Establishment of a new representative model of human ovarian cancer in mice

The patient was a 53-year-old post-menopausal woman presenting with a 2-month history of abdominal distention with a mass in the right adnexa. Cytological study of the ascites indicated presence of adenocarcinoma cells, with a serum CA125 level of 1364 U. She received intraperitoneal chemotherapy with Thiotepa twice (1^st: 30mg, 2^nd: 20mg) before surgery (cyto-reductive surgery, extrafascial hysterectomy, bilateral adnexectomy, omentectomy, appendectomy). Poorly differentiated serous adenocarcinomas of the ovary (stage FIGO IIIC) were pathologically confirmed. Informed consent was acquired from the patient.

A tumor specimen (ascites) was collected and sterilely transported on ice to the laboratory. The ascites was centrifuged at 300G for 5min. The cell pellet was immediately intraperitoneally injected into the NOD/SCID mice. The number of tumor cells from the patient inoculated into the two mice was 5x10E7.

When the animals had marked abdominal distensions, their ascites was collected from the tumor-bearing mice, and directly injected into the next mice. The mice were then monitored daily for general health status as well as the degree of abdominal extension before another passaging. Xenografts were established within 6 months after heterotransplantation in both mice.As passaging continued, the number of tumor cells we used was 1x10E7. The peritoneal cavity of each animal was examined. Tissues and organs suspected of being affected by tumor were eventually harvested for histological examinations.

After the animals were sacrificed, ascites was collected and centrifuged at 300G for 5min. The pellet was suspended in cold freezing media, consisting of DMEM with 8% DMSO and 20% fetal bovine serum (FBS). The cells were then transferred to liquid nitrogen for long-term storage. For retransplantation, the tumor suspension underwent a rapid thaw at 37°C, and was diluted in 10ml of DMEM without FBS. Then it was centrifuged, and intraperitoneally injected into the mice.

Tumors in the peritoneal cavity were fixed in 4% neutral buffered formalin, and processed for histological examination. The fixed samples were embedded in paraffin, 5μ-thick sections were cut, and H-E staining was applied. Cell suspensions from ascites were cyto-centrifuged, and smear was fixed with cold acetone. Endogenous peroxidase activity was then blocked with 3% hydrogen peroxide in PBS for 10min. The samples were subsequently rinsed with PBS for 5min. All tissues were blocked with BSA for 20min at room temperature. Primary monoclonal antibodies to vimentin (dilution 1:300), pan-cytokeatin (dilution 1:400), EMA (dilution 1:500), P53 (dilution 1:300), MMP-2 (dilution 1:300) (all supplied by Wuhan Boster Bio-Engineering Co. Ltd., Wuhan, China), and CA125 (dilution 1:400) (supplied by Beijing Zhongshan Golden Bridge Biotechnology Co. Ltd., Beijing, China) were used. Sections were incubated with the antibodies overnight at 4°C following three 5-min washes with PBS, then incubated with biotinylated secondary antibody at 37°C for 20min, and finally incubated with streptavidin-biotin-horse radish peroxidase complex at 37°C for 20min. The sections were then developed with diaminobenzidine (DAB) as a substrate. Cellular nuclei were counterstained with hematoxylin. All operations followed the manufacturers’ recommendations. Absent primary antibody during processing was used as blank control. Immunostaining was semi-quantitative as described [19].

Whole blood and ascites specimens were collected from the tumor-bearing mice, and centrifuged for 10min. The supernatants were isolated and stored at −20°C for subsequent analysis. Levels of CA125, CA199, AFP and CEA were determined.

Ascites specimens were collected and centrifuged. Cell pellet was resuspended in DMEM supplemented with 10% heat-inactivated fetal calf serum (FCS), and exposed to 0.5μg/ml demecolcine (Colcemid) for 2 hours. Cytogenetic analysis was accomplished by G-banding using a conventional Giemsa staining protocol [20]. Briefly, tumor cells were harvested and incubated in 75mM potassium chloride at 37°C. Then they were fixed with 25% acetic acid in anhydrous methanol twice. The slides were trypsinized and stained with 10% Giemsa, and 35 metaphase spreads were karyotyped.

Tumor cells in the ascites from the patient with poorly differentiated ovary papillary serous adenocarcinoma were injected into the peritoneal cavities of female NOD/SCID mice. Xenografts were established within 6 months after heterotransplantation. As passaging continued, the time to achieve a clinically evident disease decreased to 1–2 months. All mice eventually developed visible abdominal swelling (Figure 1A) and yellowish pale ascites (Figure 1B). The volume of ascites was generally 4 to 8ml per mouse. Abundant cells of clinical resemblance were seen individually or in clusters on cytological examination of the ascites (Figure 1C). The malignant features of the cells were confirmed by high nuclear-cytoplasmic ratio, nuclear multiforme and predominant nucleoli; and these histological characteristics were documented in different mouse generations. The mean viable tumor cell yield from ascites was approximately 1.5 × 10⁹ cells per mouse. The malignant cells in the ascites could be reproducibly maintained and directly passaged to subsequent hosts. The cells have been serially transplanted 8 times to date. This model may be reproduced using our cryopreserved tumor cells in liquid nitrogen free of histological changes.

The metastatic pattern was investigated, where similar results were noted in different generations. In the advanced stage of this model, the mesentery was most frequently invaded by tumor tubercles (Figure 1D). Deposits of tumor cells were found on the surfaces of peritoneal organs, such as the sub-diaphragmatic and liver linings, which replicated the initial mice (Figure 1E). The reproductive system was severely affected by tumor cells (Figure 1F). Severe unilateral or bilateral ovarian pathogenesis was found in 75% of the animals in our model. In most cases, the histomorphology and cytology of ovary changed as the condition developed from crowding and displacement to complete invasion. All suspicious areas were biopsied. No evidence of extraperitoneal spread was noted.

Human female origin of the established xenograft model was confirmed by cytogenetic analysis. Thirty-five banded metaphases were analyzed. Chromosome number ranged from 54 to 60 (median: 58), mainly trisomy (Figure 2H).

To analyze the histopathology of the tumor cells in the ascites, immunohistochemical staining was performed to determine expression of epithelial membrane antigen (EMA), Pan-cytokeatin, vimentin (VIM), PCNA, P53, CA125 and MMP-2. The results are given in Table 1 and Figure 2. It’s found that there is no significant changes in comparison with the first generation (the pictures were showed as Additional file 1: Figure S3).

In analyzing the CA125, CA199, CEA and AFP levels in serum and ascites, we found that only the CA125 level rose in both serum and ascites in the tumor-laden mice. The serous CA125 level ranged 293.4-1024.4U/ml. Supernatant of the malignant ascites was found to contain 484.2-3800U of CA125 per ml. The normal serous CA125 level (< 11 U/ml) was documented in all healthy control mice.