Patient-Derived Xenografts as Pre-Clinical Models of Response to Chemotherapy
by
Paulina Cybulska
A thesis submitted in conformity with the requirements for the degree of Master of
Science
Institute of Medical Science
University of Toronto
© Copyright by Paulina Cybulska 2014
ii
Patient-Derived Xenografts as Pre-Clinical Models of Response to Chemotherapy
Paulina Cybulska
Master of Science
Institute of Medical Science
University of Toronto
2014
Abstract Ovarian high-grade serous cancer (HGSC) is the most lethal gynecologic malignancy and well-
characterized models may improve patient outcomes. Patient-derived xenografts (PDXs)
recapitulate disease heterogeneity; however, to be useful in predicting response to novel
chemotherapeutics, they must reflect the response of the donor tissue to standard chemotherapy.
The objectives of this study were: first, to evaluate the response of PDXs’ to platinum therapy
and compare this response to that of the donor; and second, to determine whether treatment with
chemotherapy enriches for tumourigenic cells. Eighteen samples formed tumours in the
mammary fat pads of NOD-Scid-IL2Rγnull mice and were treated with Carboplatin. There was a
100% concordance between sample status and PDXs response to chemotherapy. HGS histology
was confirmed for all cases. A conclusion regarding post-chemotherapy tumourigenicity could
not be made due to inadequate statistical power. PDXs represent useful tools for evaluation of
novel therapies and identification of patients who are platinum-resistant/sensitive.
iii
Acknowledgments & Contributions First, I would like to thank both of my supervisors, Drs. Benjamin Neel and Marcus Bernardini. Ben’s incredible scientific knowledge inspires the pursuit of answers. Ben taught me to always be an inquisitive, critical thinker, not afraid of asking questions or challenging concepts. He also taught me the value of collaboration and scientific ‘consultation’. It has been a memorable, privileged journey.
I cannot understate the impact that Dr. Bernardini has had on my training, both during residency and graduate school. I look up to him professionally, as a brilliant surgeon and clinician, and personally, as a dedicated spouse and parent. I am so grateful for his encouragement, praise, and motivation (‘stay cool’).
In addition, I am so grateful to all members of my committee Dr. Blaise Clarke, Dr. Micheline Piquette, and honorary member, Dr. Amit Oza. I truly value the time you committed these projects. Your ideas and suggestions were always mindful, sensible and instrumental to completing the project.
Thank you to my defense examiners: Dr. Catherine O’Brien, Dr. Ted Brown and Dr. Barbara Vanderhyden. Thank you to Dr. Howard Mount for your time and input.
Thank you Dr. Marcus Bernardini for obtaining all of the clinical patient information. Thank you to Dr. Blaise Clarke for all of the histological annotations and scoring. Thank you to Dr. Kwan Ho Tang for all of the virus work required for the bioluminescence project and to Dr. James Jonkman for your BLI teaching.
My research could not have been initiated if it were not for the wonderful support I received from the department of Obstetrics and Gynecology. Thank you to Dr. Heather Shapiro for encouraging me to pursue this goal. My extreme gratitude to Dr. Donna Steele for her constant wise words, mentorship, support, laughs and tears. You are an amazing woman and I am blessed to have you on my side. Thank you to Drs. Alan Bocking and John Kingdom for making the clinician investigator program an option. Thank you to Dr. Satkunaratnam who allowed me to continue working in the OR, despite nervous hands.
To all of my labmates – none of my achievements (or failures!) could have been possible without your help. I am so lucky to have had the pleasure of working with you all. Thanks to Sherifa Mohamed for dealing with everything administrative. Cathy Iorio, Richard Marcotte, Gordon Chan, Kwan Ho Tang, Angel Sing, Shengqing Gu, Xionan Wang, Raquel De Souza – thank you for all of your help, particularly in the days when I didn’t know how to do simple tasks! Cathy, it has been a tremendous gift to get to know you and your family.
To Dr. Laurie Ailles and members of the Ailles lab, particularly Dr. Craig Gedye, Ella Hyatt and Keira Pereira. To Dr. Gordon Keller and members of the Keller lab- thank you for always including me in your social events. Your friendships are blessings!
Dianne Chadwick (and the Biobank team), Golnessa Mojtahedi and Sarah Rachel Katz, thank
iv
you for putting up with my constant emails and questions. Carl Virtanen- thank you for always making time for me despite your busy schedule.
Thank you to all of my new grad school friends: Alicia Tone, Jennifer Woulter, Tracy Liu, Alec Witty, Keira Pereira, Shawn Stapleton, TD MacDonald. I am grateful for all of the science discussions and social events. To my residency family: thank you for always including me in activities and remembering that I am still part of the program (both my ‘former’ colleagues and ‘new’ colleagues). Special thanks to Laura Sovran, Kelsey Mills, Louise-Helene Gagnon, Helena Frecker, Karli Mayo, and Brian Liu.
To my Ottawa family: Jennifer Brodeur, Nadine Doris, Erin Lamont, Peter Unger, Courtney Maskerine, Julie Hakim, and Robyn Phanouvong; my life would be empty without you. To my Toronto family: Justin Weissglas, Jenny Ferguson, Christa Favot, Caitlin Mckeever, Adrian Sacher, Lauren Lapointe-Shaw, Shannon Corbett, Jennica Platt: you brighten my days.
Dr. Susan Goldstein – your words of wisdom have been indispensible.
Finally, to the sister I never knew I had – Jocelyn Stewart, my pip. Getting to know and love you has been one of the greatest gifts. You made this project possible through your scientific input and amazing personality (you are the reason I dance with my arms up). I cannot wait to share more experiences with you. Thank you to Mamma Lynne and Papa Tom for incorporating me into the family.
Lastly, to my mom: nothing could be possible without you. You have always made me believe that I can achieve anything. Thanks to your courage, I am able to succeed. Kocham mocno.
v
Table of Contents !
PATIENT'DERIVED+XENOGRAFTS+AS+PRE'CLINICAL+MODELS+OF+RESPONSE+TO+
CHEMOTHERAPY+..................................................................................................................................................+I!
ACKNOWLEDGMENTS+&+CONTRIBUTIONS+................................................................................................+III!
LIST+OF+TABLES+................................................................................................................................................+VIII!LIST+OF+FIGURES+.................................................................................................................................................+IX!
ABBREVIATIONS+..................................................................................................................................................+X!CHAPTER+1+.............................................................................................................................................................+1!
OVARIAN+CANCER:+INTRODUCTION+&+OVERVIEW+...................................................................................+1!1.1!EMBRYOLOGY!OF!GONADAL!DEVELOPMENT!................................................................................................................!1!1.2!CLASSIFICATION!.................................................................................................................................................................!3!1.3!EPIDEMIOLOGY!...................................................................................................................................................................!4!1.4!PATHOGENESIS!...................................................................................................................................................................!4!1.4.1$Role$of$TP53$in$Ovarian$Cancer$Pathogenesis$..............................................................................................$6!1.4.2$Role$of$BRCA1$and$BRCA2$in$Ovarian$Cancer$Pathogenesis$...................................................................$6!
1.5!RISK!FACTORS!....................................................................................................................................................................!7!1.5.1$Patient$Characteristics$............................................................................................................................................$7!1.5.2$Risk$from$Familial$Cancer$Syndromes$..............................................................................................................$7!1.5.3$Reproductive$History$Risk$......................................................................................................................................$9!1.5.4$Inflammation$............................................................................................................................................................$11!
1.6!OVARIAN!CANCER!ETIOLOGY!........................................................................................................................................!11!1.6.1$The$Menstrual$Cycle:$A$Brief$Overview$..........................................................................................................$11!1.6.2$Ovarian$Cancer$Etiology:$The$CellNofNOrigin$Controversy$....................................................................$13!1.6.3$Ovarian$Cancer$Etiology:$Incessant$Ovulation$..........................................................................................$15!1.6.4$Ovarian$Cancer$Etiology:$The$Role$of$Hormones$......................................................................................$16!
1.7!SCREENING!.......................................................................................................................................................................!16!1.8!PRESENTATION!................................................................................................................................................................!17!1.9!DIAGNOSIS!........................................................................................................................................................................!18!1.9.1$Cancer$Antigen$125$................................................................................................................................................$18!1.9.2$Radiologic$Studies$...................................................................................................................................................$19!
vi
1.10!MANAGEMENT!OF!DISEASE!........................................................................................................................................!20!1.10.1$Management$of$Primary$Disease:$Surgery$................................................................................................$20!1.10.2$Management$of$Primary$Disease:$Chemotherapy$.................................................................................$21!1.10.3$Management$of$Recurrent$Disease$...............................................................................................................$23!
1.11!MECHANISMS!OF!THERAPY!RESISTANCE!.................................................................................................................!25!1.12!TUMOUR!HETEROGENEITY!.........................................................................................................................................!25!1.13!CURRENT!MODELS!OF!OVARIAN!CANCER!................................................................................................................!28!1.13.1$In$Vitro$Models$of$Ovarian$Cancer$................................................................................................................$28!1.13.2$In$Vivo$Models$of$Ovarian$Cancer$.................................................................................................................$29!
1.14!CURRENT!STATE!OF!THE!PROBLEM!..........................................................................................................................!33!1.15!RESEARCH!QUESTIONS!&!THESIS!OVERVIEW!.........................................................................................................!34!1.16!RESEARCH!OBJECTIVES!&!HYPOTHESES!..................................................................................................................!35!
CHAPTER+2+...........................................................................................................................................................+36!
PATIENT'DERIVED+XENOGRAFTS+AS+PRE'CLINICAL+MODELS+OF+RESPONSE+TO+
CHEMOTHERAPY+...............................................................................................................................................+36!2.1!INTRODUCTION!................................................................................................................................................................!37!2.1.1$Drug$Dose$Translation$..........................................................................................................................................$38!2.1.2$CA125$...........................................................................................................................................................................$40!2.1.3$Tumour$Initiating$Cells$........................................................................................................................................$41!
2.2!METHODS!.........................................................................................................................................................................!43!2.2.1$Identification$of$Patient$Samples$.....................................................................................................................$43!2.2.2$Tumour$Processing$&$Establishment$of$Xenografts$................................................................................$44!2.2.3$Tumour$Measurement$..........................................................................................................................................$45!2.2.4$Carboplatin$Administration$...............................................................................................................................$45!2.2.5$CA125$ELISA$&$Immunohistochemistry$........................................................................................................$45!2.2.6$Limiting$Dilution$Assays$......................................................................................................................................$46!
2.3!RESULTS!............................................................................................................................................................................!47!2.3.1$Engraftment$..............................................................................................................................................................$47!2.3.2$Patient$Demographics$..........................................................................................................................................$47!2.3.3$Mouse$Carboplatin$MTD$......................................................................................................................................$50!2.3.4$PDXs$from$PlatinumNSensitive$Patients$Show$a$NearNComplete$Response$to$Platinum$
Therapy$..................................................................................................................................................................................$51!2.3.5$PDXs$from$PlatinumNResistant$Patients$do$not$Respond$to$Platinum$Therapy$..........................$57!
vii
2.3.6$‘Prospective’$PDXs$Predict$ChemoNResponsiveness$..................................................................................$75!2.3.7$PDX$‘Recurrence’$also$Parallels$Patient$Clinical$Course$........................................................................$79!2.3.8$Summary$of$PDXs’$Response$to$Carboplatin$...............................................................................................$80!2.3.9$Primary$Tumour$Histology$is$Maintained$in$PDX$Tissue$......................................................................$81!2.3.10$PDXs’$Serum$CA125$Cannot$be$Used$to$Monitor$Disease$Progression$and/or$Response$......$81!2.3.11$PDXs$CA125$Immunohistochemistry$Parallels$Patient$Serum$CA125$Value$.............................$83!2.2.12$Effect$of$Carboplatin$Treatment$on$TICs$...................................................................................................$86!
2.4!DISCUSSION!......................................................................................................................................................................!93!2.4.1$Chemotherapy$..........................................................................................................................................................$94!2.4.2$Assessment$of$Response$........................................................................................................................................$95!2.4.3$Patients’$Clinical$Outcomes$................................................................................................................................$97!2.4.4$TumourNInitiating$Cells$......................................................................................................................................$100!
CHAPTER+3:+......................................................................................................................................................+104!
CONCLUSIONS+&+FUTURE+DIRECTIONS+...................................................................................................+104!REFERENCES+.....................................................................................................................................................+108!
APPENDIX+1:+ESTABLISHMENT+OF+AN+ORTHOTOPIC+BIOLUMINESCENT+PATIENT'DERIVED+
XENOGRAFT+MODEL+OF+OVARIAN+CANCER+...........................................................................................+138!APPENDIX+2:+DISULFIRAM'INDUCED+TOXICITY+IN+OVARIAN+CANCER+.........................................+156!
viii
List of Tables
1.0 Ovarian cancer staging
2.1 Previously published Carboplatin doses administered to mice
2.2 Patient Samples Injected: engraftment results
2.3 Patient demographics and clinical features of xenografted tumours
2.4 PDX CA125 Immunohistochemistry
2.5 Summary of LDA results for case 3654
2.6 Summary of LDA results for case 3875
2.7 Summary of LDA results for case 2555
2.8 Summary of LDA results for case 6447
2.9 Summary of LDA results for case 67326
ix
List of Figures
1.0 1.1 1.2
Classification of Ovarian Cancers The Menstrual Cycle Stochastic and hierarchical models of tumour heterogeneity
2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22 2.23 2.24 2.25
LDA method and sample preparation Establishment of Carboplatin Maximum Tolerated Dose (MTD) Patient 3670’s clinical course and PDXs’ response to platinum Patient 3748’s clinical course and PDXs’ response to platinum Patient 61382’s clinical course and PDXs’ response to platinum Patient 3654/3875’s clinical course and PDXs’ response to platinum Patient 6447’s clinical course and PDXs’ response to platinum Patient 2261’s clinical course and PDXs’ response to platinum Patient 2028’s clinical course and PDXs’ response to platinum Patient 2803’s clinical course and PDXs’ response to platinum Patient 2489’s clinical course and PDXs’ response to platinum Patient 3444’s clinical course and PDXs’ response to platinum Patient 2555’s clinical course and PDXs’ response to platinum Patient 2903’s clinical course and PDXs’ response to platinum Patient 2753’s clinical course and PDXs’ response to platinum Patient 2685’s clinical course and PDXs’ response to platinum Patient 67199’s clinical course and PDXs’ response to platinum Patient 67326’s clinical course and PDXs’ response to platinum 3670-PDXs’ response to second course of treatment with Carboplatin Tumour volume change in all xenograft cases PDXs 3670 and 3654 serum CA125 ELISA results Carboplatin sensitivity experiment for case 3654 Carboplatin sensitivity experiment for case 3875 Carboplatin sensitivity experiment for case 2555 Carboplatin sensitivity experiment for case 6447 Carboplatin sensitivity experiment for case 67326
x
Abbreviations
AMH Anti-Müllerian Hormone AA Acetaldehyde ALDH Aldehyde dehydrogenase AMG Medical Oncology trial drug AUC Area under the curve BLI Bioluminescent imaging BRAF Proto-oncogene B-Raf BRCA1 Breast Cancer 1, early onset BRCA2 Breast Cancer 2, early onset BSA Body Surface Area BSO Bilateral salpingo-oophorectomy C.I. Confidence interval CA125 Cancer antigen 125 CI Combination Index CL Corpus luteum CRP C-reactive protein CSC Cancer stem cell CT Computed tomography DMEM Dulbecco’s modified eagle medium DMSO Dimethyl sulfoxide DNA Deoxyribonucleic acid DSB Double strand break DSF Disulfiram ED Estimated dose EOC Epithelial Ovarian carcinoma FA Fanconi anemia FACS fluorescence-activated cell sorting FBS Fetal bovine serum FDA U.S. Food and Drug administration FSH Follicle Stimulating Hormone FT Fallopian tube FTE Fallopian tube epithelium γ-H2AX Gamma-histone 2AX GEMM Genetically engineered mouse model
xi
GFP Green fluorescent protein GnRH Gonadotropin-releasing hormone GOG Gynecologic Oncology Group H&E Hematoxylin and eosin HBOC Hereditary breast and ovarian cancer HBSS Hank’s buffered salt solution HGSC High-grade serous cancer HMG High mobility group HOXA Homeobox A HR Homologous recombination HRT Hormone replacement therapy IB Intrabursal IC Inhibitory Concentration ICL Interstrand crosslink IDS Interval debulking IHC Immunohistochemistry IL Interleukin IMDM Iscove’s Modified Dulbecco’s Medium IP Intraperitoneal IUD Intrauterine device IV Intravenous LAR Low anterior resection LDA Limiting dilution assay LH Luteinizing hormone LOH Loss of heterozygosity Luc Luciferase MAPK Mitogen activated protein kinase MDA Microtubule disrupting agents MFP Mammary fat pad MIS Müllerian Inhibiting Substance MMR Mismatch repair MRI Magnetic resonance imaging MTD Maximum tolerable dose MUC 16 Mucin 16 NACT Neoadjuvant chemotherapy NER Nucleotide excision repair NFκB Nuclear factor kappa light chain enhancer of activated B cells
xii
NK Natural killer NOD Non-obese diabetic NSAID Non-steroidal anti-inflammatory NSG NOD-scid-IL2Rγnull OCP Oral contraceptive pill OS Overall survival OSE Ovarian Surface epithelium PARP Poly(ADP)-ribose polymerase PAX2 Paired-box 2 PBS Phosphate buffered saline PDS Primary debulking surgery PDX Patient-derived xenograft PET Positron emission tomography PFI Platinum free interval PFS Progression free survival PID Pelvic inflammatory disease PIK3 Phosphatidylinositol 3’-kinase PLD Pegylated liposomal doxorubicin PM Princess Margaret Cancer Centre PPV Positive predictive value PTEN Phosphatase and tensin homolog RAR Retinoic acid receptor
Rb RECIST
Retinoblastoma Response Evaluation Criteria in Solid Tumours
RNA Ribonucleic acid ROI Region of interest RPMI Roswell Park Memorial Institute medium RR Relative risk SC Subcutaneous SCID Severe combined immunodeficiency SCOUT Secretory cell outgrowth SO Salpingo-oophorectomy SRY Sex-determining region STIC Serous tubal intraepithelial carcinoma TP53 Tumour protein 53 TAH Total abdominal hysterectomy
xiii
TCGA The Cancer Genome Atlas TIC Tumour initiating cell TILT Intraepithelial lesions in transition TVUS Transvaginal ultrasound UHN University Health Network WHO World Health Organization
1
Chapter 1
Ovarian Cancer: Introduction & Overview
1.1 Embryology of Gonadal Development
Knowledge of the embryology of the reproductive tract is important for understanding the genes
and cell signaling pathways involved in normal development and the alterations that lead to
disease. The third week of human embryonic life is marked by gastrulation, a process wherein
the bilaminar embryonic disk transforms into a trilaminar disk (composed of ectoderm,
mesoderm and endoderm). By the end of the third week, the mesoderm subdivides further into 3
distinct components, from medial to lateral: paraxial mesoderm, intermediate mesoderm and
lateral plate mesoderm. The urogenital ridge originates from the intermediate mesoderm and will
ultimately form the reproductive and excretory systems (DeCherney and Nathan, 2007). At
approximately 40 days of gestation, primordial germ cells migrate from the yolk sac to the
urogenital ridge, which divides further into the nephrogenic and genital ridges, respectively. The
mesonephric (Wolffian) ducts and mesonephric kidneys develop from the nephrogenic ridge,
while paired paramesonephric (Müllerian) ducts develop from invaginations of the coelomic
epithelium.
Sex determination is based upon the presence or absence of the sex-determining region (SRY)
gene and anti-Müllerian hormone (AMH). Female embryos lack the SRY gene and AMH, which
cause the development of the gonad into the ovary and the persistence of the Müllerian ducts,
respectively. The Müllerian ducts ultimately grow and fuse caudally to form the proximal
vagina, while their unfused cephalad portions become the fallopian tubes (Hoffman et al., 2012).
All proposed sites of serous cancers: the OSE, fallopian tubes and the lining of the peritoneal
cavity are derived from the coelomic epithelium (Auersperg, 2013b).
Although the embryology of the female reproductive system is fairly well described, the cellular
and molecular mechanisms that regulate its development are poorly understood. Our current
knowledge stems largely from mouse knockout studies. In mice, Müllerian duct formation
2
requires a small set of homeodomain-containing transcription factors and signaling molecules.
One of these is paired-box gene 2 (Pax2). Pax2-null mice die soon after birth and lack a
reproductive tract due to the degeneration of both the mesonephric and the paramesonephric
ducts (Kobayashi and Behringer, 2003). Similarly, Lim-1 and Emx2 are required for Müllerian
duct formation; Emx2 and Lim-1-null mutant mice lack fallopian tubes, a uterus and the upper
portion of the vagina (Kobayashi and Behringer, 2003).
Retinoic-acid signaling also plays a critical role in Müllerian duct development. Females with
compound mutations in retinoic-acid receptor genes (RARα1, RARα2, RARβ2 and RARϒ) lack
reproductive organs (Lohnes et al., 1994; Mendelsohn et al., 1994a; Mendelsohn et al., 1994b).
Likewise, Wnt gene family members, specifically Wnt4, Wnt5a and Wnt7a, are required for
normal Müllerian development in females and Müllerian duct regression in males (Miller et al.,
1998).
As stated above, female embryos lack AMH (a.k.a. Müllerian inhibiting substance, MIS), a
hormone that is both necessary and sufficient for the regression of the Müllerian duct system.
Mis-mutant male mice have testes and are virilized but have a uterus and fallopian tubes. When
Mis is overexpressed in female mice, Müllerian duct-derived organs are absent (Behringer, 1994;
Behringer et al., 1994).
The Homeobox A (Hoxa) genes are a family of highly conserved transcription factors. They are
expressed along the anterior-posterior axis of the Müllerian ducts according to their 3’-5’ order
in the HOX cluster, and specify the identity of each segment: HOXA9 is expressed in the oviduct,
HOXA10 in the uterus, HOXA11 in the uterus and cervix and HOXA13 in the cervix and upper
vagina (Taylor et al., 1997). The reproductive tract is unique in that HOX genes continue to be
expressed in these tissues in adulthood and are important not only for proper reproductive tract
development but also for appropriate adult function (Taylor, 2000). High-grade serous ovarian
cancers express HOXA9 and forced expression of Hoxa9 in undifferentiated murine ovarian
surface epithelium cells can induce the formation of ovarian HGSC (Bast et al., 2009).
3
1.2 Classification
Figure 1.0 Classification of Ovarian Cancers
The current classification scheme for ovarian tumours is based on histogenesis of the normal
ovary. Tumours of the ovary were initially thought to originate from one of the three cell types
found within gonadal tissue: sex cord-stromal cells, germ cells and surface-epithelial cells.
However, recent evidence suggests that most EOC, and likely all HGSC, do not originate from
the surface-epithelium, but rather from the fallopian tube epithelium (FTE). These details will be
discussed further in Sections 1.4 and 1.6. Furthermore, epithelial tumours are not a single
disease, but rather comprise a diverse group of tumours that can be classified based on distinctive
morphologic and molecular genetic features (Kurman and Shih Ie, 2010). These can be further
subdivided into serous, mucinous, endometrioid, clear cell and transitional cell (Figure 1.0).
Ovarian Cancer
Sex Cord Stromal
(5%)
Epithelial (90%)
Mucinous Endometrioid Serous (80%)
Low Grade (10%)
High Grade (90%)
Clear Cell Transitional
Germ Cell (5%)
4
1.3 Epidemiology
According to the WHO, ovarian cancer is the 6th most common cancer in women and the 7th
most common cause of cancer death. In 2008, about 225,000 women were diagnosed and 140,00
died from this disease worldwide (Jemal et al., 2011; Organization, 2008). In Canada, it is the 7th
most common cancer type and the 5th most common cause of cancer death, representing 2600
new cases and 1750 deaths annually, thus making it the most lethal gynecologic malignancy
(Society, 2012). Epithelial ovarian cancer is the most common gynecologic cancer, with an
overall incidence of 9.4/100,000 and an incidence of 33/100,000 in women over age 50
(Hennessy et al., 2011; Schorge et al., 2010). Based on the US cancer data, the annual incidence
is highest in white women (13.4 per 100,000), compared with Hispanic (11.3 per 100,000), black
(9.8 per 100 000) or Asian (9.8 per 100 000) women (Edwards et al., 2002). The median age of
patients with ovarian cancer is 60 years, and the average lifetime risk for women is about 1 in 70
(Cannistra, 2004). The age of diagnosis is younger in women with a hereditary ovarian cancer
syndrome. For example, women with Lynch syndrome are typically diagnosed between the ages
of 43-50 (Watson et al., 2001). Similarly, women with a BRCA1 mutation are younger at
diagnosis, as their risk of ovarian cancer is 2-3% by age 40 (King et al., 2003). Lifetime risks are
also significantly higher, with an estimate in the range of 28-66% for BRCA1 mutation and 16-
27% for BRCA2 mutation (Satagopan et al., 2002).
1.4 Pathogenesis
As mentioned, ovarian tumours are classified according to the cell types from which they are
derived (Figure 1.0). Ninety percent of ovarian tumours are classified as epithelial (Hennessy et
al., 2011); however, most recent evidence suggests that epithelial neoplasms can be further
divided into two groups: those that arise from the ovary, and those that arise from the fallopian
tube (Tone et al., 2012a). The first group, which likely originates in the ovary, consists of: low-
grade endometrioid, mucinous, clear cell, borderline and low grade serous. This group is often
classified as Type I, as they are low grade, present at an early stage and behave in a relatively
indolent manner. They develop in a slow, step-wise fashion from well-defined precursor lesions.
5
Genomically, they are characterized by mutations in KRAS, BRAF, ERBB2, PTEN, or PIK3CA
and display a low level of chromosomal disruption (Vang et al., 2013).
Conversely, the second type of EOC (Type II), typically appears in the absence of a well-defined
precursor, presents at an advanced stage and are high-grade (Kurman and Shih Ie, 2010). They
metastasize early in the disease course, with tumours found on the ovarian surface, fallopian tube
and peritoneal cavity (Crum et al., 2007b; Shih Ie and Kurman, 2004). Ovarian HGSC is largely
synonymous with Type II tumours (Bowtell, 2010). Originally thought to be mainly ovary-
derived, Type II tumours are now thought to originate from both the FTE and OSE (Auersperg,
2013a, c; Tone et al., 2012a; Tone et al., 2012b) (discussed further in Section 1.6). In contrast to
Type I tumours, molecular abnormalities in Type II tumours include mutations in the tumour
suppressors TP53 (>95% of cases) and BRCA1 or BRCA2 (20-50% of cases) (Hennessy et al.,
2010; Network, 2011; Press et al., 2008a; Schrader et al., 2012). Amplification and
overexpression of the genes in the PI3K family occur in more than 40% of type II cancers,
conferring activation of the PI3K pathway and consequently, cell proliferation and survival.
Additionally, the RB pathway is altered in >65% of cases, leading to aberrant cell cycle
regulation. Less than 1% of Type II tumours have mutations of BRAF, KRAS and PIK3CA,
which are typically found in type I tumours (Romero and Bast, 2012).
Recently, the Cancer Genome Atlas (TCGA) analyzed the genomes of 489 high-grade serous
ovarian adenocarcinomas. They concluded that the mutation spectrum of ovarian HGSC is
completely distinct form other EOC subtypes. Furthermore, on the basis of gene expression
profiles, the TCGA found at least four subtypes within ovarian HGSC, consisting of:
immunoreactive, differentiated, proliferative and mesenchymal; however, survival was not
dictated by the transcriptional subtype (Network, 2011). Verhaak et al. expanded the analysis of
the TCGA and showed that individual ovarian HGSC samples often express multiple subtype
signatures. Using this analysis, they were also able to develop the Classification of Ovarian
Cancer (CLOVAR) survival signatures with significant ability to predict outcome to therapy
(Verhaak et al., 2013).
6
1.4.1 Role of TP53 in Ovarian Cancer Pathogenesis
The Cancer Genome Atlas study showed that >95% of ovarian HGSC have TP53 mutations
(Network, 2011). TP53 mutation is an early and obligatory event in ovarian HGSC development
(Bowtell, 2010). Wild-type P53 is a nuclear phosphoprotein, encoded by a highly conserved 20-
Kb gene. The central core of the protein is made up primarily of the DNA-binding domain
required for sequence-specific DNA binding. It is largely within this binding domain that most
missense mutations occur and result in a dysfunctional protein (Bai, 2006). As a tumour
suppressor, TP53 is essential for preventing inappropriate cell proliferation. It does so by
activating diverse downstream targets that, in turn, elicit various responses such as cell cycle
check points, cell survival and apoptosis (Christie and Oehler, 2006; Hofseth et al., 2004).
Mutations that lead to loss of TP53 function, as seen in ovarian HGSC, result in failure to
activate these responses, leading to unrepaired DNA damage, increased chromosomal instability,
and ultimately, tumour formation (Buttitta et al., 1997; Tomasini et al., 2008). TP53 loss
associated with chromosomal instability is believed to contribute to the widespread copy number
change seen in ovarian HGSC (Bowtell, 2010).
1.4.2 Role of BRCA1 and BRCA2 in Ovarian Cancer Pathogenesis
Germline or somatic mutations in BRCA1 and BRCA2 are found in approximately 20-50% of
ovarian HGSCs (Castellarin et al., 2013; Hennessy et al., 2010; Network, 2011; Pal et al., 2005;
Press et al., 2008a; Zhang et al., 2011). The BRCA1 and BRCA2 genes are located on
chromosomes 17q21 and 13q12.3, respectively, and their proteins prevent malignant
transformation by helping to maintain genomic integrity through their respective roles in DNA
repair, particularly homologous recombination (HR) (Powell et al., 2013).
BRCA1 comprises 24 exons encoding a protein of 1863 amino acids (Miki et al., 1994). It
mediates HR repair of double strand breaks (DSBs) by binding directly to DSBs through
association with the abraxas-RAP80 macro-complex. This complex is ultimately responsible for
DSB resection, and functions with cell cycle checkpoints to prevent propagation of DNA damage
(Powell et al., 2013). BRCA1 has also been shown to regulate cell cycle progression via roles in
the G1/S, G2/M and spindle checkpoints (Deng, 2006).
7
BRCA2 consists of 27 exons encoding a protein of 3418 amino acids and is a member of the
Fanconi Anemia (FA) DNA repair pathway (Ramus and Gayther, 2009). The FA-BRCA DNA
repair pathway comprises 13 proteins, which act together to repair DNA interstrand crosslinks
(ICLs) and respond to stalled replication forks (Abraham et al., 2011). BRCA2 (a.k.a. FANCD1)
is integral to this process through recruitment of the recombinase RAD51 to DSBs. RAD51 is
essential for HR and is responsible for the tumour suppressive function of the repair process
(Moynahan et al., 2001).
1.5 Risk Factors
1.5.1 Patient Characteristics
Epidemiologic studies examining typical cancer risk factors, such as socioeconomic status,
alcohol consumption, smoking, and physical activity have been inconsistent with respect to
ovarian cancer risk (Zografos et al., 2004). On the other hand, age and family history have been
consistently identified as major risk factors. HGSC rates increase with older age, with mean age
at diagnosis of 55-65 years (Vang et al., 2009). Furthermore, women with a family history of
ovarian cancer are three to four times more likely to develop ovarian cancer than those without
such history (Banks, 2001). A geographic distribution for ovarian cancer has also been found,
with incidence increasing with latitudinal distance away from the equator. Some evidence point
to vitamin D deficiency as a potential mechanism for this distribution, however, the underlying
mechanisms have not been described (Bakhru et al., 2010; Garland et al., 2006; Lefkowitz and
Garland, 1994). Additionally, the risk of ovarian cancer increases for women who migrate from a
country with a low-risk incidence to a country with a higher risk incidence (Organization, 2008).
1.5.2 Risk from Familial Cancer Syndromes
Although the vast majority (~90%) of ovarian cancer cases are sporadic, the greatest and most
reliable risk factors for ovarian HGSC are a positive family history and/or a known BRCA1/2
mutation. If either gene is mutated in the germ line, the result is hereditary breast and ovarian
cancer (HBOC) syndrome, which causes a high lifetime risk to not only breast and ovarian
8
cancers, but also pancreatic, stomach, laryngeal and prostate cancers (Roy et al., 2012). The most
commonly reported BRCA1 mutations are 185delAG and 5382insC, while the most common
BRCA2 mutation is 6174delT (Boyd, 2003). These mutations are inherited in an autosomal
dominant pattern. More than 1200 mutations have been reported for BRCA1/2 with ~80% being
either frameshift or nonsense, causing the formation of a truncated protein (ACOG, 2009; Boyd,
2003).
A meta-analysis of six studies showed that 5.7% and 3.8% of all ovarian cancers are associated
with germline BRCA1 and BRCA2 mutations, respectively, although more recent studies have
found a higher mutation prevalence of 13-20% (Boyd, 2003; Schrader et al., 2012; Zhang et al.,
2011). Mutations in BRCA1 confer a risk of ovarian cancer of ~50% by age 70, with a lifetime
risk of 16-63% (Easton et al., 1995; Ford and Easton, 1995). BRCA2 mutations confer a ~10%
risk of ovarian cancer by age 70 and a lifetime risk of 20-30% (Boyd, 2003; Lakhani et al.,
2004). Risks of ovarian cancer are significantly higher in BRCA1 mutation carriers, compared
with BRCA2 (cumulative risk of 0.21 vs. 0.02 by age 50), and tumours tend to occur at a younger
age (mean age 40s/50s vs. 60 years for BRCA2 carriers) (Boyd and Rubin, 1997; Gayther et al.,
1995; King et al., 2003). These figures stand in stark contrast to the lifetime risk of ovarian
cancer of 1.4% in the general population.
Approximately 1 in 280 women carry a germline BRCA1/2 mutation; however, approximately
40% of all Ashkenazi Jewish ovarian cancer patients are BRCA1/2 mutation carriers, putting this
population at particularly high risk (Boyd, 2003). Other populations with more frequent
BRCA1/2 mutations include French Canadians and Icelanders (ACOG, 2009).
Whereas 90% of hereditary ovarian cancers are associated with a mutation in BRCA1/2, the
remainder can be attributed to mutations in the DNA mismatch repair (MMR) genes, primarily
seen in association with Lynch Syndrome (LS) (formerly: hereditary nonpolyposis colon cancer,
HNPCC). Lynch syndrome occurs due a germline mutation in one of several mismatch repair
genes (MMR): MLH1, MSH2, MSH6, PMS1 or PMS2, and like BRCA1/2, it is inherited in an
autosomal dominant pattern (Crispens, 2012). The MMR pathway is involved in the removal of
DNA base mismatches that arise either during DNA replication or as a result of DNA damage.
Mutation rates are 100- to 1000-fold greater in MMR-deficient tumour cells compared with
9
normal cells (Martin et al., 2010). As a result, the lifetime risk of ovarian cancer in women with
LS has been estimated to be 7-12%, with a younger age at diagnosis (mean age 43-49 years)
(Ketabi et al., 2011; Watson et al., 2008). Although most of these tumours are endometrioid, a
small fraction (~10%) has high-grade serous histology (Grindedal et al., 2010).
1.5.3 Reproductive History Risk
Studies examining the relationship between age of menarche and ovarian cancer risk have been
conflicting. Rodriguez et al. found that menarche after age 12 correlates significantly with lower
ovarian cancer rates, compared with menarche at a younger age, but other studies have been
inconclusive (Franceschi et al., 1991; Rodriguez et al., 1995).
The data surrounding age of menopause are similarly conflicting. A large European study found
that, compared with women whose menopause occurred at age 44 or earlier, the relative risk
(RR) for ovarian cancer was 1.4 for women having menopause between the ages of 45 and 49,
1.6 for women with menopause between 50 and 52, and 1.9 if menopause occurred after the age
of 52 (Franceschi et al., 1991). Similarly, women who had surgical menopause saw protective
effects, with a RR= 0.7 (Hartge et al., 1988). However, parity-adjusted data from the USA and
Australia found no trend in ovarian cancer risk with increasing age of menopause (Purdie et al.,
1995; Whittemore et al., 1992).
Unlike menarche and menopause, the association between parity/gravidity and ovarian cancer is
well established (Adami et al., 1994). In the 1970s, Beral et al. found a clear inverse relation
between average completed family size and mortality from ovarian cancer in various populations
of women and successive generations (Beral et al., 1978). In the Nurses Health Study, a RR of
0.84 per pregnancy was observed (Hankinson et al., 1995). Furthermore, Albrektsen et al.
showed that the RR of epithelial ovarian cancer was 0.56 for women with three pregnancies but
did not decrease with additional pregnancies (Albrektsen et al., 1996). Conversely, nulliparous
women have a marked increase in risk, reaching three times that of parous women (Bandera,
2005; Stewart et al., 2013).
10
Like gravidity, lactation suppresses ovulation and has a favorable role in ovarian cancer risk
reduction. The protection against ovarian cancer is related to total lactation time, with longer
durations conferring the greatest risk reduction (Greggi et al., 2000; Yen et al., 2003).
Likewise, use of the oral contraceptive pill (OCP) has a substantial protective effect on ovarian
cancer. Several cohort and case-control studies have demonstrated a 40% reduction of ovarian
cancer in OCP users, and this number increases to 50% if OCPs have been used for 5 years or
longer (Bosetti et al., 2002; Chiaffarino et al., 2001; La Vecchia and Franceschi, 1999). Similar
protective effects are seen with the use of an intrauterine device (IUD) (Ness et al., 2011).
Gravidity, lactation and OCPs all suppress ovulation and are thought to be protective through a
common mechanism of either ovulation suppression or altered levels of steroid hormones (Ness
and Cottreau, 1999).
Unfortunately, the link between hormone replacement therapy (HRT) and ovarian cancer has not
been as clear. Several case-control studies have shown that any history of HRT-use is related to a
modest increase (RR 1.4) in ovarian cancer risk, (Chiaffarino et al., 1999; Riman et al., 2002a;
Riman et al., 2002b); however, other studies found no such association (La Vecchia, 2006;
Lukanova and Kaaks, 2005; Sit et al., 2002). Two large prospective studies did find an
association between HRT use and ovarian cancer. The first, a Danish study looking at >900,000
women, found that compared with women who never took hormone therapy, current users of
hormones had EOC incidence rate ratios of 1.44 (95% C.I., 1.30-1.58). The researchers estimated
that 1 extra ovarian cancer occurred in approximately 8300 women taking hormone therapy each
year (Morch et al., 2009). The second study, the UK Million Women Study, found a similarly
increased risk of EOC for current HRT users, with a relative risk of 1.53 (95% C.I., 1.31-1.79)
(Beral et al., 2007). Neither of these studies examined the risk specifically for ovarian HGSC.
