SOX9 is a novel cancer stem cell marker surrogated by osteopontin in human hepatocellular carcinoma

The current lack of cancer stem cell (CSC) markers that are easily evaluated by blood samples prevents the establishment of new therapeutic strategies in hepatocellular carcinoma (HCC). Herein, we examined whether sex determining region Y-box 9 (SOX9) represents a new CSC marker, and whether osteopontin (OPN) can be used as a surrogate marker of SOX9 in HCC. In HCC cell lines transfected with a SOX9 promoter-driven enhanced green fluorescence protein gene, FACS-isolated SOX9+ cells were capable of self-renewal and differentiation into SOX9− cells, and displayed high proliferation capacity in vitro. Xenotransplantation experiments revealed that SOX9+ cells reproduced, differentiated into SOX9− cells, and generated tumors at a high frequency in vivo. Moreover, SOX9+ cells were found to be involved in epithelial-mesenchymal transition (EMT) and activation of TGFb/Smad signaling. Gain/loss of function experiments showed that SOX9 regulates Wnt/beta-catenin signaling, including cyclin D1 and OPN. Immunohistochemistry of 166 HCC surgical specimens and serum OPN measurements showed that compared to SOX9− patients, SOX9+ patients had significantly poorer recurrence-free survival, stronger venous invasion, and higher serum OPN levels. In conclusion, SOX9 is a novel HCC-CSC marker regulating the Wnt/beta-catenin pathway and its downstream target, OPN. OPN is a useful surrogate marker of SOX9 in HCC.


4
We performed single-cell culture analyses as previously described. 14,47,48 The individual isolated cells were each sorted into 96-well culture plates using FACSAria (BD Biosciences). We used a light microscope 10-16h after cell sorting to confirm that each well contained only one cell. Following cell expansion after isolation of each clone, we subjected the cells to flow cytometry.
To examine the anchorage-independent growth, 1 × 10 4 EGFP + and EGFP − cells were suspended in 2.0 mL of 0.3% agar (Wako) supplemented with culture medium. The 5 cell suspension was layered over the bottom layer of 2.0 mL of 0.6% agar. We counted the colonies 14 days after cell sorting.
To investigate the ability to form cell spheres, 1 × 10 5 EGFP + and EGFP − were seeded in 6-well ultra-low attachment plates (Corning Inc., NY, USA) in serum-free medium. We observed the spheres 5 days after cell sorting.
To investigate the differentiation ability in vivo, we performed serial transplantation. Tumors generated from single-cell-derived EGFP + or EGFP − cells were harvested, and then digested with collagenase solution for 30 min at 37°C. After rinsing the tumors in phosphate-buffered saline (PBS), we analyzed and sorted the cells using the FACSAria. Prior to the second transplantation, the harvested cells were cultured in G418-containing medium for at least 7 days to eliminate cells that originated from the host mice. Thereafter, we sorted these cells according to the EGFP fluorescence, and transplanted 1 × 10 4 EGFP + or EGFP − cells into NOD/SCID mice in the same way as in the first transplantation. Wound healing assay and migration assay Wound healing assays were used to assess capacity for cell motility. We seeded the isolated EGFP + and EGFP − cells differentiated from one EGFP + cell at a density of 1 × 10 6 cells per well in 35-mm culture dishes. On reaching full confluency, the cell layer was scratched with a 10-µL plastic tip and then cultured with low serum (2% fetal bovine serum) culture medium and 10 ng/ml TGFb. Micrographs were taken at 24 h after the scratch.
For migration assays, 8-µm-pore 24-well cell culture plates (Corning Inc.) coated with type I collagen were used. We plated 2.5 × 10 4 EGFP + and EGFP − cells in the upper chamber with serum-free medium and 10 ng/ml TGFb; in the lower chamber, normal culture medium containing 10% fetal bovine serum was added. After 48 h of incubation, the cells on the upper surface of the membrane were removed, and the cells on the lower surface were fixed with 4% paraformaldehyde and stained using the Diff-Quick staining kit (Sysmex, Kobe, Japan).  T-cell factor/lymphoid enhancer factor (TCF/LEF) luciferase assay To evaluate the activation of Wnt/beta-catenin pathway in HCC cells, we performed TCF/LEF luciferase assay in addition to immunocytochemistry of activated beta-catenin.
Cignal TCF/LEF Reporter Assay Kit (Qiagen) and Dual-Glo TM Luciferase Assay System (Promega) were used according to the manufacturer's protocols.
Enzyme-linked immunosorbent assays (ELISA) To measure the serum OPN level of HCC patients, we performed ELISA. Blood samples were collected and stored at 4°C for 1 h. Serum was separated by centrifugation (10,000 g for 10 min) at 4°C, and then stored at −80°C until analysis. We quantitated serum OPN levels, respectively, with the human OPN ELISA kits (Millipore) according to the manufacture's protocols. The OPN concentration was calculated by a standard curve. Each serum sample was tested in duplicate, and the results were averaged to maintain reliability.

Statistical analysis
The statistical analyses were performed using SPSS version 17.0 (SPSS Inc., IL, USA) and GraphPad Prism software version 5.0 (GraphPad Software Inc., San Diego, CA, USA). Student's t-test, F-test, Mann-Whitney U test, Fisher's exact test or chi-squared test, log-rank test, and repeated-measures analyses of variances were used for assessment. The tumor initiating frequency in xenotransplantation was calculated by extreme limiting dilution analysis. 54 The mean ± SD of three or more independent experiments is reported.
Recurrence-free survival (RFS) and overall survival (OS) after the operation were calculated using the Kaplan-Meier method and analyzed with the log-rank test.
Significant variables from the univariate analysis were entered in the multivariate analysis using a Cox regression model with forward stepwise selection. We plotted receiver operating characteristics (ROC) curve for serum OPN level and HCC tumor markers, and calculated area under the ROC curve (AUC). The optimal cutoff value of 12 OPN was calculated using the maximum sum of sensitivity and specificity as well as using the minimum distance to the top-left corner of the ROC curve. Statistical significance was defined as P < 0.05.
Supplemental Table 1 Tumor incidence in initial/serial transplantation initial Abbreviation: SOX9, sex determining region Y-box 9; Tx, transplantation.
Supplemental Table 2 The tumor initiating frequency in HCC cells