microRNA-9 suppresses the proliferation, invasion and metastasis of gastric cancer cells through targeting cyclin D1 and Ets1

Abstract

Recent evidence shows that altered microRNA-9 (miR-9) expression is implicated in the progression of gastric cancer. However, the exact roles and underlying mechanisms of miR-9 in the proliferation, invasion and metastasis of gastric cancer still remain unknown. In this study, miR-9 was found to be down-regulated and inversely correlated with the expression of cyclin D1 and v-ets erythroblastosis virus E26 oncogene homolog 1 (Ets1) in gastric cancer tissues and cell lines. Bioinformatics analysis revealed the putative miR-9 binding sites in the 3'-untranslated regions (3'-UTR) of cyclin D1 and Ets1 mRNA. Ectopic expression or knockdown of miR-9 resulted in responsively altered expression of cyclin D1, Ets1 and their downstream targets phosphorylated retinoblastoma and matrix metalloproteinase 9 in cultured gastric cancer cell lines SGC-7901 and AGS. In the luciferase reporter system, miR-9 directly targeted the 3'-UTR of cyclin D1 and Ets1, and these effects were abolished by mutating the miR-9 binding sites. Over-expression of miR-9 suppressed the proliferation, invasion, and metastasis of SGC-7901 and AGS cells in vitro and in vivo. Restoration of miR-9-mediated down-regulation of cyclin D1 and Ets1 by transient transfection, rescued the cancer cells from decrease in proliferation, migration and invasion. Furthermore, anti-miR-9 inhibitor promoted the proliferation, migration and invasion of gastric cancer cells, while knocking down of cyclin D1 or Ets1 partially phenocopied the effects of miR-9 over-expression. These data indicate that miR-9 suppresses the expression of cyclin D1 and Ets1 via the binding sites in their 3'-UTR, thus inhibiting the proliferation, invasion and metastasis of gastric cancer.

Figure 1. miR-9 was down-regulated and inversely correlated with the expression of cyclin D1 and Ets1 in gastric cancer tissues and cell lines.

A, B and C, real-time quantitative RT-PCR indicated that when compared to adjacent non-neoplastic mucosa (n = 86), down-regulation of miR-9 and higher transcript levels of cyclin D1 and Ets1 were detected in gastric cancer tissues (n = 86). D and E, there was an inverse correlation between miR-9 expression and transcript levels of cyclin D1 or Ets1 in gastric cancer tissues. F, lower miR-9 expression and higher cyclin D1 and Ets1 transcript levels were observed in gastric cancer cell lines (MKN-74, SGC-7901 and AGS), when compared to those in normal gastric epithelial GES-1 cells. G and H, lower miR-9 expression was observed in gastric cancer tissues with higher immunostaining of cyclin D1 or Ets1. The symbols (* and #) indicate a significant decrease and a significant increase from GES-1, respectively.

Figure 2. miR-9 suppressed the expression of cyclin D1 and Ets1 through post- transcriptional repression in gastric cancer cells.

A, scheme of the potential binding sites of miR-9 in the 3′-UTR of cyclin D1 and Ets1, locating at bases 2974–2995 and 2648–2670, respectively. B, stable transfection of miR-9 precursor resulted in decreased protein levels of cyclin D1, Ets1 and their downstream targets pRB and MMP-9 in gastric cancer SGC-7901 and AGS cells than those transfected with negative control vector (mock). C and D, the transcript levels of cyclin D1, Ets1 and MMP-9 were decreased in miR-9 precursor-transfected SGC-7901 and AGS cells, when compared to those in mock cells. E, transfection of anti-miR-9 inhibitor (100 nmol/L) resulted in increased protein levels of cyclin D1, Ets1 and their downstream targets pRB and MMP-9 in SGC-7901 and AGS cells than those transfected with negative control inhibitor (anti-NC, 100 nmol/L). F, the transcript levels of cyclin D1, Ets1, and MMP-9 were increased in SGC-7901 and AGS cells transfected with anti-miR-9 inhibitor (100 nmol/L), when compared to those transfected with anti-NC (100 nmol/L). The symbols (* and #) indicate a significant decrease and a significant increase from mock or anti-NC, respectively.

Figure 3. miR-9 directly interacted with the putative binding sites in the 3′-UTR of cyclin D1 and Ets1.

