Molecular Mechanisms Of Metastasis-supressing Effect On Gastric Cancer By RUNX3

Despite the progress in diagnosis and treatment of gastric cancer, only about 20% of patients survive to 5 years. Metastasis, the most fearful aspect of cancer, is one of the major causes of mortality for gastric cancer patients. In order to improve the treatment of this disease, it is of great importance to clearly understand the molecular mechanism of gastric cancer metastasis. All the RUNX family members, namely, RUNX1, RUNX2 and RUNX3, encode DNA bindingαsubunits which form heterodimers with the commonβsubunit CBFβand act as transcription regulators. It has been realized that all three RUNX family members play important roles in normal developmental processes and carcinogenesis. They regulate the expression of cellular genes through binding to promoters, enhancer or silencer elements, with growing list of targes that including many genes relevant to carcinogenesis.Recently, RUNX3, one of Runt-related genes, is taken as a candidate tumor suppressor gene whose deficiency is causally related with gastric cancer. In RUNX3-knock-out mice, the stomach mucosal cells showed characteristics of malignant transformation. Reactivation of RUNX3 in gastric cancer cells inhibited their growth and tumorigenicity. There have emerged numerous reports on gene methylation and silencing of RUNX3 in tumors of many other tissues, such as liver, colon, gallbladder, lung, etc. Furthermore, decrease of RUNX3 protein expression was significantly associated with inferior survival duration of gastric cancer patients. It was reported RUNX3 expression was decreased in about 60% of the analyzed primary human gastric tumors, while the reduced RUNX3 levels rising to nearly 90% among the late stage, representing highly metastatic tumors. These data indicated that RUNX3 deficiency is closely related with gastric cancer metastasis.However, the function of RUNX3 in pathogenesis of gastric cancer metastasis remains not fully understood. So in the present work, we investigated whether RUNX3 might supress the invasive and metastatic abilities of gastric cancer cells and explored the relative molecular mechanism involved.【Objectives】(1) To investigate whether RUNX3 expression level is assocciated with invasion and metastasis of gastric cancer; (2) To examine the possible molecular mechanisms underlying the metastasis-suppressing effect by RUNX3 in gastric cancer; (3) To screen the downstream effectors of RUNX3【Methods】(1) The expression level of RUNX3 in metastatic gastric cancer and non-metastatic gastric cancer were determined by immunohistochemistry assay. (2) The RUNX3-specific siRNA vector was designed and constructed. (3) The plasmids pBK-RUNX3 (kindly provided by Prof. Paul J Farrell) or pSilencer-RUNX3 was transfected into SGC7901 and MNK28 cells using lipofectamine 2000 reagent. Empty vector of pBK-CMV or pSilencer were used as control. The transfected cells were screened with G418 for 2 months, single cell colonies were obtained by limited dilution method. Semi-quantitative RT-PCR and Western blot were used to detect the RUNX3 expression in the transfected cells for confirmation of transfection. (4) The effects of RUNX3 on the motility, invasive, angiogensis and in vivo metastatic abilities of gastric cancer cell lines SGC7901 and MKN28 were respectively investigated by wound-healing assays, invasion assay, tail vein metastatic assay andⅧfactor staining. (5) The activities of MMP2 and MMP9 in the conditioned medium were visualized by performing gelatin zymography (6) The mRNA and protein expression of MMP2, MMP9 and TIMP-1 were determined by RT-PCR and Western blotting. (7) The TIMP-1 and VEGF protein levels in the culture supernatants were determined using ELISA. (8) The TIMP-1 promoter was amplified by PCR using the genomic DNA from SGC7901 cells as template. The effect on TIMP-1 promoter activity by RUNX3 was evaluated by dual luciferase reporter assay. (9)The direct interaction between RUNX3 and the TIMP-1 promoter in vivo was determined by performing ChIP assay. (10) EMSA and supper shift assay were performed to visualize the binding of RUNX3 to TIMP-1 promoter via the two putative binding sites. (11) The downstream effectors of RUNX3 were screened cDNA microarray.