Serum miR-658 induces metastasis of gastric cancer by activating PAX3-MET pathway: A population-based study

Abstract

This article has been retracted, and the online PDF replaced with this retraction notice.

1. Introduction

Gastric cancer is the third leading cause of cancer death in the world. The cancer of most metastatic cancers is largely treated by surgery associated with conventional chemotherapy, and the prognosis is poor [1]. Metastasis is the main reason for causing the death of gastric cancer patients and the research on the theory of gastric cancer metastasis and clinical examination has become very important in the diagnosis and treatment of gastric cancer.

MicroRNA (miR) has been reported to play an important role in tumor progression and metastasis [2, 3]. For instance, miR-429 has been shown to be down-regulated in gastric cancer and is directly targeted protein in gastric cancer progression [4]. MiR-1258 acts as a tumor suppressor to inhibit invasion and metastasis by targeting heparanase. Therefore, miR-1258 can be used as a new biomarker and therapeutic target for the treatment of gastric cancer [5]. MiR-658 is overexpressed in gastric cancer compared to adjacent normal tissues [6]. Furthermore, miR-658 may be associated with metastasis of gastric cancer (MGC) [7] but the molecular mechanism remains unknown.

PAX3 is a gene in the PAX family, a group of transcription factors consisting of proteins that bind to DNA sequences and control gene transcription. PAX3 is important in embryonic development. PAX3/FKHR fusion gene is usually associated with alveolar rhabdomyosarcoma, a cancer from striated muscle cells. The translocation between chromosomes 2 and 13 produces the fusion protein PAX3/FKHR as a tumor marker of rhabdomyosarcoma [8]. MiRNAs are thought to have an important role in tumor metastasis by regulatory diverse cellular pathways. The activation of PAX3-MET pathways promotes the progression of MGC [9]. MiR-658 may affect MGC by regulating PAX3-MET pathway and was investigated.

2. Materials and methods

2.1. Participants

Before all experiments, the protocols were approved by the Human Research Ethical Committee of China-Japan Union Hospital of Jilin University. The work was carried out based on the Declaration of Helsinki and signed consent form was taken from each person. From June 2016 to September 2016, ninety-eight gastric cancer patients with distant MGC (DM group) and ninety-six gastric cancer patients with no MGC (NM group) were recruited.

2.2. Inclusion criteria

All patients underwent preoperative diagnosis of advanced gastric cancer and can be treated by distal gastrectomy. The inclusion criteria included: 1) invasion of MP, SS or SE without involving other organs; 2) gastric cancer was determined by endoscopic and abdominal peritoneal computed tomography, and histological examination; 3) tumors were located in the stomach; 4) no history of gastrointestinal surgery; 5) the patients underwent distal subtotal gastrectomy.

2.3. Exclusion criteria

Exclusion criteria were provided as follows: 1) pregnant or breastfeeding women; 2) serious mental illness; 3) continuous systemic steroid therapy; 4) a history of myocardial infarction or unstable angina within 6 months; 5) cannot control the high blood pressure; 6) uncontrollable administration of diabetes or insulin; 7) Serious respiratory disease that requires sustained oxygen treatment.

2.4. Total RNA preparation

All the samples collected before the initial treatment. The blood sample was collected in a 10 cc tube with a polymer gel and a clot activator, allowed to stand at room temperature for 30–60 minutes, rotated at 3000 rpm for 10 min and equilibrated. Total RNA was extracted from samples by using Trizol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Total RNA was quantified by microfluidic analysis.

2.5. Oligonucleotide construction and lentivirus production

Gastric cancer cell lines of MGC80-3 were purchased from cell bank and were cultured in a DMEM medium (Sigma, St. Louis, MO, USA) with 10% fetal bovine serum (Hyclone, Logan, UT, USA), 100 $\mu$g/ml penicillin sodium, and 100 $\mu$g/ml streptomycin sulfate at 37${}^{\circ }$C with 5% CO${}_2$. The miR-658 mimics, miR-658 inhibitor and corresponding control oligonucleotides (available from RiboBo, Guangzhou, China) were transfected into gastric cancer cell lines of MGC80-3. The cells were grown for 48 hours and the lysates were tested by western blotting. For stable transfection, lentivirus-mediated miR-658 was constructed as previously reported [10].

