Role of Annexin A2 in Ovarian Cancer Metastasis

A Thesis Submitted for the Degree of Doctor of Philosophy by

Noor Alia Lokman

Robinson Institute, Research Centre for Reproductive Health, Discipline of Obstetrics and Gynaecology, School of Paedriatrics and Reproductive Health, Faculty of Health Sciences, The University of Adelaide, Australia

February 2014

Mak and Ayah, this is for you, I love you so much.

This thesis is dedicated to my loving parents, Norma Muhammad and Lokman Abdul Hamid.

Table of Contents List of Figures ...... 8

List of Supplementary Figures ...... 10

List of Tables ...... 11

List of Supplementary Tables ...... 12

Summary ...... 13

Declaration...... 15

Acknowledgements ...... 16

Publications Arising During Ph D Candidature ...... 17

Publications Contributing to This Thesis ...... 18

Presentations at Scientific Meeting ...... 19

Awards Arising Out of This Thesis ...... 22

Abbreviations ...... 23

1 Chapter 1: The Role of Annexin A2 in Tumourigenesis and Cancer Progression ...... 25

1.1 Abstract ...... 28

1.2 Introduction ...... 29

1.3 The structure and function of annexin A2 ...... 29

1.4 Expression of annexin A2 in cancer ...... 30

1.5 Role of annexin A2 and its binding proteins in cancer ...... 31

1.6 Role of annexin A2 in the plasminogen activator system ...... 32

1.7 Role of annexin A2 in cancer cell adhesion, invasion, and migration ...... 33

1.8 Post translational modifications and proteolytic cleavage of annexin A2 in cancer cells …………………………………………………………………………………………………34

1.9 Annexin A2 and ovarian cancer ...... 35

1.10 Annexin A2 as a potential therapeutic target in cancer ...... 36

1.11 Conclusions and future directions ...... 37

1.12 Acknowledgement...... 37

1.13 Statement of hypotheses, aims and research plan ...... 44

1.13.1 Hypotheses ...... 44

1.13.2 Aims ...... 44

1.13.3 Research plan ...... 45

2 Chapter 2: Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis ...... 46

2.1 Abstract ...... 49

2.2 Introduction ...... 50

2.3 Materials and methods ...... 51

2.3.1 Cell culture ...... 51

2.3.2 Cell motility and invasion assays ...... 51

2.3.3 Chick chorioallantoic membrane assay (CAM Assay) ...... 51

2.3.4 Immunohistochemistry ...... 52

2.3.5 Statistical analysis...... 52

2.4 Results ...... 53

2.4.1 Human ovarian cancer cell motility and invasion in vitro ...... 53

2.4.2 Human ovarian cancer cell invasion into the chick chorioallantoic membrane (CAM) …………………………………………………………………………………………...53

2.4.3 The effects of protein A neutralizing antibody on OV-90 cancer cell invasion into the CAM……………………………………………………………..……………………………..54

2.5 Discussion ...... 60

2.6 Conclusions ...... 63

2.7 Acknowledgments ...... 63

3 Chapter 3: Annexin A2 is regulated by Ovarian Cancer-Peritoneal Cell Interactions and Promotes Metastasis ...... 64

3.1 Abstract ...... 69

3.2 Introduction ...... 70

3.3 Materials and methods ...... 72

3.3.1 Patient’s tissue samples ...... 72

3.3.2 Cell culture ...... 72

3.3.3 Real-time PCR ...... 73

3.3.4 Immunohistochemistry ...... 73

3.3.5 Cytokeratin immunocytochemistry ...... 73

3.3.6 Co-culture of ovarian cancer and peritoneal cells ...... 74

3.3.7 1D-western immunoblotting ...... 74

3.3.8 2D-western immunoblotting ...... 74

3.3.9 Motility and invasion assays ...... 75

3.3.10 Adhesion assay ...... 75

3.3.11 Chick chorioallantoic membrane assay ...... 75

3.3.12 Intraperitoneal SKOV-3 xenograft mouse model...... 76

3.3.13 Immunohistochemical assessment ...... 76

3.3.14 Statistical analysis...... 77

3.4 Results ...... 78

3.4.1 Expression of annexin A2 in human ovarian cancer tissues and peritoneal cells …………………………………………………………………………………………...78

3.4.2 Morphology of primary ovarian cancer cells derived from ascites ...... 78

3.4.3 Annexin A2 expression in primary ovarian cancer cells ...... 79

3.4.4 Expression of annexin A2 in human ovarian cancer and peritoneal cell lines .... 79

3.4.5 Annexin A2 is cleaved in ovarian cancer and peritoneal cells co-culture ...... 79

3.4.6 Annexin A2 is cleaved by protease plasmin ...... 80

3.4.7 Annexin A2 promotes ovarian cancer motility and invasion and aids peritoneal adhesion of ovarian cancer cells ...... 80

3.4.8 Annexin A2 promotes ovarian cancer cell invasion in the chick chorioallantoic membrane model ...... 81

3.4.9 Annexin A2 promotes primary ovarian cancer cell invasion in the CAM model .. 81

3.4.10 Annexin A2 promotes ovarian cancer growth and metastasis ...... 81

3.5 Discussion ...... 96

3.6 Acknowledgements ...... 98

4 Chapter 4: Annexin A2 as a Diagnostic Marker for Serous Ovarian Cancer ...... 107

4.1 Abstract ...... 108

4.2 Introduction ...... 109

4.3 Materials and methods ...... 111

4.3.1 Urine, serum and plasma samples ...... 111

4.3.2 Urine sample preparation for western blot and annexin A2 ELISA ...... 111

4.3.3 1D-western immunoblotting ...... 111

4.3.4 Annexin A2 ELISA and detection of CA125 ...... 112

4.3.5 Statistical analyses ...... 112

4.4 Results ...... 113

4.4.1 Level of annexin A2 in the urine of ovarian cancer patients ...... 113

4.4.2 Blood annexin A2 levels are elevated in serous ovarian cancer patients compared with non-cancer controls ...... 113

4.4.3 CA125 levels in serous ovarian cancer patients ...... 113

4.4.4 Annexin A2 levels are weakly correlated with CA125 and able to distinguish serous ovarian cancer patients with normal CA125 levels from non-cancer controls ..... 114

4.4.5 Annexin A2 and CA125 diagnostic performance in ovarian cancer patients compared with non-cancer controls using specified cut points ...... 114

4.4.6 Annexin A2 and CA125 diagnostic performance in serous ovarian cancers compared with benign ovarian tumour using specified cutoff points ...... 115

4.4.7 Logistic regression model to evaluate the efficacy of combining annexin A2 and CA125 measurements in ovarian cancer patients and non-cancer controls ...... 115

4.5 Discussion ...... 128

4.6 Acknowledgements ...... 130

5 Chapter 5: Prognostic Significance of Annexin A2 and S100A10 in Serous Ovarian Cancer ………………………………………………………………………………………………….…133

5.1 Abstract ...... 134

5.2 Introduction ...... 135

5.3 Materials and methods ...... 137

5.3.1 Patient’s tissue microarray cohort ...... 137

5.3.2 Immunohistochemistry ...... 137

5.3.3 Immunohistochemical assessment ...... 137

5.3.4 Statistical analyses ...... 137

5.4 Results ...... 139

5.4.1 Increased stromal annexin A2 and S100A10 expression in metastatic ovarian cancer tissues compared with matching primary ovarian tumours ...... 139

5.4.2 High stromal annexin A2 expression in the cancer associated stroma predicts clinical outcome ...... 139

5.4.3 S100A10 cytoplasmic positivity is associated with reduced survival ...... 139

5.4.4 Stromal annexin A2 positivity and cytoplasmic S100A10 expression is an independent predictor of clinical outcome ...... 140

5.5 Discussion ...... 152

5.6 Acknowledgements ...... 154

6 Chapter 6: General Discussion and Conclusions ...... 156

6.1 General discussion ...... 157

6.1.1 Annexin A2 plays an important role in ovarian cancer invasion and metastasis ………………………………………………………………………………………….158 6

6.1.2 Role of annexin A2 in tumour microenvironment ...... 159

6.1.3 Clinical significance of annexin A2 in serous ovarian cancer patients ...... 161

6.1.4 Co-localization of annexin A2 and S100A10 proteins in ovarian cancer cells .. 162

6.2 Future directions ...... 163

6.3 Conclusions ...... 164

References ...... 167

Appendix ...... 185

List of Figures Figure 1.1 Annexin A2 domain structure...... 38 Figure 1.2 2D-gel electrophoresis of proteins in co-cultured ovarian cancer and peritoneal cell.... 39 Figure 1.3 Annexin A2 expression in human ovarian tumour tissues...... 40 Figure 1.4 Proposed mechanism of annexin A2 promoting cancer metastasis in the plasminogen activation system...... 41 Figure 2.1 Motility and invasion of human ovarian cancer cell lines (OVCAR-3, SKOV-3 and OV- 90) in vitro...... 55 Figure 2.2 Chick embryo development...... 56 Figure 2.3 Invasion of ovarian cancer cells in the chick chorioallantoic membrane (CAM)...... 57 Figure 2.4 Chick chorioallantoic membrane (CAM) invasion by ovarian cancer cells in day 14 chick embryos...... 58 Figure 2.5 Effects of protein A neutralizing antibody on OV-90 cancer cell invasion into the chick chorioallantoic membrane (CAM)...... 59 Figure 3.1 Annexin A2 immunostaining in human ovarian tissues and omental tissues...... 83 Figure 3.2 Morphology of primary ovarian cancer cell lines derived from patients ascites...... 85 Figure 3.3 Annexin A2 expression in primary ovarian cancer cells...... 86 Figure 3.4 Annexin A2 expression in ovarian cancer cell lines, peritoneal cell line and co-cultured ovarian cancer and peritoneal cells...... 87 Figure 3.5 The effects of annexin A2 siRNA knockdown on ovarian cancer cell motility, invasion and adhesion in vitro...... 88 Figure 3.6 Annexin A2 promotes OV-90 cell motility and invasion in vitro and in vivo...... 90 Figure 3.7 Neutralizing annexin A2 antibody inhibits primary ovarian cancer cells invasion in the CAM model...... 92 Figure 3.8 Annexin A2 promotes growth and metastasis of SKOV-3 cells...... 94 Figure 3.9 Annexin A2 neutralizing antibodies increase apoptosis...... 95 Figure 4.1 Annexin A2 expression in controls and ovarian cancer patient’s urine...... 118 Figure 4.2 Annexin A2 and CA125 levels in normal ovaries, benign ovarian tumours and serous ovarian cancers...... 120 Figure 4.3 Correlation of annexin A2 and CA125 levels...... 121 Figure 4.4 ROC curves from logistic regression analysis for annexin A2, CA125 and combined annexin A2 and CA125 measurements...... 123 Figure 5.1 Annexin A2 and S100A10 immunostaining in stage III serous ovarian cancers ...... 143

Figure 5.2 Annexin A2 and S100A10 immunostaining in primary ovarian tumour and matching metastasis tissues...... 144 Figure 5.3 High stromal annexin A2 expression is associated with poor prognosis in stage III serous ovarian cancer patients...... 146 Figure 5.4 S100A10 cytoplasmic expression is associated with reduced overall survival for stage III serous ovarian cancer patients...... 148 Figure 6.1 The mechanism of annexin A2 in cancer progression ...... 165 Figure 6.2 The effects of annexin A2 inhibitors on cancer cell invasion and metastasis ...... 166

List of Supplementary Figures

Supplementary Figure 3.1 Ion-trap mass spectrometry LC-MS/MS analysis of annexin A2 spots from the 2D gel electrophoresis...... 100 Supplementary Figure 3.2 Annexin A2 expression after knockdown with annexin A2 siRNAs. ... 101 Supplementary Figure 3.3 IVIS imaging of mice...... 102

List of Tables

Table 1.1 Annexin A2 expression in cancer...... 42 Table 1.2 Summary of the effects of annexin A2 on cancer cells...... 43 Table 2.1 Comparison of the advantages and limitations of the chick chorioallantoic membrane (CAM) and mouse model...... 62 Table 4.1 Comparison of annexin A2 and CA125 marker sensitivity, specificity, PPV and NPV for controls (normal ovaries and benign ovarian tumours) and serous ovarian cancers using specified cutoff points...... 124 Table 4.2 Comparison of annexin A2 and CA125 markers sensitivity, specificity, PPV and NPV for benign ovarian tumours and serous ovarian cancers using specified cutoff points...... 125 Table 4.3 Comparison of annexin A2 and CA125 markers in controls (normal ovaries and benign ovarian tumours) and serous ovarian cancers using logistic regression and ROC curve analyses...... 126 Table 4.4 Comparison of annexin A2 and CA125 markers in benign ovarian tumours and serous ovarian cancers using logistic regression and ROC curve analyses...... 127 Table 5.1 Correlation of annexin A2 and S100A10 immunostaining in the membrane and cytoplasm of cancer cells and cancer associated stroma in stage III primary serous ovarian cancer...... 149 Table 5.2 Correlation of annexin A2 stromal intensity with age, tumour grade and residual disease in stage III primary serous ovarian cancer...... 150 Table 5.3 Cox Regression analyses of stage III serous ovarian cancer patient cohort…………...151

List of Supplementary Tables

Supplementary Table 3.1 Annexin A2 epithelial and stromal immunostaining intensity in the normal ovaries, serous cystadenomas, serous borderline tumours and serous ovarian carcinomas (stage I-IV) tissues...... 103 Supplementary Table 3.2 Annexin A2 epithelial and stromal immunostaining intensity in the primary tumour and matching omental metastasis tissues...... 104 Supplementary Table 3.3 Annexin A2 epithelial immunostaining intensity in the omental tissues...... 105 Supplementary Table 3.4 Clinical characteristics of primary ovarian cancer cell lines derived from patients’ ascites...... 106 Supplementary Table 4.1 Clinicopathological characteristics of urine patient cohort...... 131 Supplementary Table 4.2 Clinicopathological characteristics of serum or plasma patient cohort. 132 Supplementary Table 5.1 Clinicopathological characteristics of stage III serous ovarian cancer patient tissue cohort...... 155

Summary

Ovarian cancer is the most lethal gynaecological cancer. We identified annexin A2 to be modulated in the co-culture of ovarian cancer and peritoneal cells in vitro. Annexin A2, a calcium phospholipid binding protein, has been characterized in many malignancies and mediates various cellular functions such as cell motility, invasion, proliferation, angiogenesis and cell adhesion. Annexin A2 heterotetramer consists of annexin A2 and S100A10 monomers on the cell surface and plays an important role in the plasminogen activator system that enhances cancer cell invasion and metastasis. The aim of this Ph D thesis was to investigate the role of annexin A2 in ovarian cancer metastasis using in vitro and in vivo ovarian cancer models. Annexin A2 expression was characterized in serous ovarian cancer cell lines and human serous ovarian cancer tissues. Annexin A2 inhibitors were used to evaluate the effects of annexin A2 on ovarian cancer cell motility, invasion and adhesion to the peritoneal cells. Furthermore, annexin A2 neutralizing antibodies were used to examine the role of annexin A2 in tumour invasion and metastasis using a chick chorioallantoic membrane (CAM) assay and an intraperitoneal xenograft mouse model. We evaluated whether annexin A2 can be used as a diagnostic marker by measuring blood annexin A2 levels in serous ovarian cancer patients. Moreover, annexin A2 and its binding protein, S100A10 expression were assessed using immunohistochemistry to determine their relationship with clinical outcome in a cohort of stage III serous ovarian cancers. We showed that annexin A2 immunostaining was significantly increased in cancer-associated stromal cells compared with non-malignant ovarian tissues. Annexin A2 siRNAs significantly inhibited ovarian cancer cell motility, invasion and adhesion to peritoneal cells. Moreover, annexin A2 neutralizing antibodies significantly inhibited OV-90 cell motility and invasion in vitro and in the in vivo CAM assay. Furthermore, we also demonstrated that annexin A2 neutralizing antibodies significantly inhibited the invasion of primary ovarian cancer cell lines in the CAM assay. The growth of SKOV-3/GFP Luc cells and peritoneal dissemination in nude mice was significantly inhibited by annexin A2 neutralizing antibodies. Our findings suggested that reduced tumour burden and metastatic spread was a result of reduced cell survival. Blood annexin A2 levels were increased in early stage and advanced stage ovarian cancer patients compared with women without malignancy (normal ovaries and benign ovarian tumours). We showed an improved sensitivity for detecting early stage ovarian cancer by combining annexin A2 and CA125 at the 95% and 98% specificity level. Kaplan-Meier analyses of stage III serous ovarian cancers showed that high stromal annexin A2 expression was significantly associated with

reduced progression-free survival and overall survival. Moreover, we also showed high cytoplasmic S100A10 in the cancer cells to be associated with reduced overall survival. Both, high stromal annexin A2 and high cytoplasmic S100A10, were independent predictors of overall survival in a multivariate analysis which included positive residual disease. In conclusion, our findings indicate that annexin A2 plays an important role in ovarian cancer tumourigenesis and metastasis is therefore a potential novel therapeutic target against ovarian cancer. We also demonstrated that annexin A2 has both diagnostic and prognostic significance and may be useful for serous ovarian cancer diagnosis and patient management.

Declaration

I certify that this thesis does not incorporate without acknowledgement any material previously submitted for the award of any degree or diploma in any university; and that to the best of my knowledge and belief, this work does not contain any material previously published or written by any other person except where due reference is made in the text.

I give consent to the copy of my thesis to be deposited in the University Library to be made available for loan and photocopying, subject to embargoes, and the Copyright Act 1968.

I also give permission for the digital version of my thesis to be made available on the web, via the University’s digital research repository, the ADTP, and the library catalogue, subject to embargoes.

……………………………… Noor Alia Lokman February 2014

Acknowledgements

I am grateful to a lot of people for their help and support throughout my Ph D experience. First of all, I would like thank my supervisor, Dr Carmela Ricciardelli that has given me the best supervision over the last four years. Thank you for being patience, supportive, helping me out with my data analysis and a great mentor to me. I would like to thank my co-supervisor, Prof Martin Oehler for your supervision, guidance, ideas and discussion throughout my Ph D. Thank you Carmela and Martin for evaluating my manuscripts and being committed to my research project.

I would like to thank previous and present members of Reproductive Cancer group that assist me in my Ph D studies, Dr Miranda Ween, Dr Carmen Pyragius, Wendy Bonner, Dr Alison Elder and Dr Izza Tan for their friendship, technical assistance in the lab and manuscript evaluations. I would like to thank A/Prof Peter Hoffmann and Karina Martin from Adelaide Proteomics Center for your assistance throughout my Ph D and also Prof Ray Rodgers, Dr Katja Hummitzsch and Nick Hatzirodos for their help throughout my PhD. Previous and present Ph D students for their friendship and assistance throughout my studies: Nok, Lisa, Adrian, Jacky, Siew, Macarena, Zahied, Siti Mariam, Jess, Dulama, David, Vincent and Sally. Thank you all.

I am fortunate to receive scholarships from Adelaide Graduate Fee Scholarship, Florey Foundation, Faculty of Health Sciences and Robinson Institute for supporting my Ph D journey and opportunity to present my research in conferences. Thank you also to the staff of Discipline of Obstetrics and Gynaecology for excellent resources throughout my studies. I’m also grateful to the patients at Royal Adelaide Hospital and Burnside Hospital that have given consent for their samples to be used in this study.

Thank you to my friends Ayuni, Suhana, Afifah and Solehah for your friendship and support. To my best friend, Nur Hezrin Shahrin, for her continuous support in good and hard times and being a good listener in our coffee catch-up sessions in these past few years, all the best with your baby boy, Adyan Hamzah. Finally, to my loving parents, Norma Muhammad and Lokman Abdul Hamid, for everything that they have done for me and made me who I am today. There’s no way that I can thank you enough for your endless love, support and encouragement. Mak and ayah, I’ve been blessed to have both of you in my life, I love you so much. To my one and only sister, Dr Noor Dayana for her love and strong character, I’m so proud of you. To my brothers, Adi Ikmal, Dani Aizat and Dhiya Aiman, who always had been there for me. And to my family, Cik Yah, Cik Nin, Azwa and everyone for their love and support.

Publications Arising During Ph D Candidature

1. Ricciardelli C, Ween MP, Lokman NA, Tan IA, Pyragius CE and Oehler MK, Chemotherapy-induced hyaluronan production: A novel chemoresistance mechanism in ovarian cancer. BioMed Central Cancer, 2013, 13:476. (Impact factor: 3.33)

2. Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis. Oncotarget, 2013, 4(8):1129-1211. (Impact factor: 6.636)

3. Lokman NA, Elder ASF, Ricciardelli C and Oehler MK, Chick chorioallantoic membrane (CAM) assay as an in vivo model to study the effect of newly identified molecules on ovarian cancer invasion and metastasis. Int.J.Mol.Sci., 2012,13, 9959-9970. (Impact factor: 2.464)

4. Lokman NA, Ween MP, Oehler MK and Ricciardelli C, The role of annexin A2 in tumourigenesis and cancer progression. Cancer Microenvironment, 2011, 4, 199-208.

5. Ween MP, Lokman NA, Hoffmann P, Rodgers RJ, Ricciardelli C and Oehler MK, Transforming growth factor-beta-induced protein secreted by peritoneal cells increases the metastatic potential of ovarian cancer cells, Int.J. Cancer, 2011,128, 1570-1584. (Impact factor: 6.198)

Publications Contributing to This Thesis

1. Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis. Oncotarget, 2013, 4(8):1129-1211. (Impact factor: 6.636)

2. Lokman NA, Elder ASF, Ricciardelli C and Oehler MK, Chick chorioallantoic membrane (CAM) assay as an in vivo model to study the effect of newly identified molecules on ovarian cancer invasion and metastasis. Int.J.Mol.Sci., 2012,13, 9959-9970. (Impact factor: 2.464)

3. Lokman NA, Ween MP, Oehler MK and Ricciardelli C, The role of annexin A2 in tumourigenesis and cancer progression. Cancer Microenvironment, 2011, 4, 199-208.

Presentations at Scientific Meeting 2013 Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis, Robinson Institute Research Symposium, 4th November 2013, National Wine Centre, Adelaide, South Australia, Australia. (poster)

Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis, Matrix Biology Society of Australia and New Zealand (MBSANZ), 20-23rd October 2013, McCracken Country Club, South Australia, Australia. (oral and poster)

Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis, Postgraduate Research Expo, Faculty of Health Science, University of Adelaide, 29th August 2013, National Wine Centre, Adelaide, South Australia, Australia. (poster)

Lokman NA, Elder ASF, Ween MP, Pyragius CE, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis, The Centre for Personalised Cancer Medicine Symposium ‘Driving personalised cancer medicine through next generation sequencing’, 25th July 2013, National Wine Centre, Adelaide, South Australia, Australia. (poster)

2012 Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour in ovarian cancer metastasis, Robinson Institute Research Symposium, 12th December 2012, National Wine Centre, Adelaide, South Australia, Australia. (poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour in ovarian cancer metastasis, Cold Spring Harbor Asia/International Cancer Microenvironment

Society Joint Conference on Tumour Microenvironment, 13th to 17th November 2012, Suzhou Dushu Lake Conference Center, Suzhou, China. (poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour in ovarian cancer metastasis, Frontiers in Cancer Science 2012, 5th to 8th November 2012, National University of Singapore, Singapore. (poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour in ovarian cancer metastasis, Matrix Biology Society of Australia and New Zealand (MBSANZ) 2012, 5th to 8th September 2012, Mantra Legends, Gold Coast, Queensland, Australia. (oral and poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour in ovarian cancer metastasis, 14th International Biennial Congress of the Metastasis Research Society, 2nd to 5th September 2012, Brisbane Convention Centre, Brisbane, Queensland, Australia. (poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Annexin A2 released during ovarian cancer-peritoneal cell interaction promotes a pro-metastatic cancer cell behaviour. Postgraduate Research Expo, Faculty of Health Science, University of Adelaide, 31st August 2012, The National Wine Centre, Adelaide, South Australia, Australia. (poster)

2011 Lokman NA, Ween MP, Hoffmann P, Oehler MK. and Ricciardelli C, Role of Annexin A2 in ovarian cancer metastasis. Australian Society for Medical Research (ASMR), SA Scientific Meeting, Adelaide Convention Center, 7th June 2011, South Australian Division (poster)

2010 Lokman NA, Ween MP, Hoffmann P, Oehler MK. and Ricciardelli C, Role of Annexin A2 in ovarian cancer metastasis. Matrix Biology Society of Australia and New Zealand (MBSANZ) combined

meeting with the 5th Australian Health & Medical Research Congress 2010, 14th-18th November 2010, Melbourne Convention and Exhibition Centre, Victoria, Australia. (oral and poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK and Ricciardelli C, Role of Annexin A2 in ovarian cancer metastasis, Postgraduate Research Expo, Faculty of Health Science, University of Adelaide, 1st September 2010, The National Wine Centre, Adelaide, South Australia, Australia. (poster)

Lokman NA, Ween MP, Hoffmann P, Oehler MK. and Ricciardelli C, Role of Annexin A2 in ovarian cancer metastasis. Australian Society for Medical Research (ASMR), SA Scientific Meeting, Adelaide Entertainment Centre, 9th June 2010, South Australian Division. (oral)

Awards Arising Out of This Thesis

1. High Commended - Student Poster Award, 2013 Robinson Institute Research Symposium, 4th November 2013, National Wine Centre, Adelaide, South Australia, Australia.

2. Vice-Chancellor’s Prize for Best Poster, Faculty of Health Sciences Postgraduate Research Conference, 29th August 2013, National Wine Centre, Adelaide, South Australia, Australia.

3. Best Student Poster Award, 2012 Robinson Institute Research Symposium, 12th December 2012, National Wine Centre, Adelaide, South Australia, Australia.

4. Best Student Poster Award 2012, Matrix Biology Society of Australia and New Zealand (MBSANZ), 5th to 8th September 2012, Mantra Legends, Gold Coast, Queensland, Australia.

Abbreviations

Є-ACA 6-Aminocaproic Acid Ab Antibody ABC ATP binding cassette ANXA2 Annexin A2 CA125 Cancer Antigen 125 CAM Chick Chorioallantoic Membrane CD44 Cluster of Differentiation 44 CM Conditioned Media CMI Cancer Cells and Matrigel Implant DAB Diaminobenzidine Tetrahydrochloride ECT Ectoderm ECL Enhanced chemiluminescence ECM Extracellular Matrix EDTA Ethylenediaminetetraacetic Acid EMT Epithelium-Mesenchymal Transition END Endoderm ELISA Enzyme Linked Immunosorbent Assay ERK Extracellular Signal Regulated Kinase FBS Fetal Bovine Serum FIGO International Federation of Gynaecologist and Obstetricians GFP Green Fluorescent Protein GM6001 Galardin

H2O2 Hydrogen Peroxide HE4 Human Epididymis Protein 4 HCC Hepatocellular Carcinoma HCl Hydrochloric Acid Ig Immunoglobulin MAPK Mitogen Activated Protein Kinase MES Mesoderm MMPs Matrix Metalloproteinases

N-terminal domain Amino Terminus Domain NPV Negative Predictive Values pI Isoelectric Point P11 Protein P11 PBS Phosphate Buffered Saline PBS-T Phosphate Buffered Saline with Tween-20 PPV Positive Predictive Values RMI Risk Malignancy Index ROC Receiver Operating Curve ROMA Risk of Ovarian Malignancy Algorithm RR Relative Risk RT-PCR Real-time Polymerase Chain Reaction u-PA Urokinase-Type Plasminogen Activator SDS Sodium Dodecyl Sulphate siRNAs Small Interfering RNAs SKOV-3/GFP Luc SKOV-3 cells labeled with GFP and Luciferase t-PA Tissue-Type Plasminogen Activator TGFBIp Transforming Growth Factor Inducible Protein Tris Tris(hydroxymethyl)-aminomethane TMA Tissue Microarray LC-MS/MS Liquid Chromatography-Mass Spectrometry

1 Chapter 1: The Role of Annexin A2 in Tumourigenesis and Cancer Progression

Noor A Lokman1, Miranda P Ween1, Martin K Oehler1, 2 and Carmela Ricciardelli1

Received: 6 January 2011, Accepted: 21 February 2011, Published online: 5 March 2011

Cancer Microenvironment (2011) 4:199-208

STATEMENT OF AUTHORSHIP Chapter 1: The Role of Annexin A2 in Tumourigenesis and Cancer Progression

Noor Alia Lokman (candidate) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Development and wrote the manuscript.

Certification that the statement of contribution is accurate

Signed: …………………………………………………………………………Date:…………………..

Miranda P. Ween (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Martin K. Oehler (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Carmela Ricciardelli (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work, manuscript evaluation and acted as corresponding author.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

1.1 Abstract Annexin A2 is a calcium-dependent, phospholipid-binding protein found on various cell types. It is up-regulated in various tumour types and plays multiple roles in regulating cellular functions, including angiogenesis, proliferation, apoptosis, cell migration, invasion and adhesion. Annexin A2 binds with plasminogen and tissue plasminogen activator on the cell surface, which leads to the conversion of plasminogen to plasmin. Plasmin is a serine protease which plays a key role in the activation of metalloproteinases and degradation of extracellular matrix components essential for metastatic progression. We have recently found that both annexin A2 and plasmin are increased in conditioned media of co cultured ovarian cancer and peritoneal cells. Our studies suggest that annexin A2 is part of a tumour-host signal pathway between ovarian cancer and peritoneal cells which promotes ovarian cancer metastasis. Accumulating evidence suggest that interactions between annexin A2 and its binding proteins play an important role in the tumour microenvironment and act together to enhance cancer metastasis. This article reviews the current knowledge on the biological role of annexin A2 and its binding proteins in solid malignancies including ovarian cancer.

Keywords: annexin A2, p11 protein, t-PA, plasmin and metastasis

1.2 Introduction The plasminogen activation system known to be involved in thrombolysis and wound healing plays a major role in cancer progression. The inactive enzyme plasminogen is converted to the active serine protease plasmin by plasminogen activators; tissue-type plasminogen activator (t-PA) and urokinase-plasminogen activator (u-PA). Plasmin is involved in many biological processes besides vascular thrombolysis including angiogenesis, tissue remodelling and also activates metalloproteinases (MMPs) and latent growth factors and the degradation of extracellular matrix (ECM), processes which all contribute to tumour development and metastasis [1]. A well-studied protein which is an important mediator in the plasminogen activator system is annexin A2 [2]. The interaction between annexin A2 and t-PA mediates the conversion of plasminogen to plasmin and increases the catalytic efficiency of plasmin generation by up to 60-fold [3]. Annexin A2, a calcium-binding cytoskeletal protein is localized at the extracellular surface of endothelial cells and various types of tumour cells [4, 5]. The increased expression of annexin A2 has been reported in cancers of the breast, liver, prostate and pancreas [6-9]. Additionally, annexin A2 has also been demonstrated to play a role in cancer cell migration, invasion [7, 10] and adhesion [11, 12], processes which are essential for cancer metastasis. This article reviews the current knowledge on annexin A2 expression in various malignancies including ovarian cancer and its role in regulating cancer cell behavior.

1.3 The structure and function of annexin A2 Annexin A2 belongs to the annexin family of calcium binding proteins which consists of 12 members (A1-A11 & A13). Each family member shares the conserved calcium binding motifs in the carboxyl domain. However, the individual annexin members have diverse functions due to the highly variable amino-terminal domain [13]. Annexin family members have been reported to play multiple roles in cancer including signal transduction, angiogenesis, tumour invasion and metastasis and apoptosis (reviewed in [14]). Studies have shown annexin A1, A2, A3, A4, A6, A8 and A11 to be up-regulated whilst annexin A1, A2, A7, A8, A10 and A11 are decreased in different cancer types [14]. Moreover, annexins have also been associated with chemo resistance in breast and ovarian cancer cells [15-18]. The human annexin A2 gene spans approximately 40kb on the long arm of chromosome 15 (15q21) [19]. This region has been shown to be associated with allelic imbalance in advanced breast cancer [20]. However, no annexin A2 mutations have been described to date in cancer or other diseases. Annexin A2 has a functional N-terminal domain with p11 proteins and t-PA binding

sites and a C-terminal domain consist of binding sites for calcium, phospholipids and actin filaments (Figure 1.1). Annexin A2 exits both as a monomer and a heterotetramer. The annexin A2 monomer is an intracellular 38 kDa protein whilst, annexin A2 heterotetramer consisting of two subunits of annexin A2 monomers and two subunits of p11 proteins (also known as S100A10) is localized to the plasma membrane [21]. The translocation of annexin A2 monomers from the cytoplasm to the cell surface occurs as a result of heat stress [22], tyrosine phosphorylation [23], and interaction with heat shock protein 90 alpha [24, 25]. In endothelial cells, translocation of annexin A2 to the plasma membrane due to heat stress is dependent on the expression of p11 proteins and phosphorylation of tyrosine 23 [22]. Annexin A2 also assists in the rearrangement of the actin cytoskeleton, maintaining the plasticity of the actin cytoskeleton and regulating membrane trafficking events including endocytosis and exocytosis [26-28].

1.4 Expression of annexin A2 in cancer Many studies have reported increased expression of annexin A2 in cancer tissues compared with normal tissues [6, 7, 29-31] (Table 1.1). The up-regulation of annexin A2 expression in pancreatic, colorectal, and brain tumours was directly correlated with advanced clinical stage [31-33]. Higher annexin A2 expression was also observed in metastatic breast cancer and colon cancer cells compared with the non-metastatic cells [7, 34]. In prostate cancer however, reports regarding annexin A2 have been conflicting. Whilst several studies reported a reduction of annexin A2 expression in prostate cancer tissues [35-38], a study by Banerjee et al. (2003) showed high focal membrane staining of annexin A2 in high grade prostate cancers and the prostate carcinoma cell lines, PC3 and DU145 [39]. Furthermore, a more recent study demonstrated a correlation between high annexin A2 levels and a more aggressive prostate cancer phenotype [9]. Conflicting results on annexin A2 expression have also been observed in head and neck cancers. A study by Wu et al. (2002) using 2D gel-electrophoresis of two head and neck squamous cell carcinoma cell lines derived from primary and metastatic tumours from the same patient, found annexin A2 to be up-regulated in the cell line established from the metastatic cancer cell line [40]. In contrast, Pena-Alonso et al. (2008) observed a significant down-regulation of annexin A2 mRNA and protein levels in poorly differentiated carcinomas of the head and neck by real time-PCR and immunohistochemistry [41]. These contradictory findings may be explained by different experimental techniques utilised to examine annexin A2 expression or differences between primary tumours and metastatic lesions. The different annexin A2 levels identified by Wu et al. (2002) may reflect annexin A2 protein modifications detected by mass spectrometry which are not

detectable by real-time PCR or immunohistochemistry [40]. Despite various studies having confirmed either an up or down regulation of annexin A2 in various tumour types, the exact mechanism that regulate annexin A2 expression at both mRNA and protein level are poorly understood. Annexin A2 apart from being a cell surface protein, it has also been shown to be secreted. Deora et al. (2003) reported that annexin A2 is secreted via a non-classical secretory pathway as it does not have a signal secretion sequence [22]. Secreted annexin A2 was also found in the conditioned media of co-cultured human keratinocytes and fibroblasts [42]. Annexin A2 secretion has been also reported for various tumour types. Davis et al. (1995) demonstrated high levels of secreted annexin A2 in a human pancreatic adenocarcinoma cell line (Capan-2) [43]. Zhao et al. (2003) have reported that annexin A2 secretion by rat adrenal pheochromocytoma (PC12) cells in vitro is induced by activation of both the insulin and insulin like growth factor receptor followed by reduction of intracellular annexin A2 [44]. A study by Ji et al. (2009) reported that annexin A2 was up-regulated in the serum of hepatocellular carcinoma patients compared with the serum of healthy patients by ELISA [45]. A recent study using a proteomic approach investigating the secretome of the gastric cancer cell line, SGC7901 identified annexin A2 as a secreted phosphoprotein [46]. Together these studies suggest that annexin A2 has potential diagnostic and prognostic value for several different carcinomas which needs to be further examined.

1.5 Role of annexin A2 and its binding proteins in cancer The co-localization of annexin A2 with its binding proteins including t-PA, p11 protein, tenascin C and cathepsin B facilitates the proteolytic cascade leading to the activation of pro-enzymes and selective degradation of ECM components which regulate cancer cell properties and enhance metastasis (reviewed in [5]). Several studies have demonstrated the co-expression of annexin A2 and p11 protein in tumour and endothelial cells. Annexin A2 co-localizes with p11 protein at the surface of tumour cells where plasminogen activation occurs [47]. Puisieux et al. (1996) showed a co-expression of annexin A2 and p11 protein in colon, renal and liver carcinoma cell lines [48]. Moreover, over-expression of annexin A2 and p11 protein was observed in renal cell carcinoma tissues when compared with normal kidney [31]. Zhang et al. (2010) also reported over-expression of annexin A2 and p11 protein in the invasive breast cancer cell lines compared with the non- invasive breast cancer cell lines [47]. Annexin A2 also binds to tenascin C (reviewed in [5]), a large ECM molecule which is expressed in the tumour microenvironment and plays an important role in tumourigenesis (reviewed in [49]).