Lastly, tubal ligation (TL) reduces ovarian cancer risk. This procedure does not entail removal of
any part of the fallopian tube, but simply creates an interruption in the fallopian tube lumen. The
Nurses’ Health Study reported a dramatic protective effect of TL, with an odds ratio of 0.33
(95% C.I., 0.16-0.64) (Hankinson et al., 1995). Several mechanisms have been suggested to
explain the protective effects of TL. TL may reduce blood flow to the ovary resulting in
anovulation and altered levels of hormones and growth factors. Alternatively, TL could block the
11
retrograde flow of carcinogenic or inflammatory agents into the peritoneal cavity (Sieh et al.,
2013). TL has been suggested as a feasible risk-reducing option for certain at-risk women (Narod
et al., 2001)
1.5.4 Inflammation
A growing body of literature suggests that the OSE and FTE are chronically exposed to an
inflammatory environment, due to normal physiological events (ovulation, menstruation), as well
as pathological states (endometriosis, pelvic inflammatory disease). Pro-inflammatory cytokines,
such as C-reactive protein (CRP), are present in ovulatory fluid and menstrual effluent. CRP also
is increased markedly in endometriosis (implantation of ectopic endometrial tissue in the
abdomen and pelvis) and pelvic inflammatory disease (PID; ascension of sexually transmitted
infections into the pelvis) (Hunn and Rodriguez, 2012). Inflammation produces toxic oxidants
intended to kill pathogens; however, these oxidants can also damage host DNA, proteins and
lipids, and could play a direct role in carcinogenesis (Dreher and Junod, 1996). Elevated serum
levels of CRP, prostaglandins and cytokines (specifically interleukins, IL-2, IL-4, IL-6, IL-12,
and IL-13) are elevated in women who subsequently develop EOC (Clendenen et al., 2011;
McSorley et al., 2007; Toriola et al., 2011).
Environmental toxins can enter the genital tract and also cause inflammation. This association
was first seen between talc use and EOC. A meta-analysis of 16 studies demonstrated a 33%
increase risk of EOC in talc users, particularly in HGSC (Gertig et al., 2000; Huncharek et al.,
2003). Talc particles have also been seen in ovarian neoplasms, providing some further evidence
that these can act as direct carcinogens (Henderson et al., 1979).
1.6 Ovarian Cancer Etiology
1.6.1 The Menstrual Cycle: A Brief Overview
Understanding the proposed etiologies and pathogenesis of ovarian cancers requires knowledge
of the normal physiology of the reproductive tract. The production and release of an ovum is a
12
carefully regulated process coordinated between the hypothalamus, pituitary and ovaries. In
humans, the typical 28-day cycle consists of follicle development, ovulation and luteinization
(Figure 1.1). The follicular phase comprises the first half of the cycle, during which a dominant
follicle is selected from a pool of candidate follicles. During the follicular phase, estradiol levels
rise and the endometrium proliferates. Ovulation is defined by the release of the oocyte from the
dominant follicle, and formation of the corpus luteum. In the luteal phase, the corpus luteum
produces high levels of estradiol and progesterone in order to support an impending pregnancy.
Accordingly, the endometrium transforms into a secretory state to facilitate implantation. In the
absence of pregnancy, menstruation occurs as the corpus luteum regresses. Similarly, the
fallopian tube epithelium undergoes cyclical changes comparable with the proliferative and
secretory patterns of the endometrium, with the follicular phase being ‘secretory’, and the luteal
phase being a ‘recovery’ stage (Crow et al., 1994).
Figure 1.1 The Menstrual Cycle
Coordination of the ovarian-uterine cycle involves complex interplay between the hypothalamic
release of gonadotropin-releasing hormone (GnRH), pituitary gonadotropin release and ovarian
sex steroid production (Bieber, 2006). Neurons in the arcuate nucleus of the mediobasal
13
hypothalamus secrete GnRH in a pulsatile manner into the hypophyseal portal circulation. These
act to stimulate the anterior pituitary to release gonadotropins: follicle-stimulating hormone
(FSH) and luteinizing hormone (LH), which act on their respective target cells in the gonad. The
rise of FSH in the follicular phase, which begins on the first day of menstruation, stimulates the
granulosa cells of the ovary to proliferate, and express aromatase. LH stimulates the theca cells
of the ovary to produce androgens, which are then, aromatized in the granulosa cells to produce
estradiol. Ovulation occurs ~38 hours after the LH surge and the corpus luteum (CL) develops
from the ruptured ovarian follicle. This marks the beginning of the luteal phase. Progesterone is
released from the CL and slows GnRH drive, which leads to decreased LH and FSH release. If
pregnancy does not occur, the CL regresses causing a drop in progesterone and estradiol, and
subsequent menstruation (Bieber, 2006).
1.6.2 Ovarian Cancer Etiology: The Cell-of-Origin Controversy
The cell-of-origin and pathogenesis of ovarian cancer continue to be current topics of debate.
The traditional view of ovarian carcinogenesis has been that all EOC tumours are derived from
the OSE. Specifically, following ovulation, inclusion cysts form during repair of ovulatory
‘wounds’ in the OSE. These form invaginations into the ovarian cortex, pinch off from the
surface of the ovary and sit within the ovarian stroma (Fathalla, 1971). Subsequently, these
inclusion cysts are subjected to the stimulative influence of stromal growth factors and were
proposed to undergo metaplasia, dysplasia and eventually, malignant transformation (Lukanova
and Kaaks, 2005; Risch, 1998). Supporting this, several studies reported focal p53 immuno-
positive inclusion cysts in the ovaries of women with serous carcinomas, suggesting an origin
from these sites (Folkins et al., 2008; Jarboe et al., 2008). Moreover, inactivation of Brca1 in
mouse OSE results in pre-neoplastic changes that resemble pre-malignant lesions (Clark-
Knowles et al., 2007).
More recently, an alternative hypothesis was proposed, which suggests that ‘ovarian’ HGSCs are
actually derived from the epithelium of the distal, fimbrial end of the fallopian tube (Crum et al.,
2007a; Kindelberger et al., 2007; Lee et al., 2006; Medeiros et al., 2006). This concept emerged
14
from studies identifying dysplastic lesions and in situ carcinomas in prophylactic salpingo-
oophorectomy (SO) specimens removed from high risk women (Auersperg, 2013b).
As described in Section 1.1, the fallopian tubes are paired organs derived from the Müllerian
ducts. Their main role is to transport the oocyte from the ovary for subsequent fertilization. The
most distal ends of the FT, termed fimbriae, are finger-like projections that sweep the released
ovum into the tube. The FT comprises two main epithelial cells: secretory FTE cells, which line
the tube and support oocyte survival; and ciliated cells, which facilitate transport of the oocyte
through their sweeping action (Gordts et al., 1998).
The FT model of ‘ovarian’ HGSC pathogenesis describes a step-wise process. First, FT secretory
cells are exposed to genotoxic injury, leading to TP53 mutations. Next, clonal expansion of
TP53-mutated cells results in a ‘p53 signature’ (the ‘p53 signature’ describes 12 consecutive
nuclei with strong TP53 immunopositivity in an otherwise benign-appearing FTE). A subset of
cells with the p53 signature might undergo cellular proliferation and result in fully developed
serous tubal intraepithelial carcinomas (STICs). These are characterized by frequent TP53
mutations and increased FT secretory cell proliferation (Bowtell, 2010). STICs have the
capability of either expanding locally or metastasizing to the ovarian surface or peritoneum as a
HGSC (Crum et al., 2007b). Extensive examination of fimbria in patients with ‘ovarian’ HGSC
has revealed the presence of STICs in 40-60% of FT (Kindelberger et al., 2007). Prophylactic
SOs for high-risk women have identified STICs in 5-10% of cases, with no analogous lesions
found in the ovaries of the same women (Hunn and Rodriguez, 2012).
Two additional pre-malignant lesions have also been identified. Secretory cell outgrowth
(SCOUT) is defined as a discrete expansion of 30 secretory cells without the involvement of
TP53. Tubal intraepithelial lesions in transition (TILTs) have nuclear atypia and p53 positivity
but lack the proliferation rate seen in STICs, and may represent the precursor between the
development of p53 signature and STICs. Knowledge of lesions on the SCOUT-STIC
continuum, and how each participates in carcinogenesis is limited (Delair and Soslow, 2012).
Debate surrounding the ovarian HGSC cell-of-origin continues. Those supporting the FTE-cell-
of-origin model argue that prophylactic SOs have failed to reveal a pre-malignant ovarian
pathology (Tone et al., 2012a). Others maintain that there is histologic evidence of early lesions
15
in OSE-lined inclusion cysts, with ‘p53 signatures’ analogous to those found in STICs
(Auersperg, 2013a). While arguments in favour of ovarian HGSC originating in the fallopian
tube are compelling, the possibility that the OSE and peritoneum are additional and significant
sources of these carcinomas remains (Auersperg, 2013b; Delair and Soslow, 2012).
Recently, a stem cell niche has been identified in the hilum region of the mouse ovary, in the
junction between the OSE and FTE. This niche contains cells responsible for OSE regeneration
that are prone to malignant transformation (Flesken-Nikitin et al., 2013). It remains unclear if
this niche contributes to ovarian carcinogenesis.
1.6.3 Ovarian Cancer Etiology: Incessant Ovulation
Historically, incessant ovulation has been proposed as a major contributor to the development of
EOC. Rupture of the ovulating follicle damages the OSE, requiring immediate repair, and
frequent ovulation causes continuous damage to the OSE, increasing the likelihood of mutations
during the repair process (Fathalla, 1971). This model is supported by evidence of a decrease in
EOC risk in women with suppressed ovulation as result of oral contraceptive use, pregnancy or
lactation (Section 1.5.3). A positive association between lifetime ovulatory years and risk of
ovarian cancer in pre-menopausal women has been consistently demonstrated (Tung et al.,
2005). Furthermore, EOC is rare in non-primate mammals, and this is thought to be because
humans are one of the few species that cycle continuously. The only other non-primates to
develop EOC are hens, especially those that have been hyper-ovulated to produce eggs
(Fredrickson, 1987; Land, 1993).
The ovulatory process causes a pro-inflammatory state, characterized by infiltration of
leukocytes, ROS generation and production of inflammatory cytokines (Bonello et al., 1996;
Rizzo et al., 2012). The OSE proliferates rapidly in the presence of cytokines and oxidative
stress resulting from ovulation. King et al. showed that oxidative stress causes normal mouse
OSE to undergo transformative changes (King et al., 2013). DNA damage caused by ovulation-
inducted inflammatory factors could be an important element in the transformation of damaged
OSE/FTE (Fleming et al., 2006). This possibility is supported by the decreased risk of EOC
associated with the use of non-steroidal anti-inflammatory drugs (NSAIDs) (Wernli et al., 2008).
16
1.6.4 Ovarian Cancer Etiology: The Role of Hormones
The OSE and FTE are responsive to the local hormonal environment (Risch, 1998). Multiple
studies have shown that progesterone protects against ovarian cancer development, whereas
androgens, estrogens and gonadotropins increase ovarian cancer risk (Ho, 2003; Lukanova and
Kaaks, 2005; Risch, 1998). Cramer and Welch suggested that EOC was the result of excessive
stimulation of the OSE by LH and FSH. They suggested that gonadotropins either directly
activated gonadotropin-responsive genes, contributing to malignant transformation, or indirectly
stimulated ovarian steroidogenesis (Cramer and Welch, 1983). Several observations argue
against this proposed mechanism. First, estrogen levels are highest during pregnancy, a
protective event for EOC (Risch, 1998). Second, one study found no estrogen receptors on the
OSE or in inclusion cysts (Zeimet et al., 1994). Third, women who develop EOC do not have
elevated serum gonadotropin levels (Helzlsouer et al., 1995). Conversely, progesterone is
believed to be protective against EOC. This association has been seen in multiple births (twins,
triplets, etc), which have significantly higher circulating progesterone levels and exert a higher
protective effect against EOC compared to singleton pregnancies (Whiteman et al., 2000).
1.7 Screening
The small proportion of patients diagnosed with stage 1 EOC have a 5-year survival in excess of
90% (Jacobs and Menon, 2004). However, because the large majority of patients are diagnosed
with advanced (> stage 3) disease, numerous efforts have been made to develop screening tools
utilizing imaging techniques and/or serum markers. A screening test for a disease with low
incidence, like ovarian cancer, must have good sensitivity in addition to very high specificity in
order to achieve an acceptable positive predictive value (PPV) (Skates et al., 1995). The
consequence of a false positive screening test in this case is surgical intervention with its
inherent risks. Numerous studies have looked at screening with CA125, with and without trans-
vaginal ultrasound (TVUS) (Jacobs et al., 1993; Menon et al., 1999). Unfortunately, elevated
CA125 (> 35 U/mL) is not specific for EOC and can be found in a variety of other
gynaecological conditions, including endometriosis, pregnancy, adenomyosis and polycystic
17
ovarian syndrome. In addition, CA-125 levels are typically low in early stage ovarian cancer
(Cannistra, 2004). Multiple other biomarkers, including prostasin, osteopontin, inhibin and
kallikrein have been proposed to have a role in EOC, but their utility as screening tools has not
been established (Jacobs and Menon, 2004).
The Prostate, Lung, Colon and Ovary trial looked at combining CA125 with TVUS, but found a
PPV for invasive cancer of only 23% if both tests were used in combination (Buys et al., 2005).
Most recently, Lu et al. evaluated a 2-stage ovarian cancer screening strategy, which
incorporates age and changes in CA125 over time. Women with high risk scores based on
CA125 changes were then referred for TVUS and to a gynecologic oncologist for further
evaluation. With this model, they were able to achieve a PPV of 40% and all cancers were
discovered at an early stage (Lu, 2013). Whether this will have an impact on survival remains
unanswered.
1.8 Presentation
Symptoms of ovarian cancer are related to patterns of disease spread. EOC disseminates
throughout the abdominal cavity, forming nodules on the surface of the visceral and parietal
peritoneum and omentum. EOC can also travel through lymphatics and blood vessels to nodes
and parenchyma of distant organs, most commonly, the liver and spleen. Blockage of
diaphragmatic lymphatics prevents outflow of proteinaceous fluid from the peritoneal cavity,
causing the accumulation of ascites fluid in advanced disease (Romero and Bast, 2012).
The clinical presentation of EOC may be either acute or subacute, but typically is related to
advanced stage disease, as over 80% of women have stage 3 or greater disease at diagnosis
(Clegg et al., 2002; Institute, 2013). Diagnoses based on subacute presentations are typically
made due to the incidental finding of an adnexal mass either on physical examination or by
imaging. Symptoms associated with subacute presentations are generally vague, consisting of
bloating, urinary urgency/frequency, pelvic/abdominal pain or early satiety, and symptoms do
not typically correlate with stage of disease. Conversely, acute presentations are related to
18
symptoms of pleural effusion (shortness of breath), bowel obstruction (obstipation, constipation)
or venous thromboembolism (leg swelling, leg pain, shortness of breath).
1.9 Diagnosis
Given the frequent absence of symptoms or the presence of vague complaints, clinicians must
have a high clinical suspicion to investigate and diagnose ovarian cancer. A thorough history and
physical are required, including pelvic and rectovaginal examinations. A palpable pelvic mass,
especially in the context of a post-menopausal female, must be carefully evaluated. The
definitive diagnosis requires a tissue diagnosis, obtained either at the time of surgery, from
paracenthesis/thoracenthesis or via an imaging-guided biopsy; however, radiologic and serum
studies assist in confirming the diagnosis.
1.9.1 Cancer Antigen 125
Cancer Antigen 125 (a.k.a. Mucin 16) is a serum biomarker that is elevated in greater than 80%
of EOC patients and plays an important role in detection and disease management (Duffy et al.,
2005; Hogdall et al., 2007). With sensitivity and specificity of ~70% and ~85%, respectively,
serum CA125 is used to assess the risk of malignancy in women with an adnexal mass (Nolen et
al., 2010; O'Connell et al., 1987; Partheen et al., 2011). Once diagnosed with EOC, CA125
measurements are performed routinely and are used to monitor response to treatment,
progression and recurrence (Aggarwal and Kehoe, 2010; Diaz-Padilla et al., 2012).
CA125 is a trans-membrane glycoprotein that is shed from ovarian cancers and circulates in
serum (Romero and Bast, 2012). Physiologically, it is expressed in a number of tissues derived
from coelomic epithelium, including trachea and pericardium, where it functions to provide a
lubricating barrier at mucosal surfaces (Hori et al., 2008). In ovarian cancer, CA125 is thought to
have a role in tumour cell growth, tumourigenesis, cell adhesion and metastases (Rump et al.,
2004; Theriault et al., 2011). Furthermore, in vitro, it has been implicated in sensitivity of
ovarian cancer to chemotherapy, with higher levels suggestive of platinum resistance (Boivin et
19
al., 2009) (Levanon et al., 2008). Persistent elevation of CA125 after chemotherapy indicates
residual disease with greater than 90% accuracy (Romero and Bast, 2012).
1.9.2 Radiologic Studies
Imaging studies can help to assess the presence of ascites and the extent of disease. Abdominal
and pelvic computerized tomography (CT) and transvaginal US are the main imaging modalities.
Several characteristics found on TVUS are highly suggestive of malignancy, these include: size,
morphology (septation, papillation, solid components), and Doppler flow within the mass. These
characteristics provide 90% and 88% sensitivity and specificity, respectively, when combined
with other clinical variables, such as age (Twickler and Moschos, 2010).
Computed tomography (CT) is the preferred technique for pre-surgical evaluation of patients, as
it defines the extent of disease and likelihood of successful surgical cytoreduction (Axtell et al.,
2007). On CT, advanced ovarian cancer typically appears as thick-walled, septated cysts, with
papillary projections. These are optimally visualized after administration of CT contrast dye.
Additional CT findings often include pelvic organ involvement, sidewall or peritoneal implants,
adenopathy and ascites (Iyer and Lee, 2010). Most recently, positron emission tomography
(PET) imaging has been combined with CT imaging. This combination provides the most
accurate radiologic evaluation of suspected recurrence, with a recent meta-analysis showing that
it has a sensitivity and specificity of 91% and 88%, respectively (Gu et al., 2009).
Magnetic Resonance Imaging (MRI) is used occasionally, particularly when the patient has a CT
contrast dye allergy. MRI is not as sensitive as Doppler TVUS for identifying malignant lesions
(TVUS 100%, MRI 96.%) but it is more specific (TVUS 40%, MRI 84%) (Iyer and Lee, 2010).
Consequently, MRI is typically reserved for patients with a low risk of malignancy, but an
indeterminate lesion on TVUS.
20
1.10 Management of Disease
1.10.1 Management of Primary Disease: Surgery
Surgical management is the mainstay of treatment for ovarian cancer. A surgical procedure is
necessary to obtain tissue for diagnosis, assess extent of disease and attempt optimal
cytoreduction (Mann, 2013). The standard surgical procedure includes exploratory laparotomy,
total abdominal hysterectomy (TAH), bilateral salpingo-ooporectomy (BSO), pelvic and peri-
aortic lymphadenectomy, omentectomy and multiple peritoneal biopsies. Confirming the
diagnosis is critical, as several tumours can mimic ovarian cancer, including gastrointestinal tract
tumours (with Krukenberg ovarian metastases) and breast cancer with ovarian metastases.
Extent of disease (surgical staging) determines initial treatment. Less than 25% of patients with
EOC present at early stage disease (stages 1 or 2) (Table 1.0), and their management differs from
patients with advanced (stage ≥3) disease (Young et al., 1983; Young et al., 1990). The
ACTION trial showed that after a median follow-up of 10.1 years, there were no differences
between the observation and the adjuvant chemotherapy arms in cancer-specific survival if early
stage patients were optimally debulked. However, when non-optimally debulked patients were
analyzed, cancer-specific survival was better in the adjuvant chemotherapy arm than in the
observation arm (Trimbos et al., 2010). This study examined all EOC patients and did not look
specifically at HGSC.
Unfortunately, the majority of EOC patients present with advanced stage disease. The standard
of care for these patients consists of a combination of aggressive debulking surgery followed by
platinum-based chemotherapy. Multiple studies have shown that minimum residual tumour post-
operatively is one of the most powerful determinants of survival (Allen et al., 1995; Bristow et
al., 2002; Hoskins et al., 1994). Residual tumour size greater than 2 cm is associated with a
survival of 12–16 months, compared with 40–45 months if the tumour is less than 2 cm (Mutch,
2002). Bristow et al. showed that every 10% increase in cytoreduction was associated with a
5.5% increase in median survival (Bristow et al., 2002). The goal of every surgery is for ‘optimal
cytoreduction’; however, that definition remains debatable. The most widely used definition for
‘optimal’ is residual disease less than 1 cm (Del Campo et al., 1994; Eisenkop and Spirtos,
21
2001). In addition to increasing survival, there are other benefits of surgery, including decreasing
disease-related symptoms, improving host immune competence and maximizing the effect of
chemotherapy (Covens, 2000; Merogi et al., 1997). Patients who are not good surgical
candidates due to medical comorbidities, can be considered for neoadjuvant chemotherapy
(NACT) with interval debulking surgery (IDS); however, whether this is a reasonable alternative
for all patients is currently being investigated (Morrison et al., 2012).
Table 1.0 Ovarian Cancer Staging (Adapted from Mann, 2013)
Stage Definition:
1a Tumour limited to ovary; negative washings
1b Tumour limited to both ovaries; negative washings
1c Tumour limited to both ovaries with either tumour on the surface of the ovary or positive washings
2a Implants on uterus and/or tube(s); negative washings.
2b Implant on other pelvic tissues; negative washings.
2c Pelvic extension and/or implants; positive washings.
3a Microscopic peritoneal metastasis beyond pelvis (no macroscopic tumour)
3b Microscopic peritoneal metastasis beyond pelvis ≤2 cm
3c Peritoneal metastasis beyond pelvis > 2 cm and/or regional lymph node metastasis
4 Distant metastasis (excludes peritoneal metastasis)
1.10.2 Management of Primary Disease: Chemotherapy
Cisplatin has been used since the 1970s for the treatment of various cancers, including ovarian,
testicular, head and neck and cervical. Its second-generation analog, Carboplatin, forms identical
lesions on DNA but has a notably lower spectrum of toxicities (Rabik and Dolan, 2007). Unlike
cisplatin, Carboplatin rarely results in nephrotoxicity, ototoxicity or peripheral neuropathy, but
can cause myelosuppression, particularly when administered in high doses (Rabik and Dolan,
22
2007). A large meta-analysis showed that patients with advanced ovarian cancer had virtually
identical survival durations when treated with Carboplatin or cisplatin-containing regimens, and
as such, Carboplatin use is preferred (Alberts, 1995).
Upon entering the cell, all platinating agents are aquated and become positively charged. This
highly reactive species is then capable of interacting with nucleophilic molecules within the cell,
such as DNA and RNA (Siddik, 2003). When binding to DNA, platinating agents favor the
imidazole rings of guanosine and adenosine and can form three different lesions: mono-adducts,
intra-strand crosslinks, and inter-strand crosslinks (Rabik and Dolan, 2007).
The most abundant adducts are intrastrand crosslinks between adjacent purines (Pt-GG and Pt-
AG), representing 65% and 25%, respectively, of total adducts formed (Crul et al., 2003). The
majority of monoadducts also react to form crosslinks, and these all cause steric changes in the
DNA. The physical distortions in the DNA are then recognized by over 20 individual proteins,
which transduce DNA damage signals to downstream effectors. The downstream effects
activated by cisplatin can be pro-survival or pro-apoptotic, and the relative intensity and/or
duration of each signal determines the final fate of the cell (Siddik, 2003). Intra-strand crosslinks
are believed to be the most cytotoxic because of their ability to recruit high mobility group
(HMG) proteins, which in turn recognize and bind DNA and prevent replicative bypass
(Vaisman et al., 1999). Increased HMGB1 expression is associated with platinum sensitivity (He
et al., 2000; Reeves and Adair, 2005).
Activation of cell cycle checkpoints occurs with platinum-induced DNA damage. One of these is
a transient S-phase arrest, followed by inhibition of Cdc2-cyclin A or B kinase to cause a long-
lasting G2/M arrest (He et al., 2005). Cell-cycle arrest is necessary to enable the nucleotide
excision repair (NER) complex to remove adducts and promote cell survival. However, DNA
damage exceeding a certain threshold results in activation of the caspase 9 - caspase 3 pathway
and causes cell death (Siddik, 2003).
Inter-strand crosslinks represent 5-10% of the total number of Pt-DNA adducts. DSBs resulting
from platinum-induced inter-strand adducts are thought to be repaired predominantly by
homologous recombination, which explains the increased drug sensitivity seen in cells lacking a
functional HR pathway (Crul et al., 2003). As such, women with a germline mutation in BRCA1
23
or BRCA2 show higher response rates to platinum-based chemotherapy and therefore, a longer
progression-free survival and improved overall survival (Berns and Bowtell, 2012).
Paclitaxel is the standard taxane used to treat ovarian HGSC. Taxanes are microtubule-disrupting
agents (MDA), whose cytotoxicity is not the result of any one single parameter, but rather a
combination of effects on the cell cycle and apoptosis. Paclitaxel works mainly by binding to
intracellular β-tubulin, which leads to microtubule stabilization, G2/M arrest and apoptosis
(Wang et al., 1999).
The preferred route of administration and timing of chemotherapy has recently come into
question. Standard intravenous (IV) treatment of Carboplatin (AUC 4-6) with paclitaxel (175
mg/m2) is given every 21 days for six cycles (Hennessy et al., 2009). Treatment for longer
durations has been evaluated and no clear benefit has been shown by extending beyond six
cycles (Muggia, 2011). For optimally debulked patients, intraperitoneal (IP) chemotherapy has
been shown to have an advantage over standard IV chemotherapy. GOG 172 compared an IV
regimen (IV paclitaxel and IV cisplatin) to IV and IP (IV paclitaxel, IP cisplatin, IP paclitaxel)
regimen: the IV and IP regimen showed significant benefit in both overall and progression free
survival, but had significantly higher toxicity (Armstrong et al., 2006). With standard-of-care
chemotherapy, more than 80% of patients have a significant reduction in tumour size, with 40-
60% of patients showing complete response.
1.10.3 Management of Recurrent Disease
Overall survival (OS) is measured as the time from treatment to death from any cause, and
progression free survival (PFS) is measured as the time from treatment to documented disease
progression. In ovarian cancer, PFS correlates well with OS and is often used as a surrogate to
measure clinical benefit (Bast et al., 2007). Eighty-five percent of women with stage 3 or greater
disease at presentation will recur, and 20-30% will relapse within 6 months (Berns and Bowtell,
2012). Due to these recurrence rates, the median PFS is only 18 months (Greenlee et al., 2001).
PFS is most strongly linked to the length of the platinum-free interval (PFI) as well as debulking
status (Huang et al., 2012; Rump et al., 2004; Theriault et al., 2011). The PFI is defined as the
time from first-line platinum chemotherapy to relapse. The longer this interval, the better the
response to subsequent chemotherapy and length of survival (Huang et al., 2012).
24
Patients who recur within 6 months of completing platinum therapy are classified as platinum-
resistant, and if disease progression occurs during the course of platinum therapy, these cases are
classified as platinum-refractory (Yap et al., 2009). Primary platinum-resistance occurs in 20-
40% of patients. Re-treatment with platinum in these instances yields a response rate of only 10-
20% and a response rate to other chemotherapeutic agents of only 10-25% (Agarwal and Kaye,
2003). Platinum-resistant tumours present a therapeutic dilemma, and thus, the focus with these
patients tends to be palliative rather than curative (Muggia, 2011). Second-line therapy is not
well established but typically consists of doxorubicin (caelyx), etoposide or topotecan. Given the
low success rates of these agents, patients are typically enrolled into clinical trials for multi-drug
treatment.
Women with a PFI of greater than 6 months can expect to respond to further chemotherapy. For
example, a PFI of 12 months yields a platinum retreatment response rate of 27%; conversely, a
PFI of 24 months gives a 72% response rate (Parmar et al., 2003). The term ‘platinum-sensitive’
is often given to women who recur ≥6 months after completing platinum therapy. Several trials
have looked at platinum-sensitive recurrent disease in an effort to maximize platinum therapy.
Most recently, the CALYPSO trial studied the combination of Carboplatin with either pegylated
liposomal doxorubicin (PLD) or paclitaxel. The Carboplatin/PLD arm had a longer PFS and a
better toxicity profile (Pujade-Lauraine et al., 2010). Re-treatment with platinum or platinum
analogs remains the preferred treatment choice, but carries a risk of hypersensitivity, which can
be fatal. In recurrent platinum-sensitive disease, re-treatment with non-platinum agents also
provides benefit (Muggia, 2011).
Secondary surgery in recurrent disease has a role in women with platinum-sensitive recurrence,
who have a single recurrent lesion and a prolonged PFI (Hennessy et al., 2009). In the setting of
platinum-resistant disease, secondary surgery is often palliative.
To monitor for recurrence, the National Comprehensive Cancer Network has recommended
serum CA125 assessment every 3-4 months (Mann, 2013). A rise in serum CA125 predicts
clinical relapse within 2-6 months; however, treatment on the sole basis of elevated CA125 has
not shown a survival benefit (Rustin, 2010; Rustin et al., 2010).
25
1.11 Mechanisms of Therapy Resistance
As most patients relapse, re-treatment options are important. Resistance to platinum-based
chemotherapy limits the efficacy and use of these compounds, and is believed to cause treatment
failure and death in more than 90% of patients (Agarwal and Kaye, 2003). Chemotherapy failure
can be classified into three categories: pharmacokinetic, tumour micro-environment dependent
and cancer cell-specific. Pharmacokinetic resistance is a function of drug concentration and time
of exposure. Drug resistance can result from inadequate intra-tumour drug concentration due to
patient variables, such as first-pass metabolism, renal clearance, hepatic drug metabolism and
tumour vascularity. The tumour micro-environment triggers signaling pathways in response to
various stressors. For example, cellular attachment has been shown to affect chemosensitivity,
with cell monolayers displaying different sensitivities to drugs compared to cells grown as
spheroids (Francia et al., 2005). Furthermore, hypoxia can induce chemo-resistance. Using an
ovarian cancer cell line, Tomida and Tsuruo found that drug resistance was induced in all cell
lines under stress conditions (hypoxia, glucose deprivation), and that this resistance was
reversible once the stress conditions were removed (Tomida and Tsuruo, 1999). Cancer cell-
specific resistance pertains to acquired somatic mutations and epigenetic changes within tumour
cells as they evolve. Bernardini et al., found that resistant tumours were found to have twice as
much allelic imbalance as sensitive tumours (Bernardini et al., 2005). Finally, cell specific drug
resistance can develop as a result of decreased influx or increased efflux of drug, and/or
increased drug detoxification (Rabik and Dolan, 2007). Furthermore, certain cancer cells (e.g.,
tumor-initiating cells) might be intrinsically more resistant to chemotherapy, as described in
Section 1.12.
1.12 Tumour Heterogeneity
Solid tumours display intra-tumour heterogeneity. Two models have been proposed to explain
this heterogeneity: the stochastic model and the hierarchical model. The stochastic model holds
that all malignant cells within a tumour are equally capable of generating the tumour.
Conversely, the hierarchical, ‘cancer stem cell’ model suggests that only a subset of tumour cells
26
have this capacity (Figure 1.2). These ‘cancer stem cells’ (CSCs) are defined by their distinct
gene expression, which confers the ability to self-renew, recapitulate the original tumour,
proliferate and promote recurrence (Alvero et al., 2009a; O'Brien et al., 2009a). Given these
properties, it should be possible to prospectively isolate and purify this population of tumour
cells. Tumour-initiating cells (TICs), often used as a surrogate for CSCs, are defined
operationally by their capacity to generate tumours in immunocompromised mice.
Figure 1.2 A) Stochastic and B) hierarchical models of tumour heterogeneity.
Multiple studies have shown that within a single tumour, tremendous genetic heterogeneity
exists (Anderson et al., 2011; Gerlinger et al., 2012; Notta et al., 2011; Shah et al., 2012). This
heterogeneity is further seen in primary versus metastatic tumours, at presentation versus
recurrence and even in tumours taken at the same time from different anatomic sites (Campbell
et al., 2010; Ding et al., 2012; Gerlinger et al., 2012; Yachida et al., 2010). Because the majority
of patients with ovarian HGSC have resistant and/or recurrent disease, interest into whether
CSCs are responsible for drug-resistance in ovarian cancer has grown. In this scenario, although
bulk tumour shrinkage occurs with chemotherapeutic treatment, recurrence and metastasis arise
from failure of chemotherapy to eradicate CSCs. According to the CSC model, treatment with
chemotherapy allows reservoir of CSCs to persist and subsequently reproduce the tumour
phenotype (O'Brien et al., 2009a; Zhang et al., 2008). Consequently, eradicating the tumour
would require specifically targeting the CSC population.
27
Current chemotherapeutics are thought to be unable to target CSCs; therefore, treatment with
chemotherapy is thought to actually enrich for tumourigenic cells. This has been reported in
several cancers, including pancreatic, breast and brain cancer models (Bao et al., 2006; Hermann
et al., 2007; Li et al., 2008). Dylla et al. showed that treatment of a colon cancer xenograft model
with standard chemotherapy not only enriched for a CSC-phenotype in the residual tumour, but
also increased tumourigenic cell frequency (Dylla et al., 2008).
In ovarian cancer, several studies have attempted to isolate ‘CSCs’, but most have utilized
cultured primary cells or immortalized cell lines. The first report of ‘CSCs’ was by Bapat et al.
who isolated a single tumourigenic clone from a mixed population of cells derived from the
ascites (Bapat et al., 2005). Next, Alvero et al. reported the molecular characterization of ovarian
cancer stem cells as CD44+. Again, this was primarily done in vitro with insufficient in vivo
analysis (Alvero et al., 2009a). Subsequently, using more robust assays, it was found that both
the CD44+ and CD44- fractions of primary ovarian HGSC are tumourigenic in
immunocompromised mice (Stewart, 2013).
CD133 has been studied as a marker of the CSC population in a variety of solid cancers,
including brain, colon, pancreas and lung (Dalerba et al., 2007; Eramo et al., 2008; Hermann et
al., 2007; O'Brien et al., 2007; Ricci-Vitiani et al., 2007; Singh et al., 2004). Stewart et al.,
found that ovarian HGSC conforms to the CSC model, but that there are at least two (CD133+
and CD133-), and possibly three (CD133+, CD133-, CD133+/-) CSC phenotypes in this disease
(Stewart et al., 2011). This was performed using limiting dilution assays (LDA), the gold
standard method for assessing CSC.
Although it is not within the scope of this thesis to debate the validity of the CSC model, it
should be mentioned that several strong criticisms against the model have been made. Most
notably, Kern and Shibata argue that the “mathematical support for the concept of therapeutically
useful stem cells is weak” (Kern and Shibata, 2007).
28
1.13 Current Models of Ovarian Cancer
1.13.1 In Vitro Models of Ovarian Cancer
To unravel the key determinants in ovarian cancer initiation, progression and response to
treatment, a model that faithfully recapitulates ovarian HGSC is necessary. Most preferable
would be in vitro and in vivo models that reflect the cell biology and heterogeneity of the disease.
Most studies of ovarian cancer have utilized immortalized cancer cell lines as in vitro tumour
models. There are several problems with the current in vitro models. First, cancer lines are
derived from cancer cells adapted to grow outside a normal tumour microenvironment, resulting
in changes that are distinct from the genetic stress imposed on tumours in patients (Tentler et al.,
2012). Moreover, many of these cell lines are unlikely of high-grade serous origin, and as such,
are badly suited for investigating ovarian HGSC. This is supported by the work of Domcke et al.,
who found that the most frequently cited ‘high-grade’ ovarian cancer cell lines (SKOV3, A2780,
OVCAR-3, CAOV3 and IGROV1), have genomic profiles that do not reflect those of the
primary tumour (Domcke et al., 2013). The authors argue that using cell lines with genomic
backgrounds similar to those in patient samples could at least increase the likelihood that
conclusions reached in the in vitro setting be transferable to the clinical setting. Furthermore,
many of these ‘ovarian HGSC’ cell lines either do not reproduce serous histology when
implanted into immune-compromised mice or do not give rise to xenografts at all (Press et al.,
2008b; Vaughan et al., 2011). Of the cell lines that do generate the appropriate histology in vivo,
their capacity for predicting drug response is severely lacking (Jin et al., 2010). There are a
number of problems inherent to ‘predictive’ in vitro assays, making the relationship between
tumour inhibition in vitro and a patient’s response complex (Fiebig et al., 2004).
Another in vitro method, developed with the aim of predicting drug response, is the ATP-based
tumour chemosensitivity assay (ATP-TCA). This assay measures the viability of dissociated
tumour cells, cultured in serum-free medium, against a panel of single agents or drug
combinations. Although this method indicates platinum resistance with reasonable accuracy
(~90%), this only applies to primary and not recurrent cases (Andreotti et al., 1995; Kurbacher et
al., 1995). Because no reliable data have shown this assay strategy to be reliably, clinically
useful, it has not been incorporated into clinical practice (Markman, 2011).
29
1.13.2 In Vivo Models of Ovarian Cancer
Few animal models develop spontaneous ovarian tumours, with the exception of hens. The
laying hen provides insight into the role of incessant ovulation in the development of
spontaneous ovarian adenocarcinomas (Ricci et al., 2013; Vanderhyden et al., 2003). Forty
percent of hens will develop ovarian tumours by age 4, with certain mutations analogous to
human ovarian HGSC (Mullany and Richards, 2012). Although the hen provides a good model
of determining disease initiation, progression and drug testing, the cost makes these models
fairly prohibitive. Also, a challenge of working with hens is the limited knowledge of their
molecular biology and difficulty with genetic manipulation (Lengyel et al., 2013).
1.13.2.1 Genetically Engineered Mouse Models
Mice are the primary mammalian model in cancer research due to their short generation time,
ability to bear large litters, relative ease of breeding, and advances in mouse genomics (Edinger
et al., 2002). Genetically engineered mouse models (GEMMs) were developed 20 years ago in
order to provide an efficient system in which to analyze the specific roles of oncogenes and
tumour suppressor genes in tumourigenesis in vivo (Mullany and Richards, 2012; Talmadge et
al., 2007). In GEMMs, the genetic profile of the mice is altered such that one or several genes
thought to be involved in transformation or malignancy are mutated, deleted or overexpressed
(Richmond and Su, 2008). Moreover, using Cre-lox conditional targeting, these genes can be
altered in a precise, tissue-specific manner (Garson et al., 2012).
