu et al. 2010, Jurmeister, Baumann et al. 2012,

Rana, Aloisio et al. 2018). In ovarian cancer, decreased E-cadherin expression is associated with peritoneal seeding of tumor cells and lower overall survival rate (Cho,

83 Choi et al. 2006, Vergara, Merlot et al. 2010). Here we identify a novel role for mDia2 in AJ stabilization to impact EMT in 3D ovarian cancer spheroids (Pettee,

Dvorak et al. 2014). While we did not observe an interaction between mDia2 and the cadherin molecules, mDia2 interacts with key regulators of the AJ, α-andβ-catenin to regulate E-cadherin localization to AJs.

It is interesting to note that mDia2 minimally interacts with junctional β-catenin, which was unexpected given its junction stabilizing eﬀect. It has previously been proposed that the AJ acts as a sink for cytosolic β-catenin, drawing β-catenin away from the Wnt signaling pathway with transcriptional activation of pro-migratory genes (Jeanes, Gottardi et al. 2008). Cytosolic and nuclear α-catenin can also regulate transcription, both through β-catenin binding and its regulation of nuclear actin

(Daugherty, Serebryannyy et al. 2014). Our ﬁnding that mDia2 interacts with both catenins may suggest a potential role in transcriptional regulation, bridging the AJ and Wnt signaling pathways. Future studies would aim to determine how mDia2 aﬀects Wnt/β-catenin signaling in EOC.

2.4 Conclusions

In summary, our findings indicate an essential role for mDia2 in AJ formation and stability in EOC cells. These effects are likely achieved through its interactions with and regulation of α-andβcatenin. While we demonstrate interaction, it remains uncertain whether α-catenin binding to mDia2 impacts either actin polymerization or bundling, or which domains of mDia2 and α-catenin interact. Furthermore, our current studies utilize EOC monolayers to dissect the interactions between mDia2 and proteins involved in the AJ. Assuming that these same interactions occur in 3D spheroids and given that loss of mDia2 is associated with disease progression in ovarian cancer (Creekmore, Silkworth et al. 2011), our findings support a novel mechanism

84 for EOC dissemination that should be considered in development of targeted therapy against this deadly disease.

2.5 Methods

2.5.1 Cell lines and reagents

Serous ovarian adenocarcinoma OVCA429 cells were kind gifts from Dr. Deborah

Vestal (University of Toledo, Toledo, OH) were grown in RPMI-1640 (GE Lifesciences

(Pittsburgh, PA)) containing 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin, and 100 g streptomycin. HEK293 human embryonic kidney cells were from

ATCC (Manassas, VA) and were grown in DMEM (GE Lifesciences) containing 10%

(v/v) FBS, 100 U/ml, and 100 µg streptomycin. All cells were grown in a 37OC incubator with 5% CO2. OVCA429 cells were plated at 200,000 cells per 35 mm well, grown to 70-80% con-

ﬂuence, then treated with DMSO or 40 µM SMIFH2 in DMSO (EMD Biochemicals,

Tocris Bioscience, Avonmouth) in full media for 8 hours.

2.5.2 Western blotting

Cells were harvested for Western blots using SDS lysis buﬀer (0.5 M Tris-HCl, pH 6.8, glycerol, 10% SDS (wt/vol), 0.1% bromophenol blue (wt/vol), 0.1 M dio- thiothreitol (DTT)). Lysates were separated using 4-20% gradient SDS-PAGE gels

(BioRad, Hercules, CA) and were transferred to PVDF membranes using the BioRad

Trans-Blot turbo system.

85 2.5.3 Transfection and knockdown

pCMV-driven plasmid vectors encoding GFP, GFP-mDia2 as well as Flag-mDia2 were kind gifts of Dr. Art Alberts (Van Andel Institute, Grand Rapids, MI). Plasmids encoding HA and HA-α-catenin were kind gifts of Dr. Deniz Toksoz (Tufts University,

Medford, MA). GFP-β-catenin was from Addgene (Cambridge, MA) and GFP empty vector was a kind gift from Dr. Kam Yeung (University of Toledo).

For knockdown experiments, mDia1 siRNA (J-010347-070005) and control GAPDH

(D-001140-01-05) siRNA constructs were purchased from Thermo Scientiﬁc (Waltham,

MA), and mDia2 shRNA and control pGFPVRS from Origene (Rockville, MD) respectively. To generate OVCA429 cells stably depleted of mDia2, OVCA429 cells were transfected with GFP-mDia2 shRNA constructs (Origene) using Fugene (Promega

(Madison, WI)) per manufacturers protocol and drug selected with 4 µg/ml puromycin.

Control cells were generated using empty pGFP-V-RS vector (Origene). Cells were then further selected for GFP through ﬂow cytometry using the FACS Aria Ilu High-

Speed Cell Sorter (BD Biosciences (Franklin Lakes, NJ)). Knockdown of mDia2 was conﬁrmed using Western blotting with an anti-mDia2 antibody (Proteintech, Rose- mont, IL) at 1:1000.

HEK293 cells were transiently transfected using a standard calcium phosphate transfection method (Jordan and Wurm 2004).

2.5.4 Immunoprecipitation

OVCA429 cells were grown to 70-80% conﬂuence and serum starved overnight with RPMI-1640 containing 0.1% (v/v) FBS and 100 mg streptomycin. Cells were then serum stimulated with RPMI-1640 containing 10% (v/v) FBS and 100 µgstrep- tomycin for 4 hours. Lysates were collected with NP40 lysis buﬀer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 1% NP40, 10% glycerol) with protease inhibitors (1 µMeach

86 of NaV O4, aprotinin, pepstatin, leupeptin, DTT, PMSF). Lysates were incubated with anti-mDia2 antibody (Proteintech) or control Fab fragment (Jackson Immuno- labs) at a concentration of 1 µg antibody per 1 mg lysate for 3 hours at 4OCwith shaking, followed by addition of Protein A Agarose beads (Invitrogen, Santa Cruz) for 1 hour at 4OC with shaking. Beads were washed 5 times with NP40 buﬀer then heated at 85OCinSDSlysisbuﬀerpriortoloadingingels.

HEK293 cells were grown to 70% conﬂuence and transfected with HA and GFP or

HA-α-catenin and GFP-mDia2-encoding vectors. Lysates were collected with NP40 buﬀer and 1 µM protease inhibitors 48 hours post-transfection. Lysates were incubated with anti-GFP (Abcam (Cambridge, United Kingdom)) antibody for 3 hours at 4OC then 1 hour with Protein A agarose beads, prior to washing in NP-40 buﬀer.

Western blotting was performed with 4-20% SDS-PAGE (Biorad), followed by immunoblotting with the following antibodies: rabbit anti-mDia2 (1:1000, (Proteintech), mouse anti-E-cadherin (1:1000, Cell Signaling), mouse anti-N-cadherin (BD Trans- duction Laboratories (Franklin Lakes, NJ)), rabbit anti-mDia1 (1:1000, (Proteintech), rabbit anti-α-catenin (1:2000 (Proteintech), and rabbit anti-β-catenin (1:2000 (Pro- teintech) and visualization by the ClarityTM Western ECL kit (Biorad).

2.5.5 Immunoﬂuorescence and Image Analysis

For immunofluorescence, cells grown upon glass coverslips were fixed in 4% paraformalde- hyde (PFA)/phosphate buffered saline (PBS) for 5 minutes, washed with PBS, permeated with 0.5% Triton X-100 (TX100) for 20 minutes, blocked for an hour with

3% bovine serum albumin (BSA)/PBS, and incubated with antibodies against 1:100

β-catenin (Proteintech) and 1:100 E-cadherin (Cell Signaling), or 1:100 αE-catenin

(Genetex (Irvine, CA)) and 1:100 mDia2 (Proteintech) overnight at 4OC followed by incubation with 1:200 Alexa-Fluor secondary antibodies (Invitrogen) for 2 hours at

37 OC. To visualize F-actin and nucleus, we used 1:100 Alexa Fluor 647 Phalloidin

87 (ThermoFisher Scientiﬁc (Waltham, MA)) and DAPI (Invitrogen), respectively. Cov- erslips were mounted with Fluoromount-G (SouthernBiotech (Birmingham, AL)) and visualized with an Olympus 60x UPlanFl 1.25 NA oil objective on the EVOS FL epi-

ﬂuorescence microscope (AMG/Thermo Fisher).

To quantify actin ﬁlament number and length, images of phalloidin-stained cells were uniformly processed with Photoshop and Filaquant software provided by Dr.

Konrad Engel of the University of Rostock (Matschegewski, Staehlke et al. 2012).

To determine junction continuity, a junction was characterized as continuous as described (Carramusa, Ballestrem et al. 2007). Briefly, if E-cadherin fluorescence along acell-cellcontactwasabovebackgroundfluorescenceforatleast50%ofthedis- tance between the cell vertices, that junction was considered continuous (Carramusa,

Ballestrem et al. 2007). Quantiﬁcation was performed using a custom Python script.

At least 50 junctions were counted per condition for each experiment and the experiment was performed thrice.

2.5.6 In Situ Proximity Ligation Assay (PLA)

To visualize the interactions and localization of the interactions between mDia2 and αE-catenin, we used the Duolink PLA kit (Sigma-Aldrich (St. Louis, MO)).

OVCA429 cells were ﬁxed with 4% PFA/PBS, permeated with 0.5% TX100, blocked with 3% BSA/PBS and incubated overnight with primary antibodies. The following antibodies were used: 1:100 goat anti-E-cadherin (R&D Systems (Minneapolis,

MN)), 1:100 rabbit anti-mDia2 (Proteintech), 1:100 mouse anti-αE-catenin (Gene-

Tex), 1:100 mouse anti-β-catenin (Origene), 1:100 mouse anti-E-cadherin (Cell Sig- naling). Cells were then incubated with secondary antibodies conjugated to oligonucleotides per manufacturer’s protocol. Brieﬂy, after incubation with anti-mouse and anti-rabbit secondary antibodies, cells were incubated with ligase followed by polymerase and close protein-protein interactions (<40 nm apart) were detected as ﬂuo-

88 rescent dots generated by rolling circle ampliﬁcation with complementary ﬂuorescent oligonucleotides (Debaize, Jakobczyk et al. 2017). Cell nuclei were stained with DAPI

(Invitrogen) and F-actin with 1:100 Alexa Fluor 647 Phalloidin (ThermoFisher Scien- tific). Quantification of colocalization was performed using ImageJ Particle Analysis and Colocalization plugin. Three independent fields and at least 47 cells per condition were quantified.

To visualize interactions between mDia2 and α-catenin/β-catenin, or between α- catenin and β-catenin in OVCA429 cells upon cytochalasin D treatment, cells were plated at 200,000 cells/well into a 6-well dish upon glass coverslips. At 70-80% con-

ﬂuence, cells were treated with 1 M cytochalasin D (Calbiochem (Burlington, MA)) for 30 minutes, then ﬁxed and stained for PLA pairs (mDia2/α-catenin, mDia2/β- catenin, or α-catenin/β-catenin), E-cadherin, F-actin, and DAPI as above.

2.5.7 Hanging Drop Assay

The assay was performed as described (Kim, Islam et al. 2000, McLaughlin,

Kruger et al. 2007). Brieﬂy, control or mDia2 KD OVCA429 cells were trypsinized, centrifuged, and re-suspended as single- or up to 3-cell suspensions at 2.5x105 cells/ml.

For each cell type, 20-ml droplets were pipetted onto the lids of 35 mm culture dishes and dishes were ﬁlled with 2 ml of growth media. At 0.5, 2, and 4 hours, the lids were inverted and drops were transferred to glass slides and pipetted 10 times through a

20-ml pipet tip. Three random fields were imaged using epifluorescence microscopy with an Olympus 4x UPlanFL 0.13 NA objective lens and numbers and sizes of clusters quantified. At least 200 cells were counted per condition. The experiment was performed thrice.

89 2.5.8 Calcium Switch Assay

Untransfected, control knockdown, and mDia2 KD OVCA429 cells were grown to

60-70% conﬂuence upon glass coverslips and incubated with RPMI-1640 with 0.1%

FBS, 100 U/ml penicillin, and 100 mg streptomycin, without calcium (US Biological

(Salem, MA) for 16 hours. Medium was then changed to either the same calcium- free RPMI-1640 or RPMI-1640 with 0.42 mM calcium (GE Lifesciences) and 10%

FBS and 100 µg streptomycin for 4 hours. Cells were ﬁxed with 4% PFA/PBS at given time points and stained for E-cadherin and βcatenin. Junction continuity was determined as above. At least 50 junctions were scored per condition at each time point and the experiment was performed twice.

2.5.9 Statistics

Two-tail Students t-tests were used with a 95% conﬁdence value. P-values less than 0.05 were interpreted as statistically signiﬁcant. All error bars denote standard deviations from representative experiments unless otherwise indicated. Graphs and statistics were generated from Microsoft Excel and GraphPad Prism software.

2.6 Acknowledgements

We thank members of the Eisenmann lab, Drs. Rafael Garcia-Mata, William

Maltese, Randall Ruch, Eda Yildirim-Ayan, Andrea Kalinoski, Dayanidhi Raman and Peterson G.T. Schwifty for discussion and technical guidance. We also thank the University of Toledo Department of Cancer Biology for everyones support and generosity with their time and resources, especially Nicole Bearss and Augustus Tilley for guidance with PLA experiments. We thank Dr. Sahezeel Awadia and Ashtynn

Zinn for technical assistance, Dr. Konrad Engel for use of Filaquant, and Dr. Kam

Yeung and Dr. Deniz Toksoz for their valuable time and plasmids. 90 2.7 Funding

University of Toledo Foundation, Rita T. Sheely Endowment, University of Toledo

URFO/URAF

91 Chapter 3

Discussion

3.1 mDia2 interacts with both α- and β-catenin

One novel ﬁnding of this study was that instead of interacting with α-catenin alone

(Kobielak, Pasolli et al. 2004), or with both α-catenin and E-cadherin (Grikscheit,

Frank et al. 2015) like other members of the formin family, mDia2 interacts with both

α-andβ-catenin. These interactions are not junctional, as PLA studies in the current work suggest that mDia2/α-catenin and mDia2/β-catenin interactions mainly occur outside the AJs (Fig. 2-5). Considering this, mDia2 most likely interacts with the heteromer of α-andβ-catenin as the predominant binding partner for β-catenin in the cytosol is α-catenin, as mentioned previously. This opens up the possibility for mDia2

(whether directly or via actin polymerization) to regulate β-catenin localization to the cytosol or nucleus, with eﬀects on transcription of genes such as Axin, which is associated with tumor suppressors like adenomatous polyposis coli (APC) (Nakamura,

Hamada et al. 1998).

The lack of mDia2 interaction with either E- or N-cadherin further conﬁrms that its non-junctional localization. Although other formins are known to both localize to and regulate the AJ, some, including mDia2 in our study and mDia1 ((Rao and

Zaidel-Bar 2016) regulate junctions without speciﬁcally being conﬁned to them. This

92 supports the concept that formins are essential to maintaining actin dynamics that in turn help to stabilize the junctional cadherin/β-catenin complex.

3.2 Biomechanics of EOC

It is interesting to note that from the exfoliation of EOC cells into the abdominal cavity to EOC at a secondary site, these processes all involve applications of force at the molecular level. Actin polymerization and bundling to form stress ﬁbers are also inherently associated with generation of force. Rizvi et al. showed that that the macroscopic level of shear force generated from ascites was enough to induce changes in expression proﬁles of EOC (Rizvi, Gurkan et al. 2013). At the microscopic level, tensile force changes the conformation of α-catenin to allow for new binding partners.

It would be interesting to see how the two can be linked and whether formins might have a role in mediating invasive transitions in EOC that are triggered by mechanical changes in the tumor microenvironment.

Currently, it is already known that mDia1 regulates junctional tension by reorganizing actin into actomyosin bundles at the AJ in Caco-2 colon epithelial cells

(Acharya, Wu et al. 2017). Given our finding that mDia2 is important for junctional stability and resistance to shear force in EOC cells, we could ask whether pressure and flow of ascites may have an effect on mDia2 expression or function. The answers to these questions can have significant implications for our understanding of EOC dissemination.

3.3 Limitations

Alimitationofthisworkisthelackofexperimentswithnormalmesothelialcells or surface ovarian epithelial cells for comparison to EOC. Would mDia2 be important

93 in junctional stabilization in the former and would the interactions between mDia2 and the catenins still be applicable? In vivo, the expression proﬁles of disseminated ovarian cancer cells appear to evolve (e.g., through EMT and MET) even as they maintain some of the original characteristics of the tumor. Would the interactions between mDia2 and the AJ proteins evolve with EOC progression? These may all be answerable questions in the future.

Another limitation lies in the inherent technical difficulty of triple IPs. Due to this, we cannot rule out the possibility of a triple complex of E-cadherin, β-catenin, and mDia2, even though our co-IPs failed to support this notion. This question may be better answered with direct binding assays, using different domains of mDia2 to confirm specific interactions, and rule out others.

3.4 Future Directions

There are many directions that this work can take us. Our ﬁnding that mDia2 interacts with both catenins may suggest a potential role in transcriptional regulation, bridging the AJ and Wnt signaling pathways. Future studies would aim to determine how mDia2 aﬀects Wnt/β-catenin signaling in cancer as well as normal tissue. In addition, the mechanism for mDia2-mediated AJ stabilization can be further explored.

