Cyclooxygenase-2 (COX-2) is directly involved but not decisive in proliferation of human hepatocellular carcinoma cells

Abstract

Expression of cyclooxygenase-2 (COX-2) is involved in the chronic inflammation-related development of hepatocellular carcinoma (HCC), and the use of selective COX-2 inhibitors might provide new chemoprevention strategies for HCC. However, the role of the COX-2 in hepatocarcinogenesis remains obscure, particularly as it has been primarily studied with selective COX-2 inhibitors that may affect other cellular proteins involved in cell proliferation. Therefore, we investigated the effects of the inhibition of COX-2 by the selective COX-2 inhibitor NS-398 as well as by COX-2 specific small interfering RNA (siRNA) in the human HCC cell lines Hep3B and SNU-387. These cell lines expressed COX-2, and NS-398 induced apoptosis of these cells. NS-398 inhibited more than 60% of prostaglandin E(2) (PGE2) production and cell proliferation in a concentration-dependent manner in these cells. The inhibition of proliferation was almost restored with PGE2 supplement, suggesting that NS-398 may inhibit cell growth partially through inhibition of COX-2 and PGE2 production in human HCC cells. However, treatment with NS-398 led to increased expression of COX-2 in Hep3B and SNU-387 cells. To examine the effect of COX-2 depletion on these cells, we electroporated COX-2-specific siRNAs into SNU-387 cells. We observed significant, sequence-specific reductions in COX-2 expression, PGE2 production, and cell proliferation, though the reduction in cell proliferation was less than that induced by NS-398. In conclusion, these data suggest that COX-2 itself is directly involved, though not decisively, in proliferation of human HCC cells. RNA interference may provide a useful tool for manipulating COX-2-related hepatocarcinogenesis in research and therapeutic settings.

Fig. 1

Constitutive expression of COX-2 in human HCC cell lines. COX-2 protein levels in the cell lines were determined by western blot analysis, with membranes reprobed for actin and subjected to densitometric analysis as shown below. All of the HCC cell lines used in this study (Hep3B, SNU-387, SNU-182 and SNU-475) were found to express COX-2 proteins; the greatest staining intensity for COX-2 expression (corrected by β-actin) was observed in SNU-387 cells

Fig. 2

NS-398, a selective COX-2 inhibitor, reduces hepatocellular carcinoma cell growth. SNU-387 and Hep3B cells were treated with 50--200 μM NS-398 for 24--48 h and cell viability was measured by MTT assay. The results showed that NS-398-induced inhibition of cell proliferation depended on both concentration and time. The results represent the mean ± SD of six separate experiments

Fig. 3

NS-398, a selective COX-2 inhibitor, induces apoptosis in SNU-387 and Hep3B cells. Flow cytometric analysis using the FITC-conjugated monoclonal active caspase-3 antibody showed that SNU-387 and Hep3B cells underwent apoptosis after treatment with 100 and 200 μM NS-398 for 6--12 h. Apoptotic changes were noticed after 6 h treatment with NS-398 in both SNU-387 and Hep3B cells

Fig. 4

NS-398, a selective COX-2 inhibitor, reduces cell growth partially through inhibition of PGE2 production in SNU-387 cells. SNU-387 cells were treated for 24--48 h with 100 μM NS-398, and PGE2 production was measured by specific ELISA. a Treatment with 100 μM NS-398 for 24--48 h significantly reduced PGE2 production in SNU-387 cells. b Exogenous PGE2 (3 μg/ml) partially neutralized NS-398-dependent growth inhibition in SNU-387 cells (P<0.05). SNU-387 cells were cultured with 3 μg/ml PGE2 and 100 μM NS-398 simultaneously for 24 h, and cell viability was assessed by MTT assay. The data represent the means of four independent measurements

Fig. 5

NS-398, a selective COX-2 inhibitor, increases expression of COX-2 in human HCC cell lines. Hep3B and SNU-387 cells were treated with NS-398, and COX-2 expression was measured by western blot analysis after 24 h. Treatment with 50--200 μM NS-398 increased COX-2 expression compared with untreated controls as shown by densitometric analysis based on reprobing of membranes for β-actin

Fig. 6

Electroporation effectively transfected siRNA oligonucleotides into SNU-387 HCC cells. The transfection efficiency of electroporation in SNU-387 cells was tested by using a fluoresceinyl 3′-tagged non-silencing siRNA oligo control (5′-AATTCTCCGAACGTGTCACGT-3′). a SNU-387 cells treated with fluoresceinyl siRNA were observed by phase contrast microscopy (×100) b Cytoplasmic staining with fluoresceinyl siRNA was evaluated by fluorescent microscopy (×100). Over 90% of the transfected SNU-387 cells showed fluorescent staining in the cytoplasm, suggesting that electroporation efficiently transfected the siRNA oligos into SNU-387 HCC cells

Fig. 7

Transfection of COX-2-specific siRNAs by electroporation inhibits COX-2 expression in SNU-387 HCC cells. a Electroporation of 20 μM COX-2-specific siRNA 4 (nt 1774-1792, 5′-GGGCTGTCCCTTTACTTCA-3′) into SNU-387 cells significantly reduced COX-2 protein expression after 24 h and 48 h. b Densiometry analysis revealed that COX-2 protein levels were 64% lower in siRNA 4-transfected SNU-387 cells than in siRNA-free controls. Luciferase (Luc) siRNA and the other COX-2 siRNAs 1-3 did not reduce COX-2 expression in SNU-387 cells

Fig. 8

COX-2-specific siRNAs inhibit PGE2 production and cell growth in SNU-387 HCC cells. a Treatment with a 20 μM COX-2 siRNA 4 (nt 1774-1792, 5′-GGGCTGTCCCTTTACTTCA-3′) significantly reduced constitutive PGE2 production in SNU-387 HCC cells from 419±22.1 to 252.2±11.8 pg/ml after 24 h, and from 192.7±26.4 to 192.7±26.4 pg/ml after 48 h (P<0.05). The mean and SD of six independent experiments are shown. b MTT assay shows that transfection of COX-2 siRNA 4 significantly inhibited proliferation of SNU-387 cells by 17.3% (from 1.31±0.13 to 1.09±0.12) and 24% (1.80±0.18 to 1.37±0.15) after 24 h and 48 h, respectively, as compared with the siRNA-free controls (P<0.001). The inhibition of SNU-387 cell proliferation induced by siRNA was relatively small in comparison with that induced by NS-398. The mean ± SD of 12 separate experiments are shown