Esposito et al. (2006) reported annexin A2 and tenascin C to be over-expressed at both the mRNA and protein levels in pancreatic cancers compared with normal pancreas tissues [50]. Moreover, the interaction between annexin A2 and tenascin C has been shown to regulate cell migration and to enhance cell proliferation of endothelial cells [51]. Tenascin C has also been shown to be over- expressed in malignant ovarian tumour and to regulate adhesion and migration of ovarian cancer cells [52]. Cathepsin B, a lysosomal cysteine protease on the cell surface of various tumour cells (reviewed in [5]) also binds to annexin A2. Mai et al. (2000) reported that human pro-cathepsin B interacts with the annexin A2 heterotetramer which is co-localized on the surface of human breast carcinoma and glioma cells [53]. Cavallo-Medved et al. (2009) recently reported that the co-localization of cathepsin B and annexin A2 heterotetramer were involved in ECM degradation in endothelial cells during in vitro tube formation [54]. There are also other proteins which have been reported in the literature to bind to annexin A2. Annexin A2 receptor protein which has been characterized in osteoclasts [55] plays an important role in promoting prostate cancer progression [11]. Annexin A2 binds to collagen I in epithelial cells in a calcium dependent manner [56] and annexin A2 also interacts with the transcription factor, STAT6 in the metastatic prostate cancer cells [57]. Annexin A2 also binds to progastrin in colon cancer cells [58] and the interaction between annexin A2 and progastrin mediate growth effects of progastrin in gastrointestinal cancers (reviewed in [59]). Furthermore, annexin A2 has also been reported to form complexes with other S100 proteins such as S100A4 in endothelial cells [60] and S100A6 in pancreatic cancer cells [61].

1.6 Role of annexin A2 in the plasminogen activator system Emerging evidence indicates that annexin A2 plays an important role in the plasminogen activator system which regulates cancer metastasis and angiogenesis (reviewed in [62, 63]). The interaction between annexin A2 and t-PA mediates the conversion of plasminogen to plasmin [3, 64, 65] which facilitates MMP activation and ECM degradation leading to enhanced cancer cell migration and invasion [10, 66]. These findings suggest that annexin A2 is likely to play an important role in the early dissemination model which has been recently been described by Coghlin and Murray [67]. Kinetic studies have demonstrated that binding of annexin A2 to t-PA on the cell surface of endothelial cells increases the catalytic efficiency of plasmin generation by up to 60-fold [3]. The co-localization of annexin A2 and t-PA has been demonstrated in breast and pancreatic cancer

cells [10, 66, 68]. Roda et al. (2003) reported that the annexin A2 binding site for t-PA, the LCKLSL peptide motif (amino acid 8-13), is vital for the interaction between annexin A2 and t-PA to mediate plasmin production [69]. This was confirmed by Diaz et al. (2004) who reported activation of plasminogen via specific interaction of t-PA and annexin A2 in pancreatic cancer cells using the LCKLSL annexin A2 peptide [10]. Over-expression of annexin A2 resulted in higher levels of plasmin production in acute promyelocytic leukaemia cells [70] and increased plasmin levels were also observed in invasive breast cancer cell lines when compared with non-invasive cell lines [47, 66]. Recently, the interaction between heat shock protein 90 alpha (Hsp90a), annexin A2, t-PA and plasminogen in the exosomes of human breast, glioma and fibrosarcoma cancer cell lines resulted in increased cancer cell motility due to increased plasmin production [24]. Thus, these studies suggest increased annexin A2 expression enhances plasmin generation in the cancer cells via the t-PA dependent plasminogen activation pathway.

1.7 Role of annexin A2 in cancer cell adhesion, invasion, and migration Several studies suggest that annexin A2 plays an essential role in regulating cancer cell adhesion, invasion, proliferation and migration [10, 36, 66, 71]. The effects of annexin A2 on different cancer cells are summarised in Table 1.2. Sharma et al. (2006) demonstrated that the highly invasive breast cancer cell line, MDA-MB231 expressed high annexin A2 levels and in the presence of plasminogen resulted in plasmin generation which could be blocked by a monoclonal antibody to annexin A2 [7]. However, neither plasmin production nor annexin A2 expression was observed in the non-metastatic MCF-7, breast cancer cell line [7]. Additionally, annexin A2 expression was shown to be up-regulated in MDA-MB231 breast cancer cells by basic fibroblast growth factor (bFGF) which is known to accelerate tumour growth and angiogenesis [7]. A later study by Sharma et al. (2009) confirmed the importance of annexin A2 and t-PA interactions on the cell surface of MDA-MB231 cells and demonstrated that the t-PA dependent plasmin generation was essential for cancer cell migration and neo-angiogenesis [66]. In contrast, Gillette et al. (2004) showed that over-expression of annexin A2 in osteosarcoma cells did not alter their metastatic properties such as motility, adhesion, or proliferation. This study suggested however that annexin A2 decreases osteosarcoma aggressiveness by inducing a more differentiated state [72]. Furthermore, Jung et al. (2007) demonstrated using both in vivo and in vitro studies that the annexin A2 heterotetramer regulates adhesion of hematopoietic stem cells to osteoblasts and bone marrow endothelial cells

[73]. Recently, Shiozawa et al. (2008) also reported that annexin A2 regulates adhesion and migration of prostate cancer cells to osteoblasts and endothelial cells [11]. Annexin A2 has also been demonstrated to play an important role in regulating cytoskeleton structures and actin remodeling which both are essential for cancer cell migration [74, 75]. A recent study by Pu et al. (2009) reported that annexin A2 interaction with HAb18G/CD187 (a member of immunoglobulin family) promotes invasion and migration of human hepatocellular carcinoma cells (FHCC-98 cells) in vitro [76]. siRNA studies have also confirmed that knockdown of annexin A2 expression decreases cell migration in human glioma cells [77] and decreases the invasive ability of multiple myeloma cells [71]. The silencing of annexin A2 expression in hepatocellular carcinoma cells results in a significant decrease of both MMP-2 and MMP-9 levels which decreases cancer cell invasiveness [76]. Bao et al. (2009) also showed that the loss of annexin A2 expression leads to reduced levels of pro-angiogenic molecules including MMP-2, MMP-9, MTI-MMP, and TIMP2 in multiple myeloma cell lines (U266 and RPMI18226) [71]. These pro-angiogenic molecules is essential for tumour cells properties such as angiogenesis, proliferation and invasion [71]. Annexin A2 and heat shock protein 90 alpha (Hsp90a) released in the exosomes of the tumour cells activates plasminogen and pro-MMP2 which increase plasmin production to enhance ECM remodeling and increased tumour cell motility and invasion in the tumour microenvironment [24]. These studies demonstrate that annexin A2 is a key factor in regulating MMP secretion and activation which are essential for ECM degradation and metastatic progression. Hou et al. (2008) reported that annexin A2 regulates p11 protein and plasmin levels in a mouse lymphoma (L5178Y) cell line [78]. Annexin A2 knockout in L5178Y cells exhibited reduced cell motility and invasion in vitro and in vivo [78]. In a recent study by Zhang et al. (2010) showed that the silencing of annexin A2 expression resulted in an inhibition of breast cancer cell proliferation and invasion which was also associated with a down-regulation of p11 protein and plasmin levels [47]. Together these studies indicate that annexin A2 plays an important role in promoting adhesion, migration and subsequent metastasis of cancer cells.

1.8 Post translational modifications and proteolytic cleavage of annexin A2 in cancer cells Several studies have reported post-translational modification and proteolysis of annexin A2 in cancer. Multiple isoforms of annexin A2 were identified by 2D gel electrophoresis and mass spectrometry in colorectal and oral carcinomas [79, 80]. Phosphorylation at tyrosine 23 in the N- terminal domain of annexin A2, has been shown to be present in liver [8] and pancreatic cancer

tissues but not in corresponding normal tissues [32]. Takano et al. (2008) also observed post- translational modifications of annexin A2 in a chemotherapy-resistant pancreatic cancer cell line compared with chemo-sensitive cells [81]. Zheng et al. (2009) recently reported that phosphorylation of annexin A2 at tyrosine 23 was essential for localisation of annexin A2 at the cell surface and its contribution to the invasive properties of pancreatic cancer cells [32]. Eustace et al. (2011) recently reported increased phosphorylated forms of annexin A2 in dasatinib sensitive melanoma cell lines (WM-115) compared with the dasatinib resistant cell line (WM-2660-4) [82]. Annexin A2 can also undergo proteolysis. N-terminal processing of annexin A2 in human peripheral monocytes [83] and endothelial cells [84] induced by plasmin resulted in the loss of the first 27 amino acid residues. Tomonaga et al. (2007) reported cleavage of annexin A2 in the C- terminal end of annexin A2 in the healthy bowel mucosa but not in the colorectal cancer samples [80]. Tsunezumi et al. (2008) demonstrated that treatment of human colon cancer and breast cancer cell lines with MMP-7 (matrilysin) can result in specific cleavage of annexin A2 at the N- terminal region between Lys9 and Leu10 within the t-PA binding motif [85]. This study additionally demonstrated that the first nine amino acid residues in the N-terminal domain of annexin A2 bound preferentially to t-PA over intact annexin A2 on the surface of colon cancer cells [85]. Annexin A2 cleavage induced by glycogen synthase-3 was also observed in a time-dependent manner in serum-free conditioned media of human lung epithelial cells and human epidermoid cells [86]. Whilst several studies have demonstrated post-translational modifications of annexin A2, including phosphorylation changes and annexin A2 cleavage by cancer cells, the regulation of these modifications is poorly understood. Further studies are required to evaluate the significance of these modified forms of annexin A2 in cancer.

1.9 Annexin A2 and ovarian cancer To date, there is limited knowledge about the importance of annexin A2 in ovarian cancer. A proteomic ovarian cancer biomarker study by Wang et al. (2004) using 2D liquid phase separation followed by reverse-phase high performance liquid chromatography (RP-HPLC) and matrix- assisted laser desorption or ionisation-time of flight (MALDI-TOF) mass spectrometry, identified annexin A2, amongst other proteins at three different pI values. Annexin A2 at pI 7.10, was markedly elevated in ovarian cancer cell lines when compared with normal ovarian surface epithelial cells, suggesting that a post-translational modification of annexin A2 occurs in ovarian cancer cells [87]. Furthermore, Gagne et al. (2007) also found annexin A2 to be up-regulated in the highly metastatic, TOV-112D epithelial ovarian cancer cell line compared with the TOV-81D

ovarian cancer cells with low malignant potential, using iTRAQ (isobaric tag for relative and absolute quantitation) and 2D gel electrophoresis coupled with liquid chromatography and tandem mass spectrometry [88]. A more recent microarray study by Tchagang et al. (2008) reveal increased annexin A2 gene expression in ovarian cancer tissues when compared with normal ovarian tissue [89]. Moreover, several proteomic studies have also identified annexin A2 in the secretome of human ovarian cancer cell lines [90, 91]. We have recently explored the interaction between ovarian cancer and peritoneal cells using an in vitro co culture system [92]. We compared proteins in the secretome of ovarian cancer (OVCAR-5) and peritoneal cells (LP-9) cultured separately (Figure 1.2a) and co-cultured together (Figure 1.2b). We identified annexin A2 to be up-regulated in the secretome of co-cultured ovarian cancer and peritoneal cells by 2D-gel electrophoresis and mass spectrometry (Figure 1.2b). Furthermore, we have recently reported increased plasmin levels during the co culture of ovarian cancer and peritoneal mesothelial cells in vitro [92]. Preliminary findings have demonstrated high annexin A2 immunostaining in serous ovarian cancer tissues compared with benign serous cystadenomas (Figure 1.3). Our studies suggest that increased plasmin production and increased annexin A2 secretion is part of a tumour-host signal pathway between ovarian cancer and peritoneal cells which promotes ovarian cancer metastasis. Ongoing studies in our laboratory are currently investigating the importance of annexin A2 in ovarian cancer metastasis and its potential as a novel therapeutic target for this malignancy.

1.10 Annexin A2 as a potential therapeutic target in cancer Several research groups have targeted annexin A2 to inhibit cancer progression and metastasis. Monoclonal antibodies against annexin A2 have been reported to be an effective therapeutic approach against Lewis lung carcinoma xenografts [93]. Jacovina et al. (2001) reported that a polyclonal antibody against the N-terminal domain of annexin A2 was able to block plasmin production in rat adrenal pheochromocytoma (PC-12) cells in vitro [94]. Annexin A2 has also been shown to bind to angiostatin, a powerful anti-angiogenic molecule that is generated from plasminogen processing [95]. Tuszynski et al. (2002) reported that angiostatin binds to the lysine binding domain of annexin A2 in endothelial cells results in anti-angiogenic effect [95]. Moreover, the interaction between annexin A2 and angiostatin resulted in reduced plasmin generation in Lewis lung carcinoma cells [93]. These findings suggest that plasminogen and angiostatin bind to the same annexin A2 binding site and antiangiogenic action of angiostatin may be mediated via interactions with annexin A2 [95]. The identification of angiostatin-like compounds that block

annexin A2 binding to plasminogen could be promising cancer therapeutics. More recently, Braden et al. (2009) showed that polymeric nanoparticles combined with annexin A2 siRNA vector to allow long term annexin A2 silencing could inhibit prostate cancer growth in mice [96]. Annexin A2 has also been identified as a molecular target for TM601 (a peptide with tumour-targeting and anti- angiogenic effects) in glioma, lung and pancreatic cancer cells [97]. Together these studies indicate that annexin A2 has a great potential as therapeutic target to inhibit cancer progression and metastasis.

1.11 Conclusions and future directions In conclusion, annexin A2 has been shown to play an important role in cancer cell migration, invasion, adhesion and angiogenesis processes which are essential for cancer metastasis. Emerging evidence indicates that annexin A2 plays an important role in the plasminogen activator system. The interaction between annexin A2 and t-PA mediates the conversion of plasminogen to plasmin which facilitates ECM degradation leading to enhanced cancer cell migration and invasion (Figure 1.4). Annexin A2 over-expression and proteolytic cleavage has also been demonstrated in different malignancies. The processing of annexin A2 by MMPs and other proteases results in several annexin A2 isoforms but no studies so far have investigated the significance of these isoforms in tumour progression. Additional studies are needed to further explore the interactions between annexin A2 and its binding proteins and the importance of post-translational modifications of annexin A2 to promote tumour invasion and metastasis. A greater understanding of these mechanisms could potentially lead to the development of novel therapeutics to inhibit annexin A2 function and tumour progression.

1.12 Acknowledgement This research was supported by the Ovarian Cancer Research Foundation (OCRF).

Figure 1.1 Annexin A2 domain structure.

The N-terminal domain consists of the p11 proteins and tissue plasminogen activator (t-PA) binding sites. The C-terminal domain consists of four repeating domains which each contains annexin A2 consensus sequence and the biding sites for calcium, phospholipids and F-actin.

Figure 1.2 2D-gel electrophoresis of proteins in co-cultured ovarian cancer and peritoneal cell.

A. 2D-gel electrophoresis of proteins in the secretome of ovarian cancer (OVCAR-5) and peritoneal cells (LP-9) cultured separately. B. 2D-gel electrophoresis of proteins in the secretome of OVCAR-5 and LP-9 co-culture. Circled spots that were up-regulated in the ovarian cancer- peritoneal co culture but absent in the single culture were identified to be annexin A2 by mass spectrometry.

Figure 1.3 Annexin A2 expression in human ovarian tumour tissues.

A. Benign serous cystadenoma B. Serous ovarian carcinoma. Annexin A2 immunohistochemistry using mouse anti-annexin A2 (BD Biosciences, 1/500), with a citrate buffer microwave retrieval. Magnification bar = 100 µm for all images.

Figure 1.4 Proposed mechanism of annexin A2 promoting cancer metastasis in the plasminogen activation system.

Annexin A2 heterotetramer on the cell surface binds to t-PA and activates plasminogen conversion to plasmin. Plasmin results in activation of MMPs and lead to ECM degradation. Increased annexin A2 expression results in increased plasmin generation and enhances cancer invasion and metastasis.

Table 1.1 Annexin A2 expression in cancer.

Cancer Annexin A2 Expression Ref Pancreatic cancer Over-expression of cell surface annexin A2 in pancreatic cell lines [10] Five human pancreatic cell lines had 5-15 fold higher levels of annexin A2 [6] compared with normal cells Over-expression of annexin A2 on surface of gemcitabine resistant [81] pancreatic cells Renal cell carcinoma Annexin A2 expression correlates with tumour differentiation and clinical [98] outcome Increased annexin A2 expression in metastatic tumour of clear cell renal [29] carcinoma compared with the primary tumour Prostate cancer Annexin A2 expression is lost in prostate cancer tissues [35] Expression annexin A2 is reduced in prostate cancer cells [36] Down-regulation of annexin A2 at mRNA and protein level in metastatic [37] prostate cancer Reduced annexin A2 levels in high grade tumour [38] Increased expression of annexin A2 in high grade tumour and metastatic [39] cell lines Breast cancer Increased expression of annexin A2 in invasive breast cancer cell lines [7] compared with non-invasive breast cancer cells Annexin A2 is a potential predictor of response to neoadjuvant [18] chemotherapy in breast cancer Head and neck cancer Up-regulation of annexin A2 in metastatic cancer cells compared to cells [40] derived from primary tumours Down-regulation of annexin A2 expression in moderately and poorly [41] differentiated cancer Esophageal Carcinoma Decreased expression of annexin A2 in esophageal cancer cell lines and [99] cancer tissues Angiosarcoma Increased annexin A2 in benign and malignant tissues [100] Colorectal cancer Annexin A2 over-expressed in colorectal cancer tissues [101] Brain Increased annexin A2 expression in high grade glioma cells [33, 102, 103] Oral squamous carcinoma Increased annexin A2 protein and mRNA levels in cancer cells and tissues [79, 104] Cervical carcinoma Up-regulated annexin A2 in tumour tissue compared with non-tumourous [105, 106] tissue Acute myeloid leukaemia Up-regulation of annexin A2 at the mRNA level in promyelocytic cell line [107] compared with the early myeloid cell line

Table 1.2 Summary of the effects of annexin A2 on cancer cells.

Function Observations Ref Motility Enhanced migration of hepatocellular carcinoma cells in vitro via annexin A2 [76] interaction with HAb18G/CD187 Annexin A2 binds to S100A6 protein enhances cell migration in pancreatic cancer [61] cells The interaction of annexin A2 and t-PA induced cell migration in human breast cancer [66] cell line Annexin A2 expression is vital in prostate cancer cell migration [36] Invasion Annexin A2 mediates plasmin generation in human breast cancer cells which [7] enhances cell invasion Interaction of annexin A2 and t-PA promotes pancreatic cancer cell invasion in vitro [10] Knockdown of annexin A2 expression in pancreatic cancer cells inhibits cell invasion [32] Annexin A2 promotes invasion of hepatocellular carcinoma via interaction with [76] HAb18G/CD187 in vitro Annexin A2 enhances invasiveness of multidrug resistant breast cancer cell lines [108] Adhesion Annexin A2 regulates the adhesion of prostate cancer cells [11] Annexin A2 regulates cell adhesion of RAW 117 cells (large cell lymphoma cells) to [12] the endothelial cells Phosphorylation of annexin A2 mediates baby hamster kidney cell adhesion [109] Apoptosis Annexin A2 plays a role in p53 induced apoptosis [110] Annexin A2 cleavage leads to cell apoptosis [86] Proliferation Annexin A2 affects proliferation of multiple myeloma cells [71] Annexin A2 regulates proliferation of dasatinib sensitive melanoma cells [82] Knockdown of annexin A2 down regulates p11 proteins expression which decreases [47] proliferation of invasive breast cancer cell lines Annexin A2 induces proliferation of hepatocytes [111]

1.13 Statement of hypotheses, aims and research plan Annexin A2 is one of the proteins modulated in the conditioned media of co-cultured serous ovarian cancer cells (OVCAR-5) and peritoneal cells (LP-9), identified by mass spectrometry and 2D gel electrophoresis. Annexin A2, a multifunctional calcium phospholipid binding protein has been characterized in many malignancies and reported to play an important role in tumourigenesis. Annexin A2 exists on the cell surface as annexin A2 heterotetramer with its binding protein, S100A10. To date, annexin A2 expression and its functional role has not been characterized in ovarian cancer. This study aimed to characterize annexin A2 expression and investigate the functional role of annexin A2 in serous ovarian cancer using in vitro and in vivo ovarian cancer models. We also investigated whether annexin A2 can be used as a diagnostic marker serous ovarian cancer and assessed the efficacy of annexin A2 together with ovarian cancer biomarker, CA125 in serous ovarian cancer patients. Moreover, annexin A2 and S100A10 expression were examined in a uniform cohort of stage III serous ovarian cancers to investigate its potential role as a prognostic marker for ovarian cancer.

1.13.1 Hypotheses 1. Annexin A2 expression plays an important role in ovarian cancer tumourigenesis 2. Annexin A2 inhibitors decrease ovarian cancer cell motility, invasion and adhesion to the peritoneal cells in vitro and in vivo 3. Annexin A2 can be used as a tumour marker for ovarian cancer diagnosis and prognosis

1.13.2 Aims 1. To characterize annexin A2 expression in human serous ovarian cancer tissues, serous ovarian cancer cell lines and primary ovarian cancer cell lines 2. To investigate the effects of annexin A2 inhibitors in ovarian cancer cell motility, invasion and adhesion to the peritoneal cells in vitro 3. To determine whether annexin A2 inhibitors can block ovarian cancer invasion and metastasis in in vivo model 4. To evaluate whether annexin A2 can be used as a tumour marker for ovarian cancer diagnosis 5. To assess whether either annexin A2 or S100A10 can be used as a prognosis marker for ovarian cancer

1.13.3 Research plan Annexin A2 expression in human normal ovarian surface epithelium, benign cystadenomas, borderline ovarian tumours and serous ovarian cancer tissues (stage I to IV) were characterized using annexin A2 immunohistochemistry. Annexin A2 expression was also examined in the omental tissues, matching primary and metastatic ovarian cancer tissues. Real-time PCR and western immunoblotting were performed to assess annexin A2 expression in human serous ovarian cancer cell lines (OVCAR-3, OVCAR-5, OV-90 and SKOV-3), peritoneal cell line (LP-9) and primary ovarian cancer cell lines derived from stage III serous ovarian cancer patients ascites (n=11). To examine the functional role of annexin A2, annexin A2 inhibitors including annexin A2 siRNAs and neutralizing annexin A2 antibodies were used in cell motility, invasion and adhesion assay to the peritoneal cells in vitro in OVCAR-3, OVCAR-5, OV-90 and SKOV-3 cells. Moreover, we investigated the invasion of OVCAR-3, OV-90, SKOV-3 and two primary ovarian cancer cell lines using the CAM model in vivo. In this study, we also examined the effects of neutralizing annexin A2 antibody using SKOV-3/GFP Luc cells in an intraperitoneal mouse model. Tumour development and metastasis in the nude mice were evaluated using bioluminescence imaging over a period of 36 days (n=5 per treatment group). Western immunoblotting and commercially available annexin A2 ELISA were used to examine annexin A2 urine levels in serous ovarian cancer patients and control patients. We also utilized annexin A2 ELISA to examine blood annexin A2 levels in serous ovarian cancer patients (n=86) and controls which are patients with normal ovaries (n=27) and benign ovarian tumours (n=30). We performed ROC curve analysis to determine the sensitivity and the specificity of annexin A2 and CA125 for serous ovarian cancer detection. We also investigated the expression of annexin A2 and S100A10 using immunohistochemistry in a cohort of stage III serous ovarian cancer patients (n=91) using TMAs. Kaplan Meier and Cox Regression analyses were performed to investigate if either annexin A2 or S100A10 is associated with progression-free survival and overall survival. This study assessed whether annexin A2 expression plays a role in ovarian cancer metastasis and can be used as a novel diagnosis and prognosis marker for serous ovarian cancer.

2 Chapter 2: Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis

Noor A Lokman1, Alison SF Elder1, Carmela Ricciardelli1 and Martin K Oehler1, 2

Received: 2 July 2012, in revised form: 27 July 2012, Accepted: 2 August 2012, Published online: 10 August 2012

International Journal of Molecular Science (2012) 13: 9959-9970

STATEMENT OF AUTHORSHIP Chapter 2: Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis

Noor Alia Lokman (candidate) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Performed analysis on all samples presented, interpreted the data and wrote the manuscript.

Certification that the statement of contribution is accurate

Signed: …………………………………………………………………………Date:…………………..

Alison S.F. Elder (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Interpreted the data and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Carmela Ricciardelli (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work, interpreted the data and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Martin K. Oehler (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work, manuscript evaluation and acted as corresponding author.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

2.1 Abstract The majority of ovarian cancer patients present with advanced disease and despite aggressive treatment, prognosis remains poor. Significant improvement in ovarian cancer survival will require the development of more effective molecularly targeted therapeutics. Commonly, mouse models are used for the in vivo assessment of potential new therapeutic targets in ovarian cancer. However, animal models are costly and time consuming. Other models, such as the chick embryo chorioallantoic membrane (CAM) assay, are therefore an attractive alternative. CAM assays have been widely used to study angiogenesis and tumour invasion of colorectal, prostate and brain cancers. However, there have been limited studies that have used CAM assays to assess ovarian cancer invasion and metastasis. We have therefore developed a CAM assay protocol to monitor the metastatic properties of ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) and to study the effect of potential therapeutic molecules in vivo. The results from the CAM assay are consistent with cancer cell motility and invasion observed in in vitro assays. Our results demonstrate that the CAM assay is a robust and cost effective model to study ovarian cancer cell metastasis. It is therefore a very useful in vivo model for screening of potential novel therapeutics.

Keywords: ovarian cancer; invasion; metastasis; chick chorioallantoic membrane

2.2 Introduction Ovarian cancer is the most lethal gynaecological malignancy. Most patients are diagnosed at an advanced stage when the cancer cells have already metastasized to the abdominal cavity. Ovarian cancer metastasis is characterized by the ability of ovarian cancer cells to detach from the ovary and to adhere and invade the peritoneal cell layer, which lines the organs in the abdominal cavity [112]. The development of new therapies with the aim to disrupt ovarian cancer metastasis requires the in vivo study of novel targets and molecules. However, commonly used murine models are costly and require a large number of animals as well as a long experimental time frame. An attractive alternative is the chick chorioallantoic membrane (CAM) assay. CAM assays have been widely used to study angiogenesis [113], tumour cell invasion and metastasis [114-117]. The CAM model has many advantages, such as (a) the highly vascularized nature of the CAM greatly promotes the efficiency of tumour cell grafting; (b) high reproducibility; (c) simplicity and cost effectiveness, and finally (d) as the CAM assay is a closed system, the half- life of many experimental molecules such as small peptides tends to be much longer in comparison to animal models, allowing experimental study of potential anti-metastatic compounds that are only available in small quantities [115, 118]. The CAM is composed of a multilayer epithelium; the ectoderm at the air interface, mesoderm (or stroma) and endoderm at the interface with the allantoic sac [119]. Furthermore, the CAM contains extracellular matrix proteins (ECM) such as fibronectin, laminin, collagen type I and integrin ανβȜ [120]. The presence of these extracellular matrix proteins mimics the physiological cancer cell environment. Although the CAM assay is a well-established model for studying tumour angiogenesis and invasion in malignancies such as bowel cancer [121, 122], glioma [123-125], prostate cancer [126- 128], leukemia [129] and osteosarcoma [130], there has only been one study to date that has used a CAM assay to assess ovarian cancer invasion and metastasis [131]. We recently investigated the ovarian cancer-peritoneal cell interaction and identified several novel proteins that may be involved in ovarian cancer metastasis [92, 132]. To effectively determine their function, we developed a CAM assay protocol using a range of ovarian cancer cell lines to allow the monitoring of candidate molecules on ovarian cancer cell invasion in vivo. The in vivo CAM assay data was compared with results from in vitro assays.

2.3 Materials and methods

2.3.1 Cell culture The human ovarian cancer cell lines OVCAR-3, SKOV-3 and OV-90 were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). All ovarian cancer cell lines were maintained in RPMI 1640 medium supplemented with 4 mM L-glutamine, antibiotics (100 U penicillin G, 100 µg/ml streptomycin sulfate and 0.25 µg/ml amphotericin B), (Sigma-Aldrich, St. Louis, MO, USA). OVCAR-3 and SKOV-3 cells were supplemented with 5% fetal bovine serum (FBS) (Sigma-Aldrich) and OV-90 cells were supplemented with 10% FBS. All cell lines were maintained at 37⁰C in an environment of 5% CO2.

2.3.2 Cell motility and invasion assays Cell motility and invasion assays were performed as previously described [133]. Briefly, the ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) were diluted to a cell concentration of 1 × 106 cells per ml and labeled with calcein-AM (1 µg/ml, Life Technologies, VIC, Australia). Ovarian cancer cells were mixed at room temperature for 1 hour in the dark on an oscillating platform. Ovarian cancer cells (50 µl, 40,000 cells) were loaded on top of uncoated 12 µm filter inserts (Disposable, 96-well plate, ChemoTx, Neuro Probe Inc, Gaitherburg, MD, USA) for motility assays or 12 µm filters coated with Geltrex (0.6 µl per well, 9 mg/ml, Life Technologies) for invasion assays. 10% FBS RPMI media was used as a chemo attractant. The cells were allowed to migrate and invade to the lower chamber for 6 hours. Non-migratory cells on the top of the filter were removed and the fluorescence was measured at 485-520 nm.

2.3.3 Chick chorioallantoic membrane assay (CAM Assay) Fertilized white leghorn chicken eggs were obtained from Hi-Chick, South Australia, Australia. Eggs were incubated in a MultiQuip Incubator (E2) at 37⁰C with 60% humidity. Ethics approval was obtained by the University of Adelaide Animal Ethics Committee. A small window was made in the shell on day 3 of chick embryo development under aseptic conditions. The window was resealed with adhesive tape and eggs were returned to the incubator until day 11 of chick embryo development. On day 11, OVCAR-3, SKOV-3 and OV-90 cell suspensions (9 × 105) were mixed with growth factor reduced matrigel (8.9 mg/ml, BD Biosciences, NSW, Australia) in a total volume of 30 µl. Control anti-mouse IgG (20 µg/ml) (Sigma Aldrich) and neutralizing antibody to protein A (20 µg/ml) (BD Biosciences) were mixed together with the OV-90 cancer cells and matrigel. Matrigel grafts were placed on top of the CAM and eggs were resealed and returned to the

incubator for 72 hours until day 14 (n=6 chicken embryos per cell line). Matrigel grafts with surrounding CAM were harvested from each embryo and fixed with 4% paraformaldehyde for 24 hours and embedded in paraffin. Serial sections (6 µm) were stained with haematoxylin and eosin. Slides were digitally scanned using the NanoZoomer (Hamamatsu Photonics K.K., Japan).

2.3.4 Immunohistochemistry CAM paraffin sections (6 µm) were incubated on a heat plate at 60⁰C for 2 hours. Tissue sections were dewaxed with xylene and ethanol, followed by PBS washes. Antigen retrieval was performed by using 1% protease (Sigma-Aldrich) for 10 minutes on a heat plate at 37⁰C. The endogenous peroxidase activity of the sections was quenched with 0.3% H2O2. Each tissue sample was blocked with 5% goat serum for 30 minutes before incubation with monoclonal mouse anti-human cytokeratin clone AE1/AE3 (1/50, Dako, VIC, Australia) at 4⁰C overnight. Subsequently, the tissues sections were incubated sequentially with biotinylated goat anti-mouse (1/400, Dako), followed by streptavidin-HRP conjugated (1/500, Dako) for 1 hour at room temperature. Immunoreactivity was detected using diaminobenzidine/H2O2 substrate (Sigma-Aldrich). The sections were counterstained with 10% haematoxylin (Sigma-Aldrich), dehydrated and mounted in Pertex (Medite Medizintechnik, Germany). Slides were digitally scanned using the NanoZoomer (Hamamatsu Photonics K.K.). For quantitative analysis of ovarian cancer cell invasion into the mesoderm layer, 8 to 12 CAM images from each embryo were assessed by two independent researchers [127].

2.3.5 Statistical analysis All analyses were performed using SPSS 15.0 for Windows software (SPSS, Chicago, IL, USA). The student’s t-test and one-way ANOVA were used to determine statistical significance of ovarian cancer cell motility and invasion in vitro, and ovarian cancer cell invasion in the CAM model. Statistical significance was accepted at P < 0.05.

2.4 Results

2.4.1 Human ovarian cancer cell motility and invasion in vitro We compared the motility and invasion of three ovarian cancer cells (OVCAR-3, SKOV-3 and OV- 90) using in vitro assays. We found OV-90 cells to be most invasive through an extracellular matrix and migrated faster through 12 µm pores towards a chemo attractant, compared to SKOV-3 and OVCAR-3 cells (Figure 2.1). OVCAR-3 cells were the least motile and invasive cell line in our study.

2.4.2 Human ovarian cancer cell invasion into the chick chorioallantoic membrane (CAM) We initially utilized an ex ovo method and incubated the chick embryos in plastic weigh boats as described previously [134]. The ex ovo method has the advantage of allowing the application of a larger number of matrigel grafts as a wider area of the CAM is accessible. However, the survival rate for the ex ovo method was very low and only 10% of embryos survived to day 14. The in ovo method had a survival rate of 70% on day 14. In the in ovo method, a small window is made in the shell on day 3 of chick embryo development to detach the CAM layer from the egg shell (Figure 2.2a). Ovarian cancer cells (9 x 105 cells) were mixed with matrigel to form a gel and grafted on top of the CAM of day 11 chick embryos. The chick embryos were incubated with the matrigel grafts until day 14 of development. An example of a matrigel graft on day 14 is shown in Figure 2.2b. The CAM layers; ectoderm (ET), mesoderm (M) and endoderm (ED) can be seen in Figure 2.3a. Cytokeratin immunohistochemistry was used to identify the CAM layer integrity and presence of ovarian cancer cells in the mesodermal layer. The invasion of the ovarian cancer cells through the ectoderm into the mesoderm was assessed on day 14 of chick embryo development. OVCAR-3 cytokeratin immunohistochemistry showed some damage to the ectoderm layer but minimal invasion into the CAM mesoderm (Figure 2.3b). The SKOV-3 cells showed invasion into the mesoderm layer and minimal destruction of the ectoderm layer (Figure 2.3c). OV-90 cells were the most invasive cells in the CAM assay (Figure 2.3d), which agrees with the results of the in vitro assays (Figure 2.1). Figure 2.3d shows a large number of OV- 90 cells invading into the mesoderm of the CAM, as well as the destruction of the CAM ectoderm layer. Quantitative analysis of ovarian cancer cell invasion into the CAM was performed by determining the number of images (8 to 12 sections per chick embryo) with cancer cell invasion into the CAM

mesoderm of day 14 chick embryos. Our results showed a significantly higher invasion of OV-90 and SKOV-3 cells into the CAM mesoderm, when compared with OVCAR-3 cells (Figure 2.4).

2.4.3 The effects of protein A neutralizing antibody on OV-90 cancer cell invasion into the CAM We used a neutralizing antibody against one of the novel proteins identified in our previous study that investigated ovarian cancer-peritoneal cell interactions [92, 132]. OV-90 cancer cells (9 x 105 cells) were mixed with matrigel and the control anti-mouse IgG or the neutralizing antibody against protein A before grafting onto day 11 chick embryos. Neutralizing antibody against protein A effectively inhibited OV-90 cancer cell invasion into the mesoderm of the CAM, compared with the control anti-mouse IgG, where OV-90 cancer cells invaded the mesoderm of the CAM and a destruction of the ectoderm layer was observed (Figure 2.5).

Figure 2.1 Motility and invasion of human ovarian cancer cell lines (OVCAR-3, SKOV-3 and OV-90) in vitro.

The fluorescence reading represents the cancer cells that have migrated through the pores or invaded through the extracellular matrix (Geltrex). Data represents the mean ± SEM from two independent experiments performed in quadruplicate. (a) Indicates significant difference from OVCAR-3 cells; and (b) indicates significant difference from SKOV-3 cells, P < 0.05.

Figure 2.2 Chick embryo development.

A. Day 3 chick embryo B. Ovarian cancer cells and matrigel graft on the chick chorioallantoic membrane (CAM) on day 14.

Figure 2.3 Invasion of ovarian cancer cells in the chick chorioallantoic membrane (CAM).

A. Control showing the normal structure of the CAM layers; ectoderm (ET), mesoderm (M) and endoderm (ED); OVCAR-3 (B); SKOV-3 (C); and OV-90 (D) cancer cell matrigel grafts (CM) were placed on top of the ectoderm layer and cancer cell invasion into the CAM mesoderm was assessed in day 14 chick embryos. CAM paraffin sections (6 µm) were immunostained with cytokeratin antibody. Original magnification: x200. Magnification bar = 100 µm for all images.

Figure 2.4 Chick chorioallantoic membrane (CAM) invasion by ovarian cancer cells in day 14 chick embryos.

Data generated from 48–60 images from 6 chicken embryos per cell line. Data represents the percentage of images with invasion into the mesoderm, mean ± SEM from two independent experiments. (a) Indicates significant difference from OVCAR-3 cells; and (b) indicates significant difference from SKOV-3 cells, P < 0.05.

