DEAD-Box Protein RNA-Helicase DDX6 Regulates the Expression of HER2 and FGFR2 at the Post-Transcriptional Step in Gastric Cancer Cells

The human DEAD/H-box RNA helicase DDX6 (RCK/p54) is a protein encoded by the fusion gene from the t(11;14)(q23;q32) chromosomal translocation observed in human B-cell lymphoma cell line RC-K8. DDX6 has a variety of functions such as translation initiation, pre-mRNA splicing, and ribosome assembly. However, details of the regulatory mechanism governing DDX6 and the functions of DDX6 are largely unknown. Previously, we reported that DDX6 is overexpressed in most malignant cell lines and clinical colorectal tumor samples and that DDX6 positively contributes to the pathogenesis of various cancers. In the current study, we aimed at revealing the function of DDX6 in HER2 and FGFR2 related human gastric cancer (GC) by using clinical samples and GC cell lines. DDX6 protein was overexpressed in about 60% of the clinical samples; HER2, in 35%; and FGFR2, in 30%, (n = 20). Interestingly, the DDX6 protein was overexpressed in all HER2-positive samples (n = 7), and in 83% (5 of 6) of the FGFR2-positive samples, which could reflect the contribution of DDX6 to the expression of HER2 and FGFR2. In the GC cell line MKN7, which has HER2 amplification, the knockdown of DDX6 by siR-DDX6 led to the decreased expression of the HER2 protein. On the other hand, the knockdown of HER2 did not influence the DDX6 expression. Similar results were also obtained for the KATO-III and HSC39 cell lines having amplified FGFR2 expression. The increased expression of DDX6 induced a significantly increased expression of the HER2 protein without increasing the mRNA expression. The results of an RNP Immunoprecipitation (RIP)-assay using GC cells indicated that the DDX6 protein acted as an RNA-binding protein for HER2 and FGFR2 mRNAs and positively regulated their post-transcriptional processes. These findings demonstrated that DDX6 was an upstream molecule that positively regulated the expression of HER2 and FGFR2 at the post-transcriptional step in GC cells.

85 In order to elucidate the relationship between DDX6 and the expression of HER2 and FGFR2, we examined the cell viability of MKN7, MKN45, HSC39, and KATO-III cells and the expression of FGFR2 and HER2 in them after the knockdown of DDX6 by use of siRNA for DDX6 (siR-DDX6). The number of viable cells in all cell lines tested was significantly reduced at 72 h after post transfection, even at the concentration of 1 nM siR-DDX6 ( Figure 2A). Additionally, the knockdown of DDX6 led to decreased expression levels of HER2 and FGFR2 in these cells. These results indicated that DDX6 positively regulated the expression of HER2 and FGFR2 ( Figure 2B).

DDX6 Expression after the Knockdown of FGFR2 in HSC39 and KATO-III Cells
Next, we examined the growth of HSC39 and KATO-III cells and their expression of DDX6 at 72 h after FGFR2 silencing (siR-FGFR2), the cell viability for both cell types was significantly reduced to about 40-50% of the control ( Figure 3A). On the other hand, the knockdown of FGFR2 did not change the expression levels of DDX6 in HSC39 or KATO-III cells ( Figure 3B). These results suggested that DDX6 acted upstream of FGFR2 to regulate the FGFR2 expression.

DDX6 Expression after the Knockdown of FGFR2 in HSC39 and KATO-III Cells
Next, we examined the growth of HSC39 and KATO-III cells and their expression of DDX6 at 72 h after FGFR2 silencing (siR-FGFR2), the cell viability for both cell types was significantly reduced to about 40-50% of the control ( Figure 3A). On the other hand, the knockdown of FGFR2 did not change the expression levels of DDX6 in HSC39 or KATO-III cells ( Figure 3B). These results suggested that DDX6 acted upstream of FGFR2 to regulate the FGFR2 expression.

DDX6 Expression after the Knockdown of FGFR2 in HSC39 and KATO-III Cells
Next, we examined the growth of HSC39 and KATO-III cells and their expression of DDX6 at 72 h after FGFR2 silencing (siR-FGFR2), the cell viability for both cell types was significantly reduced to about 40-50% of the control ( Figure 3A). On the other hand, the knockdown of FGFR2 did not change the expression levels of DDX6 in HSC39 or KATO-III cells ( Figure 3B). These results suggested that DDX6 acted upstream of FGFR2 to regulate the FGFR2 expression.

