The human endometrial cancer cell lines, Ishikawa and MFE280, were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS), streptomycin (100 μg/ml), and penicillin (100 IU/ml) in the presence of 5% CO₂.

Cells were incubated in phosphate-buffered saline (PBS) containing 0.5% bovine serum albumin (BSA) and 2 mM EDTA with phycoerythrin (PE)-conjugated CD133/2 (clone 293C3) antibodies (Miltenyi Biotec, Auburn, CA, USA). Mouse IgG2b-PE (Miltenyi Biotec) was used as the isotype control antibody. To identify and isolate SP cells, cells were stained with Hoechst 33342 (Sigma-Aldrich), either alone or in combination with 100 μM verapamil (Sigma-Aldrich). For flow cytometry and cell sorting, samples were analyzed using the JSAN desktop cell sorter and AppSan software (Bay Bioscience Co. Ltd., Kobe, Japan).

Cell proliferation was determined using the WST-1 reagent (Roche Diagnostics, Tokyo, Japan). Briefly, 2×10³ cells were seeded in 96-well plates and incubated in normal medium conditions at 37°C. On designated days, WST-1 reagent (10 μl) was added to each well, and the cells further incubated for 2 h at 37°C. Absorbance was measured using a microplate reader at test and reference wavelengths of 450 and 655 nm, respectively.

Diluted single cells (5×10⁴) were seeded onto 60-mm dishes containing 0.33% soft agar in DMEM supplemented with 10% heat-inactivated FBS on top of 0.5% base agar in DMEM supplemented with 10% heat-inactivated FBS. Colonies with diameters larger than 0.25 mm after 14 days incubation were counted.

Cells (1×10⁴) were seeded in 24-well plates, incubated for 24 h, then treated with designated concentrations of paclitaxel (provided by Bristol Pharmaceuticals, Tokyo, Japan). After incubation for 48 h, the cells were counted using a hemacytometer with trypan blue staining.

Total RNA was isolated from the cells using the RNeasy mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol. Complementary DNA was synthesized from 2 μg of RNA using the Omniscript RT kit (Qiagen) with random primers and amplified together with Taq polymerase (Nippon Gene, Tokyo, Japan) for the amplification of ABCG2, MT1-MMP and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Primers and PCR conditions are listed in Table I. Real-time PCR (quantitative PCR; qPCR) was performed using a LightCycler and a SYBR-Green system (Applied Biosystems, Foster City, CA, USA). Microarray analyses were performed using 3D-Gene Human Oligo chip 24k (Toray, Tokyo, Japan).

Whole cell extracts were prepared using 1X lysis buffer (Cell Signaling, Danvers, MA, USA) and concentrations were determined using the Bradford protein assay (Life Science, Hercules, CA, USA). SDS-PAGE and western blot analysis were performed as described previously, using 50–100 μg of protein (15,16). MT1-MMP antibody (clone 114-6G6, Fuji Chemical Industries Ltd, Takaoka, Japan) was used at 1:25, ABCG2 (ab3380, Abcam) and GAPDH (Abcam) at 1:10,000.

Cells were seeded and transfected with 30 nM of negative control small interfering RNA (siRNA) or human MT1-MMP siRNA oligonucleotides (Applied Biosystems) using Lipofectamine 2000 Transfection Reagent (Invitrogen) according to the manufacturer’s protocol.

The invasive ability of CD133⁺ or CD133⁻ cells was assayed in vitro using a Biocoat Matrigel Invasion Chamber (Becton-Dickinson Biosciences, Bedford, MA, USA), as described previously (17). Cells were suspended in the upper wells of Matrigel chambers, in DMEM containing 0.1% BSA. After a 22-h incubation, cells on the upper surface of the membrane were removed by wiping with cotton swabs, and cells that had migrated through the membrane to the lower surface were fixed with methanol and stained with Mayer’s Hematoxylin and eosin. The cells on the lower surface of the membrane were counted microscopically to obtain the invasion index. Chemotaxis assays were performed in the same manner, except that the filters were not coated with Matrigel, and the number of cells on the lower surface of the membrane was considered the migration index. The invasive ability of cells was described as the ratio of the invasion index to the migration index.

The supernatants of cells were subjected to gelatin zymography with Gelatin zymography kit (Primary Cell, Sapporo, Japan) according to the manufacturer’s directions.

Statistical analysis was carried out using the statistical package StatView version 5.0 (Abacus Concepts, Berkeley, CA, USA). We used the Student’s t-test for in vitro experiments. A p-value of <0.05 was considered to indicate statistical significance.

In our previous study (14), we examined the frequency of CD133⁺ cells in 6 endometrial cancer cell lines, among which Ishikawa and MFE280 cells exhibited significant levels of CD133 expression detectable by FACS analysis. Furthermore, only Ishikawa cells had distinct side population fraction, another hallmark of CSC. Thus, in the present study, we mainly used Ishikawa cells for the subsequent analyses. To characterize CD133⁺ cells in endometrial cancer, we first performed cDNA microarray analyses with Ishikawa cells. A total of 440 genes were found to be overexpressed in CD133⁺ cells at least 2-fold compared with CD133⁻ cells: genes of ABC transporters, cytokines, growth factors and invasion molecules were included in the total (Table II). In contrast, a total of 96 genes were downregulated in CD133⁺ cells at least 2-fold compared with CD133⁻ cells (Table III). Among these genes, we paid special attention to the multi-drug resistance gene ABCG2, because of the previous findings that CD133⁺ endometrial cancer cells are more resistant to chemotherapeutic agents such as paclitaxel and cisplatin than CD133⁻ cells (14). Increased expression of ABCG2 in CD133⁺ Ishikawa cells was confirmed by RT-PCR and western blot analyses (Fig. 1A).

