Background
Ovarian cancer is the fifth most commonly diagnosed cancer in women and the fourth most common cause of death from cancer [
1]. The high mortality can be attributed to the high percentage of affected women presenting at an advanced stage, with spread within the peritoneal cavity [
2,
3]. With current therapies, including surgical debulking and platinum-based chemotherapy, patients in stage III or stage IV only have a 20% chance of long-term survival [
2,
3]. Better understanding ovarian carcinoma biology, as well as the development of new therapies for the disease, has been hampered by the lack of suitable animal models.
Current ovarian cancer models fall into three broad categories: rare spontaneous carcinomas, induced tumors, and human xenografts [
4]. Although these models have allowed researchers to gain valuable insights into the biology of ovarian cancer, each model exhibits important limitations [
4,
5]. Spontaneous ovarian cancer has been observed in mice, rats, and hens [
6‐
8]. The drawback to these models is that the cancers tend to occur at an advanced age and at similar low frequencies as in humans. The low incidence and the length of time required for the development of these tumors render them of limited use for studying the biology and treatment of ovarian carcinoma. Induced tumor models circumvent these problems but create their own artificial systems, which may not accurately reflect the human disease. In one model of
in vitro transformation, ovarian surface epithelium cells are subcloned until they exhibit the loss of contact inhibition, the capacity for substrate-independent growth, cytogenetic abnormalities, and the ability to form tumors when injected subcutaneously and/or intraperitoneally into athymic mice [
9]. This model, though, fails to account for critical interactions between the cancer cells and the host. Also, it is uncertain if these cells or their malignant transformation are representative of normal human cells or clinical disease.
Animal models have been generated by expressing simian virus 40 large T antigen [
10], by inactivating p53 and Rb1 [
11], by inactivating p53 and activating an oncogene [
12], and through hormone treatment [
13‐
15]. The high rate of cancer development in these animals makes these models attractive, but they may not reliably represent human cancer because a majority of these genetic changes usually do not occur in patients. Xenografts of human cancers have undergone continuous improvement over the past twenty years [
16‐
19]. These models allow for direct examination of the human cancer but do not allow the study of the early stages of the cancer. These models also rely on an immune-deficient host, which eliminates the interaction between the cancer and the immune system.
We present a new model of ovarian carcinoma, designated FNAR, that spontaneously developed in an untreated, previously normal Lewis rat. The tumor could be serially passaged both in vivo as malignant ascites in rats and in vitro. Importantly, the biologic characteristics of the tumor closely paralleled one type of human ovarian carcinoma.
Methods
Animals
Female Lewis strain rats aged 4-6 weeks (purchased from Charles River Breeding Laboratories, Inc., Wilmington, MA) were kept in sterile micro-isolator cages and fed food and water ad libitum. The institutional guidelines of Johns Hopkins University concerning the care and use of research animals were followed. The animals were challenged intraperitoneally with graded numbers of FNAR cells and monitored daily for abdominal swelling. At various intervals after tumor challenge or when animals appeared moribund (pallor, lethargy, and marked abdominal distension), the animals were sacrificed by CO2 asphyxiation and the cells within the peritoneal cavity harvested by flushing the abdomen with 35 milliliters of sterile phosphate buffered saline (PBS, Grand Island Biological Co., Gibco BRL, Grand Island, NY). At sacrifice, the animals were examined for tumor growth and tissues taken for histological examination.
In vitro propagation and growth curve
A cell line (FNAR) that grows in vitro as an adherent monolayer was established by culture in RPMI 1640 (Gibco) supplemented with 10% fetal calf serum in 30 ml tissue culture flasks (Corning Flask 3056, Corning Inc., Corning NY). Cells used for experiments were low passage and maintained in culture for one to three months. The doubling time of the cell line was measured by plating 104 cells into macrotiter wells then harvesting and counting at 19.5, 43.5, and 115.5 hours.