Using this technology, two studies showed that concomitant inactivation of Brca1 and Trp53 or
Rb1 and Trp53 in mouse OSE can give rise to serous carcinomas; however, conflicting results
were found by other studies, which showed that specific conditional inactivation of Brca1 and
Trp53, in the presence or absence of Rb, resulted in ovarian leiomyosarcomas (Clark-Knowles et
al., 2009; Flesken-Nikitin et al., 2003; Quinn et al., 2009; Xing and Orsulic, 2006). Most
recently, Szabova et al., found that inactivation of Rb-mediated tumour suppression induced
surface epithelial proliferation and additional bi-allelic inactivation and/or missense p53
mutation in the presence or absence of Brca1/2 caused progression to stage IV disease. As in
human ovarian HGSC, these mice developed peritoneal carcinomatosis, ascites, and distant
metastases (Szabova et al., 2012).
30
Despite these impressive advances, GEMMs have not been shown to be superior to xenograft
models in a variety of tumour types and have not shown to predict clinical trial results (Francia
and Kerbel, 2010). Overall, GEMMs have seldom been used to test novel anti-cancer drugs and
the few studies that have used GEMMs against clinically effective agents have not been
encouraging. As such, these models have not yet demonstrated a role in drug discovery
(Talmadge et al., 2007).
1.13.2.2 Xenografts
The ability to successfully engraft surgically derived tumours from cancer patients has been
established for decades. A commonly used strain for maximizing engraftment is the non-obese
diabetic severe combined immunodeficiency (NOD-scid) mouse. Non-Obese Diabetic (NOD) is
an inbred mouse strain, which has low Natural Killer (NK) cell activity, and lacks circulating
complement and proper function of antigen-presenting cells (Bastide et al., 2002). The Scid
mutation is a loss-of-function allele of the Prkdc gene. Prkdc encodes the catalytic subunit of
DNA-dependent protein kinase, which has a role in resolving DNA DSBs that occur during
variable, diverse and joining [V(D)J] recombination. In the absence of V(D)J recombination,
mice lack T- and B-lymphocytes but retain innate immune functions (Laboratory, 2013).
Together, the NOD-Scid model combines multiple functional defects in adaptive and innate
immunity to maximize engraftment efficiency. Unfortunately, NOD-Scid mice are limited by
their relatively short life span (due to lymphoma), residual NK cell activity, and other
components of innate immunity, which can impede engraftment (Shultz et al., 2007).
To mitigate these problems, an immunodeficient mouse homozygous for targeted mutations at
the interleukin-2 receptor (IL-2R) γ-chain locus was generated. These mice have increased
engraftment of human tissue compared with all previously developed immunodeficient mouse
models (Ito et al., 2002). The IL-2R γ-chain is a crucial component of the receptors for IL-2, IL-
4, IL-7, IL-9, IL-15 and IL-21, required for signaling by these receptors. The absence of the IL-
2R γ-chain blocks NK cell development and results in additional defects in innate immunity
(Shultz et al., 2007; Shultz et al., 2005). NOD-Scid-IL2Rγnull (NSG) mice lack mature
lymphocytes and NK cells, have severe defects in innate immunity, do not develop thymic
lymphomas (unlike NOD-Scid), and engraft human tissues at high levels (Shultz et al., 2005).
31
Although their lack of immune response translates into high engraftment rates, it also means that
aspects of the host-tumour interaction cannot be studied using xenograft models.
Typically, the process of generating a patient-derived xenograft (PDX) consists of obtaining
fresh surgical tissue, sectioning it into small pieces, and implanting these pieces into an
immunodeficient mouse (Tentler et al., 2012). One of the main advantages of PDX models is
maintenance of the original tumour architecture and histological characteristics (Gray et al.,
2004). Unlike xenografts using tumour cell lines, the PDX-process of engraftment and expansion
appears to maintain the majority of the key genes and global pathway activity in tumours. This
has been shown in lung, colorectal, and pancreatic cancer PDX models, where a high degree of
genetic similarity between the primary cancer and the corresponding PDX tumour was observed
(Daniel et al., 2009; Fichtner et al., 2008; Jones et al., 2008; Tentler et al., 2012). In addition to
recapitulating the original tumour biology, there are several other key advantages of using PDX:
results can be obtained in a few months (compared to a GEMM, which often require as long as a
year to develop), and multiple therapies can be tested in a single tumour sample (Richmond and
Su, 2008).
For ovarian cancer, injection into either the ovarian bursa (intrabursal; IB) or the FT fimbriae
would constitute an orthotopic PDX. An intraperitoneal (IP) PDX is also considered orthotopic
and is easily generated, but represents advanced metastatic disease. A disadvantage of both IB
and IP models is difficulty of monitoring disease progression. Even with recent advances in
mouse imaging technology, the ability to accurately measure disease has not been optimized
(Ricci et al., 2013). Furthermore, the requirement of an experienced operator combined with low
throughput of the technology (typically allowing only one animal to be imaged at a time)
diminishes its practicality and affordability (Garson et al., 2005; Kung, 2007).
Due to the challenges seen with orthotopic PDXs, the standard PDX is ectopic, with cells
injected typically implanted subcutaneously (SC) or into the mammary fat pad (MFP). This
model allows close monitoring of superficial tumours but the predictive capability of SC PDXs
has been questioned (Kung, 2007). Although SC/MFP PDXs recapitulate many aspects of the
tumour, it is likely that the tumour-host microenvironment is most similar when cells are
implanted into the organs/anatomic sites from which the tumour originally arose.
32
The advantages of MFP over orthotopic PDX were previously demonstrated by the Neel
laboratory, which evaluated the peritoneum, ovarian bursa, kidney capsule and mammary fat pad
of mice as sites for xenotransplantation. They found high tumour takes from each site but
tumours were most readily and reliably detected while still relatively small in the mammary fat
pad. Mammary fat pad xenografts also recapitulated the inter- and intra-tumour heterogeneity of
primary HGSC, as assessed by histology and surface immunophenotype (Stewart, 2013; Stewart
et al., 2011). These results, combined with the impracticality of using orthotopic models,
provided the rationale for using the MFP model in my experiments.
To take advantage of the orthotopic site but to avoid some of the issues seen with ectopic PDXs,
fluorescent or bioluminescent technology has been employed. Bioluminescence imaging (BLI) is
a powerful tool for imaging of small laboratory animals, enabling the study of ongoing biological
processes in vivo. This has been increasingly adopted for tumour burden quantification in mice
(Sadikot and Blackwell, 2005). This technology will be further described in Appendix 1.
Despite these challenges, many researchers continue to utilize ectopic PDX for drug testing.
Several large-scale PDX programs have been established, and the activity of drugs in these
models was scored and compared with subsequent activity in phase II clinical trials. The general
conclusion from some of these trials was that the predictive power of PDX is variable. For
example, PDX derived from breast carcinomas predict poorly whereas PDX derived from lung
tumours have a good predictive value (Johnson et al., 2001). The criticism of those experiments
is that the pharmacology of the therapeutic was poorly understood and not optimized, thus
making the drug dosing schedule in humans unrealistic (Sausville and Burger, 2006). Two trials
showed high predictive results with PDX used agents where the pharmacology was well
understood and the human tumours were directly transplanted from the patient, not first cultured
in vitro (Fiebig et al., 2004; Peterson and Houghton, 2004). Recently, Hidalgo et al. reported a
pilot study using ectopic PDX from patients with advanced cancers, whose treatments were
selected on the basis of activity seen in the PDX. Their findings showed a remarkable correlation
between drug activity in the PDX and clinical outcome (Hidalgo et al., 2011). Ovarian cancers
were not tested in any of these studies, providing a rationale for this project.
33
Several investigators have generated PDX from ovarian cancer. The first were Ward et al. and
Massazza et al., who used fresh primary tumour ‘slurries’ and injected these IP into nude mice
(Massazza et al., 1989; Ward et al., 1987). Schumacher et al. reported that IP injection of
ovarian cancer cells from patients’ ascites into Scid mice resulted in tumour formation
(Schumacher et al., 1997; Schumacher et al., 1996). Xu et al. implanted intact ovarian cancer
samples (all EOC) into the peritoneum of Scid mice (Xu et al., 1999). In recent years, several
groups have used bulk tumour specimens to generate PDX. First, Lee et al. grafted tumour
specimens under the renal capsule of NOD-Scid mice and found that these tumours retained
major histopathological characteristics of the original tissue (Lee et al., 2005). Next, Press et al.
showed that PDX grown under the renal capsule of NOD/SCID mice showed minimal genetic
change from the original tumour based on array comparative genomic hybridization (Press et al.,
2008b). Most recently, Bankert et al. injected tumour cell aggregates IP into NSG mice and
observed tumour progression and metastasis mirroring patient disease (Bankert et al., 2011).
Although all of these authors state that these PDX are valuable for testing novel
chemotherapeutics, no one to date has validated this model for this purpose. Moreover, all
previous methods have utilized tissue fragments, which cannot be used to measure TIC
frequency. Stewart et al. established robust conditions to generate PDXs from ascites and solid
tumours in the MFP of mice, with the capacity to assess tumourigenicity (Stewart et al., 2011).
As such, this protocol was utilized for the project.
1.14 Current State of the Problem
The major problem with treatment of ovarian HGSC is that all patients receive platinum and
Taxol therapy. This ‘one size fits all’ approach is largely to blame for overall survival being
unchanged over the last 30 years. We know that the length of the treatment-free interval (TFI)
positively correlates with overall survival, and this is largely due to platinum-resistant/refractory
tumours having low response rates to re-treatment (Yap et al., 2009). It is my belief that three
major changes can alter the outcomes of patients with ovarian HGSC: improved therapeutics,
early identification of platinum-resistant/refractory patients, and the individualization of cancer
treatment.
34
Our recent expanded understanding of the pathogenesis and evolution of ovarian HGSC has led
to the development of many new molecular targeted therapies, including those targeting
angiogenesis, cell proliferation and metastases. However, individualized patient selection for the
application of these targeted therapies is essential (Yap et al., 2009).
One of the most frequently cited reasons for high failure rates of new agents in oncology clinical
trials is the paucity of pre-clinical models that recapitulate the heterogeneity of patient disease
(Johnson et al., 2001; Tentler et al., 2012). For successful drug development and establishment
of treatment strategies, pre-clinical in vivo models must have a good probability of being
predictive of similar activity in humans. Ovarian HGSC PDXs could potentially resolve some of
these issues, as these tumours have been shown to recapitulate the biological characteristics of
the primary tumour.
1.15 Research Questions & Thesis Overview
New models of ovarian HGSC are required that can ultimately improve management of this
disease. In order to improve survival, these models should identify platinum-resistant/refractory
patients and have the capacity to be used to study new therapeutic cancer treatments. To this end,
this thesis is a culmination of addressing several elements of the problem:
1) Can ovarian HGSC PDXs predict response to platinum therapy? (Chapter 2)
2) Does chemotherapy enrich for TICs in ovarian HGSC PDXs? (Chapter 2)
3) Can I generate luminescent ovarian PDXs to utilize bioluminescent imaging for
disease monitoring? (Appendix 1)
4) Can we use ovarian HGSC PDXs to evaluate response to novel anti-tumour agents?
(Appendix 2)
35
1.16 Research Objectives & Hypotheses
Because our model of ovarian HGSC PDXs recapitulates disease heterogeneity (Stewart, 2013), I
expect that these PDXs also will recapitulate response to chemotherapy. Over the past 6 years,
Dr. Neel’s laboratory has procured a large collection of primary ovarian HGSC samples, with
clinical follow-up. From this collection, I identified sensitive, resistant and prospective tumours
and engrafted these into NSG mice. The PDXs received treatment with Carboplatin and response
was measured. I hypothesized that PDXs derived from platinum-sensitive tumours would
respond to platinum therapy by displaying a reduction in tumour volume. Conversely, I expected
that PDXs derived from platinum-resistant/refractory tumours would show little or no response
when treated with platinum.
Next, I utilized these PDXs to ask whether treatment with chemotherapy enriches for CSCs. The
CSC model predicts that CSC and non-CSC have differences in biological properties and CSCs
might be intrinsically more resistant to conventional chemotherapy than non-CSCs. Each of my
PDXs was treated with Carboplatin or vehicle control, and effects on TICs was assessed by
limiting dilution assays. As reported for other tumour types, I expected TICs would be enriched
following treatment with chemotherapy.
As previously discussed, the orthotopic PDX-model (IB and/or IP) does not allow easy or
practical evaluation of disease burden and response to treatment. For these reasons, an ectopic
(mammary fat pad, MFP) model was used to evaluate chemotherapeutic response in my cohort
of PDXs. However, I also wanted to generate a luminescent ovarian PDX model to monitor
disease by imaging (Appendix 1).
Lastly, I used PDXs to evaluate the effects of drug shown to have anti-tumour effects in other
cancer types. Disulfiram (DSF), a drug commonly used to treat alcohol abuse, works by causing
an accumulation of acetaldehyde, and causing ICLs. By combining this drug with Carboplatin, I
hypothesized that the drug combination would increase the number of DNA adducts formed,
thereby enhancing the toxic effects of platinum (Appendix 2).
36
Chapter 2
Patient-Derived Xenografts as Pre-Clinical Models of Response to Chemotherapy
37
2.1 Introduction
Ovarian cancer is the 6th most common cancer in women and the 7th most common cause of
cancer death worldwide (Organization, 2008). The clinical presentation of ovarian cancer is
typically related to symptoms of advanced stage disease, as over 80% of women have stage 3 or
greater disease at diagnosis (Clegg et al., 2002; Institute, 2013). Epithelial ovarian cancer (EOC)
is the most common form of ovarian cancer, and HGSC is the commonest and most aggressive
subtype. Aggressive surgical cytoreduction and platinum-based chemotherapy are the mainstays
of treatment; however, overall survival for these patients has remained largely unchanged over
the last thirty years (Winter et al., 2007).
Clinical outcomes in ovarian HGSC are most strongly related to the length of the platinum-free
interval (PFI) as well as debulking status (Huang et al., 2012; Rump et al., 2004; Theriault et al.,
2011). Several studies have shown that minimum residual tumour post-operatively is one of the
most powerful determinants of survival (Allen et al., 1995; Bristow et al., 2002; Hoskins et al.,
1994). For example, residual tumour size of greater than 2 cm is associated with a survival of
only 12–16 months, compared with 40–45 months if the tumour is less than 2 cm (Mutch, 2002).
The platinum-free interval is defined as the time from first-line platinum chemotherapy to
relapse; the longer this interval, the better the response rate to subsequent chemotherapy and the
chance for improved survival (Huang et al., 2012).
Unfortunately, over 85% of women with advanced disease at presentation will recur. Thirty-
percent will relapse within 6 months of completing platinum therapy, and such patients are
termed ‘platinum-resistant’ (Yap et al., 2009). These patients will have response rates of only
10-20% to additional platinum therapy, and 10-25% to other chemotherapeutic agents (Agarwal
and Kaye, 2003). Early identification of platinum-resistant patients is critical in order to triage
these patients into clinical trials. Moreover, appropriate and individualized patient selection for
the application of these novel and targeted therapies is critical (Yap et al., 2009).
High failure rates plague oncology clinical trials, and one of the most frequently cited reasons is
the paucity of pre-clinical models that recapitulate the heterogeneity of patient disease (Johnson
et al., 2001; Tentler et al., 2012). According to Cespedes et al., the ideal cancer model should
38
show histologic similarity to the human tumour, progress through the same stages of disease,
involve the same genes and biochemical pathways, closely reflect the response of the human
tumour to a particular therapy and finally, predict therapeutic efficacy in assays (Cespedes et al.,
2006). The creation of such cancer models is essential for advancing translational ovarian cancer
research (Press et al., 2008b).
Traditionally, in vivo experimental approaches have relied on the use of GEMMs or xenografts
established from ovarian cancer cell lines. Neither of these types of models has demonstrated a
valuable role in drug discovery, likely because they do not reflect the genetic and cellular
diversity of human HGSC (Talmadge et al., 2007). PDXs potentially resolve some of the issues
seen with other in vivo models. Establishment of PDXs consists of obtaining surgical tissue and
implanting it into an immunodeficient mouse (Tentler et al., 2012). PDXs’ tumours have been
shown to maintain the architecture, biological properties and histological characteristics of the
original tumour (Gray et al., 2004; Stewart et al., 2011).
A robust PDX model could alter the course of patients with ovarian HGSC in two ways. First, if
PDXs reproduce the patient’s response to platinum treatment, this could allow for identification
of platinum-resistant/refractory patients. Identifying such ‘high-risk’ patients could change the
way they are triaged, followed and managed, and ultimately could increase the patient’s quality
and quantify of life. Second, these models could be used to develop and test new therapeutics
and predict the best course of treatment for a given patient. Several therapies could be tested at
the same time on a ‘library’ of PDXs derived from the same tumour, thereby tailoring the
patient’s treatment in order to establish effective drug response.
2.1.1 Drug Dose Translation
First-line chemotherapy for advanced ovarian cancer consists of Carboplatin and paclitaxel
administered IV at the maximum tolerable dose (MTD), followed by a treatment-free interval to
allow recovery of healthy tissues. The efficacy of Carboplatin is directly related to its plasma
concentration. Since Carboplatin is eliminated primarily via glomerular filtration, the
pharmacokinetics and, ultimately, the pharmacodynamics of Carboplatin are highly dependent on
renal function. Conventional fixed dosing based on body surface area (BSA) has led to
Carboplatin overdosing or, more commonly, under-dosing, which has resulted in less than
39
optimal treatment results (Alberts and Dorr, 1998). Consequently, dosing is typically
individualized by using the Calvert formula: target AUC x (creatinine clearance +25). The area
under the curve (AUC) is based on pharmacokinetic studies of the dose effect. The AUC level of
drug and/or its active metabolites correlates with toxicity and therapeutic effect and is a product
of desired serum concentration (mg/mL) and time (min) (Frei and Antman, 2000). Typically, an
AUC of 4-6 is used for treating ovarian HGSC patients, with doses corresponding to 600-900
mg/treatment (Calvert et al., 1989).
Table 2.1 Previously published Carboplatin doses administered to mice.
NSCL: non-small cell lung cancer
Typically, mouse-to-human drug dose translation is obtained through normalization of the dose
to the BSA and multiplication by the Km factor (Human equivalent dose = Animal dose x
Kmanimal/Kmhuman). The Km factor, body weight (kg) divided by BSA (m2), is used to convert the
mg/kg dose to a mg/m2 dose. The BSA correlates well with parameters of mammalian biology
(blood volume, renal function, etc.), which makes BSA normalization logical for scaling drug
doses between species (Reagan-Shaw et al., 2008). However, because dosing of Carboplatin uses
the Calvert formula rather than BSA dosing, using the Km factor for drug translation into mice is
Author Mouse strain Cancer model Carboplatin Dose
(Jandial et al., 2009; Press et al., 2008b)
NOD/SCID Ovarian & Uterine
Carboplatin IP 80 mg/kg Qweek x2
(Jandial et al., 2009) nu/nu Ovarian Carboplatin single dose: 85 mg/kg IP
(Merk et al., 2009) NMRI:nu/nu NSCLC Carboplatin IP 75mg/kg d1 and d8
(Fichtner et al., 2008) NMRI:nu/nu NSCLC Carboplatin IP 75 mg/kg d1 and d8
(Harrap et al., 1990) unspecified Ovarian cancer cell line
Carboplatin 100 mg/kg d0 & d7
(Das et al., 2008) BALB/c-nu/nu
Neurblastoma, Medulloblastoma
Carboplatin 100mg/kg IP Qweek x 3 weeks
(Merk et al., 2011) NOD/SCID NSCLC Carboplatin 75 mg/kg/d Qd 1 and 8 IP
40
adequate but imperfect. As such, I performed a thorough literature search to determine the
Carboplatin doses that had previously been used in other studies (Table 2.1).
2.1.2 CA125
CA125 (MUC 16) is a trans-membrane glycoprotein with a short intracellular domain, and a
large 22,097 amino acid extracellular domain, and an average molecular weight of ~2.5 million
Dalton (Theriault et al., 2011). It is aberrantly expressed and cleaved from the surface of ovarian
cancer cells and shed into the blood, providing a useful biomarker for disease monitoring (Wang
et al., 2008). CA125 is elevated in greater than 85% of serous HGSC patients and plays an
important role in the detection and management of patients (Duffy et al., 2005; Hogdall et al.,
2007).
CA125 is also thought to have a role in tumour cell growth, tumourigenesis, cell adhesion and
metastases (Rump et al., 2004; Theriault et al., 2011). MUC16 extends from the surface of
ovarian cancer cells and has been shown to bind to mesothelin, a protein found on the surface of
mesothelial cells lining the peritoneal cavity (Scholler et al., 2007). Theriault et al. showed that
knockdown of MUC16 in ovarian cancer cell lines inhibited tumour growth in vitro and in vivo,
and conversely, that overexpression increased tumourigenicity and metastases (Theriault et al.,
2011). Furthermore, in vitro, it has been implicated in sensitivity of ovarian cancer to
chemotherapy, with higher levels being suggestive of platinum resistance (Boivin et al., 2009).
Physiologically, mucins are involved in coating, lubricating and protecting epithelial surfaces of
internal tracts (Bafna et al., 2010). During embryonic development, CA125 is expressed in fetal
coelomic epithelia and its derivatives. In adulthood, MUC16 is normally expressed by the lining
of pleural, peritoneal and pelvic cavities, as well as in the pericardium of the heart, ocular surface
and upper respiratory tract.
Given its expression in a variety of tissues and organs, CA125 serum levels have been detected
in other types of cancers (endometrial, colon, pancreatic, breast) and in some benign conditions
(endometriosis, pelvic inflammatory disease), making it a non-specific marker for ovarian cancer
(Bast et al., 1998; Wang et al., 2008). With a positive and negative predictive value of 60% and
90%, respectively, serum CA125 levels are used to assess the risk of malignancy in women with
41
a pelvic mass (O'Connell et al., 1987). Once diagnosed with ovarian cancer, CA125
measurements are performed routinely and are used to monitor response to treatment,
progression and recurrence (Aggarwal and Kehoe, 2010; Diaz-Padilla et al., 2012). The positive
predictive value of elevated serum CA125 at suspected recurrence is >95% (Prat et al., 2009).
Decisions to continue or cease chemotherapy are made on the basis of rising CA125, even in the
absence of radiological progression. Given the importance of CA125 serum values in managing
ovarian cancer patients, I wanted to assess whether PDXs similarly express CA125. Moreover, I
wanted to see if the change in CA125 following chemotherapy could be used in conjunction with
tumour volume measurements as a corollary measure of drug response.
2.1.3 Tumour Initiating Cells
Over 85% of patients will recur following initial therapy, and the challenge in treating HGSC is
that recurrences tend to be platinum-resistant, limiting further treatment options (Kulkarni-Datar
et al., 2013). Understanding how chemo-resistant tumour cells develop and potentially cause
disease resurgence could provide a strategy to modify the current therapeutic management of
both primary and recurrent ovarian cancer.
Solid tumours, and in particular HGSC, display tremendous inter- and intra-patient cellular
heterogeneity (Bashashati et al., 2013; Burgos-Ojeda et al., 2012). Two models exist to explain
these distinct tumour cell populations. The stochastic model holds that all cells within a tumour
are equally capable of generating and sustaining the tumour. This model argues that cancer
heterogeneity is the result of selection pressures exerted on tumour cells, causing stochastic
genetic and epigenetic changes. These changes confer a survival advantage to certain cells and
the development of an increasingly more aggressive cancer over time (Aparicio and Caldas,
2013).
By contrast, the CSC model suggests that only a subpopulation of tumour cells, termed Cancer
Stem Cells (CSC), possess self-renewal ability and the capacity to initiate tumour growth and
drive tumour progression. In the CSC model, tumour heterogeneity results from differentiation of
CSC into phenotypically diverse non-tumourigenic cells (Kulkarni-Datar et al., 2013). An
essential feature of the CSC model is that tumours are organized hierarchically, so that
42
tumourigenic and non-tumourigenic cells can be prospectively differentiated from one another
according to their unique phenotypes (Burgos-Ojeda et al., 2012).
Importantly, a hierarchical organization of cells, intrinsic to the CSC model, has not been
described in ovarian cancers (Dyall et al., 2010). Furthermore, no single marker clearly identifies
the ovarian CSC population and the use of combinations of markers has yielded inconclusive
results (Foster et al., 2013). Burgos-Ojeda et al., proposed a hierarchy of ovarian CSC based on
the current literature; however, they acknowledge that conclusive markers and definitive data
have yet to be identified (Burgos-Ojeda et al., 2012).
Recent attention has focused on ovarian CSCs as a mechanism behind disease relapse and drug
resistance. It is believed that CSCs in other tumour sites evade the toxicity of standard
chemotherapy due to their slow rate of cell division and enhanced drug efflux properties (Rizzo
et al., 2011; Steg et al., 2012). As such, following treatment with chemotherapy, a reservoir of
CSCs could persist and subsequently reproduce the tumour phenotype (O'Brien et al., 2009;
Zhang et al., 2008). As a result, treatment with chemotherapy is thought to actually enrich for
tumourigenic cells. This has been shown in several cancers, including colon, pancreatic, breast
and brain cancer models (Bao et al., 2006; Hermann et al., 2007; Li et al., 2008; Dylla et al.,
2008).
Although often used interchangeably, CSCs differ from tumour initiating cells (TICs)
operationally. TICs are cells capable of initiating a tumour in immunodeficient mice. However,
according to the CSC model, TICs are organized hierarchically and can be distinguished
prospectively from non-TICs. Many cancers might contain TICs, but some could also contain
tumorigenic cells, without hierarchical organization (Zhou et al., 2009).
The ability to identify and isolate CSCs in other tumours was made possible by knowledge of
surface markers, expression patterns, and immunophenotyping of normal stem cells in some
organs or tissues (Bapat, 2010). Lack of information regarding normal stem cells in ovarian
biology has been a limiting factor in the identification of CSCs in ovarian cancer (Alvero et al.,
2009a). Several groups have identified CD133, ALDH, CD44, CD117 and CD24 as ‘stem-like’
markers; however, a hierarchy using these markers has yet to be defined (Curley et al., 2009;
Gao et al., 2010; Luo et al., 2011; Zhang et al., 2008). Stewart et al., found that ovarian HGSC
43
conforms to the CSC model, but that there are at least two (CD133+ and CD133-), and possibly
three (CD133+, CD133-, CD133+/-) CSC phenotypes in this disease (Stewart et al., 2011).
To date, enrichment or isolation of these CSCs has been carried out via multiple strategies, most
frequently utilizing immortalized cell lines (which typically are not HGSC) or cultured cells,
which calls into question the relevance of these results to primary human tumours. By definition,
CSC must have the capacity to form tumours in vivo, which again, argues for caution in
interpretation of many studies that have used in vitro experiments (Zhou et al., 2009).
Some of these studies have suggested that there is an enrichment of ovarian cancer stem-like cell
populations post-chemotherapy, supporting the concept that these cells resist conventional
chemotherapy and contribute to resurgence of disease (Figure 1.2). Alvero et al., identified a
CD44-positive cell population with CSC-like properties, that when treated with Carboplatin and
paclitaxel, differentiated to become even more resistant in response to chemotherapy (Alvero et
al., 2009b). Again, these studies were done in vitro and did not properly examine the CD44-
negative cell compartment.
The gold standard assay to measure the frequency of cells possessing a unique functional
property within a heterogeneous cell population is in vivo limiting dilution analysis (LDA). In
normal and cancer stem cell biology, this is an experimental method that is used to detect and
quantify cells with the ability to engraft a recipient mouse. For CSCs, this assay measures the
frequency of cells that are able to give rise to tumours (Eirew et al., 2010). I performed LDAs
with tumour cells obtained from Carboplatin-treated PDXs or control (saline-treated) PDXs in
order to ascertain whether treatment with Carboplatin enriches for the TIC population.
2.2 Methods
2.2.1 Identification of Patient Samples
Patient samples were identified retrospectively, to be either platinum-resistant (3654, 3875,
6447, 2555, 2028, 2462, 2803, 2489, 3444, 2903, 6259, 2685, 2753, 2261, 4070, 62143, 6259) or
platinum-sensitive (3748, 3670, 61382, 63867, 3028, 3336) at the time of acquisition. In
44
addition, samples were required to have sufficient clinical information and must have been
previously viably frozen in ample quantities in the Neel laboratory’s HGSC repository.
Prospective samples (67199, 67326, 6565, 6775) were chemo-naïve, primary samples from
patients likely to be followed at The Princess Margaret Cancer Centre.
2.2.2 Tumour Processing & Establishment of Xenografts
Human tumour specimens were obtained with informed consent from patients undergoing
surgery at the PM following a protocol approved by the University Health Network Research
Ethics Board (REB #06-0903T).
HGSC tissue samples were obtained from the University Health Network Tissue Bank. Once
confirmed by a pathologist, tumours were obtained within 5 hours of excision. To establish first
generation xenografts (X1), solid tumours were minced and digested with collagenase-
hyaluronidase (Stem Cell Technologies) in DMEM at 37°C for two hours. Red blood cells were
lysed in 0.16M ammonium chloride for approximately 5 minutes. The remaining cells were
filtered through a 70µm, mesh and viable cells were counted using trypan blue exclusion. Ascites
cells were collected by centrifugation at 300xg and processed as above. Anti-CD45 microbeads
were used to deplete leukocytes, as per the manufacturer’s instructions (Miltenyi Biotec). CD45-
depleted cells (106) in 1:1 HBSS:growth factor-reduced Matrigel (BD Biosciences) were injected
into the mammary fat pads (1 pad/mouse) of healthy, 6-8 week old, female NSG mice. Tumours
obtained from xenografts were processed as described above. Single cells were frozen viably in
10% DMSO, 15% FBS, 50% DMEM, and stored in liquid nitrogen.
To generate X2 (or later passage) xenografts, previously frozen cells were thawed rapidly in
37°C and washed in HBSS + 2% serum. Anti-H2K-biotin microbeads were used to deplete
mouse stroma, as per the manufacturer’s instructions (Miltenyi Biotec). H2K-depleted cells (106)
in 1:1 HBSS:growth factor-reduced Matrigel (BD Biosciences) were injected into the right
mammary fat pad of 10-15 NSG mice. When tumours became palpable (~200 mm3), mice were
treated with Carboplatin. Mice were sacrificed when tumours exceeded the size of 2.5cm or if
mice looked unwell or fell under criteria indicated in the humane endpoints.
Upon reaching the endpoint (either termination of treatment or humane endpoint), all mice were
45
sacrificed in accordance with the protocol procedures. Tumours were excised, a portion was
snap-frozen at −80 °C, and some tissue was fixed in 10% neutral buffered formalin and paraffin-
embedded. If there was sufficient material, cells were viably frozen according the methods
described above or were expanded into additional mice (either for LDA or future use).
2.2.3 Tumour Measurement
Tumour measurements were made every 72 hours using electronic calipers. Tumour volume was
approximated using the formula: volume = length x width2 x 0.52.
2.2.4 Carboplatin Administration
Carboplatin 10mg/mL (Hospira) was purchased from the PM pharmacy. Based on the mouse-to-
human drug translation formula (Section 2.1.1), and assuming an average human adult weight of
60kg, I estimated that an appropriate Carboplatin dose for mice to be approximately 125 mg/kg.
Given the published Carboplatin LD10 of 100mg/kg, and published literature using smaller
(<125mg/kg) doses, I performed an MTD assay (Figure 2.1), on healthy, 6-8 week female NSG
mice. From this, I concluded that 75mg/kg would be an appropriate dose. As such, experimental
mice were given 75mg/kg Carboplatin IP (in a total volume of 0.35 cc normal saline) once
weekly (day 0 & day 7) for 2 doses (dosing based on Table 2.1). Control mice received 0.35 cc
of normal saline IP.
2.2.5 CA125 ELISA & Immunohistochemistry
Serum was collected from all xenografts corresponding to specimens 3654 and 3670 at four
times: baseline (prior to inoculation with tumour cells), 6 weeks after inoculation with tumour,
pre-chemotherapy dose 1 and post-mortem (one week after 2nd chemotherapy dose). To do so,
mice were warmed with a heating lamp for approximately 10 minutes. The dorsal limbs were
shaved, and blood was drawn from the saphenous vein and collected into Eppendorf tubes. Post-
mortem blood was collected via cardiac puncture. Whole blood was stored on ice until it was
centrifuged at 16,000 rotations per minute. The serum layer was collected and stored at -20°C for
future use. A CA125 ELISA kit was purchased from Alpha Diagnostic International (Cat. No.
1820). Assays were performed according to the kit instructions. As most serum samples were
46
less than 25 µL, typically 20µL of standards, controls and serum was pipetted into each well.
Patient serum samples (obtained with consent) were used as positive controls, with known
CA125 values of 500, 50 and 5 U/mL.
To evaluate CA125 by immunohistochemistry (IHC), paraffin-embedded tissue sections were
stained with antibodies against CA125 (Leica, NCL-CA125; 1:50), according to the UHN
Pathology Research Program protocol.
2.2.6 Limiting Dilution Assays
A)
B)
Figure 2.0. Diagram showing LDA method and sample preparation. At the completion of A) Carboplatin treatment or B) control treatment (3 weeks), mice with palpable tumours were sacrificed and tumours were excised. Each tumour was dissociated into a single cell suspension. Cells were counted and bilaterally injected into the fat pads (FP) of NSG mice.
47
Upon reaching the termination of Carboplatin treatment, all mice were sacrificed in accordance
with the protocol procedures. Tumours were excised and processed as described previously. If
there were sufficient cells, LDAs were performed by injecting three cell doses into the mammary
fat pads of NSG mice (Figure 2.0). The number of fat pads injected for each dose was case- and
cell number-dependent and ranged from 2 to 8. TIC frequencies were calculated using ELDA
software (http://bioinf.wehi.edu.au/software/elda/) (Hu and Smyth, 2009).
2.3 Results
2.3.1 Engraftment
A total of 27 patient tumours were implanted between January 2012 and March 2013. Of these,
17 were classified as platinum-resistant, 6 platinum-sensitive, and 4 prospective (Table 2.2).
Eighteen out of 27 (67%) tumours (corresponding to 16 unique patients) engrafted successfully:
13/17 (76%) resistant cases, 3/6 sensitive cases (50%) and 2/4 (50%) prospective cases. Two
cases represent the same patient with samples obtained at different points in the patient’s clinical
course (3654/3875 and 2462/2803). Of the 18 successfully engrafted cases, 10/18 (56%) were
from ascites and 8/18 (44%) from solid tissue samples. One mouse had lymphoid proliferation,
without evidence of an epithelial tumour was found (6259) and was discarded. Median time from
cell injection to tumor treatment (at ~200mm3) was 15 weeks (range 3-21 weeks).
2.3.2 Patient Demographics
Patient demographics were analyzed for the 18 cases (corresponding to 16 patients) and are
summarized in Table 2.3. The mean age of all patients was 60.3 years. The mean age of
platinum-resistant, platinum-sensitive and prospective patients was 63.2, 53.6 and 54.5 years,
respectively. One patient had a known BRCA1 mutation (3444) and one patient was a BRCA1
variant of undefined significance (3748). The majority of patients (12/16; 75%) had either not
been tested or were negative. All patients in the ‘resistant’ category received their initial course
of Carboplatin/Taxol intravenously, whereas the ‘sensitive’ and ‘prospective’ patients received
either IP + IV or IP only. In addition, all sensitive and prospective patients were optimally
48
debulked. Patient 61382 was ‘optimally’ debulked, but there was a question of a liver lesion
post-operatively. In the resistant group, 4/11 (36%) were known to be optimally debulked. All
patients had stage 3C or greater, high-grade disease.
!49
!
50
2.3.3 Mouse Carboplatin MTD
The maximum tolerated dose is defined as the dose that causes no more than a 10% decrease in
body weight and does not produce mortality or toxicity. Based on a known mouse LD10 value of
100mg/kg (data not shown) and published experiments (Table 2.1), I performed an MTD assay
using a dose ranging from 20-80mg/kg to assess whether a target dose of 75mg/kg was
appropriate. Mouse body weight was measured weekly and compared to original (starting) body
weight (Figure 2.1). Based on this assay, no mouse had a >10% weight loss at any of the tested
doses. As a result, I proceeded with using 75mg/kg Carboplatin IP for our experiment.
Figure 2.1 Establishment of Carboplatin maximum tolerated dose (MTD) in NSG mice. Treatment was administered once weekly (d0, d7; black arrows) and body weights were measured weekly. Points are an average of two mice represented as percent of original starting weight.
51
2.3.4 PDXs from Platinum-Sensitive Patients Show a Near-Complete
Response to Platinum Therapy
To define response, I used the RECIST criteria definition of at least a 30% decrease in tumour
size (Eisenhauer et al., 2009). In total, PDXs were derived from 3 platinum-sensitive patients:
3670, 3748 and 61382. The clinical history of each patient will be described in detail. PDXs
derived from chemo-sensitive patients showed 77-91% (average 83%) reduction in tumour
volume, compared with the initial tumour.
Figure 2.2A shows the clinical history for patient 3670 based on her CA125 graph and Figure
2.2B shows the PDXs’ response to platinum therapy. At the time of diagnosis, patient 3670 was
a 68-year-old woman with a history of an abdominal mass and pain. This pain persisted over 3
months, until finally in October of 2009, she presented to the emergency room. A CT scan
showed a large left lower quadrant mass and evidence of bowel ischemia. She underwent urgent
surgery, which consisted of a TAH-BSO, omentectomy, debulking and sub-total colectomy. She
had stage 3C disease and was optimally debulked. Post-operatively, she received one cycle of IV
Carboplatin/Taxol, followed by 6 cycles of IP Cisplatin+IV Taxol. In November of 2011, she
had ultrasound imaging that confirmed the presence of multiple solid pelvic masses consistent
with recurrent carcinoma, as well as metastatic adenopathy in the peri-aoritc region. She
continued further chemotherapy in Barrie, Ontario. As of August 2013, the patient continues to
be alive.
52
A)
B)
Figure 2.2. Clinical course of patient 3670 and PDXs’ response to platinum. A) Clinical history of patient based on serum CA125 (U/mL) over time (months). Patient 3670 recurred 18 months after completion of initial therapy. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #3670 PDXs (n=13) to platinum therapy ± s.e. PDXs received initial dose of Carboplatin when the tumours averaged 200 mm3 (black arrow indicates time of treatment). Note 91% percent decrease in tumour volume of Carboplatin-treated PDXs, compared with initial tumour volume (paired t-test one-tailed P value = 0.0005).
