miR-492 promotes chemoresistance to CDDP and metastasis by targeting inhibiting DNMT3B and induces stemness in gastric cancer

Abstract Objective: Metastasis and chemoresistance indicate treatment fail and progresses in gastric cancer (GC) patients. However, the molecular mechanisms of chemoresistance and metastasis remain unclear in GC. Thus, identifying the biological indicators of chemoresistance and metastasis is particularly important. Materials and methods: We establish a role for miR-492 in GC metastasis and chemoresistance through experiments in vitro and in vivo. Results: We identified miR-492 overexpression in GC specimens and cell lines, the miR-492 expression level was inversely correlated with the prognosis of GC patients. The inhibition of miR-492 suppressed GC cell invasion and enhanced the sensitivity of gastric cancer cells to CDDP treatment. In contrast, miR-492 overexpression significantly stimulated GC cell invasion and contributed to chemoresistance development. In addition, our research results indicated that the inhibition of miR-492 stimulates GC stemness, and the overexpression of miR-492 induces GC stemness. Importantly, our experiments demonstrated that miR-492 inhibitor suppressed tumor formation, and the combination treatment of miR-492 inhibitor and CDDP significantly inhibited tumor growth in vivo. Furthermore, we demonstrated that miR-492 exerts its anticancer role by targeting DNMT3B in GC. Conclusions: Our results suggested that inhibiting miR-492 is a novel strategy to control tumor metastasis and chemoresistance in GC.


Introduction
Gastric cancer (GC) is a common tumor of the digestive system 1  CSCs present high resistance to chemotherapy drugs that are commonly used in the treatment of GC, including cisplatin and 4,5 . Some studies showed that CSCs are able to regenerate all of the cell types in tumors because they have stem cell-like behavior that leads to metastatic recurrence 4,6 . Therefore, CSCs are important therapeutic targets in cancer. However, the mechanism of CSCs regulation in GC remains unclear.
Dysregulated expression of miRNAs was detected in most cancer types, and dysregulation of even a single miRNA can lead to tumorigenesis and stimulate cancer progression 7,8 . Additionally, dysregulated expression of miRNAs was demonstrated in CSCs, and such aberrantly regulated miRNAs participate in the development of CSCs and maintenance of stemness 9 . Previous studies showed that MiR-492 can regulates metastatic properties of hepatoblastoma via CD44, which also is a stemness protein markers. However, its function and mechanism in GC remain unclear. So, we want to study whether miR-492 will affect the progression of gastric cancer through stemness. Here, we describe the functional role of miR-492 as a tumor promoter in regulating metastasis, chemoresistance and CSCs. Additionally, we identified DNMT3B as a target of miR-492 in GC. We also found that miR-492 is significantly overexpression in GC and participates in stemness phenotypic via post-transcriptional regulation of DNMT3b.

CD133 Flow detection
Wash the transplanted tumor tissue in a sterilized dish, wash it with hanks buffer, and cut it into pieces of tissue about 3 mm 3 ; Add 3 ml serum-free 1640 medium to the culture dish, and then add the tissue fragments Piece. Then add 3 ml collagenase (0.1%); put the petri dish into the incubator for 3 h for digestion; blow the tissue fragments and pass through a 70 μm cell sieve to make a cell suspension; (Cell testing starts from this step) suspend the cells centrifuge the solution for 10 min at 1200 rpm, remove the supernatant, and wash once with pbs; Adjust the four groups of cells obtained to 1 × 10 6 per 100 μl and transfer to EP tubes. Add 1 μl FC block antibody to each tube and incubate for 30 min, wash once with PBS. Add 5 μg of CD133-FITC antibody (abcam, USA) to each tube, and protect from ice for 30 min in the dark; add 2 ml of cell staining buffer, centrifuge at 350 g for 5 min, remove the supernatant, and repeat twice; Resuspend the cells in 0.4 ml cell staining buffer, add 2 μl million cells of nuclear dye 7-AAD, and incubate on ice for 3-5 minutes; The cells were then analyzed using flow cytometry.