2.6. Real-time PCR quantification

cDNA was synthesized from total RNA by using reverse transcription kit (Genecopoeia, MD, USA). The level of miR-658 was measured by using the primers, forward: 5’-TTGTGCTCGGTTGCCGTG-3’; reverse: 5’-GTGCAGGGTCCGAGGT-3’. the primers for PAX3 were 5’-TTACTCAAGGACGCGGTCTGT GATC-3’ (forward) and 5’-ATTGGCCCCAGCTTGC TT-3’ (reverse); MET primers were 5’-TCGTGCTCC TGTTTACCTTG-3’ (forward) and 5’-ACTGGCTGG GCTCTTC TATC-3’ (reverse). GAPDH was used as a control with forward primer 5’-AGCCTTCTCCATG GTGGTGAA-3’ (forward) and reverse primer 5’-AT CACCATCTTCCAGGAGCGA-3’. The relative-fold change was calculated by using 2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{Ct}}$ method.

2.7. Western blot analysis

The cells were washed twice with ice-cold PBS and placed on ice. Protein lysate was prepared by using the buffer with 20 mM Tris pH 7.5, 100 mM NaCl, 5 mM EDTA pH 8.0, 0.1% sodium azide, protease inhibitor cocktail (1:200, Sigma). The protein concentration was measured by a Protein Assay Kit (Pierce, Rockford, IL, USA). The samples were heated at 98${}^{\circ }$C for 10 min, and SDS-PAGE (sodium dodecyl sulphate-polyacrylamide gel electrophoresis) was used to separate the protein. The protein was transferred to the Polyvinylidene difluoride (PVDF) membrane (Millipore Corporation, Billerica MA USA) and the PVDF membrane was blocked with non-fat milk solution for one hour. The protein was then transferred to the solution with Anti-PAX3 antibody (ab180754), Anti-Met (c-Met) antibody (ab74217), and anti-GAPDH antibody (cat. no. ab9485) as a loading control from Abcam (Cambridge, MA, USA). After one-hour incubation at 22${}^{\circ }$C, the secondary antibodies Goat Anti-Rabbit IgG H&L (cat. no. ab7090, Abcam) were added. The protein levels were measured by enhanced chemiluminescence.

2.8. Cell migration test

Gastric cancer cell lines of MGC80-3 were cultured to log phase. After the cells were digested with trypsin, they were washed with serum-free culture solution and suspended in serum-free culture solution. The cell concentration was adjusted to 3 $\times$ 10${}^5$ cells/mL. Two-hundred-microliter cells were placed in the transwell chamber with 10% fetal bovine serum (FBS) culture solution to promote the cells passing through the small chamber with eight-micrometer-pore membrane. The number of cells running through membrane to reflects the ability of cells to migrate. The more cells passed through the membrane, the stronger for the ability of cell migration. Each experiment is repeated at least three times.

2.9. Cell invasion test

Transcell cells were preliminarily added with a layer of matrigel matrix, and the cells were digested with trypsin and washed with serum-free culture solution. The concentration of gastric cancer cell lines of MGC80-3 was adjusted to 5 $\times$ 10${}^5$ cells/mL. Two-hundred-microliter cells were placed in the transcell chamber to promote the cells digesting the matrix glue and running through the membrane with pores. The more cells passed through the membrane, the stronger for the ability of cell invasion. Each experiment is repeated at least three times.

2.10. Statistical analysis

All data are expressed as mean $\pm$ SD. Statistical comparisons of unpaired data and single factor analysis of variance were performed by using the Student’s t test. The relationship between two variables was performed by using Spearman’s rank correlation test. Statistical calculation was performed by using SPSS 20 (SPSS Inc., Chicago, IL, USA). There were significantly statistically differences if 0.001.

3. Results

3.1. Baseline clinical characters of gastric cancer patients

From April 2014 to March 2016, 1078 gastric cancer patients were recruited at our hospital. After inclusion and exclusion criteria, and other reasons, only one-hundred-and-ninety-four patients with gastric cancer (GC) completed the study (Fig. 1), and were supported by imaging and pathologic evidence. The patients were confirmed by the histological diagnosis of GC at initial diagnosis, and no one received treatment, including chemotherapy and radiotherapy. Tumor stages were classified according to the International Cancer Control Alliance (UICC) tumor lymph node metastasis (TNM) system. The control plasma samples were from individuals who had undergone routine physical examination and did not show disease signs. The two groups were matched by age and sex. A further investigation of MGC tissue and no MGC tissue ($>$ five-cm tumor) was performed among 194 patients. The detailed clinical features of the study participants are shown in Table 1.

Figure 1.

