Background
Ovarian cancer is the most common cause of death from gynecologic malignancies and the fifth most common cause of cancer death in women in the United States [
1]. Ovarian adenocarcinomas account for 85-90% of all cancers of the ovary. The initiating cell population for EOC remains to be exactly defined, with different evidence suggesting tumors originate from the ovarian surface epithelium (OSE), inclusion cysts lined by OSE [
2‐
5] or alternatively, the fallopian tube epithelium [
6] or components of the secondary Müllerian system, including the epithelial cells of the
rete ovarii, paraovarian/paratubal cysts, endosalpingiosis, endometriosis or endomucinosis [
7]. The lack of clarity regarding tumor origin stems from the fact that unlike epithelial cancers arising in other organs, a well-defined disease spectrum consisting of benign, invasive and metastatic lesions has not been identified for EOC. This is due at least in part to that fact that the majority of cases are identified at advanced stage when disease has spread beyond the ovary. Another reason is the morphologic complexity of common EOCs which consist of several distinct histologic subtypes; these include serous, endometrioid, mucinous and clear cell cancers.
Progress in ovarian cancer research has been slowed by the lack of suitable animal models that exhibit features of human disease. Genetically manipulable mammalian models of spontaneous ovarian cancer are rare, particularly those representing ovarian adenocarcinomas. Human and rodent models of spontaneous
ex vivo transformation of OSE have been described [
8‐
10]. One of these models, a syngeneic mouse model of EOC [
10], has been extensively used for preclinical studies of therapeutic agents and studies of the tumor microenvironment [
11‐
18]. Early attempts to produce murine EOC models using transgenic or other genetic engineering approaches resulted in the development of granulosa cell tumors [
19‐
24]. More recently, a number of laboratories have developed genetically engineered mouse (GEM) models of EOC by using
ex vivo transformation [
25,
26], transgenic [
27,
28] and conditional gene expression strategies [
29‐
31]. To date, due to the lack of a suitable GEM model expressing Cre-recombinase, the strategy most frequently employed for conditional gene expression in the ovarian epithelium involves survival surgery for intrabursal injection of recombinant Adenovirus-Cre [
29‐
34].
Recently, our group developed a spontaneous transgenic mouse model of EOC by expressing the oncogenic early region of SV40 under the transcriptional control of the Müllerian inhibiting substance type II receptor gene promoter [
27,
28]. Although SV40 TAg expression is not directly associated with the development of human cancer, its expression results in functional inactivation of the critical tumor suppressors p53 and Rb. Mutation of
TP53 is, by far, the most common genetic alteration observed in EOC, particularly the serous subtype [
35,
36]. Direct mutation or loss of
Rb or its downstream signaling mediators are also common in EOCs [
37‐
41]. Via binding and inhibition of PP2A, SV40 tag also results in activation of PI3K/AKT and mitogen activated protein kinase (MAPK) signaling [
42], pathways frequently activated in human EOC [
43]. A stable transgenic line of Tg
MISIIR-TAg mice was established in which female mice develop bilateral ovarian carcinoma with 100% penetrance [
28]. To date, this is the only GEM model that develops spontaneous EOC with pathological features of serous EOC that does not require extensive surgical manipulation to induce the phenotype. Like human EOC, female Tg
MISIIR-TAg mice with significant tumor burden exhibit no apparent symptoms of illness and disease dissemination is typically restricted to the peritoneum [
27,
28]. Murine ovarian carcinoma (MOVCAR) cell lines isolated from the ascites and primary tumors of these mice share many molecular features with human tumors [
27,
28,
44‐
48] and are well suited to experimental analysis
in vitro. With these reagents, the expression levels of specific genes can be experimentally manipulated and properties of MOVCAR cell lines can be assessed
in vitro. However, the lack of a syngeneic recipient for manipulated MOVCAR cells has limited the analysis of the
in vivo effects of genetic alterations in the model to studies in immunodeficient mice. The present study describes the identification of non-tumor prone lines of Tg
MISIIR-TAg transgenic mice that can be used as syngeneic recipients for MOVCAR cell allografts. The availability of this syngeneic model affords the opportunity to study the
in vivo effects of genetic alterations on tumor properties and on interactions between tumor cells and their microenvironment in an immunocompetent host. Moreover, this immunocompetent mouse model of EOC is suitable for studies of immune-based therapeutic strategies and vaccine development.