Transwell and osteosphere assays
Transwell and osteosphere assays were performed according to the description of Xu et al and Roscigno et al. 4,10 .

Luciferase reporter assay
The Luciferase reporter assays were performed according to the description of Roscigno et al (Roscigno et al., 2016).

DNA methylation analysis by pyrosequencing
DNA methylation analysis were performed as described by Roscigno et al. 10 . The primer sequence information in table 2.

Animal experiments
1×10 6 SGC7901 CDDP resistance cells in 100µl serum-free medium, which stably transfected with a series of miR-492 or negative control lentiviruses were constructed in our laboratory, were injected subcutaneously (s.c.) into per mouse (right back).
When the tumors reached ~50 mm 3 , which were intraperitoneally (i.p.) injected with CDDP (6 mg/kg). Following implantation, tumor volumes and body weight were

Statistical analysis
All statistical analysis using the GraphPad Prism 8 software. P< 0.05 was considered significant. Statistical significance was analysed by unpaired Student's t tests or one-way ANOVA and Duncan's multiple range tests. Kaplan-Meier survival analysis was used to calculate the overall survival rate of gastric cancer.

miR-492 expression was associated with poor clinical outcome
The data demonstrated that compared to normal gastric tissues miR-492 expression was significantly increased in GC specimens (Fig. 1A). Our data found that miR-492 was associated with clinical poor outcomes in GC patients (Fig. 1B). Consistent with these clinical data, the miR-492 expression was decreased in GC cell lines compared to the Human gastric mucosal cells GES-1 (Fig. 1C).

MiR-492 induce proliferation and metastasis in Gastric cancer cells.
The figure1 showed that miR-492 was associated with clinical poor outcomes in GC  2D). Together, our above data suggest that miR-492 promoted GC progression by inducing GC cell invasion and proliferation, suppressed the apoptosis.

miR-492 induces CSCs in GC
Because previous studies have shown that CSCs cause progression and metastasis in cancers, we investigated whether miR-492 is involved in CSCs regulation of GC. The western blot results found that the miR-492 overexpression significantly upregulated CSCs marker proteins expression, including CD133, Nanog, OCT-3/4 and BMI-1 in two cells (Fig. 3A, B). The inhibition of miR-492 expression suppressed these CSCs marker proteins expression (Fig. 3A, B). The flow cytometric assay showed that inhibiting the miR-492 expression could induce the expression of CD133, a stemness related protein. These data suggested that miR-492 exerts its tumor promoting effect partially due to the induce of CSCs in GC (Fig. 3C).

miR-492 target DNMT3B and suppress its expression
Further, to investigate how miR-492 regulates CSCs in GC, we searched for target gene candidates of miR-492 (mirdb.org) and identified DNMT3B as a candidate of miR-492 (Fig. 4A). DNMT3B is involved in cancer stemness maintenance and is closely associated with cancer proliferation and metastasis 10 . So, we chose to further study DNMT3B. To investigate whether miR-492 regulated DNMT3B, we examined DNMT3B expression levels by western blot. Our experiment results showed that DNMT3B expression was upregulated or downregulated by the inhibition or ectopic expression of miR-492, respectively, in GC cells at both the mRNA and protein levels ( Fig. 4B and C). the immunofluorescence also confirmed this result (Fig. 4D). we verified that miR-492 directly targeted the 3' UTR of DNMT3B by luciferase reporter assay (Fig. 4E). Furthermore, consistent with the cells results, the clinical sample analysis results also showed an inverse association between DNMT3B and miR-492 in GC specimens (Fig. 4F).