DDX6 Expression after the Knockdown of HER2 in MKN7 and MKN45 Cells
We also examined whether knockdown of HER2 with siRNA for HER2 (siR-HER2) would influence the expression level of DDX6 in and viability of MKN7 and MKN45 cells. The numbers of viable cells remained almost unchanged compared with that of the control cells at 72 h after the knockdown of HER2 ( Figure 3C). Additionally, the knockdown of HER2 did not affect the expression levels of DDX6 in either cell type ( Figure 3D).
These results indicated that DDX6 also functioned upstream of HER2 to regulate the expression of HER2.
2.6. HER2 and FGFR2 mRNA Levels after Knockdown of DDX6 in MKN7, MKN45, HSC39, and KATO-III Cells Furthermore, we used real-time RT-PCR to examine the mRNA levels of HER2, FGFR2, and DDX6 in siR-DDX6-transfected MKN7, MKN45, HSC39, and KATO-III cells ( Figure 4). We confirmed that the expression level of DDX6 mRNA was extremely decreased in all siR-DDX6-treated cells ( Figure 4A-D). Interestingly, the levels of HER2 and FGFR2 mRNAs in DDX6-silenced cells were slightly increased in a concentration-dependent manner compared with those of the control, whereas the protein expression levels of both FGFR2 and HER2 were significantly down-regulated in KATO-III and HSC39 cells ( Figure 2B). Additionally, similar expression profiles were observed in the cases of MKN7 and MKN45 cell lines ( Figures 2B and 4C,D). These results suggested that DDX6 acted at a post-transcriptional step for the expression of HER2 and FGFR2.

DDX6 Expression after the Knockdown of HER2 in MKN7 and MKN45 Cells
We also examined whether knockdown of HER2 with siRNA for HER2 (siR-HER2) would influence the expression level of DDX6 in and viability of MKN7 and MKN45 cells. The numbers of viable cells remained almost unchanged compared with that of the control cells at 72 h after the knockdown of HER2 ( Figure 3C). Additionally, the knockdown of HER2 did not affect the expression levels of DDX6 in either cell type ( Figure 3D).
These results indicated that DDX6 also functioned upstream of HER2 to regulate the expression of HER2.

HER2 and FGFR2 mRNA Levels after Knockdown of DDX6 in MKN7, MKN45, HSC39, and KATO-III Cells
Furthermore, we used real-time RT-PCR to examine the mRNA levels of HER2, FGFR2, and DDX6 in siR-DDX6-transfected MKN7, MKN45, HSC39, and KATO-III cells (Figure 4). We confirmed that the expression level of DDX6 mRNA was extremely decreased in all siR-DDX6-treated cells ( Figure 4A-D). Interestingly, the levels of HER2 and FGFR2 mRNAs in DDX6-silenced cells were slightly increased in a concentration-dependent manner compared with those of the control, whereas the protein expression levels of both FGFR2 and HER2 were significantly down-regulated in KATO-III and HSC39 cells ( Figure 2B). Additionally, similar expression profiles were observed in the cases of MKN7 and MKN45 cell lines ( Figures 2B and 4C,D). These results suggested that DDX6 acted at a post-transcriptional step for the expression of HER2 and FGFR2.

RNA Immunoprecipitation (RIP)-Assay Using GC Cells
In order to validate the role of DDX6 in the translational step of HER2 and FGFR2 expression, we performed a RIP-assay. Its results clearly indicated that the DDX6 protein acted as an RNA-binding protein for HER2 and FGFR2 mRNAs and it positively regulated the translational process ( Figure 5).

RNA Immunoprecipitation (RIP)-Assay Using GC Cells
In order to validate the role of DDX6 in the translational step of HER2 and FGFR2 expression, we performed a RIP-assay. Its results clearly indicated that the DDX6 protein acted as an RNAbinding protein for HER2 and FGFR2 mRNAs and it positively regulated the translational process ( Figure 5).

Discussion
Previously, we reported that oncogenic RNA helicase DDX6 was highly overexpressed in most of the malignant cell lines and clinical colorectal tumors that we had tested [7,8]. We also showed that DDX6 expression is linked to the regulation of translation of c-myc mRNA through the internal ribosome entry site (IRES) [9]. In addition, the DDX6 protein is most likely to interact with the 5-capstructure binding protein eIF4E [6,10], which is a rate-limiting factor serving as a translation regulator for initiating translation. In GC, more than 50% of the clinical GC samples we examined showed the increased expression of DDX6 compared with the normal samples from the same patients. Most of the DDX6-overexpressed samples also showed HER2 and/or FGFR2 overexpression, as estimated by Western blot analysis. Similar findings were also made for the cell lines tested. Based on such findings, we speculated that DDX6 might be involved in the overexpression of HER2 and FGFR2 proteins, which is due to the gene amplifications observed in GC. Thus, we determined that DDX6 was positively associated with translational initiation for HER2 and FGFR2 mRNAs, validated by the results of the RIP-assay and by the experiments using siRNAs for DDX6, HER2, and FGFR2 ( Figures  4-6). We also found the feedback system of the DDX6-HER2 and DDX6-FGFR2 translational process because the down-regulation of HER2 and FGFR2 protein expression by silencing DDX6 induced the up-regulation of their mRNA levels. It is clear that DDX6 contributed to the maintenance of overexpression of HER2 and FGFR2 even in the presence of genetic aberrations. However, when we introduced the DDX6 expression vector into MKN74, which barely expressed HER2 and FGFR2, their protein expression levels were not increased (data not shown). DDX6 might work to efficiently express the amplified HER2 and FGFR2. It is speculated that DDX6 is necessary to maintain the higher protein levels of these amplified HER2 and FGFR2. RIP-assay showed that the DDX6 protein directly The results are presented as the mean ± SD. ** p < 0.01; *** p < 0.001.