Figure 2.5 Effects of protein A neutralizing antibody on OV-90 cancer cell invasion into the chick chorioallantoic membrane (CAM).

OV-90 cancer cells were mixed with matrigel and (A) control anti-mouse IgG (20 µg/ml); or (B) neutralizing antibody against protein A (20 µg/ml). CAM paraffin sections (6 µm) were immunostained with a pan-cytokeratin antibody. ET = ectoderm. M = mesoderm. ED = endoderm. CM = cancer cell matrigel grafts. Original magnification: x200. Magnification bar = 100 µm for all images.

2.5 Discussion The CAM assay is a frequently applied model to study ovarian cancer angiogenesis [135-138]. However, there is only one study which has used CAM assays to assess ovarian cancer cell invasion and metastasis [131]. Chang et al. described IGROV-1 ovarian cancer cell invasion and metastasis to the posterior CAM and lungs of chick embryos [131]. We have developed a CAM assay protocol to monitor the metastatic properties of ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) and have successfully used it to study the effect of newly identified molecules in vivo. Our results show that the CAM assay is an effective model to study ovarian cancer metastasis. Importantly, our CAM model closely mimics the mode of ovarian cancer metastasis which involves ovarian cancer cell attachment and invasion into the peritoneum. The ectodermal layer of the CAM has many similarities with the peritoneum, which consists of a single layer of mesothelial cells covering the organs in the abdominal cavity. We observed a higher survival rate with the in ovo method in comparison to the ex ovo method for monitoring of ovarian cancer cell growth in the chick embryos. Various methods have been used to graft cancer cells in the CAM model; such as collagen onplants [134], plastic rings [130], and matrigel grafts [139]. Furthermore, cancer cells can also be inoculated by dropping the cell suspension on top of the CAM [140], or administered intravenously to study metastasis of cancer cells in the chick embryos [129]. Matrigel is one of the most suitable scaffolds used for implantation and grafting of cancer cells onto the CAM. In our model, ovarian cancer cells and matrigel were mixed with or without neutralizing antibodies before grafting onto the CAM of the chick embryos to assess ovarian cancer cell invasion. The grafting of the matrigel in the CAM model allows continuous visualization of the test site. Moreover, other studies have reported visible and solid tumours on the CAM of chick embryos a few days after cancer cell inoculation [122, 130]. We used histological assessment by means of a pan-cytokeratin antibody, to allow the visualization of cancer cells invading into the mesoderm. The CAM model has been previously employed to assess cancer metastasis [114, 140]. In some studies quantitative alu PCR was used to assess the presence of metastatic human cancer cells in chick embryo organs [141]. Several studies have compared both CAM assays and mouse models to assess tumour growth and metastasis. Colorectal cancer cells were reported to colonize the CAM similarly to the mouse model [121]. Strojnik et al. conducted a study to compare the histological and immunohistochemical characteristics of glioma tumour protein expression in the CAM and an established rat model. They reported a similar profile of proteins expressed in both

models [123]. In addition, the CAM model has also been used concurrently with the nude mice model to assess tumour growth of fibrosarcoma (HT1080 cells) and human squamous carcinoma (Hep3 cells) cells [142]. Lyu et al. also showed that over expression of urokinase-type plasminogen activator receptor (u-PAR) in Hep3 cells leads to an increase in cancer cell invasion in the CAM model as well as accelerated tumour growth in the SCID mice model [143]. These studies demonstrate the validity of the CAM model for in vivo analysis of cancer cell invasion and metastasis. The CAM model has many advantages. It is cost effective, allows large scale screening and is an easily reproducible in vivo model [115, 118]. A comparison of the advantages and limitations of the CAM against the mouse model are summarized in Table 2.1. The CAM model has also been used in pre-clinical screening to assess the efficacy of drugs and inhibitors on tumour growth. Hagedorn et al. reported that treatment of human glioma cells with receptor tyrosine kinase inhibitors inhibited tumour growth in a CAM model [125]. Bekes et al. demonstrated that treatment of prostate cancer (PC3 cells) with u-PA activation blocking antibody mAb-112 significantly inhibited cancer cell invasion in the CAM model [144]. Additionally, the CAM model has been used to test the efficacy of chemotherapy agents (such as doxorubicin) in human leukemia cell lines and has been shown to reduce cancer cell growth in the CAM [129]. An important limitation of the CAM model is the inability to assess cancer–immune cell interactions. Examination of cancer-immune cell interactions requires the use of transgenic ovarian cancer models [145], however, these models are not widely available, are not suitable for high throughput screening, and cannot be used with primary ovarian cancer cells derived from clinical samples. The CAM assay is therefore an attractive model to rapidly assess the effectiveness of novel candidate therapeutic drugs and the in vivo inhibition of specific tumour types and subtypes in one consistent model. We have shown that OV-90 ovarian cancer cells invade into the mesoderm of the CAM within three days of implantation, therefore making the OV-90 CAM model ideal for studying ovarian cancer invasion and metastasis. In contrast, the OVCAR-3 cells showed limited invasion in the CAM over the three days of our assays, and would therefore be suitable for studying the role of molecules that promote ovarian cancer invasion. In conclusion, the CAM model provides a high throughput in vivo model for the assessment and evaluation of candidate pro-invasive molecules as well as potential therapeutic targets for ovarian cancer.

Table 2.1 Comparison of the advantages and limitations of the chick chorioallantoic membrane (CAM) and mouse model.

In vivo Advantages Reference Limitations Reference Model Short experimental assay (days) Short observation period (days) Cannot examine cancer-immune Inexpensive cell interactions Easily reproducible and high throughput Rapid morphological changes [115] Closed system - allows assessment of Limited antibodies to chicken tissues [118] small quantities of therapeutic agents for characterization Naturally immunodeficient [118] CAM Multiple tests per individual CAM Allows large scale screening [115] Biology and physiology well known Availability of in vivo imaging [114] Allows direct visualization [119] Animals do not have to be restrained Can be used with primary human cell lines Longer observation period Long experimental length

(weeks to months) (months to years) Biology and physiology well known, [115] Costly but also complex [115] Mouse Availability of in vivo imaging [146] Mature immune system Defined genetic background [147] Reproducibility is expensive Large number of animals required

Animals have to be restrained [119]

2.6 Conclusions The CAM assay is a robust technique that can be used to monitor invasion of ovarian cancer cell lines and to assess the role of novel molecules and potential therapeutic targets. It is a valuable alternative to murine in vivo models for the study of ovarian cancer invasion and metastasis.

2.7 Acknowledgments The grant sponsors for this project were the Ovarian Cancer Research Foundation (OCRF), Australia the Cancer Council South Australia and South Australian Health and Medical Research Institute. We thank A/Prof Simon Koblar from the School of Medicine, University of Adelaide, Australia for his help with setting up the CAM assay.

3 Chapter 3: Annexin A2 is regulated by Ovarian Cancer-Peritoneal Cell Interactions and Promotes Metastasis

Noor A Lokman1, Alison SF Elder1, Miranda P Ween2, Carmen E Pyragius1, Peter Hoffmann3, Martin K Oehler1, 4 and Carmela Ricciardelli1

1 Discipline Robinson Institute, Research Centre for Reproductive Health, School of Paedriatrics and Reproductive Health, University of Adelaide, Adelaide, SA 5005, Australia 2 Research Centre for Infectious Diseases, School of Molecular Biosciences, University of Adelaide, Adelaide, SA 5005, Australia 3 Adelaide Proteomics Centre, School of Molecular and Biomedical Science, University of Adelaide, Adelaide, SA 5005, Australia 4 Department of Gynaecological Oncology, Royal Adelaide Hospital, Adelaide, SA 5005, Australia

Received: 25 June 2013, Accepted: 12 July 2013, Published online: 14 July 2013

Oncotarget (2013) 4:1199-1211

STATEMENT OF AUTHORSHIP Chapter 3: Annexin A2 is regulated by Ovarian Cancer-Peritoneal Cell Interactions and Promotes Metastasis

Noor Alia Lokman (candidate) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Performed analysis on all samples presented, interpreted the data and wrote the manuscript.

Certification that the statement of contribution is accurate

Signed: …………………………………………………………………………Date:…………………..

Alison S.F. Elder (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Performed the intraperitoneal xenograft mice study, interpreted the data and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Miranda P. Ween (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Manuscript evaluation

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Carmen E. Pyragius (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Performed immunohistochemistry for the mouse tissues and manuscript evaluation

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Peter Hoffmann (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Martin K. Oehler (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work and manuscript evaluation.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

Carmela Ricciardelli (co-author) Statement of contribution (in terms of conceptualization of the work, its realization and its documentation)

Supervised development of work, interpreted the data, manuscript evaluation and acted as corresponding author.

Certification that the statement of contribution is accurate and permission is given for the inclusion of the paper in the thesis

Signed: …………………………………………………………………………Date:…………………..

A Lokman, N.A., Elder, A.S.F., Ween, M.P., Pyragius, C.E., Hoffmann, P., Oehler, M.K. & Ricciardelli, C. (2013) Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis. Oncotarget, v. 4(8), pp. 1199-1211

NOTE: This publication is included on pages 69-106 in the print copy of the thesis held in the University of Adelaide Library.

4 Chapter 4: Annexin A2 as a Diagnostic Marker for Serous Ovarian Cancer

Noor A Lokman1, Andrew N Stephens2, Martin K Oehler1, 3 and Carmela Ricciardelli1

1 Discipline Robinson Institute, Research Centre for Reproductive Health, School of Paedriatrics and Reproductive Health, University of Adelaide, Adelaide, SA 5005, Australia 2 Prince Henry's Institute for Medical Research, Monash Medical Centre, Clayton, VIC 3168, Australia 3 Department of Gynaecological Oncology, Royal Adelaide Hospital, Adelaide, SA 5005, Australia

Manuscript in Preparation

107

4.1 Abstract Up to 70% of ovarian cancers are diagnosed at advanced stage due to unavailable screening methods and limited biomarkers to detect early stage disease. Our group recently identified annexin A2 to be secreted in the conditioned media of co-cultured ovarian cancer and peritoneal cells in vitro. Annexin A2, a calcium phospholipid binding protein plays an important role in serous ovarian cancer metastasis. This study examined the value of annexin A2 as a diagnostic biomarker in serous ovarian cancer. Annexin A2 levels were measured using an ELISA in urine samples from women with normal ovaries (n=6) and benign ovarian tumours (n=19) and serous ovarian cancer (n=36) and in blood samples from women with normal ovaries (n=27) and benign ovarian tumours (n=30) and serous ovarian cancer patients (n=86). ROC curves were generated with annexin A2 alone and in combination with the ovarian cancer marker, CA125. Annexin A2 and CA125 levels were also analyzed together using logistic regression models. There was no difference in urine annexin A2 levels of serous ovarian cancer patients compared with controls. However, blood annexin A2 levels were increased in early stage and advanced stage ovarian cancers compared with controls (P = 0.0001). Using a cutoff point of 300 ng/ml for annexin A2 and the established cutoff point of 35 U/ml for CA125, we observed an improved sensitivity of 100% and specificity of 70% when annexin A2 and CA125 measurements were combined in the controls and early stage ovarian cancer samples. By logistic regression analysis in the early stage and benign ovarian tumours, an AUC of 0.857 (95% CI: 0.753-0.960) was observed when annexin A2 and CA125 measurements combined and an improved sensitivity of 56.5% compared to either annexin A2 (21.7%) or CA125 (21.7%) alone at the 95% specificity level. Annexin A2 has potential as a useful biomarker for serous ovarian cancer diagnosis that warrants further investigation.

Keywords: Annexin A2, CA125, ovarian cancer, diagnosis, biomarker

108

4.2 Introduction Epithelial ovarian cancer is the leading cause of death from gynaecological cancer and each year it is responsible for 120,000 deaths worldwide [148]. The majority of epithelial ovarian cancers are of serous histology, followed by clear cell, endometriod, mucinous and mixed cell types [163]. The International Federation of Gynaecologist and Obstetricians (FIGO) characterized the staging of ovarian cancer as stage I - growth limited to the ovaries, stage II - growth involving one or both ovaries with pelvic extension, stage III - tumour growth with peritoneal implants in the upper abdomen and stage IV - distant metastasis to organs such as the liver, lung or brain [164]. Despite advances in surgery and in chemotherapies, no significant improvement in ovarian cancer survival has been observed over the last two decades [165]. Up to 70% of ovarian cancer patients are diagnosed at advanced stage and to date, there is no effective early detection test for ovarian cancer. CA125 is widely used as a diagnostic tool and for follow up after ovarian cancer treatment [166]. However, while CA125 is elevated in 70% advanced stage disease, it is elevated in less than 50% of early stage ovarian cancer [167] and also increased in patients with benign ovarian lesions [168, 169]. Consequently, CA125 is not a useful tumour marker for early detection of ovarian cancer. The diagnostic performance of an early diagnostic test is based on the specificity, sensitivity, positive predictive value (PPV) and negative predictive value (NPV) in detecting the disease [170]. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial reported that annual screening with CA125 and transvaginal ultrasonography (TVU) compared with standard usual treatment care has no benefit on ovarian cancer mortality rate [171]. Combinations of serum biomarkers which have a high specificity and sensitivity are required for early detection of ovarian cancer and to improve ovarian cancer survival rate. Our group utilized a proteomic approach to explore the molecules modulated in ovarian cancer and peritoneal cells interactions. From 2D gel electrophoresis and mass spectrometry, we identified novel molecules modulated in the secretome of co-cultured ovarian cancer and peritoneal cells in vitro [92] and one of the proteins was annexin A2 [132]. Annexin A2 is a calcium phospholipid binding protein that exists on the cell surface of various tumour cells [5]. We recently reported that annexin A2 plays an important role in ovarian cancer metastasis (Chapter 3) [172]. Several studies have recently investigated the potential use of annexin A2 as a biomarker. Serum annexin A2 levels was found to be elevated in hepatocellular carcinoma [45, 158, 173], breast cancer [174]

109

and multiple myeloma [175]. However, to date, no studies have investigated the potential of using annexin A2 for the diagnosis of ovarian cancer. In this study, we assessed blood annexin A2 levels in serous ovarian cancers and controls and compared its diagnostic efficacy with CA125. We utilized two established methods to evaluate the efficacy of annexin A2 and CA125 in our patient cohort; i) optimal cutoff values determined by the highest accuracy from the ROC curve and ii) logistic regression model combining annexin A2 and CA125 measurements. The diagnostic performances of either annexin A2 or CA125 alone and in combination were compared in non-cancer controls and serous ovarian cancer patients. We also compared the efficacy of the both biomarkers in benign ovarian tumours versus early stage and advanced stage serous ovarian cancers.

110

4.3 Materials and methods

4.3.1 Urine, serum and plasma samples Urine samples were collected from patients with normal ovaries (n=6), patients with benign ovarian tumours (n=19), and serous ovarian cancer patients at time of diagnosis (n=36). Urine samples were centrifuged at 3000 rpm for 10 minutes at room temperature and stored at -80⁰C until assayed. The clinicopathological characteristics for urine cohorts are detailed in Supplementary Table 4.1. Serum samples were collected from patients with normal ovaries undergoing surgery for benign gynaecological conditions (n=27), patients with benign ovarian tumours (n=30), early stage serous ovarian cancer patients at time of diagnosis (stage I and II) (n=13) and advanced stage serous ovarian cancer patients at time of diagnosis (stage III and IV) (n=63). Blood samples were collected into vacutainer plain tubes, the blood was allowed to clot, centrifuged at 3000 rpm for 10 minutes at room temperature and serum was stored at -80⁰C until assayed. An additional 10 early stage serous ovarian cancer plasma samples were obtained from Prince Henry’s Institute. All samples were collected with approval from the Royal Adelaide Hospital or Prince Henry’s Institute Human Ethics Committees and written consent was obtained from all patients involved in this study. The clinicopathological characteristics for the serum and plasma cohorts are detailed in Supplementary Table 4.2.

4.3.2 Urine sample preparation for western blot and annexin A2 ELISA For normalization of the urine samples, urine creatinine values were measured using Hitachi 912 Analysis with a CREA kit (Roche Diagnostics GmbH, Germany). Urine samples were normalized to creatinine values in Tris-buffered saline (0.5 M Tris and 1.5 M sodium chloride, pH 7.4) with the presence of protease inhibitors (Sigma-Aldrich, MO, USA) in a total volume of 2 ml and precipitated using 4-fold volume of cold acetone overnight at -20⁰C, then spun at 1500 rpm for 5 minutes. The protein pellet was washed twice with 80% cold acetone, dissolved in RIPA buffer (1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M sodium chloride, 50 mM Tris-HCl and 1 mM EDTA, pH 8.0) for western immunoblotting and in 200 μl of PBS (with 0.1% Nonidet P-40, pH 7.4) and sonicated for 30 minutes (Soniclean Pty. Ltd., SA, Australia) before assayed for annexin A2 ELISA.

4.3.3 1D-western immunoblotting Protein concentration of the urine samples was determined using the Bradford assay (Bio-Rad) and samples were stored at -20⁰C until assayed. Western immunoblotting was conducted as

111

previously described [92] with a mouse monoclonal antibody to annexin A2 (1/2000, Clone 5, BD Biosciences) and anti-mouse IgG peroxidase-conjugated secondary antibodies (1/2000, Dako), visualized by enhanced chemiluminescence and autoradiography (ECL Hyperfilm, GE Healthcare).

4.3.4 Annexin A2 ELISA and detection of CA125 Annexin A2 levels were measured using a commercial human annexin A2 ELISA kit as per the manufacturer’s instructions (USCN Life Science Inc., Wuhan, China). Briefly, 100 μl of standards, urine samples and 10-fold dilutions of serum or plasma samples in PBS (pH 7.0) were added into each 96 well in duplicates. The absorbance was measured at 450 nm using the Triad series multimode detector (Dynex technologies, VA, USA). The detection limit of this assay was 0.321 ng/ml and the intra assay and inter assay coefficient was 6.1% and 3.3% respectively. CA125 measurements in the serum were determined using the Siemens Advia Centaur XP automated analyzer by the Institute of Medical Veterinary Science (IMVS) (SA Pathology, Adelaide, SA, Australia).

4.3.5 Statistical analyses Statistical analyses were performed using SPSS for Windows (Version 20.0, SPSS Inc., Chicago, IL, USA), GraphPad Prism for Windows (Version 5.0, La Jolla, CA, USA) and SigmaPlot for Windows (Version 12.5, Systat Software Inc., Chicago, IL, USA). Receiver operating characteristics (ROC) curve were performed for annexin A2 and CA125 measurements with the controls and serous ovarian cancers to obtain the optimal cutoff values. The highest Youden index (J) calculated from equation, J = (sensitivity + specificity -1) was used to calculate the annexin A2 cutoff point. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated as proportion of true positive and true negative patients. The blood measurements for annexin A2 and CA125 were log (natural) transformed due to skewness in the logistic regression analysis and the predictive probability values for annexin A2 and CA125 alone and combined annexin A2 and CA125 were used to perform the ROC curve analyses. The statistical significance between the patient’s diagnosis groups was determined with Kruskal-Wallis test and Dunnett C post-hoc test. Spearman’s correlation test was used to determine the correlation between blood CA125 and annexin A2 levels. SigmaPlot was used to compare area under the curve (AUC) of the ROC curves for annexin A2 and CA125. Statistical significance was accepted at P < 0.05.

112

4.4 Results

4.4.1 Level of annexin A2 in the urine of ovarian cancer patients Urine annexin A2 level of serous ovarian cancer patients (P1 to P4) and benign ovarian tumour patients (P5 to P8) was examined by western immunoblotting after normalization to urine creatinine measurements (Figure 4.1a). Our findings showed strong annexin A2 bands at 37 kDa in the urine of ovarian cancer patients (P1 and P2) and weak annexin A2 bands for P3 and P4. However, we did not observe any annexin A2 bands at 37 kDa in the urine of benign ovarian tumour patients (P5 to P8). Figure 4.1b shows annexin A2 measurements in the urine of patients with normal ovaries, benign ovarian tumours and serous ovarian cancers using an annexin A2 ELISA. We did not observe any significant difference in the annexin A2 levels in the urine of serous ovarian cancer patients compared with the patient’s with benign ovarian tumours or normal ovaries (P = 0.521, Kruskal-Wallis test). Moreover, we also observed no significant difference in urine annexin A2 levels between the early stage (stage I and II) and advanced stage (stage III and IV) ovarian cancer patients (P = 0.711, Kruskal-Wallis test). These findings indicate that urine annexin A2 levels cannot be used to diagnose serous ovarian cancer.

4.4.2 Blood annexin A2 levels are elevated in serous ovarian cancer patients compared with non-cancer controls Blood annexin A2 levels were significantly increased in serous ovarian cancer patients (stage I-IV) compared with patients with normal ovaries and benign ovarian tumours (Figure 4.2a, P < 0.0001). However, no significant difference in annexin A2 levels was observed between advanced stage ovarian cancers patients compared with patients with benign ovarian tumours. We also observed that annexin A2 levels were significantly increased in stage I ovarian cancer patients compared to those with stage III and IV ovarian cancers (Figure 4.2c, P = 0.0001). We performed a ROC curve of the non-cancer controls and serous ovarian cancers to determine the optimal cutoff value for annexin A2 (Figure 4.2e, AUC=0.670, P = 0.001, 95% CI: 0.579-0.760) and the best cutoff point, 300 ng/ml had a sensitivity of 51.2% and a specificity of 78.9%.

4.4.3 CA125 levels in serous ovarian cancer patients The established tumour marker for ovarian cancer, CA125 levels was significantly increased in serous ovarian cancer patients (stage I-IV) compared to patients with normal ovaries and benign ovarian tumours (Figure 4.2b, P < 0.0001). Moreover, a significant increase in CA125 levels was observed in stage III ovarian cancer patients compared with stage I ovarian cancer patients (Figure

113

4.2d, P = 0.0132). Similar to annexin A2, we performed a ROC curve of the non-cancer controls and serous ovarian cancers. Using the established cutoff value for CA125, of 35 U/ml, we observed a sensitivity of 88.4% and specificity of 89.5% (Figure 4.2f, AUC=0.925, P < 0.0001, 95% CI: 0.879-0.972).

4.4.4 Annexin A2 levels are weakly correlated with CA125 and able to distinguish serous ovarian cancer patients with normal CA125 levels from non-cancer controls We observed a weak correlation between annexin A2 and CA125 levels (Figure 4.3a, r = 0.180, P = 0.032, Spearman’s correlation test). Annexin A2 was able to distinguish between the serous ovarian cancers with normal (<35 U/ml) CA125 levels and non-cancer controls (Figure 4.3b, AUC: 0.803, P = 0.004, 95% CI: 0.674-0.932). Moreover, annexin A2 was also able to distinguish between the serous ovarian cancers with normal CA125 levels compared with benign ovarian tumours (Figure 4.3c, AUC: 0.730, P = 0.039, 95% CI: 0.560-0.899).

4.4.5 Annexin A2 and CA125 diagnostic performance in ovarian cancer patients compared with non-cancer controls using specified cut points Using the optimal cutoff values for annexin A2 and CA125, 300 ng/ml and 35 U/ml respectively, we determined the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for elevated annexin A2 (≥300 ng/ml) or elevated CA125 alone (≥35 U/ml) and either elevated annexin A2 or elevated CA125 (Table 4.1). We observed an improvement in the sensitivity of detecting ovarian cancer when annexin A2 and CA125 were combined (98%) compared with either annexin A2 (52%) or CA125 (90%) alone, but no improvement in the specificity. Moreover, we showed no improvement in the PPV when the markers were combined (83%) but a better NPV (95%) compared with annexin A2 (52%) or CA125 (85%) alone (Table 4.1a). We also compared blood annexin A2 and CA125 levels in the non-cancer controls versus early stage ovarian cancers and observed a sensitivity of 100% by combining annexin A2 or CA125 (Table 4.1b). We showed an NPV of 100%, but no improvement in the specificity or PPV when annexin A2 and CA125 measurements were combined (Table 4.1b). Moreover, we also compared the controls with the advanced stage ovarian cancers and observed a small increase in the sensitivity of detecting ovarian cancer by combining annexin A2 and CA125 (97%) compared with annexin A2 (40%) and CA125 (95%) alone. However, we showed no improvement in the specificity or PPV by combining annexin A2 and CA125 measurements but a small improvement in the NPV (Table 4.1c).

114

4.4.6 Annexin A2 and CA125 diagnostic performance in serous ovarian cancers compared with benign ovarian tumour using specified cutoff points We also investigated whether annexin A2 could be a potential biomarker for diagnosing ovarian cancer in patients presenting with pelvic masses. Table 4.2 summarized the sensitivity, specificity, PPV and NPV of annexin A2 and CA125 in benign ovarian tumours compared with serous ovarian cancers. We observed an improved sensitivity by combining annexin A2 and CA125 measurements (98%) compared with annexin A2 (52%) or CA125 (90%) alone, but no improvements in the specificity or PPV. The NPV (89%) was also improved when annexin A2 and CA125 measurements were combined compared to annexin A2 alone (34%) and CA125 alone (73%) (Table 4.2a). When we compared benign ovarian tumours only with early stage ovarian cancers, the sensitivity and NPV of combining annexin A2 and CA125 measured was increased to 100%. We showed no improvement in the specificity or PPV by combining annexin A2 and CA125 measurements (Table 4.2b). In comparison, only a small improvement in sensitivity was observed when annexin A2 and CA125 measurements were combined (97%) in the benign ovarian tumours and advanced stage serous ovarian cancer. However, there was no improvement in the specificity, PPV or NPV of the combine measurements compared with the individual markers alone (Table 4.2c). In summary, we have demonstrated that combining blood annexin A2 and CA125 can improve the sensitivity but not the specificity in discriminating benign ovarian tumours from serous ovarian cancer in patients with ovarian lesions.

4.4.7 Logistic regression model to evaluate the efficacy of combining annexin A2 and CA125 measurements in ovarian cancer patients and non-cancer controls The blood annexin A2 and CA125 levels were log transformed, for the logistic regression model. The ROC curve analysis with non-cancer controls and all ovarian cancer patients (stage I-IV) is shown in Figure 4.4a. The AUC was slightly improved when annexin A2 and CA125 values were combined (AUC=0.94, P < 0.0001, 95% CI: 0.900-0.979) compared with annexin A2 alone (AUC=0.67, P = 0.001, 95% CI: 0.579-0.760) or CA125 alone (AUC=0.93, P < 0.0001, 95% CI: 0.879-0.972). At the 95% specificity level, the combined annexin A2 and CA125 measurements resulted in a sensitivity of 61.6% compared with annexin A2 (17.4%) or CA125 (48.8%) alone. At the 98% specificity level, the sensitivity of combining annexin A2 and CA125 was 44.2% (Table 4.3a). Non-cancer controls were also analyzed with only the early stage ovarian cancers (Figure 4.4b). ROC curve analysis showed an improved AUC when both annexin A2 and CA125 were combined

115

(AUC=0.91, P < 0.0001, 95% CI: 0.835-0.980) compared to either annexin A2 alone (AUC=0.84, P < 0.0001, 95% CI: 0.740-0.943) or CA125 alone (AUC=0.85, P < 0.0001, 95% CI: 0.750-0.951). We observed an improvement in the sensitivity of detecting early stage ovarian cancer at the 95% specificity level when annexin A2 and CA125 were combined (56.5%) compared with annexin A2 (47.8%) or CA125 (26.1%) alone. The sensitivity at the 98% specificity was also increased to 52.2% when annexin A2 and CA125 measurements were combined (Table 4.3b). Figure 4.4c showed the ROC curve of the controls and advanced stage ovarian cancer. We observed a slightly improved AUC when annexin A2 and CA125 measurements were combined (AUC=0.96, P < 0.0001, 95% CI: 0.920-0.991) compared with either annexin A2 (AUC=0.61, P < 0.043, 95% CI: 0.505-0.709) or CA125 (AUC=0.95, P < 0.0001, 95% CI: 0.916-0.989) alone. The sensitivity at the 95% specificity was only slightly increased when annexin A2 and CA125 measurements were combined but there was no improvement at the 98% specificity level (Table 4.3c). This study has shown that combining annexin A2 and CA125 measurements using logistic regression can improve the sensitivity of detecting ovarian cancer.

4.4.8 Logistic regression model to evaluate the efficacy of combining annexin A2 and CA125 measurements in ovarian cancer patients versus patients with benign ovarian tumours We also utilized logistic regression model and ROC curve analysis to combine annexin A2 and CA125 measurements in patients detected with ovarian cancer (stage I to IV) and benign ovarian tumours alone (Figure 4.4d). From the ROC curve, we observed a slightly increased AUC when annexin A2 and CA125 measurements were combined (AUC=0.894, P < 0.0001, 95% CI: 0.826- 0.963) compared with annexin A2 (AUC=0.597, P = 0.115, 95% CI: 0.479-0.715) and CA125 (AUC=0.880, P < 0.0001, 95% CI: 0.805-0.955) alone. At the 90% specificity level, we observed an improvement in the sensitivity combining both biomarkers (61.6%) compared with either annexin A2 (22.1%) or CA125 (48.8%) alone. However, at the 95% specificity level, combining annexin A2 and CA125 did not improve the sensitivity (44.2%) compared to CA125 alone which was 45.3% (Table 4.4a). Figure 4.4e shows the ROC curve analyses for the benign ovarian tumours and early stage ovarian cancers. We showed an improved AUC when both biomarkers were combined (AUC=0.857, P < 0.0001, 95% CI: 0.753-0.960) compared to either annexin A2 (AUC=0.788, P < 0.0001, 95% CI: 0.661-0.919) or CA125 alone (AUC=0.790, P < 0.0001, 95% CI: 0.661-0.919). The sensitivity of detecting early stage ovarian cancer was improved at both the 90% and 95% specificity levels when both biomarkers were combined (Table 4.4b).

116

The ROC curve analysis comparing benign ovarian tumours with advanced stage ovarian cancers demonstrates a slightly improved AUC when both biomarkers were combined (AUC=0.917, P < 0.0001, 95% CI: 0.854-0.980) compared to annexin A2 (AUC=0.527, P = 0.675, 95% CI: 0.398- 0.656) or CA125 (AUC=0.913, P < 0.0001, 95% CI: 0.848-0.978) alone (Figure 4.4f). The sensitivity of detecting ovarian cancer was slightly improved at both the 90% and 95% specificity level when annexin A2 and CA125 measurements were combined (Table 4.4c). In summary, we have shown that combining annexin A2 and CA125 measurements using logistic regression can improve the sensitivity of discriminating benign ovarian tumours from serous ovarian cancers and particularly early stage ovarian cancer.

117

Figure 4.1 Annexin A2 expression in controls and ovarian cancer patient’s urine.

A. Western immunoblotting shows annexin A2 band at 37 kDa in the urine of patients with advanced stage serous ovarian cancer (P1-P4) and low annexin A2 in the urine of patients with benign ovarian tumours (P5-P8). B. Urine annexin A2 levels measured from patients with normal ovaries (n=6), benign ovarian tumours (n=19) and serous ovarian cancers (n=36) by ELISA. The bars represent the median values for annexin A2 levels for normal ovaries (38.8 ng/ml, range: 15.2 - 45.3 ng/ml), benign ovarian tumours (26.7 ng/ml, range: 6.0 - 59.3 ng/ml) and advanced stage ovarian cancer (30.2 ng/ml, range: 12.6 - 175.5 ng/ml). (P = 0.521, Kruskal-Wallis test).

118

119

Figure 4.2 Annexin A2 and CA125 levels in normal ovaries, benign ovarian tumours and serous ovarian cancers.

Serum and plasma annexin A2 (A) and CA125 (B) levels measured from patients with normal ovaries (n=27), benign ovarian tumours (n=30), early stage serous ovarian cancers (n=23) and advanced stage serous ovarian cancers (n=63). The bars represent the median values for annexin A2 in the normal ovaries (240.7 ng/ml, range: 143.7 - 400.7 ng/ml), benign ovarian tumours (271.5 ng/ml, 137.3 - 562.8 ng/ml), early stage (400.5 ng/ml, range: 145.6 - 668.8 ng/ml) and advanced stage (275.6 ng/ml, range: 140.4 - 558.7 ng/ml). The bars represent median values for CA125 levels in the normal ovaries (10 U/ml, range: 5 - 26 U/ml), benign ovarian tumours (13 U/ml, range: 4 - 749 U/ml), early stage (122 U/ml, range: 6 - 2274 U/ml) and advanced stage (431 U/ml, range: 16 - 8456 U/ml). The comparison between the patient groups was performed by Kruskal-Wallis test and Dunnett C post-hoc test. * P < 0.0001. Annexin A2 (C) and CA125 (D) levels from patients with stage I (n=15), stage II (n=8), stage III (n=56) and stage IV (n=7) of serous ovarian cancers. The bars represents median values for annexin A2 in stage I (418.6 ng/ml, range: 255.7 - 668.8 ng/ml), stage II (334.4 ng/ml, range: 145.6 - 420.1 ng/ml), stage III (277.7 ng/ml, range: 152.4 - 558.7 ng/ml) and stage IV (274.5 ng/ml, range: 140.4 - 395.1 ng/ml). The bars represents median values for CA125 in stage I (94 U/ml, range: 9 - 2274 U/ml), stage II (126.5 U/ml, range: 6 - 1274 U/ml), stage III (417.5 U/ml, range: 16 - 8456 U/ml) and stage IV (758 U/ml, range: 17- 2715 U/ml). The comparison between the patient groups was performed by Kruskal-Wallis test and Dunnett C post- hoc test. * P < 0.05. ROC curve and area under the curve (AUC) for annexin A2 (E) and CA125 (F) of controls (normal ovaries and benign ovarian tumours, n=57) and serous ovarian cancers (n=86).

120

Figure 4.3 Correlation of annexin A2 and CA125 levels.

A. Scatterplot of blood annexin A2 and CA125 correlation in control and ovarian cancer patients (n=143) (r = 0.180, P = 0.032, Spearman’s correlation test). Dotted lines represent the established cutoff for CA125 which is 35 U/ml. B. ROC curve for ovarian cancer patients with normal CA125 levels (<35 U/ml, n=9, AUC: 0.803, 95% CI: 0.674-0.932) and control cases (normal ovaries and benign ovarian tumours, n=57).C. ROC curve for ovarian cancer patients with normal CA125 levels (<35 U/ml, n=9, AUC: 0.730, 95% CI: 0.56-0.899) and benign ovarian tumours, n=30).

121

122

Figure 4.4 ROC curves from logistic regression analysis for annexin A2, CA125 and combined annexin A2 and CA125 measurements.

A. ROC curve and area under the curve (AUC) for non-cancer controls (normal ovaries and benign ovarian tumours, n=57) versus serous ovarian cancers (n=86). B. ROC curve and AUC for non- cancer controls (normal ovaries and benign ovarian tumours, n=57) versus early stage serous ovarian cancer (n=23). C. ROC curve and AUC for non-cancer controls (normal ovaries and benign ovarian tumours, n=57) versus advanced stage serous ovarian cancers (n=63). D. ROC curve and AUC for benign ovarian tumours (n=30) versus serous ovarian cancers (n=86). E. ROC curve and AUC for benign ovarian tumours (n=30) versus early stage serous ovarian cancer (n=23). F. ROC curve and AUC for benign ovarian tumours (n=30) versus advanced stage serous ovarian cancers (n=63).