Based on the microarray analyses, we were interested in the relationship between CD133⁺ cells and SP cells, since the latter specialized populations are known to highly express ABCG2. Flow cytometry was used to examine the SP fraction in CD133⁺ or CD133⁻ Ishikawa cells. The ratio of the SP cells in CD133⁺ cells was calculated as 0.69% (±0.14%) based on the control with verapamil, whereas it was 0.49% (±0.17%) in CD133⁻ cells. These findings indicate that there are more SP cells in CD133⁺ cells (Fig. 1B).

A recent report indicated that SP cells in endometrial cancer are potential CSCs (18). Therefore, we confirmed such potential of Ishikawa-SP cells by chemosensitivity or colony-formation assay. Ishikawa-SP or non-SP cells were treated with paclitaxel at 2 or 10 nM for 48 h and the cells counted. As shown in Fig. 1C, SP cells are more resistant than non-SP cells to paclitaxel at 10 nM (Fig. 1C). A colony-formation assay was performed, in which Ishikawa-SP or non-SP cells were seeded onto soft agar and colonies larger than 0.25 mm in diameter after incubation for 14 days were counted. The SP cells showed significantly greater colony-forming ability than non-SP cells (Fig. 1D), consistent with a previous report (18).

We next examined the tumorigenic potential of SP cells with or without CD133 expression. CD133⁺/SP, CD133⁺/non-SP, CD133⁻/SP and CD133⁻/non-SP Ishikawa cells were sorted, purified and cultured in normal growth medium for 8 days, and cell growth compared by WST assay. As shown in Fig. 2A, in normal growth medium, CD133⁺/SP Ishikawa cells grew significantly faster than any other group.

To evaluate anchorage-independent growth of these cells, we assessed their colony-forming ability in soft agar. CD133⁺/SP cells formed more colonies than CD133⁺/non-SP, CD133⁻/SP or CD133⁻/non-SP cells (Fig. 2B). These results suggest that CD133⁺/SP Ishikawa cells have the highest potential as CSCs.

We previously reported CD133 to be a prognostic factor in endometrial cancer (14). The precise mechanisms, however, remained unclear, but we speculated that increased invasive ability of CD133⁺ cells might be the key. Among the genes highly expressed in CD133⁺ cells, we took particular notice of matrix metalloproteinase genes (MMPs) involved in cellular invasion. We investigated the expression of a total of 26 MMPs in both CD133⁺ and CD133⁻ Ishikawa cells and found that only MMP14 (MT1-MMP) showed higher expression (>2-fold) in CD133⁺ cells than CD133⁻ cells. We confirmed by RT-PCR and western blot analysis that MT1-MMP was preferentially expressed in CD133⁺ cells (Fig. 3A).

MT1-MMP plays a critical role in tumor invasion and metastasis. We evaluated the invasion ability of CD133⁺ or CD133⁻ endometrial cancer cells using an in vitro invasion assay. Significant differences in invasive ability were observed between CD133⁺ and CD133⁻ Ishikawa cells (50.5 vs 21.4%) (Fig. 3B), which were confirmed in another cell type (MFE280 cells): CD133⁺ MFE280 cells showed higher invasive ability than CD133⁻ cells (73.6 vs 46.1%). These results suggest that CD133⁺ endometrial cancer cells have increased invasive activity compared with CD133⁻ cells.

To confirm whether elevated MT-MMP1 expression is essential for increased invasive ability of CD133⁺ endometrial cancer cells, we performed siRNA knockdown experiments of MT1-MMP. RT-PCR and western blot analysis showed successful knockdown of MT1-MMP by siRNA (Fig. 3C). Gelatin zymography also showed significant inhibition of enzymatic activity of MT1-MMP by siRNA. In vitro invasion assay revealed that the knockdown of MT1-MMP led to decreased invasive ability in both Ishikawa and MFE280 cells, from 147.6 to 52.0% and from 38.0 to 19.1%, respectively (Fig. 3C), confirming that MT1-MMP influences their invasive capacity. We also examined the invasive ability of sorted CD133⁺ and CD133⁻ endometrial cancer cells, with or without knockdown of MT1-MMP. In CD133⁺ Ishikawa and MFE280 cells, invasive ability was significantly decreased by successful knockdown of MT1-MMT from 159.7 to 63.6%, and 10.6 to 6.3%, respectively: knockdown efficacies in mRNA expression were 60 and 45%, respectively, compared with mock transfected cells (Fig. 3D). Knockdown of CD133⁻ Ishikawa and MFE280 cells was not sufficient, and invasive ability was not affected in these cells (47.9 vs 51.6%, 3.7 vs 2.3%, respectively). This was because constitutive levels of mRNA MT1-MMP expression were much lower in CD133⁻ Ishikawa and MFE280 cells than in CD133⁺ Ishikawa and MFE280 cells. These findings indicate that increased expression of MT1-MMP in CD133⁺ endometrial cancer cells contributes to their invasive ability.