For example, one could ask, what would joint binding of mDia2 and α-catenin to the actin ﬁlament have on actin polymerization and/or actin bundling rates?

3.5 Summary of ﬁndings

In summary, our ﬁndings indicate that mDia2 is essential in AJ formation and stability in EOC cells. Depletion of mDia2 reduces resistance to shear stress in EOC spheroids. This can be attributed to interactions between mDia2 and α-andβ-

94 catenin. While we demonstrate interaction between these proteins, it remains uncertain whether α-catenin binding to mDia2 impacts either actin polymerization or bundling, or which domains of mDia2 and α-catenin interact. Interestingly, junctional localization is not necessary for mDia2s aﬀect on the AJ formation and stabilization.

It is possible that mDia2 indirectly aﬀects interactions between β-catenin and E- cadherin through their ability to polymerize and/or bundle actin ﬁlaments. On the other hand, disruption of the F-actin network did not prevent interactions between

α-andβ-catenin or between mDia2 and β-catenin. Therefore, mDia2-dependent AJ formation and stability in EOC cells may not be entirely dependent on its actin- polymerizing activity. Overall, we have found a novel interaction between mDia2 and certain AJ proteins, along with a new mechanism for EOC dissemination that should be considered in development of targeted therapy against this deadly disease.

95 References

[RN1, 2003] (2003). Drugs of choice for cancer. Treat Guidel Med Lett,1(7):41–52.

[Abe and Takeichi, 2008] Abe, K. and Takeichi, M. (2008). Eplin mediates linkage

of the cadherin catenin complex to f-actin and stabilizes the circumferential actin

belt. Proc Natl Acad Sci U S A,105(1):13–9.

[Abu-Hijleh et al., 1995] Abu-Hijleh, M. F., Habbal, O. A., and Moqattash, S. T.

(1995). The role of the diaphragm in lymphatic absorption from the peritoneal

cavity. J Anat,186(Pt3):453–67.

[Acharya et al., 2017] Acharya, B. R., Wu, S. K., Lieu, Z. Z., Parton, R. G., Grill,

S. W., Bershadsky, A. D., Gomez, G. A., and Yap, A. S. (2017). Mammalian

diaphanous 1 mediates a pathway for e-cadherin to stabilize epithelial barriers

through junctional contractility. Cell Rep,18(12):2854–2867.

[Adams et al., 1998] Adams, C. L., Chen, Y. T., Smith, S. J., and Nelson, W. J.

(1998). Mechanisms of epithelial cell-cell adhesion and cell compaction revealed

by high-resolution tracking of e-cadherin-green ﬂuorescent protein. J Cell Biol,

142(4):1105–19.

[Adams and Nelson, 1998] Adams, C. L. and Nelson, W. J. (1998). Cytomechanics

of cadherin-mediated cell-cell adhesion. Curr Opin Cell Biol,10(5):572–7.

[Adamson et al., 1992] Adamson, P., Paterson, H. F., and Hall, A. (1992). Intracel-

lular localization of the p21rho proteins. J Cell Biol,119(3):617–27.

96 [Adur et al., 2012] Adur, J., Pelegati, V. B., de Thomaz, A. A., Baratti, M. O.,

Almeida, D. B., Andrade, L. A., Bottcher-Luiz, F., Carvalho, H. F., and Cesar,

C. L. (2012). Optical biomarkers of serous and mucinous human ovarian tumor

assessed with nonlinear optics microscopies. PLoS One,7(10):e47007.

[Aghdassi et al., 2012] Aghdassi, A., Sendler, M., Guenther, A., Mayerle, J., Behn,

C. O., Heidecke, C. D., Friess, H., Buchler, M., Evert, M., Lerch, M. M., and

Weiss, F. U. (2012). Recruitment of histone deacetylases hdac1 and hdac2 by the

transcriptional repressor zeb1 downregulates e-cadherin expression in pancreatic

cancer. Gut,61(3):439–48.

[Akhmanova and Steinmetz, 2015] Akhmanova, A. and Steinmetz, M. O. (2015).

Control of microtubule organization and dynamics: two ends in the limelight. Nat

Rev Mol Cell Biol,16(12):711–26.

[Al Habyan et al., 2018] Al Habyan, S., Kalos, C., Szymborski, J., and McCaﬀrey,

L. (2018). Multicellular detachment generates metastatic spheroids during intra-

abdominal dissemination in epithelial ovarian cancer. Oncogene,37(37):5127–5135.

[Alberts, 2001] Alberts, A. S. (2001). Identiﬁcation of a carboxyl-terminal

diaphanous-related formin homology protein autoregulatory domain. J Biol Chem,

276(4):2824–30.

[Allemani et al., 2015] Allemani, C., Weir, H. K., Carreira, H., Harewood, R., Spika,

D., Wang, X. S., Bannon, F., Ahn, J. V., Johnson, C. J., Bonaventure, A., Marcos-

Gragera, R., Stiller, C., Azevedo e Silva, G., Chen, W. Q., Ogunbiyi, O. J., Rachet,

B., Soeberg, M. J., You, H., Matsuda, T., Bielska-Lasota, M., Storm, H., Tucker,

T. C., Coleman, M. P., and Group, C. W. (2015). Global surveillance of cancer

survival 1995-2009: analysis of individual data for 25,676,887 patients from 279

97 population-based registries in 67 countries (concord-2). Lancet,385(9972):977–

1010.

[Allen et al., 1987] Allen, H. J., Porter, C., Gamarra, M., Piver, M. S., and John-

son, E. A. (1987). Isolation and morphologic characterization of human ovarian

carcinoma cell clusters present in eﬀusions. Exp Cell Biol,55(4):194–208.

[Angres et al., 1996] Angres, B., Barth, A., and Nelson, W. J. (1996). Mechanism for

transition from initial to stable cell-cell adhesion: kinetic analysis of e-cadherin-

mediated adhesion using a quantitative adhesion assay. J Cell Biol,134(2):549–57.

[Arden et al., 2015] Arden, J. D., Lavik, K. I., Rubinic, K. A., Chiaia, N., Khuder,

S. A., Howard, M. J., Nestor-Kalinoski, A. L., Alberts, A. S., and Eisenmann, K. M.

(2015). Small-molecule agonists of mammalian diaphanous-related (mdia) formins

reveal an eﬀective glioblastoma anti-invasion strategy. Mol Biol Cell,26(21):3704–

18.

[Ardighieri et al., 2014] Ardighieri, L., Zeppernick, F., Hannibal, C. G., Vang, R.,

Cope, L., Junge, J., Kjaer, S. K., Kurman, R. J., and Shih Ie, M. (2014). Muta-

tional analysis of braf and kras in ovarian serous borderline (atypical proliferative)

tumours and associated peritoneal implants. J Pathol,232(1):16–22.

[Arjonen et al., 2011] Arjonen, A., Kaukonen, R., and Ivaska, J. (2011). Filopodia

and adhesion in cancer cell motility. Cell Adh Migr,5(5):421–30.

[Armstrong et al., 2006] Armstrong, D. K., Bundy, B., Wenzel, L., Huang, H. Q.,

Baergen, R., Lele, S., Copeland, L. J., Walker, J. L., Burger, R. A., and Gyne-

cologic Oncology, G. (2006). Intraperitoneal cisplatin and paclitaxel in ovarian

cancer. N Engl J Med,354(1):34–43.

[Baarlink et al., 2010] Baarlink, C., Brandt, D., and Grosse, R. (2010). Snapshot:

Formins. Cell,142(1):172,172e1. 98 [Baarlink et al., 2013] Baarlink, C., Wang, H., and Grosse, R. (2013). Nuclear

actin network assembly by formins regulates the srf coactivator mal. Science,

340(6134):864–7.

[Balkwill, 2004] Balkwill, F. (2004). Cancer and the chemokine network. Nat Rev

Cancer,4(7):540–50.

[Bartolini et al., 2008] Bartolini, F., Moseley, J. B., Schmoranzer, J., Cassimeris, L.,

Goode, B. L., and Gundersen, G. G. (2008). The formin mdia2 stabilizes micro-

tubules independently of its actin nucleation activity. J Cell Biol,181(3):523–36.

[Baum and Georgiou, 2011] Baum, B. and Georgiou, M. (2011). Dynamics of ad-

herens junctions in epithelial establishment, maintenance, and remodeling. J Cell

Biol,192(6):907–17.

[Beli et al., 2008] Beli, P., Mascheroni, D., Xu, D., and Innocenti, M. (2008). Wave

and arp2/3 jointly inhibit ﬁlopodium formation by entering into a complex with

mdia2. Nat Cell Biol,10(7):849–57.

[Bendoraite et al., 2010] Bendoraite, A., Knouf, E. C., Garg, K. S., Parkin, R. K.,

Kroh, E. M., O’Briant, K. C., Ventura, A. P., Godwin, A. K., Karlan, B. Y.,

Drescher, C. W., Urban, N., Knudsen, B. S., and Tewari, M. (2010). Regulation of

mir-200 family micrornas and zeb transcription factors in ovarian cancer: evidence

supporting a mesothelial-to-epithelial transition. Gynecol Oncol,116(1):117–25.

[Bershadsky, 2004] Bershadsky, A. (2004). Magic touch: how does cell-cell adhesion

trigger actin assembly? Trends Cell Biol,14(11):589–93.

[Bione et al., 1998] Bione, S., Sala, C., Manzini, C., Arrigo, G., Zuﬀardi, O., Banﬁ,

S., Borsani, G., Jonveaux, P., Philippe, C., Zuccotti, M., Ballabio, A., and Toniolo,

D. (1998). A human homologue of the drosophila melanogaster diaphanous gene

99 is disrupted in a patient with premature ovarian failure: evidence for conserved

function in oogenesis and implications for human sterility. Am J Hum Genet,

62(3):533–41.

[Bogdan et al., 2013] Bogdan, S., Schultz, J., and Grosshans, J. (2013). Formin’ cellu-

lar structures: Physiological roles of diaphanous (dia) in actin dynamics. Commun

Integr Biol,6(6):e27634.

[Brady-Kalnay et al., 1995] Brady-Kalnay, S. M., Rimm, D. L., and Tonks, N. K.

(1995). Receptor protein tyrosine phosphatase ptpmu associates with cadherins

and catenins in vivo. J Cell Biol,130(4):977–86.

[Braga, 2002] Braga, V. M. (2002). Cell-cell adhesion and signalling. Curr Opin Cell

Biol,14(5):546–56.

[Braga et al., 1997] Braga, V. M., Machesky, L. M., Hall, A., and Hotchin, N. A.

(1997). The small gtpases rho and rac are required for the establishment of

cadherin-dependent cell-cell contacts. J Cell Biol,137(6):1421–31.

[Brandt et al., 2007] Brandt, D. T., Marion, S., Griﬃths, G., Watanabe, T.,

Kaibuchi, K., and Grosse, R. (2007). Dia1 and iqgap1 interact in cell migration

and phagocytic cup formation. J Cell Biol,178(2):193–200.

[Burkhalter et al., 2015] Burkhalter, R. J., Westfall, S. D., Liu, Y., and Stack, M. S.

(2015). Lysophosphatidic acid initiates epithelial to mesenchymal transition and

induces beta-catenin-mediated transcription in epithelial ovarian carcinoma. J Biol

Chem,290(36):22143–54.

[Cancer Genome Atlas Research, 2011] Cancer Genome Atlas Research, N. (2011).

Integrated genomic analyses of ovarian carcinoma. Nature,474(7353):609–15.

100 [Cannistra, 2004] Cannistra, S. A. (2004). Cancer of the ovary. N Engl J Med,

351(24):2519–29.

[Carlier et al., 1999] Carlier, M. F., Ducruix, A., and Pantaloni, D. (1999). Signalling

to actin: the cdc42-n-wasp-arp2/3 connection. Chem Biol,6(9):R235–40.

[Carramusa et al., 2007] Carramusa, L., Ballestrem, C., Zilberman, Y., and Bershad-

sky, A. D. (2007). Mammalian diaphanous-related formin dia1 controls the orga-

nization of e-cadherin-mediated cell-cell junctions. J Cell Sci, 120(Pt 21):3870–82.

[Casagrande et al., 1979] Casagrande, J. T., Louie, E. W., Pike, M. C., Roy, S., Ross,

R. K., and Henderson, B. E. (1979). ”incessant ovulation” and ovarian cancer.

Lancet,2(8135):170–3.

[Casey et al., 2001] Casey, R. C., Burleson, K. M., Skubitz, K. M., Pambuccian, S. E.,

Oegema, T. R., J., Ruﬀ, L. E., and Skubitz, A. P. (2001). Beta 1-integrins regulate

the formation and adhesion of ovarian carcinoma multicellular spheroids. Am J

Pathol,159(6):2071–80.

[Cassimeris, 2002] Cassimeris, L. (2002). The oncoprotein 18/stathmin family of mi-

crotubule destabilizers. Curr Opin Cell Biol,14(1):18–24.

[Castrillon and Wasserman, 1994] Castrillon, D. H. and Wasserman, S. A. (1994). Di-

aphanous is required for cytokinesis in drosophila and shares domains of similarity

with the products of the limb deformity gene. Development,120(12):3367–77.

[Cavey and Lecuit, 2009] Cavey, M. and Lecuit, T. (2009). Molecular bases of cell-cell

junctions stability and dynamics. Cold Spring Harb Perspect Biol,1(5):a002998.

[Cavey et al., 2008] Cavey, M., Rauzi, M., Lenne, P. F., and Lecuit, T. (2008). A

two-tiered mechanism for stabilization and immobilization of e-cadherin. Nature,

453(7196):751–6. 101 [Chen et al., 2003a] Chen, V. W., Ruiz, B., Killeen, J. L., Cote, T. R., Wu, X. C.,

and Correa, C. N. (2003a). Pathology and classiﬁcation of ovarian tumors. Cancer,

97(10 Suppl):2631–42.

[Chen et al., 2003b] Chen, X., Kojima, S., Borisy, G. G., and Green, K. J. (2003b).

p120 catenin associates with kinesin and facilitates the transport of cadherin-

catenin complexes to intercellular junctions. J Cell Biol,163(3):547–57.

[Cheng et al., 2011] Cheng, L., Zhang, J., Ahmad, S., Rozier, L., Yu, H., Deng, H.,

and Mao, Y. (2011). Aurora b regulates formin mdia3 in achieving metaphase

chromosome alignment. Dev Cell,20(3):342–52.

[Chesarone et al., 2010] Chesarone, M. A., DuPage, A. G., and Goode, B. L. (2010).

Unleashing formins to remodel the actin and microtubule cytoskeletons. Nat Rev

Mol Cell Biol,11(1):62–74.

[Chhabra and Higgs, 2007] Chhabra, E. S. and Higgs, H. N. (2007). The many

faces of actin: matching assembly factors with cellular structures. Nat Cell Biol,

9(10):1110–21.

[Cho et al., 2006] Cho, E. Y., Choi, Y., Chae, S. W., Sohn, J. H., and Ahn, G. H.

(2006). Immunohistochemical study of the expression of adhesion molecules in

ovarian serous neoplasms. Pathol Int,56(2):62–70.

[Chugh et al., 2017] Chugh, P., Clark, A. G., Smith, M. B., Cassani, D. A. D.,

Dierkes, K., Ragab, A., Roux, P. P., Charras, G., Salbreux, G., and Paluch, E. K.

(2017). Actin cortex architecture regulates cell surface tension. Nat Cell Biol,

19(6):689–697.

[Chung et al., 2013] Chung, B. M., Rotty, J. D., and Coulombe, P. A. (2013). Net-

working galore: intermediate ﬁlaments and cell migration. Curr Opin Cell Biol,

25(5):600–12. 102 [Clarke-Pearson, 2009] Clarke-Pearson, D. L. (2009). Clinical practice. screening for

ovarian cancer. N Engl J Med,361(2):170–7.

[Clevers and Nusse, 2012] Clevers, H. and Nusse, R. (2012). Wnt/beta-catenin sig-

naling and disease. Cell,149(6):1192–205.

[Collaborative Group on Epidemiological Studies of Ovarian et al., 2008]

Collaborative Group on Epidemiological Studies of Ovarian, C., Beral, V.,

Doll, R., Hermon, C., Peto, R., and Reeves, G. (2008). Ovarian cancer and

oral contraceptives: collaborative reanalysis of data from 45 epidemiological

studies including 23,257 women with ovarian cancer and 87,303 controls. Lancet,

371(9609):303–14.

[Copeland et al., 2007] Copeland, S. J., Green, B. J., Burchat, S., Papalia, G. A.,

Banner, D., and Copeland, J. W. (2007). The diaphanous inhibitory do-

main/diaphanous autoregulatory domain interaction is able to mediate het-

erodimerization between mdia1 and mdia2. J Biol Chem,282(41):30120–30.

[Creekmore et al., 2011] Creekmore, A. L., Silkworth, W. T., Cimini, D., Jensen,

R. V., Roberts, P. C., and Schmelz, E. M. (2011). Changes in gene expres-

sion and cellular architecture in an ovarian cancer progression model. PLoS One,

6(3):e17676.

[Dalbeth et al., 2014] Dalbeth, N., Lauterio, T. J., and Wolfe, H. R. (2014). Mecha-

nism of action of colchicine in the treatment of gout. Clin Ther,36(10):1465–79.