123

Table 4.1 Comparison of annexin A2 and CA125 marker sensitivity, specificity, PPV and NPV for controls (normal ovaries and benign ovarian tumours) and serous ovarian cancers using specified cutoff points.

a) Controls (n=57) and serous ovarian cancers (n=86) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 52 79 79 52 63 CA125 ≥35 U/ml 90 89 93 85 90 Annexin A2 or ≥300 ng/ml or 98 70 83 95 87 CA125 ≥35 U/ml b) Controls (n=57) and early stage ovarian cancer (n=23) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 87 79 63 94 81 CA125 ≥35 U/ml 74 89 74 89 85 Annexin A2 or ≥300 ng/ml or 100 70 57 100 79 CA125 ≥35 U/ml c) Controls (n=57) and advanced stage ovarian cancer (n=63) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 40 79 68 54 58 CA125 ≥35 U/ml 95 89 91 94 84 Annexin A2 or ≥300 ng/ml or 97 70 78 95 84 CA125 ≥35 U/ml

124

Table 4.2 Comparison of annexin A2 and CA125 markers sensitivity, specificity, PPV and NPV for benign ovarian tumours and serous ovarian cancers using specified cutoff points.

a) Benign ovarian tumours (n=30) and serous ovarian cancers (n=86) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 52 70 83 34 49 CA125 ≥35 U/ml 90 80 93 73 90 Annexin A2 or ≥300 ng/ml or 98 53 86 89 83 CA125 ≥35 U/ml b) Benign ovarian tumours (n=30) and early stage ovarian cancer (n=23) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 87 70 69 88 77 CA125 ≥35 U/ml 74 80 74 80 77 Annexin A2 or ≥300 ng/ml or 100 53 62 100 74 CA125 ≥35 U/ml c) Benign ovarian tumours (n=30) and advanced stage ovarian cancer (n=63) Correctly Markers Marker cutoff Sensitivity (%) Specificity (%) PPV (%) NPV (%) classified (%) Annexin A2 ≥300 ng/ml 40 70 74 36 57 CA125 ≥35 U/ml 95 80 91 89 87 Annexin A2 or ≥300 ng/ml or 97 53 81 89 86 CA125 ≥35 U/ml

125

Table 4.3 Comparison of annexin A2 and CA125 markers in controls (normal ovaries and benign ovarian tumours) and serous ovarian cancers using logistic regression and ROC curve analyses.

a) Controls (n=57) and serous ovarian cancers (n=86)

Sensitivity (%) Markers AUC (95%CI) P-value a 95% specificity 98% specificity lnlogAnnexin A2 0.670 (0.579-0.760) < 0.0001 17.4 9 lnlogCA125 0.925 (0.879-0.972) - 48.8 45.3 lnlogAnnexin A2 and 0.939 (0.900-0.979) 0.6546 61.6 44.2 lnlogCA125 b) Controls (n=57) and early stage ovarian cancer (n=23) Sensitivity (%) Markers AUC (95%CI) P-value a 95% specificity 98% specificity lnlogAnnexin A2 0.841 (0.740-0.943) 0.9001 47.8 21.7 lnlogCA125 0.850 (0.750-0.951) - 26.1 21.7 lnlogAnnexin A2 and 0.908 (0.835-0.980) 0.3673 56.5 52.2 lnlogCA125 c) Controls (n=57) and advanced stage ovarian cancer (n=63) Sensitivity (%) Markers AUC (95%CI) P-value a 95% specificity 98% specificity lnlogAnnexin A2 0.607 (0.505-0.709) < 0.0001 6.3 4.8 lnlogCA125 0.953 (0.916-0.989) - 57.1 54 lnlogAnnexin A2 and 0.955 (0.920-0.991) 0.9185 65.1 54 lnlogCA125 a p-value for comparison of ROC-AUC to CA125

126

Table 4.4 Comparison of annexin A2 and CA125 markers in benign ovarian tumours and serous ovarian cancers using logistic regression and ROC curve analyses.

a) Benign ovarian tumours (n=30) and serous ovarian cancers (n=86)

Sensitivity (%) Markers AUC (95%CI) P-value a 90% specificity 95% specificity lnlogAnnexin A2 0.597 (0.479-0.715) < 0.0001 22.1 9.3 lnlogCA125 0.880 (0.805-0.955) - 48.8 45.3 lnlogAnnexin A2 and 0.894 (0.826-0.963) 0.7988 61.6 44.2 lnlogCA125 b) Benign ovarian tumours (n=30) and early stage ovarian cancer (n=23) Sensitivity (%) Markers AUC (95%CI) P-value a 90% specificity 95% specificity lnlogAnnexin A2 0.788 (0.661-0.916) 0.9685 52.2 21.7 lnlogCA125 0.790 (0.661-0.919) - 26.1 21.7 lnlogAnnexin A2 and 0.857 (0.753-0.960) 0.4409 56.5 56.5 lnlogCA125 c) Benign ovarian tumours (n=30) and advanced stage ovarian cancer (n=63) Sensitivity (%) Markers AUC (95%CI) P-value a 90% specificity 95% specificity lnlogAnnexin A2 0.527 (0.398-0.656) < 0.0001 11.1 4.8 lnlogCA125 0.913 (0.848-0.978) - 57.1 54 lnlogAnnexin A2 and 0.917 (0.854-0.980) 0.9408 61.9 55.6 lnlogCA125

a p-value for comparison of ROC-AUC to CA125

127

4.5 Discussion To the best of our knowledge, this is the first study that has investigated whether annexin A2 can be used as a diagnostic marker for serous ovarian cancer. In this study, we showed that i) blood levels of annexin A2 are increased in serous ovarian cancer compared with the patients with normal ovaries and benign ovarian tumours, ii) blood levels of annexin A2 are increased in early stage serous ovarian cancer compared with patients with benign ovarian tumours and iii) combining annexin A2 and CA125 measurements improves the sensitivity of detecting early stage serous ovarian cancer. Urine is one of the sources for non-invasive biomarkers for cancer detection and we investigated whether annexin A2 level is elevated in the urine of ovarian cancer patients. Western immunoblotting showed elevated annexin A2 in advanced stage ovarian cancer patients compared to benign ovarian tumours. However, we showed using ELISA that annexin A2 urine levels were not elevated in serous ovarian cancers compared to non-cancerous controls. A previous proteomic study identified annexin A2 in the urine of pancreatic cancer patients but failed to validate annexin A2 expression by western immunoblotting. It was suggested that this was due to annexin A2 post- translational modifications in the urine samples [176]. Moreover, another reason for limited or no detection of annexin A2 may also be due to the degradation of proteins in urine samples as no protease inhibitors were added during sample collections [177]. As no protease inhibitors were added to our urine samples at the time of storage, degradation of annexin A2 may also have affected our measurements. We then investigated whether blood levels of annexin A2 could be utilized as tumour biomarker in ovarian cancer. Several studies have investigated the clinical significance of annexin A2 as a diagnostic tool in cancer and evaluated blood annexin A2 levels in cancers of the liver [45, 158], breast [174] and multiple myeloma [175]. Annexin A2 was significantly elevated in the plasma of breast cancer patients [174] and in the bone marrow aspirate of multiple myeloma patients [175] compared with controls. Moreover, serum annexin A2 was also elevated in hepatocellular carcinoma (HCC) patients compared with normal individuals [45]. In a large cohort study (n=175), serum annexin A2 was significantly elevated in hepatocellular carcinoma patients compared with non-cancerous controls and cancers of the breast, lung, esophageal, gastric and colorectal [158]. We showed that blood levels of annexin A2 were elevated in early and advanced stage ovarian cancer patients compared with healthy controls and patients with benign ovarian tumours. We also observed increased blood annexin A2 levels in stage I ovarian cancers compared with advanced

128

ovarian cancer patients. The mechanism that results in a higher annexin A2 levels in the early stage serous ovarian cancers compared to advanced stage ovarian cancers is not known. Further investigations are required to examine whether blood annexin A2 levels can reflect the bulk of ovarian cancer disease and is reduced following debulking surgery and chemotherapy like CA125 levels and can be used to monitor disease progression. Previous studies reported annexin A2 having a role in chemoresistance in cancers of the breast [157, 178], pancreas [160, 179] and Lewis Lung Carcinoma [149]. It will be important to determine whether blood annexin A2 levels are also associated with chemoresistance in ovarian cancer. To evaluate the efficacy of combining both annexin A2 and CA125 to detect ovarian cancer, we utilized both specified cutoffs and logistic regression models followed by ROC curve analysis [158, 180]. Using a cutoff point of 300 ng/ml for annexin A2 and the established cutoff point of 35 U/ml for CA125, we observed an improved sensitivity of 100% and a specificity of 70% when annexin A2 and CA125 measurements were combined in the controls and early stage ovarian cancer samples. Importantly, annexin A2 was also able to distinguish between the serous ovarian cancers with normal (<35 U/ml) CA125 levels and non-cancer controls. By logistic regression, we showed that at high specificity levels (95% and 98%), there was an improvement in the sensitivity of detecting early stage ovarian cancer when annexin A2 and CA125 measurements were combined. Sun et al. showed that combining serum annexin A2 and alpha-fetoprotein (AFP) levels improved the sensitivity of detecting early stage hepatocellular carcinoma to 87.4% compared with either annexin A2 (83.2%) or AFP alone (54.7%) [158]. We also showed that combining annexin A2 and CA125 improved the sensitivities of diagnosing early stage serous ovarian cancer compared with the benign ovarian tumours at 90% and 95% specificity levels. The strength of this study is that annexin A2 levels were assessed only in serous ovarian cancer patients. However, the limitation of this study to assess the potential clinical use of annexin A2 as a biomarker for ovarian cancer is that the sample size of patients was small. These preliminary findings require further validation in a larger cohort of early and advanced stage ovarian cancer patients with other histological ovarian cancer types such as endometriod, clear cell and mucinous. Several studies have reported established methods for the improvement in ovarian cancer diagnosis such as Risk of Ovarian Malignancy Algorithm (ROMA), which includes CA125 and Human Epididymis Protein 4 (HE4) values together with menopausal status [181, 182] and Risk Malignancy Index (RMI) which incorporated the CA125, ultrasound and menopausal status [183]. Therefore, clinical information such as pre or post-menopausal status and ultrasound assessment

129

together combined annexin A2 and CA125 biomarkers may be useful to improve the sensitivity, specificity, PPV and NPV. Moreover, the potential clinical utility of annexin A2 as a diagnostic tool in high risk women, particularly post-menopausal women and women with pelvic masses needs to be further investigated together with these clinical assessments similar to ROMA and RMI to improve the diagnostic performance of annexin A2. In conclusion, we showed that combining annexin A2 and CA125 can improve the sensitivity of detecting serous ovarian cancer and particularly early stage ovarian cancer. Annexin A2 shows promise as a diagnostic biomarker for serous ovarian cancer and its potential clinical significance warrants further investigation.

4.6 Acknowledgements We thank Suzanne Edwards for her help with the logistic regression model and ROC curve analysis. The authors thank Dr Carmen Pyragius and Dr Izza Tan for their help with the annexin A2 ELISA. This research was supported by the Ovarian Cancer Research Foundation (OCRF) Australia, Cancer Council SA, Australia and the South Australian Health and Medical Research Institute (SAHMRI).

130

Supplementary Table 4.1 Clinicopathological characteristics of urine patient cohort.

Age (Years) Disease Status n Median (range)

Controls (n=6) Normal ovaries 3 51 (35-67)

Simple ovarian cyst 3

Benign ovarian tumours (n=19) Serous cystadenoma 8 58 (25-90)

Serous cystadenofibroma 5

Ovarian fibroma 3

Ovarian endometrioma 1

Ovarian leiomyoma 1

Ovarian mucinous cystadenoma 1

Serous ovarian cancers (n=36) Stage I 3 65 (37-84)

Stage II 3

Stage III 29

Stage IV 1

131

Supplementary Table 4.2 Clinicopathological characteristics of serum or plasma patient cohort.

Age (Years) Disease Status n Median (range)

Controls (n=27) Normal ovaries 9 50 (30-82)

Simple ovarian cyst 6

Menorrhagia 9

Endometrial polyp 3

Benign ovarian tumours (n=30) Serous cystadenoma 18 58 (25-90)

Serous cystadenofibroma 6

Ovarian fibroma 3

Ovarian endometrioma 1

Ovarian leiomyoma 1

Ovarian mucinous cystadenoma 1

Serous ovarian cancers (n=86) Stage I 15 65 (37-88)

Stage II 8

Stage III 56

Stage IV 7

132

5 Chapter 5: Prognostic Significance of Annexin A2 and S100A10 in Serous Ovarian Cancer

Noor A Lokman1, Carmen E Pyragius1, Andrew Ruszkiewicz2, Martin K Oehler1, 3 and Carmela Ricciardelli1

1 Discipline Robinson Institute, Research Centre for Reproductive Health, School of Paedriatrics and Reproductive Health, University of Adelaide, Adelaide, SA 5005, Australia 2 Centre of Cancer Biology, University of South Australia and Department of Anatomical Pathology, SA Pathology, SA, Australia 3 Department of Gynaecological Oncology, Royal Adelaide Hospital, Adelaide, SA 5005, Australia

Manuscript in Preparation

133

5.1 Abstract Annexin A2, a calcium phospholipid binding protein has been shown to play an important role in ovarian cancer metastasis. This study examined whether annexin A2 can be used as a prognostic marker in serous ovarian cancer. Expression of annexin A2 and its binding protein; S100A10 was assessed using immunohistochemistry in a uniform tissue microarray cohort of stage III serous ovarian cancers (n=91) and matching metastatic lesions (n=30). Kaplan-Meier and Cox Regression analyses were performed to assess the relationship between annexin A2 or S100A10 expression and clinical outcome. Kaplan-Meier and Cox Regression analyses showed high stromal annexin A2 immunostaining to be significantly associated with reduced progression-free survival and overall survival. Moreover, high cytoplasmic S100A10 staining was associated with reduced overall survival but not progression-free survival. Multivariate Cox Regression analysis showed stromal annexin A2 and cytoplasmic S100A10 levels were independent predictors of overall survival in the patients with residual disease. Annexin A2 and S100A10 are novel prognostic biomarkers for serous ovarian cancer patients could aid patient management.

Keywords: ovarian cancer, annexin A2, S100A10, prognosis

134

5.2 Introduction Ovarian cancer is the most lethal gynaecological cancer worldwide. It has been estimated that there will be 22,240 new cases of ovarian cancer and 14,030 deaths due to ovarian cancer in the United States in 2013 [148]. Up to 70% of ovarian cancer patients are diagnosed at advanced stage and the 5-year survival rate for advanced stage ovarian cancer is 27% compared to 92% for patients with early stage disease [148]. The standard treatment options for ovarian cancer are radical debulking surgery and chemotherapy such as carboplatin and paclitaxel. However, despite improvements in ovarian cancer treatment, there is a high mortality rate and poor treatment outcome for ovarian cancer due to ovarian cancer recurrence and chemoresistance. Therefore, identifying prognostic markers are important for predicting survival, to guide in the management of ovarian cancer patients and aid in the development of better therapeutic targets. Annexin A2, a calcium phospholipid binding protein was reported to be expressed in many malignancies and play an important role in tumourigenesis [132]. Previously, we have characterized annexin A2 expression in epithelial cells and cancer associated stromal cells of serous ovarian cancer tissues and in serous ovarian cancer metastatic implants [172]. Annexin A2 expression was significantly increased in the stromal cells of serous ovarian cancer tissues compared with stromal cells of normal ovaries, benign serous cystadenomas and non-invasive borderline ovarian tumours [172]. Annexin A2 has been reported to be a potential prognostic marker in various malignancies. Increased annexin A2 expression was associated with poor prognosis in multiple myeloma [175], non-small cell lung cancer [184, 185], colorectal carcinoma [101, 186], cholangiocarcinoma [187], gastric carcinoma [188] and glioma [189]. However, no studies have investigated whether annexin A2 can be used as a prognostic marker for ovarian cancer. Annexin A2 and its binding protein, S100A10 exist as an annexin A2 heterotetramer on the cell surface of various tumour cells [190]. S100A10, also known as p11 protein is a member of S100 protein family which has 2 EF-hand calcium binding motifs [191]. Moreover, S100A10 is a cytosolic protein and upon binding to annexin A2, annexin A2 heterotetramer complex translocate to the plasma membrane [192]. S100A10 has been shown to play an important role in oncogenesis [193] and S100A10 expression has been characterized in cancers of the bowel [194, 195], thyroid [196] and kidney [31]. In a recent study, S100A10 protein expression was identified in a gene signature predicting overall survival in primary serous ovarian carcinoma [197]. Moreover, high S100A10 expression has been shown to correlate with poor prognosis in Barrett’s adenocarcinoma [198].

135

S100A10 was also identified using a proteomic approach to be a potential serum biomarker for gallbladder cancer. Furthermore, increased S100A10 expression in late stage gallbladder cancer was associated with poor clinical outcome [199]. To date, no studies have reported S100A10 expression in ovarian malignancies and its relationship with ovarian cancer prognosis. In this study, we investigate whether annexin A2 and S100A10 expression has a prognostic relevance in stage III serous ovarian cancer, the most common and lethal ovarian cancer subtype.

136

5.3 Materials and methods

5.3.1 Patient’s tissue microarray cohort Archived formalin fixed paraffin tissue blocks from stage III serous primary ovarian tumours (n=91) and matching metastatic tissues (n=30), diagnosed between 1994 and 2004 were obtained from the IMVS (SA Pathology, Adelaide, SA, Australia). Tissue microarrays (TMAs) were constructed from formalin fixed paraffin tissue blocks and each tissue block was represented by triplicate 1.0 mm diameter tissue cores, with the approval of ethics committee from the Royal Adelaide Hospital. The clinicopathological characteristics of TMAs cohort are detailed in Supplementary Table 5.1.

5.3.2 Immunohistochemistry Immunohistochemistry was performed as previously described [172]. Briefly, tissue sections (5 m) underwent microwave antigen retrieval for 10 minutes at 100⁰C in a steam microwave (Sixth Sense, Whirlpool, VIC, Australia) in 10 mM citric acid buffer (pH 6.0). Sections were incubated overnight with mouse monoclonal antibody against annexin A2 (1/500, BD Biosciences, CA, USA) or mouse monoclonal antibody against S100A10 (1/2000, BD Biosciences) in blocking buffer (5% normal goat serum) at 4⁰C. Visualization of immunoreactivity was achieved using biotinylated anti- mouse immunoglobulins (1/400, Dako, NSW, Australia) and streptavidin-peroxidase conjugate (1/500, Dako). Known positive tissues were used as positive controls [172] and negative controls included tissues incubated with no primary antibody or anti-mouse immunoglobulins.

5.3.3 Immunohistochemical assessment Slides were digitally scanned using the NanoZoomer Digital Pathology System (Hamamatsu Photonics, SZK, Japan) and images were collected using NDP view imaging software (NDP scan software v2.2, Hamamatsu Photonics). The percentage of positive cancer cells with cytoplasmic and membrane staining and stromal intensity score of strong (3+), moderate (2+), weak (1+), or negative (0) were assessed by two experienced researchers (NL and CR). Positivity ≥10% was defined as high membrane or cytoplasmic immunostaining and <10% was defined as low. A score of 2+ or 3+ was defined as high stromal immunostaining and 1+ or 0 was defined as low stromal immunostaining, as described previously [172].

5.3.4 Statistical analyses All statistical analyses were performed using the SPSS for Windows software (Version 20.0, SPSS Inc., Chicago, IL, USA). Wilcoxon sign rank test was conducted to compare annexin A2 or

137

S100A10 expression in the primary ovarian cancer tissues with matching metastatic lesions. Chi- squared test was performed to determine the correlation of annexin A2 and S100A10 immunostaining in primary ovarian tumour tissues. Kaplan-Meier, univariate Cox Regression and multivariate Cox Regression analyses were performed to assess the association of annexin A2 and S100A10 expression in the stage III ovarian cancer TMA cohort with progression-free survival and overall survival. Statistical significance was accepted at P < 0.05.

138

5.4 Results

5.4.1 Increased stromal annexin A2 and S100A10 expression in metastatic ovarian cancer tissues compared with matching primary ovarian tumours Annexin A2 immunostaining patterns in primary ovarian tumours are shown in Figure 5.1a-d. Similar staining patterns to annexin A2 were observed in matching tissues samples immunostained for S100A10 (Figure 5.1e-h). A significant correlation between annexin A2 and S100A10 immunostaining in the membrane (P = 0.002), cytoplasm (P = 0.003) and stroma (P < 0.0001) were observed (Table 5.1a-c). Moreover, we showed a significant increase in stromal annexin A2 expression in metastatic ovarian cancer tissues (26/30) compared with matching primary ovarian cancer tissues (Figure 5.2a-b, P < 0.0001). Increased stromal S100A10 expression was also observed in metastatic ovarian cancer tissues (16/31) compared with matching primary tissues (Figure 5.2c-d, P = 0.001). However, our findings showed no difference in both annexin A2 or S100A10 expression in epithelial cells between the primary ovarian tumours and matching metastatic tissues (data not shown).

5.4.2 High stromal annexin A2 expression in the cancer associated stroma predicts clinical outcome Kaplan-Meier analyses showed that annexin A2 positivity in the cancer cell membrane or cytoplasm was not associated with progression-free survival (Figure 5.3a, P = 0.994 & Figure 5.3c, P = 0.568) or overall survival (Figure 5.3b, P = 0.437 & Figure 5.3d, P = 0.196). However, high annexin A2 immunostaining in the cancer associated stroma was significantly associated with reduced progression-free survival (Figure 5.3e, P = 0.009) and reduced overall survival (Figure 5.3f, P = 0.04). The 5-year progression-free survival rate for patients with high stromal annexin A2 expression was 7% compared to 29% for patients with low stromal annexin A2. Moreover, the 5- year overall survival rate for patients with high stromal annexin A2 was 16.2% compared to 41.3% for patients with low stromal annexin A2. No correlations were observed between membrane, cytoplasmic or stromal annexin A2 immunostaining and clinical and pathological parameters including patient’s age (P = 0.276), tumour grade (P = 0.449) or residual disease status (P = 0.792) (Table 5.2).

5.4.3 S100A10 cytoplasmic positivity is associated with reduced survival Kaplan-Meier analyses showed no significant relationship between S100A10 membrane positivity with progression-free survival (Figure 5.4a, P = 0.610) nor overall survival (Figure 5.4b, P = 0.431).

139

However, high S100A10 positivity in the cytoplasm was marginally associated with reduced overall survival (Figure 5.4d, P = 0.054), but not with progression-free survival (Figure 5.4c, P = 0.108). The 5-year overall survival rate for high cytoplasmic S100A10 expression was 19.1% compared to 28.6% for patients that have low S100A10 cytoplasmic expression. In contrast to stromal annexin A2, stromal S100A10 immunostaining was not associated with either progression-free survival (Figure 5.4e, P = 0.254) or overall survival (Figure 5.4f, P = 0.178).

5.4.4 Stromal annexin A2 positivity and cytoplasmic S100A10 expression is an independent predictor of clinical outcome Univariate Cox Regression analysis showed no significant relationship between patient age or tumour grade with progression-free survival or overall survival (Table 5.3a). However the presence of residual disease was associated with a 2.17 fold increased risk of disease progression (95% CI: 1.02-4.63, P = 0.043) and 2.9 fold increased risk of death due to serous ovarian cancer (95% CI: 1.28-6.57, P = 0.011). Cox Regression analysis showed annexin A2 expression in the membrane is not significantly associated with increased risk of disease progression (RR=1.02, 95% CI: 0.55- 1.75, P = 0.994) and increased risk of death (RR=1.28, 95% CI: 0.69-2.37, P = 0.438). Similarly, cytoplasmic annexin A2 was not associated with increased risk of disease progression (RR=1.2, 95% CI: 0.64-2.25, P = 0.569) or an increased risk of death (RR=1.6, 95% CI: 0.78-3.25, P = 0.200). However, Cox Regression analysis found that patients with high stromal annexin A2 had a 1.96 fold increased risk of disease progression (95 %CI: 1.17-3.25, P = 0.010) and 1.72 fold increased risk of ovarian cancer death (95 %CI: 1.02-2.9, P = 0.043; Table 5.3a). S100A10 membrane expression was not associated with increased risk of disease progression (RR=1.16, 95% CI: 0.66-2.05, P = 0.610) and increase risk of death (RR=1.27, 95% CI: 0.70-2.32, P = 0.423). High cytoplasmic S100A10 was associated with a 2.3 fold increase of ovarian cancer death, although this only reached a borderline significance (95%CI: 0.97-5.22, P = 0.06). In comparison to stromal annexin A2 expression, stromal S100A10 expression was not significantly associated with an increased risk of disease progression (RR=1.33, 95% CI: 0.81-2.20, P = 0.256) nor increased risk of ovarian cancer death (RR=1.42, 95% CI: 0.85-2.37, P = 0.180). In the multivariable Cox Regression analysis for progression-free survival, which included positive residual disease and stromal annexin A2 immunostaining, stromal annexin A2 remained an independent predictor of progression-free survival (RR=1.94, 95% CI: 1.15-3.28, P = 0.013). Moreover, multivariable Cox Regression analysis for overall survival also showed that stromal annexin A2 immunostaining overall survival (RR= 2.07, 95%CI: 1.20-3.56, P = 0.009) (Table 5.3b)

140

and cytoplasmic S100A10 were independent predictors of overall survival (RR= 2.75, 95%CI: 1.09- 6.94, P = 0.032, Table 5.3b).

141

142

Figure 5.1 Annexin A2 and S100A10 immunostaining in stage III serous ovarian cancers

A representative image of low (A) and high (B) annexin A2 immunostaining in the membrane and cytoplasm of the epithelial cells and cancer associated stroma. Low (C) and high (D) annexin A2 immunostaining in the cancer associated stroma. A representative image of low (E) and high (F) S100A10 expression in the membrane and cytoplasm of the epithelial cells and in the cancer associated stroma. Low (G) and high (H) S100A10 expression in the cancer associated stroma. Ep = cancer cells and St = peritumoural stroma. Magnification bar = 100 μm for all images.

143

Figure 5.2 Annexin A2 and S100A10 immunostaining in primary ovarian tumour and matching metastasis tissues.

A representative image of low stromal annexin A2 immunostaining (A) in primary ovarian tumour and high stromal annexin A2 immunostaining in matching metastatic tissue (B). Low stromal S100A10 immunostaining is present in primary ovarian tumour (C) compared with high stromal S100A10 immunostaining in matching metastatic tissue (D). Ep = cancer cells and St = peritumoural stroma. Magnification bar = 100 μm for all images.

144

145

Figure 5.3 High stromal annexin A2 expression is associated with poor prognosis in stage III serous ovarian cancer patients.

Kaplan Meier analyses showing the association of annexin A2 expression in membrane (A-B) and cytoplasm of epithelial cells (C-D) and cancer associated stroma (E-F) with progression-free survival (n=83) and overall survival in stage III serous ovarian cancer patients (n=87). Patients with high annexin A2 stroma expression had a significantly reduced progression-free survival (P = 0.009) and overall survival (P = 0.04). Positivity of ≥10% was defined as high membrane or cytoplasmic immunostaining and <10% was defined as low. Annexin A2 stromal immunostaining score of 0 and 1+ was defined as low annexin A2 expression and 2+ and 3+ represents high annexin A2 expression.

146

147

Figure 5.4 S100A10 cytoplasmic expression is associated with reduced overall survival for stage III serous ovarian cancer patients.

Kaplan Meier analyses showing association of S100A10 expression in membrane (A-B) and cytoplasm of epithelial cells (C-D) and cancer associated stroma (E-F) with progression-free (n=83) and overall survival in stage III serous ovarian cancer patients (n=88). Patients with high S100A10 cytoplasm expression had a significantly reduced overall survival (P = 0.054). Positivity of ≥10% was defined as high membrane or cytoplasmic immunostaining and <10% was defined as low. S100A10 stromal immunostaining score of 0 and 1+ was defined as low S100A10 expression and 2+ and 3+ represents high S100A10 expression.

148

Table 5.1 Correlation of annexin A2 and S100A10 immunostaining in the membrane and cytoplasm of cancer cells and cancer associated stroma in stage III primary serous ovarian cancer.

a) Annexin A2 membrane and S100A10 membrane Annexin A2 membrane <10% a Annexin A2 membrane ≥10% S100A10 membrane <10% b 12/22 (54.5%) 12/68 (17.6%) S100A10 membrane ≥10% 10/22 (45.5%) 56/68 (82.3%) Chi-squared test P = 0.002 b) Annexin A2 cytoplasm and S100A10 cytoplasm Annexin A2 cytoplasm <10% c Annexin A2 cytoplasm ≥10% S100A10 cytoplasm <10% d 7/15 (46.7%) 8/75 (10.7%) S100A10 cytoplasm ≥10% 8/15 (53.3%) 67/75 (89.3%) Chi-squared test P = 0.003 c) Annexin A2 stroma and S100A10 stroma Annexin A2 stroma low e Annexin A2 stroma high S100A10 stroma low f 29/39 (74.4%) 11/51 (21.6%) S100A10 stroma high 10/39 (25.6%) 40/51 (78.4%) Chi-squared test P < 0.0001 a = Annexin A2 membrane (% positive cells) as a dichotomous variable, cut point <10 vs 10 b = S100A10 membrane (% positive cells) as a dichotomous variable, cut point <10 vs 10 c = Annexin A2 cytoplasm (% positive cells) as a dichotomous variable, cut point <10 vs 10 d = S100A10 cytoplasm (% positive cells) as a dichotomous variable, cut point <10 vs 10 e = Annexin A2 stroma as dichotomous variable, cut point intensity < 0 or 1 (low) vs  2 or 3 (high) f = S100A10 stroma as dichotomous variable, cut point intensity < 0 or 1 (low) vs  2 or 3 (high)

149

Table 5.2 Correlation of annexin A2 stromal intensity with age, tumour grade and residual disease in stage III primary serous ovarian cancer.

a Annexin A2 stroma low Annexin A2 stroma high Age

<55 18/39 (46.2%) 17/51 (33.3%)

≥55 21/39 (53.8%) 34/51 (66.7%) Chi-squared test P = 0.276 Tumour grade b Well or moderate 7/36 (19.4%) 14/50 (28%) Poor 29/36 (80.6%) 36/50 (72%) Chi-squared test P = 0.449

Residual disease

No 30/37 (81%) 38/49 (77.6%) Yes 7/37 (18.9%) 11/49 (22.4%) Chi-squared test P = 0.792 a = Annexin A2 stroma as dichotomous variable, cut point intensity < 0 or 1 (low) vs  2 or 3 (high) b= Well or moderate = grade 1 or 2 and poor = grade 3

150

Table 5.3 Cox Regression analyses of stage III serous ovarian cancer patient cohort.

a) Univariate Cox Regression analyses for progression-free survival and overall survival. Progression-free survival Overall survival Variable Relative risk 95% CI P value Relative risk 95% CI P value Age a 1.31 0.79-2.15 0.297 1.41 0.84-2.35 0.190 Tumour grade b 1.24 0.69-2.21 0.470 1.34 0.75-2.42 0.323 Residual disease c 2.17 1.02-4.63 0.043 2.90 1.28-6.57 0.011 Annexin A2 mem d 1.02 0.55-1.75 0.994 1.28 0.69-2.37 0.438 Annexin A2 cyto e 1.20 0.64-2.25 0.569 1.60 0.78-3.25 0.200 Annexin A2 st f 1.96 1.17-3.25 0.010 1.72 1.02-2.90 0.043 S100A10 mem g 1.16 0.66-2.05 0.610 1.27 0.70-2.32 0.423 S100A10 cyto h 1.77 0.87-3.60 0.113 2.25 0.97-5.22 0.060 S100A10 st i 1.33 0.81-2.20 0.256 1.42 0.85-2.37 0.180

b) Multivariate Cox Regression analyses for progression-free survival and overall survival. All variables are significant in univariate analysis (n=82). Progression-free survival Overall survival Variable Relative risk 95% CI P value Relative risk 95% CI P value Residual disease 1.95 0.92-4.24 0.082 2.73 1.16-6.39 0.021 Annexin A2 st 1.94 1.15-3.28 0.013 2.07 1.20-3.56 0.009 S100A10 cyto - - - 2.75 1.09-6.94 0.032

NOTE: P values highlighted in bold indicate P < 0.05 a = Age a dichotomous variable, cut point <55 vs ≥55 b = Tumours grade (well or moderate vs poor) c = Residual disease status (negative vs positive) d = Annexin A2 membrane (% positive cells) as a dichotomous variable, cut point <10 vs 10 e = Annexin A2 cytoplasm (% positive cells) as a dichotomous variable, cut point <10 vs 10 f = Annexin A2 stroma as dichotomous variable, cut point intensity < 0 or 1 vs  2 or 3 g = S100A10 membrane (% positive cells) as a dichotomous variable, cut point <10 vs 10 h = S100A10 cytoplasm (% positive cells) as a dichotomous variable, cut point <10 vs 10 i = S100A10 stroma as dichotomous variable, cut point intensity < 0 or 1 vs  2 or 3

151

5.5 Discussion To date, this is the first study that has investigated the prognostic relevance of annexin A2 in stage III serous ovarian cancer. We demonstrate that high stromal annexin A2 expression is an independent prognostic factor for progression-free and overall survival in patients with ovarian cancer. In this study, a strong correlation of annexin A2 and S100A10 expression immunostaining was observed in stage III primary serous ovarian cancer. Annexin A2 and S100A10 co-expression has been reported in renal carcinomas [31] and thyroid neoplasms [196]. Moreover, we observed increased annexin A2 and S100A10 immunostaining in stromal cells of metastatic ovarian cancer tissues compared with the matching primary ovarian tumours. Previous studies also showed increased annexin A2 expression in the stroma of breast cancer tissues [200] and in the stroma of cervical cancer tissues after neoadjuvant chemotherapy [201]. Annexin A2 has also been shown to be overexpressed in patients of metastatic renal carcinomas tissues compared with the primary tissues [29]. Moreover, increased annexin A2 expression in the cancer of bile ducts (cholangiosarcoma) was significantly associated with lymphatic invasion, metastasis and perineural invasion [187]. Our findings showed high stromal annexin A2 expression in stage III serous ovarian cancer was significantly associated with reduced progression-free survival and overall survival. However, no relationship between annexin A2 in the membrane or cytoplasm of cancer cells and patients outcome was observed. Moreover, our results showed that high stromal annexin A2 expression independently predicted an increased risk of death for serous ovarian cancer patients with residual disease. Previous studies showed high annexin A2 expression correlates with poor prognosis and reduced survival outcome in multiple myeloma [175], non-small cell lung cancer [184, 185] and colorectal carcinoma [101, 186]. Kagawa et al. also reported that high annexin A2 expression in pancreatic cancer was associated with rapid recurrence and reduced disease free survival in pancreatic cancer patients after being treated with gemcitabine as an adjuvant chemotherapy [202]. Ohno et al. showed that high annexin A2 expression in clear-cell renal carcinomas was associated with a lower 5 year metastasis free rate and an independent predictor for metastasis [29]. Annexin A2 expression in head and neck squamous carcinomas correlates with tumour grade, however, no significant association with higher tumour stages and disease recurrence was observed [41].Contrary to those results, reduced annexin A2 tissue expression was reported in advanced disease with regional lymph node and distant metastasis in prostate cancer. Moreover,

152

low annexin A2 expression in prostate cancer was also associated with risk of recurrence and presented as an independent factor for overall survival [203]. The discrepancies of these findings may be explained due to different histological origin of the tumour and diverse role of annexin A2 in different tumour types. Our previous findings suggest that the extracellular form of annexin A2 found in the cancer associated stroma in the ovarian cancer tissues may represent a cleaved secreted form of annexin A2, which may assist ovarian cancer progression and metastasis [172]. This notion is supported by the report that cleavage of annexin A2 at lysine 10 by MMP-7 can assist tumour invasion and metastasis of colorectal and breast cancer cell lines [85]. Moreover, increased stromal annexin A2 and tenascin C expression in colorectal carcinomas was shown to be an independent factor of poor prognosis [101]. However annexin A2 lacks a signal peptide and is not secreted via the endoplasmic reticulum pathway, the mechanisms that regulate increased stromal annexin A2 are unknown. We observed high cytoplasmic S100A10 expression in stage III serous ovarian cancer was associated with overall survival (borderline significant) but not progression-free survival. Tan et al. showed high S100A10 expression in advanced stage cancer of the gall bladder to be associated with poor prognosis [199]. A large cohort study (n=228) demonstrated S100A10 gene expression was an independent predictor for tumour recurrence in stage II and III colon carcinoma patients after 5’ fluoruracil-based adjuvant chemotherapy [204]. Moreover, high S100A10 expression in patients with invasive Barrett’s adenocarcinomas (n=73) was associated with reduced progression- free survival and was found to be an independent prognostic factor in Barrett’s adenocarcinoma [198]. Our multivariate analysis showed cytoplasmic S100A10 expression in stage III serous ovarian cancer patients to be an independent factor of poor clinical outcome together with patients that have high stromal annexin A2 expression and positive for residual disease. This is the first study to demonstrate that S100A10 expression in serous ovarian cancer tissues. The strength of this study is that we examined annexin A2 and S100A10 expression in a uniform cohort of stage III serous ovarian cancer patients. It is not known whether annexin A2 and S100A10 are also expressed in other subtypes of epithelial ovarian cancers such as endometriod, clear cell and mucinous carcinomas. Previous studies a showed correlation of annexin A2 expression with increasing tumour stage in clear cell renal carcinomas [29], colorectal carcinomas [30] and squamous cell carcinomas of the lung [205]. However, further studies need to address whether annexin A2 or S100A10 expression is associated with increasing stage and ovarian

153

cancer progression in a larger patient cohort. Future studies should also address whether annexin A2 or S100A10 are potential predictors of response to treatment such as chemotherapy in ovarian cancer.

5.6 Acknowledgements We thank Mrs Wendy Bonner for her help with tissue sectioning. This study was supported by Ovarian Cancer Research Foundation (OCRF) Australia, Cancer Council SA and South Australian Health and Medical Research Institute (SAHMRI).

154

Supplementary Table 5.1 Clinicopathological characteristics of stage III serous ovarian cancer patient tissue cohort.

