Knockdown of CMTM3 promotes peritoneal metastasis of gastric cancer in vivo

In vitro experiments have demonstrated that silencing CMTM3 promoted cell migration and invasion. To further confirm the effects of CMTM3 on tumor metastasis, an animal model of peritoneal metastasis was established by the intraperitoneal injection of SGC-7901-shN or SGC-7901-sh393 cells into mice. As shown in Figure ​Figure2A,2A, a fluorescence signal could be detected after two weeks in the sh393 group, while no fluorescence signal was detected in the control shN group. Although fluorescence was eventually detected in the shN mice, the fluorescence intensity was far weaker than that in the sh393 mice. The quantified fluorescence intensity was represented by bar graphs as shown in Figure ​Figure2B.2B. The body weight of the mice was monitored every four days from the time of injection. The changes revealed that knockdown of CMTM3 promoted metastasis, which further caused a decrease in the body weight of the mice (Figure ​(Figure2C).2C). All mice were sacrificed 4 weeks after injection. As shown in Figure ​Figure2D,2D, a higher number of metastatic nodules were found in the peritoneal cavity of the sh393 group compared with the shN group. Combined with the in vitro findings, we conclude that the knockdown of CMTM3 promotes cell migration, invasion and metastasis of gastric cancer cells.

Open in a separate window

Figure 2

Silencing CMTM3 promotes tumor metastasis in vivo

A total of 4×10⁶ cells (SGC-7901-shN or SGC-7901-sh393, n=8 per group) was injected into the abdominal cavity of each mouse on Day 0. A. fluorescence visualization of tumors from representative mice from each group was detected by an IVIS Spectrum Imaging System. B. fluorescence intensity was quantified and shown by the bar graphs. C. the changes in body weight were monitored every four days after injection. D. images of the peritoneal cavity were obtained after the mice were sacrificed. Arrows indicate the disseminated tumor nodules. Disseminated nodules larger than or equal to 1 mm in diameter were counted and were shown in the bar graphs. Data are presented as the mean±s.d (*P<0.05 **P<0.01).

Twist1 is involved in CMTM3-suppressed EMT in gastric cancer

Overexpression of CMTM3 in SGC-7901 and AGS cells inhibits cell migration and invasion [12], but the mechanism remains to be studied. In the current study, we further demonstrated that silencing CMTM3 promoted cell migration and invasion. Next, we aimed to investigate the effects of CMTM3 on EMT, which is important in the initiation and promotion of cell migratory and invasive properties [14]. First, the expression of EMT markers was analyzed in CMTM3 restored SGC-7901 cells and AGS cells. As shown in Figure ​Figure3A3A and ​and3B,3B, overexpression of CMTM3 increased the expression of E-cadherin but decreased the expression of N-cadherin and Vimentin. Among the EMT-related transcription factors, the expression of Zeb1 and Snail was not obviously affected. However, Twist1 expression was significantly downregulated. Further, we confirmed the effects of CMTM3 on EMT in CMTM3-knockdown SGC-7901 cells (Figure ​(Figure3C)3C) and GES-1 cells (Figure ​(Figure3D).3D). The expression of E-cadherin was downregulated whereas the expression of N-cadherin, Vimentin and Twist1 was upregulated, but no obvious differences were observed in the expression of Zeb1 and Snail.

Open in a separate window

Figure 3

Twist1 plays a vital role in the inhibition of EMT by CMTM3

A. expression of the EMT markers E-cadherin, N-cadherin, Vimentin, Zeb1, Snail and Twist1 was analyzed in CMTM3-restored SGC-7901 cells (MOI: 100). B. expression of EMT markers was analyzed in CMTM3-restored AGS cells (MOI: 100). C. expression of EMT markers was examined in shN and sh393 of SGC-7901 cells. D. expression of EMT markers was examined in shN and sh393 of GES-1 cells. E. expression of EMT markers was analyzed in tumors from mouse model of peritoneal metastasis (4 mice were randomly selected from the 8 mice in each group) by western blot. F. representative IHC staining for the EMT markers E-cadherin, N-cadherin, Vimentin, and Twist1 (top, shN group; bottom, sh393 group, 200 × magnification) was performed. The average relative gray density is shown under each blot (A-E).