53
Figure 2.3A shows the clinical history for patient 3748 based on her CA125 values, and Figure
2.3B shows the PDXs’ response to platinum therapy. Patient 3748 was a 40-year-old women
referred to the GYN oncology clinic in October 2009 with bilateral pelvic masses. The patient
had experienced abdominal pain since August 2009, and had an ultrasound through her family
physician that revealed a vascular, 8 x 7 x 5.9 cm predominantly solid mass in the posterior cul-
de-sac, and a small 2.5 cm mass in the inferior aspect of the sigmoid thought to arise from the
left ovary. Her CA125 was 3533. In November, she underwent primary debulking surgery
(TAH/BSO, omentectomy, LAR, right diaphragmatic stripping, peri-aortic lymph node
dissection, pelvic peritonectomy) and was found to have stage 3C disease. Post-operatively, she
received 2 cycles of IV Carboplatin/Taxol, followed by 4 cycles of IP Cisplatin/IV Taxol, which
finished in May 2010. Following routine examination in October 2011, the patient was thought
to have a local recurrence at the rectovaginal septum, which was confirmed by MRI. The patient
was offered Carboplatin/Taxol but preferred to delay treatment until May 2012. She completed 6
cycles of Carboplatin/Taxol in July 2012. In July 2013, she was diagnosed with a second
recurrence with a nodule in the rectovaginal septum and a CA125 of 2000U/mL. As of August
2013, she is alive and may be re-starting Carboplatin/Taxol therapy in the near future.
54
A)
B)
Figure 2.3. Clinical course of patient 3748 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Patient 3748 recurred 17 months after initial therapy. Dashed line represents normal serum CA125 (35 U/mL) B) Response of #3748 PDXs (n=5) to platinum therapy± s.e. PDXs received initial dose of Carboplatin when the tumours averaged 300 mm3 (black arrow indicates time of treatment). Note, 77% decrease in tumour volume of Carboplatin-treated PDXs, compared with initial tumour volume (paired t-test one-tailed P value is 0.04).
55
Figure 2.4A shows the clinical history for patient 61382 based on her CA125 values and Figure
2.4B shows the PDXs’ response to platinum therapy. At the time of initial diagnosis, patient
#61382 was a 52 year-old, post-menopausal female. She had an ultrasound to investigate
symptoms of abdominal pain and vaginal discharge, which revealed bilateral adnexal masses,
with mixed solid and cystic components and no ascites or extra-ovarian disease. Her CA125 at
the time was 341. In January 2008, she underwent a TAH, BSO, pelvic peritonectomy,
appendectomy, resection of sigmoid colon, right partial diaphragmatic resection, and
omentectomy. She had stage 3C disease, which was optimally debulked. She began with one
cycle of IV chemotherapy (Carboplatin + Taxol) and switched to IP chemotherapy for five
additional cycles. Following completion of chemotherapy, the patient had a CT scan to confirm
remission; However, this scan demonstrated a ~1cm lesion on the serosa of the liver, which was
suggestive of disease and called into question whether the patient was optimally debulked. This
lesion stayed relatively stable in size over the course of several months. However, a repeat scan
in November 2009 showed a new lesion in the hilum of the spleen. In September 2010, she
underwent a debulking splenectomy with no adjuvant chemotherapy. In December 2011, the
patient had a PET scan, which showed two lesions in the liver and one peri-aortic node
suggestive of recurrent disease. The patient completed 6 cycles of second-line IV
Carboplatin/Taxol in July 2012 and continues to have no evidence of disease on CT scan.
56
A)
B)
Figure 2.4. Clinical course for patient 61382 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Patient 61382 recurred 18 months after initial therapy. Dashed line represents normal serum CA125 (35 U/mL) B) Response of #61382 PDXs (n=4) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Note 81% reduction in tumour volume of Carboplatin-treated PDXs, compared with initial tumour volume (paired t-test one-tailed P value is 0.001).
57
2.3.5 PDXs from Platinum-Resistant Patients do not Respond to Platinum
Therapy
In total, PDXs were derived from 11 platinum-resistant patients: 3654/3875, 6447, 2261,
2462/2803, 2489, 3444, 2555, 2903, 2753, 2028 and 2685. PDXs response ranged from 29%
(#6447) reduction in average tumour volume to 307% (#2555) increase in tumour volume,
despite platinum therapy. Six of the 11 cases showed some degree of reduction in tumour
volume, 1 case showed no response and 4 cases progressed in spite of platinum treatment. None
of the cases met the RECIST response criteria (30% reduction). The clinical histories of each
patient are described in detail.
Figure 2.5A shows the clinical history for patient 3654/3875 based on her CA125 values, and
Figures 2.5B and 2.5C show the response of PDXs 3654 and 3875 to platinum therapy,
respectively. At initial diagnosis, the patient was a 77-year-old referred from a transitional
hospital due to extensive abdominal carcinomatosis and elevated CA-125 (8500U/mL). She
experienced bloating, constipation, and urinary difficulties, with a 20 pound weight loss and
significant fatigue. Her past medical history was significant for a colonic adenocarcinoma treated
with colectomy 37 years prior and a renal cell carcinoma, treated by nephrectomy 9 years earlier.
For these reasons, in October 2008, she was seen in a peripheral hospital where a CT scan
showed ascites, peritoneal carcinomatosis, and a possible recurrence at her rectal stump.
Prior to commencing chemotherapy, the patient had an omental biopsy, which revealed a high-
grade metastatic carcinoma. She began to receive Carboplatin (20% reduction due to patient’s
frailty) and Taxol. She received 4 cycles prior to surgery. March 2009, she underwent a delayed
debulking surgery, consisting of a BSO, omentectomy. She was found to have stage 3C disease,
and went on to receive 4 additional cycles of chemotherapy. Unfortunately, in July 2009, she had
increasing abdominal pain and bloating, and a CT scan done in the emergency room confirmed
disease progression. Given her platinum-refractory disease, the patient was started on Caelyx
chemotherapy in September 2009, then Gemcitabine in October. Minimal response was observed
and the patient continued to decline. She had several palliative paracentheses and died in March
2010.
58
A)
B)
59
C)
Figure 2.5. Clinical course for patient 3654/3875’s and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. The patient recurred 3 months after initial therapy. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #3654 PDXs (n=7) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 200 mm3 (black arrow indicates time of treatment). C) Response of #3875 PDXs (n=5) to platinum therapy± s.e. Carboplatin-treated PDXs 3654 showed an average 21% increase in tumour volume, compared with initial tumour volume (paired t-test one-tailed P value is 0.2). PDXs 3875 showed an average 6% reduction in tumour volume of Carboplatin-treated PDXs, compared with initial tumour volume (paired t-test one-tailed P value is 0.1).
Figure 2.6A shows the clinical history for patient 6447 based on her serum CA125 values and
Figure 2.6B shows PDXs’ response to platinum therapy. Patient 6447 was a 62-year-old woman
with a 1-month history of bloating, abdominal pain and fatigue. Ultrasound imaging revealed
free fluid in the cul-de-sac and a mass measuring 3.8 x 2.5 cm, with a CA125 of 1204 U/mL. She
received single agent Carboplatin in December 2009, followed by 2 cycles of Carboplatin/Taxol.
She had interval debulking (TAH/BSO, omentectomy) surgery in April 2010, with stage 3c
disease that was optimally debulked. Post-operatively, she completed a further 3 cycles of
Carboplatin/Taxol, ending in June 2010. CT scan done February 2011 showed peritoneal nodules
and ascites. She received one cycle of IV Taxol, and then 7 more cycles of Carboplatin/Taxol,
completed in October 2011. Nevertheless, she had significant progression of disease based on
60
CT imaging. She started Gemcitabine in April 2012, but continued to decline in function. She
died in May 2012.
A)
B)
Figure 2.6 Clinical course for patient 6447 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Patient 6447 recurred 9 months after initial therapy. Dashed line represents normal serum CA125 (35 U/mL). B. Response of #6447 PDXs (n=9) to platinum therapy± s.e. PDXs received first
61
dose of Carboplatin when the tumours averaged ~300 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed an average 29% reduction in tumour volume, compared with initial tumour volume (paired t-test one-tailed P value is 0.004).
Figure 2.7A shows the clinical history for patient 2261 based on her CA125 values, and Figure
2.7B shows the PDXs’ response to platinum therapy. This patient was seen in May 2007 as a 47
year-old, who had been diagnosed previously with stage 3c disease in July 2005 at a peripheral
hospital. At that time, she had a subtotal hysterectomy, BSO and omentectomy, followed by 6
cycles of Carboplatin/Taxol. CT imaging in May 2007 showed a rectovaginal nodule, in
combination with a rising CA125. The patient opted to have her recurrence treated in Hungary,
with naturopathic measures (vitamin-B17 IV) and paracenthesis. The patient was counseled to
receive traditional chemotherapy but continued to refuse until July 2008, when she received one
dose of Carboplatin/Taxol. She died in September 2008.
A)
62
B)
Figure 2.7 Clinical course of patient 2261 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on CA125 serum measurements over time. The patient recurred 18 months after initial therapy. B) Response of #2261 PDXs (n=3) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed an average 21% increase in tumour volume, compared with initial tumour volume (paired t-test one-tailed P value is 0.08).
Figure 2.8A shows the clinical history for patient 2028 based on her CA125 levels and Figure
2.8B shows the PDXs’response to platinum therapy. Patient 2028 was seen in the gynecologic
oncology clinic in October 2007, as a 76-year-old with presumed ovarian cancer. She had a
month-long-history of abdominal distention, anorexia, early satiety, and constipation, which
prompted her to go to the emergency room, where an ultrasound showed ascites and bilateral
hydronephrosis from a solid left adnexal lesion, measuring 2.7x2.2x2.4 cm. Her CA125 was
1971 U/mL. She received five courses of neoadjuvant Carboplatin/Taxol and proceeded to
interval debulking (TAH/BSO, omentectomy, peritonectomy), which was sub-optimal, in
February 2008. She then completed 2 additional cycles of Carboplatin/Taxol in April 2008.
Unfortunately, her CA125 never fell below 136, and in July, it was 862, with continued disease
progression. In January 2009, she was treated with single agent Carboplatin. She was admitted
to the palliative care floor in March 2009 and died in April 2009.
63
A)
B)
Figure 2.8 Clinical course of patient 2028 and PDX’s response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Patient 2028 recurred 3 months after initial therapy. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #2028 PDXs (n=6) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed an average 15.5% reduction in tumour volume, compared with initial tumour volume (paired t-est one-tailed P value is 0.03).
64
Figure 2.9A shows the clinical history for patient 2803 based on her CA125 levels and Figure
2.8B shows the PDXs’ response to platinum therapy. This patient was initially seen in 2004, as a
57-year-old with abdominal pain. An ultrasound demonstrated a large adnexal mass, and the
patient’s CA125 was elevated at 313U/mL. This patient received 4 cycles of neoadjuvant
chemotherapy, followed by interval debulking surgery (BSO, omentectomy). She had stage 3C
disease, and received 4 adjuvant courses of chemotherapy, completed in September 2004. In
October 2004, a CT showed minimal residual lesions in the pelvis and omentum. For this reason,
the patient started second line Caelyx in November 2004. She received 6 doses of Caelyx,
ending March 2005. In March 2007, she began to experience flank pain, and a CT showed
lymphadenopathy and moderate ascites. In July 2007, she was started on gemcitabine, but due to
the development of a rash, was switched to Topotecan for 8 cycles, ending in December 2007.
The patient was then re-started on cisplatin for 6 cycles, ending July 2008 due to ototoxicity. As
a consequence of her persistent disease, the patient was started on etoposide in September 2008,
but stopped in October due to disease progression. She died in August 2009.
A)
65
B)
Figure 2.9 Clinical course of patient 2803 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on CA125 serum measurements over time. The patient progressed on therapy. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #2803 PDXs (n=6) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 100 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed an average 6% increase in tumour volume, compared with initial tumour volume (paired t-est one-tailed P value is 0.3).
Figure 2.10A shows the clinical history for patient 2489 based on her CA125 level and Figure
2.10B shows the PDXs’ response to platinum therapy. Patient 2489 was seen in clinic in January
2008. At the time, she was 48-years-old, with a biopsy-proven papillary serous carcinoma,
thought to be ovarian in origin. She presented in December of 2007 with shortness of breath. A
chest x-ray showed a right pleural effusion, which was tapped and showed malignant cells. She
also had an omental biopsy, which showed papillary serous carcinoma and an elevated CA125
(2107 U/mL). She was given seven courses of Carboplatin/Taxol, ending in June 2008. Due to
disease progression, she began avastin and cyclophosphamide in September 2008.
Unfortunately, a breast lump confirmed to be a synchronous breast primary carcinoma was
discovered in November 2008. In March 2009, the patient began third line therapy in the form
of Caelyx, but continued to have ascites and a right pleural effusion. She was admitted to the
palliative care unit in July 2009 and died in August.
66
A)
B)
Figure 2.10 Clinical course of patient 2489 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. The patient progressed on therapy. B. Response of #2489 PDXs (n=4) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged ~150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed an average 11% reduction in tumour volume, compared with initial tumour volume (paired t-est one-tailed P value is 0.04).
67
Figure 2.11A shows the clinical history for patient 3444 based on her CA125 levels and Figure
2.11B shows the PDX’s response to platinum therapy. Patient 3444 was initially treated for
serous ovarian cancer in 2007. At that time, she received 3 cycles of neoadjuvant
Carboplatin/Taxol, followed by interval debulking surgery, and a further six cycles of
Carboplatin/Taxol, completed in February 2008. A routine CT scan performed in September
2008 showed an irregular mass adjacent to the rectosigmoid, which increased in size on a follow-
up scan. This mass received radiotherapy in 15 fractions completed in April 2009. The patient’s
past medical history was significant for a T1N2 receptor-positive breast cancer, treated in 1994
with a lumpectomy followed by chemotherapy, radiation and five years of tamoxifen treatment.
Given the breast and ovarian cancers, the patient underwent genetic testing and was found to be
BRCA1 mutation carrier. It was felt that the best way to deal with this recurrent pelvic mass
would be to perform a secondary surgery, which occurred in July 2009. The patient completed
another six courses of Carboplatin/Taxol, which finished in 2011. In November 2012, she was
found to have enlarging para-aortic nodes, and was started back on chemotherapy, which she
finished in March 2013. The patient had a third surgery in March 2013, in the form of a bilateral
para-aortic lymphadenectomy and common iliac node dissection. All nodes were positive for
disease. She resumed chemotherapy at her local hospital and is alive as of August 2013.
68
A)
B)
Figure 2.11 Clinical course of patient 3444 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. The patient progressed on therapy. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #3444 PDXs (n=7) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). 3444 PDXs showed no reduction in tumour volume when treated with Carboplatin, compared with initial tumour volume (paired t-est one-tailed P value is 0.5).
69
Figure 2.12A shows the clinical history for patient 2555 based on her serum CA125 levels and
Figure 2.12B shows the PDXs’response to platinum therapy. Patient 2555 was seen in June 2008
at age 77. Her history dated back to August 2007 when she first noted a hard right-sided inguinal
mass. The mass gradually increased in size but the patient did not seek medical attention until a
year later. She went on to have a CT scan in April 2008, which showed several masses in both
the left and right inguinal regions. In addition, there were enlarged nodes along the left and right
common iliac arteries. Biopsy of the groin node was performed in May 2008, and showed
metastatic carcinoma, consistent with an ovarian origin. CA125 was 1289 U/mL. The patient
remained largely asymptomatic, aside from some leg swelling as a result of the nodal
compression. She received one cycle of single agent Carboplatin in August, then presented to the
ER shortly thereafter with abdominal pain from advanced disease. She died in September 2008.
A)
70
B)
Figure 2.12 Clinical course of patient 2555 and PDX’s response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of #2555 PDXs (n=6) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 250 mm3 (black arrow indicates time of treatment). 2555 PDXs showed a 307% increase in tumour volume when treated with Carboplatin, compared with initial tumour volume (paired t-est one-tailed P value is 0.02).
Figure 2.13A shows the clinical history for patient 2903 based on her CA125 levels and Figure
2.13B shows the PDXs’ response to platinum therapy. This patient was referred from an outside
hospital. She was 57-years-old with stage 4 disease, and had previously received three lines of
chemotherapy. She was initially diagnosed in 2004 with ascites, a pelvic mass and a CA125 of
3344. She received 3 cycles of neoadjuvant Carboplatin/Taxol, followed by a
TAH/BSO/omentectomy in April 2005. She received 3 additional cycles of Carboplatin/Taxol.
Her CA125 was undetectable at the completion of treatment in August 2005. One year later, in
July 2006, she re-developed ascites and had a CA125 of 2500U/mL. She was then treated with 6
cycles of liposomal doxorubicin, which ended in December 2006. In April 2007, she developed
ascites again, with a CA125 of 1483U/mL and was treated with cisplatin plus VP-16 for 2 cycles.
In January 2008, she was started on Letrozole, but her CA125 continued to rise and CT imaging
71
showed peritoneal metastases. In July 2008, she started on a clinical trial of AMG 386 (an anti-
angiopoietin 1 and 2 drug) plus Taxol, but progressed on this and was started on Etoposide in
mid-January 2009. She continued on Etoposide treatment for a very short period and was
enrolled in a PARP-inhibitor trial, but had progression of disease. She died shortly thereafter, in
May 2010.
A)
72
B)
Figure 2.13 Clinical course of patient 2903 and PDX’s response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of PDXs (n=4) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed a 10% reduction in tumour volume when treated with Carboplatin, compared to initial tumour volume (paired t-est one-tailed P value is 0.1).
Figure 2.14A shows the clinical history for patient 2753 based on her CA125 levels and Figure
2.14B shows the PDXs’ response to platinum therapy. Patient 2753 was first seen in the fall of
2008 as a 58-year-old referred for a pelvic mass and extensive ascites. She had surgery in late
October 2008, consisting of TAH/BSO/omentectomy and debulking, but was sub-optimally
debulked due to diffuse disease. The patient received six cycles of adjuvant Carboplatin/Taxol
and two cycles of single agent Carboplatin, but continued to have residual pelvic disease at the
completion of treatment in May 2009. The patient pursued some travel prior to initiating second
line treatment with Caelyx in November. She had continued disease progression and started third
line therapy (Gemcitabine) in May 2010 and fourth line therapy (Topotecan) in August 2010.
The patient is alive as of August 2013.
73
A)
B)
Figure 2.14 Clinical course of patient 2753 and PDX’s response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of xenografts (n=4) to platinum therapy ± s.e. PDXs received first dose of Carboplatin when the tumours averaged 400 mm3 (black arrow indicates time of treatment). Experiment had to be terminated prematurely as tumour volume was approaching endpoint. After one dose, PDXs did not respond to Carboplatin therapy and continued to grow at a rate similar to that of controls.
74
Figure 2.15A shows the clinical history for patient 2685 based on her CA125 levels and Figure
2.15B shows the PDX’s response to platinum therapy. This patient was initially seen in June
2008, at the age of 72, when her family physician found clinical and radiological evidence of
ovarian cancer. The patient had symptoms of bloating and abdominal discomfort. An ultrasound
showed a moderate amount of ascites, and blood work revealed a CA125 of 4759 U/mL. The
patient had a previous history of breast cancer, treated with surgery in 2001 and tamoxifen,
ending in 2006. She had a paracenthesis confirming serous adenocarcinoma, consistent with an
ovarian origin. She completed four cycles of IV neoadjuvant chemotherapy in September 2008,
prior to her surgical procedure in October 2008 (BSO, omentectomy). She was sub-optimally
debulked, with residual disease on the diaphragm, large bowel and cul-de-sac. She received four
additional cycles of Carboplatin/Taxol, completed in January 2009. CT scan performed shortly
after the completion of chemotherapy showed increased peritoneal disease and an enlarging
pelvic mass. The patient died in March 2009.
A)
75
B)
Figure 2.15 Clinical course of patient 2685 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on serum CA125 measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of xenografts (n=4) to platinum therapy± s.e. PDXs received first dose of Carboplatin when the tumours averaged 200 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed a 37% increase in tumour volume compared with initial tumour volume (paired t-est one-tailed P value is 0.03).
2.3.6 ‘Prospective’ PDXs Predict Chemo-Responsiveness
PDXs from two patient samples (67199 and 67326) were derived and treated prospectively (i.e.,
prior to knowing the clinical outcomes of the patients). Xenografts derived from both patients
showed a marked reduction in tumour volume (90% and 63% for 67199 and 67326,
respectively), suggesting that these patients would be platinum-sensitive. After following their
clinical course for over 1 year, both patients have not recurred (>12 months), confirming that
both patients have platinum-sensitive disease.
Figure 2.16A shows the clinical course of patient 67199. The patient was 51-years-old when
initially seen in the gynecologic-oncology clinic with CT scan results showing a solid mass with
hyper-vascularity, as well as ascites and omental caking. At the end of March, she received a
TAH/BSO, peritonectomy, diaphragm resection, and para-aortic lymph node sampling. She was
76
optimally debulked and had stage 4 disease. She received one cycle of IV Carboplatin/Taxol,
followed by 5 cycles of IP cisplatin/IV Taxol. In June, her CA125 rose to 63 (from 36 in March),
suggesting an early indication of recurrence; however, a CT scan in July showed no evidence of
disease. Her current PFI is 17 months. The PDXs had a near complete response to platinum
therapy. Given that the patient’s PFI is >17 months, the PDXs correctly predicted the platinum-
response of this patient.
A)
77
B)
Figure 2.16 Clinical course of patient 67199 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on CA125 serum measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of PDXs (n=5) to platinum therapy ± s.e. PDXs received first dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed a 90% reduction in tumour volume, compared with initial tumour volume (paired t-est one-tailed P value is 0.03).
Similarly, patient 67326 was initially seen in the gynecologic oncology clinic in March 2012, as
a 58-year-old female with a left adnexal mass and a CA125 of 92 U/mL. The patient had been
feeling otherwise asymptomatic, aside from some mild abdominal distention. In April 2012, she
underwent a TAH/BSO, LAR, omentectomy, pelvic and para-aortic lymph node dissection. She
had stage 3c disease and was optimally debulked. Post-operatively, she received 6 cycles of IP
cisplatin and IV Taxol day 1 and IP Taxol day 8. Last seen in April 2013, she had no evidence
of disease. Her current PFI is 16 months. Figure 2.17 shows the patient’s clinical course and the
PDXs’ response to platinum.
78
A)
B)
Figure 2.17 Clinical course of patient 67326 and PDXs’ response to platinum. A) CA125 graph showing clinical history of patient based on CA125 serum measurements over time. Dashed line represents normal serum CA125 (35 U/mL). B) Response of PDXs (n=4) to platinum therapy ± s.e. PDXs received first dose of Carboplatin when the tumours averaged 400 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed a 63% reduction in tumour volume, compared with initial tumour volume (paired t-est one-tailed P value is 0.05).
79
2.3.7 PDX ‘Recurrence’ also Parallels Patient Clinical Course
PDX 3670 was derived from a platinum-sensitive patient, and these PDXs showed a 90%
reduction in average tumour volume upon platinum treatment (Figure 2.2B). At the completion
of the experiment, several mice (n=4) without palpable tumours were not sacrificed in order to
see if the tumours would re-grow. One-hundred forty-four days after the completion of
treatment, tumours in these mice became palpable again, and they were subjected to a second
round of Carboplatin therapy. Figure 2.18 shows the response of the ‘re-grown’ xenografts to
platinum treatment. Again, the xenografts showed a complete response to platinum therapy,
parallels the patient’s response.
Figure 2.18 #3670 PDXs response to second course of treatment with platinum. PDXs re-formed tumours 144 days after completion of initial therapy. PDXs (n=2) received ‘third’ dose of Carboplatin when the tumours averaged 150 mm3 (black arrow indicates time of treatment). Carboplatin-treated PDXs showed a 90% reduction in tumour volume, compared with initial tumour volume.
80
2.3.8 Summary of PDXs’ Response to Carboplatin
Figure 2.19 summarizes the results for all PDXs tested. Cases classified as ‘resistant’ showed a
range in response to Carboplatin treatment, from a 29% reduction (#6447) to a 305% increase
(#2555). The sensitive cases and prospective cases (which turned out to be sensitive), showed a
60-90% reduction in tumour volume. Figure 2.19B shows that when the treatment groups are
divided between ‘sensitive’ and ‘resistant’ cases, those classified as platinum-sensitive show a
significant reduction in average tumour volume, compared to resistant cases. There is no
difference in final average tumour volume between sensitive and resistant cases among the
untreated controls.
A)
81
B)
Figure 2.19. Tumour volume change in all tested PDXs. A) Waterfall plot showing percent change [(TVfinal-TVinitial)/TVinitial x 100] following Carboplatin treatment in all PDX cases. In total, 13 resistant and 5 sensitive cases were treated. Two of the sensitive cases were treated prospectively. Dashed line represents 30% RECIST definition. Note that none of the resistant PDXs (0/13) met the RECIST definition. B) Average tumour volume change following Carboplatin or vehicle treatment ± s.e. in sensitive (light grey) and resistant (dark grey) cases. ** = p value < 0.005, ns= non-significant using Mann Whitney test.
2.3.9 Primary Tumour Histology is Maintained in PDX Tissue
All PDX tissue was sent for histologic examination and review by a gynecologic pathologist. All
cases were confirmed to be HGSC, reflecting the histology of the primary tumour.
2.3.10 PDXs’ Serum CA125 Cannot be Used to Monitor Disease
Progression and/or Response
Using two cases (3670 and 3654), blood was collected from control (n=3) and treatment mice
(n=7) at 4 various times: baseline (prior to tumour inoculation); at 6 weeks; prior to first dose of
chemo (12 weeks); at endpoint (14 weeks). Patient serum CA125 at sample acquisition was 50
82
U/mL and 3725 U/mL for cases #3670 and #3654, respectively (Table 2.4). Although there was
seemingly a temporal change in mouse serum CA125 as demonstrated by absolute absorbance
values (Figure 2.20A), this corresponded to a clinically irrelevant CA125 value of ~10 U/mL
when plotted on the standard curve (Figure 2.20B). Furthermore, temporal changes seen in
Figure 2.20A were seen in both treated and control mice, suggesting normal fluctuations not
related to treatment.
A)
B)
Figure 2.20. PDX 3670 and 3654 CA125 levels A) CA125 ELISA showing absolute absorbance values measured over time ± s.e. Serum was collected at four time points, one of which corresponds to prior to administration of chemotherapy (black arrow). For both cases, the control group consisted of n=3, and treatment group, n=7. B. Absorbance values from graph (A) plotted on CA125 standard curve. All values correspond to CA125 < 10 U/mL (red triangle).
83
2.3.11 PDXs CA125 Immunohistochemistry Parallels Patient Serum CA125
Value
CA125 IHC staining was obtained on 15 PDX samples and compared with patient serum values
from the time of sample acquisition (Table 2.4). PDXs’ tissue obtained from patients with a low
serum CA125 (<100 U/mL) had minimal positive cell staining (cases: 3670, 2803, 2261, 2803).
Conversely, PDXs from patients with high serum CA125 (>500 U/mL) have considerably
greater CA125 staining (cases: 2903, 3654, 6447, 2489, 2028, 67199). However, there does not
appear to be a strong positive correlation, as some PDXs from high CA125 patients showed only
a moderate amount of staining (cases: 2555, 3748). Table 2.4. Patient serum CA125 value and PDX CA125 immunohistochemistry.
Patient ID Patient Serum CA125
PDX CA125 IHC (10x) PDX CA125 IHC(20x)
3444 10
3670 50$
2803 80$
84
Patient ID Patient Serum CA125$
PDX CA125 IHC (10x) PDX CA125 IHC (20x)
2261 90
67326 425
2903 600
2555 830
6447
1400
85
Patient ID Patient Serum CA125
PDX CA125 IHC (10x) PDX CA125 IHC (20x)
2489 1780
3875 3000
2028 3500
3748 3500
3654 3725
86
Patient ID
Patient$Serum$CA125$ PDX CA125 (10x) PDX CA125 IHC (20x)
67199 7630$
2.2.12 Effect of Carboplatin Treatment on TICs
Six cases were used for LDAs; however, one was subsequently found to be a mouse lymphoma
and results were discarded. Four of the 5 cases were from platinum-resistant patients and one
was from a ‘prospective’ patient, who was ultimately found to be platinum-sensitive. Cases
3654, 3875, 2555 and 6447 are all ascites specimens, whereas 67326 was a solid tumour sample.
Additional clinical details can be found in Table 2.2. In addition, LDA PDXs were passage 4
(x4) for case 3654, x3 for case 3875, x5 for case 2555, and x2 for cases 6447 and 67326.
Only one case (3654) showed a significant enrichment for CSC’s after in vivo treatment with
chemotherapy (Figure 2.21, Table 2.5). At the time of the LDA, treated mice had an average
tumour volume of 300mm3 and control mice of 600 mm3. This case was derived from a chemo-
resistant patient. TIC frequency increased from 1/18,000 to 1/3,000 following Carboplatin
treatment, with a p-value of 0.03.
Three cases (3875, 6447 and 2555), all derived from platinum-resistant patients, yielded
insignificant results (Tables 2.6-2.7). Both control and treated PDXs from case 3875 had an
average tumour volume of ~250 mm3 (no difference between groups) at the time of the LDA
(Figure 2.22). For case 3875, two PDXs (one mouse) died prior to tumour formation, making the
results difficult to interpret.
87
Control PDXs from case 6447 had an average tumour volume of 800 mm3, compared with 200
mm3 in the treatment group (Figure 2.24). This case had the largest number of pooled tumours
(n=5 in control group; n=9 in treatment group) and showed the most response to chemotherapy
of all of the platinum-resistant cases (29% reduction in tumour volume). There was a trend
toward enrichment in the chemo-treated PDXs (1/700 versus 1/1,400), but this was not
statistically significant.
At the time of the LDA, PDXs from case 2555, for both treatment and control, had the largest
average tumour volume of all cases (1800 mm3 and 500 mm3, respectively) (Figure 2.23). This
case grew very rapidly (from implantation to palpable tumours = 3 weeks) and aggressively, as
indicated by a TIC frequency of 1/1 in the treated group and 1/36 in the control group. The TIC
frequency in both control and treated groups was much higher than in any other case.
The LDA for one case (67326) showed significant enrichment (p<0.005) for CSCs in the control
arm of the experiment. These PDXs were derived from a platinum-sensitive patient (Figure
2.24). The tumour volume of the control and treated groups averaged 750 mm3 and 200 mm3,
respectively.
88
Figure 2.21 Carboplatin sensitivity experiment for case #3654. Carboplatin administration occurred at day 0 and day 7 (168 hours) (black arrows). Tumours from all mice were pooled for LDA experiments on day 14 (red arrow). Table 2.5 summarizes the results of the LDA experiments (xenografts formed/xenografts injected).
89
Figure 2.22 Carboplatin sensitivity experiment for case #3875. Carboplatin administration occurred at day 0 and day 7 (168 hours) (black arrows). Tumours from all mice were pooled for LDA experiments on day 14 (red arrow). Table 2.6 summarizes the results of the LDA experiments (xenografts formed/xenografts injected).
90
Figure 2.23 Carboplatin sensitivity experiment for case #2555. Carboplatin administration occurred at day 0 and day 7 (168 hours) (black arrows). Tumours from all mice were pooled for LDA experiments on day 14 (red arrow). Table 2.7 summarizes the results of the LDA experiments (xenografts formed/xenografts injected).
91
Figure 2.24 Carboplatin sensitivity experiment for case #6447. Carboplatin administration occurred at day 0 and day 7 (168 hours) (black arrows). Tumours from all mice were pooled for LDA experiments on day 14 (red arrow). Table 2.8 summarizes the results of the LDA experiments (xenografts formed/xenografts injected).
92
Figure 2.25 Carboplatin sensitivity experiment for case #67326. Carboplatin administration occurred at day 0 and day 7 (168 hours) (black arrows). Tumours from all mice were pooled for LDA experiments on day 14 (red arrow). Table 2.9 summarizes the results of the LDA experiments (xenografts formed/xenografts injected).
93
2.4 Discussion
The value of an in vivo model depends on the extent to which its characteristics reflect the
properties of the cancer. The aim of this study was to establish a pre-clinical model of ovarian
high-grade serous cancer that maintains a correlation with the characteristics of the clinical
disease. My findings suggest that, despite the tremendous cellular heterogeneity of the primary
tumour and the diverse clinical course of each patient, PDXs correctly reproduce the patient
response to platinum therapy. The PDXs’ response (or lack thereof) to platinum therapy
correlated with sample status (i.e., platinum-resistant or platinum-sensitive) in 100% of the cases.
I established and treated 18 cases, corresponding to 16 unique patients. For patients in whom
Carboplatin was ineffective, irrespective of whether this was primary or acquired resistance, the
corresponding PDXs showed no response when treated with platinum. Conversely, treatment of
PDXs derived from platinum-sensitive patients showed near-eradication of the tumour.
Furthermore, after completion of therapy, tumours for one platinum sensitive case (3670) re-
grew. Upon re-treatment with Carboplatin, this tumour was once again sensitive to the effects of
platinum. This also is in keeping with the patient response: the patient was re-treated with
platinum at her recurrence 18 months after completing initial therapy (data not shown). Four
years after her initial diagnosis, the patient continues to be alive, suggesting that her tumour had
a good response to platinum re-treatment.
Most impressively, the PDXs were able to predict the chemo-responsiveness of two patients for
who the clinical course of disease was not known at the time of the experiments. Cases #67199
and #67326 were primary, chemo-naïve samples engrafted at the time of the patients’ primary
surgery. PDXs derived from both patients showed a nearly complete response to platinum
therapy. Based on my PDX results, I predicted that each patient’s respective PFI would be >12
months, in keeping with platinum sensitivity. To date, both patients have a PFI of >17 months.
Three elements of this model that are worth discussing in detail: first, is the choice and dose of
chemotherapy administered; second, is the method used to determine PDX response; and third, is
the correlation with patients’ clinical outcomes.
94
2.4.1 Chemotherapy
Standard chemotherapy treatment for patients with ovarian cancer is platinum based using either
Carboplatin or Cisplatin. Originally cisplatin was the agent of choice, however, data from
numerous clinical trials of advanced ovarian cancer have now documented that IV Carboplatin is
equivalent to IV Cisplatin in activity, but causes considerably fewer side effects, including
ototoxicity, neurotoxicity, and nephrotoxicity. Standard IV treatment with Carboplatin and
paclitaxel is administered every 21 days for six cycles (Hennessy et al., 2009). More recently,
the preferred timing and route of administration have come into question. For optimally debulked
patients, IP chemotherapy has shown to have an advantage over standard IV chemotherapy, with
Cisplatin being the preferred analog (Armstrong et al., 2006). Despite some of the current
debates surrounding chemotherapy administration, platinum-based chemotherapy remains the
cornerstone of treatment of ovarian HGSC.
For my PDX experiments, mice were treated with IP Carboplatin, using a dose of 75mg/kg/dose.
Carboplatin was selected over Cisplatin because 15/16 patients in my cohort received
Carboplatin during their clinical course, whereas only one received exclusively Cisplatin
(#67326). Because Cisplatin and Carboplatin exert the same effects on DNA, I expect that the
results observed in this study would be very similar if IP Cisplatin (at the appropriate dose) had
been used instead. IP administration was used because attempts at mouse tail (IV) administration
were unsuccessful. These treatments were difficult and imprecise, particularly when
administering the second dose. After receiving the first chemotherapy dose, mice were vaso-
constricted, causing drug to spill out of the puncture site. This same phenomenon was observed
when attempting to collect blood for serum CA125 analysis, and as such, could only be done
post-mortem. In humans, it has been shown that through IP administration there is a high IV drug
concentration, therefore, I am confident that the IP mode of drug administration in the PDXs is
adequate in delivering the drug to the tumour.
As previously stated, Carboplatin is administered based on the Calvert formula, so that each
patient receives an individualized dose (Section 2.1.1). Because of this, there is no ‘standard
dose’; however, the average dose ranges between 600-900mg/treatment. Converting this to a
mg/kg dose (~10mg/kg), and then using the Km factor to determine the mg/m2 dose, I estimated
95
that the approximate equivalent Carboplatin dose for mice is 125 mg/kg. Although using the Km
factor provides a reasonable approach for estimating human-equivalent/mouse-equivalent doses,
this factor does not take into account the mouse strain. Compared with other strains, NSG mice
are more sensitive to DNA damage. As such, I searched the literature to determine previously
published Carboplatin doses, administered specifically to immunocompromised mice. Because
no study utilized NSG mice, I performed an MTD assay to ascertain an appropriate drug dose
(Figure 2.0). Based on these results, I decided to use a conservative Carboplatin dose of 75
mg/kg/mouse, appreciating the fact that I could be under dosing the mice, and thereby having
less chemotherapeutic effect than expected. However, the results demonstrate that this dose was
appropriate to discriminate between platinum-resistant versus platinum-sensitive cases.
2.4.2 Assessment of Response
This study showed that PDXs predict response to platinum-therapy. However, in order to draw
this conclusion, several methods of assessing ‘response’ were evaluated. First, I attempted to
validate the use of PDX serum CA125 as a surrogate marker for chemo-responsiveness. Second,
I examined the use of BLI and other imaging modalities, and finally I examined actual tumour
size, which was possible due to the MFP tumour site. Ultimately, tumour size measurement
proved to be the most robust method.
In patients, CA125 levels at the time of diagnosis are of limited prognostic significance;
however, changes in CA125 in response to treatment have proven useful. Decreasing levels of
CA125 after cytoreductive surgery and during initial courses of chemotherapy have been used as
an indicator of clinical outcome (Gupta and Lis, 2009). Patients with advanced disease have a
significantly lower OS when CA125 is persistently elevated after three courses of neoadjuvant
chemotherapy (Van Dalen et al., 2000a; van Dalen et al., 2000b). Considering the success of
CA125 in determining response to chemotherapy in humans, I wanted to see if this method could
be used in PDXs.
In my assay, I used patient serum samples as positive controls (500, 50 and 5 U/mL), and these
mapped appropriately to the standard curve, suggesting that the ELISA was sensitive enough to
detect low CA125 levels. However, PDXs’ serum CA125 value was equivalent to that of the
negative control (un-injected NSG mice). Additionally, CA125 levels did not decrease with
96
chemotherapy administration, and there was no link to a clinically significant value. There are
multiple potential reasons for this discrepancy. It is possible that mouse factors do no allow the
protein to enter circulation (e.g. due to large protein size). Second, it could be that CA125
undergoes additional modifications in the mouse that render it undetectable by the ELISA assay.