The flow-chart of the present study. DM, distant metastasis gastric cancer. NM, no metastasis cancer.

Table 1.

Baseline characters of participants

	NM	DM	Chi-square statistic/$t$-value	$P$ values
Gender (Male/Female)	98 (67/31)	96 (66/30)	0.003	0.954
Age (Years)	61.5 $\pm$ 12.3	59.7 $\pm$ 11.2	0.136	0.738
SBP (mm Hg)	121.2 $\pm$ 14.6	124.7 $\pm$ 16.1	0.329	0.476
DBP (mm Hg)	85.3 $\pm$ 6.8	87.5 $\pm$ 9.8	0.3032	0.512
BMI	25.3 $\pm$ 3.8	25.6 $\pm$ 4.2	0.056	0.854
TC (mmol/L)	5.2 $\pm$ 0.8	5.4 $\pm$ 0.6	0.114	0.758
TG (mmol/L)	2.5 $\pm$ 0.7	2.4 $\pm$ 0.9	0.164	0.685
LDL-C (mmol/L)	2.1 $\pm$ 0.5	2.3 $\pm$ 0.6	0.277	0.598
HDL-C (mmol/L)	1.9 $\pm$ 0.4	1.7 $\pm$ 0.5	0.204	0.673
TNM stages
I, cases (%)	19 (19.4)	0 (0)	–	–
II, cases (%)	40 (40.8)	0 (0)	–	–
III, cases (%)	39 (39.8)	0 (0)	–	–
IV, cases (%)	0 (0)	96 (100)	–	–
Metastasis
Liver, cases (%)	0 (0)	58 (60.4)	–	–
Lung, cases (%)	0 (0)	30 (31.3)	–	–
Bone, cases (%)	0 (0)	8 (8.3)	–	–

Note: SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; TG, triglyceride; TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TNM, tumor, node and metastasis.

3.2. The levels of serum miR-658, PAX3 and MET is associated with gastric cancer development

qRT-PCR analysis showed that the mRNA level of serum miR-658 was lower in NM group than in DM group ( 0.001, Fig. 2A). Meanwhile, the mRNA levels of PAX3 ( 0.001, Fig. 2B) and MET ( 0.001, Fig. 2C) were also lower in NM group than DG group. Western blot analysis showed that the protein levels of PAX3 ( 0.001, Fig. 2B) and MET ( 0.001, Fig. 2C) were lower in NM group than DG group. Western Blot analysis showed that the protein levels of PAX3 (Fig. 2D) and MET (Fig. 2E) were higher in MGC than in NM group ( 0.001). The results suggest that the levels of serum miR-658, PAX3 and MET is associated with gastric cancer development.

Figure 2.

The levels of miR-658, PAX3 and MET in different groups. A: relative mRNA levels of miR-658 in different groups. B: relative mRNA levels of PAX3 in different groups. C: relative mRNA levels of MET in different groups. D: relative protein levels of PAX3 in different groups. E: relative protein levels of MET in different groups. NM, the patients with gastric cancers. MGC, the patients with metastatic gastric cancers. There is significantly statistical difference if 0.001.

3.3. The levels of miR-658 affect the levels of PAX3 and MET

To explore the effects of miR-658 on the levels of PAX3 and MET, miR-658 was overexpressed or blocked by using its mimic or inhibitor at a cell level. The results showed miR-658 was successfully overexpressed or silenced when compared with a control group (Fig. 3A). On the other hand the mRNA levels of PAX3 ( 0.001, Fig. 3B) and MET ( 0.001, Fig. 3C) are also higher in miR-658 overexpressed group or lower in miR-658 silence group when compared with a control group. Western blot analysis shows that the protein levels of PAX3 ( 0.001, Fig. 3D) and MET ( 0.001, Fig. 3E) are also higher in miR-658 overexpressed group or lower in miR-658 silence group when compared with a NM group. The results suggest that the levels of miR-658 affect the levels of PAX3 and MET.

Figure 3.

The effects of miR-658 on the levels of PAX3 and MET. A: relative mRNA levels of miR-658. B: relative mRNA levels of PAX3. C: relative mRNA levels of MET in different groups. D: relative protein levels of PAX3. E: relative protein levels of MET. Control, gastric cancer cell lines of MGC80-3; miR-658, MGC80-3 cells were transfected with miR-658 mimic; Anti-miR-658, gastric cancer cell lines of MGC80-3 was transfected with miR-658 inhibitor. There is significant difference if 0.001.