Discussion
Previously, we reported that oncogenic RNA helicase DDX6 was highly overexpressed in most of the malignant cell lines and clinical colorectal tumors that we had tested [7,8]. We also showed that DDX6 expression is linked to the regulation of translation of c-myc mRNA through the internal ribosome entry site (IRES) [9]. In addition, the DDX6 protein is most likely to interact with the 5-cap-structure binding protein eIF4E [6,10], which is a rate-limiting factor serving as a translation regulator for initiating translation. In GC, more than 50% of the clinical GC samples we examined showed the increased expression of DDX6 compared with the normal samples from the same patients. Most of the DDX6-overexpressed samples also showed HER2 and/or FGFR2 overexpression, as estimated by Western blot analysis. Similar findings were also made for the cell lines tested. Based on such findings, we speculated that DDX6 might be involved in the overexpression of HER2 and FGFR2 proteins, which is due to the gene amplifications observed in GC. Thus, we determined that DDX6 was positively associated with translational initiation for HER2 and FGFR2 mRNAs, validated by the results of the RIP-assay and by the experiments using siRNAs for DDX6, HER2, and FGFR2 (Figures 4-6). We also found the feedback system of the DDX6-HER2 and DDX6-FGFR2 translational process because the down-regulation of HER2 and FGFR2 protein expression by silencing DDX6 induced the up-regulation of their mRNA levels. It is clear that DDX6 contributed to the maintenance of overexpression of HER2 and FGFR2 even in the presence of genetic aberrations. However, when we introduced the DDX6 expression vector into MKN74, which barely expressed HER2 and FGFR2, their protein expression levels were not increased (data not shown). DDX6 might work to efficiently express the amplified HER2 and FGFR2. It is speculated that DDX6 is necessary to maintain the higher protein levels of these amplified HER2 and FGFR2. RIP-assay showed that the DDX6 protein directly or indirectly influences the post-transcriptional step of HER2 and FGFR2. High-throughput sequencing of RNA isolated by crosslinking immunoprecipitation (HITS-CLIP) or photoactivatable-ribonucleoside-enhanced crosslinking and immunoprecipitation (PAR-CLIP) may directly indicate the RBP binding motif.

Clinical Samples
All clinical samples were obtained from patients who had undergone surgery for resection at the Department of Surgical Oncology, Gifu University Hospital (Gifu, Japan). Whether the tissue was tumorous or not was determined by the surgeon. All samples were immediately snap-frozen in liquid nitrogen and stored at −80 °C until RNA or protein extraction could be performed. All samples were histopathologically confirmed by H&E staining. The pathologic tumor staging was determined according to the Japanese Gastric Cancer Association (2011; see Table 1).
This study and collection and distribution of the samples were conducted in accordance with the principles of the 1975 Declaration of Helsinki after receiving approval from the Institutional Review Board of the Gifu University Graduate School of Medicine. (Approval number: 28-508 (23 March 2017)).