Also, it could be it could be that the ectopic mammary fat pad model does not allow CA125 to be
cleaved into the serum and perhaps, an orthotopic model using the peritoneal cavity would
resolve this issue. Other groups have successfully quantified serum CA125 using IP PDXs
(unpublished results, Paul Haluska; Ronny Drapkin). Considering these data, and the fact that
PDXs tumours did show positive CA125 IHC staining, means that CA125 could have a role in
future models; however, for this project, CA125 could not be used as a corollary of platinum-
response.
The next step was to evaluate the ability to monitor tumour load longitudinally by using BLI.
This method provides sensitive, quantitative, and non-invasive detection of luciferase-positive
cells. I attempted to generate luciferase-expressing PDXs to track tumour growth and metastasis
(Appendix 1). Although these PDXs initially displayed a luciferase signal, ultimately, no tumour
was seen histologically. This method might become useful in monitoring disease progression and
response to treatment, but requires further development and optimization, and as such, could not
be used in this study.
Although using BLI was unsuccessful in my study, I also evaluated the utility of other imaging
modalities. Previously, the Neel laboratory assessed several sites for xenotransplantation, and
found that tumours were most readily and reliably detected in the MFP (Stewart, 2013).
Moreover, xenotransplantation into the ovarian bursa generally resulted in the development of
ascites, without formation of solid tumours (unpublished results, Jocelyn Stewart). To use an IB
PDX model and monitor disease via longitudinal measurements of ascites (by CT, MRI and/or
U/S) is not practical for many reasons. First, the ability to accurately quantify disease has not
been optimized at our institution. Furthermore, the requirement of an experienced operator,
combined with the high cost and low throughput of the technology decreases its practicality and
affordability.
97
These combined observations provided the rationale for using the MFP site, where tumor
measurements can easily be done with calipers. ‘Response’ was evaluated according to the
RECIST definition. These criteria were developed to assess the change in tumour burden
following the use of therapeutics, and is the considered the standard for assessing response in a
clinical trial setting. According to RECIST, a 30%, or greater, reduction in the target lesions
constitutes a partial response. Based on this, a reduction in PDXs tumour volume of >30% was
interpreted as ‘response’, whereas anything below 30% was characterized as ‘no response’ (i.e.,
progressive/stable disease).
2.4.3 Patients’ Clinical Outcomes
Clinical response to therapy in patients is typically assessed by changes in serum CA125 levels,
which are inexpensive, reproducible, and quantitative measurements. CA125 is elevated in ~90%
of ovarian HGSC, and in these patients, CA125 is the most useful serum marker for monitoring
response to chemotherapy and detecting disease recurrence (Bast et al., 2005). For these reasons,
I used serum CA125 levels as the primary mode of assessment of tumor response in patients
included in this study. One patient in the study (#3444) was a CA125 non-producer, and in this
case, imaging was used to determine response and to diagnose recurrence.
Standard definitions of platinum-resistance and sensitivity in ovarian cancer are based on time-
to-recurrence after completion of platinum treatment. This is a practical definition that has been
used to guide clinical management; however, results obtained in this project question this
definition. For example, there is a group of patients who have an immediate, complete decline in
CA125 in response to platinum therapy (primary and/or recurrent tumours). Riedinger et al.
showed that CA125 normalization after the first course of chemotherapy is an independent
prognostic factor for achieving a complete response and for improving OS, indicating that these
tumours are dramatically sensitive to platinum-based therapy (Riedinger et al., 2007). By
contrast, in another group, CA125 normalizes slowly and gradually over the course of
chemotherapy (corrected for tumour burden). If patients from both groups recur within 6 months
of completion of treatment, they are both treated uniformly and considered platinum-resistant.
The limitation of this definition is highlighted by one case (#6447) tested in my cohort.
Case #6447 showed a 29% reduction in tumour volume, which I classified as ‘no response’.
98
However, given the standard error, this represents a borderline partial response according to the
RECIST definition. Looking closely at the clinical course for this patient, she had a good
response to Carboplatin/Taxol therapy, but recurred 8 months following treatment. Because she
was classified as ‘platinum sensitive’ (>6 months to recurrence), she was re-treated with
platinum/Taxol, and again, showed a good response as demonstrated by a complete and
immediate normalization of her CA125. Six-months later, the patient had a rise in CA125 and
was treated with Gemcitabine, to which she did not respond. The PDXs’ response (29% tumour
reduction) suggests that this patient might have had some response to 3rd line Carboplatin/Taxol
treatment, and this could have been more effective than treatment with Gemcitabine.
Case 6447 highlights the challenges that clinicians face in the monitoring and treatment of
patients. The 6-month cutoff used to classifying patients as platinum-resistant or sensitive is
clearly imperfect and based on a few small, old studies (Gore et al., 1990; Markman et al.,
1991). Certainly, platinum refractory patients are unlikely to benefit from platinum therapy;
however, some patients characterized as ‘platinum-resistant’ might still display a clinical benefit
in receiving second-line platinum. Although this response might be small, it is likely comparable,
and perhaps, better than other second-line agents.
Another problem with the current definition of platinum-resistance is that it can often be related
to the frequency of clinical surveillance and not necessarily the biology of the tumour. For
example, two patients (A and B) might recur within 6 months of completion of therapy; Patient
A is seen at the time of recurrence (6 months) but Patient B’s recurrence is diagnosed at 8
months, because she is only seen every 4 months. Due to the arbitrary cut-off of 6 months to
classify platinum resistance, Patient A will likely receive second-line agents, whereas Patient B
will be re-treated with platinum (provided there are no contra-indications), and might have a
clinical outcome different from Patient A. Case 6447 illustrates this scenario; the patient’s first
recurrence was documented 8 months following the completion of chemotherapy. However, she
did not have any investigations (e.g., blood work, imaging) performed during those 8 months,
and her ‘real’ recurrence time could have been months earlier. Because she was diagnosed at 8
months, she was eligible to receive platinum therapy.
As platinum-based therapy is the most successful chemotherapy option for patients, the current
99
classification based on time-to-recurrence might be doing patients a disservice. The response to
platinum is likely to be unique to each recurrence and should be identified longitudinally
through the course of the patient’s disease, therefore providing patients with the chance to gain
some additional benefits from this agent. To return to the example of case 6447, according to the
clinical record, this patient’s 3rd recurrence (<6 months) was characterized as platinum-resistant.
However, my study demonstrates that the cells from 6447, when grown as PDXs, show
borderline-response. This suggests that this patient might have benefitted from a third course of
platinum therapy, despite being classified as platinum-resistant. In the context of this
study, PDXs derived from multiple recurrences of the same patient, or treated with several
rounds of platinum, might serve to assess the responsiveness of the tumour to platinum and other
second-line and/or investigational drugs, in order to determine which treatment(s) would be most
effective.
As experimental models, PDXs are not without limitations. First, they require ample amounts of
fresh tumour material. Typically, in ovarian cancer, this is not a constraint, but might be in the
context of patients who have received neoadjuvant chemotherapy. Second, PDXs are costly to
maintain compared to traditional cell lines. In addition, their genomic stability with increasing
passage remains to be defined. However, these limitations also apply to cell lines, which
although cheaper to maintain, have ultimately cost billions of dollars in failed drug development
projects. Unlike cell lines and cell-line derived xenografts, PDXs maintain the architecture of the
original tumour. Third, in addition, there is drift of stromal components from human to mouse,
genetic drift away from the primary tumour after numerous transplantations, loss of the immune
microenvironment, and difficulties with orthotopic transplantation (Sausville and Burger, 2006).
Last, PDXs’ engraftment rates are variable (~60% in this study) and those that engraft can take
up to 4-6 months (3-21 weeks in our study) to be ready for treatment. That time period can
correspond to recurrence time for some patients. Nevertheless, because multiple therapies,
including platinum, can be tested in parallel, PDXs could have the capacity to predict the next
most effective drug for recurrences.
The PFI is one of the strongest predictors of overall survival: the longer this interval lasts, the
better the response rate to subsequent chemotherapy (Huang et al., 2012). The ability of PDXs to
predict a long versus short PFI could have clinical implications, particularly for patients who are
100
likely to have a short PFI. Patients with short PFIs (i.e., platinum refractoriness/resistance)
typically have poor response rates to subsequent therapy and these are the patients whose
treatment options are very limited. Identifying these patients early in the course of their disease
can potentially alter their clinical course. First, it could determine how frequently these patients
are monitored. Currently, there is no standard protocol for how to assess patients once they have
completed initial therapy. Many clinicians monitor their patients every 3-6 months, but this can
only capture advanced recurrent disease. Because PDXs provide evidence regarding patients’
chemo-responsiveness, it may be beneficial to monitor patients identified as platinum-resistant
(based on their respective PDXs) more closely and rigorously for evidence of disease recurrence.
Second, PDXs could allow novel drugs and chemotherapeutic options to be evaluated for each
patient. Platinum-resistant patients require targeted or novel therapies for their disease, and
PDXs might resolve which therapies are best suited for each specific patient. Current knowledge
regarding the clinical management of ovarian cancer suggests that targeted agents must be used
in combination to select the right drugs for the right patient at the right time (Romero and Bast,
2012). ‘Personalized cancer therapy’ rather than ‘one size fits all’ is required to increase survival
in ovarian cancer patients, and PDXs may serve to resolve this issue.
2.4.4 Tumour-Initiating Cells
The majority of patients with ovarian HGSC have resistant and/or recurrent disease, and as such,
interest into whether CSCs are responsible for drug-resistance in ovarian cancer has grown. In
this scenario, although bulk tumour shrinkage occurs with chemotherapeutic treatment,
recurrence and metastasis arise from failure of chemotherapy to eradicate CSCs (O'Brien et al.,
2009b). Several studies examining other cancer sites, have reported that current
chemotherapeutics are unable to target CSCs; consequently, treatment with chemotherapy might
enrich for tumourigenic cells.
Because I was using a chemotherapy-based PDX model, it seemed reasonable to examine TIC
frequency following the chemotherapy-response experiments. Specifically, I wanted to address
whether the pool of TICs was enriched following chemotherapy compared to untreated tumours.
Unfortunately, partly due to small sample size, my results yielded inconclusive data: one case
enriched for TICs following chemotherapy, three cases gave statistically insignificant data, and
101
one case enriched for TICs in the control group. If I assume that ovarian HGSC conforms to the
CSC model, important limitations and considerations obtained from these experiments could
improve future projects of this kind.
Some limitations of this study arise in the methods. For successful LDA studies, total cell
number is important and inadequate numbers may explain my inconclusive results. At the
completion of chemotherapy treatment, many cases had an insufficient number of viable cells to
perform an LDA. Consequently, I could only use samples that generated many PDXs
simultaneously (in order to pool all tumour cells) or were relatively refractory to platinum
treatment (so that tumours at the completion of treatment had high cell numbers). A
consideration for future work would be to interrogate the effect of various smaller doses of
chemotherapy over time, as opposed to the administering the near-MTD that I utilized. Perhaps
using a lower chemotherapy dose (or one dose) and sacrificing mice at earlier, incremental time
periods, would still allow for a gradual enrichment of TICs. Doing so would allow more PDXs to
be generated per cell dose (this study had 4 PDXs/dose for the majority of cases), which was a
major limitation. For this assay to be robust, the experiment requires 8-12 PDXs per cell dose, at
a minimum.
TICs are relatively rare in many solid tumours and this number can be misestimated depending
on the nature of the PDX-assay. Tumourigenicity assays performed using NSG mice (compared
with NOD/SCID) and/or heavily passaged PDXs considerably increase the TIC frequency in
some cancers (Ishizawa et al., 2010; Magee et al., 2012). In ovarian HGSC, the Neel laboratory
found that the ability to enrich for TICs deteriorated upon PDXs passage. They also found that
TIC frequency was at most ~10-fold higher in NSG versus NOD/SCID mice, and even then, it
represented ≤0.02% of the CD45-depleted ovarian HGSC cells. In addition, they showed that the
median TIC frequency in ascites was ~1/10,000 and ~1/90,000 for solid tumours, but that this
ranged considerably between various tumours (Stewart, 2013; Stewart et al., 2011). Given this
tremendous range, it would be useful to know in advance the approximate TIC frequency to
ensure that the appropriate cell doses are used. This would avoid scenarios, such as those seen in
my study, where cell doses are too small or too large to provide meaningful results.
102
In this study, as in others that use PDXs, one cannot exclude the possibility that the mouse
microenvironment causes a misestimation of the TIC frequency. Incompatibilities between
mouse ligands and human receptors may impair survival and proliferation of human cells. In
addition, because the human cells are ectopically transplanted into mice, the differences in the
environment may modify the engraftment of cells and tumourigenic potential (Shackleton et al.,
2009). Tumour stroma has been shown to be critical in the development of a TIC niche, where
TICs can be maintained and allowed to proliferate and future experiments in this area should
explore the role stromal compartments play in maintaining tumour growth.
Furthermore, it would be extremely valuable to identify a marker (or marker combination) of
TICs. The CSC model has been carefully tested in a small subset of cancers and is often assumed
to apply widely to other cancer types (Shackleton et al., 2009). In conjunction with functional
assays, cell markers could help in understanding tumourigenesis, particularly in the context of
post-treatment with chemotherapy. Specifically, provided the marker expression was not affected
by chemotherapy (an experimental variable that would have to be tested), one could assay TIC
markers in the post-chemotreated PDXs, providing a correlative surrogate for TIC enrichment
after chemotherapy.
Despite the limitations described, variability in these results might be due to the known inter-
patient heterogeneity of this disease. It was shown by the TCGA (The Cancer Genome Atlas
Research Network, 2011) that patient tumours could exhibit expression profiles that reflect
different subtypes of HGSC. It is possible that the PDXs that I used in this study might fall into
different subtypes and that each subtype might have a differential sensitivity of TIC to
chemotherapy. In addition, enrichment in the control group might simply reflect the healthiness
of the tissues after treatment. It is possible that, despite cells “scoring” as alive by trypan blue
exclusion, the chemotherapy-treated cells are actually much less healthy than the control group.
Again, this might be PDXs-dependent.
Finally, the lack of significant enrichment of TIC after treatment of some cases might suggest
that conventional chemotherapy equally targets tumour-initiating and non-initiating populations
– this might suggest that there is no correlation between the treatment-resistant populations of
cells and the TIC population. Alternatively, it could suggest that, given the proposed level of
103
clonal evolution at the time of diagnosis (Bashashati et al., 2013), there are multiple populations
(at similar frequencies) that have tumour-initiating ability, but that are differentially resistant to
chemotherapy. Consistent with these data, no study has shown that a prospectively
phenotypically identifiable population of cells has the unique ability to initiate HGSC. Based on
my limited results, I could not make any conclusions about the nature of TIC dynamics in
ovarian HGSC PDXs in response to chemotherapy.
Reliable, predictive models are required when a treatment decision must be made, which is of
greatest importance for resistant and/or recurrent patients. I conclude that ovarian high-grade
serous patient-derived PDXs closely reflect the histology and platinum sensitivity of original
donor tissue. This newly validated model represents an effective tool for the identification of
platinum-resistant and platinum-sensitive patients. Furthermore, this model may serve as a
predictive tool for the investigation and development of novel and targeted therapeutics.
104
Chapter 3:
Conclusions & Future Directions
Ovarian HGSC is the leading cause of mortality amongst all gynecologic malignancies. Due to a
lack of specific symptoms, patients typically present with advanced stage disease. Treatment for
all patients follows a similar formula: all patients receive debulking surgery and platinum-based
chemotherapy. Most patients have a good response to initial therapy, but nearly all recur. If the
recurrence occurs within 6 months of completion of initial therapy or during the course of initial
therapy, these patients are classified as ‘platinum-resistant’ or ‘platinum-refractory’,
respectively. Second-line treatment for patients with platinum resistant/refractory disease is
expected to produce low levels of response and is aimed at extending the symptom-free interval
and improving quality of life (Fung Kee Fung, 2011). In these situations, patients are offered
non-platinum-based mono-therapy (often with paclitaxel, topotecan, doxorubicin or gemcitabine)
and the opportunity to participate in clinical trials.
Patients that recur more than 6 months after completing initial therapy are considered ‘platinum
sensitive’. Upon recurrence, they are offered re-treatment with Carboplatin/Taxol (provided there
are no contraindications), or platinum in combination with another drug (usually gemcitabine or
liposomal doxorubicin). Response to re-treatment with platinum is dependent on the length of the
PFI: the longer the interval, the greater the likelihood of response to second-line platinum
therapy (Huang et al., 2012).
This treatment ‘protocol’ and the arbitrary use of six months to distinguish platinum-resistant
from platinum-sensitive patients, has resulted in a stagnant OS of ~30% since the 1970s. To date,
we have no methods to identify platinum-sensitive patients from those that are platinum
resistant/refractory. Furthermore, despite strategies that have included the use of
chemosensitivity assays, we are unable to optimize second-line treatment, as we cannot predict
which second-line therapy will yield a response in recurrent disease. Early identification of
platinum-resistant/refractory patients and the personalization of cancer treatment might improve
survival in ovarian HGSC patients.
105
The value and utility of any in vivo cancer model rests on its capacity to accurately reflect the
properties of the disease. The aim of my study was to establish a pre-clinical model of ovarian
high-grade serous cancer that maintains a high correlation with the characteristics of the clinical
disease. The Neel laboratory previously established conditions to generate PDXs form primary
tumours and ascites with high efficiency. These PDXs recapitulate the primary tumours with
respect to their histological appearance and expression of markers by flow cytometry and
immunohistochemistry (Stewart, 2013; Stewart et al., 2011).
To further validate these PDXs as clinically relevant pre-clinical models of ovarian HGSC, I
assessed their response to Carboplatin treatment (Chapter 2). To do so, I retrospectively
identified platinum-resistant and platinum-sensitive ovarian HGSC patients, engrafted these into
NSG mice, and tested the PDXs’ response to platinum therapy. Using the RECIST criteria to
define response, PDXs derived from platinum resistant patients showed no response (i.e. stable
or progressive disease) in response to Carboplatin treatment. Conversely, PDXs derived from
platinum sensitive patients showed a striking response to platinum therapy, as seen by marked
reductions in tumour volumes.
Furthermore, I generated PDXs from ‘prospective’ patients (n=2). These were patients from
whom I generated PDXs from primary tumours obtained at the time of their primary surgery,
when their clinical course was yet unknown. I followed these patients prospectively, predicting
that based on the PDXs’ response to platinum therapy, these patients would be sensitive to
platinum treatment corresponding to a PFI > 6 months. Remarkably, both patients are platinum-
sensitive, with PFI’s >12 months and neither has yet recurred. These findings support the
capacity of the PDX model to predict the patient’s clinical course.
Interestingly, following treatment and elimination of palpable tumours, one PDX case re-grew
tumours 3.5 months after completion of therapy. Upon re-treatment with platinum, this case
proved to be, once again, sensitive to platinum, paralleling the patient’s response. Although I
only have one case from which to draw conclusions, this may suggest that not only can PDXs
predict initial response to platinum treatment, but also upon re-treatment, can further predict how
the patient’s recurrence will respond to second-line platinum therapy. This finding highlights
that perhaps following treatment of PDXs, we are selecting for cellular clones that are clinically
106
relevant to the patient’s recurrent disease.
My findings suggest that despite the tremendous cellular heterogeneity of the primary tumour
and the diverse clinical course of each patient, patient-derived PDXs reproduce patient response
to platinum therapy. Combined with the PDX profiling performed by Stewart, this PDX system
has proven to be a valid pre-clinical model that may be useful for defining personalized
approaches to ovarian HGSC therapy (Stewart, 2013; Stewart et al., 2011).
Along these lines, I explored whether PDXs could be used to investigate combination drug
studies. To do so, I tested the effects of Carboplatin in combination with Disulfiram (DSF), a
drug used to treat alcohol abuse, which has shown anti-cancer effects in several cancer models
(Appendix 2). Although exact drug doses require optimization, mice treated with both DSF and
Carboplatin show a trend toward reduced tumor growth rates. Repurposing DSF for ovarian
cancer therapy might represent a novel method to enhance the cytotoxic effects of platinum
therapy. Future work should focus on elucidating the possible mechanisms responsible for DSF’s
effects.
To examine the role of tumourigenicity following chemotherapy, I used PDXs to assess the
nature of TICs following platinum-treatment. As mentioned previously, over 85% of patients
will recur following initial therapy and often, the recurrences tend to be platinum-resistant.
Understanding how chemoresistant tumour cells develop and potentially cause disease
resurgence could provide a strategy for modification of the current therapeutic management of
both primary and recurrent ovarian cancer. A possible explanation for recurrence is that
following treatment, there are remaining cancer cells, which comprise a mixed tumour cell
population. Within this mixed population is hypothesized to exist a distinct subpopulation of
cells (CSCs) that escape the effects of chemotherapy, and as a result, treatment with
chemotherapy is thought to actually enrich for tumourigenic cells.
Limited results do not allow me to derive conclusions about the nature of TIC dynamics in
ovarian high-grade serous cancer PDXs following chemotherapy. Despite the limitations of the
assay (described in Chapter 2), variability in these results might be due to the known inter-patient
heterogeneity of this disease or could suggest that conventional chemotherapy equally targets
tumour-initiating and non-initiating cells. There might even be multiple populations with
107
tumour-initiating abilities (at similar frequencies) that are differentially resistant to
chemotherapy. Future work should focus on identification and isolation of ovarian HGSC cells
with tumourigenic capacity in order to determine their responses to therapy.
Overall this thesis demonstrated that PDXs are a useful, pre-clinical model of ovarian HGSC
with the capacity to predict response to platinum therapy. Ideally, in the future, every patient
with ovarian HGSC should have a cohort of PDXs created at the time of diagnosis. This cohort
would enable: 1) assessment of the response of the tumour to platinum therapy (i.e. predict
platinum-resistance or sensitivity); 2) assessment of the response of the tumour to additional
targeted and/or novel therapeutics; and 3) assessment of the response of the recurrent tumour to
re-treatment with platinum and/or targeted therapeutics. This model also allows the assessment
of TIC dynamics and drug-combination studies. Together, my findings positively impact ovarian
cancer research and ultimately, the patients suffering from this disease.
108
References Abraham, J., Balbo, S., Crabb, D., and Brooks, P.J. (2011). Alcohol metabolism in human cells causes DNA damage and activates the Fanconi anemia-breast cancer susceptibility (FA-BRCA) DNA damage response network. Alcoholism, clinical and experimental research 35, 2113-2120.
ACOG (2009). ACOG Practice Bulletin No. 103: Hereditary breast and ovarian cancer syndrome. Obstetrics and gynecology 113, 957-966.
Adami, H.O., Hsieh, C.C., Lambe, M., Trichopoulos, D., Leon, D., Persson, I., Ekbom, A., and Janson, P.O. (1994). Parity, age at first childbirth, and risk of ovarian cancer. Lancet 344, 1250-1254.
Agarwal, R., and Kaye, S.B. (2003). Ovarian cancer: strategies for overcoming resistance to chemotherapy. Nature reviews Cancer 3, 502-516.
Aggarwal, P., and Kehoe, S. (2010). Serum tumour markers in gynaecological cancers. Maturitas 67, 46-53.
Alberts, D.S. (1995). Carboplatin versus cisplatin in ovarian cancer. Seminars in oncology 22, 88-90.
Alberts, D.S., and Dorr, R.T. (1998). New Perspectives on an Old Friend: Optimizing Carboplatin for the Treatment of Solid Tumors. The oncologist 3, 15-34.
Albrektsen, G., Heuch, I., and Kvale, G. (1996). Reproductive factors and incidence of epithelial ovarian cancer: a Norwegian prospective study. Cancer causes & control : CCC 7, 421-427.
Allen, D.G., Heintz, A.P., and Touw, F.W. (1995). A meta-analysis of residual disease and survival in stage III and IV carcinoma of the ovary. European journal of gynaecological oncology 16, 349-356.
Alvero, A.B., Chen, R., Fu, H.H., Montagna, M., Schwartz, P.E., Rutherford, T., Silasi, D.A., Steffensen, K.D., Waldstrom, M., Visintin, I., et al. (2009a). Molecular phenotyping of human ovarian cancer stem cells unravels the mechanisms for repair and chemoresistance. Cell Cycle 8, 158-166.
Alvero, A.B., Fu, H.H., Holmberg, J., Visintin, I., Mor, L., Marquina, C.C., Oidtman, J., Silasi, D.A., and Mor, G. (2009b). Stem-like ovarian cancer cells can serve as tumor vascular progenitors. Stem Cells 27, 2405-2413.
Anderson, K., Lutz, C., van Delft, F.W., Bateman, C.M., Guo, Y., Colman, S.M., Kempski, H., Moorman, A.V., Titley, I., Swansbury, J., et al. (2011). Genetic variegation of clonal architecture and propagating cells in leukaemia. Nature 469, 356-361.
Andreotti, P.E., Cree, I.A., Kurbacher, C.M., Hartmann, D.M., Linder, D., Harel, G., Gleiberman, I., Caruso, P.A., Ricks, S.H., Untch, M., et al. (1995). Chemosensitivity testing of
109
human tumors using a microplate adenosine triphosphate luminescence assay: clinical correlation for cisplatin resistance of ovarian carcinoma. Cancer research 55, 5276-5282.
Aparicio, S., and Caldas, C. (2013). The implications of clonal genome evolution for cancer medicine. The New England journal of medicine 368, 842-851.
Armstrong, D.K., Bundy, B., Wenzel, L., Huang, H.Q., Baergen, R., Lele, S., Copeland, L.J., Walker, J.L., and Burger, R.A. (2006). Intraperitoneal cisplatin and paclitaxel in ovarian cancer. The New England journal of medicine 354, 34-43.
Auersperg, N. (2013a). The origin of ovarian cancers--hypotheses and controversies. Front Biosci (Schol Ed) 5, 709-719.
Auersperg, N. (2013b). The origin of ovarian carcinomas: A developmental view. Gynecologic Oncology 130, 452-454.
Auersperg, N. (2013c). Ovarian surface epithelium as a source of ovarian cancers: unwarranted speculation or evidence-based hypothesis? Gynecol Oncol 130, 246-251.
Axtell, A.E., Lee, M.H., Bristow, R.E., Dowdy, S.C., Cliby, W.A., Raman, S., Weaver, J.P., Gabbay, M., Ngo, M., Lentz, S., et al. (2007). Multi-institutional reciprocal validation study of computed tomography predictors of suboptimal primary cytoreduction in patients with advanced ovarian cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25, 384-389.
Bafna, S., Kaur, S., and Batra, S.K. (2010). Membrane-bound mucins: the mechanistic basis for alterations in the growth and survival of cancer cells. Oncogene 29, 2893-2904.
Bai, L.Z., W.G. (2006). p53: Structure, Function and Therapeutic Application. Journal of Cancer Molecules 2, 141-153.
Bakhru, A., Mallinger, J.B., Buckanovich, R.J., and Griggs, J.J. (2010). Casting light on 25-hydroxyvitamin D deficiency in ovarian cancer: a study from the NHANES. Gynecol Oncol 119, 314-318.
Bandera, C.A. (2005). Advances in the understanding of risk factors for ovarian cancer. The Journal of reproductive medicine 50, 399-406.
Bankert, R.B., Balu-Iyer, S.V., Odunsi, K., Shultz, L.D., Kelleher, R.J., Jr., Barnas, J.L., Simpson-Abelson, M., Parsons, R., and Yokota, S.J. (2011). Humanized mouse model of ovarian cancer recapitulates patient solid tumor progression, ascites formation, and metastasis. PloS one 6, e24420.
Banks, E. (2001). The epidemiology of ovarian cancer. Methods in molecular medicine 39, 3-11.
Bao, S., Wu, Q., McLendon, R.E., Hao, Y., Shi, Q., Hjelmeland, A.B., Dewhirst, M.W., Bigner, D.D., and Rich, J.N. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444, 756-760.
110
Bapat, S.A. (2010). Human ovarian cancer stem cells. Reproduction 140, 33-41.
Bapat, S.A., Mali, A.M., Koppikar, C.B., and Kurrey, N.K. (2005). Stem and progenitor-like cells contribute to the aggressive behavior of human epithelial ovarian cancer. Cancer research 65, 3025-3029.
Bashashati, A., Ha, G., Tone, A., Ding, J., Prentice, L.M., Roth, A., Rosner, J., Shumansky, K., Kalloger, S., Senz, J., et al. (2013). Distinct evolutionary trajectories of primary high-grade serous ovarian cancers revealed through spatial mutational profiling. The Journal of pathology 231, 21-34.
Bast, R.C., Jr., Badgwell, D., Lu, Z., Marquez, R., Rosen, D., Liu, J., Baggerly, K.A., Atkinson, E.N., Skates, S., Zhang, Z., et al. (2005). New tumor markers: CA125 and beyond. International journal of gynecological cancer : official journal of the International Gynecological Cancer Society 15 Suppl 3, 274-281.
Bast, R.C., Jr., Hennessy, B., and Mills, G.B. (2009). The biology of ovarian cancer: new opportunities for translation. Nature reviews Cancer 9, 415-428.
Bast, R.C., Jr., Xu, F.J., Yu, Y.H., Barnhill, S., Zhang, Z., and Mills, G.B. (1998). CA 125: the past and the future. The International journal of biological markers 13, 179-187.
Bast, R.C., Thigpen, J.T., Arbuck, S.G., Basen-Engquist, K., Burke, L.B., Freedman, R., Horning, S.J., Ozols, R., Rustin, G.J., Spriggs, D., et al. (2007). Clinical trial endpoints in ovarian cancer: report of an FDA/ASCO/AACR Public Workshop. Gynecol Oncol 107, 173-176.
Bastide, C., Bagnis, C., Mannoni, P., Hassoun, J., and Bladou, F. (2002). A Nod Scid mouse model to study human prostate cancer. Prostate cancer and prostatic diseases 5, 311-315.
Behringer, R.R. (1994). The in vivo roles of mullerian-inhibiting substance. Current topics in developmental biology 29, 171-187.
Behringer, R.R., Finegold, M.J., and Cate, R.L. (1994). Mullerian-inhibiting substance function during mammalian sexual development. Cell 79, 415-425.
Beral, V., Bull, D., Green, J., and Reeves, G. (2007). Ovarian cancer and hormone replacement therapy in the Million Women Study. Lancet 369, 1703-1710.
Beral, V., Fraser, P., and Chilvers, C. (1978). Does pregnancy protect against ovarian cancer? Lancet 1, 1083-1087.
Bernardini, M., Lee, C.H., Beheshti, B., Prasad, M., Albert, M., Marrano, P., Begley, H., Shaw, P., Covens, A., Murphy, J., et al. (2005). High-resolution mapping of genomic imbalance and identification of gene expression profiles associated with differential chemotherapy response in serous epithelial ovarian cancer. Neoplasia 7, 603-613.
Berns, E.M., and Bowtell, D.D. (2012). The changing view of high-grade serous ovarian cancer. Cancer research 72, 2701-2704.
111
Bieber, E.J.H., I.R.; Sanfilippo, J.S. (2006). Clinical Gynecology (Northwestern University: Churchill LIvinstone).
Boivin, M., Lane, D., Piche, A., and Rancourt, C. (2009). CA125 (MUC16) tumor antigen selectively modulates the sensitivity of ovarian cancer cells to genotoxic drug-induced apoptosis. Gynecol Oncol 115, 407-413.
Bonello, N., McKie, K., Jasper, M., Andrew, L., Ross, N., Braybon, E., Brannstrom, M., and Norman, R.J. (1996). Inhibition of nitric oxide: effects on interleukin-1 beta-enhanced ovulation rate, steroid hormones, and ovarian leukocyte distribution at ovulation in the rat. Biology of reproduction 54, 436-445.
Bosetti, C., Negri, E., Trichopoulos, D., Franceschi, S., Beral, V., Tzonou, A., Parazzini, F., Greggi, S., and La Vecchia, C. (2002). Long-term effects of oral contraceptives on ovarian cancer risk. International journal of cancer Journal international du cancer 102, 262-265.
Bowtell, D.D. (2010). The genesis and evolution of high-grade serous ovarian cancer. Nature reviews Cancer 10, 803-808.
Boyd, J. (2003). Specific keynote: hereditary ovarian cancer: what we know. Gynecol Oncol 88, S8-10; discussion S11-13.
Boyd, J., and Rubin, S.C. (1997). Hereditary ovarian cancer: molecular genetics and clinical implications. Gynecol Oncol 64, 196-206.
Brar, S.S., Grigg, C., Wilson, K.S., Holder, W.D., Jr., Dreau, D., Austin, C., Foster, M., Ghio, A.J., Whorton, A.R., Stowell, G.W., et al. (2004). Disulfiram inhibits activating transcription factor/cyclic AMP-responsive element binding protein and human melanoma growth in a metal-dependent manner in vitro, in mice and in a patient with metastatic disease. Molecular cancer therapeutics 3, 1049-1060.
Bristow, R.E., Tomacruz, R.S., Armstrong, D.K., Trimble, E.L., and Montz, F.J. (2002). Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: a meta-analysis. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 20, 1248-1259.
Brooks, P.J., Enoch, M.A., Goldman, D., Li, T.K., and Yokoyama, A. (2009). The alcohol flushing response: an unrecognized risk factor for esophageal cancer from alcohol consumption. PLoS medicine 6, e50.
Brutkiewicz, S., Mendonca, M., Stantz, K., Comerford, K., Bigsby, R., Hutchins, G., Goebl, M., and Harrington, M. (2007). The expression level of luciferase within tumour cells can alter tumour growth upon in vivo bioluminescence imaging. Luminescence : the journal of biological and chemical luminescence 22, 221-228.
Burgos-Ojeda, D., Rueda, B.R., and Buckanovich, R.J. (2012). Ovarian cancer stem cell markers: prognostic and therapeutic implications. Cancer letters 322, 1-7.
112
Buttitta, F., Marchetti, A., Gadducci, A., Pellegrini, S., Morganti, M., Carnicelli, V., Cosio, S., Gagetti, O., Genazzani, A.R., and Bevilacqua, G. (1997). p53 alterations are predictive of chemoresistance and aggressiveness in ovarian carcinomas: a molecular and immunohistochemical study. British journal of cancer 75, 230-235.
Buys, S.S., Partridge, E., Greene, M.H., Prorok, P.C., Reding, D., Riley, T.L., Hartge, P., Fagerstrom, R.M., Ragard, L.R., Chia, D., et al. (2005). Ovarian cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial: findings from the initial screen of a randomized trial. American journal of obstetrics and gynecology 193, 1630-1639.
Calvert, A.H., Newell, D.R., Gumbrell, L.A., O'Reilly, S., Burnell, M., Boxall, F.E., Siddik, Z.H., Judson, I.R., Gore, M.E., and Wiltshaw, E. (1989). Carboplatin dosage: prospective evaluation of a simple formula based on renal function. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 7, 1748-1756.
Campbell, P.J., Yachida, S., Mudie, L.J., Stephens, P.J., Pleasance, E.D., Stebbings, L.A., Morsberger, L.A., Latimer, C., McLaren, S., Lin, M.L., et al. (2010). The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 467, 1109-1113.
Canadian Cancer Society. (2012). Canadian Cancer Society's Steering Committee on Cancer Statistics. In Canadian Cancer Statistics 2012 (Toronto, ON: Canadian Cancer Society).
Cannistra, S.A. (2004). Cancer of the ovary. The New England journal of medicine 351, 2519-2529.
Castellarin, M., Milne, K., Zeng, T., Tse, K., Mayo, M., Zhao, Y., Webb, J.R., Watson, P.H., Nelson, B.H., and Holt, R.A. (2013). Clonal evolution of high-grade serous ovarian carcinoma from primary to recurrent disease. The Journal of pathology 229, 515-524.
Cen, D., Gonzalez, R.I., Buckmeier, J.A., Kahlon, R.S., Tohidian, N.B., and Meyskens, F.L., Jr. (2002). Disulfiram induces apoptosis in human melanoma cells: a redox-related process. Molecular cancer therapeutics 1, 197-204.
Cespedes, M.V., Casanova, I., Parreno, M., and Mangues, R. (2006). Mouse models in oncogenesis and cancer therapy. Clinical & translational oncology : official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico 8, 318-329.
Chen, D., Cui, Q.C., Yang, H., and Dou, Q.P. (2006). Disulfiram, a clinically used anti-alcoholism drug and copper-binding agent, induces apoptotic cell death in breast cancer cultures and xenografts via inhibition of the proteasome activity. Cancer research 66, 10425-10433.
Chiaffarino, F., Parazzini, F., La Vecchia, C., Marsico, S., Surace, M., and Ricci, E. (1999). Use of oral contraceptives and uterine fibroids: results from a case-control study. British journal of obstetrics and gynaecology 106, 857-860.
Chiaffarino, F., Pelucchi, C., Parazzini, F., Negri, E., Franceschi, S., Talamini, R., Conti, E., Montella, M., and La Vecchia, C. (2001). Reproductive and hormonal factors and ovarian
113
cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 12, 337-341.
Chong, C.R., and Sullivan, D.J., Jr. (2007). New uses for old drugs. Nature 448, 645-646.
Chou, T.C., and Talalay, P. (1984). Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Advances in enzyme regulation 22, 27-55.
Christie, M., and Oehler, M.K. (2006). Molecular pathology of epithelial ovarian cancer. The journal of the British Menopause Society 12, 57-63.
Clark-Knowles, K.V., Garson, K., Jonkers, J., and Vanderhyden, B.C. (2007). Conditional inactivation of Brca1 in the mouse ovarian surface epithelium results in an increase in preneoplastic changes. Experimental cell research 313, 133-145.
Clark-Knowles, K.V., Senterman, M.K., Collins, O., and Vanderhyden, B.C. (2009). Conditional inactivation of Brca1, p53 and Rb in mouse ovaries results in the development of leiomyosarcomas. PloS one 4, e8534.
Clegg, L.X., Li, F.P., Hankey, B.F., Chu, K., and Edwards, B.K. (2002). Cancer survival among US whites and minorities: a SEER (Surveillance, Epidemiology, and End Results) Program population-based study. Archives of internal medicine 162, 1985-1993.
Clendenen, T.V., Lundin, E., Zeleniuch-Jacquotte, A., Koenig, K.L., Berrino, F., Lukanova, A., Lokshin, A.E., Idahl, A., Ohlson, N., Hallmans, G., et al. (2011). Circulating inflammation markers and risk of epithelial ovarian cancer. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 20, 799-810.
Cordero, A.B., Kwon, Y., Hua, X., and Godwin, A.K. (2010). In vivo imaging and therapeutic treatments in an orthotopic mouse model of ovarian cancer. Journal of visualized experiments : JoVE.
Covens, A.L. (2000). A critique of surgical cytoreduction in advanced ovarian cancer. Gynecol Oncol 78, 269-274.