Then, the expression of EMT markers was analyzed in tumor samples from in vivo peritoneal metastases model by western blot (Figure ​(Figure3E)3E) and IHC staining (Figure ​(Figure3F).3F). The expression of E-cadherin was clearly downregulated, while the expression of N-cadherin, Vimentin and Twist1 was upregulated. Together, these changes suggested that Twist1 is a crucial transcription factor in the inhibition of EMT by CMTM3.

Knockdown of CMTM3 promotes cell migration in a STAT3-dependent manner

To elucidate the mechanisms that underlie the inhibition of tumor metastasis and EMT by CMTM3, the Akt, Erk1/2 and STAT3 signaling pathways, which play important roles in cell migration, tumor metastasis and EMT, were analyzed. The levels of total and phosphorylated Akt (Ser473) and STAT3 (Tyr705) were detected in CMTM3 overexpressed SGC-7901 (Figure ​(Figure4A)4A) and AGS cells (Figure ​(Figure4B),4B), and phosphorylated Akt (Ser473), Erk1/2 (Thr202/Tyr204) and STAT3 (Tyr705) were detected in knockdown SGC-7901 (Figure ​(Figure4C)4C) and GES-1 cells (Figure ​(Figure4D).4D). No significant difference was observed in the level of p-Akt in these cells, but p-Erk1/2 was increased in CMTM3 knockdown cells, which was in accordance with our overexpression results [12]. Moreover, p-STAT3 was significantly decreased in CMTM3 restored SGC-7901 and AGS cells, and increased in CMTM3 knockdown SGC-7901 and GES-1 cells. It has been well studied that p- Janus-activated kinase 2 (Jak2) could activate STAT3, and the Jak2/STAT3 signaling pathway is involved in the development of cancers and EMT [25]. Thus, we detected the effect of CMTM3 on p-Jak2 and found that p-Jak2 was obviously decreased in CMTM3 restored cells and increased in CMTM3 knockdown cells (Figure ​(Figure4A4A–4D). Further, we confirmed the status of p-Akt, p-Erk1/2 and p-STAT3 in tumors from the peritoneal metastases mouse model by IHC. Similarly, no significant differences in p-Akt were found between the two groups, while higher levels of p-Erk1/2 and p-STAT3 were observed in sh393 mice (Figure ​(Figure4E).4E). To explore the roles of the Erk1/2 and STAT3 signaling pathways in cell migration after knockdown of CMTM3, the Erk1/2 inhibitor U0126 (10 μM) or the Jak2/STAT3 inhibitor JSI-124 (0.1 μM) was applied. The levels of p-Erk1/2 (Thr202/Tyr204) (Figure ​(Figure4F)4F) and p-STAT3 (Tyr705) (Figure ​(Figure4G)4G) were dramatically reduced by their respective inhibitors. According to the wound healing assay, treatment with the Erk1/2 inhibitor U0126 did not abrogate the tumor-promoting properties induced by knockdown of CMTM3 (Figure ​(Figure4H).4H). Transwell assay revealed that although U0126 could attenuate cell migration in both shN and sh393 cells, it did not reverse the cell migration promoted by silence of CMTM3 (Figure ​(Figure4I).4I). However, treatment with the Jak2/STAT3 inhibitor JSI-124 completely inhibited the migration of SGC-7901 cells promoted by knockdown of CMTM3 in both the wound healing assay (Figure ​(Figure4J)4J) and the Transwell assay (Figure ​(Figure4K).4K). Thus, we conclude that silencing CMTM3 promotes cell migration in a STAT3-dependent manner.

Open in a separate window

Figure 4

The STAT3 signaling pathway plays a crucial role in promoting migration of CMTM3-knockdown SGC-7901 cells