A, scheme and sequence of the intact miR-9 binding site (Wt) and its mutation (Mut) within the luciferase reporter vectors. B, stable transfection of miR-9 precursor into SGC-7901 and AGS cells resulted in decreased luciferase activities of 3′-UTR reporter of cyclin D1 and Ets1 than those transfected with negative control vector (mock), which were abolished by mutation in the putative miR-9 binding sites. C, transfection of anti-miR-9 inhibitor (100 nmol/L) into SGC-7901 and AGS cells increased the luciferase activities than those transfected with anti-NC (100 nmol/L), while mutation of miR-9 recognition site abolished these effects. The symbols (* and #) indicate a significant decrease and a significant increase from mock or anti-NC, respectively.

Figure 4. Ectopic expression of miR-9 abolished the in vitro proliferation of gastric cancer cells through targeting cyclin D1.

A and B, western blot and real-time quantitative RT-PCR indicated that transfection of cyclin D1, but not of Ets1, restored the down-regulation of cyclin D1 induced by miR-9 over-expression in gastric cancer SGC-7901 and AGS cells, when compared to those transfected with negative control vector (mock). C, in colony formation assay, miR-9 over-expression attenuated the growth of SGC-7901 and AGS cells than that of mock cells, which was rescued by transfection of cyclin D1, but not of Ets1. D, flow cytometry indicated that miR-9 over-expression induced the cell cycle arrest at G₀/G₁ phase in SGC-7901 and AGS cells than in mock cells, which was rescued by transfection of cyclin D1, but not of Ets1. The symbol (*) indicates a significant decrease from mock.

Figure 5. Over-expression of miR-9 attenuated the in vitro migration and invasion of gastric cancer cells through targeting Ets1.

A and B, western blot and real-time quantitative RT-PCR indicated that transfection of Ets1, but not of cyclin D1, restored the down-regulation of Ets1 induced by miR-9 over-expression in gastric cancer SGC-7901 and AGS cells, when compared to those transfected with negative control vector (mock). C, in transwell migration assay, gastric cancer cells stably transfected with miR-9 precursor presented an impaired migration capacity than that in mock cells, which was rescued by transfection of Ets1, but not of cyclin D1. D, in matrigel invasion assay, miR-9 over-expression attenuated the invasion capabilities of SGC-7901 and AGS cells than those of mock cells, which was rescued by transfection of Ets1, but not of cyclin D1. The symbol (*) indicates a significant decrease from mock.

Figure 6. miR-9 attenuated the growth and metastasis of gastric cancer cells in vivo.

A and B, hypodermic injection of SGC-7901 or AGS cells stably transfected with miR-9 precursor into athymic nude mice (n = 5) resulted in decreased size and weight of subcutaneous xenograft tumors than those transfected with negative control vector (mock, n = 5). C, hematoxylin and eosin (H&E) and immunohistochemical staining revealed that stable transfection of miR-9 precursor resulted in decreased expression of cyclin D1, Ets1, and MMP-9 within tumors. Scale bars: 100 µm. D, the CD31-positive mean vessel density was decreased within tumors formed by cancer cells stably transfected by miR-9 precursor. Scale bars: 100 µm. E, in the experimental metastasis assay, cancer cells stably transfected with miR-9 precursor established significantly fewer metastatic lung colonies (arrowheads) in athymic nude mice (n = 6). Scale bars: 100 µm. The symbol (*) indicates a significant decrease from mock.

Figure 7. Knockdown of cyclin D1 and Ets1 suppressed the proliferation, migration, and invasion of gastric cancer cells in vitro.

A, transfection of si-CCND1 (100 nmol/L) into gastric cancer SGC-7901 and AGS cells resulted in decreased expression of cyclin D1 and its downstream target pRB than those transfected with scramble siRNA (si-Scb, 100 nmol/L). B, knockdown of cyclin D1 suppressed the proliferation of SGC-7901 and AGS cells than those transfected with si-Scb. C, knockdown of cyclin D1 induced cell cycle arrest at G₀/G₁ phase than those transfected with si-Scb. D, transfection of si-Ets1 (100 nmol/L) into gastric cancer cells resulted in decreased expression of Ets1 and its downstream gene MMP-9 than those transfected with si-Scb (100 nmol/L). E, the migration capabilities of SGC-7901 and AGS cells were decreased by transfection of si-Ets1 than those transfected with si-Scb. F, the invasion capabilities of SGC-7901 and AGS cells were reduced by transfection of si-Ets1 than those transfected with si-Scb. The symbol (*) indicates a significant decrease from si-Scb.