[Daugherty et al., 2014] Daugherty, R. L., Serebryannyy, L., Yemelyanov, A., Flozak,

A. S., Yu, H. J., Kosak, S. T., deLanerolle, P., and Gottardi, C. J. (2014). alpha-

catenin is an inhibitor of transcription. Proc Natl Acad Sci U S A,111(14):5260–5.

[Davidson et al., 2015] Davidson, B., Holth, A., Hellesylt, E., Tan, T. Z., Huang,

R. Y., Trope, C., Nesland, J. M., and Thiery, J. P. (2015). The clinical role of 103 epithelial-mesenchymal transition and stem cell markers in advanced-stage ovarian

serous carcinoma eﬀusions. Hum Pathol,46(1):1–8.

[De Troyer et al., 2009] De Troyer, A., Leduc, D., Cappello, M., Mine, B., Gevenois,

P. A., and Wilson, T. A. (2009). Mechanisms of the inspiratory action of the

diaphragm during isolated contraction. J Appl Physiol (1985),107(6):1736–42.

[Debaize et al., 2017] Debaize, L., Jakobczyk, H., Rio, A. G., Gandemer, V., and

Troadec, M. B. (2017). Optimization of proximity ligation assay (pla) for detection

of protein interactions and fusion proteins in non-adherent cells: application to

pre-b lymphocytes. Mol Cytogenet,10:27.

[Delanoe-Ayari et al., 2004] Delanoe-Ayari, H., Al Kurdi, R., Vallade, M., Gulino-

Debrac, D., and Riveline, D. (2004). Membrane and acto-myosin tension promote

clustering of adhesion proteins. Proc Natl Acad Sci U S A,101(8):2229–34.

[Desai and Mitchison, 1997] Desai, A. and Mitchison, T. J. (1997). Microtubule poly-

merization dynamics. Annu Rev Cell Dev Biol,13:83–117.

[Desai et al., 2013] Desai, R., Sarpal, R., Ishiyama, N., Pellikka, M., Ikura, M., and

Tepass, U. (2013). Monomeric alpha-catenin links cadherin to the actin cytoskele-

ton. Nat Cell Biol,15(3):261–73.

[Dettenhofer et al., 2008] Dettenhofer, M., Zhou, F., and Leder, P. (2008). Formin

1-isoform iv deﬁcient cells exhibit defects in cell spreading and focal adhesion for-

mation. PLoS One,3(6):e2497.

[DeWard and Alberts, 2009] DeWard, A. D. and Alberts, A. S. (2009). Ubiquitin-

mediated degradation of the formin mdia2 upon completion of cell division. J Biol

Chem,284(30):20061–9.

104 [DeWard et al., 2010] DeWard, A. D., Eisenmann, K. M., Matheson, S. F., and Al-

berts, A. S. (2010). The role of formins in human disease. Biochim Biophys Acta,

1803(2):226–33.

[Di Vizio et al., 2009] Di Vizio, D., Kim, J., Hager, M. H., Morello, M., Yang, W., La-

fargue, C. J., True, L. D., Rubin, M. A., Adam, R. M., Beroukhim, R., Demichelis,

F., and Freeman, M. R. (2009). Oncosome formation in prostate cancer: associa-

tion with a region of frequent chromosomal deletion in metastatic disease. Cancer

Res,69(13):5601–9.

[Dizon et al., 2011] Dizon, D. S., Sill, M. W., Gould, N., Rubin, S. C., Yamada,

S. D., Debernardo, R. L., Mannel, R. S., Eisenhauer, E. L., Duska, L. R., and

Fracasso, P. M. (2011). Phase i feasibility study of intraperitoneal cisplatin and

intravenous paclitaxel followed by intraperitoneal paclitaxel in untreated ovarian,

fallopian tube, and primary peritoneal carcinoma: a gynecologic oncology group

study. Gynecol Oncol,123(2):182–6.

[Do et al., 2007] Do, T. V., Symowicz, J. C., Berman, D. M., Liotta, L. A., Petricoin,

E. F., r., Stack, M. S., and Fishman, D. A. (2007). Lysophosphatidic acid down-

regulates stress ﬁbers and up-regulates pro-matrix metalloproteinase-2 activation

in ovarian cancer cells. Mol Cancer Res,5(2):121–31.

[Domchek et al., 2016] Domchek, S. M., Aghajanian, C., Shapira-Frommer, R.,

Schmutzler, R. K., Audeh, M. W., Friedlander, M., Balmana, J., Mitchell, G.,

Fried, G., Stemmer, S. M., Hubert, A., Rosengarten, O., Loman, N., Robertson,

J. D., Mann, H., and Kaufman, B. (2016). Eﬃcacy and safety of olaparib monother-

apy in germline brca1/2 mutation carriers with advanced ovarian cancer and three

or more lines of prior therapy. Gynecol Oncol,140(2):199–203.

105 [Doubeni et al., 2016] Doubeni, C. A., Doubeni, A. R., and Myers, A. E. (2016).

Diagnosis and management of ovarian cancer. Am Fam Physician,93(11):937–44.

[Drees et al., 2005] Drees, F., Pokutta, S., Yamada, S., Nelson, W. J., and Weis, W. I.

(2005). Alpha-catenin is a molecular switch that binds e-cadherin-beta-catenin and

regulates actin-ﬁlament assembly. Cell,123(5):903–15.

[Drew et al., 2016] Drew, Y., Ledermann, J., Hall, G., Rea, D., Glasspool, R., High-

ley, M., Jayson, G., Sludden, J., Murray, J., Jamieson, D., Halford, S., Acton, G.,

Backholer, Z., Mangano, R., Boddy, A., Curtin, N., and Plummer, R. (2016). Phase

2multicentretrialinvestigatingintermittentandcontinuousdosingschedulesofthe

poly(adp-ribose) polymerase inhibitor rucaparib in germline brca mutation carriers

with advanced ovarian and breast cancer. Br J Cancer,114(12):e21.

[D’Souza-Schorey, 2005] D’Souza-Schorey, C. (2005). Disassembling adherens junc-

tions: breaking up is hard to do. Trends Cell Biol,15(1):19–26.

[Ehrlich et al., 2002] Ehrlich, J. S., Hansen, M. D., and Nelson, W. J. (2002). Spatio-

temporal regulation of rac1 localization and lamellipodia dynamics during epithelial

cell-cell adhesion. Dev Cell,3(2):259–70.

[Eisenmann et al., 2009] Eisenmann, K. M., Dykema, K. J., Matheson, S. F., Kent,

N. F., DeWard, A. D., West, R. A., Tibes, R., Furge, K. A., and Alberts, A. S.

(2009). 5q- myelodysplastic syndromes: chromosome 5q genes direct a tumor-

suppression network sensing actin dynamics. Oncogene,28(39):3429–41.

[Eisenmann et al., 2007a] Eisenmann, K. M., Harris, E. S., Kitchen, S. M., Holman,

H. A., Higgs, H. N., and Alberts, A. S. (2007a). Dia-interacting protein modulates

formin-mediated actin assembly at the cell cortex. Curr Biol,17(7):579–91.

[Eisenmann et al., 2007b] Eisenmann, K. M., West, R. A., Hildebrand, D., Kitchen,

S. M., Peng, J., Sigler, R., Zhang, J., Siminovitch, K. A., and Alberts, A. S. (2007b). 106 Tcellresponsesinmammaliandiaphanous-relatedforminmdia1knock-outmice.

J Biol Chem,282(34):25152–8.

[Ellenbroek and Collard, 2007] Ellenbroek, S. I. and Collard, J. G. (2007). Rho gt-

pases: functions and association with cancer. Clin Exp Metastasis,24(8):657–72.

[Erami et al., 2015] Erami, Z., Timpson, P., Yao, W., Zaidel-Bar, R., and Anderson,

K. I. (2015). There are four dynamically and functionally distinct populations of

e-cadherin in cell junctions. Biol Open,4(11):1481–9.

[Etienne-Manneville et al., 2005] Etienne-Manneville, S., Manneville, J. B., Nicholls,

S., Ferenczi, M. A., and Hall, A. (2005). Cdc42 and par6-pkczeta regulate the

spatially localized association of dlg1 and apc to control cell polarization. J Cell

Biol,170(6):895–901.

[Fackler and Grosse, 2008] Fackler, O. T. and Grosse, R. (2008). Cell motility

through plasma membrane blebbing. J Cell Biol,181(6):879–84.

[Falconer et al., 2015] Falconer, H., Yin, L., Gronberg, H., and Altman, D. (2015).

Ovarian cancer risk after salpingectomy: a nationwide population-based study. J

Natl Cancer Inst,107(2).

[Farley et al., 2013] Farley, J., Brady, W. E., Vathipadiekal, V., Lankes, H. A., Cole-

man, R., Morgan, M. A., Mannel, R., Yamada, S. D., Mutch, D., Rodgers, W. H.,

Birrer, M., and Gershenson, D. M. (2013). Selumetinib in women with recurrent

low-grade serous carcinoma of the ovary or peritoneum: an open-label, single-arm,

phase 2 study. Lancet Oncol,14(2):134–40.

[Favaro et al., 2003] Favaro, P. M., de Souza Medina, S., Traina, F., Basseres, D. S.,

Costa, F. F., and Saad, S. T. (2003). Human leukocyte formin: a novel protein

expressed in lymphoid malignancies and associated with akt. Biochem Biophys Res

Commun,311(2):365–71. 107 [Feki et al., 2009] Feki, A., Berardi, P., Bellingan, G., Major, A., Krause, K. H.,

Petignat, P., Zehra, R., Pervaiz, S., and Irminger-Finger, I. (2009). Dissemination

of intraperitoneal ovarian cancer: Discussion of mechanisms and demonstration of

lymphatic spreading in ovarian cancer model. Crit Rev Oncol Hematol,72(1):1–9.

[Fernandez-Borja et al., 2005] Fernandez-Borja, M., Janssen, L., Verwoerd, D.,

Hordijk, P., and Neefjes, J. (2005). Rhob regulates endosome transport by pro-

moting actin assembly on endosomal membranes through dia1. J Cell Sci,118(Pt

12):2661–70.

[Fessenden et al., 2018] Fessenden, T. B., Beckham, Y., Perez-Neut, M., Ramirez-

San Juan, G., Chourasia, A. H., Macleod, K. F., Oakes, P. W., and Gardel, M. L.

(2018). Dia1-dependent adhesions are required by epithelial tissues to initiate

invasion. J Cell Biol,217(4):1485–1502.

[Firat-Karalar et al., 2011] Firat-Karalar, E. N., Hsiue, P. P., and Welch, M. D.

(2011). The actin nucleation factor jmy is a negative regulator of neuritogene-

sis. Mol Biol Cell,22(23):4563–74.

[Firat-Karalar and Welch, 2011] Firat-Karalar, E. N. and Welch, M. D. (2011). New

mechanisms and functions of actin nucleation. Curr Opin Cell Biol,23(1):4–13.

[Fischer-Colbrie et al., 1997] Fischer-Colbrie, J., Witt, A., Heinzl, H., Speiser, P.,

Czerwenka, K., Sevelda, P., and Zeillinger, R. (1997). Egfr and steroid receptors

in ovarian carcinoma: comparison with prognostic parameters and outcome of pa-

tients. Anticancer Res,17(1B):613–9.

[Fishman et al., 2001] Fishman, D. A., Liu, Y., Ellerbroek, S. M., and Stack, M. S.

(2001). Lysophosphatidic acid promotes matrix metalloproteinase (mmp) activa-

tion and mmp-dependent invasion in ovarian cancer cells. Cancer Res,61(7):3194–9.

108 [Fletcher and Mullins, 2010] Fletcher, D. A. and Mullins, R. D. (2010). Cell mechan-

ics and the cytoskeleton. Nature,463(7280):485–92.

[Florian and Mitchison, 2016] Florian, S. and Mitchison, T. J. (2016). Anti-

microtubule drugs. Methods Mol Biol,1413:403–21.

[Frost et al., 2009] Frost, A., Unger, V. M., and De Camilli, P. (2009). The bar

domain superfamily: membrane-molding macromolecules. Cell,137(2):191–6.

[Fujita et al., 2002] Fujita, Y., Krause, G., Scheﬀner, M., Zechner, D., Leddy, H. E.,

Behrens, J., Sommer, T., and Birchmeier, W. (2002). Hakai, a c-cbl-like protein,

ubiquitinates and induces endocytosis of the e-cadherin complex. Nat Cell Biol,

4(3):222–31.

[Fukagawa et al., 2015] Fukagawa, A., Ishii, H., Miyazawa, K., and Saitoh, M. (2015).

deltaef1 associates with dnmt1 and maintains dna methylation of the e-cadherin

promoter in breast cancer cells. Cancer Med,4(1):125–35.

[Fukata et al., 2003] Fukata, M., Nakagawa, M., and Kaibuchi, K. (2003). Roles of

rho-family gtpases in cell polarisation and directional migration. Curr Opin Cell

Biol,15(5):590–7.

[Galinska-Rakoczy et al., 2009] Galinska-Rakoczy, A., Wawro, B., and Strzelecka-

Golaszewska, H. (2009). New aspects of the spontaneous polymerization of actin

in the presence of salts. J Mol Biol,387(4):869–82.

[Gao et al., 2009] Gao, G., Chen, L., Dong, B., Gu, H., Dong, H., Pan, Y., Gao, Y.,

and Chen, X. (2009). Rhoa eﬀector mdia1 is required for pi 3-kinase-dependent

actin remodeling and spreading by thrombin in platelets. Biochem Biophys Res

Commun,385(3):439–44.

109 [Gasman et al., 2003] Gasman, S., Kalaidzidis, Y., and Zerial, M. (2003). Rhod reg-

ulates endosome dynamics through diaphanous-related formin and src tyrosine ki-

nase. Nat Cell Biol,5(3):195–204.

[Gasteier et al., 2005] Gasteier, J. E., Schroeder, S., Muranyi, W., Madrid, R., Beni-

chou, S., and Fackler, O. T. (2005). Fhod1 coordinates actin ﬁlament and micro-

tubule alignment to mediate cell elongation. Exp Cell Res,306(1):192–202.

[Gates and Peifer, 2005] Gates, J. and Peifer, M. (2005). Can 1000 reviews be wrong?

actin, alpha-catenin, and adherens junctions. Cell,123(5):769–72.

[Georgiou et al., 2008] Georgiou, M., Marinari, E., Burden, J., and Baum, B. (2008).

Cdc42, par6, and apkc regulate arp2/3-mediated endocytosis to control local ad-

herens junction stability. Curr Biol,18(21):1631–8.

[Gershenson, 2013] Gershenson, D. M. (2013). The life and times of low-grade serous

carcinoma of the ovary. Am Soc Clin Oncol Educ Book.

[Gershenson et al., 2006] Gershenson, D. M., Sun, C. C., Lu, K. H., Coleman, R. L.,

Sood, A. K., Malpica, A., Deavers, M. T., Silva, E. G., and Bodurka, D. C. (2006).

Clinical behavior of stage ii-iv low-grade serous carcinoma of the ovary. Obstet

Gynecol,108(2):361–8.

[Gershenson et al., 2015] Gershenson, D. M., Sun, C. C., and Wong, K. K. (2015).

Impact of mutational status on survival in low-grade serous carcinoma of the ovary

or peritoneum. Br J Cancer,113(9):1254–8.

[Gil et al., 2008] Gil, O. D., Lee, C., Ariztia, E. V., Wang, F. Q., Smith, P. J.,

Hope, J. M., and Fishman, D. A. (2008). Lysophosphatidic acid (lpa) promotes

e-cadherin ectodomain shedding and ovca429 cell invasion in an upa-dependent

manner. Gynecol Oncol,108(2):361–9.

110 [Gloushankova et al., 1997] Gloushankova, N. A., Alieva, N. A., Krendel, M. F., Bon-

der, E. M., Feder, H. H., Vasiliev, J. M., and Gelfand, I. M. (1997). Cell-cell contact

changes the dynamics of lamellar activity in nontransformed epitheliocytes but not

in their ras-transformed descendants. Proc Natl Acad Sci U S A,94(3):879–83.

[Goh and Ahmed, 2012] Goh, W. I. and Ahmed, S. (2012). mdia1-3 in mammalian

ﬁlopodia. Commun Integr Biol,5(4):340–4.

[Gorelik et al., 2011] Gorelik, R., Yang, C., Kameswaran, V., Dominguez, R., and

Svitkina, T. (2011). Mechanisms of plasma membrane targeting of formin mdia2

through its amino terminal domains. Mol Biol Cell,22(2):189–201.

[Grabowski et al., 2016] Grabowski, J. P., Harter, P., Heitz, F., Pujade-Lauraine, E.,

Reuss, A., Kristensen, G., Ray-Coquard, I., Heitz, J., Traut, A., Pﬁsterer, J., and

du Bois, A. (2016). Operability and chemotherapy responsiveness in advanced

low-grade serous ovarian cancer. an analysis of the ago study group metadatabase.

Gynecol Oncol,140(3):457–62.

[Gregory et al., 2008] Gregory, P. A., Bert, A. G., Paterson, E. L., Barry, S. C.,

Tsykin, A., Farshid, G., Vadas, M. A., Khew-Goodall, Y., and Goodall, G. J.