【Results】1.RUNX3 expression is down-regulated in metastatic gastric cancer. To investigate the relationship between the expression of RUNX3 and metastasis of gastric cancer, we compared expression of RUNX3 in primary sites from 83 patients enduring non-metastatic gastric cancer with those from 40 patients enduring metastatic gastric cancer. The positive rate of RUNX3 expression in non-metastatic gastric cancer was 62.7% (52/83), lower than 42.5% (17/40) in metastatic gastric cancer. When considered staining scores, the average staining score in metastatic gastric cancer was significantly higher than that in non-metastatic gastric cancer (1.25±0.48 versus 0.57±0.26, P<0.05). These results suggested that RUNX3 expression level was negatively related to gastric cancer metastasis.2.Reduced RUNX3 expression is associated with increased metastatic potential of gastric cancer cells in vitro and in vivoRT-PCR and Western blot suggested that transfecting SGC7901 and MKN28 cells with pBK/RUNX3 increased the RUNX3 expression in the cells. And transfecting pSilencer-RUNX3 decreased the RUNX3 expression in the SGC7901 cells.Upregulation of RUNX3 in SGC7901 and MKN28 cells decreased their abilities to migrate on matrigel, and reduced RUNX3 in SGC7901 cells inceased its mobility. The RUNX3-transfected cells also showed significantly lower invasive potency than that of the control. The RUNX3 transfection reduced the number of invasive cells by 42.5% and 40.3% in SGC7901 cells and MKN28 cells, respectively. The difference in invasiveness was found to be statistically significant (P<0.05 in both cell lines). Similarly, inhibition of RUNX3 expression in SGC7901 cells by RUNX3 siRNA constructs promoted the cell invasive potency significantly compared with the control (P<0.05). MTT assay also showed that the survival rates of the detached SGC7901 and MKN28 cells transfected with pBK-RUNX3 were markedly decrease as compared with control group ( P < 0.05) . Survival rate of SGC7901 cells transfected with pSilencer-RUNX3 were significantly increase as compared with control (P < 0.05). The above results suggested that RUNX3 play a negative role in the gastric cancer cell resistant to anoilkis.RUNX3 inhibited angiogenesis in xenograft tumor in nude mice (p<0.05). We performed IHC analysis on various tumor sections to assess angiogenesis in mice. The MVD (miro-vessel density) of tumors in mice subcutaneously inoculated with gastric cancer cells trasfected with pBK-RUNX3 were lower than that trasfected with pBK-CMV (6.50±2.13 vs 17.30±8.32, 23.72±9.13 vs 50.25±15.46 respectively in SGC7901 and MKN28, P < 0.05 in both cell lines). And similarly, the MVD in tumor of mice subcutaneously inoculated with SGC7901/pSilencer-RUNX3 were higher than that of SGC7901/pSilencer (28.50±12.73 vs 12.83±4.92 P < 0.05). Tail vein metastatic assay in nude mice was further adopted to examine the suppressing effect on metastatic abilities of gastric cancer cells by RUNX3. Compared with control cells transfected with empty vectors, the injection of the cells transfected with pBK-RUNX3 via a tail vein of athymic nude mice led to significantly less visible tumors in the liver and lung surface (P< 0.05). Conformably, the injection of the SGC7901 cells transfected with pSilencer-RUNX3 led to much more visible tumors in the liver and lung surface than that of control cells (P<0.05). The results above suggest that RUNX3 could suppress the moltility, invasiveness, angiogenesis and metastatic potential of gastric cancer cells.3. RUNX3 regulates the expression level of downstream genes that are relevant to tumor metastasis.To study the possible role of MMPs in RUNX3-induced inhibition of cell invasion, we first analyzed the regulation on the expression and activity of MMP2 ,9 by RUNX3. The RT-PCR and Western blot were used to examine the mRNA and protein expression of MMP2 ,9 in the gastric cancer cells. We found no significant difference in MMP2 or MMP9 expression between cells with up-regulated RUNX3 (or down- regulated RUNX3) and their respective controlsAt the same time, the activities of MMP2, 9 in the serum-free conditioned medium were determined by gelatin zymography. MMP9 that exhibited collagen-degrading activity was detected in the culture supernatant of the SGC7901 and MKN28 cells. Compared with the control cells, the activity of the MMP9 enzyme in pBK-RUNX3 transfectants was significantly lower (both P<0.05 in SGC7901 and MKN28). And the activity of the MMP9 enzyme in pSilencer-RUNX3 transfectants was significantly higher than that of the control cells (P<0.05). While the activity of MMP2 is relatively low compared with that of MMP9 and showed no significant difference between cells with up-regulated RUNX3 (or down- regulated RUNX3) and their respective controls. Above results indicated that instead of regulating the transcription of MMP9, the enhanced expression of RUNX3 attenuate the matrix degrading activity of MMP9 in gastric cancer cells.TIMP1 expression is up-regulated by RUNX3 at transcriptional level. To examine the expression level of the TIMP1 which is considered