3.4. The correlation between serum miR-658 and PAX3 and MET

Spearman’s rank correlation coefficient test shows that there is strong positive relation between serum level of miR-658 and mRNA PAX3 (Fig. 4A), MET (Fig. 4B), and protein levels of PAX3 (Fig. 4C) and MET (Fig. 4D). With the increase of miR-658, the levels of PAX3 and MET are also increased.

Figure 4.

The relationship between miR-658 and PAX3 or MET. A: the relationship for the serum levels between miR-658 and PAX3. B: the relationship for the serum levels between miR-658 and MET. Statistical analysis was performed by using Spearman’s rank correlation test. The value falls between 0.5 and 1, there is a strong positive correlation. There is significantly statistical difference if 0.001.

3.5. Migration test

The overexpression of miR-658 significantly promoted cell migration (Fig. 5A). In contrast, the silence of miR-658 significantly reduced cell migration (Fig. 5B). The results suggest that miR-658 expression affect cell migration.

Figure 5.

The effects of miR-658 on the migration of gastric cancer cells. The overexpression of miR-658 significantly promoted cell migration. In contrast, the silence of miR-658 significantly reduced cell migration.

3.6. Invasion test

The overexpression of miR-658 significantly promoted cell migration (Fig. 6A). In contrast, the silence of miR-658 significantly reduced cell migration (Fig. 6B). The results suggest that miR-658 expression affect cell migration.

Figure 6.

The effects of miR-658 on the invasion of gastric cancer cells. The overexpression of miR-658 significantly promoted cell invasion. In contrast, the silence of miR-658 significantly reduced cell invasion.

4. Discussion

To study cancer metastasis, it is important to establish an ideal model. In this study, we used repeated transwell methods to isolate invasive and non-invasive subpopulations from established human NM cell lines and have been successfully applied to explore the metastasis of gastric cells. In vitro showed that the established cell lines had significant invasive and metastatic abilities. Meanwhile, the overexpression of miR-658 promotes the migration (Fig. 5) and invasion (Fig. 6) of gastric cancer cells. Thus, the level of miR-658 may be related to metastases.

Recently, miRNA has been reported to promote or inhibit tumor metastasis [5, 11], providing a new perspective for exploring the mechanism of cancer metastasis. Nevertheless, the role of miRNAs in gastric cancer metastases remains widely unclear. Here, we obtained metastatic-associated miRNAs based on a well-established metastatic cell model. The up-regulation of miR-658 in metastatic indicates that miR-658 increase may be a common event of tumor occurrence. In this study, we focused on the effect of miR-658 on NM transfer and demonstrated that miR-658 acts as a tumor promoter for NM metastases. We demonstrate that PAX3 and MET are key downstream targets for miR-658. It is known that PAX3 triggers neoplastic development by maintaining cells in a undifferentiated and proliferative state [12]. However, previous work shows that PAX3 is a tumor suppressor in thyroid cancer by activating PI3K/Akt and MAPK pathways [13]. In the present study, we find that PAX3 is usually expressed at high levels in the patients with metastatic gastric cancer. The molecular mechanism for causing different functional roles of PAX3 in different cancers remains unclear.

MET gene is located on chromosome 7q31.2. Evaluated level of MET in primary uveal melanoma is related to metastatic development, and MET is released as a soluble ectodomain via a disintegrin and metalloprotease (ADAM) 10- and ADAM17 cleaving. MET may be a potential biomarker for diagnosing metastatic cancer [14]. In the similar case, our work also shows that MET is highly expressed in the patients with metastatic gastric cancers (Fig. 2). On the other hand, PAX3 mediates MET induction [15]. PAX3/MET signaling pathway plays an important role in metastatic gastric cancer [16]. Present findings demonstrate that overexpression of miR-658 increases the levels of PAX3/MET and silence of miR-658 reduces the levels of PAX3/MET (Fig. 3). There is a positively relationship between the level of miR-658 and PAX3/MET (Fig. 4). All these results suggest that the levels of miR-658 affecting PAX3/MET pathway.

In sum, our results show that the level of miR-658 is also positively associated with the levels of PAX3 and MET, all of which are related with gastric cancer progression. MiR-658 mimic and silence demonstrate that high-level miR-658 promotes the migration and invasion of gastric cells. MiR-658 is a potential biomarker as the predictor of gastric cancer with metastasis by activating PAX3/MET signaling pathway. To confirm that, further research is needed in the future.