A-B. the phosphorylated and total Akt, STAT3 and Jak2 were analyzed in CMTM3 restored cells by western blot. C-D. the phosphorylated and total Akt, Erk1/2, STAT3 and Jak2 were analyzed in CMTM3 knockdown cells by western blot. E. representative IHC staining for p-Akt, p-Erk1/2 and p-STAT3 in tumors from mouse model of peritoneal metastasis (top, shN group; bottom, sh393 group, 200 × magnification) was performed. F. shN- and sh393-SGC-7901 cells were treated with the Erk1/2 inhibitor U0126 (10 μM) for 12 hours, and the level of p-Erk1/2 was analyzed by western blot. G. shN- and sh393-SGC-7901 cells were treated with the Jak2/STAT3 inhibitor JSI-124 (0.1 μM) for 24 hours, and the level of p-STAT3 was analyzed by western blot. H. wound-healing assay in shN and sh393 of SGC-7901 cells in the presence or absence of U0126 (10 μM); photographs were obtained at the indicated time points after the scratch was generated (100 × magnification). I. A Transwell assay of SGC-7901 cells in the presence or absence of U0126 (10 μM); photographs were obtained after 12 hours of incubation (100 × magnification). J. wound-healing assay in shN and sh393 of SGC-7901 cells in the presence or absence of JSI-124 (0.1 μM); photographs were obtained at the indicated time points after the scratch was generated (100 × magnification). K. A Transwell assay of SGC-7901 cells in the presence or absence of JSI-124 (0.1 μM); photographs were obtained after 12 hours of incubation (100 × magnification). Data are presented as the mean±s.d (*P<0.05 **P<0.01).

Western blot and antibodies

Western blotting was performed as previously described [11]. Antibodies to MMP2, Erk1/2 and phosphor-Erk1/2 (Thr202/Tyr204), Akt (pan) and phosphor-Akt (Ser473), STAT3 and phosphor-STAT3 (Tyr705), Jak2 and phosphor-Jak2 (Y1007/1008), N-cadherin, Vimentin, Snail (CST, MA, USA), E-cadherin (BD, USA), Twist1 (SAB, USA), Zeb1 (Santa Cruz Biotechnology) and Ki67 (Abcam, Cambridge, UK) were purchased commercially. The rabbit anti-CMTM3 antibody was prepared and purified in our laboratory [12].

Real-time PCR

Real-time PCR was performed as previously described [12].

Adenovirus construction and cell infection

The construction, generation, purification and infection of the ad5-null (vector-containing adenovirus, defined as Mock) and ad5-CMTM3 vectors were performed as previously described [12].

Lentivirus transduction

The pGIPZ-lentiviral shRNAmir vectors that target human CMTM3 and the nonsilencing pGIPZ control vector (shN) were synthesized by Open Biosystems (Thermo Fisher Scientific, Inc.). They each contain a Turbo green fluorescent protein (GFP) reporter and express a puromycin-resistant gene. The lentiviruses were packaged via the co-transduction of HEK293T cells with the lentiviral vector plasmid containing shRNAmir and the packaging plasmids psPAX2 and pLP/VSVG using Vigofect transfection reagent (Vigorous Biotechnology, China). Sixty hours after transfection, the supernatants that contained the lentivirus were collected. The target cells SGC-7901 and GES-1 were infected with the collected supernatants according to the manufacturer's protocol. A total of 400 ng/mL or 600 ng/mL of puromycin (Cellgro, USA) was added to the medium of SGC-7901 and GES-1 cells respectively to select the transfected cells. For a higher knockdown efficiency, the highest GFP-expressing cells were selected and obtained by cell sorting.

siRNA transfection

To knock down endogenous STAT3 and Twist1, the following target sequences were constructed in a small interfering RNA (siRNA) vector: si-STAT3-1^#, 5ʹ-GCAAGAUUCAGACCCUCAATT-3ʹ, si-STAT3-2^#, 5ʹ-CCAACAAUCCCAAGAAUGUTT-3ʹ; si-Twist1-1^#: 5ʹ-GCAAGAUUCAGACCCUCAATT-3ʹ, si-Twist1-2^#: 5ʹ-GGAGUCCGCAGUCUUACGATT-3ʹ. A scrambled siRNA (Scr) sequence was used as a negative control: 5ʹ-UUCUCCGAACGUGUCACGUTT-3ʹ. All these siRNAs and the positive control siRNA FAM were purchased commercially (Gene Pharma Inc., Shanghai, China). Cells were transfected with siRNA 24 hours later at a final concentration of 50 nM by Lipofectamine 3000 transfection reagent (Life Technologies, Carlsbad, CA) for further experiments. The suppression efficiency of si-STAT3 or si-Twist1 was analyzed by western blot at 48 or 72 hours, respectively.

