Hoechst 33342 and Mitomycin C from Streptomyces Caespitosus, luteinizing hormone (LH) and follicle stimulating hormone (FSH) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Mouse monoclonal antibody against p53 and osteopontin and rabbit polyclonal antibodies against BAG-1 and BAX were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Goat anti-mouse IgG-HRP and goat anti-rabbit IgG-HRP were purchased from Pierce Chemical Co. Ltd. (Woburn, MA, USA) and Chemicon International Co. Ltd. (Temecula, CA, USA), respectively.

KGN cells were obtained from RIKEN Bioresource Center (Tsukuba, Japan). Early (fewer than P10) and advanced (more than P47) passage KGN cells were maintained as described previously 21. For detailed characterisation, the cells were cultured in a medium supplemented with 10% charcoal/dextran-treated FBS (delipidated FBS, Thermo Fishers Scientific Inc., South Logan, UT, USA), 0.1 IU/ml LH or 1IU/ml FSH. Cells were routinely passaged after brief exposure to 0.25% trypsin with 0.02% EDTA in PBS (-).

The morphology of KGN cells was observed under a phase-contrast microscope (IX-71; Olympus, Tokyo, Japan). The cellular proliferation was measured by population doubling of KGN cells in a logarithmic growth phase at a starting concentration of 2 × 10⁵ cells/dish in 60 mm Petri dishes (Falcon, BD Japan, Tokyo, Japan). The number of KGN cells was determined at 24 h interval for 4 days while changing the medium every other day. To determine the proliferation rate of KGN cells at Day 4, a colorimetric TetraColor ONE ELISA kit (SEIKAGAKU CORPORATION, Tokyo, Japan) was utilised. Cells were initially seeded at a density of 1 × 10⁴ cells/well in 96-well plates and the subsequent assay was carried out according to the manufacturer's instructions. For invasion assay, KGN cells (1 × 10⁵ cells/well) suspended in serum-free medium were seeded on the upper chamber coated with Matrigel (200 μg/ml, Becton Dickinson, San Jose, CA, USA). FBS (10%) was added to the medium in the lower chamber and incubated for 12 h to allow cell invasion, and then cells were fixed with 100% methanol. To exclude the possibility that invasiveness was overestimated by distinct proliferation capacity of different passage cells, they were treated with 10 μg/ml Mitomycin C for 2 h at 37°C in a CO₂ incubator and subjected to the invasion assay. The cells retained in the upper chamber were scraped off and stained with 10 μM Hoechst 33342 for 30 min at room temperature. The number of cells that invaded through the chamber was counted using an epi-fluorescence microscope. The experiments described above were triplicated.

The chromosomes were examined in exponentially growing early and advanced passage KGN cells (early: P6; advanced: P53) in an in vitro culture. The karyotype of 10 KGN cells was analysed by standard trypsin G-banding, as described previously 1.

Cells were lysed in RIPA buffer containing protease inhibitor cocktails I and II (1:100 dilution, Sigma Chemical Co) for 2 h at 4°C, and the insoluble materials were removed by centrifugation (15,000 rpm, 15 min). The protein concentration of samples was determined using a micro BCA assay kit (Pierce Chemical Co. Ltd., Rockford, IL, USA). Ten or fifty micrograms of samples per lane were separated on a 12.5% reducing SDS polyacrylamide gel, and transferred onto a PVDF membrane (Immobilon-P, Millipore Japan, Tokyo, Japan). After blocking with 10% goat serum plus 90% Blockace (SnowBrand, Tokyo, Japan) at room temperature for 1 h, the membranes were treated with the primary antibody and subsequently with the secondary antibody. SuperSignal West Femto Maximum Sensitivity Substrate (Pierce, Rockford, IL, USA) was used for visualisation, and the signal was developed on an X-ray film (Amersham Bioscience, Piscataway, NJ). The band intensity of each protein was measured using ADOBE Photoshop Element 3.0 software and the background was subtracted. It was then divided by the band intensity of β-actin for normalisation.

The experiments were approved by the Animal Ethics Committee at the National Center for Child Health and Development, Japan. Six-week-old female BALB/c Foxn1/Foxn1 mice (nude mice) were purchased from Sankyo Labo Service Co. Ltd. (Tokyo, Japan). The early or advanced passage KGN cells (5 × 10⁶ cells) were harvested, resuspended in 200 μl of PBS, and injected s.c. into both flanks of each mouse (early; n = 9, advanced; n = 9). Mice were euthanized and examined for tumour generation after 3 months, and the number of nodules formed in the bowel was counted. The experiments were quadruplicated.

For light microscopic analysis, the metastases formed in the bowel were fixed with 4% formaldehyde, paraffin-embedded, sectioned at 4 μm intervals and stained with hematoxylin-eosin.