Transfection Experiments
MKN7, MKN45, HSC39, and KATO-III cells were seeded into 6-well plates at a concentration of 0.5 × 10 5 per well (10-30% confluence) on the day before the transfection. The sequence of siR-DDX6 was the following: 5′-GGAUAUUAUUCUCACGCUACCUAAA-3′. siR-HER2 and siR-FGFR2 were obtained from Invitrogen. They were used for the transfection of the cells, which was achieved by using cationic liposomes and Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA), according to the HER2 belongs to the EGFR family and plays a key role in cell survival and proliferation [11,12]. Amplification of the HER2 gene causes the overexpression of the HER2 protein, which causes cancer cell survival, growth, and proliferation through the activation of the PI3K-AKT and MAPK-ERK pathways [13,14]. HER2 overexpression is observed to range from 9 to 38% of study samples, depending on the location of the cancers and their histology [13,[15][16][17][18][19]. In our study, the frequency of overexpression of HER2 was almost 35%, being in the higher part of the range. The role of HER2 as a prognostic biomarker of GC is still controversial. Some studies have reported a positive correlation between HER2 overexpression and the clinical features of the tumor, such as tumor size, lymph node metastasis, local invasion, and cancer stage; whereas other studies insisted no relationship between them [20][21][22][23][24][25][26]. HER2 overexpression has become a very important biomarker because the ToGa clinical phase 3 trial indicated that HER2-positive patients can get survival benefits by treatment with trastuzumab [27].
The FGFR2 gene at human chromosome 10q26 encodes FGFR2b and FGFR2c isoforms that function as FGF receptors by separate expression domains and ligand specificity. Overexpression of FGFR2 was reported to be oncogenic and associated with chemoresistant in tumor cells. Single-nucleotide polymorphisms (SNPs) within intron 2 of the FGFR2 gene are associated with breast cancer through allelic FGFR2 upregulation [28]. A missense mutation or increased copy number of the FGFR2 gene is present in breast cancer and GC to activate FGFR2 signaling. It is reported that an FGFR2 copy number gain is detected in 9% of whole cancers. In GC, FGFR2 gene amplification ranges from 4 to 10% of samples; and FGFR2 protein is overexpressed in 30-40% of cases as judged by immunohistochemistry (IHC) [29,30]. When we estimated the expression of the FGFR2 protein by Western blotting, the rate of FGFR2 overexpression was almost 30%, similar to that observed in other studies. One important question is to ask what character of mRNA does DDX6 specifically affects in the translation initiation step. We considered that DDX6 might recognize the IRES structure of the 3' UTR of mRNAs, serving as an RNA-binding protein (RBP) at the IRES structure by binding motifs of DDX6, possibly its N-terminal motif Ia and Ib, and C-terminal IV and V. It is reported that the IRES requires an oligopyrimidine and AUG unit structure ( Figure 6A) [31]. We examined whether HER2 and FGFR2 have such an IRES structure by conducting a Centroidfold scan. Centroidfold is a web-available software that can predict and visualize RNA 2D structure (http://rtools.cbrc.jp/centroidfold/). As a result, we predicted that HER2 and FGFR2 might have the IRES structure ( Figure 6A). Our results indicated that DDX6 RNA helicase promoted the translation of FGFR2 and HER2 mRNAs, which are frequently mutated or over-expressed in GC cells ( Figure 6B), and functioned upstream of their expression pathways.

Clinical Samples
All clinical samples were obtained from patients who had undergone surgery for resection at the Department of Surgical Oncology, Gifu University Hospital (Gifu, Japan). Whether the tissue was tumorous or not was determined by the surgeon. All samples were immediately snap-frozen in liquid nitrogen and stored at −80 • C until RNA or protein extraction could be performed. All samples were histopathologically confirmed by H&E staining. The pathologic tumor staging was determined according to the Japanese Gastric Cancer Association (2011; see Table 1).
This study and collection and distribution of the samples were conducted in accordance with the principles of the 1975 Declaration of Helsinki after receiving approval from the Institutional Review Board of the Gifu University Graduate School of Medicine. (Approval number: 28-508 (23 March 2017)).

Transfection Experiments
MKN7, MKN45, HSC39, and KATO-III cells were seeded into 6-well plates at a concentration of 0.5 × 10 5 per well (10-30% confluence) on the day before the transfection. The sequence of siR-DDX6 was the following: 5 -GGAUAUUAUUCUCACGCUACCUAAA-3 . siR-HER2 and siR-FGFR2 were obtained from Invitrogen. They were used for the transfection of the cells, which was achieved by using cationic liposomes and Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's Lipofection protocol. The nonspecific control RNA (HSS, Hokkaido, Japan) sequence was: 5'-GUAGGAGUAGUGAAAGGCC-3', which was used as a control for nonspecific effects [35].
We used the web-available soft Image J in the case of calculating densitometric values.

RIP-Assay
The RIP-assay was performed by using a RIP-Assay kit (MBL, MEDICAL and BIOLOGICAL LABORATORIES CO., Nagoya, Japan). Cell lysates were derived from treated cells and reacted with beads labeled with the anti-DDX6 antibody. The target protein bound to the beads was then isolated, and the mRNAs binding the target protein were measured by qRT-PCR following the manufacturer's procedures.

Statistics
Each examination was performed in triplicate. Statistical significances of differences were evaluated by performing the two-sided Student's t-test. The values were presented as the mean ± standard deviation. A p-value < 0.05 was considered to be statistically significant.

Conclusions
Our results afforded the conclusion that DDX6 acted as an upstream molecule that positively regulated the expression of HER2 and FGFR2 at the post-transcriptional step in GC cells. Therefore, a DDX6 inhibitor such as a low-molecular agent and a siRNA targeting DDX6 would be promising drugs for HER2 and FGFR2-positive GCs.