Methods
Transgenic mice and backcrosses
All procedures involving mice were approved by the Fox Chase Cancer Center (FCCC) Institutional Animal Care and Use Committee (IACUC) and all mice were maintained under specific pathogen free conditions. Individual transgenic Tg
MISIIR-TAg founder mice were generated in the FCCC Transgenic Facility in a first generation hybrid genetic background of C57BL/6 and C3H (B6C3F1) and genotyped by PCR amplification as previously described [
27]. Transgenic founders were crossed with wild type C57BL/6 mice (obtained from the FCCC Laboratory Animal Facility) to establish breeding lines. Relevant lines of EOC-prone and non-tumor-prone Tg
MISIIR-TAg mice were maintained as hemizygotes and backcrossed for a minimum of ten generations to wild type C57BL/6 mice to generate genetically pure lines of C57BL/6 Tg
MISIIR-TAg mice.
Cell lines and culture conditions
Pure C57BL/6 MOVCAR cell lines, including MOVCAR 12, 5009 [
49], 5025, 5183, 5438, 5447 and 5612, were established from bulk ascites isolated from individual ovarian tumor-bearing C57BL/6 Tg
MISIIR-TAg mice as previously described [
27]. Tumorigenic spontaneously transformed murine ovarian surface epithelial cell (MOSEC) lines ID-8, IF-5 and IG-10 were a gift from Dr. Katherine Roby, University of Kansas Medical Center, and ID-8 cells stably overexpressing murine VEGF164 were a gift from Dr. George Coukos, University of Pennsylvania. All MOVCAR and MOSEC cells were maintained in DMEM supplemented with 4% FBS, 1× Insulin/Transferrin/Selenium-A (ITS, supplied as 100× stock from Gibco/Invitrogen), penicillin/streptomycin (100 units/mL and 100 μg/mL, respectively) and 2 mM
l-glutamine and incubated at 37°C in 5% CO
2. Culture medium was changed once weekly and cells were trypsinized and passaged at 4-5 day intervals when they reached confluence. MOVCAR cells were prepared for
in vivo injection as described [
49]. For
in vivo imaging, cells were transduced with a retroviral construct encoding the firefly luciferase gene (pWZL-Luc, generously provided by Dr. Maureen Murphy, FCCC) using standard methods.
Immunoblot and immunoprecipitation
To prepare lysates for immunoblot analysis, cells were washed with cold PBS, lysed with M-PER m ammalian p rotein e xtraction r eagent (Thermo Scientific, Rockford, IL) supplemented with a cocktail of protease inhibitors (Complete Mini, Roche, Indianapolis, IN) and protein concentration was determined by BCA method (Thermo Scientific, Rockford, IL). Equal amounts of protein samples were resolved by SDS-PAGE gel electrophoresis on 12% acrylamide gels and transferred to polyvinylidene difluoride membrane (Immobilon, Millipore Corp., Bedford, MA). Membranes were blocked in 5% milk and 0.1% Tween-20 in 1× PBS for 1 h prior to incubation with primary antibodies recognizing SV40 TAg (Pab 101) and mouse p53 (Pab 240) obtained from Santa Cruz Biotechnology, Inc. at 1:1000 dilution. Horseradish peroxidase-conjugated secondary antibodies were used according to manufacturer's protocols. Immunoreactivity was visualized using the ECL system and was exposed to BioMax MR film (Eastman Kodak Co.).
For immunoprecipitation, cells were grown in 100-mm plates and lysed in 1 ml M-PER mammalian protein extraction reagent. The whole cell lysates were incubated with SV40 TAg antibody (Pab101) at a dilution of 1:100 at 4°C overnight with constant mixing. Protein A beads (40 μl) were added and mixed for 3 h at 4°C. Immunoprecipitates were then washed 5 times with M-PER mammalian protein extraction reagent and pellets resuspended in Laemli buffer for protein electrophoresis and immunoblot blot analysis performed as described above with antibodies against TAg and p53.
Cell cycle analysis
Cells were prepared for cell cycle analysis using the fluorescent nuclear stain propidium iodide and fluorescent sorting was carried out using the Guava Personal Cell Analysis machine exactly as described by the manufacturer (Guava Technologies).
RNA preparation, quantitative reverse transcription PCR
Total RNA was isolated from MOVCAR cells using the RNA Easy Mini Kit (Qiagen). With the assistance of the FCCC Genomics Facility, levels of Mdm2 mRNA expression were evaluated by real-time quantitative reverse transcription PCR (qRT-PCR) using Taqman technology with probe sets for Mdm2 and Hprt1 obtained from Applied Biosystems, Carlsbad, CA.
Quantitation of secreted VEGF by ELISA
Cells (5 × 105) were plated in triplicate in 6-well dishes and grown in complete medium for 72 hours. The conditioned culture medium was removed and the level of secreted VEGF present in the medium was determined by ELISA using the Mouse VEGF Quantikine Elisa Kit (R&D systems, Minneapolis, MN). After removal of the conditioned culture supernatant, cells were immediately rinsed with PBS, trypsinized and the number of cells present in each well was counted. Secreted VEGF levels were normalized to the total number of cells present in the sample to determine the amount of VEGF/104 cells. Three independent assays were performed and the amount of secreted VEGF/104 cells expressed as the mean value for each cell line tested.