Cell proliferation assay

Cell proliferation was measured by a Cell Counting Kit-8 (CCK8) detection kit (Dojindo Molecular Technologies, Japan). The cells were seeded at a concentration of 2000 cells per well in a 96-well plate. At the indicated time points, 10 μL CCK-8 solution was added to each well followed by an incubation at 37°C for 100 min. Absorbance values of all wells were then determined at 450 nm.

Plate colony formation assay

The cells were seeded into a 6-well plate at a density of 400 cells per well with 400 ng/mL puromycin. The selective medium was replaced every 4 days. After 10 days, the cell colonies were fixed in 4% paraformaldehyde/PBS and stained with 2% crystal violet. The colonies (≥50 cells per colony) were counted and imaged under a light microscope.

Soft-agar colony formation assay

The cells were seeded into a 12-well plate at a density of 500 cells per well and were suspended in RPMI-1640 with 0.3% agar and 10% FBS. The cells were layered in RPMI-1640 with 0.6% agar, 10% FBS and 400 ng/mL puromycin. After 2 weeks, the colonies were counted and imaged under a light microscope.

Wound-healing assay

Cells were seeded in a 6-well plate and grown to 80% confluence. A wound was scrapped with a sterilized pipette tip in the middle of the cell monolayer carefully. Photomicrographs were taken 48 hours after scrapping. The wound width in the microscopic pictures were measured by ImageJ software version 1.37 (National Institutes of Health, USA) at different time points. The percentage of wound healing was calculated from the initial wound width at 0 h.

Cell migration and invasion assay

Cell migration and invasion were analyzed in Transwell chambers (8-μm pore size, BD Biosciences, NJ, USA). The cells were cultured in serum-free RPMI-1640 medium overnight before the initiation of the experiments. The following day, 1×10⁵ (for the invasion assay) or 5×10⁴ (for the migration assay) cells in 0.25 mL serum-free RPMI-1640 medium were seeded into the upper chamber, which was pre-coated with (for the invasion assay) or without (for the migration assay) Matrigel. 0.5 mL RPMI-1640 with 10% FBS was added to the lower chamber. After incubation at 37°C, the chambers were disassembled, and the membranes were fixed in 4% paraformaldehyde/PBS for 10 min and stained with 2% crystal violet for 10 min. The number of cells was counted and images were obtained under a microscope (100 × magnification).

Gelatin zymography

Cells were incubated for 16 hours in serum-free medium. The supernatants were collected for Gelatin Zymography. 0.1% gelatin was used as a substrate. Following electrophoresis, the gels were washed with the renaturing buffer (2.5% Triton X-100) and then incubated for 24 hours at 37°C in the developing buffer (10 mM CaCl2, 15 mM NaCl, 2 mM NaN3 and 50 mM Tris-HCl, pH 7.5). The gels were stained with Coomassie Blue for 1 hour and destained at an appropriate time. Odyssey Infrared Imager (LI-COR Bioscience) was used to scan the gels.

Animal model of peritoneal metastasis

Five to six-week-old female BALB/c nude mice were purchased from Vital River (Beijing, China). The mice were bred and maintained under specific pathogen-free conditions, provided with sterilized food and water and housed in a barrier facility with a 12 h light/dark cycle. The mice were randomly divided into two groups and were injected with SGC-7901-shN (n=8) or -sh393 (n=8) cells that expressed TurboGFP. In all, 4×10⁶ cells were injected into the abdominal cavity of the mice along with 250μl PBS. The body weight of the mice was measured every 4 days. Tumor growth and metastasis were monitored in vivo via weekly measurement of the fluorescence intensity by the IVIS Spectrum Imaging System (PerkinElmer, USA). After all mice were sacrificed on day 28, the visible disseminated nodules were counted. The tumors were excised, cut into blocks, fixed in 10% formalin, and embedded in paraffin for IHC or were snap-frozen in liquid nitrogen for western blot.

Immunohistochemistry

Immunohistochemistry analysis was performed as previously described [11].

We thank Prof. Dalong Ma for his general advice. This work was supported by grants from the National Natural Science Foundation of China (81472626) and the Beijing Natural Science.