Cramer, D.W., and Welch, W.R. (1983). Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. Journal of the National Cancer Institute 71, 717-721.
Crispens, M.A. (2012). Endometrial and ovarian cancer in lynch syndrome. Clinics in colon and rectal surgery 25, 97-102.
Crow, J., Amso, N.N., Lewin, J., and Shaw, R.W. (1994). Morphology and ultrastructure of fallopian tube epithelium at different stages of the menstrual cycle and menopause. Hum Reprod 9, 2224-2233.
114
Crul, M., van Waardenburg, R.C., Bocxe, S., van Eijndhoven, M.A., Pluim, D., Beijnen, J.H., and Schellens, J.H. (2003). DNA repair mechanisms involved in gemcitabine cytotoxicity and in the interaction between gemcitabine and cisplatin. Biochemical pharmacology 65, 275-282.
Crum, C.P., Drapkin, R., Kindelberger, D., Medeiros, F., Miron, A., and Lee, Y. (2007a). Lessons from BRCA: the tubal fimbria emerges as an origin for pelvic serous cancer. Clinical medicine & research 5, 35-44.
Crum, C.P., Drapkin, R., Miron, A., Ince, T.A., Muto, M., Kindelberger, D.W., and Lee, Y. (2007b). The distal fallopian tube: a new model for pelvic serous carcinogenesis. Current opinion in obstetrics & gynecology 19, 3-9.
Cui, K., Xu, X., Zhao, H., and Wong, S.T. (2008). A quantitative study of factors affecting in vivo bioluminescence imaging. Luminescence : the journal of biological and chemical luminescence 23, 292-295.
Curley, M.D., Therrien, V.A., Cummings, C.L., Sergent, P.A., Koulouris, C.R., Friel, A.M., Roberts, D.J., Seiden, M.V., Scadden, D.T., Rueda, B.R., et al. (2009). CD133 expression defines a tumor initiating cell population in primary human ovarian cancer. Stem Cells 27, 2875-2883.
D'Andrea, A.D. (2010). Susceptibility pathways in Fanconi's anemia and breast cancer. The New England journal of medicine 362, 1909-1919.
Dalerba, P., Dylla, S.J., Park, I.K., Liu, R., Wang, X., Cho, R.W., Hoey, T., Gurney, A., Huang, E.H., Simeone, D.M., et al. (2007). Phenotypic characterization of human colorectal cancer stem cells. Proceedings of the National Academy of Sciences of the United States of America 104, 10158-10163.
Daniel, K.G., Chen, D., Orlu, S., Cui, Q.C., Miller, F.R., and Dou, Q.P. (2005). Clioquinol and pyrrolidine dithiocarbamate complex with copper to form proteasome inhibitors and apoptosis inducers in human breast cancer cells. Breast cancer research : BCR 7, R897-908.
Daniel, V.C., Marchionni, L., Hierman, J.S., Rhodes, J.T., Devereux, W.L., Rudin, C.M., Yung, R., Parmigiani, G., Dorsch, M., Peacock, C.D., et al. (2009). A primary xenograft model of small-cell lung cancer reveals irreversible changes in gene expression imposed by culture in vitro. Cancer research 69, 3364-3373.
Das, B., Antoon, R., Tsuchida, R., Lotfi, S., Morozova, O., Farhat, W., Malkin, D., Koren, G., Yeger, H., and Baruchel, S. (2008). Squalene selectively protects mouse bone marrow progenitors against cisplatin and carboplatin-induced cytotoxicity in vivo without protecting tumor growth. Neoplasia 10, 1105-1119.
DeCherney, A.H., and Nathan, L. (2007). Current Diagnosis & Treatment Obstetrics & Gynecology, Tenth edn (McGraw-Hill).
115
Del Campo, J.M., Felip, E., Rubio, D., Vidal, R., Bermejo, B., Colomer, R., and Zanon, V. (1994). Long-term survival in advanced ovarian cancer after cytoreduction and chemotherapy treatment. Gynecol Oncol 53, 27-32.
Delair, D., and Soslow, R.A. (2012). Key features of extrauterine pelvic serous tumours (fallopian tube, ovary, and peritoneum). Histopathology 61, 329-339.
Deng, C.X. (2006). BRCA1: cell cycle checkpoint, genetic instability, DNA damage response and cancer evolution. Nucleic acids research 34, 1416-1426.
Diaz-Padilla, I., Razak, A.R., Minig, L., Bernardini, M.Q., and Maria Del Campo, J. (2012). Prognostic and predictive value of CA-125 in the primary treatment of epithelial ovarian cancer: potentials and pitfalls. Clinical & translational oncology : official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico 14, 15-20.
Ding, L., Ley, T.J., Larson, D.E., Miller, C.A., Koboldt, D.C., Welch, J.S., Ritchey, J.K., Young, M.A., Lamprecht, T., McLellan, M.D., et al. (2012). Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature 481, 506-510.
Domcke, S., Sinha, R., Levine, D.A., Sander, C., and Schultz, N. (2013). Evaluating cell lines as tumour models by comparison of genomic profiles. Nature communications 4, 2126.
Dreher, D., and Junod, A.F. (1996). Role of oxygen free radicals in cancer development. Eur J Cancer 32A, 30-38.
Duffy, M.J., Bonfrer, J.M., Kulpa, J., Rustin, G.J., Soletormos, G., Torre, G.C., Tuxen, M.K., and Zwirner, M. (2005). CA125 in ovarian cancer: European Group on Tumor Markers guidelines for clinical use. International journal of gynecological cancer : official journal of the International Gynecological Cancer Society 15, 679-691.
Dyall, S., Gayther, S.A., and Dafou, D. (2010). Cancer stem cells and epithelial ovarian cancer. Journal of oncology 2010, 105269.
Dylla, S.J., Beviglia, L., Park, I.K., Chartier, C., Raval, J., Ngan, L., Pickell, K., Aguilar, J., Lazetic, S., Smith-Berdan, S., et al. (2008). Colorectal cancer stem cells are enriched in xenogeneic tumors following chemotherapy. PloS one 3, e2428.
Easton, D.F., Ford, D., and Bishop, D.T. (1995). Breast and ovarian cancer incidence in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. American journal of human genetics 56, 265-271.
Edinger, M., Cao, Y.A., Hornig, Y.S., Jenkins, D.E., Verneris, M.R., Bachmann, M.H., Negrin, R.S., and Contag, C.H. (2002). Advancing animal models of neoplasia through in vivo bioluminescence imaging. Eur J Cancer 38, 2128-2136.
Edwards, B.K., Howe, H.L., Ries, L.A., Thun, M.J., Rosenberg, H.M., Yancik, R., Wingo, P.A., Jemal, A., and Feigal, E.G. (2002). Annual report to the nation on the status of cancer, 1973-1999, featuring implications of age and aging on U.S. cancer burden. Cancer 94, 2766-2792.
116
Eirew, P., Stingl, J., and Eaves, C.J. (2010). Quantitation of human mammary epithelial stem cells with in vivo regenerative properties using a subrenal capsule xenotransplantation assay. Nature protocols 5, 1945-1956.
Eisenhauer, E.A., Therasse, P., Bogaerts, J., Schwartz, L.H., Sargent, D., Ford, R., Dancey, J., Arbuck, S., Gwyther, S., Mooney, M., et al. (2009). New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 45, 228-247.
Eisenkop, S.M., and Spirtos, N.M. (2001). What are the current surgical objectives, strategies, and technical capabilities of gynecologic oncologists treating advanced epithelial ovarian cancer? Gynecol Oncol 82, 489-497.
Eneanya, D.I., Bianchine, J.R., Duran, D.O., and Andresen, B.D. (1981). The actions of metabolic fate of disulfiram. Annual review of pharmacology and toxicology 21, 575-596.
Eramo, A., Lotti, F., Sette, G., Pilozzi, E., Biffoni, M., Di Virgilio, A., Conticello, C., Ruco, L., Peschle, C., and De Maria, R. (2008). Identification and expansion of the tumorigenic lung cancer stem cell population. Cell death and differentiation 15, 504-514.
Faiman, M.D., Dodd, D.E., Minor, S.S., and Hanzlik, R. (1978). Radioactive and nonradioactive methods for the in vivo determination of disulfiram, diethyldithiocarbamate, and diethyldithiocarbamate-methyl ester. Alcoholism, clinical and experimental research 2, 366-369.
Faiman, M.D., Jensen, J.C., and Lacoursiere, R.B. (1984). Elimination kinetics of disulfiram in alcoholics after single and repeated doses. Clinical pharmacology and therapeutics 36, 520-526.
Fathalla, M.F. (1971). Incessant ovulation--a factor in ovarian neoplasia? Lancet 2, 163.
Fichtner, I., Rolff, J., Soong, R., Hoffmann, J., Hammer, S., Sommer, A., Becker, M., and Merk, J. (2008). Establishment of patient-derived non-small cell lung cancer xenografts as models for the identification of predictive biomarkers. Clinical cancer research : an official journal of the American Association for Cancer Research 14, 6456-6468.
Fiebig, H.H., Maier, A., and Burger, A.M. (2004). Clonogenic assay with established human tumour xenografts: correlation of in vitro to in vivo activity as a basis for anticancer drug discovery. Eur J Cancer 40, 802-820.
Fleming, J.S., Beaugie, C.R., Haviv, I., Chenevix-Trench, G., and Tan, O.L. (2006). Incessant ovulation, inflammation and epithelial ovarian carcinogenesis: revisiting old hypotheses. Molecular and cellular endocrinology 247, 4-21.
Flesken-Nikitin, A., Choi, K.C., Eng, J.P., Shmidt, E.N., and Nikitin, A.Y. (2003). Induction of carcinogenesis by concurrent inactivation of p53 and Rb1 in the mouse ovarian surface epithelium. Cancer research 63, 3459-3463.
Flesken-Nikitin, A., Hwang, C.I., Cheng, C.Y., Michurina, T.V., Enikolopov, G., and Nikitin, A.Y. (2013). Ovarian surface epithelium at the junction area contains a cancer-prone stem cell niche. Nature 495, 241-245.
117
Folkins, A.K., Jarboe, E.A., Saleemuddin, A., Lee, Y., Callahan, M.J., Drapkin, R., Garber, J.E., Muto, M.G., Tworoger, S., and Crum, C.P. (2008). A candidate precursor to pelvic serous cancer (p53 signature) and its prevalence in ovaries and fallopian tubes from women with BRCA mutations. Gynecol Oncol 109, 168-173.
Ford, D., and Easton, D.F. (1995). The genetics of breast and ovarian cancer. British journal of cancer 72, 805-812.
Foster, R., Buckanovich, R.J., and Rueda, B.R. (2013). Ovarian cancer stem cells: Working towards the root of stemness. Cancer letters 338, 147-157.
Franceschi, S., La Vecchia, C., Booth, M., Tzonou, A., Negri, E., Parazzini, F., Trichopoulos, D., and Beral, V. (1991). Pooled analysis of 3 European case-control studies of ovarian cancer: II. Age at menarche and at menopause. International journal of cancer Journal international du cancer 49, 57-60.
Francia, G., Green, S.K., Bocci, G., Man, S., Emmenegger, U., Ebos, J.M., Weinerman, A., Shaked, Y., and Kerbel, R.S. (2005). Down-regulation of DNA mismatch repair proteins in human and murine tumor spheroids: implications for multicellular resistance to alkylating agents. Molecular cancer therapeutics 4, 1484-1494.
Francia, G., and Kerbel, R.S. (2010). Raising the bar for cancer therapy models. Nature biotechnology 28, 561-562.
Fredrickson, T.N. (1987). Ovarian tumors of the hen. Environmental health perspectives 73, 35-51.
Frei, E., Antman, K.H. (2000). Principles of Dose, Schedule, and Combination Chemotherapy., Vol Chapter 40, 5th Edition. edn (Hamilton (ON): BC Decker; 2000: BC Decker
Fung Kee Fung, M., Kenney, E., Francis, J., Mackay, H., and members of the Gynecologic Cancer Disease SIte Group. (2011). Optimal Chemotherapy for recurrent ovarian cancer. In Program in Evidence-based Care (Toronto (ON): Cancer Care Ontario).
Gao, M.Q., Choi, Y.P., Kang, S., Youn, J.H., and Cho, N.H. (2010). CD24+ cells from hierarchically organized ovarian cancer are enriched in cancer stem cells. Oncogene 29, 2672-2680.
Garland, C.F., Mohr, S.B., Gorham, E.D., Grant, W.B., and Garland, F.C. (2006). Role of ultraviolet B irradiance and vitamin D in prevention of ovarian cancer. American journal of preventive medicine 31, 512-514.
Garson, K., Gamwell, L.F., Pitre, E.M., and Vanderhyden, B.C. (2012). Technical challenges and limitations of current mouse models of ovarian cancer. Journal of ovarian research 5, 39.
Garson, K., Shaw, T.J., Clark, K.V., Yao, D.S., and Vanderhyden, B.C. (2005). Models of ovarian cancer--are we there yet? Molecular and cellular endocrinology 239, 15-26.
118
Gaval-Cruz, M., and Weinshenker, D. (2009). mechanisms of disulfiram-induced cocaine abstinence: antabuse and cocaine relapse. Molecular interventions 9, 175-187.
Gayther, S.A., Warren, W., Mazoyer, S., Russell, P.A., Harrington, P.A., Chiano, M., Seal, S., Hamoudi, R., van Rensburg, E.J., Dunning, A.M., et al. (1995). Germline mutations of the BRCA1 gene in breast and ovarian cancer families provide evidence for a genotype-phenotype correlation. Nature genetics 11, 428-433.
Gerlinger, M., Rowan, A.J., Horswell, S., Larkin, J., Endesfelder, D., Gronroos, E., Martinez, P., Matthews, N., Stewart, A., Tarpey, P., et al. (2012). Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. The New England journal of medicine 366, 883-892.
Gertig, D.M., Hunter, D.J., Cramer, D.W., Colditz, G.A., Speizer, F.E., Willett, W.C., and Hankinson, S.E. (2000). Prospective study of talc use and ovarian cancer. Journal of the National Cancer Institute 92, 249-252.
Gordts, S., Campo, R., Rombauts, L., and Brosens, I. (1998). Endoscopic visualization of the process of fimbrial ovum retrieval in the human. Hum Reprod 13, 1425-1428.
Gore, M.E., Fryatt, I., Wiltshaw, E., and Dawson, T. (1990). Treatment of relapsed carcinoma of the ovary with cisplatin or carboplatin following initial treatment with these compounds. Gynecol Oncol 36, 207-211.
Gray, D.R., Huss, W.J., Yau, J.M., Durham, L.E., Werdin, E.S., Funkhouser, W.K., Jr., and Smith, G.J. (2004). Short-term human prostate primary xenografts: an in vivo model of human prostate cancer vasculature and angiogenesis. Cancer research 64, 1712-1721.
Greenlee, R.T., Hill-Harmon, M.B., Murray, T., and Thun, M. (2001). Cancer statistics, 2001. CA: a cancer journal for clinicians 51, 15-36.
Greggi, S., Parazzini, F., Paratore, M.P., Chatenoud, L., Legge, F., Mancuso, S., and La Vecchia, C. (2000). Risk factors for ovarian cancer in central Italy. Gynecol Oncol 79, 50-54.
Grindedal, E.M., Renkonen-Sinisalo, L., Vasen, H., Evans, G., Sala, P., Blanco, I., Gronwald, J., Apold, J., Eccles, D.M., Sanchez, A.A., et al. (2010). Survival in women with MMR mutations and ovarian cancer: a multicentre study in Lynch syndrome kindreds. Journal of medical genetics 47, 99-102.
Gu, P., Pan, L.L., Wu, S.Q., Sun, L., and Huang, G. (2009). CA 125, PET alone, PET-CT, CT and MRI in diagnosing recurrent ovarian carcinoma: a systematic review and meta-analysis. European journal of radiology 71, 164-174.
Gupta, D., and Lis, C.G. (2009). Role of CA125 in predicting ovarian cancer survival - a review of the epidemiological literature. Journal of ovarian research 2, 13.
119
Hankinson, S.E., Colditz, G.A., Hunter, D.J., Willett, W.C., Stampfer, M.J., Rosner, B., Hennekens, C.H., and Speizer, F.E. (1995). A prospective study of reproductive factors and risk of epithelial ovarian cancer. Cancer 76, 284-290.
Harrap, K.R., Jones, M., Siracky, J., Pollard, L.A., and Kelland, L.R. (1990). The establishment, characterization and calibration of human ovarian carcinoma xenografts for the evaluation of novel platinum anticancer drugs. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 1, 65-76.
Harrison, H., Farnie, G., Howell, S.J., Rock, R.E., Stylianou, S., Brennan, K.R., Bundred, N.J., and Clarke, R.B. (2010). Regulation of breast cancer stem cell activity by signaling through the Notch4 receptor. Cancer research 70, 709-718.
Hartge, P., Hoover, R., McGowan, L., Lesher, L., and Norris, H.J. (1988). Menopause and ovarian cancer. American journal of epidemiology 127, 990-998.
He, G., Siddik, Z.H., Huang, Z., Wang, R., Koomen, J., Kobayashi, R., Khokhar, A.R., and Kuang, J. (2005). Induction of p21 by p53 following DNA damage inhibits both Cdk4 and Cdk2 activities. Oncogene 24, 2929-2943.
He, Q., Liang, C.H., and Lippard, S.J. (2000). Steroid hormones induce HMG1 overexpression and sensitize breast cancer cells to cisplatin and carboplatin. Proceedings of the National Academy of Sciences of the United States of America 97, 5768-5772.
Helzlsouer, K.J., Alberg, A.J., Gordon, G.B., Longcope, C., Bush, T.L., Hoffman, S.C., and Comstock, G.W. (1995). Serum gonadotropins and steroid hormones and the development of ovarian cancer. JAMA : the journal of the American Medical Association 274, 1926-1930.
Henderson, W.J., Hamilton, T.C., and Griffiths, K. (1979). Talc in normal and malignant ovarian tissue. Lancet 1, 499.
Hennessy, B.T., Coleman, R.L., and Markman, M. (2009). Ovarian cancer. Lancet 374, 1371-1382.
Hennessy, B.T., Suh, G.K., and Markman, M. (2011). The MD Anderson Manual of Medical Oncology (McGraw-Hill).
Hennessy, B.T., Timms, K.M., Carey, M.S., Gutin, A., Meyer, L.A., Flake, D.D., 2nd, Abkevich, V., Potter, J., Pruss, D., Glenn, P., et al. (2010). Somatic mutations in BRCA1 and BRCA2 could expand the number of patients that benefit from poly (ADP ribose) polymerase inhibitors in ovarian cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 28, 3570-3576.
Hermann, P.C., Huber, S.L., Herrler, T., Aicher, A., Ellwart, J.W., Guba, M., Bruns, C.J., and Heeschen, C. (2007). Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer. Cell stem cell 1, 313-323.
120
Hidalgo, M., Bruckheimer, E., Rajeshkumar, N.V., Garrido-Laguna, I., De Oliveira, E., Rubio-Viqueira, B., Strawn, S., Wick, M.J., Martell, J., and Sidransky, D. (2011). A pilot clinical study of treatment guided by personalized tumorgrafts in patients with advanced cancer. Molecular cancer therapeutics 10, 1311-1316.
Ho, S.M. (2003). Estrogen, progesterone and epithelial ovarian cancer. Reproductive biology and endocrinology : RB&E 1, 73.
Hoffman, B.L., Schorge, J.O., Shchaffer, J.I., Halvorson, L.M., Bradshaw, K.D., Cunningham, F.G., and Calver, L.E. (2012). Williams Gynecology, 2 edn (McGraw-Hill).
Hofseth, L.J., Hussain, S.P., and Harris, C.C. (2004). p53: 25 years after its discovery. Trends in pharmacological sciences 25, 177-181.
Hogdall, E.V., Christensen, L., Kjaer, S.K., Blaakaer, J., Kjaerbye-Thygesen, A., Gayther, S., Jacobs, I.J., and Hogdall, C.K. (2007). CA125 expression pattern, prognosis and correlation with serum CA125 in ovarian tumor patients. From The Danish "MALOVA" Ovarian Cancer Study. Gynecol Oncol 104, 508-515.
Hori, Y., Nishida, K., Yamato, M., Sugiyama, H., Soma, T., Inoue, T., Maeda, N., Okano, T., and Tano, Y. (2008). Differential expression of MUC16 in human oral mucosal epithelium and cultivated epithelial sheets. Experimental eye research 87, 191-196.
Hoskins, W.J., McGuire, W.P., Brady, M.F., Homesley, H.D., Creasman, W.T., Berman, M., Ball, H., and Berek, J.S. (1994). The effect of diameter of largest residual disease on survival after primary cytoreductive surgery in patients with suboptimal residual epithelial ovarian carcinoma. American journal of obstetrics and gynecology 170, 974-979; discussion 979-980.
Hu, Y., and Smyth, G.K. (2009). ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays. Journal of immunological methods 347, 70-78.
Huang, H., Li, Y., Liu, J., Zheng, M., Feng, Y., Hu, K., Huang, Y., and Huang, Q. (2012). Screening and identification of biomarkers in ascites related to intrinsic chemoresistance of serous epithelial ovarian cancers. PloS one 7, e51256.
Huncharek, M., Geschwind, J.F., and Kupelnick, B. (2003). Perineal application of cosmetic talc and risk of invasive epithelial ovarian cancer: a meta-analysis of 11,933 subjects from sixteen observational studies. Anticancer research 23, 1955-1960.
Hunn, J., and Rodriguez, G.C. (2012). Ovarian cancer: etiology, risk factors, and epidemiology. Clinical obstetrics and gynecology 55, 3-23.
Ishizawa, K., Rasheed, Z.A., Karisch, R., Wang, Q., Kowalski, J., Susky, E., Pereira, K., Karamboulas, C., Moghal, N., Rajeshkumar, N.V., et al. (2010). Tumor-initiating cells are rare in many human tumors. Cell stem cell 7, 279-282.
121
Ito, M., Hiramatsu, H., Kobayashi, K., Suzue, K., Kawahata, M., Hioki, K., Ueyama, Y., Koyanagi, Y., Sugamura, K., Tsuji, K., et al. (2002). NOD/SCID/gamma(c)(null) mouse: an excellent recipient mouse model for engraftment of human cells. Blood 100, 3175-3182.
Iyer, V.R., and Lee, S.I. (2010). MRI, CT, and PET/CT for ovarian cancer detection and adnexal lesion characterization. AJR American journal of roentgenology 194, 311-321.
Jackson Laboratory. (2013). http://jaxmice.jax.org/nod-scid-gamma/nsg-questions-and-answers.html - What is a humanized mouse.
Jacobs, I., Davies, A.P., Bridges, J., Stabile, I., Fay, T., Lower, A., Grudzinskas, J.G., and Oram, D. (1993). Prevalence screening for ovarian cancer in postmenopausal women by CA 125 measurement and ultrasonography. BMJ 306, 1030-1034.
Jacobs, I.J., and Menon, U. (2004). Progress and challenges in screening for early detection of ovarian cancer. Molecular & cellular proteomics : MCP 3, 355-366.
Jandial, D.D., Messer, K., Farshchi-Heydari, S., Pu, M., and Howell, S.B. (2009). Tumor platinum concentration following intraperitoneal administration of cisplatin versus carboplatin in an ovarian cancer model. Gynecol Oncol 115, 362-366.
Jarboe, E.A., Folkins, A.K., Drapkin, R., Ince, T.A., Agoston, E.S., and Crum, C.P. (2008). Tubal and ovarian pathways to pelvic epithelial cancer: a pathological perspective. Histopathology 53, 127-138.
Jemal, A., Bray, F., Center, M.M., Ferlay, J., Ward, E., and Forman, D. (2011). Global cancer statistics. CA: a cancer journal for clinicians 61, 69-90.
Jenkins, D.E., Oei, Y., Hornig, Y.S., Yu, S.F., Dusich, J., Purchio, T., and Contag, P.R. (2003a). Bioluminescent imaging (BLI) to improve and refine traditional murine models of tumor growth and metastasis. Clinical & experimental metastasis 20, 733-744.
Jenkins, D.E., Yu, S.F., Hornig, Y.S., Purchio, T., and Contag, P.R. (2003b). In vivo monitoring of tumor relapse and metastasis using bioluminescent PC-3M-luc-C6 cells in murine models of human prostate cancer. Clinical & experimental metastasis 20, 745-756.
Jin, K., Teng, L., Shen, Y., He, K., Xu, Z., and Li, G. (2010). Patient-derived human tumour tissue xenografts in immunodeficient mice: a systematic review. Clinical & translational oncology : official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico 12, 473-480.
Johansson, B. (1992). A review of the pharmacokinetics and pharmacodynamics of disulfiram and its metabolites. Acta psychiatrica Scandinavica Supplementum 369, 15-26.
Johnson, J.I., Decker, S., Zaharevitz, D., Rubinstein, L.V., Venditti, J.M., Schepartz, S., Kalyandrug, S., Christian, M., Arbuck, S., Hollingshead, M., et al. (2001). Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. British journal of cancer 84, 1424-1431.
122
Jones, S., Zhang, X., Parsons, D.W., Lin, J.C., Leary, R.J., Angenendt, P., Mankoo, P., Carter, H., Kamiyama, H., Jimeno, A., et al. (2008). Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 321, 1801-1806.
Kee, Y., and D'Andrea, A.D. (2010). Expanded roles of the Fanconi anemia pathway in preserving genomic stability. Genes & development 24, 1680-1694.
Kerbel, R.S. (2003). Human tumor xenografts as predictive preclinical models for anticancer drug activity in humans: better than commonly perceived-but they can be improved. Cancer Biol Ther 2, S134-139.
Kern, S.E., and Shibata, D. (2007). The fuzzy math of solid tumor stem cells: a perspective. Cancer research 67, 8985-8988.
Ketabi, Z., Bartuma, K., Bernstein, I., Malander, S., Gronberg, H., Bjorck, E., Holck, S., and Nilbert, M. (2011). Ovarian cancer linked to Lynch syndrome typically presents as early-onset, non-serous epithelial tumors. Gynecol Oncol 121, 462-465.
Kindelberger, D.W., Lee, Y., Miron, A., Hirsch, M.S., Feltmate, C., Medeiros, F., Callahan, M.J., Garner, E.O., Gordon, R.W., Birch, C., et al. (2007). Intraepithelial carcinoma of the fimbria and pelvic serous carcinoma: Evidence for a causal relationship. The American journal of surgical pathology 31, 161-169.
King, M.C., Marks, J.H., and Mandell, J.B. (2003). Breast and ovarian cancer risks due to inherited mutations in BRCA1 and BRCA2. Science 302, 643-646.
King, S.M., Quartuccio, S.M., Vanderhyden, B.C., and Burdette, J.E. (2013). Early transformative changes in normal ovarian surface epithelium induced by oxidative stress require Akt upregulation, DNA damage and epithelial-stromal interaction. Carcinogenesis 34, 1125-1133.
Klerk, C.P., Overmeer, R.M., Niers, T.M., Versteeg, H.H., Richel, D.J., Buckle, T., Van Noorden, C.J., and van Tellingen, O. (2007). Validity of bioluminescence measurements for noninvasive in vivo imaging of tumor load in small animals. BioTechniques 43, 7-13, 30.
Kobayashi, A., and Behringer, R.R. (2003). Developmental genetics of the female reproductive tract in mammals. Nature reviews Genetics 4, 969-980.
Kulkarni-Datar, K., Orsulic, S., Foster, R., and Rueda, B.R. (2013). Ovarian tumor initiating cell populations persist following paclitaxel and carboplatin chemotherapy treatment in vivo. Cancer letters.
Kung, A.L. (2007). Practices and pitfalls of mouse cancer models in drug discovery. Advances in cancer research 96, 191-212.
Kurbacher, C.M., Bruckner, H.W., Andreotti, P.E., Kurbacher, J.A., Sass, G., and Krebs, D. (1995). In vitro activity of titanocenedichloride versus cisplatin in four ovarian carcinoma cell lines evaluated by a microtiter plate ATP bioluminescence assay. Anti-cancer drugs 6, 697-704.
123
Kurman, R.J., and Shih Ie, M. (2010). The origin and pathogenesis of epithelial ovarian cancer: a proposed unifying theory. The American journal of surgical pathology 34, 433-443.
La Vecchia, C. (2006). Estrogen-progestogen replacement therapy and ovarian cancer: an update. Eur J Cancer Prev 15, 490-492.
La Vecchia, C., and Franceschi, S. (1999). Oral contraceptives and ovarian cancer. Eur J Cancer Prev 8, 297-304.
Lakhani, S.R., Manek, S., Penault-Llorca, F., Flanagan, A., Arnout, L., Merrett, S., McGuffog, L., Steele, D., Devilee, P., Klijn, J.G., et al. (2004). Pathology of ovarian cancers in BRCA1 and BRCA2 carriers. Clinical cancer research : an official journal of the American Association for Cancer Research 10, 2473-2481.
Land, J.A. (1993). Ovulation, ovulation induction and ovarian carcinoma. Bailliere's clinical obstetrics and gynaecology 7, 455-472.
Langevin, F., Crossan, G.P., Rosado, I.V., Arends, M.J., and Patel, K.J. (2011). Fancd2 counteracts the toxic effects of naturally produced aldehydes in mice. Nature 475, 53-58.
Lee, C.H., Xue, H., Sutcliffe, M., Gout, P.W., Huntsman, D.G., Miller, D.M., Gilks, C.B., and Wang, Y.Z. (2005). Establishment of subrenal capsule xenografts of primary human ovarian tumors in SCID mice: potential models. Gynecol Oncol 96, 48-55.
Lee, Y., Medeiros, F., Kindelberger, D., Callahan, M.J., Muto, M.G., and Crum, C.P. (2006). Advances in the recognition of tubal intraepithelial carcinoma: applications to cancer screening and the pathogenesis of ovarian cancer. Advances in anatomic pathology 13, 1-7.
Lefkowitz, E.S., and Garland, C.F. (1994). Sunlight, vitamin D, and ovarian cancer mortality rates in US women. International journal of epidemiology 23, 1133-1136.
Lengyel, E., Burdette, J.E., Kenny, H.A., Matei, D., Pilrose, J., Haluska, P., Nephew, K.P., Hales, D.B., and Stack, M.S. (2013). Epithelial ovarian cancer experimental models. Oncogene.
Levanon, K., Crum, C., and Drapkin, R. (2008). New insights into the pathogenesis of serous ovarian cancer and its clinical impact. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 26, 5284-5293.
Li, X., Lewis, M.T., Huang, J., Gutierrez, C., Osborne, C.K., Wu, M.F., Hilsenbeck, S.G., Pavlick, A., Zhang, X., Chamness, G.C., et al. (2008). Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy. Journal of the National Cancer Institute 100, 672-679.
Lin, J., Haffner, M.C., Zhang, Y., Lee, B.H., Brennen, W.N., Britton, J., Kachhap, S.K., Shim, J.S., Liu, J.O., Nelson, W.G., et al. (2011). Disulfiram is a DNA demethylating agent and inhibits prostate cancer cell growth. The Prostate 71, 333-343.
Liu, P., Kumar, I.S., Brown, S., Kannappan, V., Tawari, P.E., Tang, J.Z., Jiang, W., Armesilla, A.L., Darling, J.L., and Wang, W. (2013). Disulfiram targets cancer stem-like cells and reverses
124
resistance and cross-resistance in acquired paclitaxel-resistant triple-negative breast cancer cells. British journal of cancer.
Lohnes, D., Mark, M., Mendelsohn, C., Dolle, P., Dierich, A., Gorry, P., Gansmuller, A., and Chambon, P. (1994). Function of the retinoic acid receptors (RARs) during development (I). Craniofacial and skeletal abnormalities in RAR double mutants. Development 120, 2723-2748.
Loo, T.W., and Clarke, D.M. (2000). Blockage of drug resistance in vitro by disulfiram, a drug used to treat alcoholism. Journal of the National Cancer Institute 92, 898-902.
Lovborg, H., Oberg, F., Rickardson, L., Gullbo, J., Nygren, P., and Larsson, R. (2006). Inhibition of proteasome activity, nuclear factor-KappaB translocation and cell survival by the antialcoholism drug disulfiram. International journal of cancer Journal international du cancer 118, 1577-1580.
Lu, K.H.S., S.; Hernandez, M. A.; Bedi, D.; Bevers, T.; Leeds, L.; Moore, R.; Granai, C.; Harris, S.; Newland, W.; Adeyinka, O.; Geffen, J.' Deavers, M. T.; Sun, C. C.; Horick, N.; Fritsche, H.; Bast, R. C. (2013). A 2-stage ovarian cancer screening strategy using the Risk of Ovarian Cancer Algorithm (ROCA) identifies early-stage incident cancers and demonstrates high positive predictive value. . Cancer.
Lukanova, A., and Kaaks, R. (2005). Endogenous hormones and ovarian cancer: epidemiology and current hypotheses. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 14, 98-107.
Luo, L., Zeng, J., Liang, B., Zhao, Z., Sun, L., Cao, D., Yang, J., and Shen, K. (2011). Ovarian cancer cells with the CD117 phenotype are highly tumorigenic and are related to chemotherapy outcome. Experimental and molecular pathology 91, 596-602.
Ma, S., Lee, T.K., Zheng, B.J., Chan, K.W., and Guan, X.Y. (2008). CD133+ HCC cancer stem cells confer chemoresistance by preferential expression of the Akt/PKB survival pathway. Oncogene 27, 1749-1758.
Magee, J.A., Piskounova, E., and Morrison, S.J. (2012). Cancer stem cells: impact, heterogeneity, and uncertainty. Cancer cell 21, 283-296.
Mann, W.J.C., E.; Valea, F.A. (2013). UpToDate (Waltham, MA: UpToDate).
Marikovsky, M., Ziv, V., Nevo, N., Harris-Cerruti, C., and Mahler, O. (2003). Cu/Zn superoxide dismutase plays important role in immune response. J Immunol 170, 2993-3001.
Markman, M. (2011). Counterpoint: chemosensitivity assays for recurrent ovarian cancer. Journal of the National Comprehensive Cancer Network : JNCCN 9, 121-124.
Markman, M., Rothman, R., Hakes, T., Reichman, B., Hoskins, W., Rubin, S., Jones, W., Almadrones, L., and Lewis, J.L., Jr. (1991). Second-line platinum therapy in patients with
125
ovarian cancer previously treated with cisplatin. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 9, 389-393.
Martin, S.A., Lord, C.J., and Ashworth, A. (2010). Therapeutic targeting of the DNA mismatch repair pathway. Clinical cancer research : an official journal of the American Association for Cancer Research 16, 5107-5113.
Massazza, G., Tomasoni, A., Lucchini, V., Allavena, P., Erba, E., Colombo, N., Mantovani, A., D'Incalci, M., Mangioni, C., and Giavazzi, R. (1989). Intraperitoneal and subcutaneous xenografts of human ovarian carcinoma in nude mice and their potential in experimental therapy. International journal of cancer Journal international du cancer 44, 494-500.
McSorley, M.A., Alberg, A.J., Allen, D.S., Allen, N.E., Brinton, L.A., Dorgan, J.F., Pollak, M., Tao, Y., and Helzlsouer, K.J. (2007). C-reactive protein concentrations and subsequent ovarian cancer risk. Obstetrics and gynecology 109, 933-941.
Medeiros, F., Muto, M.G., Lee, Y., Elvin, J.A., Callahan, M.J., Feltmate, C., Garber, J.E., Cramer, D.W., and Crum, C.P. (2006). The tubal fimbria is a preferred site for early adenocarcinoma in women with familial ovarian cancer syndrome. The American journal of surgical pathology 30, 230-236.
Mendelsohn, C., Larkin, S., Mark, M., LeMeur, M., Clifford, J., Zelent, A., and Chambon, P. (1994a). RAR beta isoforms: distinct transcriptional control by retinoic acid and specific spatial patterns of promoter activity during mouse embryonic development. Mechanisms of development 45, 227-241.
Mendelsohn, C., Lohnes, D., Decimo, D., Lufkin, T., LeMeur, M., Chambon, P., and Mark, M. (1994b). Function of the retinoic acid receptors (RARs) during development (II). Multiple abnormalities at various stages of organogenesis in RAR double mutants. Development 120, 2749-2771.
Menon, U., Talaat, A., Jeyarajah, A.R., Rosenthal, A.N., MacDonald, N.D., Skates, S.J., Sibley, K., Oram, D.H., and Jacobs, I.J. (1999). Ultrasound assessment of ovarian cancer risk in postmenopausal women with CA125 elevation. British journal of cancer 80, 1644-1647.
Merk, J., Rolff, J., Becker, M., Leschber, G., and Fichtner, I. (2009). Patient-derived xenografts of non-small-cell lung cancer: a pre-clinical model to evaluate adjuvant chemotherapy? European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery 36, 454-459.
Merk, J., Rolff, J., Dorn, C., Leschber, G., and Fichtner, I. (2011). Chemoresistance in non-small-cell lung cancer: can multidrug resistance markers predict the response of xenograft lung cancer models to chemotherapy? European journal of cardio-thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery 40, e29-33.
Merogi, A.J., Marrogi, A.J., Ramesh, R., Robinson, W.R., Fermin, C.D., and Freeman, S.M. (1997). Tumor-host interaction: analysis of cytokines, growth factors, and tumor-infiltrating lymphocytes in ovarian carcinomas. Human pathology 28, 321-331.
126
Miki, Y., Swensen, J., Shattuck-Eidens, D., Futreal, P.A., Harshman, K., Tavtigian, S., Liu, Q., Cochran, C., Bennett, L.M., Ding, W., et al. (1994). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266, 66-71.
Miller, C., Pavlova, A., and Sassoon, D.A. (1998). Differential expression patterns of Wnt genes in the murine female reproductive tract during development and the estrous cycle. Mechanisms of development 76, 91-99.
Morch, L.S., Lokkegaard, E., Andreasen, A.H., Kruger-Kjaer, S., and Lidegaard, O. (2009). Hormone therapy and ovarian cancer. JAMA : the journal of the American Medical Association 302, 298-305.
Morrison, B.W., Doudican, N.A., Patel, K.R., and Orlow, S.J. (2010). Disulfiram induces copper-dependent stimulation of reactive oxygen species and activation of the extrinsic apoptotic pathway in melanoma. Melanoma research 20, 11-20.
Morrison, J., Haldar, K., Kehoe, S., and Lawrie, T.A. (2012). Chemotherapy versus surgery for initial treatment in advanced ovarian epithelial cancer. Cochrane Database Syst Rev 8, CD005343.
Moynahan, M.E., Pierce, A.J., and Jasin, M. (2001). BRCA2 is required for homology-directed repair of chromosomal breaks. Molecular cell 7, 263-272.