Characteristics n FIGO Stage III 91

Histological type Serous 91

Age at diagnosis (years) 91

Median ± SD 59 ± 12

Range 24-86

Histological grade a Well 4 Moderate 17 Poor 70

Residual disease No 68

Yes 18

No data 5 a= Well = grade 1, moderate = grade 2 and poor = grade 3

155

6 Chapter 6: General Discussion and Conclusions

156

6.1 General discussion Ovarian cancer is a complex and heterogeneous disease with a late stage presentation [112]. Despite advancements in surgical interventions, chemotherapies and novel drug developments, ovarian cancer patients have a poor survival outcome and a high mortality rate [165]. Ovarian cancer cell implantation onto the peritoneal lining is a vital step in ovarian cancer metastasis. However, the mechanisms involved in this process are poorly characterized. Our group has recently explored the interactions between ovarian cancer and peritoneal cells using an in vitro co- culture system [92]. One of the proteins identified by 2D-gel electrophoresis and mass spectrometry to be regulated by ovarian cancer-peritoneal cell interactions was annexin A2 [132]. Annexin A2, a multifunctional calcium phospholipid binding protein exists as an annexin A2 heterotetramer with S100A10 proteins on the cell surface [21] and has been shown to mediate cancer progression through the plasminogen activation system [2]. The aims of this study were to characterize annexin A2 expression in human serous ovarian cancer tissues, cell lines and primary ovarian cancer cells and to investigate the effects of annexin A2 siRNAs and neutralizing annexin A2 antibody in ovarian cancer cell motility, invasion and adhesion to the peritoneal cells in vitro. This study also determined whether annexin A2 inhibitors could block ovarian cancer invasion and metastasis in in vivo models. Moreover, we also aimed to evaluate whether annexin A2 can be used as a tumour marker for ovarian cancer diagnosis and to assess whether annexin A2 or S100A10 can be used as prognostic markers. Annexin A2 has been characterized in various malignancies and had been shown to play an important role in tumourigenesis [132]. This is the first study which characterizes and investigates the role of annexin A2 in ovarian cancer and provides further validation on annexin A2, as one of the novel proteins modulated from co-cultured ovarian cancer and peritoneal cells in vitro. Previously, we have identified Transforming Growth Factor Inducible Protein (TGFBIp) in the co- culture system to play a role in ovarian cancer cell motility, invasion and adhesion to the peritoneal cells [92]. Chapter 1 is a published literature review which summarizes annexin A2 expression in various malignancies and its functional role in cancer cell motility, invasion, adhesion, proliferation and apoptosis [132]. Chapter 2 details the development of the chick chorioallantoic membrane (CAM) model to assess the invasion of OVCAR-3, SKOV-3 and OV-90 cells in vivo which was published as a research paper [156]. The CAM model has been used previously to study cancer cell invasion and metastasis [114, 126], however, limited studies have used the CAM model to assess ovarian

157

cancer cell invasion [131]. We utilized the CAM model not only to assess the invasion of established ovarian cancer cell lines, but also to examine invasion of primary ovarian cancer cell lines as detailed in chapter 3. Moreover, this work highlights that the CAM model is a robust model to investigate the effects of novel inhibitors on ovarian cancer cell invasion.

6.1.1 Annexin A2 plays an important role in ovarian cancer invasion and metastasis Chapter 3 is a published research paper on annexin A2 expression in ovarian cancer tissues and cell lines and summarizes the results on the role of annexin A2 in ovarian cancer invasion and metastasis using in vitro and in vivo ovarian cancer models [172]. We reported annexin A2 was expressed in serous ovarian cancer cells (Figure 3.1d) and a significant increase of annexin A2 expression in ovarian cancer associated stromal cells compared with normal ovaries, benign ovarian cystadenomas and borderline ovarian tumours (Supplementary Table 3.1). Interestingly, we also observed increased annexin A2 expression in ovarian cancer cells located adjacent to the peritoneal cells in the metastatic omental implants (Figure 3.1f). Annexin A2 siRNAs significantly inhibited ovarian cancer cell motility, invasion and adhesion to the peritoneal cells (Figure 3.5). Previous studies have shown that annexin A2 plays a role in cancer cell motility and invasion [76, 150, 151] and cell adhesion [11]. We demonstrated that annexin A2 neutralizing antibodies could significantly inhibit invasion of OV-90 cells and two primary ovarian cancer cell lines in the in vivo CAM model. We reported a 3.6-fold reduction of invasion into the CAM mesoderm with the annexin A2 antibody treated OV-90 cells compared to control (Figure 3.6g) and inhibition of primary ovarian cancer cells invasion into the CAM mesoderm by 4-fold compared with the control mouse IgG (Figure 3.7e). Moreover, we utilized an intraperitoneal xenograft mouse model using SKOV-3/GFP Luc cells to assess annexin A2 role in ovarian cancer invasion and metastasis in vivo. Our findings showed that mice treated with annexin A2 neutralizing antibodies had a reduced tumour burden and peritoneal metastasis compared with control mice treated with mouse IgG (Figure 3.8 & Supplementary Figure 3.3). These observations were consistent with previous studies as mice treated with annexin A2 inhibitors showed reduced tumour growth and progression mediated by the plasminogen activation system in gliomas [150] and breast cancer [162]. We found that the reduced tumour burden in annexin A2 antibody treated mice was due to an increase in cell apoptosis and reduced cell survival (Figure 3.9). The mechanism of cell apoptosis mediated by annexin A2 is poorly characterized, although previous studies have also reported that reduced annexin A2 levels resulted in increased cancer cell apoptosis [71, 86, 150]. Sharma et al. have

158

shown reduced tumour burden in breast cancer cells of annexin A2 treated mice is a result of inhibition of neoangiogenesis [162] and Zhai et al. showed a reduced gliomas tumour growth is a result of reduced cell proliferation [150]. However, we did not observe any significant changes in either cancer cell proliferation or angiogenesis in the tumour tissues of the annexin A2 treated mice compared with the control mice. This discrepancy may be due to differences in cell lines used between these studies. A secreted form of annexin A2 was identified by mass spectrometry in the conditioned media of co- cultured ovarian cancer and peritoneal cells missing the N-terminal domain (Supplementary Figure 3.1). Sato et al. also reported annexin A2 to be secreted in the conditioned media of pancreatic cancer cell lines [160]. However, the molecular mechanisms involved in regulating annexin A2 secretion are poorly characterized. Our findings suggest that the secreted form of annexin A2 may be both cleaved and/or modified by post translational mechanisms as we observed changes in size and pI of annexin A2 isoforms (Figure 3.4c-d). We showed that plasmin cleaved recombinant annexin A2 (Figure 3.4e) and previously we observed an increase of plasmin level in the conditioned media of co-cultured ovarian cancer and peritoneal cells [92]. Therefore, we proposed that annexin A2 mediates cancer cell invasion via the plasminogen activation pathway (Figure 1.4). Zheng et al. recently reported that tyrosine 23 phosphorylation of annexin A2, was important in enhancing pancreatic ductal adenocarcinoma cells invasion and metastasis in vitro and in vivo and promotes epithelium-mesenchymal transition (EMT) via the TGFβ-Rho-mediated [32]. Annexin A2 also was shown to mediate breast cancer cells invasion and metastasis via t-PA dependent plasmin generation pathway [66]. Moreover, annexin A2 knockdown in glioma cell lines decreases glioma cells invasion and tumourigenesis which is associated with decreased membrane-bound plasmin activity [150]. Together, these findings highlight that annexin A2 plays a vital role in promoting tumourigenesis via the plasminogen activation pathway.

6.1.2 Role of annexin A2 in tumour microenvironment Annexin A2 heterotetramer localizes with CD44 [26] and t-PA [3] which promotes cancer progression via the plasminogen activation pathway. Oliferenko et al. reported interaction of annexin A2 with CD44 in the lipid rafts of mammary epithelial cells [26] and CD44 expression was also observed in serous ovarian cancer tissues and plays a role in ovarian cancer metastasis (reviewed in [206]). Annexin A2 heterotetramer mediates malignant progression i) via interactions with t-PA, cathepsin B, tenascin C and collagen I and ii) its role in regulating plasmin generation which result in proteolysis of basement membrane that releases and activate growth factors,

159

activation of proteases such as u-PA, MMPs and degradation of ECM proteins (reviewed in [5]). The extracellular matrix remodeling mediated by annexin A2 in the plasminogen activation system may be critical for cancer cell invasion, angiogenesis and metastasis [1, 62, 207]. Our findings highlight that annexin A2 plays an important role in ovarian cancer tumourigenesis in vitro and in vivo and we proposed a model of the mechanisms of annexin A2 which promotes cancer cell invasion, metastasis and tumour growth (Figure 6.1). Annexin A2 heterotetramer localizes with cell surface receptors and extracellular matrix (ECM) proteins such as CD44 [26], tenascin-C [51], cathepsin-B [53] and tissue-plasminogen activator (t-PA) [3]. Annexin A2 and t-PA interactions mediated conversion of plasminogen to plasmin [3] and increases plasmin levels in the tumour microenvironment which results in the activation of growth factors, activation of proteases such as MMP-2 and MMP-9 [151] and degradation of ECM proteins [5] that enhances cancer cell invasion, metastasis and tumour growth. Annexin A2 increases cancer cell migration through the regulation of actin cytoskeletal rearrangement [151]. Moreover, Shiozawa et al. reported that annexin A2 increases cancer cell proliferation and cell survival via the ERK1/2 and MAPK pathway [11]. Annexin A2 is also a substrate for src kinase which regulates phosphorylation of annexin A2 at tyrosine 23 and phosphorylated annexin A2 have been shown to enhance cancer cell invasion [32]. In this study, we found highly abundant annexin A2 and S100A10 expression in the cancer associated stroma of serous ovarian cancer. Stromal annexin A2 and S100A10 may interact with other ECM proteins and proteases such as cathepsin B [53], plasminogen [65] and tenascin C [208] in the tumour microenvironment. Ovarian cancer stroma consists of various ECM components and pro-inflammatory factors that play an important role in the development and progression of ovarian cancer (reviewed in [209]). A recent study identified S100A10 as one of the proteins increased in co-culture of ovarian tumour-stromal cells suggesting that it plays a role in the stromal-epithelial cell crosstalk which is important for ovarian cancer progression [210]. Moreover, the tumour stroma has been shown to be a novel source of cancer biomarkers associated with poor prognosis [211]. Figure 6.2 summarized the effects of annexin A2 inhibitors such as annexin A2 siRNAs and neutralizing annexin A2 antibodies on cancer cell invasion and metastasis. Down-regulation of annexin A2, decreases plasmin levels consequently decreases extracellular matrix remodeling and cancer cell invasion and metastasis. Zhao et al. reported that decreased annexin A2 levels resulted in decreased MMP-2 and MMP-9 levels and reduced hepatocellular carcinoma cell

160

invasion and metastasis [151]. Moreover, reduced levels of annexin A2 resulted in decreased cancer cell migration [151] and increased cancer cell apoptosis [150]. Therefore, annexin A2 has important roles in cancer cell invasion, migration and apoptosis.

6.1.3 Clinical significance of annexin A2 in serous ovarian cancer patients In chapter 4, we assessed whether annexin A2 can be used as a diagnostic marker for ovarian cancer. We observed that blood levels of annexin A2 were significantly elevated in serous ovarian cancer patients compared with non-cancer controls. Importantly, blood levels of annexin A2 were elevated in early stage ovarian cancer compared to patients with benign ovarian tumours. We showed that combining annexin A2 and CA125 measurements improved the sensitivity of detecting early stage ovarian cancer at the 95% specificity level. Sun et al. reported that combined serum annexin A2 and alpha-fetoprotein (AFP) measurements improved the sensitivity of hepatocellular carcinoma (HCC) diagnosis [158]. Our preliminary findings in a small cohort of patients also showed that annexin A2 has the clinical potential to improve the sensitivity of distinguishing early serous ovarian cancer from benign ovarian tumours and in cancers with normal CA125 levels. However, this observation needs to be further validated in a larger cohort to assess the utility of combined annexin A2 and CA125 measurements. Chapter 5 demonstrated the potential use of either annexin A2 or S100A10 as a prognostic marker for ovarian cancer. We found that stromal annexin A2 expression was associated with reduced progression-free and overall survival in stage III serous ovarian cancer cohort. Emoto et al. reported annexin A2 expression in the stromal cells of colorectal cancer [101] and secreted form of annexin A2 has been shown in pancreatic cancer cell lines [43]. However, the association of annexin A2 expression in the ovarian cancer associated stromal cells and the secreted form of annexin A2 in the conditioned media [172] is poorly understood. Therefore, future studies are required to determine the significance of annexin A2 and S100A10 expression with other extracellular matrix proteins in the ovarian cancer associated stromal cells and importance of cleaved or secreted annexin A2 in ovarian cancer progression. Our findings suggest that annexin A2 and S100A10 could be used as novel prognostic marker for serous ovarian cancer patients and targeting stromal annexin A2 has a potential clinical application in the treatment of stage III ovarian cancer. We showed no significant difference of annexin A2 expression with increasing clinical stages of ovarian cancer (stage I to IV) (Supplementary Table 3.1). Characterization of annexin A2 expression should also be performed on type I (low grade tumours) and type II (high grade

161

tumours) ovarian cancer subtypes [212]. However, in our patient cohort (n=91), we did not observe any relationship between annexin A2 stromal intensity in the well or moderate tumour grade versus the poor differentiated high grade tumours (Table 5.2). Ovarian cancer is a heterogeneous disease and poor survival rate of ovarian cancer patients demands a better approach and treatment strategies [213]. Classification of ovarian cancer into molecular subtypes and profile through high- throughput genomic approaches [214, 215] may enable us to understand better the pathogenesis of the disease and assist in the development of effective novel therapies.

6.1.4 Co-localization of annexin A2 and S100A10 proteins in ovarian cancer cells A significant correlation between annexin A2 and S100A10 expression in the epithelial cells and stromal cells was observed in stage III serous ovarian cancer tissues (Figure 5.1 & Supplementary Table 5.1).The expression of annexin A2 on the cell surface is dependent on the co-localization with S100A10 protein [22] and previous studies have showed depletion of annexin A2 resulted in loss of S100A10 expression at the mRNA levels in breast cancer cells [47] and murine lymphoma cells [78]. He et al. showed the cytoplasmic level of S100A10 would determine the levels of cell surface of annexin A2 heterotetramer, as binding of annexin A2 stabilized intracellular S100A10 and masking autonomous S100A10 polyubiquitination signal that led to proteasomal degradation [216]. S100A10 protein is also required for src kinase mediated tyrosine phosphorylation of annexin A2 which translocate both annexin A2 and S100A10 proteins on the cell surface [22, 216]. Moreover, a recent study reported no changes in annexin A2 levels but increased in phosphorylated annexin A2 levels associated with src kinase and S100A10 expression in prednisolone resistance acute lymphoblastic leukemia cells compared with prednisolone sensitive cells [217]. Myrvang et al. also showed protein interactions between annexin A2 and S100A10 in MDA-MB-231 cells mediates breast cancer cell adhesion to the microvascular endothelial cells [218]. Therefore, these studies showed it is important to characterize the expression and roles of both proteins as annexin A2 and S100A10 as they may have synergistic functions in mediating cancer cells behavior.

162

6.2 Future directions This study highlights annexin A2 playing an important role in ovarian cancer progression; however, the mechanisms whereby annexin A2 interacts with other ECM proteins and proteases in mediating ovarian cancer metastasis are poorly understood. Future studies are required to investigate the mechanism of annexin A2 mediated t-PA dependent plasmin generation and the importance of tyrosine phosphorylated annexin A2 in ovarian cancer invasion and metastasis. Moreover, S100A10 has been shown to play an important role in invasion of bowel [195] and lung cancers [219]; therefore, the functional role of S100A10 in ovarian cancer motility, invasion and adhesion needs to be further evaluated. Recent studies showed that SKOV-3 cell line, which has been extensively used in ovarian cancer studies was unlikely to represent high-grade serous ovarian cancer based on their molecular profile characterization [220]. Therefore, other high-grade serous ovarian cancer cell lines and primary ovarian cancer cell lines should also be used in future preclinical research to examine annexin A2 expression and its functional role in serous ovarian cancer invasion and metastasis. Our findings show that annexin A2 has clinical significance in ovarian cancer diagnosis and prognosis. However, our findings are only limited to serous ovarian cancer patients. Annexin A2 expression remains to be characterized in other ovarian cancers such as the endometriod, clear cell and mucinous subtypes. Moreover, the sensitivity of combined annexin A2 and CA125 blood levels in ovarian cancer diagnosis could be improved with additional tumour biomarkers or patients’ clinical information such as ultrasound assessment and menopausal status. It will be also important to determine whether blood annexin A2 levels can be used to detect ovarian cancer recurrence and associated with the development of chemoresistance. Several studies reported annexin A2 playing a role in chemoresistance in pancreatic cancer cells [81, 160, 179] and breast cancer cells [108, 178, 221]. Limited studies have reported the relationship between annexin A2 with ATP binding cassette (ABC) transporters in chemoresistance and warrants further investigation [221]. Moreover, the association between annexin A2 or S100A10 expression with increasing clinical stages of ovarian cancer warrants further investigation in larger patient cohort. Future studies should investigate the effect of annexin A2 neutralizing antibodies on ovarian cancer cell survival together with ovarian cancer chemotherapy drugs including carboplatin or paclitaxel.

163

6.3 Conclusions Altogether, our findings strongly suggest that annexin A2 plays an important role in ovarian cancer invasion and metastasis. Annexin A2 is expressed in primary and metastatic serous ovarian cancer tissues, serous ovarian cancer cell lines and primary ovarian cancer cell lines. We identified a cleaved and secreted form of annexin A2 in the co-cultured ovarian cancer-peritoneal cells conditioned media. Moreover, annexin A2 siRNAs inhibited ovarian cancer cell motility, invasion and adhesion to the peritoneal cells in vitro. We also showed neutralizing annexin A2 antibodies inhibited OV-90 cells and primary ovarian cancer cells invasion in the CAM model. Moreover, neutralizing annexin A2 antibodies inhibited SKOV-3 cells tumour growth and peritoneal metastasis in the intraperitoneal xenograft mice model by increasing cancer cell apoptosis. Annexin A2 has an important clinical significance in ovarian cancer diagnosis and prognosis. Blood annexin A2 levels were elevated in serous ovarian cancer compared with patients with normal ovaries and benign ovarian tumours. Moreover, combined annexin A2 and CA125 levels improved the sensitivity of early stage ovarian cancer detection. High stromal annexin A2 expression was associated with reduced progression-free survival and overall survival in stage III serous ovarian cancer. Moreover, multivariate analysis showed high stromal annexin A2 expression to be an independent predictor of progression-free survival and overall survival in stage III serous ovarian cancer patient cohort together with the presence of residual disease. Annexin A2 plays an important role in ovarian cancer invasion and metastasis and has diagnostic and prognostic significance; and is therefore a novel promising biomarker that warrants further evaluation.

164

Figure 6.1 The mechanism of annexin A2 in cancer progression

Annexin A2 with p11 protein (S100A10) forms an annexin A2 heterotetramer complex on the plasma membrane which co-localize with CD44 [26], tenascin-C [51], cathepsin-B [53] and tissue- plasminogen activator (t-PA) [3]. Annexin A2 mediated t-PA dependent plasmin generation leads to activation of growth factors, activation of proteases including MMPs [151] and degradation of extracellular matrix (ECM) proteins [5] which increases cancer cell invasion, metastasis and tumour growth. Annexin A2 increases cancer cell proliferation and cell survival via the ERK1/2 and MAPK pathway [11]. Annexin A2 is a substrate for src kinase and regulates tyrosine 23 phosphorylation of annexin A2 and enhances cancer cell invasion [32]. Annexin A2 also plays a role in actin cytoskeletal rearrangement and regulates cancer cell migration [151].

165

Figure 6.2 The effects of annexin A2 inhibitors on cancer cell invasion and metastasis

Annexin A2 with p11 protein (S100A10) forms an annexin A2 heterotetramer complex on the plasma membrane which co-localized with CD44 molecule. Annexin A2 siRNAs and neutralizing annexin A2 antibodies (ANXA2 Ab) in which inhibit annexin A2 levels or function and inhibit the conversion of tissue-plasminogen activator (t-PA) dependent plasminogen activation on the cancer cells resulting in reduced plasmin levels. Consequently, decrease extracellular matrix remodelling due to inactivation of growth factors, proteases and MMPs result in decrease cancer cell invasion and metastasis. Inhibiting annexin A2 also decreases cancer cell migration and increases cancer cell apoptosis.

166

References 1. Ranson, M. and N.M. Andronicos, Plasminogen binding and cancer: promises and pitfalls. Front Biosci, 2003. 8: p. 294-304. 2. Hajjar, K.A. and S. Krishnan, Annexin II: A Mediator of the Plasmin /Plasminogen Activator System. Trends Cardiovas Med, 1999. 9(5): p. 128-138. 3. Cesarman, G.M., C.A. Guevara, and K.A. Hajjar, An endothelial cell receptor for plasminogen/tissue plasminogen activator (t-PA). II. Annexin II-mediated enhancement of t-PA-dependent plasminogen activation. J Biol Chem, 1994. 269(33): p. 21198-21203. 4. Hajjar, K.A., et al., Interaction of the Fibrinolytic Receptor, Annexin II, with the Endothelial Cell Surface. ESSENTIAL ROLE OF ENDONEXIN REPEAT 2. J Biol Chem, 1996. 271(35): p. 21652-21659. 5. Mai, J., D.M. Waisman, and B.F. Sloane, Cell surface complex of cathepsin B/annexin II tetramer in malignant progression. Biochim Biophys Acta 2000. 1477(1-2): p. 215-230. 6. Vishwanatha, J.K., et al., Enhanced expression of annexin II in human pancreatic carcinoma cells and primary pancreatic cancers. Carcinogenesis, 1993. 14(12): p. 2575- 2579. 7. Sharma, M.R., et al., Angiogenesis-associated protein annexin II in breast cancer: Selective expression in invasive breast cancer and contribution to tumor invasion and progression. Exp Mol Path, 2006. 81(2): p. 146-156. 8. Mohammad, H.S., et al., Annexin A2 expression and phosphorylation are up-regulated in hepatocellular carcinoma. Int J Oncol, 2008. 33(6): p. 1157-1163. 9. Inokuchi, J., et al., Annexin A2 positively contributes to the malignant phenotype and secretion of IL-6 in DU145 prostate cancer cells. Int J Cancer, 2009. 124(1): p. 68-74. 10. Diaz, V.M., et al., Specific interaction of tissue-type plasminogen activator (t-PA) with annexin II on the membrane of pancreatic cancer cells activates plasminogen and promotes invasion in vitro. Gut, 2004. 53(7): p. 993-1000. 11. Shiozawa, Y., et al., Annexin II/annexin II receptor axis regulates adhesion, migration, homing, and growth of prostate cancer. J Cell Biochem, 2008. 105(2): p. 370-380. 12. Tressler, R.J., et al., Extracellular annexin II is associated with divalent cation-dependent tumor cell-endothelial cell adhesion of metastatic RAW117 large-cell lymphoma cells. J Cell Biochem, 1993. 53(3): p. 265-276.

167

13. Gerke, V. and S.E. Moss, Annexins: From Structure to Function. Physiol Rev, 2002. 82(2): p. 331-371. 14. Mussunoor, S. and G.I. Murray, The role of annexins in tumour development and progression. J Pathol, 2008. 216(2): p. 131-140. 15. Yan, X., et al., Increased Expression of Annexin A3 Is a Mechanism of Platinum Resistance in Ovarian Cancer. Cancer Research, 2010. 70(4): p. 1616-1624. 16. Kim, A., et al., Targeting annexin A4 to counteract chemoresistance in clear cell carcinoma of the ovary. Expert Opinion on Therapeutic Targets, 2010. 14(9): p. 963-971. 17. Song, J., et al., Suppression of annexin A11 in ovarian cancer: implications in chemoresistance. Neoplasia, 2009. 11(6): p. 605-614. 18. Chuthapisith, S., et al., Annexins in human breast cancer: Possible predictors of pathological response to neoadjuvant chemotherapy. Eur J Cancer, 2009. 45(7): p. 1274- 1281. 19. Huebner, K., et al., Chromosomal localization of the human genes for lipocortin I and lipocortin II. Oncogene Res, 1988. 2(4): p. 299-310. 20. Rhiem, K., et al., Chromosomal region 15q21.1 is a frequent target of allelic imbalance in advanced breast carcinomas. International Journal of Cancer, 2003. 106(1): p. 74-77. 21. Waisman, D.M., Annexin II tetramer: structure and function. Mol Cell Biochem, 1995. 149(1): p. 301-322. 22. Deora, A.B., et al., An Annexin 2 Phosphorylation Switch Mediates p11-dependent Translocation of Annexin 2 to the Cell Surface. J Biol Chem, 2004. 279(42): p. 43411- 43418. 23. Bellagamba, C., et al., Tyrosine phosphorylation of annexin II tetramer is stimulated by membrane binding. J Biol Chem, 1997. 272(6): p. 3195-3199. 24. McCready, J., et al., Secretion of extracellular hsp90alpha via exosomes increases cancer cell motility: a role for plasminogen activation. BMC Cancer, 2010. 10: p. 294-303. 25. Lei, H., G. Romeo, and A. Kazlauskas, Heat Shock Protein 90{alpha}-Dependent Translocation of Annexin II to the Surface of Endothelial Cells Modulates Plasmin Activity in the Diabetic Rat Aorta. Circ Res, 2004. 94(7): p. 902-909. 26. Oliferenko, S., et al., Analysis of CD44-containing lipid rafts: Recruitment of annexin II and stabilization by the actin cytoskeleton. J Cell Biol, 1999. 146(4): p. 843-854.

168

27. Schnitzer, J.E., J. Liu, and P. Oh, Endothelial Caveolae Have the Molecular Transport Machinery for Vesicle Budding, Docking, and Fusion Including VAMP, NSF, SNAP, Annexins, and GTPases. J Biol Chem, 1995. 270(24): p. 14399-14404. 28. Kapoor, M., et al., Evidence for Annexin II-S100A10 Complex and Plasmin in Mobilization of Cytokine Activity of Human TrpRS. J Biol Chem, 2008. 283(4): p. 2070-2077. 29. Ohno, Y., et al., Annexin II represents metastatic potential in clear-cell renal cell carcinoma. Br J Cancer, 2009. 101(2): p. 287-294. 30. Duncan, R., et al., Characterisation and protein expression profiling of annexins in colorectal cancer. British Journal of Cancer, 2008. 98(2): p. 426-433. 31. Domoto, T., et al., Evaluation of S100A10, annexin II and B-FABP expression as markers for renal cell carcinoma. Cancer Sci, 2007. 98(1): p. 77-82. 32. Zheng, L., et al., Tyrosine 23 phosphorylation-dependent cell-surface localization of annexin A2 is required for invasion and metastases of pancreatic cancer. PLoS One, 2011. 6(4): p. e19390. 33. Reeves, S.A., et al., Developmental Regulation of Annexin II (Lipocortin 2) in Human Brain and Expression in High Grade Glioma. Cancer Res, 1992. 52(24): p. 6871-6876. 34. Yeatman, T.J., et al., Expression of annexins on the surfaces of non-metastatic and metastatic human and rodent tumor cells. Clin Exp Metastasis, 1993. 11(1): p. 37-44. 35. Chetcuti, A., et al., Loss of Annexin II Heavy and Light Chains in Prostate Cancer and Its Precursors. Cancer Res, 2001. 61(17): p. 6331-6334. 36. Liu, J.W., et al., Annexin II expression is reduced or lost in prostate cancer cells and its re- expression inhibits prostate cancer cell migration. Oncogene, 2003. 22(10): p. 1475-1485. 37. Xin, W., et al., Dysregulation of the Annexin Family Protein Family Is Associated with Prostate Cancer Progression. Am J Pathol, 2003. 162(1): p. 255-261. 38. Yee, D.S., et al., Reduced Annexin II Protein Expression in High-Grade Prostatic Intraepithelial Neoplasia and Prostate Cancer. Arch Path Lab Med, 2007. 131(6): p. 902- 908. 39. Banerjee, A., et al., Expression of biomarkers modulating prostate cancer angiogenesis: Differential expression of annexin II in prostate carcinomas from India and USA. Mol Cancer, 2003. 2(1): p. 34.

169

40. Wu, W., et al., Identification and validation of metastasis-associated proteins in head and neck cancer cell lines by two-dimensional electrophoresis and mass spectrometry. Clin Exp Metastasis, 2002. 19(4): p. 319-326. 41. Pena-Alonso, E., et al., Annexin A2 localizes to the basal epithelial layer and is down- regulated in dysplasia and head and neck squamous cell carcinoma. Cancer Lett, 2008. 263(1): p. 89-98. 42. Karimi-Busheri, F., et al., Expression of a releasable form of annexin II by human keratinocytes. J Cell Biochem, 2002. 86(4): p. 737-747. 43. Davis, R.G. and J.K. Vishwanatha, Detection of secreted and intracellular annexin II by a radioimmunoassay. J Immunol Methods, 1995. 188(1): p. 91-95. 44. Zhao, W.Q., et al., Secretion of Annexin II via Activation of Insulin Receptor and Insulin- like Growth Factor Receptor. J Biol Chem, 2003. 278(6): p. 4205-4215. 45. Ji, N.Y., et al., Evaluation of annexin II as a potential serum marker for hepatocellular carcinoma using a developed sandwich ELISA method. Int J Mol Med, 2009. 24(6): p. 765- 771. 46. Yan, G.R., et al., Characterization of Phosphoproteins in Gastric Cancer Secretome. OMICS, 2010. 15(1-2): p. 83-90. 47. Zhang, J., et al., Silencing of the annexin II gene down-regulates the levels of S100A10, c- Myc, and plasmin and inhibits breast cancer cell proliferation and invasion. Saudi Med J, 2010. 31(4): p. 374-381. 48. Puisieux, A., J. Ji, and M. Ozturk, Annexin II up-regulates cellular levels of p11 protein by a post-translational mechanisms. Biochem J, 1996. 313(1): p. 51-55. 49. Orend, G. and R. Chiquet-Ehrismann, Tenascin-C induced signaling in cancer. Cancer Letters, 2006. 244(2): p. 143-163. 50. I, E., et al., Tenascin C and annexin II expression in the process of pancreatic carcinogenesis. The Journal of Pathology, 2006. 208(5): p. 673-685. 51. Chung, C., J. Murphy-Ullrich, and H. Erickson, Mitogenesis, cell migration, and loss of focal adhesions induced by tenascin-C interacting with its cell surface receptor, annexin II. Mol Biol Cell, 1996. 7(6): p. 883-892. 52. Wilson, K.E., et al., Regulation and function of the extracellular matrix protein tenascin-C in ovarian cancer cell lines. Br J Cancer, 1999. 80(5-6): p. 685-692.

170

53. Mai, J., et al., Human Procathepsin B Interacts with the Annexin II Tetramer on the Surface of Tumor Cells. J Biol Chem, 2000. 275(17): p. 12806-12812. 54. Cavallo-Medved, D., et al., Live-cell imaging demonstrates extracellular matrix degradation in association with active cathepsin B in caveolae of endothelial cells during tube formation. Exp Cell Res, 2009. 315(7): p. 1234-1246. 55. Lu, G., et al., Cloning and Characterization of the Annexin II Receptor on Human Marrow Stromal Cells. J Biol Chem, 2006. 281(41): p. 30542-30550. 56. Wirl, G. and R. Schwartz-Albiez, Collagen-Binding proteins of mammary epithelial cells are related to Ca2+-and phospholipid-binding annexins. Journal of Cellular Physiology, 1990. 144(3): p. 511-522. 57. Das, S., et al., Signal transducer and activator of transcription 6 (STAT6) is a novel interactor of annexin A2 in prostate cancer cells. Biochemistry, 2010. 49(10): p. 2216- 2226. 58. Singh, P., et al., Annexin II binds progastrin and gastrin-like peptides, and mediates growth factor effects of autocrine and exogenous gastrins on colon cancer and intestinal epithelial cells. Oncogene, 2007. 26(3): p. 425-440. 59. Singh, P., Role of Annexin-II in GI cancers: Interaction with gastrins/progastrins. Cancer Letters, 2007. 252(1): p. 19-35. 60. Semov, A., et al., Metastasis-associated Protein S100A4 Induces Angiogenesis through Interaction with Annexin II and Accelerated Plasmin Formation. J Biol Chem, 2005. 280(21): p. 20833-20841. 61. Nedjadi, T., et al., S100A6 binds to annexin 2 in pancreatic cancer cells and promotes pancreatic cancer cell motility. Br J Cancer, 2009. 101(7): p. 1145-1154. 62. Kwaan, H.C., The plasminogen-plasmin system in malignancy. Cancer Metastasis Rev, 1992. 11(3): p. 291-311. 63. Sharma, M.C. and M. Sharma, The Role of Annexin II in Angiogenesis and Tumor Progression: A Potential Therapeutic Target. Curr Pharm Des, 2007. 13(35): p. 3568- 3575. 64. Kassam, G., et al., The p11 Subunit of the Annexin II Tetramer Plays a Key Role in the Stimulation of t-PA-Dependent Plasminogen Activation Biochemistry, 1998. 37(48): p. 16958-16966.

171

65. Hajjar, K.A., A.T. Jacovina, and J. Chacko, An endothelial cell receptor for plasminogen/tissue plasminogen activator. I. Identity with annexin II. J Biol Chem, 1994. 269(33): p. 21191-21197. 66. Sharma, M., R.T. Ownbey, and M.C. Sharma, Breast cancer cell surface annexin II induces cell migration and neoangiogenesis via tPA dependent plasmin generation. Exp Mol Path, 2010. 88(2): p. 278-286. 67. Coghlin, C. and G.I. Murray, Current and emerging concepts in tumour metastasis. J Pathol, 2010. 222(1): p. 1-15. 68. Paciucci, R., et al., The plasminogen activator system in pancreas cancer: role of t-PA in the invasive potential in vitro. Oncogene, 1998. 16(5): p. 625-633. 69. Roda, O., et al., New Insights into the tPA-Annexin A2 Interaction. IS ANNEXIN A2 CYS8 THE SOLE REQUIREMENT FOR THIS ASSOCIATION? J Biol Chem, 2003. 278(8): p. 5702-5709. 70. Zhang, X., et al., Arsenic trioxide, retinoic acid and Ara-c regulated the expression of annexin II on the surface of APL cells, a novel co-receptor for plasminogen/tissue plasminogen activator. Thromb Res, 2002. 106(1): p. 63-70. 71. Bao, H., et al., Overexpression of Annexin II affects the proliferation, apoptosis, invasion and production of proangiogenic factors in multiple myeloma. Int J Hematol, 2009. 90(2): p. 177-185. 72. Gillette, J.M., D.C. Chan, and S.M. Nielsen-Preiss, Annexin 2 expression is reduced in human osteosarcoma metastases. J Cell Biochem, 2004. 92(4): p. 820-832. 73. Jung, Y., et al., Annexin II expressed by osteoblasts and endothelial cells regulates stem cell adhesion, homing, and engraftment following transplantation. Blood, 2007. 110(1): p. 82-90. 74. Hayes, M.J., et al., Regulation of actin dynamics by annexin 2. EMBO J, 2006. 25: p. 1816-1826. 75. Thiel, C., M. Osborn, and V. Gerke, The tight association of the tyrosine kinase substrate annexin II with the submembranous cytoskeleton depends on intact p11- and Ca(2+)- binding sites. J Cell Sci, 1992. 103(3): p. 733-742. 76. Pu, Z., et al., Annexin II promotes invasion and migration of human hepatocellular carcinoma cells in vitro via its interaction with HAb18G/CD147. Cancer Sci, 2009. 101(2): p. 387-395.

172

77. Tatenhorst, L., et al., Knockdown of annexin 2 decreases migration of human glioma cells in vitro. Neuropathol Appl Neurobiol, 2006. 32(3): p. 271-277. 78. Hou, Y., et al., Annexin A2 Regulates the Levels of Plasmin, S100A10 and Fascin in L5178Y Cells. Cancer Invest, 2008. 26(8): p. 809-815. 79. Chen, J., et al., Proteomics of buccal squamous cell carcinoma: The involvement of multiple pathways in tumorigenesis. Proteomics, 2004. 4(8): p. 2465-2475. 80. Tomonaga, T., et al., Identification of Altered Protein Expression and Post-Translational Modifications in Primary Colorectal Cancer by Using Agarose Two-Dimensional Gel Electrophoresis. Clin Cancer Res, 2004. 10(6): p. 2007-2014. 81. Takano, S., et al., Annexin II Overexpression Predicts Rapid Recurrence after Surgery in Pancreatic Cancer Patients Undergoing Gemcitabine-Adjuvant Chemotherapy. Ann Surg Oncol, 2008. 15(11): p. 3157-3168. 82. Eustace, A.J., et al., 2D-DIGE analysis of phospho-enriched fractions from dasatinib- treated melanoma cell lines. J Proteomics, 2011. 74(4): p. 490-501. 83. Laumonnier, Y., et al., Identification of the annexin A2 heterotetramer as a receptor for the plasmin-induced signaling in human peripheral monocytes. Blood, 2006. 107(8): p. 3342- 3349. 84. He, K.L., et al., Feedback regulation of endothelial cell surface plasmin generation by PKC dependent phosphorylation of annexin A2. J Biol Chem, 2010. 286(17): p. 15428-15439. 85. Tsunezumi, J., et al., Matrilysin (matrix metalloprotease-7) cleaves membrane-bound annexin II and enhances binding of tissue-type plasminogen activator to cancer cell surfaces. FEBS J, 2008. 275(19): p. 4810-4823. 86. Wang, C.Y., et al., Glycogen Synthase Kinase-3 and Omi/HtrA2 Induce Annexin A2 Cleavage followed by Cell Cycle Inhibition and Apoptosis. Mol Biol Cell, 2009. 20(19): p. 4153-4161. 87. Wang, H., et al., Comprehensive proteome analysis of ovarian cancers using liquid phase separation, mass mapping and tandem mass spectrometry: a strategy for identification of candidate cancer biomarkers. Proteomics, 2004. 4(8): p. 2476-2495. 88. Gagne, J.P., et al., Comparative proteome analysis of human epithelial ovarian cancer. Proteome Sci, 2007. 5(1): p. 16. 89. Tchagang, A.B., et al., Early detection of ovarian cancer using group biomarkers. Mol Cancer Ther, 2008. 7(1): p. 27-37.