(2008). The mir-200 family and mir-205 regulate epithelial to mesenchymal tran-

sition by targeting zeb1 and sip1. Nat Cell Biol,10(5):593–601.

[Grikscheit et al., 2015] Grikscheit, K., Frank, T., Wang, Y., and Grosse, R. (2015).

Junctional actin assembly is mediated by formin-like 2 downstream of rac1. J Cell

Biol,209(3):367–76.

[Grikscheit and Grosse, 2016] Grikscheit, K. and Grosse, R. (2016). Formins at the

junction. Trends Biochem Sci,41(2):148–159.

[Grisham et al., 2013] Grisham, R. N., Iyer, G., Garg, K., Delair, D., Hyman, D. M.,

111 Zhou, Q., Iasonos, A., Berger, M. F., Dao, F., Spriggs, D. R., Levine, D. A., Agha-

janian, C., and Solit, D. B. (2013). Braf mutation is associated with early stage

disease and improved outcome in patients with low-grade serous ovarian cancer.

Cancer,119(3):548–554.

[Guadamillas et al., 2011] Guadamillas, M. C., Cerezo, A., and Del Pozo, M. A.

(2011). Overcoming anoikis–pathways to anchorage-independent growth in can-

cer. J Cell Sci,124(Pt19):3189–97.

[Gumbiner, 2005] Gumbiner, B. M. (2005). Regulation of cadherin-mediated adhe-

sion in morphogenesis. Nat Rev Mol Cell Biol,6(8):622–34.

[Gundersen and Bulinski, 1988] Gundersen, G. G. and Bulinski, J. C. (1988). Selec-

tive stabilization of microtubules oriented toward the direction of cell migration.

Proc Natl Acad Sci U S A,85(16):5946–50.

[Gupta et al., 2013] Gupta, N., Xu, Z., El-Sehemy, A., Steed, H., and Fu, Y. (2013).

Notch3 induces epithelial-mesenchymal transition and attenuates carboplatin-

induced apoptosis in ovarian cancer cells. Gynecol Oncol,130(1):200–6.

[Gupton et al., 2007] Gupton, S. L., Eisenmann, K., Alberts, A. S., and Waterman-

Storer, C. M. (2007). mdia2 regulates actin and focal adhesion dynamics and

organization in the lamella for eﬃcient epithelial cell migration. J Cell Sci,120(Pt

19):3475–87.

[Haga and Ridley, 2016] Haga, R. B. and Ridley, A. J. (2016). Rho gtpases: Regula-

tion and roles in cancer cell biology. Small GTPases,7(4):207–221.

[Hager et al., 2012] Hager, M. H., Morley, S., Bielenberg, D. R., Gao, S., Morello,

M., Holcomb, I. N., Liu, W., Mouneimne, G., Demichelis, F., Kim, J., Solomon,

K. R., Adam, R. M., Isaacs, W. B., Higgs, H. N., Vessella, R. L., Di Vizio, D., and

112 Freeman, M. R. (2012). Diaph3 governs the cellular transition to the amoeboid

tumour phenotype. EMBO Mol Med,4(8):743–60.

[Han et al., 2014] Han, S. P., Gambin, Y., Gomez, G. A., Verma, S., Giles, N.,

Michael, M., Wu, S. K., Guo, Z., Johnston, W., Sierecki, E., Parton, R. G.,

Alexandrov, K., and Yap, A. S. (2014). Cortactin scaﬀolds arp2/3 and wave2

at the epithelial zonula adherens. J Biol Chem,289(11):7764–75.

[Hanker et al., 2012] Hanker, L. C., Loibl, S., Burchardi, N., Pﬁsterer, J., Meier, W.,

Pujade-Lauraine, E., Ray-Coquard, I., Sehouli, J., Harter, P., and du Bois, A.

(2012). The impact of second to sixth line therapy on survival of relapsed ovarian

cancer after primary taxane/platinum-based therapy. Ann Oncol,23(10):2605–12.

[Harris et al., 2006] Harris, E. S., Rouiller, I., Hanein, D., and Higgs, H. N. (2006).

Mechanistic diﬀerences in actin bundling activity of two mammalian formins, frl1

and mdia2. J Biol Chem,281(20):14383–92.

[Harris and Tepass, 2008] Harris, K. P. and Tepass, U. (2008). Cdc42 and par pro-

teins stabilize dynamic adherens junctions in the drosophila neuroectoderm through

regulation of apical endocytosis. J Cell Biol,183(6):1129–43.

[Harris and Tepass, 2010] Harris, T. J. and Tepass, U. (2010). Adherens junctions:

from molecules to morphogenesis. Nat Rev Mol Cell Biol,11(7):502–14.

[Haslehurst et al., 2012] Haslehurst, A. M., Koti, M., Dharsee, M., Nuin, P., Evans,

K., Geraci, J., Childs, T., Chen, J., Li, J., Weberpals, J., Davey, S., Squire, J.,

Park, P. C., and Feilotter, H. (2012). Emt transcription factors snail and slug

directly contribute to cisplatin resistance in ovarian cancer. BMC Cancer,12:91.

[Helgeson et al., 2014] Helgeson, L. A., Prendergast, J. G., Wagner, A. R., Rodnick-

Smith, M., and Nolen, B. J. (2014). Interactions with actin monomers, actin

113 ﬁlaments, and arp2/3 complex deﬁne the roles of wasp family proteins and cortactin

in coordinately regulating branched actin networks. J Biol Chem,289(42):28856–

69.

[Helwani et al., 2004] Helwani, F. M., Kovacs, E. M., Paterson, A. D., Verma, S., Ali,

R. G., Fanning, A. S., Weed, S. A., and Yap, A. S. (2004). Cortactin is necessary

for e-cadherin-mediated contact formation and actin reorganization. J Cell Biol,

164(6):899–910.

[Herrmann and Aebi, 2000] Herrmann, H. and Aebi, U. (2000). Intermediate ﬁla-

ments and their associates: multi-talented structural elements specifying cytoar-

chitecture and cytodynamics. Curr Opin Cell Biol,12(1):79–90.

[Herzog et al., 2011] Herzog, D., Loetscher, P., van Hengel, J., Knusel, S., Brake-

busch, C., Taylor, V., Suter, U., and Relvas, J. B. (2011). The small gtpase rhoa is

required to maintain spinal cord neuroepithelium organization and the neural stem

cell pool. J Neurosci,31(13):5120–30.

[Higashi et al., 2008] Higashi, T., Ikeda, T., Shirakawa, R., Kondo, H., Kawato, M.,

Horiguchi, M., Okuda, T., Okawa, K., Fukai, S., Nureki, O., Kita, T., and Ho-

riuchi, H. (2008). Biochemical characterization of the rho gtpase-regulated actin

assembly by diaphanous-related formins, mdia1 and daam1, in platelets. J Biol

Chem,283(13):8746–55.

[Higashida et al., 2004] Higashida, C., Miyoshi, T., Fujita, A., Oceguera-Yanez, F.,

Monypenny, J., Andou, Y., Narumiya, S., and Watanabe, N. (2004). Actin

polymerization-driven molecular movement of mdia1 in living cells. Science,

303(5666):2007–10.

[Higgs, 2005] Higgs, H. N. (2005). Formin proteins: a domain-based approach. Trends

Biochem Sci,30(6):342–53. 114 [Higgs and Peterson, 2005] Higgs, H. N. and Peterson, K. J. (2005). Phylogenetic

analysis of the formin homology 2 domain. Mol Biol Cell,16(1):1–13.

[Hill et al., 1995] Hill, C. S., Wynne, J., and Treisman, R. (1995). The rho family

gtpases rhoa, rac1, and cdc42hs regulate transcriptional activation by srf. Cell,

81(7):1159–70.

[Hirokawa et al., 1983] Hirokawa, N., Keller, T. C., r., Chasan, R., and Mooseker,

M. S. (1983). Mechanism of brush border contractility studied by the quick-freeze,

deep-etch method. J Cell Biol,96(5):1325–36.

[Homem and Peifer, 2008] Homem, C. C. and Peifer, M. (2008). Diaphanous reg-

ulates myosin and adherens junctions to control cell contractility and protrusive

behavior during morphogenesis. Development,135(6):1005–18.

[Hong et al., 2013] Hong, S., Troyanovsky, R. B., and Troyanovsky, S. M. (2013).

Binding to f-actin guides cadherin cluster assembly, stability, and movement. J

Cell Biol,201(1):131–43.

[Hotulainen and Lappalainen, 2006] Hotulainen, P. and Lappalainen, P. (2006).

Stress ﬁbers are generated by two distinct actin assembly mechanisms in motile

cells. J Cell Biol,173(3):383–94.

[Howard and Hyman, 2003] Howard, J. and Hyman, A. A. (2003). Dynamics and

mechanics of the microtubule plus end. Nature,422(6933):753–8.

[Howitt et al., 2015] Howitt, B. E., Hanamornroongruang, S., Lin, D. I., Conner,

J. E., Schulte, S., Horowitz, N., Crum, C. P., and Meserve, E. E. (2015). Evi-

dence for a dualistic model of high-grade serous carcinoma: Brca mutation status,

histology, and tubal intraepithelial carcinoma. Am J Surg Pathol,39(3):287–93.

115 [Hu et al., 2014] Hu, J., Lu, J., Lian, G., Ferland, R. J., Dettenhofer, M., and Sheen,

V. L. (2014). Formin 1 and ﬁlamin b physically interact to coordinate chondrocyte

proliferation and diﬀerentiation in the growth plate. Hum Mol Genet,23(17):4663–

73.

[Hunn and Rodriguez, 2012] Hunn, J. and Rodriguez, G. C. (2012). Ovarian cancer:

etiology, risk factors, and epidemiology. Clin Obstet Gynecol,55(1):3–23.

[Iino et al., 2001] Iino, R., Koyama, I., and Kusumi, A. (2001). Single molecule imag-

ing of green ﬂuorescent proteins in living cells: E-cadherin forms oligomers on the

free cell surface. Biophys J,80(6):2667–77.

[Imai et al., 2004] Imai, T., Horiuchi, A., Shiozawa, T., Osada, R., Kikuchi, N.,

Ohira, S., Oka, K., and Konishi, I. (2004). Elevated expression of e-cadherin and

alpha-, beta-, and gamma-catenins in metastatic lesions compared with primary

epithelial ovarian carcinomas. Hum Pathol,35(12):1469–76.

[Ireton et al., 2002] Ireton, R. C., Davis, M. A., van Hengel, J., Mariner, D. J.,

Barnes, K., Thoreson, M. A., Anastasiadis, P. Z., Matrisian, L., Bundy, L. M.,

Sealy, L., Gilbert, B., van Roy, F., and Reynolds, A. B. (2002). A novel role for

p120 catenin in e-cadherin function. J Cell Biol,159(3):465–76.

[Ishiyama et al., 2010] Ishiyama, N., Lee, S. H., Liu, S., Li, G. Y., Smith, M. J., Re-

ichardt, L. F., and Ikura, M. (2010). Dynamic and static interactions between p120

catenin and e-cadherin regulate the stability of cell-cell adhesion. Cell,141(1):117–

28.

[Isogai et al., 2015] Isogai, T., van der Kammen, R., and Innocenti, M. (2015). Smifh2

has eﬀects on formins and p53 that perturb the cell cytoskeleton. Sci Rep,5:9802.

[Ivanov et al., 2005] Ivanov, A. I., Hunt, D., Utech, M., Nusrat, A., and Parkos, C. A.

116 (2005). Diﬀerential roles for actin polymerization and a myosin ii motor in assembly

of the epithelial apical junctional complex. Mol Biol Cell,16(6):2636–50.

[Iwaya et al., 2007] Iwaya, K., Oikawa, K., Semba, S., Tsuchiya, B., Mukai, Y., Ot-

subo, T., Nagao, T., Izumi, M., Kuroda, M., Domoto, H., and Mukai, K. (2007).

Correlation between liver metastasis of the colocalization of actin-related protein

2and3complexandwave2incolorectalcarcinoma.Cancer Sci,98(7):992–9.

[Jaaback et al., 2016] Jaaback, K., Johnson, N., and Lawrie, T. A. (2016). Intraperi-

toneal chemotherapy for the initial management of primary epithelial ovarian can-

cer. Cochrane Database Syst Rev,(1):CD005340.

[Jeanes et al., 2008] Jeanes, A., Gottardi, C. J., and Yap, A. S. (2008). Cadherins

and cancer: how does cadherin dysfunction promote tumor progression? Oncogene,

27(55):6920–9.

[Jegou et al., 2013] Jegou, A., Carlier, M. F., and Romet-Lemonne, G. (2013).

Formin mdia1 senses and generates mechanical forces on actin ﬁlaments. Nat Com-

mun,4:1883.

[Jelovac and Armstrong, 2011] Jelovac, D. and Armstrong, D. K. (2011). Recent

progress in the diagnosis and treatment of ovarian cancer. CA Cancer J Clin,

61(3):183–203.

[Ji et al., 2008] Ji, P., Jayapal, S. R., and Lodish, H. F. (2008). Enucleation of

cultured mouse fetal erythroblasts requires rac gtpases and mdia2. Nat Cell Biol,

10(3):314–21.

[Jia et al., 2018] Jia, D., Nagaoka, Y., Katsumata, M., and Orsulic, S. (2018). In-

ﬂammation is a key contributor to ovarian cancer cell seeding. Sci Rep,8(1):12394.

117 [Jiang, 2017] Jiang, J. (2017). Diaphanous-related formin-3 overexpression inhibits

the migration and invasion of triple-negative breast cancer by inhibiting rhoa-gtp

expression. Biomed Pharmacother,94:439–445.

[Jordan and Wurm, 2004] Jordan, M. and Wurm, F. (2004). Transfection of adherent

and suspended cells by calcium phosphate. Methods,33(2):136–43.

[Jordan, 2002] Jordan, M. A. (2002). Mechanism of action of antitumor drugs that

interact with microtubules and tubulin. Curr Med Chem Anticancer Agents,2(1):1–

17.

[Jurmeister et al., 2012] Jurmeister, S., Baumann, M., Balwierz, A., Keklikoglou,

I., Ward, A., Uhlmann, S., Zhang, J. D., Wiemann, S., and Sahin, O. (2012).

Microrna-200c represses migration and invasion of breast cancer cells by targeting

actin-regulatory proteins fhod1 and ppm1f. Mol Cell Biol,32(3):633–51.

[Kajiyama et al., 2007] Kajiyama, H., Shibata, K., Terauchi, M., Yamashita, M.,

Ino, K., Nawa, A., and Kikkawa, F. (2007). Chemoresistance to paclitaxel induces

epithelial-mesenchymal transition and enhances metastatic potential for epithelial

ovarian carcinoma cells. Int J Oncol,31(2):277–83.

[Kaldawy et al., 2016] Kaldawy, A., Segev, Y., Lavie, O., Auslender, R., Sopik, V.,

and Narod, S. A. (2016). Low-grade serous ovarian cancer: A review. Gynecol

Oncol,143(2):433–438.

[Kamasani et al., 2007] Kamasani, U., Duhadaway, J. B., Alberts, A. S., and Pren-

dergast, G. C. (2007). mdia function is critical for the cell suicide program triggered

by farnesyl transferase inhibition. Cancer Biol Ther,6(9):1422–7.

[Kametani and Takeichi, 2007] Kametani, Y. and Takeichi, M. (2007). Basal-to-

apical cadherin ﬂow at cell junctions. Nat Cell Biol,9(1):92–8.

118 [Karki et al., 2002] Karki, S., Ligon, L. A., DeSantis, J., Tokito, M., and Holzbaur,

E. L. (2002). Plac-24 is a cytoplasmic dynein-binding protein that is recruited to

sites of cell-cell contact. Mol Biol Cell,13(5):1722–34.

[Karst and Drapkin, 2010] Karst, A. M. and Drapkin, R. (2010). Ovarian cancer

pathogenesis: a model in evolution. J Oncol,2010:932371.

[Katayama et al., 2012] Katayama, K., Leslie, J. R., Lang, R. A., Zheng, Y., and

Yoshida, Y. (2012). Left-right locomotor circuitry depends on rhoa-driven organiza-

tion of the neuroepithelium in the developing spinal cord. J Neurosci,32(30):10396–

407.

[Kato et al., 2001] Kato, T., Watanabe, N., Morishima, Y., Fujita, A., Ishizaki, T.,

and Narumiya, S. (2001). Localization of a mammalian homolog of diaphanous,

mdia1, to the mitotic spindle in hela cells. J Cell Sci,114(Pt4):775–84.

[Katoh and Katoh, 2004] Katoh, M. and Katoh, M. (2004). Identiﬁcation and char-

acterization of human diaph3 gene in silico. Int J Mol Med,13(3):473–8.

[Kenny et al., 2014] Kenny, H. A., Chiang, C. Y., White, E. A., Schryver, E. M.,

Habis, M., Romero, I. L., Ladanyi, A., Penicka, C. V., George, J., Matlin, K., Mon-

tag, A., Wroblewski, K., Yamada, S. D., Mazar, A. P., Bowtell, D., and Lengyel,

E. (2014). Mesothelial cells promote early ovarian cancer metastasis through ﬁ-

bronectin secretion. J Clin Invest,124(10):4614–28.

[Kim et al., 2016] Kim, D., Jung, J., You, E., Ko, P., Oh, S., and Rhee, S. (2016).

mdia1 regulates breast cancer invasion by controlling membrane type 1-matrix

metalloproteinase localization. Oncotarget,7(14):17829–43.