MiR-492 inhibitor induce the apoptosis and suppress the metastasis by target DNMT3B
We next transfected mimics specific for miR-492 inhibitor and siDNMT3B plasmid AGS and SGC-7901 cells. The CCK-8 assays showed that gastric cancer cells proliferation was decreased in the miR-492 inhibitor group, but it was restored in the siDNMT3B group compare to the miR-492 inhibitor group (Fig. 5A, B). The cell apoptosis rate was higher when miR-492 was suppressed, but lower when combine siDNMT3B (Fig. 5C). Consistent with the apoptosis analysis, knockdown DNMT3B could reverse the suppression of metastasis by miR-492 inhibitor (Fig. 5D). Together, our data suggest that miR-492 overexpression promoted cancer progression in gastric cancer.

miR-492 target DNMT3B modulates GC stemness
As we know, DNMT3b is a major regulator of Nanog and Oct 3/4 expression and inhibits their expression during embryogenesis through their methylation activity 10 .

Downregulated miR-492 reverse chemoresistance of CDDP via target DNMT3B in vitro
The CSCs closely related to cancer chemoresistance. Therefore, we investigated whether downregulated of miR-492 could promote cisplatin killing of gastric cancer.
We examined the miR-492 expression in both SGC7901/SGC7901 CDDP resistance and AGS/AGS CDDP resistance cell lines via RT-PCR, and the data showed that miR-492 overexpression in the CDDP resistance cell lines (Fig. 7A). And then, the CCK-8 proliferation assay of miR-492 inhibitors transfected in SGC7901CDDP and AGSCDDP cells, knockdown miR-492 could increase promote cisplatin killing of gastric cancer cisplatin resistance cells (Fig. 7B). Overexpression of DNMT3B by DNMT3B plasmid could promote cisplatin killing of gastric cancer cisplatin resistance cells (Fig. 7C), Overexpression of DNMT3B combine CDDP (5ug/ml) could promote apoptosis rate of gastric cancer cisplatin resistance cells, the apoptosis assay analysis by flow cytometric assay (Fig. 7D).

MiR-492 inhibitor significantly inhibits tumorigenesis and chemoresistance of CDDP in vivo
Then, we investigated the effects of miR-492 inhibitor on tumorigenesis and chemoresistance in vivo. As shown in Fig 8A and 8B, tumor volume and tumor weight were significantly decreased when miR-492 inhibitor combine with CDDP compare to the other groups. silencing of miR-492 expression combine CDDP could decrease the percent of CD133 high cells in xenograft ( Fig 8C). As expected, the expression levels of Ki67 were lower in the miR-492 mimics combine cisplatin treatment group compared to the control (Fig 8D). Taken together, the silence of miR-492 dramatically restored the resistance of GC cells to chemotherapy, and inhibits GC metastasis through suppressing GC stemness by targeting DNMT3B

Discussion
The chemoresistance and metastasis in GC patients indicates poor outcome 11 . We used a series of experiments to study the effects of miR-492 on metastasis and chemoresistance in GC. Here, we found that silence of miR-492 expression significantly correlated with good clinical outcomes. In addition, our study showed that miR-492 inhibitor suppressed GC cell invasion and chemoresistance in vitro.
Furthermore, the knockdown of miR-492 enhanced the chemosensitivity of GC cells to CDDP.
Next, we clarified the mechanism of miR-492 promote chemoresistance and metastasis. Accumulated evidence has shown that increased cancer stemness can stimulate cancer metastasis and induce chemoresistance [12][13][14] . Our data showed that miR-492 overexpression stimulated GC stemness. More importantly, our experiments showed that the silence of miR-492 significantly suppressed GC stemness populations in tumor tissues, suggesting that miR-492 plays an promote cancer role partially due to the induce of GC CSCs. In addition, we investigated the mechanism of miR-492 regulate CSCs in GC. We identify DNMT3B as a target gene of miR-492 in GC.
Previous studies have shown that DNMT3b is a regulator of Nanog and Oct 3/4 expression and inhibits their expression during embryogenesis through their methylation activity 15 . DNMT3B is a CSCs marker and involved in CSCs regulation.
Our data showed that DNMT3B expression was increased or decreased in GC cells by the inhibition or ectopic expression of miR-492, respectively. miR-492 directly targets