173

90. Gunawardana, C.G., et al., Comprehensive Analysis of Conditioned Media from Ovarian Cancer Cell Lines Identifies Novel Candidate Markers of Epithelial Ovarian Cancer. J Proteome Res, 2009. 8(10): p. 4705-4713. 91. Faca, V.M., et al., Proteomic analysis of ovarian cancer cells reveals dynamic processes of protein secretion and shedding of extra-cellular domains. PLoS One, 2008. 3(6): p. e2425. 92. Ween, M.P., et al., Transforming growth factor-beta-induced protein secreted by peritoneal cells increases the metastatic potential of ovarian cancer cells. Int J Cancer, 2011. 128(7): p. 1570-1584. 93. Sharma, M.R., et al., Antibody-directed targeting of angiostatin's receptor annexin II inhibits Lewis Lung Carcinoma tumor growth via blocking of plasminogen activation: Possible biochemical mechanism of angiostatin's action. Exp Mol Path, 2006. 81(2): p. 136-145. 94. Jacovina, A.T., et al., Neuritogenesis and the Nerve Growth Factor-induced Differentiation of PC-12 Cells Requires Annexin II-mediated Plasmin Generation. J Biol Chem, 2001. 276(52): p. 49350-49358. 95. Tuszynski, G.P., et al., Angiostatin Binds to Tyrosine Kinase Substrate Annexin II through the Lysine-Binding Domain in Endothelial Cells. Microvascular Research, 2002. 64(3): p. 448-462. 96. Braden, A.R., et al., Polymeric nanoparticles for sustained down-regulation of annexin A2 inhibit prostate tumor growth. J Nanosci Nanotechnol, 2009. 9(5): p. 2856-2865. 97. Kesavan, K., et al., Annexin A2 Is a Molecular Target for TM601, a Peptide with Tumor- targeting and Anti-angiogenic Effects. J Biol Chem, 2010. 285(7): p. 4366-4374. 98. Zimmermann, U., et al., Expression of annexin II in conventional renal cell carcinoma is correlated with Fuhrman grade and clinical outcome. Virchows Arch, 2004. 445(4): p. 368- 374. 99. Qi, Y.J., et al., Proteomic identification of malignant transformation-related proteins in esophageal squamous cell carcinoma. J Cell Biochem, 2008. 104(5): p. 1625-1635. 100. Syed, S.P., et al., Angiostatin receptor annexin II in vascular tumors including angiosarcoma. Hum Pathol, 2007. 38(3): p. 508-513. 101. Emoto, K., et al., Annexin II overexpression correlates with stromal tenascin-C overexpression. Cancer, 2001. 92(6): p. 1419-1426.

174

102. Nygaard, S.J., et al., Expression of annexin II in glioma cell lines and in brain tumor biopsies. J Neurooncol, 1998. 38(1): p. 11-18. 103. Roseman, B.J., et al., Annexin II marks astrocytic brain tumors of high histologic grade. Oncol Res, 1994. 6(12): p. 561-567. 104. Zhong, L.P., et al., Increased expression of Annexin A2 in oral squamous cell carcinoma. Arch Oral Biol, 2009. 54(1): p. 17-25. 105. Hellman, K., et al., Differential tissue-specific protein markers of vaginal carcinoma. Br J Cancer, 2009. 100(8): p. 1303-1314. 106. Bae, S.M., et al., Two-dimensional gel analysis of protein expression profile in squamous cervical cancer patients. Gynecol Oncol, 2005. 99(1): p. 26-35. 107. Olwill, S.A., et al., Annexin II cell surface and mRNA expression in human acute myeloid leukaemia cell lines. Thromb Res, 2005. 115(1-2): p. 109-114. 108. Zhang, F., et al., Anxa2 plays a critical role in enhanced invasiveness of the multidrug resistant human breast cancer cells. J Proteome Res, 2009. 8(11): p. 5041-5047. 109. Rescher, U., et al., Tyrosine phosphorylation of annexin A2 regulates Rho-mediated actin rearrangement and cell adhesion. J Cell Sci, 2008. 121(13): p. 2177-2185. 110. Huang, Y., et al., Involvement of Annexin A2 in p53 induced apoptosis in lung cancer. Mol Cell Biochem, 2008. 309(1): p. 117-123. 111. Masaki, T., et al., Involvement of annexin I and annexin II in hepatocyte proliferation: can annexins I and II be markers for proliferative hepatocytes? Hepatology, 1994. 20(2): p. 425-435. 112. Lengyel, E., Ovarian cancer development and metastasis. Am J Pathol, 2010. 177(3): p. 1053-1064. 113. Richardson, M. and G. Singh, Observations on the use of the avian chorioallantoic membrane (CAM) model in investigations into angiogenesis. Current Drug Targets- Cardiovascular; Hematological Disorders, 2003. 3(2): p. 155-185. 114. Deryugina, E.I. and J.P. Quigley, Chick embryo chorioallantoic membrane model systems to study and visualize human tumor cell metastasis. Histochem Cell Biol, 2008. 130(6): p. 1119-1130. 115. Tufan, A.C. and N.L. Satiroglu-Tufan, The chick embryo chorioallantoic membrane as a model system for the study of tumor angiogenesis, invasion and development of anti- angiogenic agents. Curr Cancer Drug Targets, 2005. 5(4): p. 249-266.

175

116. Ossowski, L., In vivo invasion of modified chorioallantoic membrane by tumor cells: the role of cell surface-bound urokinase. J Cell Biol, 1988. 107(6): p. 2437-2445. 117. Zhai, Y., et al., Gene expression analysis of preinvasive and invasive cervical squamous cell carcinomas identifies HOXC10 as a key mediator of invasion. Cancer Res, 2007. 67(21): p. 10163-10172. 118. Cimpean, A.M., D. Ribatti, and M. Raica, The chick embryo chorioallantoic membrane as a model to study tumor metastasis. Angiogenesis, 2008. 11(4): p. 311-319. 119. Valdes, T.I., D. Kreutzer, and F. Moussy, The chick chorioallantoic membrane as a novel in vivo model for the testing of biomaterials. Journal of Biomedical Materials Research, 2002. 62(2): p. 273-282. 120. Giannopoulou, E., et al., X-rays modulate extracellular matrix in vivo. International Journal of Cancer, 2001. 94(5): p. 690-698. 121. Subauste, M.C., et al., Evaluation of metastatic and angiogenic potentials of human colon carcinoma cells in chick embryo model systems. Clin Exp Metastasis, 2009. 26(8): p. 1033-1047. 122. Demir, R., et al., Malignant progression of invasive tumour cells seen in hypoxia present an accumulation of beta-catenin in the nucleus at the tumour front. Exp Mol Pathol, 2009. 87(2): p. 109-116. 123. Strojnik, T., et al., Experimental model and immunohistochemical comparison of U87 human glioblastoma cell xenografts on the chicken chorioallantoic membrane and in rat brains. Anticancer Res, 2010. 30(12): p. 4851-4860. 124. Balciuniene, N., et al., Histology of human glioblastoma transplanted on chicken chorioallantoic membrane. Medicina (Kaunas), 2009. 45(2): p. 123-131. 125. Hagedorn, M., et al., Accessing key steps of human tumor progression in vivo by using an avian embryo model. Proc Natl Acad Sci U S A, 2005. 102(5): p. 1643-1648. 126. Kobayashi, T., et al., A chick embryo model for metastatic human prostate cancer. European urology, 1998. 34(2): p. 154-160. 127. Wittig-Blaich, S.M., et al., Matrix-dependent regulation of AKT in Hepsin-overexpressing PC3 prostate cancer cells. Neoplasia, 2011. 13(7): p. 579-589. 128. Conn, E.M., et al., Comparative Analysis of Metastasis Variants Derived from Human Prostate Carcinoma Cells: Roles in Intravasation of VEGF-Mediated Angiogenesis and uPA-Mediated Invasion. The American Journal of Pathology, 2009. 175(4): p. 1638-1652.

176

129. Taizi, M., et al., A novel and rapid in vivo system for testing therapeutics on human leukemias. Exp Hematol, 2006. 34(12): p. 1698-708. 130. Balke, M., et al., Morphologic characterization of osteosarcoma growth on the chick chorioallantoic membrane. BMC Research Notes, 2010. 3(1): p. 58. 131. Chang, H.L., et al., Mullerian inhibiting substance inhibits invasion and migration of epithelial cancer cell lines. Gynecol Oncol, 2011. 120(1): p. 128-134. 132. Lokman, N.A., et al., The role of annexin A2 in tumorigenesis and cancer progression. Cancer Microenviron, 2011. 4(2): p. 199-208. 133. Ween, M.P., et al., Versican induces a pro-metastatic ovarian cancer cell behavior which can be inhibited by small hyaluronan oligosaccharides. Clin Exp Metastasis, 2011. 28(2): p. 113-125. 134. Deryugina, E.I. and J.P. Quigley, Chapter 2. Chick embryo chorioallantoic membrane models to quantify angiogenesis induced by inflammatory and tumor cells or purified effector molecules. Methods Enzymol, 2008. 444: p. 21-41. 135. Manenti, L., et al., Circulating plasma vascular endothelial growth factor in mice bearing human ovarian carcinoma xenograft correlates with tumor progression and response to therapy. Mol Cancer Ther, 2005. 4(5): p. 715-725. 136. Fang, J., et al., Apigenin inhibits tumor angiogenesis through decreasing HIF-1α and VEGF expression. Carcinogenesis, 2006. 28(4): p. 858-864. 137. Luo, H., et al., Kaempferol inhibits angiogenesis and VEGF expression through both HIF dependent and independent pathways in human ovarian cancer cells. Nutr Cancer, 2009. 61(4): p. 554-563. 138. Liu, L.Z., et al., Acacetin inhibits VEGF expression, tumor angiogenesis and growth through AKT/HIF-1alpha pathway. Biochem Biophys Res Commun, 2011. 413(2): p. 299- 305. 139. Kisker, O., et al., Generation of multiple angiogenesis inhibitors by human pancreatic cancer. Cancer research, 2001. 61(19): p. 7298. 140. Petitclerc, E. and P.C. Brooks, The chick embryo metastasis model. Methods Mol Med, 2001. 58: p. 173-187. 141. Zijlstra, A., et al., A quantitative analysis of rate-limiting steps in the metastatic cascade using human-specific real-time polymerase chain reaction. Cancer Res, 2002. 62(23): p. 7083-7092.

177

142. Aguirre-Ghiso, J.A., L. Ossowski, and S.K. Rosenbaum, Green fluorescent protein tagging of extracellular signal-regulated kinase and p38 pathways reveals novel dynamics of pathway activation during primary and metastatic growth. Cancer research, 2004. 64(20): p. 7336. 143. Lyu, M.A., et al., Over-expression of urokinase receptor in human epidermoid-carcinoma cell line (HEp3) increases tumorigenicity on chorio-allantoic membrane and in severe- combined-immunodeficient mice. Int J Cancer, 1998. 77(2): p. 257-263. 144. Bekes, E.M., et al., Activation of pro-uPA is critical for initial escape from the primary tumor and hematogenous dissemination of human carcinoma cells. Neoplasia, 2011. 13(9): p. 806-821. 145. Mullany, L.K. and J.A.S. Richards, Minireview: Animal Models and Mechanisms of Ovarian Cancer Development. Endocrinology, 2012. 153(4): p. 1585-1592. 146. Ray, P., et al., Noninvasive imaging reveals inhibition of ovarian cancer by targeting CXCL12-CXCR4. Neoplasia (New York, NY), 2011. 13(12): p. 1152. 147. Sale, S. and S. Orsulic, Models of ovarian cancer metastasis: Murine models. Drug Discov Today Dis Models, 2006. 3(2): p. 149-154. 148. Siegel, R., D. Naishadham, and A. Jemal, Cancer statistics, 2013. CA: A Cancer Journal for Clinicians, 2013. 63(1): p. 11-30. 149. Madureira, P.A., et al., Annexin A2 is a novel cellular redox regulatory protein involved in tumorigenesis. Oncotarget, 2011. 2(12): p. 1075-1093. 150. Zhai, H., et al., Annexin A2 promotes glioma cell invasion and tumor progression. The Journal of Neuroscience, 2011. 31(40): p. 14346-14360. 151. Zhao, P., et al., Annexin II promotes invasion and migration of human hepatocellular carcinoma cells in vitro via its interaction with HAb18G/CD147. Cancer science, 2010. 101(2): p. 387-395. 152. Sodek, K.L., et al., Identification of pathways associated with invasive behavior by ovarian cancer cells using multidimensional protein identification technology (MudPIT). Mol Biosyst, 2008. 4(7): p. 762-773. 153. Bengtsson, S., et al., Large-Scale Proteomics Analysis of Human Ovarian Cancer for Biomarkers. Journal of Proteome Research, 2007. 6(4): p. 1440-1450. 154. Livak, K.J. and T.D. Schmittgen, Analysis of Relative Gene Expression Data Using Real- Time Quantitative PCR and the 2−ΔΔCT Method. Methods, 2001. 25(4): p. 402-408.

178

155. Ricciardelli, C., et al., The ADAMTS1 protease gene is required for mammary tumor growth and metastasis. Am J Pathol, 2011. 179(6): p. 3075-3085. 156. Lokman, N.A., et al., Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis. Int J Mol Sci, 2012. 13(8): p. 9959-9970. 157. Shetty, P.K., et al., Reciprocal Regulation of Annexin A2 and EGFR with Her-2 in Her-2 Negative and Herceptin-Resistant Breast Cancer. PLoS One, 2012. 7(9): p. e44299. 158. Sun, Y., et al., Annexin A2 is a discriminative serological candidate in early hepatocellular carcinoma. Carcinogenesis, 2013. 34(3): p. 595-604. 159. Deryugina, E.I. and J.P. Quigley, Cell Surface Remodeling by Plasmin: A New Function for an Old Enzyme. Journal of Biomedicine and Biotechnology, 2012. 2012. 160. Sato, T., et al., Extracellular release of annexin II from pancreatic cancer cells and resistance to anticancer drug-induced apoptosis by supplementation of recombinant annexin II. Pancreas, 2012. 41(8): p. 1247-1254. 161. Brownstein, C., et al., Annexin II ; A Mediator of Cell Surface-Specific Plasmin Generation. Annals of the New York Academy of Sciences, 2001. 947: p. 143-156. 162. Sharma, M., M.R. Blackman, and M.C. Sharma, Antibody-directed neutralization of annexin II (ANX II) inhibits neoangiogenesis and human breast tumor growth in a xenograft model. Experimental and Molecular Pathology, 2012. 92(1): p. 175-184. 163. Lim, D. and E. Oliva, Precursors and pathogenesis of ovarian carcinoma. Pathology, 2013. 45(3): p. 229-242. 164. Shepherd, J.H., Revised FIGO staging for gynaecological cancer. BJOG: An International Journal of Obstetrics & Gynaecology, 1989. 96(8): p. 889-892. 165. Coleman, R.L., et al., Latest research and treatment of advanced-stage epithelial ovarian cancer. Nat Rev Clin Oncol, 2013. 10(4): p. 211-224. 166. Bast Jr, R.C., et al., A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. New England Journal of Medicine, 1983. 309(15): p. 883-887. 167. Mann, W.J., et al., Preoperative Serum CA-125 Levels in Patients With Surgical Stage I Invasive Ovarian Adenocarcinoma. Journal of the National Cancer Institute, 1988. 80(3): p. 208-209.

179

168. Shiau, C.-S., et al., Ovarian endometrioma associated with very high serum CA-125 levels. Chang Gung medical journal, 2003. 26(9): p. 695-699. 169. Buamah, P., Benign conditions associated with raised serum CA-125 concentration. Journal of Surgical Oncology, 2000. 75(4): p. 264-265. 170. Akobeng, A.K., Understanding diagnostic tests 1: sensitivity, specificity and predictive values. Acta Paediatr, 2007. 96(3): p. 338-341. 171. Buys, S.S., et al., Effect of screening on ovarian cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Randomized Controlled Trial. JAMA, 2011. 305(22): p. 2295-2303. 172. Lokman, N.A., et al., Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis. Oncotarget, 2013. 4 (8): p. 1199-1211. 173. Zhang, H.J., et al., Expression characteristics and diagnostic value of annexin A2 in hepatocellular carcinoma. World J Gastroenterol, 2012. 18(41): p. 5897-5904. 174. Jeon, Y.R., et al., Identification of annexin II as a novel secretory biomarker for breast cancer. PROTEOMICS, 2013. 13(21): p. 3145-3156. 175. Seckinger, A., et al., Clinical and prognostic role of annexin A2 in multiple myeloma. Blood, 2012. 120(5): p. 1087-1094. 176. Weeks, M.E., et al., Analysis of the urine proteome in patients with pancreatic ductal adenocarcinoma. Proteomics Clin Appl, 2008. 2(7-8): p. 1047-1057. 177. Zhou, H., et al., Collection, storage, preservation, and normalization of human urinary exosomes for biomarker discovery. Kidney Int, 2006. 69(8): p. 1471-1476. 178. Zhang, F., et al., P-glycoprotein associates with Anxa2 and promotes invasion in multidrug resistant breast cancer cells. Biochem Pharmacol, 2013. 87(2): p. 292-302. 179. Gong, X.-G., et al., Gemcitabine resistance induced by interaction between alternatively spliced segment of tenascin-C and annexin A2 in pancreatic cancer cells. Biological and Pharmaceutical Bulletin, 2010. 33(8): p. 1261-1267. 180. Skates, S.J., et al., Preoperative sensitivity and specificity for early-stage ovarian cancer when combining cancer antigen CA-125II, CA 15-3, CA 72-4, and macrophage colony- stimulating factor using mixtures of multivariate normal distributions. Journal of clinical oncology, 2004. 22(20): p. 4059-4066.

180

181. Sandri, M.T., et al., Comparison of HE4, CA125 and ROMA algorithm in women with a pelvic mass: correlation with pathological outcome. Gynecol Oncol, 2013. 128(2): p. 233- 238. 182. Moore, R.G., et al., A novel multiple marker bioassay utilizing HE4 and CA125 for the prediction of ovarian cancer in patients with a pelvic mass. Gynecol Oncol, 2009. 112(1): p. 40-46. 183. Jacobs, I., et al., A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer. Br J Obstet Gynaecol, 1990. 97(10): p. 922-929. 184. Jia, J.-W., et al., Clinical significance of annexin II expression in human non-small cell lung cancer. Tumor Biology, 2013. 34(3): p. 1767-1771. 185. Wang, C.-Y., et al., Annexin A2 Silencing Induces G2 Arrest of Non-small Cell Lung Cancer Cells through p53-dependent and -independent Mechanisms. Journal of Biological Chemistry, 2012. 287(39): p. 32512-32524. 186. Yang, T., et al., Prognostic and Diagnostic Significance of Annexin A2 in Colorectal Cancer. Colorectal Disease, 2013. 15(7): p. 373-381. 187. Yonglitthipagon, P., et al., Up-regulation of annexin A2 in cholangiocarcinoma caused by Opisthorchis viverrini and its implication as a prognostic marker. International Journal for Parasitology, 2010. 40(10): p. 1203-1212. 188. EMOTO, K., et al., Annexin II overexpression is correlated with poor prognosis in human gastric carcinoma. Anticancer research, 2001. 21(2B): p. 1339-1345. 189. Gao, H., et al., Correlation of expression levels of ANXA2, PGAM1, and CALR with glioma grade and prognosis. Journal of Neurosurgery, 2013. 118(4): p. 846-853. 190. Bharadwaj, A., et al., Annexin A2 heterotetramer: structure and function. Int J Mol Sci, 2013. 14(3): p. 6259-6305. 191. Donato, R., S100: a multigenic family of calcium-modulated proteins of the EF-hand type with intracellular and extracellular functional roles. Int J Biochem Cell Biol, 2001. 33(7): p. 637-668. 192. Zobiack, N., V. Gerke, and U. Rescher, Complex formation and submembranous localization of annexin 2 and S100A10 in live HepG2 cells. FEBS Letters, 2001. 500(3): p. 137-140.

181

193. Madureira, P.A., et al., The biochemistry and regulation of S100A10: a multifunctional plasminogen receptor involved in oncogenesis. J Biomed Biotechnol, 2012. 2012: p. 353687. 194. Shang, J., et al., S100A10 as a novel biomarker in colorectal cancer. Tumor Biology, 2013. 34(6): p. 3785-3790. 195. Zhang, L., D.K. Fogg, and D.M. Waisman, RNA Interference-mediated Silencing of the S100A10 Gene Attenuates Plasmin Generation and Invasiveness of Colo 222 Colorectal Cancer Cells. Journal of Biological Chemistry, 2004. 279(3): p. 2053-2062. 196. Ito, Y., et al., S100A10 expression in thyroid neoplasms originating from the follicular epithelium: contribution to the aggressive characteristic of anaplastic carcinoma. Anticancer Res, 2007. 27(4C): p. 2679-2683. 197. Gillet, J.-P., et al., Multidrug resistance–linked gene signature predicts overall survival of patients with primary ovarian serous carcinoma. Clinical Cancer Research, 2012. 18(11): p. 3197-3206. 198. Elsner, M., et al., MALDI imaging mass spectrometry reveals COX7A2, TAGLN2 and S100-A10 as novel prognostic markers in Barrett's adenocarcinoma. Journal of Proteomics, 2012. 75(15): p. 4693-4704. 199. Tan, Y., et al., Proteomic-based analysis for identification of potential serum biomarkers in gallbladder cancer. Oncology Reports, 2011. 26(4): p. 853-859. 200. Pavlides, S., et al., The reverse Warburg effect: aerobic glycolysis in cancer associated fibroblasts and the tumor stroma. Cell Cycle, 2009. 8(23): p. 3984-4001. 201. Jin, L., et al., Immunohistochemical expression of Annexin A2 and S100A proteins in patients with bulky stage IB–IIA cervical cancer treated with neoadjuvant chemotherapy. Gynecologic Oncology, 2012. 126(1): p. 140-146. 202. Kagawa, S., et al., Akt/mTOR signaling pathway is crucial for gemcitabine resistance induced by Annexin II in pancreatic cancer cells. Journal of Surgical Research, 2012. 178(2): p. 758-767. 203. Ding, T., et al., Down-regulation of annexin II in prostate cancer is associated with Gleason score, recurrence, metastasis and poor prognosis. Mol Med Rep, 2010. 3(5): p. 781-787.

182

204. Giráldez, M.D., et al., Gene-expression signature of tumor recurrence in patients with stage II and III colon cancer treated with 5′fluoruracil-based adjuvant chemotherapy. International Journal of Cancer, 2013. 132(5): p. 1090-1097. 205. Yao, H., et al., Identification of metastasis associated proteins in human lung squamous carcinoma using two-dimensional difference gel electrophoresis and laser capture microdissection. Lung Cancer, 2008. 65(2009): p. 41-48. 206. Ween, M.P., M.K. Oehler, and C. Ricciardelli, Role of Versican, Hyaluronan and CD44 in Ovarian Cancer Metastasis. Int J Mol Sci, 2011. 12(2): p. 1009-29. 207. McMahon, B. and H.C. Kwaan, The plasminogen activator system and cancer. Pathophysiol Haemost Thromb, 2008. 36(3-4): p. 184-194. 208. Chung, C.Y. and H.P. Erickson, Cell surface annexin II is a high affinity receptor for the alternatively spliced segment of tenascin-C. J Cell Biol, 1994. 126(2): p. 539-548. 209. Furuya, M., Ovarian cancer stroma: pathophysiology and the roles in cancer development. Cancers (Basel), 2012. 4(3): p. 701-724. 210. Fu, S., et al., Stromal–Epithelial Crosstalk Provides a Suitable Microenvironment for the Progression of Ovarian Cancer Cells in Vitro. Cancer Investigation, 2013. 31(9): p. 616- 624. 211. Sund, M. and R. Kalluri, Tumor stroma derived biomarkers in cancer. Cancer and Metastasis Reviews, 2009. 28(1-2): p. 177-183. 212. Ricciardelli, C. and M.K. Oehler, Diverse molecular pathways in ovarian cancer and their clinical significance. Maturitas, 2009. 62(3): p. 270-275. 213. Vaughan, S., et al., Rethinking ovarian cancer: recommendations for improving outcomes. Nat Rev Cancer, 2011. 11(10): p. 719-25. 214. Tothill, R.W., et al., Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome. Clin Cancer Res, 2008. 14(16): p. 5198-208. 215. Integrated genomic analyses of ovarian carcinoma. Nature, 2011. 474(7353): p. 609-15. 216. He, K.L., et al., Endothelial cell annexin A2 regulates polyubiquitination and degradation of its binding partner S100A10/p11. J Biol Chem, 2008. 283(28): p. 19192-19200. 217. Spijkers-Hagelstein, J.A.P., et al., Src kinase-induced phosphorylation of annexin A2 mediates glucocorticoid resistance in MLL-rearranged infant acute lymphoblastic leukemia. Leukemia, 2013. 27(5): p. 1063-1071.

183

218. Myrvang, H.K., et al., Protein interactions between surface annexin A2 and S100A10 mediate adhesion of breast cancer cells to microvascular endothelial cells. FEBS Letters, 2013. 587(19): p. 3210-3215. 219. Yang, X., N.C. Popescu, and D.B. Zimonjic, DLC1 interaction with S100A10 mediates inhibition of in vitro cell invasion and tumorigenicity of lung cancer cells through a RhoGAP-independent mechanism. Cancer research, 2011. 71(8): p. 2916-2925. 220. Domcke, S., et al., Evaluating cell lines as tumour models by comparison of genomic profiles. Nat Commun, 2013. 4: p. 2126. 221. Zhang, H.C., et al., Identification of the Interaction between P-Glycoprotein and Anxa2 in Multidrug-resistant Human Breast Cancer Cells. Cancer Biol Med, 2012. 9(2): p. 99-104.

184

7 Appendix

Publications Contributing to This Thesis

1. The Role of Annexin A2 in Tumourigenesis and Cancer Progression 2. Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis 3. Annexin A2 is regulated by Ovarian Cancer-Peritoneal Cell Interactions and Promotes Metastasis

185

A Lokman, N.A., Ween, M.P., Oehler, M.K. & Ricciardelli, C. (2011) The role of annexin A2 in tumorigenesis and cancer progression. Cancer Microenvironment, v. 4(2), pp. 199-208

NOTE: This publication is included on pages 186-195 in the print copy of the thesis held in the University of Adelaide Library.

It is also available online to authorised users at:

http://doi.org/10.1007/s12307-011-0064-9

Int. J. Mol. Sci. 2012, 13, 9959-9970; doi:10.3390/ijms13089959 OPEN ACCESS International Journal of Molecular Sciences ISSN 1422-0067 www.mdpi.com/journal/ijms Article Chick Chorioallantoic Membrane (CAM) Assay as an In Vivo Model to Study the Effect of Newly Identified Molecules on Ovarian Cancer Invasion and Metastasis

Noor A. Lokman 1, Alison S. F. Elder 1, Carmela Ricciardelli 1 and Martin K. Oehler 1,2,*

1 Research Centre for Reproductive Health, School of Paediatrics and Reproductive Health, Robinson Institute, University of Adelaide, Adelaide, South Australia 5000, Australia; E-Mails: [email protected] (N.A.L.); [email protected] (A.S.F.E.); [email protected] (C.R.) 2 Department of Gynaecological Oncology, Royal Adelaide Hospital, Adelaide, South Australia 5000, Australia

* Author to whom correspondence should be addressed; E-Mail: [email protected]; Tel.: +61-8-8222-4816; Fax: +61-8-8222-5952.

Received: 2 July 2012; in revised form: 27 July 2012 / Accepted: 2 August 2012 / Published: 10 August 2012

Abstract: The majority of ovarian cancer patients present with advanced disease and despite aggressive treatment, prognosis remains poor. Significant improvement in ovarian cancer survival will require the development of more effective molecularly targeted therapeutics. Commonly, mouse models are used for the in vivo assessment of potential new therapeutic targets in ovarian cancer. However, animal models are costly and time consuming. Other models, such as the chick embryo chorioallantoic membrane (CAM) assay, are therefore an attractive alternative. CAM assays have been widely used to study angiogenesis and tumor invasion of colorectal, prostate and brain cancers. However, there have been limited studies that have used CAM assays to assess ovarian cancer invasion and metastasis. We have therefore developed a CAM assay protocol to monitor the metastatic properties of ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) and to study the effect of potential therapeutic molecules in vivo. The results from the CAM assay are consistent with cancer cell motility and invasion observed in in vitro assays. Our results demonstrate that the CAM assay is a robust and cost effective model to study ovarian cancer cell metastasis. It is therefore a very useful in vivo model for screening of potential novel therapeutics. Int. J. Mol. Sci. 2012, 13 9960

Keywords: ovarian cancer; invasion; metastasis; chick chorioallantoic membrane

1. Introduction

Ovarian cancer is the most lethal gynecological malignancy. Most patients are diagnosed at an advanced stage when the cancer cells have already metastasized to the abdominal cavity. Ovarian cancer metastasis is characterized by the ability of ovarian cancer cells to detach from the ovary and to adhere and invade the peritoneal cell layer, which lines the organs in the abdominal cavity [1]. The development of new therapies with the aim to disrupt ovarian cancer metastasis requires the in vivo study of novel targets and molecules. However, commonly used murine models are costly and require a large number of animals as well as a long experimental time frame. An attractive alternative is the chick chorioallantoic membrane (CAM) assay. CAM assays have been widely used to study angiogenesis [2], tumor cell invasion and metastasis [3–6]. The CAM model has many advantages, such as (a) the highly vascularized nature of the CAM greatly promotes the efficiency of tumor cell grafting; (b) high reproducibility; (c) simplicity and cost effectiveness, and finally (d) as the CAM assay is a closed system, the half-life of many experimental molecules such as small peptides tends to be much longer in comparison to animal models, allowing experimental study of potential anti-metastatic compounds that are only available in small quantities [4,7]. The CAM is composed of a multilayer epithelium; the ectoderm at the air interface, mesoderm (or stroma) and endoderm at the interface with the allantoic sac [8]. Furthermore, the CAM contains extracellular matrix proteins (ECM) such as fibronectin, laminin, collagen type I and integrin ανβз [9]. The presence of these extracellular matrix proteins mimics the physiological cancer cell environment. Although the CAM assay is a well established model for studying tumour angiogenesis and invasion in malignancies such as bowel cancer [10,11], glioma [12–14], prostate cancer [15–17], leukemia [18] and osteosarcoma [19], there has only been one study to date that has used a CAM assay to assess ovarian cancer invasion and metastasis [20]. We recently investigated the ovarian cancer-peritoneal cell interaction and identified several novel proteins that may be involved in ovarian cancer metastasis [21,22]. To effectively determine their function, we developed a CAM assay protocol using a range of ovarian cancer cell lines to allow the monitoring of candidate molecules on ovarian cancer cell invasion in vivo. The in vivo CAM assay data was compared with results from in vitro assays.

2. Results

2.1. Human Ovarian Cancer Cell Motility and Invasion In Vitro

We compared the motility and invasion of three ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) using in vitro assays. We found OV-90 cells to be most invasive through an extracellular matrix and migrated faster through 12 µm pores towards a chemo attractant, compared to SKOV-3 and OVCAR-3 cells (Figure 1). OVCAR-3 cells were the least motile and invasive cell line in our study.

Int. J. Mol. Sci. 2012, 13 9961

Figure 1. Motility and invasion of human ovarian cancer cell lines (OVCAR-3, SKOV-3 and OV-90) in vitro. The fluorescence reading represents the cancer cells that have migrated through the pores or invaded through the extracellular matrix (Geltrex). Data represents the mean ± SEM from two independent experiments performed in quadruplicate. (a) Indicates significant difference from OVCAR-3 cells; and (b) indicates significant difference from SKOV-3 cells, p < 0.05.

2.2. Human Ovarian Cancer Cell Invasion into the Chick Chorioallantoic Membrane (CAM)

We initially utilized an ex ovo method and incubated the chick embryos in plastic weigh boats as described previously [23]. The ex ovo method has the advantage of allowing the application of a larger number of matrigel grafts as a wider area of the CAM is accessible. However, the survival rate for the ex ovo method was very low and only 10% of embryos survived to day 14. The in ovo method had a survival rate of 70% on day 14. In the in ovo method, a small window is made in the shell on day 3 of chick embryo development to detach the CAM layer from the egg shell (Figure 2a). Ovarian cancer cells (9 × 105 cells) were mixed with matrigel to form a gel and grafted on top of the CAM of day 11 chick embryos. The chick embryos were incubated with the matrigel grafts until day 14 of development. An example of a matrigel graft on day 14 is shown in Figure 2b.

Figure 2. (a) Day 3 chick embryo; (b) Ovarian cancer cells and matrigel graft on the chick chorioallantoic membrane (CAM) on day 14.

The CAM layers; ectoderm (ET), mesoderm (M) and endoderm (ED) can be seen in Figure 3a. Cytokeratin immunohistochemistry was used to identify the CAM layer integrity and presence of ovarian cancer cells in the mesodermal layer.

Int. J. Mol. Sci. 2012, 13 9962

Figure 3. Invasion of ovarian cancer cells in the chick chorioallantoic membrane (CAM). (a) Control showing the normal structure of the CAM layers; ectoderm (ET), mesoderm (M) and endoderm (ED); (b) OVCAR-3; (c) SKOV-3; and (d) OV-90 cancer cell matrigel grafts (CM) were placed on top of the ectoderm layer and cancer cell invasion into the CAM mesoderm was assessed in day 14 chick embryos. CAM paraffin sections (6 µm) were immunostained with cytokeratin antibody. Original magnification ×200, scale bar 100 µm.

The invasion of the ovarian cancer cells through the ectoderm into the mesoderm was assessed on day 14 of chick embryo development. OVCAR-3 cytokeratin immunohistochemistry showed some damage to the ectoderm layer but minimal invasion into the CAM mesoderm (Figure 3b). The SKOV-3 cells showed invasion into the mesoderm layer and minimal destruction of the ectoderm layer (Figure 3c). OV-90 cells were the most invasive cells in the CAM assay (Figure 3d), which agrees with the results of the in vitro assays (Figure 1). Figure 3d shows a large number of OV-90 cells invading into the mesoderm of the CAM, as well as the destruction of the CAM ectoderm layer. Quantitative analysis of ovarian cancer cell invasion into the CAM was performed by determining the number of images (8 to 12 sections per chick embryo) with cancer cell invasion into the CAM mesoderm of day 14 chick embryos. Our results showed a significantly higher invasion of OV-90 and SKOV-3 cells into the CAM mesoderm, when compared with OVCAR-3 cells (Figure 4).

2.3. The Effects of Protein a Neutralizing Antibody on OV-90 Cancer Cell Invasion into the CAM

We used a neutralizing antibody against one of the novel proteins identified in our previous study investigating ovarian cancer-peritoneal cell interactions [21,22]. OV-90 cancer cells (9 × 105 cells) were mixed with matrigel and the control anti-mouse IgG or the neutralizing antibody against protein A before grafting onto day 11 chick embryos. Neutralizing antibody against protein A

Int. J. Mol. Sci. 2012, 13 9963 effectively inhibited OV-90 cancer cell invasion into the mesoderm of the CAM, compared with the control anti-mouse IgG, where OV-90 cancer cells invaded the mesoderm of the CAM and a destruction of the ectoderm layer was observed (Figure 5).

Figure 4. Chick chorioallantoic membrane (CAM) invasion by ovarian cancer cells in day 14 chick embryos. Data generated from 48–60 images from 6 chicken embryos per cell line. Data represents the percentage of images with invasion into the mesoderm, mean ± SEM from two independent experiments. (a) Indicates significant difference from OVCAR-3 cells; and (b) indicates significant difference from SKOV-3 cells, p < 0.05.

Figure 5. Effects of protein A neutralizing antibody on OV-90 cancer cell invasion into the chick chorioallantoic membrane (CAM). OV-90 cancer cells were mixed with matrigel and (a) control anti-mouse IgG (20 µg/mL); or (b) neutralizing antibody against protein A (20 µg/mL). CAM paraffin sections (6 µm) were immunostained with a pan-cytokeratin antibody. ET = ectoderm. M = mesoderm. ED = endoderm. CM = cancer cell matrigel grafts. Original magnification ×200, scale bar 100 µm.