The small intestine with or without metastases was obtained from four mice in each group (control, the 'early', and the 'advanced' groups) and total RNA was isolated with the use of ISOGEN RNA isolation reagent (Nippon Gene Co. Ltd., Tokyo, Japan). Reverse transcription was conducted, as described previously 21. The expression of p53 was monitored by PCR amplification. The specific primers for human p53 and GAPDH and cycles used for each gene were as follows; p53 forward primer: CAGCCAAGTCTGTGACTTGCACGTAC, p53 reverse primer: CTATGTCGAAAAGTGTTTCTGTCATC, 35 cycles; GAPDH forward primer: ACCACAGTCCATGCCATCAC, GAPDH reserve primer: TCCACCACCCTGTTGCTGTA, 28 cycles. The conditions for PCR amplification were: initial denaturation at 94°C for 5 min, followed by the indicated cycles of denaturation at 94°C for 45 s, annealing at 57°C for 45 s, and elongation at 72°C for 1 min, with a final extension of 72°C for 15 min. The image analysis of the PCR product was performed as described previously 21.

The average data on population doubling, proliferation, invasion and metastasis were presented as means with SEM from three independent experiments. Statistical analyses were conducted by Student's or modified t-test using Microsoft Excel software. Differences were considered statistically significant when P < 0.05.

KGN cells appeared to grow faster during passages in our preliminary experiments. For detailed characterisation of KGN cells, early and advanced passage cells were cultured, and the proliferation capacity was monitored. Growth of the advanced passage cells turned out to be significantly faster than that of the early passage cells, determined by population doubling evaluation (Fig 1A and 1B, about 2-fold increase at Days 3 and 4, P < 0.05, P < 0.01). The behavioural changes of KGN cells at different passages occurred without any morphological changes at Days 1 and 4 (Fig 1A).

Next, we conducted an in vitro invasion assay to assess the characteristic features of the tumours. Along with increased proliferative activity, the advanced passage cells were 2-fold more invasive than the early passage cells (Fig 1C, P < 0.01). Since the advanced passage cells were more proliferative than the early passage cells, KGN cells were treated with Mitomycin C to prevent proliferation during the invasion assay. Loss of proliferative capacity of the advanced passage cells was confirmed by the TetraColor ONE proliferation assay (OD450; untreated early passage cells: 0.10; untreated advanced: 0.18; treated early: 0.09; treated advanced 0.10). Under the conditions of the assay, the advanced passage cells remained more invasive than the early passage cells (Fig 1D, P < 0.01), suggesting that the increased invasiveness was not related to proliferation.

Because serum factors and hormones may be involved in GCT progression, KGN cells were cultured in a medium containing delipidated FBS, LH or FSH, and subjected to the proliferation assay. As shown in Fig 2, none of these treatments affected the characteristics of the early and advanced passage cells.

G-banded karyotype analyses of 10 early and advanced KGN cells revealed that all the cells exhibited 45 chromosomes (data not shown), as described previously 1.

Subsequently, we examined the expression level of malignant tumour markers in the early and advanced passage cells (Fig 3A, B). Initially, we determined the expression level of p53, which has been reported to be a prognostic marker of metastasis in granulosa cell tumours 922. As expected, the expression level of p53 was up-regulated in the advanced passage cells. Next, we performed immunoblots for osteopontin, which is a biomarker of tumour progression 23, and found that it markedly increased in the advanced passage cells. Furthermore, we examined the expression of BAX and BAG-1, which are known biomarkers of apoptosis 24. As expected, BAX was up-regulated, whereas BAG-1 was down-regulated in the advanced passage cells, suggesting that the advanced passage cells might be resistant to apoptosis. These results strongly indicate that KGN cells became aggressive during passages in vitro.

Further characterisation of KGN cells was performed using an in vivo tumour growth assay in nude mice. As shown in Fig 4A, the xenograft of the advance passage cells developed faster than that of the early passage cells at the region where KGN cells were injected, and larger clumps under the skin were observed (in the small windows of Fig 4A). More interestingly, KGN cells at both the early and advanced passages metastasized from the subcutaneous transplanted region to the bowel, especially in the submucosa of the small intestine, but not to the other tissues (metastases in the bowel were seen in 7 out of 9 mice transplanted) (Fig 4B). Although an increased number of nodules were formed in the bowel by transplantation of the advanced passage cells (Fig 4C, P < 0.05), the size of nodules obtained using both passage cells was similar (3 mm × 3 mm).

To confirm the origin of nodules, the expression of a GCT-specific marker in the nodules was examined by RT-PCR using human-specific primers. As shown in Fig 4D, human p53 mRNA was found to be expressed in the nodules, suggesting that the nodules originated from KGN cells. Moreover, histological analysis revealed that the cells at both passages exhibited a coffee bean-like nuclear appearance, which is typically observed in specimens of granulosa cell origin (Fig 5). These results demonstrated that KGN cells metastasized from the subcutaneous region to the bowel.