The MHCC97H human HCC cell line was cultured in DMEM medium supplemented with 10% fetal calf serum (Sigma Chemical Co., St. Louis, MO, USA) and 1% penicillin-streptomycin G (Invitrogen Life Technologies, Carlsbad, CA, USA). In all experiments, the cells were cultured at 37°C in a humidified 5% CO₂/95% air atmosphere. To identify and isolate the SP and MP fractions, we used flow cytometric analysis as described previously (6). Briefly, the cells were preincubated at 37°C for 15 min and stained with Hoechst 33342 dye (Sigma Chemical Co.) at 6 μg/ml. The cells were then incubated for 90 min at 37°C alone or with 50 μM verapamil (Sigma Chemical Co.), and then stained with 2 μg/ml of propidium iodide (BD Pharmingen, San Diego, CA, USA) to label dead cells. The cells were then filtered through a 40-μm strainer (BD Falcon) and maintained at 4°C before analysis and sorting using a FACSAria flow cytometer (BD Falcon). Samples of sorted cells were reanalyzed to examine the sorting purities.

Total RNA, including miRNAs, was isolated from MCHH97H cells with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Expression of hsa-miR-21 was analyzed with the miScript system (Qiagen, Valencia, CA, USA), which consists of the miScript Reverse Transcription kit, miScript Primer assays and miScript SYBR-Green PCR kit, according to the manufacturer’s instructions. Small nuclear RNA U6 was used for normalization. Quantitative RT-PCR (RT-qPCR) was run on the ABI PRISM 7700 Sequence Detector (Applied Biosystems, Foster City, CA, USA). Reactions were run in triplicate. The ΔΔCt method was used for relative quantification of the gene expression to determine miR-21 expression levels.

The cells were lysed in lysis buffer as previously described (15) by incubating for 20 min at 4°C. The protein concentration was determined using the Bio-Rad assay system (Bio-Rad Laboratories, Hercules, CA, USA). Total proteins were fractionated using SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat dried milk or bovine serum albumin in 1X TBS buffer containing 0.1% Tween-20 and then incubated with the appropriate primary antibodies. Horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG was used as the secondary antibody and the protein bands were detected using the enhanced chemiluminescence detection system (Amersham Pharmacia Biotech, Amersham, UK). Quantification of the western blots was performed using laser densitometry and the relative protein expression was then normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels.

JSI-124 (cucurbitacin-I, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) was dissolved in DMSO. Cells were treated with the appropriate volume of DMSO for the vehicle control. Following treatment with different concentrations of JSI-124 or DMSO, the cells were washed twice in PBS and resuspended in 500 μl of 1X binding buffer prior to incubation with 5 μl of Annexin V and 10 μl of PI. The cells were then analyzed by using flow cytometry after incubation for 5–10 min in the dark. Early apoptotic cells were stained with Annexin V alone whereas necrotic and late apoptotic cells were stained with both Annexin V and PI.

SP cells were resuspended and maintained at 37°C. Human CXCL12 (Peprotech, Inc., Rocky Hill, NJ, USA) or PBS were added to cell suspensions and aliquots were removed at the indicated times and immediately fixed in 4% paraformaldehyde for 10 min. After washing, the samples were stained with FITC Phalloidin (Molecular Probes, Eugene, OR, USA) stain F-actin and analyzed by flow cytometry. The relative F-actin index was determined as the ratio of the F-actin level of different cells treated with CXCL12 to cells treated with PBS.

Cell migration was analyzed with non-Matrigel-coated Transwell cell culture chambers (Millipore, Billerica, MA, USA). Cell invasion was analyzed with Matrigel-coated Transwell cell culture chambers (Millipore). Cells treated with different reagents (5×10⁴ cells/well) were serum-starved for 24 h and plated in the upper insert of a 24-well chamber in serum-free medium. A medium with or without CXCL12 chemokine was added to the well. After incubation for 24 h, the cells on the upper side of the filters were mechanically removed by scrubbing with a cotton swab, after which the membrane was fixed with 4% formaldehyde for 10 min at room temperature and stained with 0.5% crystal violet for 10 min. Invasive or migrated cells were counted at a magnification of ×200 from 10 different fields of each filter.

Each experiment was repeated at least three times. The data were summarized and presented as means ± SD. The differences among means were statistically analyzed using a t-test. Statistical analyses were performed using SPSS 13.0 software (Chicago, IL, USA). P<0.05 was considered statistically significant.

Using flow cytometry, we identified and successfully isolated SP and MP cell populations from MHCC97H cell lines. The SP gate was defined as the region where cells were absent in the presence of verapamil, an agent that blocks the efflux of Hoechst 33342. The SP cells accounted for 2.97±0.33 of the total cells in the MHCC97H cell line. The purities of sorted SP and MP cells were >97%. As previously reported (6), MHCC97H SP cells exhibited stem-like characteristics including tumorigenic potential and chemoresistance migration and invasion abilities, which may lead to relapse and metastasis formation.