the specific inhibitor of MMP-9 activity, we performed RT-PCR to detect the mRNA expression level and Western blot to detect the protein expression. We found that transfection of pBK-RUNX3 induced significantly stronger expression of TIMP-1 in SGC7901/pBK-RUNX3 and MKN28/pBK-RUNX3 cells. However, weaker expression of TIMP-1 was observed in cells transfected with pSilencer-RUNX3 as comparing with the control cells.The TIMP-1 levels secreted in the conditioned culture medium of gastric cancer cells were detected by performing ELISA We found that TIMP-1 production in the medium of the gastric cells transfected with pBK-RUNX3 was signifiantly higher than that of the respective control cells (16.5 ng/ml vs10.6 ng/ml in SGC7901, P<0.05; 14.9 ng/ml vs 7.2ng/ml in MKN28, P<0.05). Conformably, TIMP-1 level in the medium of SGC7901 /pSilencer -RUNX3 deceased compared with the control cells SGC7901/pSilencer (5.7 ng/ml vs 12.5 ng/ml, P<0.05).The VEGF levels secreted in the conditioned culture medium of gastric cancer cells were detected by performing ELISA. The VEGF production in the medium of the gastric cells transfected with pBK-RUNX3 was signifiantly lower than that of the respective control cells (1270±580 pg/millioncell vs 2950±870pg/millioncell in SGC7901, P<0.05; 2060±790 pg/millioncell vs 4300±1130 pg/millioncell in MKN28, P<0.05). VEGF level in the medium of SGC7901/pSilencer-RUNX3 increased when compared with the control cells SGC7901/pSilencer (6500±1730 pg/millioncell vs 3100±1066 pg/millioncell P<0.05).The total RNA of SGC7901/pBK-RUNX3 and SGC7901/pBK-CMV were extracted. The quality of extracted RNA was evaluated by agarose electrophoresis and UV absorbance spectroscopy. After microarray hybridization and data normalization, 14 genes were showed up-regulated while 109 genes were down-regulated by RUNX3. Among those genes, the expression level of CRKⅡ was evaluated by RT-PCR. It was confirmed that CRKⅡexpression was reduced by RUNX3, which indicated that inhibiting effect of RUNX3 on gastric cancer cell mobility was at least partially mediated by CRKⅡdownregulation.4. RUNX3 regulated TIMP-1 expression in gastric cancer cells by directly interacting with TIMP-1 promoterTo investigate the possibility that RUNX3 increased TIMP-1 expression by stimulating the TIMP-1 promoter activity, the dual luciferase reporter assay was performed. The transfection of different doses of pBK-RUNX3 plasmid led to a significantly increase of TIMP-1 promoter activity compared with the empty vector transfection. The result indicated that RUNX3 caused transactivation of TIMP-1 promoter and then up-regulated the mRNA and protein expression of TIMP-1.Further analysis of TIMP-1 revealed two putative RUNX binding sites was contained in the TIMP-1 promoter (-327 to -321; -66 to -61). Therefore, we tried to determine whether RUNX3 binds to the regions of the TIMP-1 promoter in vivo by performing ChIP assay. The chromatin fragments from SGC7901 cells and MKN28/ pBK-RUNX3 were immunoprecipitated with specific anti-RUNX3 antibody or with normal rabbit serum as a negative control. We found that both exogenously expressed RUNX3 in MKN28/ pBK-RUNX3 cells and endogenous RUNX3 in SGC7901 cells could be recruited to the two sites.Moreover, EMSA was conducted to demonstrate whether RUNX3 could bind to the two putative binding sites. The nuclear extracts of SGC7901 cells were prepared. Biotin-labeled probes including the RUNX3-binding site were incubated with the nuclear extracts, and DNA-protein complexes were analyzed with PAGE. For both sites, shifted bands of labeled probes were detected with nuclear extracts from SGC7901 cells. These bands were dramatically inhibited when competitive cold probes were included in the reaction, while the mutated cold probes could not inhibit the specific binding between labeled probes and nuclear protein. The shifted band was supershifted by addition of the anti-RUNX3 antibody. These observations suggest that the band was a specific complex formed by RUNX3 and the probes including the RUNX3-binding sites and RUNX3 did bind to the putative RUNX binding sites inTIMP-1gene promoter.【Conclusion】We provide evidence that RUNX3 expression was down-regulated in metastatic gastric cancer tissues than metastatic gastric cancer tissues. Our study presented evidence for RUNX3-mediated suppression of gastric cancer metastasis by inhibting the invasiveness, mobility and angiogenesis of cancer cells. We also provided novel molecular mechanism that at least partly contributes to the metastasis inhibiting activity of RUNX3. These data may be helpful for beneficial application of RUNX3 to diagnostics and therapeutics of gastric cancer metastasis.