3. Discussion

The CAM assay is a frequently applied model to study ovarian cancer angiogenesis [24–27]. However, there is only one study which has used CAM assays to assess ovarian cancer cell invasion and metastasis [20]. Chang et al. described IGROV-1 ovarian cancer cell invasion and metastasis to

Int. J. Mol. Sci. 2012, 13 9964 the posterior CAM and lungs of chick embryos [20]. We have developed a CAM assay protocol to monitor the metastatic properties of ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) and have successfully used it to study the effect of newly identified molecules in vivo. Our results show that the CAM assay is an effective model to study ovarian cancer metastasis. Importantly, our CAM model closely mimics the mode of ovarian cancer metastasis which involves ovarian cancer cell attachment and invasion into the peritoneum. The ectodermal layer of the CAM has many similarities with the peritoneum, which consists of a single layer of mesothelial cells covering the organs in the abdominal cavity. We observed a higher survival rate with the in ovo method in comparison to the ex ovo method for monitoring of ovarian cancer cell growth in the chick embryos. Various methods have been used to graft cancer cells in the CAM model; such as collagen onplants [23], plastic rings [19], and matrigel grafts [28]. Furthermore, cancer cells can also be inoculated by dropping the cell suspension on top of the CAM [29], or administered intravenously to study metastasis of cancer cells in the chick embryos [18]. Matrigel is one of the most suitable scaffolds used for implantation and grafting of cancer cells onto the CAM. In our model, ovarian cancer cells and matrigel were mixed with or without neutralizing antibodies before grafting onto the CAM of the chick embryos to assess ovarian cancer cell invasion. The grafting of the matrigel in the CAM model allows continuous visualization of the test site. Moreover, other studies have reported visible and solid tumors on the CAM of chick embryos a few days after cancer cell inoculation [11,19]. We used histological assessment by means of a pan-cytokeratin antibody, to allow the visualization of cancer cells invading into the mesoderm. The CAM model has been previously employed to assess cancer metastasis [3,29]. In some studies quantitative alu PCR was used to assess the presence of metastatic human cancer cells in chick embryo organs [30]. Several studies have compared both CAM assays and mouse models to assess tumor growth and metastasis. Colorectal cancer cells were reported to colonize the CAM similarly to the mouse model [10]. Strojnik et al. conducted a study to compare the histological and immunohistochemical characteristics of glioma tumor protein expression in the CAM and an established rat model. They reported a similar profile of proteins expressed in both models [12]. In addition, the CAM model has also been used concurrently with the nude mice model to assess tumor growth of fibrosarcoma (HT1080 cells) and human squamous carcinoma (Hep3 cells) cells [31]. Lyu et al. also showed that over expression of urokinase-type plasminogen activator receptor (u-PAR) in Hep3 cells leads to an increase in cancer cell invasion in the CAM model as well as accelerated tumor growth in the SCID mice model [32]. These studies demonstrate the validity of the CAM model for in vivo analysis of cancer cell invasion and metastasis. The CAM model has many advantages. It is cost effective, allows large scale screening and is an easily reproducible in vivo model [4,7]. A comparison of the advantages and limitations of the CAM against the mouse model are summarized in Table 1.

Int. J. Mol. Sci. 2012, 13 9965

Table 1. Comparison of the advantages and limitations of the chick chorioallantoic membrane (CAM) and mouse model.

In vivo Advantages Reference Limitations Reference model Short experimental assay (days) Short observation period (days) Cannot examine cancer-immune Inexpensive cell interactions Easily reproducible and high throughput Rapid morphological changes [4] Closed system—allows assessment of Limited antibodies to chicken tissues [7] small quantities of therapeutic agents for characterization Naturally immunodeficient [7] CAM Multiple tests per individual CAM Allows large scale screening [4] Biology and physiology well known Availability of in vivo imaging [3] Allows direct visualization [8] Animals do not have to be restrained Can be used with primary human cell lines Longer observation period (weeks to Long experimental length (months to

months) years) Biology and physiology well known, [4] Costly but also complex [4] Mouse Availability of in vivo imaging [33] Mature immune system Defined genetic background [34] Reproducibility is expensive Large number of animals required

Animals have to be restrained [8]

The CAM model has also been used in pre-clinical screening to assess the efficacy of drugs and inhibitors on tumor growth. Hagedorn et al. reported that treatment of human glioma cells with receptor tyrosine kinase inhibitors inhibited tumor growth in a CAM model [14]. Bekes et al., demonstrated that treatment of prostate cancer (PC3 cells) with u-PA activation blocking antibody mAb-112 significantly inhibited cancer cell invasion in the CAM model [35]. Additionally, the CAM model has been used to test the efficacy of chemotherapy agents (such as doxorubicin) in human leukemia cell lines and has been shown to reduce cancer cell growth in the CAM [18]. An important limitation of the CAM model is the inability to assess cancer–immune cell interactions. Examination of cancer–immune cell interactions requires the use of transgenic ovarian cancer models [36], however, these models are not widely available, are not suitable for high throughput screening, and cannot be used with primary ovarian cancer cells derived from clinical samples. The CAM assay is therefore an attractive model to rapidly assess the effectiveness of novel candidate therapeutic drugs and the in vivo inhibition of specific tumor types and subtypes in one consistent model. We have shown that OV-90 ovarian cancer cells invade into the mesoderm of the CAM within three days of implantation, therefore making the OV-90 CAM model ideal for studying ovarian cancer invasion and metastasis. In contrast, the OVCAR-3 cells showed limited invasion in the CAM over the

Int. J. Mol. Sci. 2012, 13 9966 three days of our assays, and would therefore be suitable for studying the role of molecules that promote ovarian cancer invasion. In conclusion, the CAM model provides a high throughput in vivo model for the assessment and evaluation of candidate pro-invasive molecules as well as potential therapeutic targets for ovarian cancer.

4. Materials and Methods

4.1. Cell Culture

The human ovarian cancer cell lines OVCAR-3, SKOV-3 and OV-90 were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). All ovarian cancer cell lines were maintained in RPMI 1640 medium supplemented with 4 mM L-glutamine, antibiotics (100 U penicillin G, 100 µg/mL streptomycin sulfate and 0.25 µg/mL amphotericin B), (Sigma-Aldrich, St. Louis, MO, USA). OVCAR-3 and SKOV-3 cells were supplemented with 5% fetal bovine serum (FBS) (Sigma-Aldrich) and OV-90 cells were supplemented with 10% FBS. All cell lines were maintained at

37 °C in an environment of 5% CO2.

4.2. Cell Motility and Invasion Assays

Cell motility and invasion assays were performed as previously described [37]. Briefly, the ovarian cancer cells (OVCAR-3, SKOV-3 and OV-90) were diluted to a cell concentration of 1 × 106 cells/mL and labeled with calcein-AM (1 µg/mL, Life Technologies, VIC, Australia). Ovarian cancer cells were mixed at room temperature for 1 hour in the dark on an oscillating platform. Ovarian cancer cells (50 µL, 40,000 cells) were loaded on top of uncoated 12 µm filter inserts (Disposable, 96-well plate, ChemoTx, Neuro Probe Inc, Gaitherburg, MD, USA) for motility assays or 12 µm filters coated with Geltrex (0.6 µL/well, 9 mg/mL, Life Technologies) for invasion assays. 10% FBS RPMI media was used as a chemo attractant. The cells were allowed to migrate and invade to the lower chamber for 6 hours. Non-migratory cells on the top of the filter were removed and the fluorescence was measured at 485–520 nm.

4.3. Chick Chorioallantoic Membrane Assay (CAM Assay)

Fertilized white leghorn chicken eggs were obtained from Hi-Chick, South Australia, Australia. Eggs were incubated in a MultiQuip Incubator (E2) at 37 °C with 60% humidity. Ethics approval was obtained by the University of Adelaide Animal Ethics Committee. A small window was made in the shell on day 3 of chick embryo development under aseptic conditions. The window was resealed with adhesive tape and eggs were returned to the incubator until day 11 of chick embryo development. On day 11, OVCAR-3, SKOV-3 and OV-90 cell suspensions (9 × 105) were mixed with growth factor reduced matrigel (8.9 mg/mL, BD Biosciences, NSW, Australia) in a total volume of 30 µL. Control anti-mouse IgG (20 µg/mL) (Sigma Aldrich) and neutralizing antibody to protein A (20 µg/mL) (BD Biosciences) were mixed together with the OV-90 cancer cells and matrigel. Matrigel grafts were placed on top of the CAM and eggs were resealed and returned to the incubator for 72 hours until day 14 (n = 6 chicken embryos per cell line). Matrigel grafts with surrounding CAM were harvested from each embryo and fixed with 4% paraformaldehyde for 24 hours and embedded in paraffin. Serial

Int. J. Mol. Sci. 2012, 13 9967 sections (6 µm) were stained with hematoxylin and eosin. Slides were digitally scanned using the NanoZoomer (Hamamatsu Photonics K.K., Japan).

4.4. Immunohistochemistry

CAM paraffin sections (6 µm) were incubated on a heat plate at 60 °C for 2 hours. Tissue sections were dewaxed with xylene and ethanol, followed by PBS washes. Antigen retrieval was performed by using 1% protease (Sigma-Aldrich) for 10 minutes on a heat plate at 37 °C. The endogenous peroxidase activity of the sections was quenched with 0.3% H2O2. Each tissue sample was blocked with 5% goat serum for 30 minutes before incubation with monoclonal mouse anti-human cytokeratin clone AE1/AE3 (1:50, Dako, VIC, Australia) at 4 °C overnight. Subsequently, the tissues sections were incubated sequentially with biotinylated goat anti-mouse (1:400, Dako), followed by streptavidin-HRP conjugated (1:500, Dako) for 1 hour at room temperature. Immunoreactivity was detected using diaminobenzidine/H2O2 substrate (Sigma-Aldrich). The sections were counterstained with 10% haematoxylin (Sigma-Aldrich), dehydrated and mounted in Pertex (Medite Medizintechnik, Germany). Slides were digitally scanned using the NanoZoomer (Hamamatsu Photonics K.K.). For quantitative analysis of ovarian cancer cell invasion into the mesoderm layer, 8 to 12 CAM images from each embryo were assessed by two independent researchers [16].

4.5. Statistical Analysis

All analyses were performed using SPSS 15.0 for Windows software (SPSS, Chicago, IL, USA). The student’s t-test and one-way ANOVA were used to determine statistical significance of ovarian cancer cell motility and invasion in vitro, and ovarian cancer cell invasion in the CAM model. Statistical significance was accepted at p < 0.05.

5. Conclusions

The CAM assay is a robust technique that can be used to monitor invasion of ovarian cancer cell lines and to assess the role of novel molecules and potential therapeutic targets. It is a valuable alternative to murine in vivo models for the study of ovarian cancer invasion and metastasis.

Acknowledgments

The grant sponsors for this project were the Ovarian Cancer Research Foundation (OCRF), Australia, Cancer Council South Australia and South Australian Health and Medical Research Institute. We thank Simon Koblar from the School of Medicine, University of Adelaide, Australia for his help with setting up the CAM assay.

References

1. Lengyel, E. Ovarian cancer development and metastasis. Am. J. Pathol. 2010, 177, 1053–1064. 2. Richardson, M.; Singh, G. Observations on the use of the avian chorioallantoic membrane (CAM) model in investigations into angiogenesis. Curr. Drug Targets-Cardiovasc. Hematol. Disord. 2003, 3, 155–185.

Int. J. Mol. Sci. 2012, 13 9968

3. Deryugina, E.I.; Quigley, J.P. Chick embryo chorioallantoic membrane model systems to study and visualize human tumor cell metastasis. Histochem. Cell Biol. 2008, 130, 1119–1130. 4. Tufan, A.C.; Satiroglu-Tufan, N.L. The chick embryo chorioallantoic membrane as a model system for the study of tumor angiogenesis, invasion and development of anti-angiogenic agents. Curr. Cancer Drug Targets 2005, 5, 249–266. 5. Ossowski, L. In vivo invasion of modified chorioallantoic membrane by tumor cells: The role of cell surface-bound urokinase. J. Cell Biol. 1988, 107, 2437–2445. 6. Zhai, Y.; Kuick, R.; Nan, B.; Ota, I.; Weiss, S.J.; Trimble, C.L.; Fearon, E.R.; Cho, K.R. Gene expression analysis of preinvasive and invasive cervical squamous cell carcinomas identifies HOXC10 as a key mediator of invasion. Cancer Res. 2007, 67, 10163–10172. 7. Cimpean, A.M.; Ribatti, D.; Raica, M. The chick embryo chorioallantoic membrane as a model to study tumor metastasis. Angiogenesis 2008, 11, 311–319. 8. Valdes, T.I.; Kreutzer, D.; Moussy, F. The chick chorioallantoic membrane as a novel in vivo model for the testing of biomaterials. J. Biomed. Mater. Res. 2002, 62, 273–282. 9. Giannopoulou, E.; Katsoris, P.; Hatziapostolou, M.; Kardamakis, D.; Kotsaki, E.; Polytarchou, C.; Parthymou, A.; Papaioannou, S.; Papadimitriou, E. X-Rays modulate extracellular matrix in vivo. Int. J. Cancer 2001, 94, 690–698. 10. Subauste, M.C.; Kupriyanova, T.A.; Conn, E.M.; Ardi, V.C.; Quigley, J.P.; Deryugina, E.I. Evaluation of metastatic and angiogenic potentials of human colon carcinoma cells in chick embryo model systems. Clin. Exp. Metastasis 2009, 26, 1033–1047. 11. Demir, R.; Dimmler, A.; Naschberger, E.; Demir, I.; Papadopoulos, T.; Melling, N.; Sturzl, M.; Hohenberger, W. Malignant progression of invasive tumour cells seen in hypoxia present an accumulation of beta-catenin in the nucleus at the tumour front. Exp. Mol. Pathol. 2009, 87, 109–116. 12. Strojnik, T.; Kavalar, R.; Barone, T.A.; Plunkett, R.J. Experimental model and immunohistochemical comparison of U87 human glioblastoma cell xenografts on the chicken chorioallantoic membrane and in rat brains. Anticancer Res. 2010, 30, 4851–4860. 13. Balciuniene, N.; Tamasauskas, A.; Valanciute, A.; Deltuva, V.; Vaitiekaitis, G.; Gudinaviciene, I.; Weis, J.; von Keyserlingk, D.G. Histology of human glioblastoma transplanted on chicken chorioallantoic membrane. Medicina (Kaunas) 2009, 45, 123–131. 14. Hagedorn, M.; Javerzat, S.; Gilges, D.; Meyre, A.; de Lafarge, B.; Eichmann, A.; Bikfalvi, A. Accessing key steps of human tumor progression in vivo by using an avian embryo model. Proc. Natl. Acad. Sci. USA 2005, 102, 1643–1648. 15. Kobayashi, T.; Koshida, K.; Endo, Y.; Imao, T.; Uchibayashi, T.; Sasaki, T.; Namiki, M. A chick embryo model for metastatic human prostate cancer. Eur. Urol. 1998, 34, 154–160. 16. Wittig-Blaich, S.M.; Kacprzyk, L.A.; Eismann, T.; Bewerunge-Hudler, M.; Kruse, P.; Winkler, E.; Strauss, W.S.; Hibst, R.; Steiner, R.; Schrader, M.; et al. Matrix-dependent regulation of AKT in Hepsin-overexpressing PC3 prostate cancer cells. Neoplasia 2011, 13, 579–589. 17. Conn, E.M.; Botkjaer, K.A.; Kupriyanova, T.A.; Andreasen, P.A.; Deryugina, E.I.; Quigley, J.P. Comparative analysis of metastasis variants derived from human prostate carcinoma cells: Roles in intravasation of VEGF-mediated angiogenesis and uPA-mediated invasion. Am. J. Pathol. 2009, 175, 1638–1652.

Int. J. Mol. Sci. 2012, 13 9969

18. Taizi, M.; Deutsch, V.R.; Leitner, A.; Ohana, A.; Goldstein, R.S. A novel and rapid in vivo system for testing therapeutics on human leukemias. Exp. Hematol. 2006, 34, 1698–1708. 19. Balke, M.; Neumann, A.; Kersting, C.; Agelopoulos, K.; Gebert, C.; Gosheger, G.; Buerger, H.; Hagedorn, M. Morphologic characterization of osteosarcoma growth on the chick chorioallantoic membrane. BMC Res. Notes 2010, 3, doi:10.1186/1756-0500-3-58. 20. Chang, H.L.; Pieretti-Vanmarcke, R.; Nicolaou, F.; Li, X.; Wei, X.; MacLaughlin, D.T.; Donahoe, P.K. Mullerian inhibiting substance inhibits invasion and migration of epithelial cancer cell lines. Gynecol. Oncol. 2011, 120, 128–134. 21. Ween, M.P.; Lokman, N.A.; Hoffmann, P.; Rodgers, R.J.; Ricciardelli, C.; Oehler, M.K. Transforming growth factor-beta-induced protein secreted by peritoneal cells increases the metastatic potential of ovarian cancer cells. Int. J. Cancer 2011, 128, 1570–1584. 22. Lokman, N.A.; Ween, M.P.; Oehler, M.K.; Ricciardelli, C. The role of annexin A2 in tumorigenesis and cancer progression. Cancer Microenviron. 2011, 4, 199–208. 23. Deryugina, E.I.; Quigley, J.P. Chapter 2. Chick embryo chorioallantoic membrane models to quantify angiogenesis induced by inflammatory and tumor cells or purified effector molecules. Methods Enzymol. 2008, 444, 21–41. 24. Manenti, L.; Riccardi, E.; Marchini, S.; Naumova, E.; Floriani, I.; Garofalo, A.; Dossi, R.; Marrazzo, E.; Ribatti, D.; Scanziani, E.; et al. Circulating plasma vascular endothelial growth factor in mice bearing human ovarian carcinoma xenograft correlates with tumor progression and response to therapy. Mol. Cancer Ther. 2005, 4, 715–725. 25. Fang, J.; Zhou, Q.; Liu, L.Z.; Xia, C.; Hu, X.; Shi, X.; Jiang, B.H. Apigenin inhibits tumor angiogenesis through decreasing HIF-1α and VEGF expression. Carcinogenesis 2006, 28, 858–864. 26. Luo, H.; Rankin, G.O.; Liu, L.; Daddysman, M.K.; Jiang, B.H.; Chen, Y.C. Kaempferol inhibits angiogenesis and VEGF expression through both HIF dependent and independent pathways in human ovarian cancer cells. Nutr. Cancer 2009, 61, 554–563. 27. Liu, L.Z.; Jing, Y.; Jiang, L.L.; Jiang, X.E.; Jiang, Y.; Rojanasakul, Y.; Jiang, B.H. Acacetin inhibits VEGF expression, tumor angiogenesis and growth through AKT/HIF-1alpha pathway. Biochem. Biophys. Res. Commun. 2011, 413, 299–305. 28. Kisker, O.; Onizuka, S.; Banyard, J.; Komiyama, T.; Becker, C.M.; Achilles, E.G.; Barnes, C.M.; O’Reilly, M.S.; Folkman, J.; Pirie-Shepherd, S.R. Generation of multiple angiogenesis inhibitors by human pancreatic cancer. Cancer Res. 2001, 61, 7298–7304. 29. Petitclerc, E.; Brooks, P.C. The chick embryo metastasis model. Methods Mol. Med. 2001, 58, 173–187. 30. Zijlstra, A.; Mellor, R.; Panzarella, G.; Aimes, R.T.; Hooper, J.D.; Marchenko, N.D.; Quigley, J.P. A quantitative analysis of rate-limiting steps in the metastatic cascade using human-specific real-time polymerase chain reaction. Cancer Res. 2002, 62, 7083–7092. 31. Aguirre-Ghiso, J.A.; Ossowski, L.; Rosenbaum, S.K. Green fluorescent protein tagging of extracellular signal-regulated kinase and p38 pathways reveals novel dynamics of pathway activation during primary and metastatic growth. Cancer Res. 2004, 64, 7336–7345.

Int. J. Mol. Sci. 2012, 13 9970

32. Lyu, M.A.; Choi, Y.K.; Park, B.N.; Kim, B.J.; Park, I.K.; Hyun, B.H.; Kook, Y.H. Over-expression of urokinase receptor in human epidermoid-carcinoma cell line (HEp3) increases tumorigenicity on chorio-allantoic membrane and in severe-combined-immunodeficient mice. Int. J. Cancer 1998, 77, 257–263. 33. Ray, P.; Lewin, S.A.; Mihalko, L.A.; Schmidt, B.T.; Luker, K.E.; Luker, G.D. Noninvasive imaging reveals inhibition of ovarian cancer by targeting CXCL12-CXCR4. Neoplasia (New York, NY, USA) 2011, 13, 1152–1161. 34. Sale, S.; Orsulic, S. Models of ovarian cancer metastasis: Murine models. Drug Discov. Today Dis. Models 2006, 3, 149–154. 35. Bekes, E.M.; Deryugina, E.I.; Kupriyanova, T.A.; Zajac, E.; Botkjaer, K.A.; Andreasen, P.A.; Quigley, J.P. Activation of pro-uPA is critical for initial escape from the primary tumor and hematogenous dissemination of human carcinoma cells. Neoplasia 2011, 13, 806–821. 36. Mullany, L.K.; Richards, J.A.S. Minireview: Animal models and mechanisms of ovarian cancer development. Endocrinology 2012, 153, 1585–1592. 37. Ween, M.P.; Hummitzsch, K.; Rodgers, R.J.; Oehler, M.K.; Ricciardelli, C. Versican induces a pro-metastatic ovarian cancer cell behavior which can be inhibited by small hyaluronan oligosaccharides. Clin. Exp. Metastasis 2011, 28, 113–125.

© 2012 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).

www.impactjournals.com/oncotarget/ Oncotarget, August, Vol.4, No 8

Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis

Noor A. Lokman1, Alison SF. Elder1, Miranda P. Ween2, Carmen E. Pyragius1, Peter Hoffmann3, Martin K. Oehler1,4,* and Carmela Ricciardelli 1,* 1 Robinson Institute, Research Centre for Reproductive Health, School of Paedriatrics and Reproductive Health, University of Adelaide, Adelaide, SA, Australia 2 Research Centre for Infectious Diseases, School of Molecular Biosciences, University of Adelaide, Adelaide, SA, Australia 3 Adelaide Proteomics Centre, School of Molecular and Biomedical Science, University of Adelaide, Adelaide, SA, Australia 4 Department of Gynaecological Oncology, Royal Adelaide Hospital, Adelaide, SA, Australia * The last two authors share senior authorship. Correspondence to: Carmela Ricciardelli, email: [email protected] Keywords: annexin A2, ovarian cancer, invasion, chick chorioallantoic membrane assay, xenograft model Received: June 25, 2013 Accepted: July 12, 2013 Published: July 14, 2013

This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT: Our recent research identified the protein annexin A2 to be regulated by ovarian cancer-peritoneal cell interactions. This study investigated the role of annexin A2 in ovarian cancer metastasis and its potential utility as a novel therapeutic target, using in vitro and in vivo ovarian cancer models. Annexin A2 expression was examined by qRT-PCR and western blotting in ovarian cancer cell lines and immunohistochemistry in serous ovarian carcinoma tissues. Annexin A2 siRNAs were used to evaluate the effects of annexin A2 suppression on ovarian cancer cell adhesion, motility, and invasion. Furthermore, annexin A2 neutralizing antibodies were used to examine the role of annexin A2 in tumor invasion and metastasis in vivo using a chick chorioallantoic membrane assay and an intraperitoneal xenograft mouse model. Strong annexin A2 immunostaining was observed in 90% (38/42) of the serous ovarian cancer cells and was significantly increased in the cancer-associated stroma compared to non-malignant ovarian tissues. Annexin A2 siRNA significantly inhibited the motility and invasion of serous ovarian cancer cells and adhesion to the peritoneal cells. Annexin A2 neutralizing antibodies significantly inhibited OV-90 cell motility and invasion in vitro and in vivo using the chick chorioallantoic membrane assay. The growth of SKOV-3 cells and their peritoneal dissemination in nude mice was significantly inhibited by annexin A2 neutralizing antibodies. Annexin A2 plays a critical role in ovarian cancer metastasis and is therefore a potential novel therapeutic target against ovarian cancer.

INTRODUCTION cancer survival rates have not changed significantly. An increase of the ovarian cancer survival rate will require Ovarian cancer is the most lethal gynecological the successful development of more effective molecularly cancer and ranks as the fifth most common cause of targeted therapies. cancer-related death in women in the western world. It Ovarian cancer has a distinct predisposition for has been estimated that there will be 22,240 new cases of metastasizing via shedding of cancerous cells from the ovarian cancer and 14,030 deaths due to ovarian cancer ovary into the peritoneal cavity and implanting onto the in the United States in 2013 [1]. Despite improvements peritoneum that lines the pelvic organs. Once ovarian in the surgical treatment and the development of new cancer cells adhere to the peritoneal cells, they migrate chemotherapeutic agents over the last 10 years, ovarian through the peritoneal layer and invade local organs. The www.impactjournals.com/oncotarget 1199 Oncotarget 2013; 4: 1199-1211 local invasion of organs, such as the bowel, eventually cancer cells in the omentum (Fig. 1F). No staining was results in the death of the patient. observed in the absence of the primary antibody (insert, Our group has recently explored the interactions Fig. 1A). Stromal annexin A2 immunostaining in the between ovarian cancer-peritoneal cells using an in vitro invasive serous ovarian carcinomas (stage I to IV) was co-culture system [2]. One of the proteins identified by 2D significantly increased compared with normal ovaries, gel electrophoresis and mass spectrometry to be regulated serous cystadenomas, or serous borderline tumors (P < by ovarian cancer-peritoneal cell interactions was annexin 0.0001). No difference was observed between the intensity A2 [3]. of the annexin A2 immunostaining in the cancer cells (P Annexin A2 is a multifunctional calcium = 0.510) or percentage of annexin A2 positive cancer phospholipid binding protein which binds to collagen I, cells (P = 0.248) between the different patients groups cathepsin B and tenascin-C [4], assists in maintaining the (Supplementary Table 1). No difference in annexin A2 plasticity and rearrangement of the actin cytoskeleton [5] immunostaining was observed between primary tumor and and a cellular redox regulatory protein [6]. Annexin A2 matching omental metastasis tissues (Supplementary Table also plays an important role in the plasminogen activation 2). However, in metastatic omental implants, a higher system and acts as a tissue plasminogen activator (t-PA) proportion of ovarian cancer cells immediately adjacent receptor on the cell surface of endothelial and cancer to the peritoneal cells were annexin A2 positive (8/9) cells, which mediates the conversion of plasminogen into compared to ovarian cancers cells at a greater distance plasmin [7, 8]. from the peritoneum (4/9) (P = 0.046, Pearson Chi Square) Various studies have found increased annexin A2 (Supplementary Table 3). tissue levels in malignancies of the breast, pancreas, oropharynx, liver, kidney, and bowel (reviewed by [3]). Expression of annexin A2 in human ovarian Annexin A2 has been shown to promote cell invasion cancer and peritoneal cell lines in malignancies of the breast, brain, liver, and pancreas [9-12] and enhances cell motility and cell adhesion of prostate and hepatocellular carcinoma cells [12, 13]. Real-time PCR results showed annexin A2 was However, the knowledge on the role of annexin A2 in expressed in human serous ovarian cancer cell lines ovarian cancer is very limited. It was identified to be (OVCAR-3, OVCAR-5, SKOV-3 and OV-90) and the upregulated in ovarian cancer cell lines with high invasive peritoneal cell line, LP-9 (Fig. 2A). This was supported capacity compared to those with low invasive properties by 1D-western immunoblotting in the cell lysates as an [14]. Moreover, a large scale proteomic study identified annexin A2 band at 37 kDa was observed in all cell lines annexin A2 to be upregulated in ovarian cancers when (Fig. 2B). These findings confirm that annexin A2 is compared with normal ovarian tissue and benign lesions produced by both ovarian cancer cells and peritoneal cells. [15]. This study investigated annexin A2 expression in serous ovarian cancer tissues and cell lines and performed Annexin A2 is cleaved in ovarian cancer and functional in vitro and in vivo studies to examine its role peritoneal cells co-culture in ovarian cancer cell adhesion, motility, invasion and metastasis. Annexin A2 was found to be increased in the secretome of co-cultured OVCAR-5 and LP-9 cells RESULTS using 2D gel electrophoresis and mass spectrometry [3]. Annexin A2 expression was further examined by western immunoblotting in the conditioned media (CM) from Expression of annexin A2 in human ovarian OVCAR-3, OVCAR-5, and SKOV-3 cells either cultured cancer tissues and peritoneal cells alone or in co-culture experiments with LP-9 cells after 48 h. A 37 kDa annexin A2 band at low abundance was observed in the CM from OVCAR-3, OVCAR-5, Immunohistochemistry results showed positive and SKOV-3 cells. In the LP-9 CM, two distinct bands immunostaining of annexin A2 in the epithelial cells corresponding to annexin A2 at approximately 37 and 36 of the normal surface epithelium (Fig. 1A), serous kDa were observed. A shift in the 37 kDa band to a 35 cystadenomas (Fig. 1B) and serous borderline ovarian kDa form was observed when LP-9 cells were co-cultured tumors (Fig. 1C). In serous ovarian cancer cells, with OVCAR-3, OVCAR-5 or SKOV3 cells (Fig. 2C). annexin A2 immunostaining was present predominantly A similar finding was observed with OV-90 and LP-9 in the membrane and cytoplasm but high annexin A2 cells (data not shown). Mass spectrometry of annexin immunostaining was also noted in the cancer associated A2 from the CM of the co-cultured OVCAR-5 and LP-9 stroma (Fig. 1D). Strong annexin A2 immunostaining (n=4) did not identify any annexin A2 peptides in the was observed in the peritoneal cells of the omentum N-terminal domain (Supplementary Fig. 1). 2D-western (Fig. 1E) and in the peritoneal cells adjacent to ovarian immunoblotting was performed to confirm the presence www.impactjournals.com/oncotarget 1200 Oncotarget 2013; 4: 1199-1211 of processed or cleaved annexin A2 in the co-cultured treated with MMP inhibitor, GM-6001, a broad spectrum LP-9 and OVCAR-5 CM compared with LP-9 CM protease inhibitor cocktail as well as plasmin inhibitors, cultured alone (Fig. 2D). Multiple annexin A2 spots at ε-ACA and α2-antiplasmin. In the co-culture experiment, 37 kDa and 35 kDa and pI values ranging from 6 to 8 in only α2-antiplasmin was able to partially block annexin the co-cultured OVCAR-5 and LP-9 CM compared with A2 cleavage (Fig. 2F). The other inhibitors used in the the CM of LP-9 cells alone were observed. This result is co-culture experiment had no effect on annexin A2 consistent with the annexin A2 processing observed in the processing. 1D-western immunoblotting. Annexin A2 promotes ovarian cancer motility and Annexin A2 is cleaved by protease plasmin invasion and aids peritoneal adhesion of ovarian cancer cells Plasmin digestion experiment demonstrated that recombinant annexin A2 is cleaved by plasmin and To determine whether annexin A2 promotes ovarian annexin A2 cleavage was partially inhibited following cancer metastasis, we examined the effects of knocking treatment with α2-antiplasmin (Fig. 2E). To evaluate down annexin A2 expression on ovarian cancer cell whether annexin A2 processing was mediated by other adhesion to the peritoneal cells, motility, and invasion. proteases, co-cultured OVCAR-5 and LP-9 cells were

Figure 1: Annexin A2 immunostaining in human ovarian tissues and omental tissues. Annexin A2 immunostaining is present in the epithelial cells of normal surface epithelium (A), serous cystadenomas (B) and serous borderline ovarian tumors (C). Annexin A2 immunostaining is present in membrane and cytoplasm of ovarian cancer cells and cancer associated stroma of stage 3 serous ovarian cancer tissues (D). Annexin A2 immunostaining is observed in peritoneal cells of the omentum (E) and serous ovarian cancer cells located adjacent to the peritoneal cells in omental implants (F). Insert in (A) is the negative control with no primary antibody. Magnification bar = 100 µm for all images. www.impactjournals.com/oncotarget 1201 Oncotarget 2013; 4: 1199-1211 Real-time PCR results showed annexin A2 siRNAs also observed a significant decrease in OV-90 cell motility effectively knocked down annexin A2 expression up (P = 0.0016) and invasion (P = 0.0031) after treatment to 70% compared with negative control siRNA and with annexin A2 neutralizing antibody, compared with the non-treated OVCAR-5 cells (Supplementary Fig. 2A). mouse IgG antibody (Fig. 4A). Treatment of OVCAR-5 Moreover, the knockdown of annexin A2 expression cells with annexin A2 neutralizing antibody also showed a with siRNA A into OVCAR-3, OVCAR-5, SKOV-3, and significant decrease in cell motility and invasion compared OV-90 cells were confirmed by western immunoblotting with the mouse IgG antibody (data not shown). (Supplementary Fig. 2B). Annexin A2 siRNA treated OVCAR-5, OV-90, SKOV-3, and OVCAR-3 cells showed Annexin A2 promotes ovarian cancer cell invasion significantly decreased motility (OVCAR-5 P( = 0.0008), in the chick chorioallantoic membrane model OV-90 (P < 0.0001), SKOV-3 (P < 0.0001) and OVCAR-3 (P = 0.0069) and invasion (OVCAR-5 (P = 0.0047), OV- 90 (P = 0.0047), SKOV-3 (P < 0.0001) and OVCAR-3 (P Ovarian cancer cell invasion was assessed using = 0.0002), compared with cells treated with the negative the CAM (chick chorioallantoic membrane) assay. OV- control siRNA (Fig. 3A-D). Furthermore, annexin A2 90 cells were mixed together with matrigel and placed siRNA treated SKOV-3, OVCAR-5, and OV-90 cells onto the CAM of the chick embryos (Fig. 4B). To had significantly decreased adhesion to LP-9 cells when evaluate the effects of annexin A2 on ovarian cancer cell compared with cells treated with negative control siRNA invasion in vivo, OV-90 cells were treated with the mouse (Fig. 3A-C, OVCAR-5 (P = 0.0005), OV-90 (P < 0.0001), IgG antibody and annexin A2 neutralizing antibody. SKOV-3 (P < 0.0001)). However, there was no difference Haematoxylin and eosin staining and pan-cytokeratin in the adhesion to the peritoneal cells in annexin A2 immunostaining of OV-90 cells treated with the mouse siRNA treated OVCAR-3 cells (Fig. 3D, P = 0.9057). We IgG antibody showed invasion of OV-90 cancer cells through the ectoderm into the mesoderm of the CAM and a

Figure 2: Annexin A2 expression in ovarian cancer cell lines, peritoneal cell line and co-cultured ovarian cancer and peritoneal cells. (A) Annexin A2 expression in OV-90, SKOV-3, OVCAR-5, OVCAR-3 and LP-9 cell lines determined by real-time PCR and were assessed using 2-ΔΔCT quantitation method. Data represents triplicate determinations ± SEM from 2 independent experiments. (B) Western immunoblotting shows annexin A2 band at 37 kDa in cell lysates of OV-90, SKOV-3, OVCAR-5, OVCAR-3 and LP-9. β-actin was used as loading control. (C) Western immunoblotting shows two isoforms of annexin A2 at 37 kDa and 36 kDa bands in the conditioned media (CM) of LP-9 cells alone and a 35 kDa annexin A2 band present in the CM of co-cultured LP-9 and ovarian cancer cells. (D) 2D-western immunoblotting showed multiple annexin A2 spots at 37 kDa in the CM of LP-9 cells alone and annexin A2 isoforms at 37 kDa and 35 kDa in the CM of co-cultured LP-9 and OVCAR-5 cells over a pI values range of 6 to 8. (E) Annexin A2 western immunoblotting of recombinant annexin A2 with GST tag (~63.4 kDa) and a cleaved annexin A2 at ~36 kDa in the presence of plasmin that was partially inhibited by α2-antiplasmin. (F) Annexin A2 western immunoblotting of the CM of co-cultured OVCAR-5 and LP-9 cells treated with α2- antiplasmin, protease inhibitor cocktail, ε-aminocaproic acid (ε-ACA), GM-6001 and DMSO. www.impactjournals.com/oncotarget 1202 Oncotarget 2013; 4: 1199-1211 destruction of the ectoderm layer (Fig. 4C,D). In contrast, Whitney U test). OV-90 cells treated with annexin A2 neutralizing antibody We investigated the possible mechanisms for exhibited minimal invasion through the ectoderm and the reduced tumor burden in the annexin A2 antibody mesoderm of the CAM (Fig. 4E,F). Quantitative analysis treatment group by analysing tumor cell proliferation, showed that annexin A2 antibody significantly inhibited cell apoptosis and angiogenesis. Neutralizing annexin OV-90 cancer cell invasion into the CAM mesoderm. A2 antibody had no effect on SKOV-3 cell proliferation Treatment with annexin A2 antibody resulted in a 3.6 fold (Ki67, Fig. 6A), but significantly increased SKOV-3 cell reduction in cancer cell invasion, compared with OV-90 apoptosis (active caspase 3, Fig. 6B). No difference in cancer cells treated with mouse IgG antibody (Fig. 4G, P vascular density was observed between the two treatment = 0.004, Mann-Whitney U test). groups (CD34, Fig. 6C). Our findings suggest the reduced tumor burden and metastatic spread in the annexin Annexin A2 promotes ovarian cancer growth and A2 antibody treatment group is a result of reduced cell metastasis survival.