STAT3 and miR-21 are known to be overexpressed in carcinogenesis and metastasis. We assessed the expression levels of STAT3, phospho-STAT3 and miR-21 in SP cells as compared to MP cells, respectively. RT-qPCR was performed and a higher expression of miR-21 was observed in SP compared to MP cells (Fig. 1A). Western blot results showed that the levels of STAT3 and phospho-STAT3 were higher in SP compared to MP cells (Fig. 1B). These data indicated that STAT3, phospho-STAT3 and miR-21 were overexpressed in SP cells of MHCC97H cell line.

Accumulating evidence indicated that STAT3 plays a key role in the maintenance of stem-like cancer cells associated with tumor recurrence, metastasis and chemo-resistance (16). As the effects of STAT3 on HCC SP cells were to be determined, we used JSI-124 to inhibit STAT3 and observed the role of STAT3 on MHCC97H SP cell metastatic capacity.

STAT3 inhibition induced by JSI-124 was evaluated by western blotting. JSI-124 was dissolved in DMSO. The cells were treated with the appropriate volume of PBS or DMSO for the vehicle control. The results showed that phospho-STAT3 protein levels were downregulated significantly when treated with different concentrations of JSI-124 (50 and 100 nM). The protein levels decreased with the increase of JSI-124 concentrations (Fig. 2A). In order to exclude the possibility that the downregulation of phospho-STAT3 was caused by cell apoptosis, we evaluated JSI-124-treated cells by Annexin V/PI analysis. The results did not indicate significant cell apoptosis in either concentration (Fig. 2B).

Actin polymerization was required by cell polarization. Our previous data indicated that MHCC97H SP cells contained higher actin polymerization levels than MP cells when treated with CXCL12. In the present study, we determined whether STAT3 affected actin polymerization of SP cells. As shown in Fig. 3, we treated MHCC97H SP cells with PBS, DMSO and 50 or 100 nM JSI-124, respectively, and then detected the F-actin levels of different cells after 3, 5 and 10 min stimulation with CXCL12. Actin polymerization levels of cells treated with PBS or DMSO were similar to each other. Actin polymerization levels of cells treated with 50 or 100 nM JSI-124 were significantly lower than cells treated with PBS at the studied time points (P<0.05, P<0.01). This result suggested that the higher actin polymerization level of SP cells treated with CXCL12 was reduced by STAT3 inhibitor JSI-124 (Fig. 3A).

To examine the possible differences in chemotaxis towards CXCL12 between SP cells treated with different reagents, we performed a Transwell-based migration assay. The results indicated that SP cells showed higher migration ability towards CXCL12 rather than PBS (P<0.05). However, when treated with 50 or 100 nM JSI-124, the chemotaxis towards CXCL12 decreased significantly (P<0.01) (Fig. 3B).

We previously reported MHCC97H SP cells exhibited higher abilities of migration and invasion than MP cells and may be important in HCC metastasis. We wondered whether STAT3 affect the metastasis capacities of SP cells. Migration and invasion cell numbers were counted through the Transwell system. The results showed that 50 and 100 nM JSI-124 decreased the SP cell abilities of migration and invasion (P<0.05) (Fig. 3C).

It was reported that STAT3 binds to multiple sites in the miR-21 promoter and was required for the induction of transcription in normal or cancer cells (9,17,18). Our data indicated STAT3 phosphorylation affected metastasis-related capacities in vitro of HCC SP cells. We also improved the repression of miR-21-inhibited SP cell migration and invasion in vitro due to upregulation of its target tumor-suppressor genes, i.e., PTEN, RECK or PDCD4. However, whether the effects of phospho-STAT3 on metastasis-related capacities of MHCC97H SP cells were mediated by miR-21 and its targets remained to be determined. We measured the expression of miR-21 in cells treated with PBS (Basal), DMSO, and 50 or 100 nM JSI-124 through RT-qPCR. It was observed that 50 and 100 nM JSI-124 repressed the expression of miR-21 (P<0.05, compared with Basal) (Fig. 4A). Western blot results showed that the levels of PTEN, RECK and PDCD4 were increased with the increasing concentrations of JSI-124 (Fig. 4B).

JSI-124 inhibited phospho-STAT3, while miR-21 and its target gene expression were increased. We determined whether STAT3 was able to regulate PTEN, RECK and PDCD4 via miR-21. MHCC97H SP cells were treated with PBS (Basal), and 100 nM JSI-124 with or without transfection with miR-21 mimics. Transfection with miR-21 mimics led to upregulated miR-21 expression, whereas the increased protein expression of PTEN, RECK and PDCD4 was markedly reduced.

We also examined whether inhibition of phospho-STAT3 is involced in the migration and invasion of MHCC97H SP cells via miR-21 and its targets. Western blotting indicated that 100 nM JSI-124 did not induce SP cell migration or invasion when miR-21 expression was upregulated (Fig. 5A). In addition, cell migration and invasion were increased by silencing PTEN, RECK or PDCD4 and these increases were reduced by JSI-124 (all P<0.05) (Fig. 5B). These results indicated that the metastatic effect of phospho-STAT3 is partly mediated through miR-21 and its negative regulation of PTEN, RECK and PDCD4 expression.