Oophorectomy and MOVCAR cell allografts
Four to six week-old ovarian tumor-prone Tg
MISIIR-TAg mice were anesthetized by i.p. injection of 95 μl per 10 gram body weight of 10 mg/mL Ketamine hydrochloride and 1 mg/mL Xylazine hydrochloride in sterile saline and subjected to oophorectomy using a standard asceptic surgical procedure commonly used for transgenic embryo injection to expose the ovarian fat pad and ovary (described in detail in [
49]). Once exposed, a small incision was made in the ovarian bursa that enabled removal of the resident ovary and/or fallopian tube. The ovarian bursa was sealed with surgical glue and the reproductive tract returned through the incision in the body wall. The surgical incision was closed with wound clips. The same surgical procedure was used for orthotopic (i.b.) injection of MOVCAR cells into recipient mice. Methods for i.b. and i.p. (pseudo-orthotopic) injections of MOVCAR cells were previously described in detail [
49].
Preparation and analysis of tissues, histology and immunohistochemistry
All mice were euthanized by CO2 asphyxiation, necropsied and examined for gross abnormalities. Pathologically altered organs, entire reproductive tracts and representative specimens of multiple organs and tissues, including the brain, lung, liver, kidney, spleen, pancreas and intestine were removed at necropsy, fixed in 10% (v/v) neutral buffered formalin (NBF) overnight, transferred to 70% ethanol and paraffin-embedded. In mice with evident tumor, specimens of the tumor tissue were also excised, snap frozen in liquid N2 and stored at -80°C. For histological analysis, 5 μm formalin fixed paraffin embedded tissue sections were cut for either H&E staining or immunohistochemistry (IHC). Histopathological analysis was performed by a pathologist with expertise in human and murine malignancies (AKS).
Sections of tumor tissue for IHC staining were cut on SuperFrost Plus charged slides (Fisher). Unstained sections were deparaffinized, subjected to antigen retrieval and stained with antibody against SV40 TAg (Pab 101, 1:100) as described [
27].
Bioluminescent imaging (BLI)
For detection of
in vivo growth of pWZL-Luc transduced MOVCAR tumor cells, mice were anesthetized with 2% isofluorane and given i.p. injections of 100 mg/kg luciferin substrate (Caliper Life Sciences) ten minutes prior to imaging using the IVIS Spectrum
in vivo imaging system (Caliper Life Sciences) as described [
49]. Image analysis was performed and total flux emission (photons/second) in the region of interest (ROI) was determined using the Living Image Software for the IVIS Spectrum.
Discussion
Utilization of animal models with an intact immune system is critical for the evaluation of immune-based therapeutic strategies and vaccine development. An SV40 TAg transgenic model of prostate cancer [
55] has been used to study the effects of combining blockade of cytotoxic T lymphocyte antigen 4 (CTLA-4) and vaccination with granulocyte macrophage colony stimulating factor (GM-CSF;Gvax) and subsequent derivatives of this vaccine strategy [
56‐
60]. The C57BL/6 syngeneic mouse ovarian cancer model developed by Roby et al, [
10] has been used for studies of the contribution of cells in the tumor microenvironment, including epithelial-stromal cell interactions, VEGF induced-effects on tumor vasculature and tumor cell-secreted factors that stimulate cytokine production, macrophage infiltration and vascularization that favor tumor growth and progression [
14,
15,
18]. Similar studies would be difficult to impossible to conduct in immunodeficient mice. The availability of an additional syngeneic mouse model of EOC will allow cross-comparison of mouse models and validation of key findings.
The functional utility of animal models of human cancer depends largely on the extent to which the animal model recapitulates the histology and biological behavior of the disease in humans. Many transgenic tumor models have been developed using the immediate early region of the SV40 virus containing the potently oncogenic large and small T antigen (
TAg and
tag) genes [
55,
61‐
66]. The continued utility of SV40 TAg models in studying cancer is underscored by seminal contributions to our understanding of the "angiogenic switch" [
67‐
71] and tumor progression and invasion [
72]. Importantly, a recent study [
73] identified an integrated gene expression signature from three distinct TAg mouse models (i.e., mammary, prostate and lung cancer models) that is comparable to a signature associated with the aggressive biological behavior and prognosis for several human epithelial tumors, including breast cancers. Results from this study showed that tumors arising in TAg-based mouse models share common features of gene expression with human cancer and are relevant preclinical models [
73].