[Kim et al., 2014] Kim, J., Morley, S., Le, M., Bedoret, D., Umetsu, D. T., Di Vizio,

D., and Freeman, M. R. (2014). Enhanced shedding of extracellular vesicles from

119 amoeboid prostate cancer cells: potential eﬀects on the tumor microenvironment.

Cancer Biol Ther,15(4):409–18.

[Kim et al., 2018] Kim, J., Park, E. Y., Kim, O., Schilder, J. M., Coﬀey, D. M., Cho,

C. H., and Bast, R. C., J. (2018). Cell origins of high-grade serous ovarian cancer.

Cancers (Basel),10(11).

[Kim et al., 2000] Kim, J. B., Islam, S., Kim, Y. J., Prudoﬀ, R. S., Sass, K. M.,

Wheelock, M. J., and Johnson, K. R. (2000). N-cadherin extracellular repeat 4

mediates epithelial to mesenchymal transition and increased motility. J Cell Biol,

151(6):1193–206.

[Kindelberger et al., 2007] Kindelberger, D. W., Lee, Y., Miron, A., Hirsch, M. S.,

Feltmate, C., Medeiros, F., Callahan, M. J., Garner, E. O., Gordon, R. W., Birch,

C., Berkowitz, R. S., Muto, M. G., and Crum, C. P. (2007). Intraepithelial carci-

noma of the ﬁmbria and pelvic serous carcinoma: Evidence for a causal relationship.

Am J Surg Pathol,31(2):161–9.

[Kishikawa et al., 2008] Kishikawa, M., Suzuki, A., and Ohno, S. (2008). apkc en-

ables development of zonula adherens by antagonizing centripetal contraction of

the circumferential actomyosin cables. J Cell Sci,121(Pt15):2481–92.

[Klymenko et al., 2017] Klymenko, Y., Kim, O., and Stack, M. S. (2017). Complex

determinants of epithelial: Mesenchymal phenotypic plasticity in ovarian cancer.

Cancers (Basel),9(8).

[Kobielak and Fuchs, 2004] Kobielak, A. and Fuchs, E. (2004). Alpha-catenin: at

the junction of intercellular adhesion and actin dynamics. Nat Rev Mol Cell Biol,

5(8):614–25.

[Kobielak et al., 2004] Kobielak, A., Pasolli, H. A., and Fuchs, E. (2004). Mam-

120 malian formin-1 participates in adherens junctions and polymerization of linear

actin cables. Nat Cell Biol,6(1):21–30.

[Kodama et al., 2004] Kodama, A., Lechler, T., and Fuchs, E. (2004). Coordinating

cytoskeletal tracks to polarize cellular movements. J Cell Biol,167(2):203–7.

[Kon et al., 2008] Kon, S., Tanabe, K., Watanabe, T., Sabe, H., and Satake, M.

(2008). Clathrin dependent endocytosis of e-cadherin is regulated by the arf6gap

isoform smap1. Exp Cell Res,314(7):1415–28.

[Koshiyama et al., 2014] Koshiyama, M., Matsumura, N., and Konishi, I. (2014).

Recent concepts of ovarian carcinogenesis: type i and type ii. Biomed Res Int,

2014:934261.

[Kotsopoulos et al., 2010] Kotsopoulos, J., Baer, H. J., and Tworoger, S. S. (2010).

Anthropometric measures and risk of epithelial ovarian cancer: results from the

nurses’ health study. Obesity (Silver Spring),18(8):1625–31.

[Kourtidis et al., 2017] Kourtidis, A., Lu, R., Pence, L. J., and Anastasiadis, P. Z.

(2017). A central role for cadherin signaling in cancer. Exp Cell Res,358(1):78–85.

[Kourtidis et al., 2013] Kourtidis, A., Ngok, S. P., and Anastasiadis, P. Z. (2013).

p120 catenin: an essential regulator of cadherin stability, adhesion-induced signal-

ing, and cancer progression. Prog Mol Biol Transl Sci,116:409–32.

[Kovacs et al., 2002] Kovacs, E. M., Ali, R. G., McCormack, A. J., and Yap, A. S.

(2002). E-cadherin homophilic ligation directly signals through rac and phos-

phatidylinositol 3-kinase to regulate adhesive contacts. J Biol Chem,277(8):6708–

18.

[Kovacs et al., 2011] Kovacs, E. M., Verma, S., Ali, R. G., Ratheesh, A., Hamilton,

N. A., Akhmanova, A., and Yap, A. S. (2011). N-wasp regulates the epithelial

121 junctional actin cytoskeleton through a non-canonical post-nucleation pathway.

Nat Cell Biol,13(8):934–43.

[Kovar, 2006] Kovar, D. R. (2006). Molecular details of formin-mediated actin as-

sembly. Curr Opin Cell Biol,18(1):11–7.

[Krainer et al., 2013] Krainer, E. C., Ouderkirk, J. L., Miller, E. W., Miller, M. R.,

Mersich, A. T., and Blystone, S. D. (2013). The multiplicity of human formins:

Expression patterns in cells and tissues. Cytoskeleton (Hoboken),70(8):424–38.

[Krendel and Bonder, 1999] Krendel, M. F. and Bonder, E. M. (1999). Analysis of

actin ﬁlament bundle dynamics during contact formation in live epithelial cells.

Cell Motil Cytoskeleton,43(4):296–309.

[Kuhn and Geyer, 2014] Kuhn, S. and Geyer, M. (2014). Formins as eﬀector proteins

of rho gtpases. Small GTPases,5:e29513.

[Kurman and Shih Ie, 2016] Kurman, R. J. and Shih Ie, M. (2016). The dualistic

model of ovarian carcinogenesis: Revisited, revised, and expanded. Am J Pathol,

186(4):733–47.

[Kurrey et al., 2009] Kurrey, N. K., Jalgaonkar, S. P., Joglekar, A. V., Ghanate,

A. D., Chaskar, P. D., Doiphode, R. Y., and Bapat, S. A. (2009). Snail and

slug mediate radioresistance and chemoresistance by antagonizing p53-mediated

apoptosis and acquiring a stem-like phenotype in ovarian cancer cells. Stem Cells,

27(9):2059–68.

[Lambert et al., 2002] Lambert, M., Choquet, D., and Mege, R. M. (2002). Dynamics

of ligand-induced, rac1-dependent anchoring of cadherins to the actin cytoskeleton.

J Cell Biol,157(3):469–79.

122 [Lammers et al., 2008] Lammers, M., Meyer, S., Kuhlmann, D., and Wittinghofer,

A. (2008). Speciﬁcity of interactions between mdia isoforms and rho proteins. J

Biol Chem,283(50):35236–46.

[Lamouille et al., 2014] Lamouille, S., Xu, J., and Derynck, R. (2014). Molecu-

lar mechanisms of epithelial-mesenchymal transition. Nat Rev Mol Cell Biol,

15(3):178–96.

[Lash et al., 2013] Lash, L. L., Wallar, B. J., Turner, J. D., Vroegop, S. M., Kilkuskie,

R. E., Kitchen-Goosen, S. M., Xu, H. E., and Alberts, A. S. (2013). Small-molecule

intramimics of formin autoinhibition: a new strategy to target the cytoskeletal

remodeling machinery in cancer cells. Cancer Res,73(22):6793–803.

[Latiﬁ et al., 2012] Latiﬁ, A., Luwor, R. B., Bilandzic, M., Nazaretian, S., Stenvers,

K., Pyman, J., Zhu, H., Thompson, E. W., Quinn, M. A., Findlay, J. K., and

Ahmed, N. (2012). Isolation and characterization of tumor cells from the ascites

of ovarian cancer patients: molecular phenotype of chemoresistant ovarian tumors.

PLoS One,7(10):e46858.

[Le et al., 2003] Le, X. F., Hittelman, W. N., Liu, J., McWatters, A., Li, C., Mills,

G. B., and Bast, R. C., J. (2003). Paclitaxel induces inactivation of p70 s6 ki-

nase and phosphorylation of thr421 and ser424 via multiple signaling pathways in

mitosis. Oncogene,22(4):484–97.

[le Duc et al., 2010] le Duc, Q., Shi, Q., Blonk, I., Sonnenberg, A., Wang, N., Leck-

band, D., and de Rooij, J. (2010). Vinculin potentiates e-cadherin mechanosensing

and is recruited to actin-anchored sites within adherens junctions in a myosin ii-

dependent manner. J Cell Biol,189(7):1107–15.

[Leader et al., 2002] Leader, B., Lim, H., Carabatsos, M. J., Harrington, A., Ecsedy,

123 J., Pellman, D., Maas, R., and Leder, P. (2002). Formin-2, polyploidy, hypofertility

and positioning of the meiotic spindle in mouse oocytes. Nat Cell Biol,4(12):921–8.

[LeCorgne et al., 2018] LeCorgne, H., Tudosie, A. M., Lavik, K., Su, R., Becker,

K. N., Moore, S., Walia, Y., Wisner, A., Koehler, D., Alberts, A. S., Williams, F. E.,

and Eisenmann, K. M. (2018). Diﬀerential toxicity of mdia formin-directed func-

tional agonists and antagonists in developing zebraﬁsh. Front Pharmacol,9:340.

[Lee et al., 2018] Lee, J. Y., Kim, S., Kim, Y. T., Lim, M. C., Lee, B., Jung, K. W.,

Kim, J. W., Park, S. Y., and Won, Y. J. (2018). Changes in ovarian cancer survival

during the 20 years before the era of targeted therapy. BMC Cancer,18(1):601.

[Lee et al., 2011] Lee, Y. C., Cheng, C. J., Bilen, M. A., Lu, J. F., Satcher, R. L.,

Yu-Lee, L. Y., Gallick, G. E., Maity, S. N., and Lin, S. H. (2011). Bmp4 promotes

prostate tumor growth in bone through osteogenesis. Cancer Res,71(15):5194–203.

[Leibfried et al., 2008] Leibfried, A., Fricke, R., Morgan, M. J., Bogdan, S., and Bel-

laiche, Y. (2008). Drosophila cip4 and wasp deﬁne a branch of the cdc42-par6-apkc

pathway regulating e-cadherin endocytosis. Curr Biol,18(21):1639–48.

[Li et al., 2011] Li, D., Hallett, M. A., Zhu, W., Rubart, M., Liu, Y., Yang, Z., Chen,

H., Haneline, L. S., Chan, R. J., Schwartz, R. J., Field, L. J., Atkinson, S. J., and

Shou, W. (2011). Dishevelled-associated activator of morphogenesis 1 (daam1) is

required for heart morphogenesis. Development,138(2):303–15.

[Li and Higgs, 2005] Li, F. and Higgs, H. N. (2005). Dissecting requirements for auto-

inhibition of actin nucleation by the formin, mdia1. J Biol Chem,280(8):6986–92.

[Li et al., 2010] Li, Y., Zhu, X., Zeng, Y., Wang, J., Zhang, X., Ding, Y. Q., and

Liang, L. (2010). Fmnl2 enhances invasion of colorectal carcinoma by inducing

epithelial-mesenchymal transition. Mol Cancer Res,8(12):1579–90.

124 [Ligon and Holzbaur, 2007] Ligon, L. A. and Holzbaur, E. L. (2007). Microtubules

tethered at epithelial cell junctions by dynein facilitate eﬃcient junction assembly.

Traﬃc,8(7):808–19.

[Ligon et al., 2001] Ligon, L. A., Karki, S., Tokito, M., and Holzbaur, E. L. (2001).

Dynein binds to beta-catenin and may tether microtubules at adherens junctions.

Nat Cell Biol,3(10):913–7.

[Lin et al., 2002] Lin, S. X., Gundersen, G. G., and Maxﬁeld, F. R. (2002). Export

from pericentriolar endocytic recycling compartment to cell surface depends on

stable, detyrosinated (glu) microtubules and kinesin. Mol Biol Cell,13(1):96–109.

[Ling et al., 2007] Ling, K., Bairstow, S. F., Carbonara, C., Turbin, D. A., Hunts-

man, D. G., and Anderson, R. A. (2007). Type i gamma phosphatidylinositol

phosphate kinase modulates adherens junction and e-cadherin traﬃcking via a di-

rect interaction with mu 1b adaptin. J Cell Biol,176(3):343–53.

[Liu et al., 2012] Liu, Y., Burkhalter, R., Symowicz, J., Chaﬃn, K., Ellerbroek, S.,

and Stack, M. S. (2012). Lysophosphatidic acid disrupts junctional integrity and

epithelial cohesion in ovarian cancer cells. J Oncol,2012:501492.

[Lizarraga et al., 2009] Lizarraga, F., Poincloux, R., Romao, M., Montagnac, G.,

Le Dez, G., Bonne, I., Rigaill, G., Raposo, G., and Chavrier, P. (2009). Diaphanous-

related formins are required for invadopodia formation and invasion of breast tumor

cells. Cancer Res,69(7):2792–800.

[Lodish, 2008] Lodish, H. (2008). Molecular cell biology. Freeman, New York.

[Lynch et al., 1997] Lynch, E. D., Lee, M. K., Morrow, J. E., Welcsh, P. L., Leon,

P. E., and King, M. C. (1997). Nonsyndromic deafness dfna1 associated with

mutation of a human homolog of the drosophila gene diaphanous. Science,

278(5341):1315–8. 125 [MacDonald et al., 2009] MacDonald, B. T., Tamai, K., and He, X. (2009).

Wnt/beta-catenin signaling: components, mechanisms, and diseases. Dev Cell,

17(1):9–26.

[Magie et al., 1999] Magie, C. R., Meyer, M. R., Gorsuch, M. S., and Parkhurst,

S. M. (1999). Mutations in the rho1 small gtpase disrupt morphogenesis and seg-

mentation during early drosophila development. Development,126(23):5353–64.

[Manning et al., 2000] Manning, T. J., J., Parker, J. C., and Sontheimer, H. (2000).

Role of lysophosphatidic acid and rho in glioma cell motility. Cell Motil Cytoskele-

ton,45(3):185–99.

[Matschegewski et al., 2012] Matschegewski, C., Staehlke, S., Birkholz, H., Lange,

R., Beck, U., Engel, K., and Nebe, J. B. (2012). Automatic actin ﬁlament quan-

tiﬁcation of osteoblasts and their morphometric analysis on microtextured silicon-

titanium arrays. Materials (Basel),5(7):1176–95.

[McAlpine et al., 2014] McAlpine, J. N., Hanley, G. E., Woo, M. M., Tone, A. A.,

Rozenberg, N., Swenerton, K. D., Gilks, C. B., Finlayson, S. J., Huntsman, D. G.,

Miller, D. M., and Ovarian Cancer Research Program of British, C. (2014). Op-

portunistic salpingectomy: uptake, risks, and complications of a regional initiative

for ovarian cancer prevention. Am J Obstet Gynecol,210(5):471e1–11.

[McCluggage, 2011] McCluggage, W. G. (2011). Morphological subtypes of ovarian

carcinoma: a review with emphasis on new developments and pathogenesis. Pathol-

ogy,43(5):420–32.

[McCrea and Gottardi, 2016] McCrea, P. D. and Gottardi, C. J. (2016). Beyond

beta-catenin: prospects for a larger catenin network in the nucleus. Nat Rev Mol

Cell Biol,17(1):55–64.

126 [McGuire et al., 1996] McGuire, W. P., Hoskins, W. J., Brady, M. F., Kucera, P. R.,

Partridge, E. E., Look, K. Y., Clarke-Pearson, D. L., and Davidson, M. (1996).

Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients

with stage iii and stage iv ovarian cancer. N Engl J Med,334(1):1–6.

[McLaughlin et al., 2007] McLaughlin, M. E., Kruger, G. M., Slocum, K. L., Crowley,

D., Michaud, N. A., Huang, J., Magendantz, M., and Jacks, T. (2007). The nf2

tumor suppressor regulates cell-cell adhesion during tissue fusion. Proc Natl Acad

Sci U S A,104(9):3261–6.

[McTiernan, 2008] McTiernan, A. (2008). Mechanisms linking physical activity with

cancer. Nat Rev Cancer,8(3):205–11.

[Mei et al., 2010] Mei, L., Chen, H., Wei, D. M., Fang, F., Liu, G. J., Xie, H. Y.,

Wang, X., Zou, J., Han, X., and Feng, D. (2010). Maintenance chemotherapy for

ovarian cancer. Cochrane Database Syst Rev,(9):Cd007414.

[Meng et al., 2008] Meng, W., Mushika, Y., Ichii, T., and Takeichi, M. (2008). An-

chorage of microtubule minus ends to adherens junctions regulates epithelial cell-

cell contacts. Cell,135(5):948–59.

[Meng et al., 2004] Meng, W., Numazaki, M., Takeuchi, K., Uchibori, Y., Ando-

Akatsuka, Y., Tominaga, M., and Tominaga, T. (2004). Dip (mdia interacting

protein) is a key molecule regulating rho and rac in a src-dependent manner. EMBO

J,23(4):760–71.

[Mertens et al., 2005] Mertens, A. E., Rygiel, T. P., Olivo, C., van der Kammen, R.,

and Collard, J. G. (2005). The rac activator tiam1 controls tight junction biogenesis

in keratinocytes through binding to and activation of the par polarity complex. J

Cell Biol,170(7):1029–37.