Muggia, F.M.B., L. (2011). Textbook of Gynaecological Oncology, Vol 2nd (Milan).
Mullany, L.K., and Richards, J.S. (2012). Minireview: animal models and mechanisms of ovarian cancer development. Endocrinology 153, 1585-1592.
Mutch, D.G. (2002). Surgical management of ovarian cancer. Seminars in oncology 29, 3-8.
Narod, S.A., Sun, P., Ghadirian, P., Lynch, H., Isaacs, C., Garber, J., Weber, B., Karlan, B., Fishman, D., Rosen, B., et al. (2001). Tubal ligation and risk of ovarian cancer in carriers of BRCA1 or BRCA2 mutations: a case-control study. Lancet 357, 1467-1470.
National Cancer Institute. (2013). SEER Cancer Statistics Review, 1975-2010. In http://seercancergov/csr/1975_2010, N.A. Howlader N, Krapcho M, Garshell J, Neyman N, Altekruse SF, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Cho H, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA ed. (Bethesda, MD).
Ness, R.B., and Cottreau, C. (1999). Possible role of ovarian epithelial inflammation in ovarian cancer. Journal of the National Cancer Institute 91, 1459-1467.
Ness, R.B., Dodge, R.C., Edwards, R.P., Baker, J.A., and Moysich, K.B. (2011). Contraception methods, beyond oral contraceptives and tubal ligation, and risk of ovarian cancer. Annals of epidemiology 21, 188-196.
127
Nolen, B., Velikokhatnaya, L., Marrangoni, A., De Geest, K., Lomakin, A., Bast, R.C., Jr., and Lokshin, A. (2010). Serum biomarker panels for the discrimination of benign from malignant cases in patients with an adnexal mass. Gynecol Oncol 117, 440-445.
Notta, F., Mullighan, C.G., Wang, J.C., Poeppl, A., Doulatov, S., Phillips, L.A., Ma, J., Minden, M.D., Downing, J.R., and Dick, J.E. (2011). Evolution of human BCR-ABL1 lymphoblastic leukaemia-initiating cells. Nature 469, 362-367.
O'Brien, C.A., Kreso, A., and Dick, J.E. (2009a). Cancer stem cells in solid tumors: an overview. Seminars in radiation oncology 19, 71-77.
O'Brien, C.A., Pollett, A., Gallinger, S., and Dick, J.E. (2007). A human colon cancer cell capable of initiating tumour growth in immunodeficient mice. Nature 445, 106-110.
O'Brien, C.S., Howell, S.J., Farnie, G., and Clarke, R.B. (2009b). Resistance to endocrine therapy: are breast cancer stem cells the culprits? Journal of mammary gland biology and neoplasia 14, 45-54.
O'Connell, G.J., Ryan, E., Murphy, K.J., and Prefontaine, M. (1987). Predictive value of CA 125 for ovarian carcinoma in patients presenting with pelvic masses. Obstetrics and gynecology 70, 930-932.
O'Neill, K., Lyons, S.K., Gallagher, W.M., Curran, K.M., and Byrne, A.T. (2010). Bioluminescent imaging: a critical tool in pre-clinical oncology research. The Journal of pathology 220, 317-327.
Pal, T., Permuth-Wey, J., Betts, J.A., Krischer, J.P., Fiorica, J., Arango, H., LaPolla, J., Hoffman, M., Martino, M.A., Wakeley, K., et al. (2005). BRCA1 and BRCA2 mutations account for a large proportion of ovarian carcinoma cases. Cancer 104, 2807-2816.
Parmar, M.K., Ledermann, J.A., Colombo, N., du Bois, A., Delaloye, J.F., Kristensen, G.B., Wheeler, S., Swart, A.M., Qian, W., Torri, V., et al. (2003). Paclitaxel plus platinum-based chemotherapy versus conventional platinum-based chemotherapy in women with relapsed ovarian cancer: the ICON4/AGO-OVAR-2.2 trial. Lancet 361, 2099-2106.
Partheen, K., Kristjansdottir, B., and Sundfeldt, K. (2011). Evaluation of ovarian cancer biomarkers HE4 and CA-125 in women presenting with a suspicious cystic ovarian mass. Journal of gynecologic oncology 22, 244-252.
Peterson, J.K., and Houghton, P.J. (2004). Integrating pharmacology and in vivo cancer models in preclinical and clinical drug development. Eur J Cancer 40, 837-844.
Powell, C.B., Swisher, E.M., Cass, I., McLennan, J., Norquist, B., Garcia, R.L., Lester, J., Karlan, B.Y., and Chen, L. (2013). Long term follow up of BRCA1 and BRCA2 mutation carriers with unsuspected neoplasia identified at risk reducing salpingo-oophorectomy. Gynecol Oncol 129, 364-371.
128
Prat, A., Parera, M., Adamo, B., Peralta, S., Perez-Benavente, M.A., Garcia, A., Gil-Moreno, A., Martinez-Palones, J.M., Baselga, J., and del Campo, J.M. (2009). Risk of recurrence during follow-up for optimally treated advanced epithelial ovarian cancer (EOC) with a low-level increase of serum CA-125 levels. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 20, 294-297.
Press, J.Z., De Luca, A., Boyd, N., Young, S., Troussard, A., Ridge, Y., Kaurah, P., Kalloger, S.E., Blood, K.A., Smith, M., et al. (2008a). Ovarian carcinomas with genetic and epigenetic BRCA1 loss have distinct molecular abnormalities. BMC cancer 8, 17.
Press, J.Z., Kenyon, J.A., Xue, H., Miller, M.A., De Luca, A., Miller, D.M., Huntsman, D.G., Gilks, C.B., McAlpine, J.N., and Wang, Y.Z. (2008b). Xenografts of primary human gynecological tumors grown under the renal capsule of NOD/SCID mice show genetic stability during serial transplantation and respond to cytotoxic chemotherapy. Gynecol Oncol 110, 256-264.
Pujade-Lauraine, E., Wagner, U., Aavall-Lundqvist, E., Gebski, V., Heywood, M., Vasey, P.A., Volgger, B., Vergote, I., Pignata, S., Ferrero, A., et al. (2010). Pegylated liposomal Doxorubicin and Carboplatin compared with Paclitaxel and Carboplatin for patients with platinum-sensitive ovarian cancer in late relapse. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 28, 3323-3329.
Purdie, D., Green, A., Bain, C., Siskind, V., Ward, B., Hacker, N., Quinn, M., Wright, G., Russell, P., and Susil, B. (1995). Reproductive and other factors and risk of epithelial ovarian cancer: an Australian case-control study. Survey of Women's Health Study Group. International journal of cancer Journal international du cancer 62, 678-684.
Quinn, B.A., Brake, T., Hua, X., Baxter-Jones, K., Litwin, S., Ellenson, L.H., and Connolly, D.C. (2009). Induction of ovarian leiomyosarcomas in mice by conditional inactivation of Brca1 and p53. PloS one 4, e8404.
Rabik, C.A., and Dolan, M.E. (2007). Molecular mechanisms of resistance and toxicity associated with platinating agents. Cancer treatment reviews 33, 9-23.
Ramus, S.J., and Gayther, S.A. (2009). The contribution of BRCA1 and BRCA2 to ovarian cancer. Molecular oncology 3, 138-150.
Reagan-Shaw, S., Nihal, M., and Ahmad, N. (2008). Dose translation from animal to human studies revisited. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 22, 659-661.
Reeves, R., and Adair, J.E. (2005). Role of high mobility group (HMG) chromatin proteins in DNA repair. DNA repair 4, 926-938.
Rehemtulla, A., Stegman, L.D., Cardozo, S.J., Gupta, S., Hall, D.E., Contag, C.H., and Ross, B.D. (2000). Rapid and quantitative assessment of cancer treatment response using in vivo bioluminescence imaging. Neoplasia 2, 491-495.
129
Ricci, F., Broggini, M., and Damia, G. (2013). Revisiting ovarian cancer preclinical models: implications for a better management of the disease. Cancer treatment reviews 39, 561-568.
Ricci-Vitiani, L., Lombardi, D.G., Pilozzi, E., Biffoni, M., Todaro, M., Peschle, C., and De Maria, R. (2007). Identification and expansion of human colon-cancer-initiating cells. Nature 445, 111-115.
Rich, J.N., and Bao, S. (2007). Chemotherapy and cancer stem cells. Cell stem cell 1, 353-355.
Richmond, A., and Su, Y. (2008). Mouse xenograft models vs GEM models for human cancer therapeutics. Disease models & mechanisms 1, 78-82.
Riedinger, J.M., Bonnetain, F., Basuyau, J.P., Eche, N., Larbre, H., Dalifard, I., Wafflart, J., Ricolleau, G., and Pichon, M.F. (2007). Change in CA 125 levels after the first cycle of induction chemotherapy is an independent predictor of epithelial ovarian tumour outcome. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 18, 881-885.
Riman, T., Dickman, P.W., Nilsson, S., Correia, N., Nordlinder, H., Magnusson, C.M., and Persson, I.R. (2002a). Risk factors for invasive epithelial ovarian cancer: results from a Swedish case-control study. American journal of epidemiology 156, 363-373.
Riman, T., Dickman, P.W., Nilsson, S., Correia, N., Nordlinder, H., Magnusson, C.M., Weiderpass, E., and Persson, I.R. (2002b). Hormone replacement therapy and the risk of invasive epithelial ovarian cancer in Swedish women. Journal of the National Cancer Institute 94, 497-504.
Risch, H.A. (1998). Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. Journal of the National Cancer Institute 90, 1774-1786.
Rizzo, A., Roscino, M.T., Binetti, F., and Sciorsci, R.L. (2012). Roles of reactive oxygen species in female reproduction. Reproduction in domestic animals = Zuchthygiene 47, 344-352.
Rizzo, S., Hersey, J.M., Mellor, P., Dai, W., Santos-Silva, A., Liber, D., Luk, L., Titley, I., Carden, C.P., Box, G., et al. (2011). Ovarian cancer stem cell-like side populations are enriched following chemotherapy and overexpress EZH2. Molecular cancer therapeutics 10, 325-335.
Rodriguez, C., Calle, E.E., Coates, R.J., Miracle-McMahill, H.L., Thun, M.J., and Heath, C.W., Jr. (1995). Estrogen replacement therapy and fatal ovarian cancer. American journal of epidemiology 141, 828-835.
Romero, I., and Bast, R.C., Jr. (2012). Minireview: human ovarian cancer: biology, current management, and paths to personalizing therapy. Endocrinology 153, 1593-1602.
Roy, R., Chun, J., and Powell, S.N. (2012). BRCA1 and BRCA2: different roles in a common pathway of genome protection. Nature reviews Cancer 12, 68-78.
130
Rump, A., Morikawa, Y., Tanaka, M., Minami, S., Umesaki, N., Takeuchi, M., and Miyajima, A. (2004). Binding of ovarian cancer antigen CA125/MUC16 to mesothelin mediates cell adhesion. The Journal of biological chemistry 279, 9190-9198.
Rustin, G.J. (2010). What surveillance plan should be advised for patients in remission after completion of first-line therapy for advanced ovarian cancer? International journal of gynecological cancer : official journal of the International Gynecological Cancer Society 20, S27-28.
Rustin, G.J., van der Burg, M.E., Griffin, C.L., Guthrie, D., Lamont, A., Jayson, G.C., Kristensen, G., Mediola, C., Coens, C., Qian, W., et al. (2010). Early versus delayed treatment of relapsed ovarian cancer (MRC OV05/EORTC 55955): a randomised trial. Lancet 376, 1155-1163.
Sadikot, R.T., and Blackwell, T.S. (2005). Bioluminescence imaging. Proceedings of the American Thoracic Society 2, 537-540, 511-532.
Sakai, W., Swisher, E.M., Jacquemont, C., Chandramohan, K.V., Couch, F.J., Langdon, S.P., Wurz, K., Higgins, J., Villegas, E., and Taniguchi, T. (2009). Functional restoration of BRCA2 protein by secondary BRCA2 mutations in BRCA2-mutated ovarian carcinoma. Cancer research 69, 6381-6386.
Satagopan, J.M., Boyd, J., Kauff, N.D., Robson, M., Scheuer, L., Narod, S., and Offit, K. (2002). Ovarian cancer risk in Ashkenazi Jewish carriers of BRCA1 and BRCA2 mutations. Clinical cancer research : an official journal of the American Association for Cancer Research 8, 3776-3781.
Sato, A., Klaunberg, B., and Tolwani, R. (2004). In vivo bioluminescence imaging. Comparative medicine 54, 631-634.
Sauna, Z.E., Shukla, S., and Ambudkar, S.V. (2005). Disulfiram, an old drug with new potential therapeutic uses for human cancers and fungal infections. Molecular bioSystems 1, 127-134.
Sausville, E.A., and Burger, A.M. (2006). Contributions of human tumor xenografts to anticancer drug development. Cancer research 66, 3351-3354, discussion 3354.
Scholler, N., Garvik, B., Hayden-Ledbetter, M., Kline, T., and Urban, N. (2007). Development of a CA125-mesothelin cell adhesion assay as a screening tool for biologics discovery. Cancer letters 247, 130-136.
Schorge, J.O., Modesitt, S.C., Coleman, R.L., Cohn, D.E., Kauff, N.D., Duska, L.R., and Herzog, T.J. (2010). SGO White Paper on ovarian cancer: etiology, screening and surveillance. Gynecol Oncol 119, 7-17.
Schrader, K.A., Hurlburt, J., Kalloger, S.E., Hansford, S., Young, S., Huntsman, D.G., Gilks, C.B., and McAlpine, J.N. (2012). Germline BRCA1 and BRCA2 mutations in ovarian cancer: utility of a histology-based referral strategy. Obstetrics and gynecology 120, 235-240.
131
Schumacher, U., Adam, E., Dietl, J., and Horny, H.P. (1997). Immunophenotype of human ovarian malignancies (cystadenocarcinomata and mixed mullerian tumor) established in SCID mice. Experimental and molecular pathology 64, 103-113.
Schumacher, U., Adam, E., Horny, H.P., and Dietl, J. (1996). Transplantation of a human ovarian cystadenocarcinoma into severe combined immunodeficient (SCID) mice--formation of metastases without significant alteration of the tumour cell phenotype. International journal of experimental pathology 77, 219-227.
Shackleton, M., Quintana, E., Fearon, E.R., and Morrison, S.J. (2009). Heterogeneity in cancer: cancer stem cells versus clonal evolution. Cell 138, 822-829.
Shah, S.P., Roth, A., Goya, R., Oloumi, A., Ha, G., Zhao, Y., Turashvili, G., Ding, J., Tse, K., Haffari, G., et al. (2012). The clonal and mutational evolution spectrum of primary triple-negative breast cancers. Nature 486, 395-399.
Shaw, T.J., Senterman, M.K., Dawson, K., Crane, C.A., and Vanderhyden, B.C. (2004). Characterization of intraperitoneal, orthotopic, and metastatic xenograft models of human ovarian cancer. Molecular therapy : the journal of the American Society of Gene Therapy 10, 1032-1042.
Shian, S.G., Kao, Y.R., Wu, F.Y., and Wu, C.W. (2003). Inhibition of invasion and angiogenesis by zinc-chelating agent disulfiram. Molecular pharmacology 64, 1076-1084.
Shih Ie, M., and Kurman, R.J. (2004). Ovarian tumorigenesis: a proposed model based on morphological and molecular genetic analysis. The American journal of pathology 164, 1511-1518.
Shultz, L.D., Ishikawa, F., and Greiner, D.L. (2007). Humanized mice in translational biomedical research. Nature reviews Immunology 7, 118-130.
Shultz, L.D., Lyons, B.L., Burzenski, L.M., Gott, B., Chen, X., Chaleff, S., Kotb, M., Gillies, S.D., King, M., Mangada, J., et al. (2005). Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2R gamma null mice engrafted with mobilized human hemopoietic stem cells. J Immunol 174, 6477-6489.
Siddik, Z.H. (2003). Cisplatin: mode of cytotoxic action and molecular basis of resistance. Oncogene 22, 7265-7279.
Sieh, W., Salvador, S., McGuire, V., Weber, R.P., Terry, K.L., Rossing, M.A., Risch, H., Wu, A.H., Webb, P.M., Moysich, K., et al. (2013). Tubal ligation and risk of ovarian cancer subtypes: a pooled analysis of case-control studies. International journal of epidemiology 42, 579-589.
Singh, S.K., Hawkins, C., Clarke, I.D., Squire, J.A., Bayani, J., Hide, T., Henkelman, R.M., Cusimano, M.D., and Dirks, P.B. (2004). Identification of human brain tumour initiating cells. Nature 432, 396-401.
132
Sit, A.S., Modugno, F., Weissfeld, J.L., Berga, S.L., and Ness, R.B. (2002). Hormone replacement therapy formulations and risk of epithelial ovarian carcinoma. Gynecol Oncol 86, 118-123.
Skates, S.J., Xu, F.J., Yu, Y.H., Sjovall, K., Einhorn, N., Chang, Y., Bast, R.C., Jr., and Knapp, R.C. (1995). Toward an optimal algorithm for ovarian cancer screening with longitudinal tumor markers. Cancer 76, 2004-2010.
Sladek, N.E. (2003). Human aldehyde dehydrogenases: potential pathological, pharmacological, and toxicological impact. Journal of biochemical and molecular toxicology 17, 7-23.
Soling, A., and Rainov, N.G. (2003). Bioluminescence imaging in vivo - application to cancer research. Expert opinion on biological therapy 3, 1163-1172.
Steg, A.D., Bevis, K.S., Katre, A.A., Ziebarth, A., Dobbin, Z.C., Alvarez, R.D., Zhang, K., Conner, M., and Landen, C.N. (2012). Stem cell pathways contribute to clinical chemoresistance in ovarian cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 18, 869-881.
Stewart, D.J., Verma, S., and Maroun, J.A. (1987). Phase I study of the combination of disulfiram with cisplatin. American journal of clinical oncology 10, 517-519.
Stewart, J. (2013). Organizing Cellular Heterogeneity in High-Grade Serous Cancer. In Medical Biophysics (Toronto: University of Toronto), pp. 198.
Stewart, J.M., Shaw, P.A., Gedye, C., Bernardini, M.Q., Neel, B.G., and Ailles, L.E. (2011). Phenotypic heterogeneity and instability of human ovarian tumor-initiating cells. Proceedings of the National Academy of Sciences of the United States of America 108, 6468-6473.
Stewart, L.M., Holman, C.D., Aboagye-Sarfo, P., Finn, J.C., Preen, D.B., and Hart, R. (2013). In vitro fertilization, endometriosis, nulliparity and ovarian cancer risk. Gynecol Oncol 128, 260-264.
Szabova, L., Yin, C., Bupp, S., Guerin, T.M., Schlomer, J.J., Householder, D.B., Baran, M.L., Yi, M., Song, Y., Sun, W., et al. (2012). Perturbation of Rb, p53, and Brca1 or Brca2 cooperate in inducing metastatic serous epithelial ovarian cancer. Cancer research 72, 4141-4153.
Talmadge, J.E., Singh, R.K., Fidler, I.J., and Raz, A. (2007). Murine models to evaluate novel and conventional therapeutic strategies for cancer. The American journal of pathology 170, 793-804.
Taylor, H.S. (2000). The role of HOX genes in the development and function of the female reproductive tract. Seminars in reproductive medicine 18, 81-89.
Taylor, H.S., Vanden Heuvel, G.B., and Igarashi, P. (1997). A conserved Hox axis in the mouse and human female reproductive system: late establishment and persistent adult expression of the Hoxa cluster genes. Biology of reproduction 57, 1338-1345.
133
Tentler, J.J., Tan, A.C., Weekes, C.D., Jimeno, A., Leong, S., Pitts, T.M., Arcaroli, J.J., Messersmith, W.A., and Eckhardt, S.G. (2012). Patient-derived tumour xenografts as models for oncology drug development. Nature reviews Clinical oncology 9, 338-350.
The Cancer Genome Atlas Research Network, (2011). Integrated genomic analyses of ovarian carcinoma. Nature 474, 609-615.
Theodossiou, T., Hothersall, J.S., Woods, E.A., Okkenhaug, K., Jacobson, J., and MacRobert, A.J. (2003). Firefly luciferin-activated rose bengal: in vitro photodynamic therapy by intracellular chemiluminescence in transgenic NIH 3T3 cells. Cancer research 63, 1818-1821.
Theriault, C., Pinard, M., Comamala, M., Migneault, M., Beaudin, J., Matte, I., Boivin, M., Piche, A., and Rancourt, C. (2011). MUC16 (CA125) regulates epithelial ovarian cancer cell growth, tumorigenesis and metastasis. Gynecol Oncol 121, 434-443.
Tiffen, J.C., Bailey, C.G., Ng, C., Rasko, J.E., and Holst, J. (2010). Luciferase expression and bioluminescence does not affect tumor cell growth in vitro or in vivo. Molecular cancer 9, 299.
Tomasini, R., Mak, T.W., and Melino, G. (2008). The impact of p53 and p73 on aneuploidy and cancer. Trends in cell biology 18, 244-252.
Tomida, A., and Tsuruo, T. (1999). Drug resistance mediated by cellular stress response to the microenvironment of solid tumors. Anti-cancer drug design 14, 169-177.
Tone, A.A., Salvador, S., Finlayson, S.J., Tinker, A.V., Kwon, J.S., Lee, C.H., Cohen, T., Ehlen, T., Lee, M., Carey, M.S., et al. (2012a). The role of the fallopian tube in ovarian cancer. Clinical advances in hematology & oncology : H&O 10, 296-306.
Tone, A.A., Virtanen, C., Shaw, P., and Brown, T.J. (2012b). Prolonged postovulatory proinflammatory signaling in the fallopian tube epithelium may be mediated through a BRCA1/DAB2 axis. Clinical cancer research : an official journal of the American Association for Cancer Research 18, 4334-4344.
Toriola, A.T., Grankvist, K., Agborsangaya, C.B., Lukanova, A., Lehtinen, M., and Surcel, H.M. (2011). Changes in pre-diagnostic serum C-reactive protein concentrations and ovarian cancer risk: a longitudinal study. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 22, 1916-1921.
Trimbos, B., Timmers, P., Pecorelli, S., Coens, C., Ven, K., van der Burg, M., and Casado, A. (2010). Surgical staging and treatment of early ovarian cancer: long-term analysis from a randomized trial. Journal of the National Cancer Institute 102, 982-987.
Tung, K.H., Wilkens, L.R., Wu, A.H., McDuffie, K., Nomura, A.M., Kolonel, L.N., Terada, K.Y., and Goodman, M.T. (2005). Effect of anovulation factors on pre- and postmenopausal ovarian cancer risk: revisiting the incessant ovulation hypothesis. American journal of epidemiology 161, 321-329.
134
Twickler, D.M., and Moschos, E. (2010). Ultrasound and assessment of ovarian cancer risk. AJR American journal of roentgenology 194, 322-329.
Vaisman, A., Lim, S.E., Patrick, S.M., Copeland, W.C., Hinkle, D.C., Turchi, J.J., and Chaney, S.G. (1999). Effect of DNA polymerases and high mobility group protein 1 on the carrier ligand specificity for translesion synthesis past platinum-DNA adducts. Biochemistry 38, 11026-11039.
Van Dalen, A., Favier, J., Baumgartner, L., Hasholzner, U., De Bruijn, H., Dobbler, D., Dombi, V.H., Fink, D., Giai, M., McGing, P., et al. (2000a). Serum levels of CA 125 and TPS during treatment of ovarian cancer. Anticancer research 20, 5107-5108.
van Dalen, A., Favier, J., Burges, A., Hasholzner, U., de Bruijn, H.W., Dobler-Girdziunaite, D., Dombi, V.H., Fink, D., Giai, M., McGing, P., et al. (2000b). Prognostic significance of CA 125 and TPS levels after 3 chemotherapy courses in ovarian cancer patients. Gynecol Oncol 79, 444-450.
van der Horst, G., van Asten, J.J., Figdor, A., van den Hoogen, C., Cheung, H., Bevers, R.F., Pelger, R.C., and van der Pluijm, G. (2011). Real-time cancer cell tracking by bioluminescence in a preclinical model of human bladder cancer growth and metastasis. European urology 60, 337-343.
Vanderhyden, B.C., Shaw, T.J., and Ethier, J.F. (2003). Animal models of ovarian cancer. Reproductive biology and endocrinology : RB&E 1, 67.
Vang, R., Shih Ie, M., and Kurman, R.J. (2009). Ovarian low-grade and high-grade serous carcinoma: pathogenesis, clinicopathologic and molecular biologic features, and diagnostic problems. Advances in anatomic pathology 16, 267-282.
Vang, R., Shih Ie, M., and Kurman, R.J. (2013). Fallopian tube precursors of ovarian low- and high-grade serous neoplasms. Histopathology 62, 44-58.
Vaughan, S., Coward, J.I., Bast, R.C., Jr., Berchuck, A., Berek, J.S., Brenton, J.D., Coukos, G., Crum, C.C., Drapkin, R., Etemadmoghadam, D., et al. (2011). Rethinking ovarian cancer: recommendations for improving outcomes. Nature reviews Cancer 11, 719-725.
Verhaak, R.G., Tamayo, P., Yang, J.Y., Hubbard, D., Zhang, H., Creighton, C.J., Fereday, S., Lawrence, M., Carter, S.L., Mermel, C.H., et al. (2013). Prognostically relevant gene signatures of high-grade serous ovarian carcinoma. The Journal of clinical investigation 123, 517-525.
Wang, L.G., Liu, X.M., Kreis, W., and Budman, D.R. (1999). The effect of antimicrotubule agents on signal transduction pathways of apoptosis: a review. Cancer chemotherapy and pharmacology 44, 355-361.
Wang, W., McLeod, H.L., and Cassidy, J. (2003). Disulfiram-mediated inhibition of NF-kappaB activity enhances cytotoxicity of 5-fluorouracil in human colorectal cancer cell lines. International journal of cancer Journal international du cancer 104, 504-511.
135
Wang, Y., Cheon, D.J., Lu, Z., Cunningham, S.L., Chen, C.M., Luo, R.Z., Xing, D., Orsulic, S., Bast, R.C., Jr., and Behringer, R.R. (2008). MUC16 expression during embryogenesis, in adult tissues, and ovarian cancer in the mouse. Differentiation; research in biological diversity 76, 1081-1092.
Ward, B.G., Wallace, K., Shepherd, J.H., and Balkwill, F.R. (1987). Intraperitoneal xenografts of human epithelial ovarian cancer in nude mice. Cancer research 47, 2662-2667.
Watson, P., Butzow, R., Lynch, H.T., Mecklin, J.P., Jarvinen, H.J., Vasen, H.F., Madlensky, L., Fidalgo, P., and Bernstein, I. (2001). The clinical features of ovarian cancer in hereditary nonpolyposis colorectal cancer. Gynecol Oncol 82, 223-228.
Watson, P., Vasen, H.F., Mecklin, J.P., Bernstein, I., Aarnio, M., Jarvinen, H.J., Myrhoj, T., Sunde, L., Wijnen, J.T., and Lynch, H.T. (2008). The risk of extra-colonic, extra-endometrial cancer in the Lynch syndrome. International journal of cancer Journal international du cancer 123, 444-449.
Wernli, K.J., Newcomb, P.A., Hampton, J.M., Trentham-Dietz, A., and Egan, K.M. (2008). Inverse association of NSAID use and ovarian cancer in relation to oral contraceptive use and parity. British journal of cancer 98, 1781-1783.
Whiteman, D.C., Murphy, M.F., Cook, L.S., Cramer, D.W., Hartge, P., Marchbanks, P.A., Nasca, P.C., Ness, R.B., Purdie, D.M., and Risch, H.A. (2000). Multiple births and risk of epithelial ovarian cancer. Journal of the National Cancer Institute 92, 1172-1177.
Whittemore, A.S., Harris, R., and Itnyre, J. (1992). Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. II. Invasive epithelial ovarian cancers in white women. Collaborative Ovarian Cancer Group. American journal of epidemiology 136, 1184-1203.
Winter, W.E., 3rd, Maxwell, G.L., Tian, C., Carlson, J.W., Ozols, R.F., Rose, P.G., Markman, M., Armstrong, D.K., Muggia, F., and McGuire, W.P. (2007). Prognostic factors for stage III epithelial ovarian cancer: a Gynecologic Oncology Group Study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25, 3621-3627.
World Helath Organization, (2008). World Cancer Report 2008. In World Health Organization P. Boyle, and B. Levin, eds.
Xing, D., and Orsulic, S. (2006). A mouse model for the molecular characterization of brca1-associated ovarian carcinoma. Cancer research 66, 8949-8953.
Xu, Y., Silver, D.F., Yang, N.P., Oflazoglu, E., Hempling, R.E., Piver, M.S., and Repasky, E.A. (1999). Characterization of human ovarian carcinomas in a SCID mouse model. Gynecol Oncol 72, 161-170.
Yachida, S., Jones, S., Bozic, I., Antal, T., Leary, R., Fu, B., Kamiyama, M., Hruban, R.H., Eshleman, J.R., Nowak, M.A., et al. (2010). Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 467, 1114-1117.
136
Yap, T.A., Carden, C.P., and Kaye, S.B. (2009). Beyond chemotherapy: targeted therapies in ovarian cancer. Nature reviews Cancer 9, 167-181.
Yen, M.L., Yen, B.L., Bai, C.H., and Lin, R.S. (2003). Risk factors for ovarian cancer in Taiwan: a case-control study in a low-incidence population. Gynecol Oncol 89, 318-324.
Young, R.C., Decker, D.G., Wharton, J.T., Piver, M.S., Sindelar, W.F., Edwards, B.K., and Smith, J.P. (1983). Staging laparotomy in early ovarian cancer. JAMA : the journal of the American Medical Association 250, 3072-3076.
Young, R.C., Walton, L.A., Ellenberg, S.S., Homesley, H.D., Wilbanks, G.D., Decker, D.G., Miller, A., Park, R., and Major, F., Jr. (1990). Adjuvant therapy in stage I and stage II epithelial ovarian cancer. Results of two prospective randomized trials. The New England journal of medicine 322, 1021-1027.
Zabala, M., Alzuguren, P., Benavides, C., Crettaz, J., Gonzalez-Aseguinolaza, G., Ortiz de Solorzano, C., Gonzalez-Aparicio, M., Kramer, M.G., Prieto, J., and Hernandez-Alcoceba, R. (2009). Evaluation of bioluminescent imaging for noninvasive monitoring of colorectal cancer progression in the liver and its response to immunogene therapy. Molecular cancer 8, 2.
Zeimet, A.G., Muller-Holzner, E., Marth, C., and Daxenbichler, G. (1994). Immunocytochemical versus biochemical receptor determination in normal and tumorous tissues of the female reproductive tract and the breast. The Journal of steroid biochemistry and molecular biology 49, 365-372.
Zhang, S., Balch, C., Chan, M.W., Lai, H.C., Matei, D., Schilder, J.M., Yan, P.S., Huang, T.H., and Nephew, K.P. (2008). Identification and characterization of ovarian cancer-initiating cells from primary human tumors. Cancer research 68, 4311-4320.
Zhang, S., Royer, R., Li, S., McLaughlin, J.R., Rosen, B., Risch, H.A., Fan, I., Bradley, L., Shaw, P.A., and Narod, S.A. (2011). Frequencies of BRCA1 and BRCA2 mutations among 1,342 unselected patients with invasive ovarian cancer. Gynecol Oncol 121, 353-357.
Zhou, B.B., Zhang, H., Damelin, M., Geles, K.G., Grindley, J.C., and Dirks, P.B. (2009). Tumour-initiating cells: challenges and opportunities for anticancer drug discovery. Nature reviews Drug discovery 8, 806-823.
Zografos, G.C., Panou, M., and Panou, N. (2004). Common risk factors of breast and ovarian cancer: recent view. International journal of gynecological cancer : official journal of the International Gynecological Cancer Society 14, 721-740.
137
Appendix 1: Establishment of an Orthotopic Bioluminescent Patient-Derived Xenograft Model of
Ovarian Cancer
138
A1 Introduction
To elucidate key determinants in ovarian cancer initiation, progression and response to treatment,
models that recapitulate the disease are required. This demands both in vitro and, more
importantly, in vivo models that reflect the cell biology and heterogeneity of the disease. Most
studies of ovarian carcinogenesis and drug response utilize immortalized cancer cell lines as in
vitro tumour models; however, many of these ‘ovarian HGSC’ cell lines are either not actually
serous histology, do not reproduce serous histology when implanted into immune-compromised
mice or do not give rise to xenografts at all (Domcke et al., 2013).
Few animal models develop spontaneous ovarian tumours, with the exception being hens and
macaques; however, using these for drug development is not feasible given their cost and/or lack
of appropriate reagents (Ricci et al., 2013). As a consequence, mice have become the primary
mammalian model in cancer research due to their short generation time, ability to bear large
litters, relative ease of breeding, and advances in mouse genomics (Edinger et al., 2002). Two
main types of mouse have been utilized for cancer research: xenografts and genetically
engineered mouse models (GEMMs).
Engrafting human tumour cells in immunodeficient mice to generate xenografts has been
established for over 30 years (Kung, 2007). Typically, these models have used immortalized cell
lines that are then injected subcutaneously (SC) into mice. To complement xenografts, GEMMs
were developed to provide an efficient system in which to analyze the specific roles of
oncogenes and tumour suppressor genes in tumourigenesis in vivo (Mullany and Richards, 2012;
Talmadge et al., 2007). The extent to which either of these systems represent the heterogeneity
of HGSC is unclear, and their role in drug development has proven limited (Talmadge et al.,
2007).
Optimal preclinical models are those that reduce or eliminate the inconsistency between the
original patient tumour and the mouse (Jin et al., 2010). Patient derived xenografts (PDXs) do so
by utilizing patient tissue obtained at the time of surgery, as opposed to cells or cell lines grown
in culture. As a result, PDX models maintain the original tumour architecture, and histological
characteristics, as well as key genes and signaling pathways (Gray et al., 2004). This has been
139
shown in lung, colorectal, and pancreatic cancer PDX models, where a high degree of genetic
similarity between the primary cancer and the corresponding PDX tumour was seen (Daniel et
al., 2009; Fichtner et al., 2008; Jones et al., 2008; Tentler et al., 2012).
A limitation of many PDX models is the location of tumour formation. The majority of xenograft
studies inoculate cells or implant tumours ectopically into the subcutaneous layer or into the
mammary fat pad (i.e., not in their anatomical site of origin). These animal models typically rely
on caliper measurements, which are relatively rapid and simple to perform and allow easy
monitoring of disease progression. Nevertheless, there are limitations to this approach;
experiments based on caliper data do not account for areas of necrosis and edema, and so do not
necessarily assess the effect of treatment on the number of viable cells(Jenkins et al., 2003a;
Jenkins et al., 2003b). Although SC/MFP PDXs recapitulate many aspects of the tumour,
orthotopic implantation of tumour cells greatly enhances metastasis, which some argue is
required to make PDXs most predictive of clinical response (Kerbel, 2003).
Orthotopic ovarian xenografts (injected into the rodent ovarian bursa) provide the appropriate
microenvironment to influence cell behavior, these xenografts provide a challenge with respect
to monitoring disease initiation, progression and recurrence (Shaw et al., 2004). Unlike
orthotopic xenografts derived from cell lines, orthotopic PDXs produce diffuse ascites and lack
ovarian tumours. Furthermore, similar to patients, mice do not show signs of disease prior to the
development of discernable ascites, which represents advanced stage (Stewart, unpublished
results). This makes the ovarian orthotopic PDXs difficult to use for monitoring response to
therapy.
A newer method for in vivo assessment involves the use of bioluminescent reporter genes and
bioluminescence imaging (BLI). This method allows sensitive and quantitative detection of cells
non-invasively in small research animals (Edinger et al., 2002). Bioluminescence refers to the
enzymatic generation of visible light and the firefly luciferase gene is the most commonly used
in animal models. Luciferase oxidizes luciferin in the presence of ATP and oxygen to form an
electronically excited oxy-luciferin (Figure 4.0). Light is emitted following relaxation of the
excited oxy-luciferin to its ground state, and BLI relies on external detection of this signal
(O'Neill et al., 2010). Light emission is proportional to tumour volume, increases as the cell
140
population multiplies and is derived solely from metabolically active, transformed tumour cells
(Rehemtulla et al., 2000). The main obstacle with BLI is the scattering of light by tissues, which
make quantification of photons more difficult. Additionally, there is reduced luciferase activity
in large tumours due to hypoxia and necrosis (Soling and Rainov, 2003).
Figure 4.0 The luciferin reaction.
One group was able to generate a luc-expressing primary breast cancer PDX model and was able
to detect early metastases (Harrison et al., 2010). Several groups have done this with a variety of
other cancers, including ovarian, but all used immortalized cell lines (Cordero et al., 2010; van
der Horst et al., 2011; Zabala et al., 2009). I wanted to exploit this technology to derive luc-
expressing ovarian HGSC cells.
Luciferin ATP Oxyluciferin + Light
Luciferase
141
A2 Methods
Virus Preparation
Approximately 8 hours before transfection, 293T cells were seeded and incubated in a 15 cm
dish at 18 x 106 cells. Media was changed prior to transfection with IMDM, 10% FBS, Penicillin
(100U/ml), Streptomycin (100U/ml) and Glutamine (22.5 ml final volume). A plasmid DNA mix
was prepared by adding: ENV plasmid (VSV-G), packaging plasmid (CMV ΔR8.74), REV
plasmid, and transfer vector plasmid (courtesy of Erno Weinholds). Finally, 125μl of 2.5M CaCl2
was added and incubated at room temperature for 5 minutes. The GFP-Luciferase plasmid was
generously provided by Dr. Joseph Wu (Stanford University). The precipitate was formed by the
drop-wise addition of 1.25ml of 2X HBS solution to the DNA-TE-CaCl2 mixture while vortexing
at full speed. The CaPi-precipitated plasmid DNA was allowed to stay on the cells for 14 hours,
after which the media was replaced with fresh media for virus collection to begin. The cell
supernatant was collected 30 hours after changing the media.
Ascites Samples Preparation
Human ascites specimens were obtained with informed consent from patients undergoing
surgery at PMH, following a protocol approved by the University Health Network Research
Ethics Board (REB #06-0903T). Primary, bulk, chemo-naive ascites was plated in 75mL vented
flasks, at a ratio of 1:1 with media (RPMI with 20% FBS, Penicillin (100U/ml), Streptomycin
(100U/ml) and amphotericin B (2 µg/mL) (Qiagen). Cells were allowed to adhere for 48-72
hours, after which the media was changed every 48 hours. Once the cells were actively
proliferating (20-40 days after plating), high titer virus was added for 24 hours. Cells were then
washed, and collected by centrifugation. A small aliquot was removed for fluorescence-activated
cell sorting assessment.
