HT29 and HCT116 cells were grown in RPMI-1640 (Gibco-BRL) supplemented with 10% fetal bovine serum (FBS) as we previously described (22). For DFX treatments, we utilized concentrations of 0, 50, 100 and 200 μM respectively, of DFX supplemented with 10% FBS. Forty-eight hours later, cell morphological changes were observed or cell lysates were extracted to perform western blot analysis.

Western blotting and immunofluorescence assays were carried out as previously described (23). Antibodies to HIF-1α (Novus Biologicals, Littleton, CO, USA), E-cadherin and vimentin were applied in the immunofluorescence staining (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Vimentin (prediluted), plakoglobin, N-cadherin and glyceraldehyde-3-phosphate dehydrogenase (Abcam, Cambridge, UK) were all commercial products.

Exponentially growing cells were seeded in coverslips and then exposed to 100 μM of DFX. Forty-eight hours later, coverslips were fixed in 4% paraformaldehyde for 15 min, washed in phosphate-buffered saline (PBS) and running water, followed by incubation with hematoxylin solution for 5 min and washed under running water. Subsequently, the coverslips were dipped in acid-ethanol, washed again, and then stained in eosin-ethanol (1% eosin in 80% ethanol) for 3 min, subjected to sequential dehydration and mounted. The coverslips were observed under an Olympus microscope (Olympus Optical Co., Ltd., Tokyo, Japan).

Cells growing on coverslips were fixed with 4% paraformaldehyde at room temperature for 15 min, washed with PBS and permeabilized with 0.1% Triton X-100 for 5 min. After blocking with 1% bovine serum albumin (BSA) for 30 min, cells were incubated with E-cadherin and vimentin antibodies for 2 h. After washing with PBS, IgG Texas Red- (TR-) and IgG-FITC-labeled secondary antibodies were added and incubated for another 1 h. Cells were then rinsed with PBS followed by Hoechst 33258 staining to reveal nuclei. Fluorescence images were captured and analyzed using the Olympus CKX41 fluorescence microscope (Olympus Optical Co., Ltd.).

Substrates for adhesion assays included 25 μg/ml fibronectin or 10 mg/ml heat-inactivated BSA. After overnight coating, wells were rinsed and blocked with 10 mg/ml heat-inactivated BSA as the negative control. Cells were treated with or without 100 μM DFX for 48 h, harvested, resuspended in serum-free medium, and 5×10⁴ live cells were seeded into single wells of a 96-well plate. Cells were then incubated for 2 h at 37°C. Subsequently, the detached cells were washed away, and attached cells were determine by MTT using a microplate reader.

Cell migration and invasion assays were performed using Transwell inserts (BD Biosciences, Bedford, MA, USA) in a similar manner. The difference was that the upper surface of the chambers in the invasion assay was pre-coated with Matrigel (BD Biosciences), while not in the migration assay. Standard protocol was followed. In brief, cells were pre-treated with or without 100 μM DFX for 24 h. Then cells were digested, washed and seeded into the inner chamber (5,000 cells/chamber) in medium without FBS. RPMI-1640 plus 15% FBS was added to the lower chamber, and the plate was incubated for 24 h. Cells on the upper surface of the filter were scraped away with a rubber scraper, while cells on the lower surface were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet, and then counted under a light microscope. For the wound healing assay, DFX-treated cells were seeded in 6-well plates. After scratching the monolayer, cells were washed with PBS, cultured in RPMI-1640 plus 10% FBS, and photographed under a ×10 objective len at the indicated time points. All of the above assays were carried out in duplicate, and the results presented are the means of 5 random fields from each well.

Cells that were treated with or without DFX (100 μM) were harvested and pipetted well to become a single-cell suspension, and then mixed with regular medium containing 0.7% agar resulting in a final agar concentration of 0.35%. The above cell suspensions were immediately plated in 6-well plates coated with 0.6% agar in regular medium and cultured for 14 days. Tumor cell colonies were observed and counted under a dissecting microscope.

Statistical analysis was carried out using SPSS 13.0 for Windows (SPSS, Inc.). Quantitative variables were analyzed using the t-test. P<0.05 was considered to indicate a statistically significant result.