127 [Miow et al., 2015] Miow, Q. H., Tan, T. Z., Ye, J., Lau, J. A., Yokomizo, T., Thiery,

J. P., and Mori, S. (2015). Epithelial-mesenchymal status renders diﬀerential re-

sponses to cisplatin in ovarian cancer. Oncogene,34(15):1899–907.

[Miyake et al., 2006] Miyake, Y., Inoue, N., Nishimura, K., Kinoshita, N., Hosoya,

H., and Yonemura, S. (2006). Actomyosin tension is required for correct recruitment

of adherens junction components and zonula occludens formation. Exp Cell Res,

312(9):1637–50.

[Monteiro et al., 2013] Monteiro, P., Rosse, C., Castro-Castro, A., Irondelle, M.,

Lagoutte, E., Paul-Gilloteaux, P., Desnos, C., Formstecher, E., Darchen, F., Per-

rais, D., Gautreau, A., Hertzog, M., and Chavrier, P. (2013). Endosomal wash

and exocyst complexes control exocytosis of mt1-mmp at invadopodia. J Cell Biol,

203(6):1063–79.

[Moore et al., 1970] Moore, P. B., Huxley, H. E., and DeRosier, D. J. (1970). Three-

dimensional reconstruction of f-actin, thin ﬁlaments and decorated thin ﬁlaments.

J Mol Biol,50(2):279–95.

[Nakagawa et al., 2001] Nakagawa, M., Fukata, M., Yamaga, M., Itoh, N., and

Kaibuchi, K. (2001). Recruitment and activation of rac1 by the formation of e-

cadherin-mediated cell-cell adhesion sites. J Cell Sci,114(Pt10):1829–38.

[Nakamura et al., 1998] Nakamura, T., Hamada, F., Ishidate, T., Anai, K., Kawa-

hara, K., Toyoshima, K., and Akiyama, T. (1998). Axin, an inhibitor of the wnt

signalling pathway, interacts with beta-catenin, gsk-3beta and apc and reduces the

beta-catenin level. Genes Cells,3(6):395–403.

[Neuhaus et al., 1983] Neuhaus, J. M., Wanger, M., Keiser, T., and Wegner, A.

(1983). Treadmilling of actin. J Muscle Res Cell Motil,4(5):507–27.

128 [Noren et al., 2001] Noren, N. K., Niessen, C. M., Gumbiner, B. M., and Burridge,

K. (2001). Cadherin engagement regulates rho family gtpases. J Biol Chem,

276(36):33305–8.

[Olson et al., 1995] Olson, M. F., Ashworth, A., and Hall, A. (1995). An essential

role for rho, rac, and cdc42 gtpases in cell cycle progression through g1. Science,

269(5228):1270–2.

[Oza et al., 2015] Oza, A. M., Cibula, D., Benzaquen, A. O., Poole, C., Mathijssen,

R. H., Sonke, G. S., Colombo, N., Spacek, J., Vuylsteke, P., Hirte, H., Mahner, S.,

Plante, M., Schmalfeldt, B., Mackay, H., Rowbottom, J., Lowe, E. S., Dougherty,

B., Barrett, J. C., and Friedlander, M. (2015). Olaparib combined with chemother-

apy for recurrent platinum-sensitive ovarian cancer: a randomised phase 2 trial.

Lancet Oncol,16(1):87–97.

[Ozols, 2006] Ozols, R. F. (2006). Systemic therapy for ovarian cancer: current status

and new treatments. Semin Oncol,33(2Suppl6):S3–11.

[Parkes and Kennedy, 2016] Parkes, E. E. and Kennedy, R. D. (2016). Clinical ap-

plication of poly(adp-ribose) polymerase inhibitors in high-grade serous ovarian

cancer. Oncologist,21(5):586–93.

[Paul and Pollard, 2009] Paul, A. S. and Pollard, T. D. (2009). Review of the mech-

anism of processive actin ﬁlament elongation by formins. Cell Motil Cytoskeleton,

66(8):606–17.

[Peinado et al., 2004] Peinado, H., Ballestar, E., Esteller, M., and Cano, A. (2004).

Snail mediates e-cadherin repression by the recruitment of the sin3a/histone

deacetylase 1 (hdac1)/hdac2 complex. Mol Cell Biol,24(1):306–19.

[Pellegrin and Mellor, 2005] Pellegrin, S. and Mellor, H. (2005). The rho family gt-

pase rif induces ﬁlopodia through mdia2. Curr Biol,15(2):129–33. 129 [Peng et al., 2007] Peng, J., Kitchen, S. M., West, R. A., Sigler, R., Eisenmann,

K. M., and Alberts, A. S. (2007). Myeloproliferative defects following targeting of

the drf1 gene encoding the mammalian diaphanous related formin mdia1. Cancer

Res,67(16):7565–71.

[Peng et al., 2003] Peng, J., Wallar, B. J., Flanders, A., Swiatek, P. J., and Alberts,

A. S. (2003). Disruption of the diaphanous-related formin drf1 gene encoding mdia1

reveals a role for drf3 as an eﬀector for cdc42. Curr Biol,13(7):534–45.

[Perez-Moreno and Fuchs, 2006] Perez-Moreno, M. and Fuchs, E. (2006). Catenins:

keeping cells from getting their signals crossed. Dev Cell,11(5):601–12.

[Petrova et al., 2012] Petrova, Y. I., Spano, M. M., and Gumbiner, B. M. (2012).

Conformational epitopes at cadherin calcium-binding sites and p120-catenin phos-

phorylation regulate cell adhesion. Mol Biol Cell,23(11):2092–108.

[Pettee et al., 2014] Pettee, K. M., Dvorak, K. M., Nestor-Kalinoski, A. L., and

Eisenmann, K. M. (2014). An mdia2/rock signaling axis regulates invasive egress

from epithelial ovarian cancer spheroids. PLoS One,9(2):e90371.

[Phillips et al., 2012] Phillips, K. G., Velasco, C. R., Li, J., Kolatkar, A., Luttgen,

M., Bethel, K., Duggan, B., Kuhn, P., and McCarty, O. J. (2012). Optical quan-

tiﬁcation of cellular mass, volume, and density of circulating tumor cells identiﬁed

in an ovarian cancer patient. Front Oncol,2:72.

[Phng et al., 2015] Phng, L. K., Gebala, V., Bentley, K., Philippides, A., Wacker,

A., Mathivet, T., Sauteur, L., Stanchi, F., Belting, H. G., Aﬀolter, M., and Ger-

hardt, H. (2015). Formin-mediated actin polymerization at endothelial junctions

is required for vessel lumen formation and stabilization. Dev Cell,32(1):123–32.

[Pokutta et al., 2008] Pokutta, S., Drees, F., Yamada, S., Nelson, W. J., and Weis,

130 W. I. (2008). Biochemical and structural analysis of alpha-catenin in cell-cell con-

tacts. Biochem Soc Trans,36(Pt2):141–7.

[Polakis, 2000] Polakis, P. (2000). Wnt signaling and cancer. Genes Dev,

14(15):1837–51.

[Pollard, 2016] Pollard, T. D. (2016). Actin and actin-binding proteins. Cold Spring

Harb Perspect Biol,8(8).

[Pollard et al., 2000] Pollard, T. D., Blanchoin, L., and Mullins, R. D. (2000). Molec-

ular mechanisms controlling actin ﬁlament dynamics in nonmuscle cells. Annu Rev

Biophys Biomol Struct,29:545–76.

[Pon et al., 2008] Pon, Y. L., Zhou, H. Y., Cheung, A. N., Ngan, H. Y., and Wong,

A. S. (2008). p70 s6 kinase promotes epithelial to mesenchymal transition through

snail induction in ovarian cancer cells. Cancer Res,68(16):6524–32.

[Porter et al., 2016] Porter, A. P., Papaioannou, A., and Malliri, A. (2016). Deregu-

lation of rho gtpases in cancer. Small GTPases,7(3):123–38.

[Pradeep et al., 2014] Pradeep, S., Kim, S. W., Wu, S. Y., Nishimura, M., Chaluvally-

Raghavan, P., Miyake, T., Pecot, C. V., Kim, S. J., Choi, H. J., Bischoﬀ, F. Z.,

Mayer, J. A., Huang, L., Nick, A. M., Hall, C. S., Rodriguez-Aguayo, C., Zand,

B., Dalton, H. J., Arumugam, T., Lee, H. J., Han, H. D., Cho, M. S., Rupaimoole,

R., Mangala, L. S., Sehgal, V., Oh, S. C., Liu, J., Lee, J. S., Coleman, R. L.,

Ram, P., Lopez-Berestein, G., Fidler, I. J., and Sood, A. K. (2014). Hematogenous

metastasis of ovarian cancer: rethinking mode of spread. Cancer Cell,26(1):77–91.

[Prakasam et al., 2006] Prakasam, A. K., Maruthamuthu, V., and Leckband, D. E.

(2006). Similarities between heterophilic and homophilic cadherin adhesion. Proc

Natl Acad Sci U S A,103(42):15434–9.

131 [Prat, 2012] Prat, J. (2012). Ovarian carcinomas: ﬁve distinct diseases with diﬀer-

ent origins, genetic alterations, and clinicopathological features. Virchows Arch,

460(3):237–49.

[Pruthi et al., 2010] Pruthi, S., Gostout, B. S., and Lindor, N. M. (2010). Identiﬁca-

tion and management of women with brca mutations or hereditary predisposition

for breast and ovarian cancer. Mayo Clin Proc,85(12):1111–20.

[Pruyne et al., 2002] Pruyne, D., Evangelista, M., Yang, C., Bi, E., Zigmond, S.,

Bretscher, A., and Boone, C. (2002). Role of formins in actin assembly: nucleation

and barbed-end association. Science,297(5581):612–5.

[Przybycin et al., 2010] Przybycin, C. G., Kurman, R. J., Ronnett, B. M., Shih Ie,

M., and Vang, R. (2010). Are all pelvic (nonuterine) serous carcinomas of tubal

origin? Am J Surg Pathol,34(10):1407–16.

[Pustilnik et al., 1999] Pustilnik, T. B., Estrella, V., Wiener, J. R., Mao, M., Eder,

A., Watt, M. A., Bast, R. C., J., and Mills, G. B. (1999). Lysophosphatidic acid

induces urokinase secretion by ovarian cancer cells. Clin Cancer Res,5(11):3704–10.

[Qi et al., 2014] Qi, H., Sun, B., Zhao, X., Du, J., Gu, Q., Liu, Y., Cheng, R.,

and Dong, X. (2014). Wnt5a promotes vasculogenic mimicry and epithelial-

mesenchymal transition via protein kinase calpha in epithelial ovarian cancer. Oncol

Rep,32(2):771–9.

[Qiu et al., 1995] Qiu, R. G., Chen, J., McCormick, F., and Symons, M. (1995). A

role for rho in ras transformation. Proc Natl Acad Sci U S A,92(25):11781–5.

[Raich et al., 1999] Raich, W. B., Agbunag, C., and Hardin, J. (1999). Rapid

epithelial-sheet sealing in the caenorhabditis elegans embryo requires cadherin-

dependent ﬁlopodial priming. Curr Biol,9(20):1139–46.

132 [Ramalingam et al., 2010] Ramalingam, N., Zhao, H., Breitsprecher, D., Lap-

palainen, P., Faix, J., and Schleicher, M. (2010). Phospholipids regulate localization

and activity of mdia1 formin. Eur J Cell Biol,89(10):723–32.

[Rana et al., 2018] Rana, M. K., Aloisio, F. M., Choi, C., and Barber, D. L. (2018).

Formin-dependent tgf-beta signaling for epithelial to mesenchymal transition. Mol

Biol Cell,29(12):1465–1475.

[Rao et al., 2014] Rao, J. N., Madasu, Y., and Dominguez, R. (2014). Mechanism of

actin ﬁlament pointed-end capping by tropomodulin. Science,345(6195):463–7.

[Rao and Zaidel-Bar, 2016] Rao, M. V. and Zaidel-Bar, R. (2016). Formin-mediated

actin polymerization at cell-cell junctions stabilizes e-cadherin and maintains mono-

layer integrity during wound repair. Mol Biol Cell,27(18):2844–56.

[Rathinam et al., 2011] Rathinam, R., Berrier, A., and Alahari, S. K. (2011). Role

of rho gtpases and their regulators in cancer progression. Front Biosci (Landmark

Ed),16:2561–71.

[Reid et al., 2017] Reid, B. M., Permuth, J. B., and Sellers, T. A. (2017). Epidemi-

ology of ovarian cancer: a review. Cancer Biol Med,14(1):9–32.

[Ridley, 2006] Ridley, A. J. (2006). Rho gtpases and actin dynamics in membrane

protrusions and vesicle traﬃcking. Trends Cell Biol,16(10):522–9.

[Ridley, 2011] Ridley, A. J. (2011). Life at the leading edge. Cell,145(7):1012–22.

[Ridley, 2013] Ridley, A. J. (2013). Rhoa, rhob and rhoc have diﬀerent roles in cancer

cell migration. J Microsc,251(3):242–9.

[Rivera et al., 2008] Rivera, F., Vega-Villegas, M. E., and Lopez-Brea, M. F. (2008).

Cetuximab, its clinical use and future perspectives. Anticancer Drugs,19(2):99–

113. 133 [Rizvi et al., 2013] Rizvi, I., Gurkan, U. A., Tasoglu, S., Alagic, N., Celli, J. P.,

Mensah, L. B., Mai, Z., Demirci, U., and Hasan, T. (2013). Flow induces epithelial-

mesenchymal transition, cellular heterogeneity and biomarker modulation in 3d

ovarian cancer nodules. Proc Natl Acad Sci U S A,110(22):E1974–83.

[Rizvi et al., 2009] Rizvi, S. A., Neidt, E. M., Cui, J., Feiger, Z., Skau, C. T., Gardel,

M. L., Kozmin, S. A., and Kovar, D. R. (2009). Identiﬁcation and characteriza-

tion of a small molecule inhibitor of formin-mediated actin assembly. Chem Biol,

16(11):1158–68.

[Rohatgi et al., 1999] Rohatgi, R., Ma, L., Miki, H., Lopez, M., Kirchhausen, T.,

Takenawa, T., and Kirschner, M. W. (1999). The interaction between n-wasp

and the arp2/3 complex links cdc42-dependent signals to actin assembly. Cell,

97(2):221–31.

[Rojas et al., 2016] Rojas, V., Hirshﬁeld, K. M., Ganesan, S., and Rodriguez-

Rodriguez, L. (2016). Molecular characterization of epithelial ovarian cancer: Im-

plications for diagnosis and treatment. Int J Mol Sci,17(12).

[Romero and Bast, 2012] Romero, I. and Bast, R. C., J. (2012). Minireview: human

ovarian cancer: biology, current management, and paths to personalizing therapy.

Endocrinology,153(4):1593–602.

[Rosano et al., 2005] Rosano, L., Spinella, F., Di Castro, V., Nicotra, M. R., Dedhar,

S., de Herreros, A. G., Natali, P. G., and Bagnato, A. (2005). Endothelin-1 pro-

motes epithelial-to-mesenchymal transition in human ovarian cancer cells. Cancer

Res,65(24):11649–57.

[Rosero et al., 2013] Rosero, A., Zarsky, V., and Cvrckova, F. (2013). Atfh1 formin

mutation aﬀects actin ﬁlament and microtubule dynamics in arabidopsis thaliana.

J Exp Bot,64(2):585–97. 134 [Roy and Cowden Dahl, 2018] Roy, L. and Cowden Dahl, K. D. (2018). Can stemness

and chemoresistance be therapeutically targeted via signaling pathways in ovarian

cancer? Cancers (Basel),10(8).

[Rubtsova et al., 1998] Rubtsova, S. N., Kondratov, R. V., Kopnin, P. B., Chumakov,

P. M., Kopnin, B. P., and Vasiliev, J. M. (1998). Disruption of actin microﬁlaments

by cytochalasin d leads to activation of p53. FEBS Lett,430(3):353–7.

[Sagot et al., 2002] Sagot, I., Rodal, A. A., Moseley, J., Goode, B. L., and Pellman,

D. (2002). An actin nucleation mechanism mediated by bni1 and proﬁlin. Nat Cell

Biol,4(8):626–31.

[Sahai and Marshall, 2002] Sahai, E. and Marshall, C. J. (2002). Rock and dia have

opposing eﬀects on adherens junctions downstream of rho. Nat Cell Biol,4(6):408–

15.

[Sakata et al., 2007] Sakata, D., Taniguchi, H., Yasuda, S., Adachi-Morishima, A.,

Hamazaki, Y., Nakayama, R., Miki, T., Minato, N., and Narumiya, S. (2007).

Impaired t lymphocyte traﬃcking in mice deﬁcient in an actin-nucleating protein,

mdia1. J Exp Med,204(9):2031–8.

[Sako et al., 1998] Sako, Y., Nagafuchi, A., Tsukita, S., Takeichi, M., and Kusumi,

A. (1998). Cytoplasmic regulation of the movement of e-cadherin on the free cell

surface as studied by optical tweezers and single particle tracking: corralling and

tethering by the membrane skeleton. J Cell Biol,140(5):1227–40.