Orthotopic Xenograft Establishment
All experiments involving mice were carried out under the UHN animal protocol #1239, and
142
utilized 6-8 week-old female NSG mice. Aliquots of 105 luc-ascites tumour cells were mixed 1:1
by volume with Matrigel for mouse injections. Mice were anaesthetized using isoflorane vapor.
The skin was disinfected using an alcohol swab. A dorsolateral 1 cm incision was made on the
top of the spleen. Using curved forceps, the fat pad surrounding the ovary was isolated and the
cell suspension (10μL) was injected into the ovarian bursa. An absorbable suture was used to
suture the skin. Sterile saline (300cc) was administered SC, and the animals were given oral
antibiotics post-operatively for one week.
In vivo Imaging
72 hours after orthotopic injection, in vivo imaging was performed using the Xenogen IVIS
Imaging system. Mice were injected IP with 250 μL of 5mg/mL d-luciferin dissolved in PBS
(50mg/kg), placed in the anesthesia chamber and imaged 20 minutes following injection (based
on a previously determined luciferin kinetics curve). Imaging was repeated at 7, 30, 60, 90 and
120 days, at which point the mice were sacrificed. Total photon flux (in regions of interest) was
measured in photons/second (p/s), using IVIS Living Image ® software.
Histology
Tumours were collected, formalin fixed and paraffin-embedded. Slides were obtained and
stained for H&E. Immunohistochemistry using anti-GFP antibody (Abcam, cat. No. #6556) was
performed. All slides were reviewed by a gynecologic pathologist.
143
A3 Results
Cultured ascites cells show epithelial morphology and are readily
transduced
Sample 2844 was grown in culture for 40 days prior to the addition of high-titre virus. At the
time virus was added, the cells were rapidly proliferating and showed epithelial morphology
(Figure 4.1A) Transduction with lenti-virus generated 85.4%-GFP positive cells (Figure 4.1B).
Sample 3393 was cultured for 20 days prior to lenti-virus, was rapidly proliferating and showed
epithelial morphology (Figure 4.1C). Viral transduction yielded 89.3% GFP-positive cells
(Figure 4.1D). Once cells were GPF-positive, they were no longer maintained in culture and
were injected immediately into mice.
PDXs Show a Luciferin Signal 72 Hours after Inoculation
Mouse imaging was performed 72 hours after inoculation. Figure 4.2 shows that both mice
generated from samples 2844 and 3393 showed a positive luciferase signal, however the region
of interest (ROI) measurement was considerably stronger for 3393 PDXs (maximum total photon
flux of 1e6 p/s in 2844 compared with 2e8 p/s in 3393). Similarly, the signal was stronger for
luc-3393 IP PDXs, compared with luc-2844 (1e9 versus 2e8 p/s) (Figure 4.3).
A)
B)
10x
144
C)
D)
Figure 4.1. Morphology of cultured ascites cells and flow cytometer plot showing percent GFP transduction. A) 2844 ascites after 40 days of culture, B) Flow cytometer plot showing 85.4% transduction of sample 2844. C) Phase contrast of sample 3393 showing after 20 days in culture showing epithelial morphology and D) FACS graph showing 89.3% transfection efficiency.
A)
B)
Figure 4.2 Intrabursal PDXs 72 hours after inoculation with luc- ascites cells. A) Luc-2844 IB PDXs showing regions of interest measuring 5e5,1e6 and 3e5 p/s, respectively. B) Luc-3393 IB PDXs showing regions of interest. Mouse 3393-1 ROI measures 2e8 and mouse 3393-2 measures 6e7 p/s.
10x
145
A)
B)
C)
D)
Figure 4.3 Intraperitoneal PDXs 72 hours after inoculation with luc-ascites cells. Luc-2844 IP PDXs showing regions of interest measuring A) 2e8 p/s and B) 8e6 p/s. Luc-3393 IP PDXs showing regions of interest measuring C) 1e9 p/s and D) 3e7.
Figure 4.4 Luciferase signal measured over time in luc-2844 IP PDXs. Images were taken at A) 30 days, B) 60 days, C) 90 days and D) 120 days (experiment termination)
A
146
Figure 4.5 Luciferase signal measured over time in luc-3393 IP PDXs. Images were taken at A) 30 days, B) 60 days, C) 90 days and D) 120 days (experiment termination)
Table 4.1 Initial and final ROI values for 2844 and 3393 IP PDXs
IP PDXs ROI initial (p/s) ROI final (p/s)
2844 -1 2.0e8 3.3e5
2844 -2 8.5e6 4.2e5
3393 -1 1.2e9 3.1e5
3393 -2 3.2e7 2.8e5
Decreased luciferase signal over time in IP and IB PDXs
Mouse imaging was performed every 30 days. Figures 4.4 and 4.5 show the progressive
weakening of the luciferase signal over time for both 2844 and 3393 IP PDXs. Initial and final
ROI measurements are listed in table 4.1. Although 2844 and 3393 IP PDXs started out with
147
considerably different initial ROI measurements, at the termination of the experiment, their
respective ROIs were all similar (~3e5p/s).
Similar to the IP PDxs, IB PDXs showed a strong signal at 72 hours (Figures 4.1A,B); however
this signal faded over time (Figures 4.6 and 4.7). At the completion of the experiment, no luc-
2844 PDXs generated a luciferase signal. A weak signal could be seen in two luc-3933 IB PDXs
(Figure 4.7), corresponding to an ROI of 1-2e5.
Figure 4.9 shows a DLIT image of PDXs luc-3393-1 at 72 hours (panels A-D) and at endpoint
(panels E-H). Diffuse Light Imaging Tomography (DLIT) utilizes the data obtained from a
filtered 2D bioluminescent sequence (by measuring the effects of tissue absorption at different
wavelengths to determine depth and flux) in combination with a surface topography to represent
the bioluminescent source in a 3D tissue.
Figure 4.6 Luciferase signal measured over time in luc-2844 IB PDXs. Images were taken at A) 30 days, B) 60 days, C) 90 days and D) 120 days (experiment termination)
148
Figure 4.7 Luciferase signal measured over time in luc-3393 IB PDXs. Images were taken at A) 30 days, B) 60 days, C) 90 days and D) 120 days (experiment termination). ROIs remained largely unchanged from 30 days to 120 days.
149
Figure 4.8 Luciferase signal measured over time in luc-3393-1 IB PDX. Images were taken at A) 72 hours after inoculation and B) 120 days (experiment termination). The initial ROI 72 hours after inoculation was 1.7e8 and faded to 3.8e5 by 4 months.
150
Figure 4.9. 3D-Reconstruction of bioluminescence images (DLIT) of luc-3393-1 IB PDX. A), B), C) are coronal, sagittal and transaxial cuts, respectively, of the PDX taken at 72 hours after inoculation. D) Represents a 3D reconstructive image of luc-3393-1 at 72 hours. E-G) are coronal, sagittal and transaxial cuts, respectively, of the PDX taken at endpoint 4 months. H) 3D reconstruction of luc-3393-1 at 4 months with luciferase signal in the area of the adnexa. Digital mouse ovaries are registered onto the image by the software, and are meant to represent the location of ovaries in a typical mouse.
151
IB PDXs have an ‘Ovarian Mass’ at Necropsy
All IP and IB xenografts were carefully dissected looking for evidence of adnexal and extra-
ovarian disease. No IP xenograft showed evidence of visible disease. Conversely, all IB
xenografts, even those that did not have a luciferase signal, had an ovarian/adnexal mass on the
left side (where cells were inoculated) (representative image shown of luc-3393-1, Figure 4.10).
All other abdominal organs were examined and did not show gross evidence of disease. No
xenografts had ascites.
Figure 4.10 Necroscopic examination of luc-3393-1 IB xenograft. At initial dissection (A) a large ovarian mass could be seen. All other organs appeared normal. Comparison of normal (B) right ovary and fallopian tube (FT) compared to ‘ovarian mass’ arising from inoculated left ovary (C).
152
Histology
Despite the ‘ovarian mass’ seen at necropsy, histological examination by H&E (Figure 4.11)
showed no abnormal or tumour cells within the ‘mass’, ovary, or fallopian tube.
Figure 4.11 Histologic examination of PDX luc-3393-1 IB. H&E stain at A) 4x and B) 10x show a normal fallopian tube and ovary with no evidence of high-grade serous histology. GFP staining of xenograft luc-3393-1 IB fallopian tube and ovary at C) 4x and D) 10x magnifications show no positive GFP cells.
153
A4 Discussion
Nearly all primary and recurrent ovarian HGSC presents at advanced stage. Knowledge of the
mechanisms of metastasis, and the development of technologies to detect them are critical. Pre-
clinical models that more closely mirror human disease have an important role in evaluating
putative cancer therapies (Kung, 2007). Clinically relevant animal models that incorporate:
disease heterogeneity and allow monitoring of cancer growth/metastasis and treatment efficacy
continue to be lacking.
The high sensitivity of BLI has been demonstrated in several studies and its main strength is the
ability to monitor tumour load longitudinally. However, it is poorly suited for determination of
absolute tumour mass and the exact locations of tumours due to its limited spatial resolution
(Klerk et al., 2007). The sensitivity of BLI is dependent on several factors, such as the number of
cells expressing the reporter gene, the efficiency of the promoter and the stability of luciferase
expression (Sato et al., 2004). Although the relationship between the bioluminescence signal and
viable cancer cell load has been validated in several cancer cell line studies, a strong correlation
in orthotopic models, particularly those with larger tumours, warrants further investigation
(Zabala et al., 2009). Zabala et al., showed that larger tumours tend to have less copies of
luciferase plasmid per genome, suggesting that stromal cells are more abundant in these tumours
thereby making the BLI signal not representative of true tumour burden (Zabala et al., 2009).
Several barriers in obtaining reproducible quantification of BLI have been identified. Cui et al.,
found that the concentration of luciferin injected, the kinetics of luciferin (from time of injection
to maximal signal) and light scattering inside the animal (due to anatomical factors, tissue depth,
signal impedance, etc.) can all vary the efficiency of light transmission. In addition, animal
positioning (prone versus supine) and even the type of anaesthetic used can generate significant
variability in the BLI signal (Cui et al., 2008; Sato et al., 2004).
Perhaps the major disadvantage of BLI is that cells must be genetically modified to carry the
luciferase gene. The effects of gene transfection on cancer cell health, behavior, and
tumourogenicity are not well defined, and as such, this makes a direct translation of data difficult
(Klerk et al., 2007). Tiffen et al., found that a high level of luciferase expression did not have
154
detrimental effects on cancer cell growth in vitro or in vivo; however this experiment used cell
lines and not primary cells (Tiffen et al., 2010). Other reports suggest that production of
intracellular light is cytotoxic to cancer cells (Brutkiewicz et al., 2007; Theodossiou et al., 2003).
Furthermore, there could also be effects of serum-containing media on ascites cells. I attempted
to expose the cells minimally to culture medium, so virus was added at various times to multiple
ascites samples (data not shown). Short periods of culture(< 1 week), resulted in no transduction;
high transduction was observed only when cells were in culture for 2-4 weeks. Effects of
prolonged culture on primary cells is undefined, but may affect tumourigenicity. This might
explain why no tumour cells were seen by histology. This is supported by the control mice,
which were inoculated with cells grown in culture but without virus. No tumours were seen in
these mice suggesting perhaps that the in vitro culture conditions reduce and/or eliminate
tumourigenicity.
In our study, a strong luciferase signal was generated in both intrabursal and intraperitoneal luc-
2844 and luc-3393 PDXs 72 hours after inoculation, however this signal dissipated over time,
and at the endpoint of my study, I only had two xenografts that maintained a positive, albeit
weak, luciferin signal (intrabursal luc-3393-1 and luc-3393-2). These mice appeared to have an
‘ovarian mass’ at the time of necropsy, however, histologically, there was no evidence of HGSC
or any abnormal cells. Its possible that the initial strong signal was generated from a small
population of cells with multiple viral genome integrations. Over time, it appears that these cells,
although viable, are unable to cause tumours. Optimization of this protocol may allow for
monitoring of tumour burden over time and in response to various treatments.
155
Appendix 2: Disulfiram-Induced Toxicity in Ovarian Cancer
156
A2.1 Introduction
Ovarian cancer is the sixth most common cancer in women and the most lethal gynecologic
malignancy (Institute, 2013). Approximately 80% of ovarian cancers have high-grade serous
histology: these appear in the absence of a well-defined precursor, present at an advanced stage
and are highly aggressive (Kurman and Shih Ie, 2010). They metastasize early in the disease
course, with tumour generally found on the ovarian surface, fallopian tubes and widely
disseminated throughout the peritoneal cavity (Crum et al., 2007b; Shih Ie and Kurman, 2004).
Despite improvements in debulking surgery and generally good initial responses to
chemotherapy, five-year survival for these patients has remained largely unchanged.
Following initial therapy with Carboplatin/Cisplatin and Taxol, more than 85% of patients will
recur. The challenge in treating ovarian HGSC is that recurrences tend to be drug-resistant,
thereby making treatment options more limited and less effective (Kulkarni-Datar et al., 2013).
Thus, novel therapies are needed for patients with both newly diagnosed and relapsed ovarian
HGSC; however, the development of new drugs is lengthy and costly. It has been estimated that
it takes an average of 15 years and US$800 million to bring a single drug to market and only 20-
30 new compounds are approved annually, many of which are not chemotherapeutics (Chong
and Sullivan, 2007). Therefore, repurposing existing drugs prescribed for treating other diseases
is a potentially useful approach that circumvents this lengthy and costly FDA process.
Combining currently employed chemotherapeutics with existing compounds could be an
important strategy in enhancing drug efficacy.
Disulfiram (Tetraethylthiuram disulfide; Antabuse®) is a commercially available drug that has
been used for decades as aversion therapy in the treatment of alcohol abuse (Sauna et al., 2005).
Ethanol is converted to acetaldehyde (AA) by alcohol dehydrogenase, but is further metabolized
to acetate by aldehyde dehydrogenase (ALDH) (Figure 5.0). While many other members of the
aldehyde dehydrogenase superfamily are capable of catalyzing the oxidation of AA to acetic
acid, ALDH2 is thought to be the most important (Sladek, 2003). Disufliram (DSF) reportedly
has numerous enzymatic targets, but it works primarily through irreversibly inhibiting ALDH2
(Eneanya et al., 1981). The presence of DSF and ethanol causes an accumulation of acetate,
157
yielding an unpleasant physical response (flushing, tachycardia, nausea) and thus, an aversion
reaction (Gaval-Cruz and Weinshenker, 2009).
Figure 3.0 Ethanol metabolism and disulfiram mechanism of action.
Given its long history, the toxicity and pharmacological properties of DSF have been well
described. DSF is absorbed after oral ingestion and rapidly reduced to its corresponding thiol
metabolite, diethyldiethiocarbamic acid (DDC). It is believed that both DDC and DSF are
effective at inhibiting ALDH (Eneanya et al., 1981). In addition, DDC is a potent copper
chelator and can affect the activity of copper-dependent enzymes, such as cytochrome oxidases
(Gaval-Cruz and Weinshenker, 2009). Several studies have shown that there is marked
intersubject variability in plasma concentrations of DSF and its metabolites, likely caused by its
high lipid solubility (Faiman et al., 1984).
Numerous studies have shown that DSF has antitumour activity against a number of cancer cell
lines, including melanoma, colorectal, prostate and breast (Brar et al., 2004; Cen et al., 2002;
Daniel et al., 2005; Wang et al., 2003). Although the exact mechanisms of action of DSF have
not been well characterized, it is believed to exert anti-tumour effects through via a multitude of
mechanisms, including reduction of angiogenesis, inhibition of DNA topoisomerases and NFκB
and mitochondrial membrane permeabilization (Table 5.1). Lovborg et al., found DSF to be very
active against ovarian cancer and breast cancer cell lines and patient samples in vitro and
claimed that concentrations needed to induce cytotoxicity in patients can be safely reached
(Lovborg et al., 2006). Furthermore, several groups have investigated these effects in vivo. Using
Disulfiram
ADH ALDH
Ethanol Acetaldehyde Acetate
DNA Damage
158
MDA-MB-231 breast cancer xenografts, Chen et al. showed that DSF inhibited tumour growth
by up to 74% compared with controls (Chen et al., 2006). In addition, several studies
demonstrated that DSF and its metabolites potentiate the effects of existing anticancer drugs,
including cyclophosphamide, colchicine and 5-FU (Wang et al., 2003). Most recently, Liu et al.
showed that DSF may reverse chemo-resistance and have a direct effect on CSC (Liu et al.,
2013).
In light of the evidence of DSF’s anti-cancer activity, a phase I study was initiated in 1987 in an
effort to increase the therapeutic index of cisplatin in a variety of cancers (prostate, lung, thyroid,
adrenal, pharynx, melanoma, adenocarcinoma of unknown primary). Acute, reversible confusion
was the only noted side effect of DSF, and this was only seen in patients treated with high doses
(3,000mg/m2). Furthermore, no toxic effects of cisplatin were observed by the addition of DSF.
Based on these promising results, further studies were to be initiated (Stewart et al., 1987).
Table 3.1. Published effects of disulfiram.
Cell type (Author) Reported Mechanisms of Action
Colorectal Cancer Cell lines (Wang et al., 2003) Inhibits activation of NFkB
Melanoma cell lines (Cen et al., 2002; Morrison et al., 2010)
Induces apoptosis via redox related mitochondrial membrane permeabilization; production of ROS
Lung & bladder cancers cell lines (Shian et al., 2003)
Inhibits matrix metalloproteinases
Prostate cancer cell lines (Lin et al., 2011) DNA methyltransferase inhibitor
Embryonic Kidney Cells (Loo and Clarke, 2000) Blocks maturation of p-glycoprotein membrane pump
Breast Cancer cell lines (Chen et al., 2006) Inhibits proteosomal activity
Endothelial cells (Marikovsky et al., 2003) Reduces angiogenesis
Aldehydes are not only the main product of the disulfiram+ethanol reaction, but are also
produced endogenously as a byproduct of cellular metabolism. Numerous studies have shown
159
that cells exposed to AA accumulate DNA damage. AA generates several DNA adducts that can
block DNA replication and cause DNA interstrand crosslinks (ICLs), which might play a role in
the pathogenesis of some cancers (Brooks et al., 2009); (Abraham et al., 2011). The Fanconi
Anemia-BRCA (FA-BRCA) DNA repair pathway has a crucial role in counteracting AA induced
genotoxicity and is required for cellular resistance to the toxic effects of AA (Langevin et al.,
2011). The FA-BRCA DNA repair pathway comprises 13 proteins that act together to repair
DNA ICLs and respond to stalled replication forks (Abraham et al., 2011). It is also essential to
counteract aceteladehyde-induced genotoxcity in mice, as shown by Langevin et al. In their
study, Aldh2-/-Fancd2-/- mouse embryos were extremely sensitive to ethanol exposure in utero
and rapidly developed bone marrow failure post-natally (Langevin et al., 2011).
The effects seen in the absence of a functional FA-BRCA pathway, combined with ALDH2
deficiency, as demonstrated by Langevin et al., provided the rationale for this study. One of the
FA proteins frequently mutated in ovarian cancer is BRCA2 (FANCD1), and it is important in
repair through homologous recombination. Similarly, BRCA1 also complexes with three FA
proteins to coordinate HR repair (D'Andrea, 2010; Kee and D'Andrea, 2010). Combining
germline and somatic, BRCA1 and BRCA2 mutations are found in approximately 20-50% of
ovarian HGSCs (Castellarin et al., 2013; The Cancer Genome Atlas Research Network, 2011;
Pal et al., 2005; Press et al., 2008a). Therefore, I reasoned that cells defective in homologous
recombination (BRCA2null: PEO1 cell line) would be more susceptible to the toxic effects of
increased levels of acetaldehyde, which could be achieved by inhibiting ALDH2, via DSF. In
combination with platinum, DSF causes increased DNA adducts and should enhance the effects
of platinum, particularly in the BRCA2null cells. In this study, I investigate the actions of DSF in
combination with Carboplatin in a panel of ovarian cancer cell lines and in PDXs.
160
A2.2 Methods
A2.2.1 Cell Culture
The established human ovarian cancer cell lines OV 90, OV 1946, OV 1369, TOV 1369, PEO1,
PEO4, PEO23, TOV 2223g were purchased from ATCC (or obtained from R. Rottapel) and used
for the studies. OV90 in DMEM with 10% FBS, Penicillin (100U/ml), Streptomycin (100U/ml);
PEO1, PEO4, PEO23 were cultured in RPMI containing 10% FBS, Penicillin (100U/ml),
Streptomycin (100U/ml), and OV 1369, TOV 1369, OV 1946, TOV 2223G were cultured in
DMEM 10% FBS, Penicillin (100U/ml), Streptomycin (100U/ml). Cells were incubated at 37°C
and 5% CO2 tissue culture incubator.
For drug testing, 5-10 x103 cells/well were culture in 96-well plates overnight and treated with
indicated doses of Carboplatin or DSF for 72 hours. Carboplatin was dissolved directly in media;
Tetraethylthiuram disulfide (DSF) (Sigma Aldrich) was dissolved in DMSO (1000x) and added
to media (1:1000). IC50 values were determined by Alamar Blue assay (Invitrogen, Burlington,
ON, Canada). At 72 hours, Alamar Blue was added (10% of total volume) and incubated for 4
hours. The fluorescence was measured using a spectrophotometer with an excitation setting of
530 nm and emission at 590 nm (Spectra MAX Gemini EM, Molecular Devices).
A2.2.2 Combination Studies
For combination treatment, cells cultured overnight were treated with various concentrations of
Carboplatin/DSF at a ratio of 1:1, based on IC50 data generated from previous experiments. After
72 hours of incubation, cells were subjected to an Alamar Blue assay. The combination index
(CI) was calculated using CalcuSyn Software (Biosoft, Cambridge, UK) CI<1 and CI>1
indicates synergism or antagonism, respectively, and CI=1 indicates an additive effect (Chou and
Talalay, 1984).
A2.2.3 Cell Imaging Immunofluorescence: After 24 hours of drug exposure, PEO1 and PEO4 cells were fixed with
4% paraformaldehyde in PBS, permeabilized with PBS containing 0.4% Triton-100 and blocked
with 100% ADB. Fixed and permeablized cells were incubated with primary anti-γH2AX
161
antibody (Upstate Biotechnology) at 4C overnight, followed by a 1-hour incubation at room
temperature with the secondary antibody (Alexa Fluor 488 goat anti-mouse; Invitrogen).
Irradiation with 4Gy was used as the positive control. Images were obtained using Axio-vision
software and nuclear counts were quantified using Foci-Counter (http://focicounter.sourceforge.
net).
For Incucyte imaging, cells were seeded in 96-well plates at a density of 10,000 cells/well and
cultured overnight. Carboplatin, DSF and Carboplatin+DSF were serially diluted in growth
medium containing YOYO-1 (Life Technologies) at a final concentration of 0.1 µM in media, as
described above. Cells were placed in an IncuCyte™ FLR (Essen Bioscience) with a 10X
objective in a standard cell culture incubator at 37˚C and 5% CO2. Two images per well were
collected every 2-3 hours in both phase contrast and fluorescence-modes for 72 hours.
A2.2.4 PDX Studies
Patient case #2555 was thawed and prepared as described previously. 106 cells in 1:1
HBSS:growth factor-reduced Matrigel (BD Biosciences) were injected into the right mammary
fat pad NOD/SCID mice. When xenografts reached ~100mm3, mice were treated with either DSF
200mg/kg PO by oral gavage (Sigma, Israel) dissolved in DMSO/corn oil; Carboplatin 75mg/kg
IP; both Carboplatin 75mg/kg IP and DSF 200mg/kg PO (gavage) or vehicle control
(DMSO/corn oil PO and 300 cc saline IP). DSF was prepared as described: 5mg of DSF was
dissolved in 20µL DMSO, then mixed in 180µL corn oil (Brar et al., 2004). DSF was intended to
be administered by oral gavage every Monday, Wednesday, Friday; however, because it was felt
that the mice had significant weight loss at this dose, only the Monday and Wednesday doses
were administered. Carboplatin was administered every Monday. Mice were sacrificed when
tumours were greater 1500mm3 (within 2 weeks of initiation of treatment).
162
A2.3 Results
A2.3.1 Combination Treatment with DSF/Carboplatin shows Synergistic
Effects in PEO4 cells
To begin to assess the effect of DSF and Carboplatin on BRCA2 mutant ovarian cancer cells, I
treated two isogenic cancer cell lines: PEO1 (BRCA2null) and PEO4 (BRCA2wt) (Sakai et al.,
2009) with increasing doses of DSF and Carboplatin to determine their respective half-maximal
inhibitory concentration (IC50). Alamar Blue was used as an indicator of metabolic activity and
cellular health (Figure 5.1A, 5.1B). The Carboplatin and DSF IC50s for PEO1 were ~60 µM and
20 µM, respectively. The Carboplatin and DSF IC50s for PEO4 were ~120 µM and 25 µM,
respectively. I next asked whether combination treatment with DSF and Carboplatin would
demonstrate synergy. I expected the PEO1 cell line (BRCA2null) to be more sensitive to the
combined effects of Carboplatin and DSF. Surprisingly, the PEO4 cells, with a functional
BRCA2 protein, were found to be more sensitive in combination drug studies (Figure 5.2). This
effect was synergistic based on the combination index (CI) as determined by the Chou-Talalay
method (Figure 3) (Chou and Talalay, 1984). The CI for PEO1 cells was >1 (not synergistic),
whereas the CI for PEO4 cells was <1 (synergistic) at all effective dose (ED) levels (Table 5.2).
A)
B)
Figure 5.1. A) DSF and B) Carboplatin IC50 dose-response curves for PEO1 and PEO4 cell lines. PEO1 (light grey diamonds) and PEO4 cells (dark grey squares) were treated with indicated concentrations of a) DSF and b) Carboplatin, and effects on cell number were inferred by an Alamar Blue assay. Results represent the average of three experiments expressed as a mean percentage of untreated control cells. The Carboplatin and DSF IC50s for PEO1 are ~60 µM and 20 µM, respectively. The Carboplatin and DSF IC50s for PEO4 are ~120 µM and 25 µM, respectively.
163
A) PEO1 (BRCA2null)
B) PEO4 (BRCA2wt)
Figure 5.2. Effects of DSF and Carboplatin combinations on A) PEO1 and B) PEO4 cell lines. PEO1 and PEO4 cells were exposed to graded concentrations of Carboplatin or DSF, either alone or in combination at a ratio of 1:1 of for 72 h, followed by analysis of viability by an Alamar Blue assay. Each data point represents the mean ±�SD of at least three determinations. Carboplatin (black circles), disulfiram (diamond), Carboplatin + disulfiram (square, dotted line). A) PEO1 (BRCA2null)
B) PEO4 (BRCA2wt)
Figure 5.3. Isobolograms of DSF/Carboplatin combinations. Isobologram analysis of the combination of DSF/Carboplatin in A) PEO1 and B) PEO4 cells. The individual doses of Carboplatin and DSF to achieve 90% (blue line), 75% (green line) and 50% (red line) growth inhibition were plotted on the x- and y-axes. Combination index (CI) values calculated using Calcusyn software are represented by points above (indicate antagonism) or below the lines (indicate synergy). (X symbol) ED50, (plus sign) ED75 and (open dotted circle) ED90.
164
Table 5. 2. Combination Index (CI) values for PEO1 and PEO4 cell lines.
Next, I performed an analysis of γ-H2AX foci formation. PEO1 and PEO4 cells were incubated
in 61µM of Carboplatin and 25µM of DSF for 24 hours, after which γ-H2AX immunofluorescent
staining was performed (Figures 5.4 & 5.5) and quantified (Figure 5.6). Again, DSF in
combination with Carboplatin exerts a synergistic effect on PEO4 cells and an additive effect in
PEO1 cells, as seen by γ-H2AX foci formation. Treatment with DSF or Carboplatin produced an
average of 1.5 and 4.8 γ-H2AX foci, respectively (Figure 5.6). In combination drug treatment, an
average number of 8.38 γ-H2AX foci were seen, representing synergistic effects.
Furthermore, I used YoYo-1 staining to quantify cell death. YoYo-1 is a cell impermeant
cyanine dimer nucleic acid stain that binds to dsDNA. When added to the culture medium,
YoYo-1 fluorescently stains the nuclear DNA of cells that have lost plasma membrane integrity.
With Incucyte technology, I visualized and quantified cell death as a result of combination drug
treatment (Figures 5.8 and 5.9). There was a greater amount of cell death seen in the PEO4 cell
line with combination drug treatment (object counts >8000) than would be expected based on
either drug alone (object count = 2000 for both DSF and Carboplatin at 72hr). Conversely,
although the PEO1 cell line had a similar total amount of cell death (object count = 8000), the
effects of Carboplatin and DSF at 72hr were 3500, and 3800 respectively and were not
synergistic.
ED CI PEO1 CI PEO4
50 2.2 0.27
75 1.5 0.40
90 1.1 0.60
165
A)
B)
C)
D)
E)
Figure 5.4. Formation of γ-H2AX foci in drug treated PEO1 (BRCA2null) cells. Each panel shows the DAPI-stained nucleus (blue) and anti-γ-H2AX antibody (green). Panels: A) negative control; B) positive control (4Gy); C) DSF 25µM; D) Carboplatin 61µM; E) combination of DSF and Carboplatin.
166
A)
B)
C)
D)
E)
Figure 5.5. Formation of γ-H2AX foci in drug treated PEO4 (BRCA2wt) cells. Each panel shows the DAPI-stained nucleus (blue) and anti-y-H2AX antibody (green). Panels: A) negative control; B) positive control (4Gy); C) Disulfiram 25µM; D) Carboplatin 61µM; E) combination of DSF and Carboplatin.
167
Figure 5.6. Quantification of γ-H2AX foci in drug-treated PEO1 and PEO4 cells. . Data shown represent the average number of foci formed ± s.e.
168
t = 2 hr t = 72 hr
Disulfiram
Carboplatin
Combination
Figure 5.7. Effect of DSF and Carboplatin treatment on PEO1 cell proliferation. YoYo-1 fluorescent nuclear counts in PEO1 cells at t= 2 hours and termination of experiment (t=72 hours). DSF 25µM, Carboplatin 61.25µM and drug combination.
169
t = 2 hr t = 72 hr
Disulfiram
Carboplatin
Combination
Figure 5.8. Effect of DSF and Carboplatin treatment on PEO4 cell proliferation. YoYo-1 fluorescent nuclear counts in PEO1 cells at t= 2 hours and termination of experiment (t=72 hours). DSF 25µM, Carboplatin 61.25µM and drug combination.
170
A)
B)
Figure 5.9. Quantification of YoYo-1 fluorescence in A) PEO1 and B) PEO4 cell lines. Negative control (yellow square with cross); DSF (light grey diamond), Carboplatin (dark grey circle), combination treatment (black square).
171
A2.3.2 Many Ovarian Cancer Cell Lines show Synergistic Effects of
DSF/Carboplatin in vitro
Given the above results, I wanted to explore whether similar synergy between DSF and
Carboplatin could be observed in additional ovarian cancer cell lines. To do so, I first determined
the IC50 values for DSF and Carboplatin for the following cell lines: OV 1369, TOV 1369, OV
1946, PEO23, TOV2223G, A2780, OV 90, and SKOV3 (data not shown). Next, I performed
combination studies to assess whether the combined effects of DSF and Carboplatin (Figure
5.10). I excluded SKOV3 and A2780 as a recent publication identified these cell lines as
‘unlikely to be high-grade serous’ (Domcke et al., 2013). Of the six additional cell lines tested,
synergy between DSF and Carboplatin was observed in 5 cell lines (Figure 5.11): OV 1369,
TOV 1369, TOV 2223G, PEO23 and OV 1946.
172
Figure 5.10: Effects of Carboplatin and DSF on human ovarian cancer cell lines. OV 1369, TOV 1369, OV 1946, PEO23, TOV2223G, and OV90 cells were exposed to graded concentrations of Carboplatin or DSF, alone or in combination, at a ratio of 1:1 of for 72 hr. Viability was analyzed by Alamar Blue assay. Each data point represents the mean ±�SD of at least three determinations. Carboplatin (black circles), disulfiram (diamond), Carboplatin + disulfiram (square, dotted line).
OV 1369
TOV 1369
OV 1946
TOV 2223G
PEO23
OV 90
173
OV 1369
TOV 1369
OV 1946
TOV 2223G
PEO23
OV90
Figure 5.11. Isobolograms of drug combinations. Isobologram analysis of the combination of Carboplatin and DSF in OV 1369, TOV 1369, OV 1946, TOV 2223G, PEO23, and OV 90 cells. The individual doses of Carboplatin and DSF to achieve 90% (blue line), 75% (green line) and 50% (red line) growth inhibition are plotted on the x- and y-axes. Combination index (CI) values, calculated using Calcusyn software, are represented by points above (indicate antagonism) or below the lines (indicate synergy). (X symbol) ED50, (plus sign) ED75 and (open dotted circle) ED90.
174
Table 5.3. Combination Index (CI) values for all cell lines evaluated.
ED PEO1 PEO4 OV1369 TOV1369 TOV2223G PEO23 OV1946 OV90
50 2.2 0.3 0.4 0.4 0.1 0.2 0.2 4.2
75 1.5 0.4 0.2 0.5 0.1 0.3 0.2 3.1
90 1.1 0.6 0.2 0.6 0.1 0.5 0.1 2.4
A2.3.3 DSF/Carboplatin Combination Slows Tumour Growth
The doses of DSF required to inhibit ovarian cancer growth in short term assays in vitro were
higher than those predicted to be achievable in vivo. However, with 6/8 evaluated ovarian cancer
cell lines displayed synergy in combination drug treatments of DSF and Carboplatin, I proceeded
to evaluate the effects of these drugs in vivo. I decided to use PDXs, as opposed to cell line-
derived xenografts for reasons previously discussed (Section 1.14.2). The clinical characteristics
of patient 2555 are listed in Table 2.3.
When tumours reached ~200mm3 (Figure 5.12.A), DSF/Carboplatin treatment was initiated at
doses similar to those used to treat adult humans. Mice treated with both drugs demonstrated
slower growth, compared with control mice (Figure 5.12C). The mean tumour volume of
DSF/Carboplatin-treated group was lower than the control group on day 9: 218 mm3 versus 998
mm3 (unpaired t-test p = 0.04) (Figure 5.12B).
Although mice treated with DSF+Carboplatin had the lowest tumour volume, they also had the
lowest percent survival (Figure 5.12.D; 5/9 mice died within 24 hours of drug administration. I
believe this result to be the product of two main factors: a non-optimized disulfiram dose and/or
poor oral gavage drug delivery. I did not perform an MTD analysis during this study given that
an oral dose had previously been published for another strain of mice. In retrospect, the oral
dose for NSG mice should have been optimized, particularly in the context of drug combinations.
175
A)
B)
C)
176
D)
Figure 5.12 DSF/Carboplatin inhibit tumour growth in vivo. Patient 2555 NOD/SCID PDXs were treated with DSF (200 mg/kg PO, q48 hr x 2 doses), Carboplatin (75 mg/kg IP x 1 dose), vehicle (10% DMSO in corn oil PO, saline IP) or combination of drugs (DSF 200mg/kg PO + Carboplatin 75 mg/kg IP). A) Pre-treatment tumour volume. B) Post-treatment tumour volume. P values (two tailed) were calculated using students t-test. C) Average tumour volume versus time for each treatment group ± s.e. DSF/Carboplatin group showed slower tumour growth and smaller tumour volume at endpoint. D) Kaplan-Meier survival curve of NOD/SCID mice from control (black line; n=5), Carboplatin IP (red line, n=5), DSF PO (green dashed line, n=5) or combination drug (grey dashed line, n=9).
177
A2.4 Discussion
Approximately 30-50% of ovarian HGSC have defective homologous recombination. I
undertook this study to examine whether cells lacking a functional BRCA2 would be more
sensitive to the combined cytotoxic effects of Carboplatin and DSF drug treatment. Using two
isogenic cell lines, one with a functional BRCA2 protein (PEO4) and one without (PEO1) (Sakai
et al., 2009), I evaluated the effects of combining DSF with Carboplatin. In contrast to
expectations, a synergistic response was observed in the PEO4 cells, which harbor a functional
homologous repair pathway. This effect was seen by increased γ-H2AX foci formation, as well
as increased nuclear fluorescence using a marker of cell death. This prompted us to explore the
effects of Carboplatin with DSF in 8 additional cell lines (2 were ultimately excluded given
recent evidence suggesting they are not ovarian high-grade serous). In total, 8 cell lines were
tested and 6 showed in vitro synergy (as indicated by the CI) between Carboplatin and DSF.
Concentrations of DSF in man have been reported to range between 0.5 µg/mL to 10 µg/mL,
depending on dosing schedules, body weight and renal function. This corresponds to DSF
plasma concentrations ranging from 1 – 35 µM and 10- 600nM for its metabolites (Faiman et al.,
1978; Johansson, 1992). This makes concentrations needed to achieve cytotoxicity in vitro nearly
achievable in vivo.
Based on this data, I assessed the effects observed in vitro in an in vivo PDX model, using a case
known to be highly aggressive. A 200mg/kg/d oral dose of DSF has been previously published in
NOD/SCID mice and corresponds to what would be expected based on a human-equivalent drug
dose translation (Reagan-Shaw et al., 2008) However, although this dose seemed appropriate,
this study did not use DSF in combination with other drugs (Brar et al., 2004). When used in
combination with Carboplatin, this DSF dose was too toxic and resulted in the death of >50% of
mice in the combined-treatment group. An MTD study using DSF/Carboplatin needs to be
performed to elucidate the optimal DSF dose.
A wide range of anti-cancer effects have been described in a multitude of cancers (Table 5.1) and
it was outside the scope of this project to elucidate the mechanism(s) of DSF’s activity against
178
ovarian cancer cells. Future projects should focus on assessing how DSF is able to potentiate the
effects of Carboplatin, both in vitro and in vivo.
Numerous studies further support a role for the use of DSF in the treatment of a number of
cancers. For example, phase II clinical trials are underway examining the effects of DSF for the
treatment of prostate and lung cancers (ClinicalTrials.gov identifiers: NCT01118741 and
NCT00312810). Our results suggest that targeting ovarian cancer cells with DSF, in addition to
standard chemotherapeutics, may be an effective novel method to treat ovarian cancer. Our
findings suggest that clinical trials incorporating DSF and platinum therapy should be pursued in
ovarian cancer.