Female transgenic C57BL/6 Tg
MISIIR-TAg-DR26 mice develop spontaneous bilateral ovarian carcinoma with 100% penetrance [
28]. Tumor progression in these mice is characterized by widespread peritoneal dissemination and the development of malignant ascites and tumor morphology and histology of the tumors closely resembles high-grade serous adenocarcinomas, the most common histologic subtype of EOC detected in women. Tumors and cell lines derived from primary tumors and ascites of tumor bearing mice exhibit several characteristics in common with human EOC cell lines and tumors including AKT/mTOR activation, COX1 overexpression and VEGF overexpression and secretion ([
28,
44‐
47] and the present study). In addition, a verapimil-sensitive Hoescht dye-excluding ovarian carcinoma side population (SP), a potential population of ovarian cancer initiating cells, was identified in MOVCAR cell lines [
48]. Ovarian tumors arising in C57BL/6 Tg
MISIIR-TAg-DR26 mice are sensitive to standard combination platinum and taxane chemotherapy and to mTOR inhibition with Everolimus (RAD001) [
28,
45]. These observations underscore the potential utility of these transgenic mice for preclinical evaluation of therapeutic agents. However, reflecting its relation to the biology of human EOC, tumor formation in this transgenic model is also stochastic, resulting in variation in the latency of tumor formation and time to metastasis. This necessitates relatively large cohorts of mice and non-invasive longitudinal
in vivo imaging such as MRI to optimize results of therapeutics studies.
To overcome the limitations encountered with spontaneous tumor development, we isolated individual transgenic lines of non-tumor prone C57BL/6 Tg
MISIIR-TAg transgenic mice that can serve as syngeneic immunocompetent hosts for allografted TAg expressing MOVCAR cells isolated from tumor bearing C57BL/6 Tg
MISIIR-TAg-DR26 mice. Syngeneic mouse models of EOC in which spontaneously transformed ID-8 MOSEC grown as allografts in C57BL/6 recipients [
10] or HM-1 tumor cells grown as allografts in B6C3F1 recipients [
74] have been previously described. These syngeneic models have been used successfully for preclinical evaluation of therapeutic agents and studies of the role of the tumor microenvironment on ovarian tumor growth and progression [
11‐
18,
75]; however, these models each rely on single mouse ovarian carcinoma cell lines in which the underlying molecular mechanisms of malignant transformation remain undefined.
The ease of establishment of TAg-transformed MOVCAR cell lines in culture has enabled the isolation of a large number of distinct cell lines, several of which are described in the present study. Although derived from an inbred strain of mice, the stochastic manner in which tumors arise in C57BL/6 TgMISIIR-TAg-DR26 mice results in intrinsic differences in MOVCAR cell lines derived from individual tumor-bearing mice. MOVCAR cell lines grown in culture exhibit different growth rates and expression of proteins associated with EOC, such as levels of secreted VEGF. These cell lines also exhibit differences when grown in vivo. For example, some cells lead to very rapid growth and production of voluminous malignant ascites, whereas other cells are slower growing and produce less ascites. Interestingly, the cell lines that result in the highest levels of ascites production in vivo are the cell lines that exhibit the highest levels of VEGF secretion in vitro. These observations suggest that although the primary oncogenic stimulus driving tumorigenesis in C57BL/6 TgMISIIR-TAg-DR26 transgenic mice is the same in all animals, there are likely additional genetic, epigenetic and/or gene expression alterations that contribute to ovarian tumor progression, and identification of these alterations may contribute to our understanding of human EOC. Moreover, once identified, the role of specific alterations in gene function in ovarian tumorigenesis can be studied in these cell lines as they are readily amenable to direct manipulation using established strategies for ectopic gene expression or RNA interference.
With regard to preclinical evaluation of novel therapeutic agents, our syngeneic mouse model of EOC provides several advantages. First, tumors are grown in fully immunocompetent mice enabling the evaluation of vaccine and immune-based therapeutic strategies. Second, TgMISIIR-TAg-Low transgenic mice have been fully backcrossed to a pure C57BL/6 genetic background, exhibit normal fertility and lifespan and do not develop tumors. Thus, large cohorts of mice can be established for synchronous allograft initiation without interference of tumor growth initiated from the host. Third, the availability of multiple distinct MOVCAR cells lines for evaluation avoids issues of cell line-specific effects, and because MOVCAR cells are easily manipulated in culture, on-target effects of therapeutics can be confirmed in parallel using RNAi based strategies for direct target knockdown. Finally, the ability to easily express reporter genes in MOVCAR cells facilitates strategies for non-invasive in vivo optical imaging such as bioluminescent, fluorescent and near infrared fluorescent imaging.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BAQ, FX, LB and LM conducted the studies and participated in the data analysis. XH performed oophorectomies, ovarian transplants and orthotopic implantation of tumor cells and AKS conducted the histopathological evaluation of tumor tissues. DCC conceived and designed experiments, analyzed the data and wrote the manuscript. All authors have read and approved the final manuscript.