Flow Cytometric Analysis
Flow cytometry was utilized to assess in vitro FNAR cells for expression of known phenotypic markers. Briefly, 5 × 105 tumor cells were incubated in polystyrene tubes. Analysis of the intracellular antigens estrogen receptor α, progesterone receptor, and androgen receptor first required fixation in 2% formaldehyde (Polysciences, Warrington, PA) in phosphate buffered saline (PBS, Gibco Invitrogen, Carlsbad, CA) for 15 minutes at 4°C followed by permeabilization with 0.1% Triton-X-100 (Sigma-Aldrich, St. Louis, MO) in PBS for 15 minutes at 4°C. The cells were then incubated for 30 minutes at 4°C with commercially purchased murine monoclonal antibodies. The concentrations of antibodies used are as follows: estrogen receptor (ER) α at 8 μg/106 cells (Abcam, Cambridge, MA), progesterone receptor (PR) at 16 μg/106 cells (Affinity Bioreagents, Golden, CO), or androgen receptor (AR) at 2 μg/106 cells (Pharmingen, San Diego, CA). The cells were washed and counterstained with phycoerythrin (PE) rat anti-mouse IgG1 (Becton Dickinson, San Jose, CA) at 125 ng/106 cells for 30 minutes at 4°C. Commercially purchased murine monoclonal antibody to the rat c-neu oncogene product (Calbiochem, San Diego, CA) was used at 1 μg/106 cells and was counterstained with PE rat anti-mouse IgG2a+b (Becton Dickinson, San Jose, CA) at 30 ng/106 cells for 30 minutes at 4°C. Tumor cells incubated with secondary antibody alone served as a negative control. Epithelial cell adhesion molecule (EPCAM) expression was analyzed using a PE-conjugated antibody (Santa Cruz, Santa Cruz, CA) at 1 μg/106 cells with mouse IgG1-PE as a negative control (Becton Dickinson, San Jose, CA). A commercially available rabbit polyclonal antibody to CA125 (Abbiotec, San Diego, CA) was used at 2 μg/106 cells and counterstained with 1 μg/106 cells APC goat anti-rabbit IgG (Invitrogen Molecular Probes, Carlsbad, CA). The cells were analyzed on a Becton-Dickinson FACSCalibur flow cytometer and data was analyzed using FlowJo (Tree Star, Inc, Ashland, OR).
Immunocytochemistry
FNAR cells were plated onto four-well CultureSlides (BD Falcon, San Jose, CA). Cells were fixed in 2% formaldehyde in PBS for 20 minutes followed by permeabilization in 0.5% Triton X-100 in PBS for 10 minutes. Cells were then incubated with a mouse monoclonal antibody to beta-catenin conjugated to Cy3 (Abcam, Cambridge, MA) at 6 μg/ml for one hour and counterstained with 500 ng/ml DAPI for five minutes (Invitrogen Molecular Probes, Carlsbad, CA). Images were captured using the Nikon Eclipse E800 (Tokyo, Japan) at 200× magnification with standard filters for DAPI and Cy3, the DS-QiMc digital camera (Nikon, Tokyo, Japan), and the Advanced Research Elements AR 3.0 software (Nikon, Tokyo, Japan).