[Sander et al., 1999] Sander, E. E., ten Klooster, J. P., van Delft, S., van der Kam-

men, R. A., and Collard, J. G. (1999). Rac downregulates rho activity: reciprocal

balance between both gtpases determines cellular morphology and migratory be-

havior. J Cell Biol,147(5):1009–22.

135 [Santisteban et al., 2009] Santisteban, M., Reiman, J. M., Asiedu, M. K., Behrens,

M. D., Nassar, A., Kalli, K. R., Haluska, P., Ingle, J. N., Hartmann, L. C., Manjili,

M. H., Radisky, D. C., Ferrone, S., and Knutson, K. L. (2009). Immune-induced

epithelial to mesenchymal transition in vivo generates breast cancer stem cells.

Cancer Res,69(7):2887–95.

[Sanvitale et al., 2013] Sanvitale, C. E., Kerr, G., Chaikuad, A., Ramel, M. C., Mo-

hedas, A. H., Reichert, S., Wang, Y., Triﬃtt, J. T., Cuny, G. D., Yu, P. B., Hill,

C. S., and Bullock, A. N. (2013). A new class of small molecule inhibitor of bmp

signaling. PLoS One,8(4):e62721.

[Sarmiento et al., 2008] Sarmiento, C., Wang, W., Dovas, A., Yamaguchi, H., Sidani,

M., El-Sibai, M., Desmarais, V., Holman, H. A., Kitchen, S., Backer, J. M., Alberts,

A., and Condeelis, J. (2008). Wasp family members and formin proteins coordinate

regulation of cell protrusions in carcinoma cells. J Cell Biol,180(6):1245–60.

[Satoh and Tominaga, 2001] Satoh, S. and Tominaga, T. (2001). mdia-interacting

protein acts downstream of rho-mdia and modiﬁes src activation and stress ﬁber

formation. J Biol Chem,276(42):39290–4.

[Sawada et al., 2008] Sawada, K., Mitra, A. K., Radjabi, A. R., Bhaskar, V., Kistner,

E. O., Tretiakova, M., Jagadeeswaran, S., Montag, A., Becker, A., Kenny, H. A.,

Peter, M. E., Ramakrishnan, V., Yamada, S. D., and Lengyel, E. (2008). Loss

of e-cadherin promotes ovarian cancer metastasis via alpha 5-integrin, which is a

therapeutic target. Cancer Res,68(7):2329–39.

[Schirenbeck et al., 2005] Schirenbeck, A., Bretschneider, T., Arasada, R., Schleicher,

M., and Faix, J. (2005). The diaphanous-related formin ddia2 is required for the

formation and maintenance of ﬁlopodia. Nat Cell Biol,7(6):619–25.

136 [Schmeler and Gershenson, 2008] Schmeler, K. M. and Gershenson, D. M. (2008).

Low-grade serous ovarian cancer: a unique disease. Curr Oncol Rep,10(6):519–23.

[Shagieva et al., 2012] Shagieva, G. S., Domnina, L. V., Chipysheva, T. A., Ermilova,

V. D., Chaponnier, C., and Dugina, V. B. (2012). Actin isoforms and reorganization

of adhesion junctions in epithelial-to-mesenchymal transition of cervical carcinoma

cells. Biochemistry (Mosc),77(11):1266–76.

[Shankar and Nabi, 2015] Shankar, J. and Nabi, I. R. (2015). Actin cytoskeleton

regulation of epithelial mesenchymal transition in metastatic cancer cells. PLoS

One,10(3):e0119954.

[Shibue and Weinberg, 2017] Shibue, T. and Weinberg, R. A. (2017). Emt, cscs, and

drug resistance: the mechanistic link and clinical implications. Nat Rev Clin Oncol,

14(10):611–629.

[Shimotsuma et al., 1993] Shimotsuma, M., Shields, J. W., Simpson-Morgan, M. W.,

Sakuyama, A., Shirasu, M., Hagiwara, A., and Takahashi, T. (1993). Morpho-

physiological function and role of omental milky spots as omentum-associated lym-

phoid tissue (oalt) in the peritoneal cavity. Lymphology,26(2):90–101.

[Shimoyama et al., 1992] Shimoyama, Y., Nagafuchi, A., Fujita, S., Gotoh, M., Take-

ichi, M., Tsukita, S., and Hirohashi, S. (1992). Cadherin dysfunction in a human

cancer cell line: possible involvement of loss of alpha-catenin expression in reduced

cell-cell adhesiveness. Cancer Res,52(20):5770–4.

[Shinohara et al., 2012] Shinohara, R., Thumkeo, D., Kamijo, H., Kaneko, N.,

Sawamoto, K., Watanabe, K., Takebayashi, H., Kiyonari, H., Ishizaki, T., Fu-

ruyashiki, T., and Narumiya, S. (2012). A role for mdia, a rho-regulated actin nu-

cleator, in tangential migration of interneuron precursors. Nat Neurosci,15(3):373–

80, S1–2. 137 [Shtutman et al., 2008] Shtutman, M., Chausovsky, A., Prager-Khoutorsky, M.,

Schiefermeier, N., Boguslavsky, S., Kam, Z., Fuchs, E., Geiger, B., Borisy, G. G.,

and Bershadsky, A. D. (2008). Signaling function of alpha-catenin in microtubule

regulation. Cell Cycle,7(15):2377–83.

[So et al., 2015] So, K. A., Min, K. J., Hong, J. H., and Lee, J. K. (2015). Interleukin-

6 expression by interactions between gynecologic cancer cells and human mes-

enchymal stem cells promotes epithelial-mesenchymal transition. Int J Oncol,

47(4):1451–9.

[Soong et al., 2018] Soong, T. R., Kolin, D. L., Teschan, N. J., and Crum, C. P.

(2018). Back to the future? the fallopian tube, precursor escape and a dualistic

model of high-grade serous carcinogenesis. Cancers (Basel),10(12).

[Soslow et al., 2012] Soslow, R. A., Han, G., Park, K. J., Garg, K., Olvera, N.,

Spriggs, D. R., Kauﬀ, N. D., and Levine, D. A. (2012). Morphologic patterns

associated with brca1 and brca2 genotype in ovarian carcinoma. Mod Pathol,

25(4):625–36.

[Spence et al., 2012] Spence, H. J., Timpson, P., Tang, H. R., Insall, R. H., and

Machesky, L. M. (2012). Scar/wave3 contributes to motility and plasticity of lamel-

lipodial dynamics but not invasion in three dimensions. Biochem J,448(1):35–42.

[Spinella et al., 2004] Spinella, F., Rosano, L., Elia, G., Di Castro, V., Natali, P. G.,

and Bagnato, A. (2004). Endothelin-1 stimulates cyclooxygenase-2 expression in

ovarian cancer cells through multiple signaling pathways: evidence for involvement

of transactivation of the epidermal growth factor receptor. J Cardiovasc Pharmacol,

44 Suppl 1:S140–3.

[Stehbens et al., 2006] Stehbens, S. J., Paterson, A. D., Crampton, M. S., Shewan,

A. M., Ferguson, C., Akhmanova, A., Parton, R. G., and Yap, A. S. (2006). Dy- 138 namic microtubules regulate the local concentration of e-cadherin at cell-cell con-

tacts. J Cell Sci,119(Pt9):1801–11.

[Sun et al., 2015] Sun, B. O., Fang, Y., Li, Z., Chen, Z., and Xiang, J. (2015). Role

of cellular cytoskeleton in epithelial-mesenchymal transition process during cancer

progression. Biomed Rep,3(5):603–610.

[Sundfeldt et al., 1997] Sundfeldt, K., Piontkewitz, Y., Ivarsson, K., Nilsson, O.,

Hellberg, P., Brannstrom, M., Janson, P. O., Enerback, S., and Hedin, L. (1997).

E-cadherin expression in human epithelial ovarian cancer and normal ovary. Int J

Cancer,74(3):275–80.

[Svitkina et al., 2003] Svitkina, T. M., Bulanova, E. A., Chaga, O. Y., Vignjevic,

D. M., Kojima, S., Vasiliev, J. M., and Borisy, G. G. (2003). Mechanism of ﬁlopodia

initiation by reorganization of a dendritic network. J Cell Biol,160(3):409–21.

[Szabova et al., 2012] Szabova, L., Yin, C., Bupp, S., Guerin, T. M., Schlomer, J. J.,

Householder, D. B., Baran, M. L., Yi, M., Song, Y., Sun, W., McDunn, J. E.,

Martin, P. L., Van Dyke, T., and Diﬁlippantonio, S. (2012). Perturbation of rb,

p53, and brca1 or brca2 cooperate in inducing metastatic serous epithelial ovarian

cancer. Cancer Res,72(16):4141–53.

[Takeya et al., 2008] Takeya, R., Taniguchi, K., Narumiya, S., and Sumimoto, H.

(2008). The mammalian formin fhod1 is activated through phosphorylation by rock

and mediates thrombin-induced stress ﬁbre formation in endothelial cells. EMBO

J,27(4):618–28.

[Tanizaki et al., 2010] Tanizaki, H., Egawa, G., Inaba, K., Honda, T., Nakajima, S.,

Moniaga, C. S., Otsuka, A., Ishizaki, T., Tomura, M., Watanabe, T., Miyachi,

Y., Narumiya, S., Okada, T., and Kabashima, K. (2010). Rho-mdia1 pathway is

139 required for adhesion, migration, and t-cell stimulation in dendritic cells. Blood,

116(26):5875–84.

[Theriault et al., 2007] Theriault, B. L., Shepherd, T. G., Mujoomdar, M. L., and

Nachtigal, M. W. (2007). Bmp4 induces emt and rho gtpase activation in human

ovarian cancer cells. Carcinogenesis,28(6):1153–62.

[Thiery et al., 2009] Thiery, J. P., Acloque, H., Huang, R. Y., and Nieto, M. A.

(2009). Epithelial-mesenchymal transitions in development and disease. Cell,

139(5):871–90.

[Thumkeo et al., 2013] Thumkeo, D., Watanabe, S., and Narumiya, S. (2013). Phys-

iological roles of rho and rho eﬀectors in mammals. Eur J Cell Biol,92(10-11):303–

15.

[Tilbury and Campagnola, 2015] Tilbury, K. and Campagnola, P. J. (2015). Appli-

cations of second-harmonic generation imaging microscopy in ovarian and breast

cancer. Perspect Medicin Chem,7:21–32.

[Tsang et al., 2013] Tsang, Y. T., Deavers, M. T., Sun, C. C., Kwan, S. Y., Kuo,

E., Malpica, A., Mok, S. C., Gershenson, D. M., and Wong, K. K. (2013). Kras

(but not braf) mutations in ovarian serous borderline tumour are associated with

recurrent low-grade serous carcinoma. J Pathol,231(4):449–56.

[Vasey et al., 2004] Vasey, P. A., Jayson, G. C., Gordon, A., Gabra, H., Coleman, R.,

Atkinson, R., Parkin, D., Paul, J., Hay, A., Kaye, S. B., and Scottish Gynaecolog-

ical Cancer Trials, G. (2004). Phase iii randomized trial of docetaxel-carboplatin

versus paclitaxel-carboplatin as ﬁrst-line chemotherapy for ovarian carcinoma. J

Natl Cancer Inst,96(22):1682–91.

[Vasioukhin et al., 2000] Vasioukhin, V., Bauer, C., Yin, M., and Fuchs, E. (2000).

140 Directed actin polymerization is the driving force for epithelial cell-cell adhesion.

Cell,100(2):209–19.

[Vega et al., 2012] Vega, F. M., Colomba, A., Reymond, N., Thomas, M., and Ri-

dley, A. J. (2012). Rhob regulates cell migration through altered focal adhesion

dynamics. Open Biol,2(5):120076.

[Vega et al., 2011] Vega, F. M., Fruhwirth, G., Ng, T., and Ridley, A. J. (2011). Rhoa

and rhoc have distinct roles in migration and invasion by acting through diﬀerent

targets. J Cell Biol,193(4):655–65.

[Vergara et al., 2010] Vergara, D., Merlot, B., Lucot, J. P., Collinet, P., Vinatier, D.,

Fournier, I., and Salzet, M. (2010). Epithelial-mesenchymal transition in ovarian

cancer. Cancer Lett,291(1):59–66.

[Verma et al., 2004] Verma, S., Shewan, A. M., Scott, J. A., Helwani, F. M., den

Elzen, N. R., Miki, H., Takenawa, T., and Yap, A. S. (2004). Arp2/3 activity

is necessary for eﬃcient formation of e-cadherin adhesive contacts. J Biol Chem,

279(32):34062–70.

[Vicente-Manzanares et al., 2003] Vicente-Manzanares, M., Rey, M., Perez-Martinez,

M., Yanez-Mo, M., Sancho, D., Cabrero, J. R., Barreiro, O., de la Fuente, H., Itoh,

K., and Sanchez-Madrid, F. (2003). The rhoa eﬀector mdia is induced during t cell

activation and regulates actin polymerization and cell migration in t lymphocytes.

J Immunol,171(2):1023–34.

[Wakatsuki et al., 2000] Wakatsuki, A., Okatani, Y., and Ikenoue, N. (2000). Eﬀects

of combination therapy with estrogen plus simvastatin on lipoprotein metabolism

in postmenopausal women with type iia hypercholesterolemia. Atherosclerosis,

150(1):103–11.

141 [Wallar and Alberts, 2003] Wallar, B. J. and Alberts, A. S. (2003). The formins:

active scaﬀolds that remodel the cytoskeleton. Trends Cell Biol,13(8):435–46.

[Wallar et al., 2007] Wallar, B. J., Deward, A. D., Resau, J. H., and Alberts, A. S.

(2007). Rhob and the mammalian diaphanous-related formin mdia2 in endosome

traﬃcking. Exp Cell Res,313(3):560–71.

[Wallar et al., 2006] Wallar, B. J., Stropich, B. N., Schoenherr, J. A., Holman, H. A.,

Kitchen, S. M., and Alberts, A. S. (2006). The basic region of the diaphanous-

autoregulatory domain (dad) is required for autoregulatory interactions with the

diaphanous-related formin inhibitory domain. J Biol Chem,281(7):4300–7.

[Warner and Longmore, 2009] Warner, S. J. and Longmore, G. D. (2009). Distinct

functions for rho1 in maintaining adherens junctions and apical tension in remod-

eling epithelia. J Cell Biol,185(6):1111–25.

[Watanabe et al., 1999] Watanabe, N., Kato, T., Fujita, A., Ishizaki, T., and Naru-

miya, S. (1999). Cooperation between mdia1 and rock in rho-induced actin reor-

ganization. Nat Cell Biol,1(3):136–43.

[Watanabe et al., 2008] Watanabe, S., Ando, Y., Yasuda, S., Hosoya, H., Watanabe,

N., Ishizaki, T., and Narumiya, S. (2008). mdia2 induces the actin scaﬀold for the

contractile ring and stabilizes its position during cytokinesis in nih 3t3 cells. Mol

Biol Cell,19(5):2328–38.

[Waterman-Storer et al., 2000] Waterman-Storer, C. M., Salmon, W. C., and

Salmon, E. D. (2000). Feedback interactions between cell-cell adherens junctions

and cytoskeletal dynamics in newt lung epithelial cells. Mol Biol Cell,11(7):2471–

83.

[Welch and Mullins, 2002] Welch, M. D. and Mullins, R. D. (2002). Cellular control

of actin nucleation. Annu Rev Cell Dev Biol,18:247–88. 142 [Wen et al., 2004] Wen, Y., Eng, C. H., Schmoranzer, J., Cabrera-Poch, N., Morris,

E. J., Chen, M., Wallar, B. J., Alberts, A. S., and Gundersen, G. G. (2004). Eb1

and apc bind to mdia to stabilize microtubules downstream of rho and promote

cell migration. Nat Cell Biol,6(9):820–30.

[Wentzensen et al., 2016] Wentzensen, N., Poole, E. M., Trabert, B., White, E., Ar-

slan, A. A., Patel, A. V., Setiawan, V. W., Visvanathan, K., Weiderpass, E., Adami,

H. O., Black, A., Bernstein, L., Brinton, L. A., Buring, J., Butler, L. M., Chamosa,

S., Clendenen, T. V., Dossus, L., Fortner, R., Gapstur, S. M., Gaudet, M. M.,

Gram, I. T., Hartge, P., Hoﬀman-Bolton, J., Idahl, A., Jones, M., Kaaks, R.,

Kirsh, V., Koh, W. P., Lacey, J. V., J., Lee, I. M., Lundin, E., Merritt, M. A.,

Onland-Moret, N. C., Peters, U., Poynter, J. N., Rinaldi, S., Robien, K., Rohan,

T., Sandler, D. P., Schairer, C., Schouten, L. J., Sjoholm, L. K., Sieri, S., Swerdlow,

A., Tjonneland, A., Travis, R., Trichopoulou, A., van den Brandt, P. A., Wilkens,

L., Wolk, A., Yang, H. P., Zeleniuch-Jacquotte, A., and Tworoger, S. S. (2016).

Ovarian cancer risk factors by histologic subtype: An analysis from the ovarian

cancer cohort consortium. J Clin Oncol,34(24):2888–98.

[Whittemore et al., 1988] Whittemore, A. S., Wu, M. L., Paﬀenbarger, R. S., J., Sar-

les, D. L., Kampert, J. B., Grosser, S., Jung, D. L., Ballon, S., and Hendrickson,

M. (1988). Personal and environmental characteristics related to epithelial ovar-

ian cancer. ii. exposures to talcum powder, tobacco, alcohol, and coﬀee. Am J

Epidemiol,128(6):1228–40.