To select a suitable cell candidate to induce HIF-1α expression and EMT, we first detected the endogenous expression level of HIF-1α in 6 colorectal cancer cell lines. As shown in Fig. 1A, the level was much lower in the HT29 and HCT116 cells when compared with that in the SW480, SW620, SW1116 and DLD1 cells. Therefore, HT29 and HCT116 cells were applied in our further experiments.

Next, we incubated HT29 and HCT116 cells with DFX at doses of 0, 50, 100 and 200 μM, respectively, for 48 h, to determine the maximum efficacy at inducing EMT in both cell lines. Treatment with DFX resulted in marked HIF-1α expression in a dose-dependent manner at the range of 0 to 100 μM as shown in Fig. 1B. No significant difference between cells was noted at 100 and 200 μM. Thus, the concentration of 100 μM was used in the following experiments.

Considering that the adhesion of cancer cells to the extracellular matrix (ECM) is the first step in tumor metastasis (24,25), we first evaluated the effect of DFX on cell adhesion. HT29 and HCT116 cells with or without DFX treatment were incubated with one major ECM component in fibronectin-coated 96-well plates for 2 h. We found that DFX treatment increased cancer cell adhesion to fibronectin by 1.96- and 2.23-fold in HT29 and HCT116 cells, respectively, when compared to the cells without DFX treatment (Fig. 2A). Next, we detected the effect of DFX on cell migration and invasion. As shown in Fig. 2B and C, DFX significantly increased migration and invasion of HT29 (2.5±0.5 vs. 1±0.2; P<0.05) and HCT116 cells (1.8±0.2 vs. 1±0.1, P<0.05), when compared with the untreated control cells. Similarly, the wound healing assay also indicated that DFX significantly induced cell migration at 24 and 48 h, respectively (Fig. 2D, P<0.05). Collectively, these observations indicate that DFX induced cell adhesion, migration and invasion abilities in HT29 and HCT116 CRC cells.

Morphological change plays an important role in many cellular processes such as migration, differentiation and apoptosis (22,23,26). We then investigated whether the increased motility and invasive ability of CRC cells was coupled with any morphological change. H&E staining results revealed that treatment with DFX resulted in dramatic morphological alterations in HT29 cells at both high-density and low-density as shown in Fig. 3A. In fact, cells treated with DFX displayed flattened, spread morphology. Cells became more isolated and looked like spindle-shaped and fibroblast-like cells. In contrast the untreated HT29 cells were much more round and had very tight connections. Similar results were found in soft agar assay experiments. As shown in Fig. 3B, cell-cell contacts through the ‘pseudopod’ were present in DFX-treated cells. The edges of these cells commonly showed several finger-like protrusions which may aid in cell motility. In contrast, cells without DFX treatment had a clear boundary and resembled scattered spheres. At the same time, DFX significantly increased anchorage-independent cell growth by 3.1-fold (Fig. 3C).

In addition to alterations in the functions (adhesion, migration and invasion) and cell morphology, the altered expression of EMT markers is another typical feature of EMT and cancer metastasis. We evaluated the effect of DFX on EMT epithelial markers (E-cadherin and plakoglobin) and mesenchymal markers (vimentin and N-cadherin). Western blot analysis indicated that E-cadherin and plakoglobin expression was significantly decreased, while vimentin and N-cadherin expression was increased after DFX treatment for 24 or 48 h (Fig. 4A). As a critical adhesion molecular, E-cadherin is reported as the most important epithelial marker of EMT. We then observed the expression and localization of E-cadherin and vimentin in cells with or without DFX treatment by immunofluorescence staining. As shown in Fig. 4B, the expression of E-cadherin in the cell membrane was markedly decreased in the DFX-treated HT29 cells, which indicated lower cell adhesion ability. At the same time, vimentin expression was upregulated. Furthermore, DFX-treated cells were able to spread much wider on coverslips which was consistent with the H&E staining results (Fig. 3A). These findings suggest that DFX treatment promotes EMT progression in CRC.