Gene Expression Analysis by cDNA Microarrays
RNA was extracted and purified from cell lysates of 1-5 × 105 in vitro FNAR tumor cells and the REH cell line of normal rat endothelial cells, as a control, with 500 μl Trizol reagent (Invitrogen, Carlsbad, CA). Tissue samples were frozen in liquid nitrogen and pulverized with a mortar and pestle. The powder was dissolved in Trizol and centrifuged. Purified RNA was dissolved in 20 μl diethyl-pyrocarbonate-treated distilled water. The resulting RNA was analyzed at the Johns Hopkins microarray core. RNA from control and experimental samples was processed using the RNA amplification protocol described by Affymetrix (Affymetrix Expression Manual). Briefly, 5 μg of total RNA was used to synthesize first strand cDNA using the SuperScript Choice System (Invitrogen, Carlsbad, California) and oligonucleotide primers with 24 oligo-dT plus the T7 promoter (Proligo LLC, Boulder, Colorado). Following the double stranded cDNA synthesis, the product was purified by phenol-chloroform extraction and biotinilated anti-sense cRNA was generated through in vitro transcription using the BioArray RNA High Yield Transcript Labeling kit (ENZO Life Sciences Inc., Farmingdale, New York). Fifteen μg of the biotinilated cRNA was fragmented at 94°C for 35 minutes in buffer (100 mM Tris-acetate, pH 8.2, 500 mM potassium acetate, and 150 mM magnesium acetate), and 10 μg of total fragmented cRNA was hybridized to the Affymetrix GeneChip rat 230 2.0 array (Santa Clara, CA) for 16 hours at 45°C with constant rotation (60 rpm). Affymetrix Fluidics Station 450 was then used to wash and stain the chips with a streptavidin-phycoerythrin conjugate. The staining was then amplified as follows: blocking was performed using goat IgG, then a biotinilated anti-streptavidin antibody (goat) was bound to the initial staining, and amplification was completed by the addition of a streptavidin-phycoerythrin conjugate. Fluorescence was detected using the Affymetrix 3000 7G GeneArray Scanner and image analysis of each GeneChip was done through the GeneChip Operating System 1.4.0 (GCOS) software from Affymetrix using the standard default settings. For comparison between different chips, global scaling was used to scale all probesets to a user defined target intensity (TGT) of 150.
Quantitative RT-PCR for Cytokine Expression
Quantitative RT-PCR (Taqman, Applied Biosystems, ABI, Foster City, CA) was utilized to assess levels of cytokine mRNA transcripts of
in vitro FNAR cells as previously described [
20]. The oligonucleotide primers and fluoresceinated probes for the cytokine genes (IL-6, IL-12, and IL-18), ER, PR, and stathmin were purchased from ABI. Data were analyzed in real-time with Sequencer Detection version 1.6 software, with the results normalized against mRNA transcripts for the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GADPH).
Discussion
We present here a model of ovarian carcinoma, designated FNAR, that arose spontaneously in a normal Lewis rat. Importantly, FNAR's biology closely parallels the human disease. IP transplantation into rats produces malignant ascites and peritoneal carcinomatosis, leading to death at 5-6 months. The tumor only develops in the peritoneal cavity, suggesting the tumor microenvironment is intact during formation. Cells from the tumor can be easily passaged in vitro, and the cell line shows similar growth characteristics when returned to rats. Its morphology and expression of EPCAM are consistent with an epithelial carcinoma, and like human ovarian carcinoma, it expresses her-2/neu, sex hormone receptors, and characteristic cytokines. FNAR also displays a similar gene expression pattern to the human disease. Consistent with the endometrioid subtype, FNAR cells show cell-surface expression of CA125 and nuclear expression of β-catenin.
The FNAR model may address many of the limitation of current model systems for ovarian carcinoma. Rats transplanted with FNAR consistently become moribund by 5-6 months, avoiding the low frequency and long latency of spontaneous animal models. Xenografts of primary human tumors in immunodeficient mice are perhaps the most attractive current model [
16‐
19]. Although spontaneous human cancers can be studied and used to test treatments in these mice, the study of immunotherapeutic approaches is problematic. Conversely, FNAR develops in immunocompetent rats, allowing the study of immunotherapeutic approaches. The expression of all three sex hormone receptors and her-2/neu also allows for manipulations of these pathways using this model. However, the application of this model to the treatment of human disease remains to be established.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AS, RJ, and AH designed research. AS, CT, BR, and JB performed research. RG, DA, and CT gave assistance in analyzing model. AS, RJ, AH, and BR wrote manuscript. All authors have read and approved the final manuscript.