[Witz et al., 2001] Witz, C. A., Montoya-Rodriguez, I. A., Cho, S., Centonze, V. E.,

Bonewald, L. F., and Schenken, R. S. (2001). Composition of the extracellular

matrix of the peritoneum. J Soc Gynecol Investig,8(5):299–304.

[Wolf et al., 2003] Wolf, K., Mazo, I., Leung, H., Engelke, K., von Andrian, U. H.,

Deryugina, E. I., Strongin, A. Y., Brocker, E. B., and Friedl, P. (2003). Compen- 143 sation mechanism in tumor cell migration: mesenchymal-amoeboid transition after

blocking of pericellular proteolysis. J Cell Biol,160(2):267–77.

[Wong et al., 2010] Wong, K. K., Tsang, Y. T., Deavers, M. T., Mok, S. C., Zu, Z.,

Sun, C., Malpica, A., Wolf, J. K., Lu, K. H., and Gershenson, D. M. (2010). Braf

mutation is rare in advanced-stage low-grade ovarian serous carcinomas. Am J

Pathol,177(4):1611–7.

[Wu et al., 2008] Wu, C., Cipollone, J., Maines-Bandiera, S., Tan, C., Karsan, A.,

Auersperg, N., and Roskelley, C. D. (2008). The morphogenic function of e-

cadherin-mediated adherens junctions in epithelial ovarian carcinoma formation

and progression. Diﬀerentiation,76(2):193–205.

[Wu et al., 1988] Wu, M. L., Whittemore, A. S., Paﬀenbarger, R. S., J., Sarles, D. L.,

Kampert, J. B., Grosser, S., Jung, D. L., Ballon, S., Hendrickson, M., and Mohle-

Boetani, J. (1988). Personal and environmental characteristics related to epithelial

ovarian cancer. i. reproductive and menstrual events and oral contraceptive use.

Am J Epidemiol,128(6):1216–27.

[Wu et al., 2015] Wu, Y., Kanchanawong, P., and Zaidel-Bar, R. (2015). Actin-

delimited adhesion-independent clustering of e-cadherin forms the nanoscale build-

ing blocks of adherens junctions. Dev Cell,32(2):139–54.

[Wynshaw-Boris et al., 1997] Wynshaw-Boris, A., Ryan, G., Deng, C. X., Chan,

D. C., Jackson-Grusby, L., Larson, D., Dunmore, J. H., and Leder, P. (1997). The

role of a single formin isoform in the limb and renal phenotypes of limb deformity.

Mol Med,3(6):372–84.

[Wyse et al., 2012] Wyse, M. M., Lei, J., Nestor-Kalinoski, A. L., and Eisenmann,

K. M. (2012). Dia-interacting protein (dip) imposes migratory plasticity in mdia2-

dependent tumor cells in three-dimensional matrices. PLoS One,7(9):e45085. 144 [Xiao et al., 2005] Xiao, K., Garner, J., Buckley, K. M., Vincent, P. A., Chiasson,

C. M., Dejana, E., Faundez, V., and Kowalczyk, A. P. (2005). p120-catenin regu-

lates clathrin-dependent endocytosis of ve-cadherin. Mol Biol Cell,16(11):5141–51.

[Xing et al., 2007] Xing, H., Wang, S., Weng, D., Chen, G., Yang, X., Zhou, J.,

Xu, G., Lu, Y., and Ma, D. (2007). Knock-down of p-glycoprotein reverses taxol

resistance in ovarian cancer multicellular spheroids. Oncol Rep,17(1):117–22.

[Yamada and Nelson, 2007] Yamada, S. and Nelson, W. J. (2007). Localized zones of

rho and rac activities drive initiation and expansion of epithelial cell-cell adhesion.

J Cell Biol,178(3):517–27.

[Yamada et al., 2005] Yamada, S., Pokutta, S., Drees, F., Weis, W. I., and Nelson,

W. J. (2005). Deconstructing the cadherin-catenin-actin complex. Cell,123(5):889–

901.

[Yamaguchi, 2012] Yamaguchi, H. (2012). Pathological roles of invadopodia in cancer

invasion and metastasis. Eur J Cell Biol,91(11-12):902–7.

[Yamana et al., 2006] Yamana, N., Arakawa, Y., Nishino, T., Kurokawa, K., Tanji,

M., Itoh, R. E., Monypenny, J., Ishizaki, T., Bito, H., Nozaki, K., Hashimoto,

N., Matsuda, M., and Narumiya, S. (2006). The rho-mdia1 pathway regulates cell

polarity and focal adhesion turnover in migrating cells through mobilizing apc and

c-src. Mol Cell Biol,26(18):6844–58.

[Yamazaki et al., 2007] Yamazaki, D., Oikawa, T., and Takenawa, T. (2007). Rac-

wave-mediated actin reorganization is required for organization and maintenance

of cell-cell adhesion. J Cell Sci,120(Pt1):86–100.

[Yanagisawa et al., 2004] Yanagisawa, M., Kaverina, I. N., Wang, A., Fujita, Y.,

Reynolds, A. B., and Anastasiadis, P. Z. (2004). A novel interaction between kinesin

and p120 modulates p120 localization and function. J Biol Chem,279(10):9512–21. 145 [Yang et al., 2007] Yang, C., Czech, L., Gerboth, S., Kojima, S., Scita, G., and Svitk-

ina, T. (2007). Novel roles of formin mdia2 in lamellipodia and ﬁlopodia formation

in motile cells. PLoS Biol,5(11):e317.

[Yang and Svitkina, 2011] Yang, C. and Svitkina, T. (2011). Filopodia initiation:

focus on the arp2/3 complex and formins. Cell Adh Migr,5(5):402–8.

[Yap and Kovacs, 2003] Yap, A. S. and Kovacs, E. M. (2003). Direct cadherin-

activated cell signaling: a view from the plasma membrane. J Cell Biol,160(1):11–

[Yasuda et al., 2004] Yasuda, S., Oceguera-Yanez, F., Kato, T., Okamoto, M., Yone-

mura, S., Terada, Y., Ishizaki, T., and Narumiya, S. (2004). Cdc42 and mdia3

regulate microtubule attachment to kinetochores. Nature,428(6984):767–71.

[Yeung et al., 2015] Yeung, T. L., Leung, C. S., Yip, K. P., Au Yeung, C. L., Wong,

S. T., and Mok, S. C. (2015). Cellular and molecular processes in ovarian cancer

metastasis. a review in the theme: Cell and molecular processes in cancer metas-

tasis. Am J Physiol Cell Physiol,309(7):C444–56.

[Yingling et al., 2004] Yingling, J. M., Blanchard, K. L., and Sawyer, J. S. (2004).

Development of tgf-beta signalling inhibitors for cancer therapy. Nat Rev Drug

Discov,3(12):1011–22.

[Yonemura, 2011a] Yonemura, S. (2011a). Cadherin-actin interactions at adherens

junctions. Curr Opin Cell Biol,23(5):515–22.

[Yonemura, 2011b] Yonemura, S. (2011b). A mechanism of mechanotransduction at

the cell-cell interface: emergence of alpha-catenin as the center of a force-balancing

mechanism for morphogenesis in multicellular organisms. Bioessays,33(10):732–6.

146 [Yonemura, 2017] Yonemura, S. (2017). Actin ﬁlament association at adherens junc-

tions. J Med Invest,64(1.2):14–19.

[Yonemura et al., 1995] Yonemura, S., Itoh, M., Nagafuchi, A., and Tsukita, S.

(1995). Cell-to-cell adherens junction formation and actin ﬁlament organization:

similarities and diﬀerences between non-polarized ﬁbroblasts and polarized epithe-

lial cells. J Cell Sci,108(Pt1):127–42.

[Yonemura et al., 2010] Yonemura, S., Wada, Y., Watanabe, T., Nagafuchi, A., and

Shibata, M. (2010). alpha-catenin as a tension transducer that induces adherens

junction development. Nat Cell Biol,12(6):533–42.

[Yu et al., 2008] Yu, X., Liu, L., Cai, B., He, Y., and Wan, X. (2008). Suppression

of anoikis by the neurotrophic receptor trkb in human ovarian cancer. Cancer Sci,

99(3):543–52.

[Zalcman et al., 1995] Zalcman, G., Closson, V., Linares-Cruz, G., Lerebours, F.,

Honore, N., Tavitian, A., and Olofsson, B. (1995). Regulation of ras-related rhob

protein expression during the cell cycle. Oncogene,10(10):1935–45.

[Zhang et al., 2014] Zhang, D., Yang, R., Wang, S., and Dong, Z. (2014). Paclitaxel:

new uses for an old drug. Drug Des Devel Ther,8:279–84.

[Zhang et al., 2005] Zhang, J., Betson, M., Erasmus, J., Zeikos, K., Bailly, M.,

Cramer, L. P., and Braga, V. M. (2005). Actin at cell-cell junctions is composed

of two dynamic and functional populations. J Cell Sci,118(Pt23):5549–62.

[Zhang et al., 1999] Zhang, J., Shehabeldin, A., da Cruz, L. A., Butler, J., Somani,

A. K., McGavin, M., Kozieradzki, I., dos Santos, A. O., Nagy, A., Grinstein,

S., Penninger, J. M., and Siminovitch, K. A. (1999). Antigen receptor-induced

activation and cytoskeletal rearrangement are impaired in wiskott-aldrich syndrome

protein-deﬁcient lymphocytes. J Exp Med,190(9):1329–42. 147 [Zhou et al., 2017] Zhou, B., Xu, H., Xia, M., Sun, C., Li, N., Guo, E., Guo, L.,

Shan, W., Lu, H., Wu, Y., Li, Y., Yang, D., Weng, D., Meng, L., Hu, J., Ma, D.,

Chen, G., and Li, K. (2017). Overexpressed mir-9 promotes tumor metastasis via

targeting e-cadherin in serous ovarian cancer. Front Med,11(2):214–222.

[Zhou et al., 2009] Zhou, F., Leder, P., Zuniga, A., and Dettenhofer, M. (2009).

Formin1 disruption confers oligodactylism and alters bmp signaling. Hum Mol

Genet,18(13):2472–82.

[Ziske et al., 2016] Ziske, M. A., Pettee, K. M., Khaing, M., Rubinic, K., and Eisen-

mann, K. M. (2016). Smifh2-mediated mdia formin functional inhibition poten-

tiates chemotherapeutic targeting of human ovarian cancer spheroids. Biochem

Biophys Res Commun,472(1):33–9.

148 Appendix A

Investigation of the respiratory diaphragm as a key site of EOC invasion and metastasis

A.1 Introduction

In addition to promoting EOC dissemination within the peritoneal cavity, ascites is known to contribute to increased intra-abdominal pressure (IAP) in ovarian cancer patients. This increased IAP causes increased passive tension and elastance of the respiratory diaphragm in animal models (De Troyer, Leduc et al. 2009), along with remodeling of extracellular matrix (ECM). This, combined with the fact that EOC clusters mainly adhere to the submesothelial ECM beneath the mesothelial cell layer lining the abdominal cavity and internal organs, rather than the mesothelial cells themselves, suggests an important role for diaphragm stretch and ECM remodeling in disease progression.

More speciﬁcally, EOC adhesion to the submesothelial matrix depends on β1- integrin association with collagen, ﬁbronectin, and laminin (Casey, Burleson et al.

2001). Indeed, in both breast and ovarian cancer speciﬁc patterns of collagen align-

149 ment with respect to the tumor are associated with increased invasiveness (Adur,

Pelegati et al. 2012, Tilbury and Campagnola 2015). With respect to the diaphragm, additional structural properties, such as the patency of the lymphatic stomata (open- ings to the submesothelial lymphatic lacunae) are also altered by stretch (Abu-Hijleh,

Habbal et al. 1995). Thus over time, increased IAP from ascites is predicted to promote EOC invasion through ECM remodeling as well as increasing the patency of lymphatic stomata. In this section, we focus on the eﬀects of diaphragm stretch upon collagen alignment and its implications for EOC invasion.

A.2 Methods: Development of an ex vivo diaphragm

stretch model

EOC spheroids were generated from SKOV3 cells (an ovarian cystadenocarcinoma cell line derived from patient ascites) which express GFP (GFP-SKOV3) using the spheroid generation method described by Pettee et al., 2014. After isolation of two diaphragms from wild type BALB/c female mice and staining the tissue for cellular viability using Draq7, half of one diaphragm (left side) was subjected to 15% constant applied strain in the anterior-posterior direction using a custom-designed apparatus

O for one hour at 37 Cand5%CO2 while the other diaphragm was kept unstretched but pinned to maintain its original length after dissection. The explants were kept

O in RPMI media with 10% FBS/PBS at 37 C and 5% CO2 until 15-16 hours later, when spheroids (each consisting of about 4000 cells) were implanted on them and allowed to adhere for 1 hour before media was added. Images were acquired every 24 hours for over 72 hours with a Leica TCS SP5 multiphoton laser scanning confocal microscope with second harmonic generation (SHG), 10x objective.

The spheroids on both explants were measured using Metamorph software on the x-y axis projection along the length and width, which were then averaged for an

150 average diameter. The area was taken by calculating the area of a circle using the average diameter. Each of the areas at the 3 time points were normalized to the area measured at 24 hours.

A.3 Results

To determine if diaphragm stretch aﬀected EOC invasion and spread, we seeded

GFP-SKOV3 spheroids on unstretched mouse diaphragms and diaphragms stretched to 1.15 times unstretched length. At 24 hours, we observed a clearing of collagen un- derneath the spheroids for both conditions (Fig. A-1D, A-2D). Interestingly, collagen bundles in the stretched diaphragm appeared more ordered and aligned in one axis compared to the unstretched.

Figure A-1: Spheroid invasion at 48 hours on stretched diaphragm explant. A. GFP-SKOV3 spheroid of 4000 cells (green). Lines were drawn to measure the length of the longest and shortest axes. B. Draq7 staining of SKOV3 spheroids (red). C. Transmitted light of SKOV3 spheroids (white) D. Second harmonic generation (SHG) imaging of SKOV3 spheroids (gray) E. Overlay of channels shown in A-D.

Areas of EOC spheroids as determined by average diameter (of the longest and shortest diameters across on x-y projections) were followed over 72 hours. For each

151 Figure A-2: Spheroid invasion at 48 hours on unstretched diaphragm explant. A. GFP-SKOV3 spheroid of 4000 cells (green).Lines were drawn to measure the length of the longest and shortest axes. B. Draq7 staining of SKOV3 spheroids (red). C. Transmitted light of SKOV3 spheroids (white) D. Second harmonic generation (SHG) imaging of SKOV3 spheroids (gray) E. Overlay of channels shown in A-D.

condition, size was normalized to area for that condition at 24 hours after seeding

(see Fig. A-3). For both conditions, spheroid size increased over time and there were no signiﬁcant diﬀerences in area of spheroid spread at each time point according to a paired t-test (p-value = 0.236).

We also noted single cells migrating in the direction of collagen alignment away from the spheroid core, which was more pronounced for the stretched diaphragm

(Fig. A-4). Additionally, on the stretched diaphragms, GFP-SKOV3 cells detached from spheroids also look more elongated compared to the more rounded GFP-SKOV3 cells on the unstretched diaphragm. Although depths of spheroid invasion were also measured, it was hard to compare measurements due to the inherent changes in the diaphragm thickness as seen in the xz projections. Such measurements are difficult because the images of the diaphragm taken at different angles show different ruffles of the diaphragm. We did observe that over time, the spheroid implanted on top of the stretched diaphragm appeared more flattened at the center compared to the

152 Figure A-3: Normalized areas of GFP-SKOV3 invasion on stretched and unstretched diaphragms. There is no signiﬁcant diﬀerence between the two at 24, 48, and 72 hours.

Figure A-4: GFP-SKOV3 spheroids demonstrate collagen alignment and egress from the spheroid as single cells and clusters at 24 hours post-implantation. A: Stretched diaphragm image taken using second harmonic generation at 5.49x zoom. Scale bar = 100 µm. B: Unstretched diaphragm image taken at 11.25x zoom. Scale bar = 25 µm.

spheroid on the unstretched diaphragm, which maintained a domed conﬁguration

(Fig. A-5A-B).

Limitations of these experiments include the fact that as the diaphragm may be stretched not only by ascites but also by respirations, a more physiological model must take into account constant baseline variations in strain. Also for statistical signiﬁcance, the experiment would have to be performed for at least 3 times with 3 diaphragm halves for stretched and for unstretched conditions. Nevertheless, these results provide preliminary evidence that suggests mechanical strain promotes colla-

153 gen remodeling of the diaphragm, which may contribute to increased EOC invasion.

Figure A-5: X-Z projections of GFP-SKOV3 spheroids seeded on stretched and unstretched diaphragm explants. A. Heights of GFP- SKOV3 spheroids (green) seeded on stretched diaphragms are measured at 24 (i), 48 (ii), and 72 hours (iii) post-seeding. B. Heights of GFP-SKOV3 spheroids seeded on unstretched diaphragms are measured at 24 (i), 48 (ii), and 72 hours (iii) post-seeding. The underlying collagen of each spheroid is imaged by second harmonic generation (SHG) as shown in the bottom channel.

154