Dual action of a selective cyclooxygenase-2 inhibitor on vascular endothelial growth factor expression in human hepatocellular carcinoma cells: novel involvement of discoidin domain receptor 2

Abstract

Purpose: Vascular endothelial growth factor (VEGF) greatly contributes to the progression of hepatocellular carcinoma (HCC). It is reported that a selective cyclooxygenase-2 (COX-2) inhibitor inhibits cellular proliferation and may attenuate VEGF expression in HCC. We propose that different cascades in the VEGF pathway respond to COX-2 inhibition, depending on the cell types.

Methods: The six human HCC cell lines--Hep3B, SNU387, SNU182, SNU423, SNU449, and PLC/PRF5--were cultured under normoxic and hypoxic conditions. Cells were treated with a selective COX-2 inhibitor (NS-398) and discoidin domain receptor 2 (DDR2) siRNA, and microarray analysis was performed.

Results: NS-398 inhibited HCC proliferation and decreased the expression level of VEGF in HCC cells only under normoxia conditions. In hypoxia conditions, VEGF expression level in Hep3B cell was suppressed, while that in SNU387 cell was increased by NS-398 (P < 0.001). The NS-398-induced increase in VEGF expression in SNU387 cell was associated with the up-regulation of the DDR2 gene. NS-398-treated SNU series cells and PLC/PRF5 cells displayed a robust increase in DDR2 mRNA expression. Also, transfection with DDR2 siRNA decreased the VEGF expression level of SNU387, 423, 449 cells under hypoxia conditions (P < 0.05). In vivo chromatin immunoprecipitation assay demonstrated that NS-398 induces the enhancement of HIF-1α binding on VEGF promoter, leading to the increase in VEGF gene expression in hypoxic conditions. There is strong evidence that it is related to the DDR2 gene expression in SNU387 cells.

Conclusion: These findings disclose a novel cell-dependent regulatory mechanism of VEGF involving DDR2 gene in HCC cells.

Fig. 1

Effects of hypoxia on the proliferation of hepatocellular carcinoma cell lines. Hep3B, SNU387, SNU182, SNU423, SNU449, and PLC/PRF5 cells were cultured for 24, 36, and 48 h in normoxia or hypoxia conditions. Proliferation was measured using the MTT assay. Results are the representative of three experiments. *P < 0.05 versus normoxia-exposed cells at various time points

Fig. 2

Effect of a selective COX-2 inhibitor, NS-398, on cell viability of normoxia or hypoxia-conditioned hepatocellular carcinoma cell lines. Hep3B, SNU387, SNU182, SNU423, SNU449, and PLC/PRF5 cells were treated with various concentrations of NS-398 (50 μM, 100 μM) for 24 h in normoxia or hypoxia conditions. Cell viability was measured with the MTT assay. Results are the representative of three independent experiments. *P < 0.05 versus normoxia-exposed HCC cells treated with NS-398 (50, 100 μM), respectively. ^# P < 0.05 versus untreated HCC cells in normoxia

Fig. 3

Effect of the COX-2 inhibitor, NS-398, on VEGF expression in hepatocellular carcinoma cell lines. Hep3B, SNU387, SNU182, SNU423, SNU449, and PLC/PRF5 were maintained under normoxia conditions and subjected to hypoxia (0.1% O₂) at the indicated NS-398 concentrations (100 μM). The VEGF concentration in media was determined using enzyme-linked immunosorbent assay (ELISA), as described in “Materials and methods”. *P < 0.05 versus untreated HCC cells in hypoxic conditions. ^# P < 0.05 versus untreated HCC cells in normoxia

Fig. 4

Validation of the DDR2 gene expression with the NS-398 treated hepatocellular carcinoma cell lines. Hep3B, SNU387, SNU182, SNU423, SNU449, and PLC/PRF5 were plated on 6-well plates, grown to 70–80% confluency, and treated with 50 and 100 μM of NS-398 for 24 h in hypoxic conditions. DDR2 mRNA expression was analyzed with real-time quantitative RT-PCR. Results are the representative of three independent experiments. *P < 0.05 versus NS-398 untreated HCC cells in hypoxic conditions

Fig. 5

Validation of the association between VEGF expression and DDR2 gene in hypoxic-conditioned hepatocellular carcinoma cell lines. Transfection with 20 μM DDR2 siRNA-1 (positions 2,368–2,386 5′-CCAAACAUCAUCCAUCUAU-3′) and DDR2 siRNA-2 (positions 703–721 5′-CCUGAUGACCUGAAGGAGU-3′) in hypoxia-conditioned human HCC cells. After 24 h, the VEGF concentration in media was analyzed using the enzyme-linked immunosorbent assay (ELISA). *P < 0.05 versus SNU387, SNU423, and SNU449 cells transfected with non-target siRNA

Fig. 6

Chromatin immunopreicipitation (ChIP) assay in Hep3B and SNU387 cells: HIF-1α binds to the HIF-1α binding site of the VEGF promoter in vivo. Hep3B and SNU387 cells were plated onto the 150-mm dishes and were treated with NS-398(100 μM) under hypoxic condition for 24 h. Soluble chromatin was immunoprecipitated with HIF-1α antibody. Additionally, the ChIP assay in SNU387 cell was treated with NS-398(100 μM) and DDR2 siRNA under hypoxic condition. The bands represented that immunoprecipitated DNA was analyzed by PCR using the primers corresponding to the HIF-1α binding element of the VEGF promoter. Under panels of the (a) and (b), pictures are the quantification of each band in the ChIP assay data relative to the input control were graphed from the results of the different experiments. P is positive control and represents PCR products using the VEGF-expressed SNU449 genomic DNA (gDNA) as a template. N is negative control and then immunoprecipitated with normal IgG antibody instead of HIF-1α antibody