DISCUSSION We utilized an intraperitoneal xenograft mouse model to assess the role of annexin A2 in ovarian cancer To our knowledge, this is the first study to investigate metastasis in vivo. The extensive tumor development and the role of annexin A2 in ovarian cancer invasion and peritoneal metastasis observed with mouse IgG treatment metastasis. Here we show i) high annexin A2 expression (Fig. 5A) was significantly reduced in mice treated with in 90% of serous ovarian cells and increased annexin A2 annexin A2 neutralizing antibody (Fig. 5B). Tumor burden levels in the cancer-associated stroma, ii) suppression of measured by bioluminescence was significantly reduced annexin A2 expression significantly reduces migration and by annexin A2 neutralizing antibody treatment compared invasion of 4 ovarian cancer cell lines, iii) down regulation with mouse IgG antibody treatment over a 36 day period of annexin A2 reduces adhesion of ovarian cancer cells (Fig. 5C, Supplementary Fig. 3 and P < 0.05, Mann-

Figure 3: The effects of annexin A2 siRNA knockdown on ovarian cancer cell motility, invasion and adhesion in vitro. (A) OVCAR-5, (B) OV-90, (C) SKOV-3 and (D) OVCAR-3 cell motility, invasion and adhesion to the peritoneal cells after treatment with annexin A2 siRNA. Data are expressed as a percentage of the negative control siRNA, mean ± SEM of quadruplicates from 3 independent experiments (n=12). *, significantly different from control P( < 0.05, Student t- test). www.impactjournals.com/oncotarget 1203 Oncotarget 2013; 4: 1199-1211 to the peritoneal cells, and iv) annexin A2 neutralizing pivotal role in ovarian cancer progression and metastasis. antibodies blocks migration and invasion of ovarian Ovarian cancer cell implantation onto the peritoneal cancer cells, both in vitro and in vivo. Collectively these lining is a vital step in ovarian cancer metastasis. However, results provide strong evidence that annexin A2 plays a the mechanisms involved in this process are poorly

Figure 4: Annexin A2 promotes OV-90 cell motility and invasion in vitro and in vivo. (A) Cell motility and invasion of OV-90 cells treated with mouse IgG antibody or annexin A2 neutralizing antibody in vitro. Data are expressed as a percentage of control mouse IgG, mean ± SEM of quadruplicates from 3 independent experiments (n=12). *, significantly different from control P( < 0.05, Student t-test). (B) Chick chorioallantoic membrane (CAM) implanted with matrigel and OV-90 cancer cells (CMI) at day 14 of chick embryo development. The CAM layers ectoderm (ECT), mesoderm (MES) and endoderm (END) are evident. Haematoxylin and eosin and pan- cytokeratin immunohistochemistry of the CAM implanted with OV-90 cells in the presence of mouse IgG antibody (C,D) and annexin A2 neutralizing antibody (E,F). (G) Quantitative analysis of OV-90 cells invasion into the CAM. Data represents mean ± SEM from 72-80 images (n=12 chick embryos per treatment group from 2 independent experiments). *, significantly different from control (P < 0.05, Mann- Whitney U test). Magnification bar = 100 µm for all images. www.impactjournals.com/oncotarget 1204 Oncotarget 2013; 4: 1199-1211 characterized. In this study, annexin A2 was regulated as A2 expression in both the membrane and cytoplasm of a result of ovarian cancer and peritoneal cell interactions. ovarian cancer cells, but high annexin A2 levels were also Although annexin A2 over expression has been observed observed in the cancer associated stroma. Strong annexin in various cancer types such as those of the pancreas [16], A2 expression in stromal cells was observed for all clinical colorectal [17], breast [11, 18], prostate [13], liver [19] stages of human ovarian cancer (stage I to IV) compared and brain [9, 20], to date, no information on the expression with normal ovaries, serous cystadenomas and borderline of annexin A2 in human ovarian cancer tissues and its ovarian tumors. Increased levels of annexin A2 were also functional role in ovarian cancer have been published. present in ovarian cancer cells located adjacent to the Our immunohistochemical studies showed annexin peritoneal cells in the omental implants tissues.

Figure 5: Annexin A2 promotes growth and metastasis of SKOV-3 cells. Tumor burden of mice treated with mouse IgG antibody (A) or anti-annexin A2 antibody (B) on day 36 using IVIS imaging and at autopsy, with haematoxylin and eosin staining showing the tumor morphology. Images represent 1 sec acquisition time, and the photon emission transmitted from mice was captured and quantitated in photons/s/cm2/sr. (C) Tumor growth and metastasis measured by bioluminescence. Arrow indicates the time point of weekly antibody administration until day 36 (*P < 0.05, Mann-Whitney U test). Data represents mean ±SEM (n=5 mice per treatment group). Magnification bar = 100 µm for all images. www.impactjournals.com/oncotarget 1205 Oncotarget 2013; 4: 1199-1211 Annexin A2 was shown to be expressed in human proteases such as glycogen synthase-3 [24] and MMP- ovarian cancer and peritoneal cell lines. We observed a 7 [25]. We confirmed that plasmin cleaved recombinant shift in the full length 37 kDa annexin A2 band to a 35 annexin A2 and we demonstrated that annexin A2 cleavage kDa isoform in the CM of co-cultured ovarian cancer and in co-cultured ovarian cancer and peritoneal cells could be peritoneal cells. Furthermore, peptide mass fingerprinting partially blocked by α2-antiplasmin. of the annexin A2 protein in the co-cultured CM failed Our findings suggest that the extracellular form of to detect any peptides containing the first 30 amino acids annexin A2 found in the cancer associated stroma in the of the N-terminal domain of annexin A2. These findings ovarian cancer tissues may represent a cleaved secreted suggest that annexin A2 is cleaved at the N-terminal form of annexin A2, which may assist ovarian cancer domain as a result of the ovarian cancer-peritoneal cell progression and metastasis. Since annexin A2 lacks a co-culture environment. signal peptide and is not secreted via the endoplasmic We investigated whether annexin A2 cleavage in reticulum pathway, the mechanism that regulates annexin the ovarian cancer peritoneal cell co-culture could be A2 secretion remains unknown. It has been reported inhibited by specific proteases. Plasmin is one of the previously that cleavage of annexin A2 at lysine 10 by proteases at the cell surface involved in remodeling the MMP-7 can assist tumor invasion and metastasis of tumor microenvironment [21]. We previously showed colorectal and breast cancer cell lines [25]. Moreover, that plasmin was upregulated in the CM of co-cultured annexin A2 isoforms and cleaved annexin A2 were ovarian cancer and peritoneal cells [2]. Previous studies observed in normal and squamous cell carcinoma tissues, reported annexin A2 cleavage by plasmin at lysine 27 in confirmed by 2D-western immunoblotting [26]. However, the N-terminal domain in monocytes [22], and between further studies are required to investigate annexin A2 lysine 307 and arginine 308 in the C-terminal domain cleavage mechanisms, the functional role of cleaved in endothelial cells [23]. Annexin A2 is also cleaved by annexin A2 and post-translational modifications of

Figure 6: Annexin A2 neutralizing antibodies increase apoptosis. Tumor sections of mice treated with mouse IgG and anti- annexin A2 antibodies were immunostained with (A) Ki67, (B) cleaved caspase 3 and (C) CD34. Data represents the median from 10 images per mice (n=6 for mouse IgG and n=5 for anti-annexin A2 treatment groups). *, significantly different from control P( < 0.05, Mann-Whitney U test). Magnification bar = 100 µm for all images. www.impactjournals.com/oncotarget 1206 Oncotarget 2013; 4: 1199-1211 annexin A2 in ovarian cancer. In conclusion, our findings demonstrate that annexin In this study, we have demonstrated that suppression A2 plays an important role in ovarian cancer metastasis. of annexin A2 using siRNA decreased ovarian cancer cell Anti-annexin A2 antibodies significantly blocked ovarian adhesion to the peritoneal cells, cell motility, and invasion cancer cell invasion in the CAM model and cancer cell in vitro. These observations concur with previous studies peritoneal dissemination in the intraperitoneal xenograft demonstrating decreased cell motility and invasion in mouse model. Annexin A2 is therefore a promising novel cancers of the pancreas [10, 27], breast [28], brain [9], therapeutic target against ovarian cancer. liver [12] and prostate [29] following treatment with annexin A2 siRNA. Moreover, our study demonstrated MATERIALS AND METHODS annexin A2 plays a role in ovarian cancer cell adhesion to the peritoneal cells. Shiozawa et al. reported that annexin A2 has a role in regulating prostate cancer cell adhesion to Patient’s tissue samples osteoblasts and endothelial cells using annexin A2 siRNA [13]. Silencing of annexin A2 expression significantly reduced cell adhesion of hepatocellular carcinoma cells Archived formalin fixed paraffin tissue blocks from [12] and cell adhesion of myeloid cells to human and 20 matching primary tumors and their metastatic implants, murine osteoblasts cells [30]. Braden et al. also reported 10 cases each of early stage serous ovarian cancer (stage that down-regulation of annexin A2 expression using I and II), advanced stage serous ovarian cancer (stage polymeric nanoparticles in prostate cancer cell line inhibits III and IV) and borderline ovarian tumors, 16 normal tumor growth in nude mice [29]. Therefore, annexin A2 ovaries, 11 serous cystadenomas and 9 omental implants contributes to cancer progression by enhancing cancer cell were obtained from the Institute of Medical Veterinary motility, invasion and adhesion. Science (SA Pathology), Adelaide, South Australia, The functional role of annexin A2 in ovarian cancer Australia. Tissue microarrays (TMAs) were constructed cell motility and invasion was also assessed using annexin from formalin-fixed, paraffin embedded tumor material A2 neutralizing antibody. We showed OV-90 cells treated with the approval from the Royal Adelaide Hospital ethics with neutralizing annexin A2 antibody significantly committee and each tissue block was represented by decreased motility and invasion in vitro and invasion triplicate 1.0 mm diameter tissue cores. in vivo in the chick embryo CAM model. Annexin A2 neutralizing antibodies also significantly inhibited Cell culture growth and metastasis of SKOV-3 cells in nude mice. Our finding are consistent with previous reports that have The human serous ovarian cancer cell lines shown annexin A2 neutralizing antibodies to inhibit cell OVCAR-3, SKOV-3 and OV-90 were purchased from migration, invasion, and to block plasminogen activation American Type Culture Collection (ATCC, VA, USA). of breast cancer cells [11] and monocytes [31] in vitro. OVCAR-5 cells were obtained from Dr Thomas Hamilton Moreover, anti-annexin A2 antibodies have been shown (Fox Chase Cancer Center, PA, USA) and the peritoneal to inhibit pancreatic cancer metastasis in a mouse model cells, LP-9 were purchased from Coriell Cell Repositories of pancreatic ductal adenocarcinoma [10] and to inhibit (NJ, USA). All cell lines were maintained as previously tumor growth in breast cancer and Lewis Lung Carcinoma described [2]. The SKOV-3/GFP-Luc cells (Cell Biolabs xenograft mouse models [32, 33]. A significant reduction Inc., CA, USA) were maintained in RPMI 1640 medium of glioma tumor growth and progression in the annexin A2 supplemented with 4mM L-glutamine, antibiotics (100 knockout mice model associated with a decrease in cancer U penicillin G, 100 µg/ml streptomycin sulfate and 0.25 cell invasion, angiogenesis and proliferation has also been µg/ml amphotericin B) and supplemented with 5% fetal reported [9]. Tumor growth of fibrosarcoma (HT1080) bovine serum (FBS) (Sigma-Aldrich, MO, USA). All cell and lung cancer (A459) cell lines in NOD-SCID mice lines were maintained at 37oC in an environment of 5% has also been shown to be inhibited following annexin A2 CO . depletion [6]. Furthermore, Zhai et al. showed a decrease 2 in glioma tumor burden as a result of an increase in Real-time PCR apoptosis in tumors of the annexin A2 knockout mice [9]. Similarly, we observed an increase in cell apoptosis, but no difference in the cell proliferation or vascular density Total RNA was extracted using TRIzol (Invitrogen, in the tumor sections of mice with SKOV-3/GFP-Luc cells NSW, Australia) and each RNA sample was reverse treated with anti-annexin A2 antibody. The in vivo findings transcribed using SuperScriptTM III Reverse Transcriptase from our CAM model and intraperitoneal xenograft mouse (Invitrogen), as per manufacturer’s instructions. Real- model were consistent with our in vitro observations time PCR was performed in triplicates using human demonstrating annexin A2 promotes ovarian cancer cell annexin A2 validated primers (QIAGEN, Australia) and motility, invasion and tumor growth. SYBR Green PCR master mix (7900HT Fast Real-Time www.impactjournals.com/oncotarget 1207 Oncotarget 2013; 4: 1199-1211 PCR System, Applied Biosystems, NSW, Australia). The at 37oC in the presence and absence of α2-antiplasmin (0.2 cycling parameters were: 50oC for 2 min, 95oC for 10 µM, Calbiochem). min followed by 40 cycles of 95oC for 15 sec and then 60oC for 1 min. The CT values were normalised relative 2D-western immunoblotting to the housekeeping gene, L19 and were calibrated to the OVCAR-3 cells using the 2-(ΔΔCT) quantitation method. The total protein concentration of CM from LP-9 alone and co-cultured OVCAR-5 and LP-9 was determined Immunohistochemistry using EZQ protein assay (Invitrogen). The immobilized pH gradient (IPG) strips (11cm, GE Healthcare, NJ, For the annexin A2 immunohistochemistry, tissue USA) were rehydrated in rehydration buffer containing sections (5 µm) underwent microwave antigen retrieval 1.2% DeStreak, 0.5% C.A, 1% bromophenol blue and (5 min 750 W, 15 min 350 W) in 10 mM citric acid 4% thiourea-urea-CHAPS (TUC) buffer overnight in buffer (pH 6.5), and were incubated overnight with IPGPhorII (GE Healthcare). A total of 50 µg protein mouse monoclonal antibody to annexin A2 (1/500, BD sample was reduced with 1 M DTT, 0.8% IPG buffer, TUC Biosciences, CA, USA) in blocking buffer (5% normal buffer and bromophenol blue and applied to the IPG strip goat serum) at 4°C. Visualization of immunoreactivity was and isoelectric focusing was carried out using IPGPhorII at achieved using biotinylated anti-mouse immunoglobulins 20oC as previously described [2]. Western immunoblotting (1/400, Dako, NSW, Australia), streptavidin-peroxidise was performed as per 1D-western immunoblotting. conjugate (1/500, Dako) and diaminobenzidine tetrahydrochloride (DAB) (Sigma-Aldrich), as previously Motility and invasion assays described [2]. Normal prostate tissue was used as a positive control [34] and negative controls included no primary antibody and mouse IgG controls. For the Cell motility and invasion assays were performed as xenograft mouse model experiment, mouse tumor tissue previously described [36]. For the annexin A2 knockdown sections were immunostained with active caspase-3 (rabbit experiments, ovarian cancer cells were transfected with polyclonal, 1/200, Cell Signaling Technology, MA, USA), annexin A2 siRNAs (siRNA ID: s1384 (A) and s1385 Ki67 (rabbit monoclonal, 1/400, Epitomics, CA, USA) and (B), 10 nM ) or negative siRNA (Ambion, TX, USA) with CD34 (rat monoclonal, 1/100, clone MEC 14.7, Abcam, Oligofectamine (Invitrogen) for 48 h. Ovarian cancer MA, USA), as previously described [35]. cells were also pre-treated with annexin A2 neutralizing antibody (20 µg/ml, BD Biosciences) or mouse IgG antibody (20 µg/ml, Sigma-Aldrich) for 2 h. Cells were Co-culture of ovarian cancer and peritoneal cells labelled with calcein-AM (1 µg/ml, Invitrogen) and were loaded onto uncoated 12 µm filter inserts (96-well LP-9 cells were cultured in 6 well plates until they plate, ChemoTx, Neuro Probe, MD, USA) for migration reached confluency. OVCAR-5, SKOV-3, OV-90 and assays or 12 µm filters coated with Geltrex (0.6 µl/well, OVCAR-3 cells were added (2 x 105 cells/well) to the Invitrogen) for invasion assays. The cells were allowed LP-9 monolayer and co-cultured in direct contact for to migrate and invade to the lower chamber for 6 h and 48 h before the collection of the cell lysates and CM, as the bottom fluorescence of migratory cells was measured previously described [2]. OVCAR-5 and LP-9 cells were at 485-520 nm using the Triad series multimode detector also co-cultured in the presence or absence of protease (Dynex Technologies, VA, USA). inhibitor cocktail (1/200, Sigma-Aldrich), MMP inhibitor, GM-6001 (20 µM, Calbiochem, CA, USA) as well as Adhesion assay plasmin inhibitors, ε-ACA (150 mM, Sigma-Aldrich) and α2-antiplasmin (0.2 µM, Calbiochem). Adhesion assay were performed as previously described [2]. Briefly, ovarian cancer cells were transfected 1D-western immunoblotting with either annexin A2 siRNA A or negative siRNA for 48 h and labelled with calcein-AM. Ovarian cancer cells Western immunoblotting was conducted as were added to the LP-9 cell monolayer and were allowed previously described [2] with a mouse monoclonal to adhere for 8 min. The fluorescence of adhered cells was antibody to annexin A2 (1/2000, BD Biosciences) or measured at 485-520 nm using the Triad series multimode β-actin antibody (1/2000, Abcam) and anti-mouse IgG detector (Dynex Technologies). peroxidase-conjugated secondary antibodies (1/2000, Dako). For the plasmin digestion, recombinant annexin A2 protein with a GST-tag (0.5 µg, Abnova, Taiwan) was incubated with plasmin (0.2 U/ml, Sigma-Aldrich) for 3 h www.impactjournals.com/oncotarget 1208 Oncotarget 2013; 4: 1199-1211 Chick chorioallantoic membrane assay (2+), weak (1+), or negative (0). A score of 0 or 1+ was defined as low annexin A2 immunostaining and a score or 2+ or 3+ was defined as high annexin A2 immunostaining. Fertilized white leghorn chicken eggs (Hi Chick, The percentage of annexin A2 positive cancer cells were SA, Australia) were maintained at 37°C in 60% relative independently assessed in the ovarian cancer tissues. humidity in a Multiquip Incubator E2 (Multiquip Pty. For the CAM assay, quantitative analysis to assess OV- Ltd., NSW Australia). Approval was obtained by the 90 cancer cell invasion was performed on 8 to 12 CAM University of Adelaide Animal Ethics Committee. On images for each embryo as previously described [37]. For day 3 of chick embryo development, a small opening the mouse tumor tissue sections, ten random images of was made under aseptic conditions in the egg shell. To each tissue were captured at 40x magnification for Ki67 investigate the effects of annexin A2 on ovarian cancer and activated caspase-3 and at 20x magnification for cell invasion in the CAM model, OV-90 cells (9x104 cells) CD34. Colour threshold detection by the AnalySIS-ProTM were mixed with matrigel (8.9 mg/mL, BD Biosciences) software (Soft Imaging System, Germany) was used to with either annexin A2 neutralizing antibody (20 µg/ determine positive (brown pixels) and negative (purple ml, BD Biosciences) or mouse IgG antibody (20 µg/ml, pixels) stained cells. Data was expressed as percentage of Sigma-Aldrich) in a total volume of 30 µl and placed on positive pixels (positive brown stained divided by negative the CAM of day 11 chick embryos (n=6 chick embryos purple stained area) for Ki67 and percentage of positive per treatment group). The invasion of the cancer cells CD34 pixels of the total tumor area. For active caspase through the ectoderm into the mesoderm was assessed 3, the number of positive cells in ten fields were counted on day 14 of chick embryo development in paraffin- manually and expressed as a percentage of the total tumor embedded CAM sections stained with haematoxylin, eosin area. The mouse treatment groups were blinded until and pan-cytokeratin immunohistochemistry, as previously completion of all analyses. described [37]. Statistical analysis Intraperitoneal SKOV-3 xenograft mouse model

All statistical analyses were performed using SPSS An intraperitoneal SKOV-3 xenograft non-invasive 19.0 for Windows (SPSS Inc., IL, USA). The Chi-Square and whole-body bioluminescent imaging model was test was used to determine statistical significance between used with the approval obtained by the University of annexin A2 immunostaining in the ovarian and omental Adelaide Animal Ethics Committee. Nude mice were tissue groups. The Student’s t-test and one-way ANOVA injected intraperitoneally with SKOV-3/GFP-Luc cells with Dunnett C or Dunnett T post-hoc tests were used (1x108 cells/0.5ml) and were treated weekly with either to determine statistical significance between control and 100 µg mouse IgG antibody or 100 µg anti-annexin treatment groups. For the CAM assay and xenograft A2 antibodies (BD Biosciences). Mice were injected mouse model experiments, the Mann-Whitney U test was intraperitoneally with D-luciferin (Caliper Life Sciences, used to determine the significance between control and MA, USA) solution at 150 mg luciferin/kg body weight treatment groups. Statistical significance was accepted at and then gas-anesthetized with isoflurane (Bomac Pty. P < 0.05. Ltd., NSW, Australia). Precisely 10,15 and 20 min following D-luciferin injection, images were acquired for 0.5 to 10 sec and the photon emission transmitted from ACKNOWLEDGEMENTS mice was captured and quantitated in photons/s/cm2/sr using the IVIS Imaging System 100 (Xenogen Imaging We thank Dr. Thomas Hamilton (Fox Chase Cancer Technology, CA, USA) with living image software (Igor Center, Philadelphia, PA, USA) for kindly providing the Pro version 2.5). At autopsy, tumor tissues were fixed with OVCAR-5 cell line. The authors thank Helen Hughes for 4% paraformaldehyde (Sigma-Aldrich) and embedded into collection of archival tissue blocks, Dr. Shalini Jindal for paraffin blocks. pathological assessment and Dr. Aleksandra Ochnick for the TMA construction. We thank Mrs Wendy Bonner for Immunohistochemical assessment her help with tissue sectioning, Mr Adrian Kaczmarek for his technical assistance with the xenograft mouse model and Ms Izza Tan for her help with the image analysis. We Slides were digitally scanned using the NanoZoomer also thank Professor Simon Koblar for his assistance in (Hamamatsu Photonics, Japan) and images were setting up the CAM assay. This research was supported captured using NDP view imaging software (Hamamatsu by the Ovarian Cancer Research Foundation (OCRF) Photonics). The immunostaining intensity of annexin A2 Australia, Cancer Council SA and the South Australian in the epithelial and stromal compartments was assessed Health and Medical Research Institute (SAHMRI). using a manual scoring method: strong (3+), moderate www.impactjournals.com/oncotarget 1209 Oncotarget 2013; 4: 1199-1211 REFERENCES EA, Wang J, Lu G, Roodman GD, Loberg RD, Pienta KJ and Taichman RS. Annexin II/annexin II receptor axis 1. Siegel R, Naishadham D and Jemal A. Cancer statistics, regulates adhesion, migration, homing, and growth of 2013. CA: A Cancer Journal for Clinicians. 2013; 63(1):11- prostate cancer. J Cell Biochem. 2008; 105(2):370-380. 30. 14. Sodek KL, Evangelou AI, Ignatchenko A, Agochiya 2. Ween MP, Lokman NA, Hoffmann P, Rodgers RJ, M, Brown TJ, Ringuette MJ, Jurisica I and Kislinger T. Ricciardelli C and Oehler MK. Transforming growth factor- Identification of pathways associated with invasive behavior beta-induced protein secreted by peritoneal cells increases by ovarian cancer cells using multidimensional protein the metastatic potential of ovarian cancer cells. Int J Cancer. identification technology (MudPIT). Mol Biosyst. 2008; 2011; 128(7):1570-1584. 4(7):762-773. 3. Lokman NA, Ween MP, Oehler MK and Ricciardelli C. The 15. Bengtsson S, Krogh M, Szigyarto CA-K, Uhlen M, role of annexin A2 in tumorigenesis and cancer progression. Schedvins K, Silfverswärd C, Linder S, Auer G, Alaiya A Cancer Microenviron. 2011; 4(2):199-208. and James P. Large-Scale Proteomics Analysis of Human 4. Mai J, Waisman DM and Sloane BF. Cell surface complex Ovarian Cancer for Biomarkers. Journal of Proteome of cathepsin B/annexin II tetramer in malignant progression. Research. 2007; 6(4):1440-1450. Biochim Biophys Acta 2000; 1477(1-2):215-230. 16. Diaz VM, Hurtado M, Thomson TM, Reventos J and 5. Oliferenko S, Paiha K, Harder T, Gerke V, Schwarzler C, Paciucci R. Specific interaction of tissue-type plasminogen Schwarz H, Beug H, Gunthert U and Huber LA. Analysis of activator (t-PA) with annexin II on the membrane of CD44-containing lipid rafts: Recruitment of annexin II and pancreatic cancer cells activates plasminogen and promotes stabilization by the actin cytoskeleton. J Cell Biol. 1999; invasion in vitro. Gut. 2004; 53(7):993-1000. 146(4):843-854. 17. Duncan R, Carpenter B, Main LC, Telfer C and Murray 6. Madureira PA, Hill R, Miller VA, Giacomantonio C, Lee GI. Characterisation and protein expression profiling of PW and Waisman DM. Annexin A2 is a novel cellular annexins in colorectal cancer. British Journal of Cancer. redox regulatory protein involved in tumorigenesis. 2008; 98(2):426-433. Oncotarget. 2011; 2(12):1075-1093. 18. Shetty PK, Thamake SI, Biswas S, Johansson SL and 7. Cesarman GM, Guevara CA and Hajjar KA. An endothelial Vishwanatha JK. Reciprocal Regulation of Annexin A2 cell receptor for plasminogen/tissue plasminogen activator and EGFR with Her-2 in Her-2 Negative and Herceptin- (t-PA). II. Annexin II-mediated enhancement of t-PA- Resistant Breast Cancer. PLoS One. 2012; 7(9):e44299. dependent plasminogen activation. J Biol Chem. 1994; 19. Sun Y, Gao G, Cai J, Wang Y, Qu X, He L, Liu F, Zhang 269(33):21198-21203. Y, Lin K and Ma S. Annexin A2 is a discriminative 8. Sharma M, Ownbey RT and Sharma MC. Breast cancer serological candidate in early hepatocellular carcinoma. cell surface annexin II induces cell migration and Carcinogenesis. 2013; 34(3):595-604. neoangiogenesis via tPA dependent plasmin generation. 20. Reeves SA, Chavez-Kappel C, Davis R, Rosenblum M Exp Mol Path. 2010; 88(2):278-286. and Israel MA. Developmental Regulation of Annexin II 9. Zhai H, Acharya S, Gravanis I, Mehmood S, Seidman (Lipocortin 2) in Human Brain and Expression in High RJ, Shroyer KR, Hajjar KA and Tsirka SE. Annexin A2 Grade Glioma. Cancer Res. 1992; 52(24):6871-6876. promotes glioma cell invasion and tumor progression. The 21. Deryugina EI and Quigley JP. Cell Surface Remodeling by Journal of Neuroscience. 2011; 31(40):14346-14360. Plasmin: A New Function for an Old Enzyme. Journal of 10. Zheng L, Foley K, Huang L, Leubner A, Mo G, Olino Biomedicine and Biotechnology. 2012; 2012. K, Edil BH, Mizuma M, Sharma R and Le DT. Tyrosine 22. Laumonnier Y, Syrovets T, Burysek L and Simmet T. 23 phosphorylation-dependent cell-surface localization Identification of the annexin A2 heterotetramer as a receptor of annexin A2 is required for invasion and metastases of for the plasmin-induced signaling in human peripheral pancreatic cancer. PLoS One. 2011; 6(4):e19390. monocytes. Blood. 2006; 107(8):3342-3349. 11. Sharma MR, Koltowski L, Ownbey RT, Tuszynski GP and 23. Hajjar KA, Jacovina AT and Chacko J. An endothelial cell Sharma MC. Angiogenesis-associated protein annexin II in receptor for plasminogen/tissue plasminogen activator. I. breast cancer: Selective expression in invasive breast cancer Identity with annexin II. J Biol Chem. 1994; 269(33):21191- and contribution to tumor invasion and progression. Exp 21197. Mol Path. 2006; 81(2):146-156. 24. Wang CY, Lin YS, Su WC, Chen CL and Lin CF. Glycogen 12. Zhao P, Zhang W, Tang J, Ma XK, Dai JY, Li Y, Jiang JL, Synthase Kinase-3 and Omi/HtrA2 Induce Annexin A2 Zhang SH and Chen ZN. Annexin II promotes invasion and Cleavage followed by Cell Cycle Inhibition and Apoptosis. migration of human hepatocellular carcinoma cells in vitro Mol Biol Cell. 2009; 20(19):4153-4161. via its interaction with HAb18G/CD147. Cancer science. 25. Tsunezumi J, Yamamoto K, Higashi S and Miyazaki K. 2010; 101(2):387-395. Matrilysin (matrix metalloprotease-7) cleaves membrane- 13. Shiozawa Y, Havens AM, Jung Y, Ziegler AM, Pedersen bound annexin II and enhances binding of tissue-type www.impactjournals.com/oncotarget 1210 Oncotarget 2013; 4: 1199-1211 plasminogen activator to cancer cell surfaces. FEBS J. 28(2):113-125. 2008; 275(19):4810-4823. 37. Lokman NA, Elder AS, Ricciardelli C and Oehler MK. 26. Chen J, He QY, Yuen AP and Chiu JF. Proteomics of Chick Chorioallantoic Membrane (CAM) Assay as an buccal squamous cell carcinoma: The involvement of In Vivo Model to Study the Effect of Newly Identified multiple pathways in tumorigenesis. Proteomics. 2004; Molecules on Ovarian Cancer Invasion and Metastasis. Int 4(8):2465-2475. J Mol Sci. 2012; 13(8):9959-9970. 27. Nedjadi T, Kitteringham N, Campbell F, Jenkins RE, Park BK, Navarro P, Ashcroft F, Tepikin A, Neoptolemos JP and Costello E. S100A6 binds to annexin 2 in pancreatic cancer cells and promotes pancreatic cancer cell motility. Br J Cancer. 2009; 101(7):1145-1154. 28. Zhang J, Guo B, Zhang Y, Cao J and Chen T. Silencing of the annexin II gene down-regulates the levels of S100A10, c-Myc, and plasmin and inhibits breast cancer cell proliferation and invasion. Saudi Med J. 2010; 31(4):374- 381. 29. Braden AR, Kafka MT, Cunningham L, Jones H and Vishwanatha JK. Polymeric nanoparticles for sustained down-regulation of annexin A2 inhibit prostate tumor growth. J Nanosci Nanotechnol. 2009; 9(5):2856-2865. 30. Jung Y, Wang J, Song J, Shiozawa Y, Havens A, Wang Z, Sun YX, Emerson SG, Krebsbach PH and Taichman RS. Annexin II expressed by osteoblasts and endothelial cells regulates stem cell adhesion, homing, and engraftment following transplantation. Blood. 2007; 110(1):82-90. 31. Brownstein C, Falcone D, Jacovina A and Hajjar K. Annexin II ; A Mediator of Cell Surface-Specific Plasmin Generation. Annals of the New York Academy of Sciences. 2001; 947:143-156. 32. Sharma M, Blackman MR and Sharma MC. Antibody- directed neutralization of annexin II (ANX II) inhibits neoangiogenesis and human breast tumor growth in a xenograft model. Experimental and Molecular Pathology. 2012; 92(1):175-184. 33. Sharma MR, Rothman V, Tuszynski GP and Sharma MC. Antibody-directed targeting of angiostatin’s receptor annexin II inhibits Lewis Lung Carcinoma tumor growth via blocking of plasminogen activation: Possible biochemical mechanism of angiostatin’s action. Exp Mol Path. 2006; 81(2):136-145. 34. Inokuchi J, Narula N, Yee DS, Skarecky DW, Lau A, Ornstein DK and Tyson DR. Annexin A2 positively contributes to the malignant phenotype and secretion of IL-6 in DU145 prostate cancer cells. Int J Cancer. 2009; 124(1):68-74. 35. Ricciardelli C, Frewin KM, Tan Ide A, Williams ED, Opeskin K, Pritchard MA, Ingman WV and Russell DL. The ADAMTS1 protease gene is required for mammary tumor growth and metastasis. Am J Pathol. 2011; 179(6):3075-3085. 36. Ween MP, Hummitzsch K, Rodgers RJ, Oehler MK and Ricciardelli C. Versican induces a pro-metastatic ovarian cancer cell behavior which can be inhibited by small hyaluronan oligosaccharides. Clin Exp Metastasis. 2011; www.impactjournals.com/oncotarget 1211 Oncotarget 2013; 4: 1199-1211 Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis – Lokman et al

Supplementary Figure 1: Ion-trap mass spectrometry LC-MS/MS analysis of annexin A2 spots from the 2D gel electrophoresis. No annexin A2 peptides (red peptides: positive identified sequence via Mascot search) were observed in the N-terminal domain of annexin A2 (amino acid 1-30) in the protein spots of the 2D gel electrophoresis-silver stained gel from OVCAR-5 and LP-9 co-cultured cells conditioned media [3].

Supplementary Figure 2: Annexin A2 expression after knockdown with annexin A2 siRNAs. (A) Annexin A2 real-time PCR expression of OVCAR-5 cells after treatment with annexin A2 siRNA A and siRNA B compared with the negative control siRNA and non-treated cells, assessed using 2-ΔΔCT quantitation method. Data represents triplicate determinations ± SEM from 2 independent experiments. (B) Annexin A2 expression of OVCAR-5. SKOV-3, OVCAR-3 and OV-90 cells after knockdown with annexin A2 siRNA A and negative control siRNA confirmed by 1D-western immunoblotting. β-actin was used as a loading control.

Supplementary Figure 3: IVIS imaging of mice. Mice were treated once weekly for three weeks with either mouse IgG (top row), or anti-annexin A2 antibody (bottom row). Images represent tumor growth on days 5, 12, 19, 26, 30 and 36, assessed using the IVIS Imaging. Images represent 1 sec acquisition time, and the photon emission transmitted from mice was captured and quantitated in photons/s/cm2/sr.

Annexin A2 is regulated by ovarian cancer-peritoneal cell interactions and promotes metastasis – Lokman et al

Supplementary Table 1: Annexin A2 epithelial and stromal immunostaining intensity in the normal ovaries, serous cystadenomas, serous borderline tumors and serous ovarian carcinomas (stage I-IV) tissues. Patient Annexin A2 stromal Annexin A2 epithelial staining age staining (years) Tissue n Intensity % of positive cells Intensity Median a b a b (range) Low High ≤10 ≥10 Low High

Normal 52 2/16 14/16 1/16 15/16 14/16 2/16 16 ovaries (40-83) (12.5%) (87.5%) (6.3%) (93.8%) (87.5%) (12.5%) Serous cysta 50 4/11 7/11 2/11 9/11 10/11 1/11 11 denomas (39-66) (36.3%) (63.6%) (18.1%) (81.8%) (90.9%) (9%) Serous 50 0/10 10/10 0/10 10/10 8/10 2/10 borderline 10 (32-81) (0%) (100%) (0%) (100%) (80%) (20%) tumors Serous 64 4/42 38/42 9/42 33/42 14/42 28/42 ovarian 42 (27-86) (10%) (90%) (21.5%) (78.5%) (32.8%) (67.3%) carcinomas (Stage I-IV) Chi-squared P = 0.510 P = 0.248 P < 0.0001 test

a Annexin A2 low intensity score; no staining (0) and weak staining (1+) b Annexin A2 high intensity score; moderate staining (2+) and strong staining (3+)

Supplementary Table 2: Annexin A2 epithelial and stromal immunostaining intensity in the primary tumor and matching omental metastasis tissues. Patient Annexin A2 Annexin A2 epithelial staining age stromal staining (years) % of positive Intensity Intensity Tissue n Median cells (range) Lowa Highb ≤10 ≥10 Lowa Highb

Primary 0/21 21/21 7/21 14/21 9/21 12/21 tumor 21 (0%) (100%) (33.3%) (66.7%) (42.9%) (57.2%)

68 Matching (49-87) 3/18 15/18 10/18 8/18 4/18 14/18 omental metastasis 18 (16.7%) (83.3%) (55.6%) (44.4%) (22.2%) (77.8%)

Fisher’s P = 0.089 P = 0.206 P = 0.307 exact test

a Annexin A2 low intensity score; no staining (0) and weak staining (1+) b Annexin A2 high intensity score; moderate staining (2+) and strong staining (3+)

Supplementary Table 3: Annexin A2 epithelial immunostaining intensity in the omental tissues. Patient Annexin A2 epithelial staining age Omental Tissue n (years) Intensity % of positive cells Median a b (range) Low High ≤10 ≥10

Cancer cells 58 1/9 8/9 1/9 8/9 adjacent to the 9 (40-70) (11.1%) (88.9%) (11.1%) (88.9%) peritoneum Cancer cells 58 4/9 5/9 5/9 4/9 distant from the 9 (40-70) (44.4%) (55.5%) (55.5%) (44.4%) peritoneum Chi-squared test P = 0.114 P = 0.046

a Annexin A2 low intensity score; no staining (0) and weak staining (1+) b Annexin A2 high intensity score; moderate staining (2+) and strong staining (3+)