Oncotarget

Research Papers:

Oncogenic roles of the SETDB2 histone methyltransferase in gastric cancer

PDF |  HTML  |  Supplementary Files  |  How to cite  |  Order a Reprint

Oncotarget. 2016; 7:67251-67265. https://doi.org/10.18632/oncotarget.11625

Metrics: PDF 836 views  |   HTML 1424 views  |   ?  

Taketo Nishikawaji, Yoshimitsu Akiyama, Shu Shimada, Kazuyuki Kojima, Tatsuyuki Kawano, Yoshinobu Eishi, Yasuhito Yuasa and Shinji Tanaka _

Abstract

Taketo Nishikawaji1, Yoshimitsu Akiyama1, Shu Shimada1, Kazuyuki Kojima2, Tatsuyuki Kawano3, Yoshinobu Eishi4, Yasuhito Yuasa1, Shinji Tanaka1

1Department of Molecular Oncology, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan

2Department of Surgical Oncology, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan

3Department of Surgery, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan

4Department of Human Pathology, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan

Correspondence to:

Shinji Tanaka, email: tanaka.monc@tmd.ac.jp

Yoshimitsu Akiyama, email: yakiyama.monc@tmd.ac.jp

Keywords: gastric cancer, histone methyltransferase, H3K9me3, SETDB2

Received: March 18, 2016    Accepted: August 11, 2016    Published: August 26, 2016

ABSTRACT

SETDB2 is a histone H3 lysine 9 (H3K9) tri-methyltransferase that is involved in transcriptional gene silencing. Since it is still unknown whether SETDB2 is linked to carcinogenesis, we studied alterations and functions of SETDB2 in human gastric cancers (GCs). SETDB2 protein was highly expressed in 30 of 72 (41.7%) primary GC tissues compared with their normal counterparts by immunohistochemistry. SETDB2 overexpression was significantly associated with the late stage of GCs (P<0.05) and poor prognosis of GC patients (P<0.05). The GC cell lines with SETDB2 knockdown and overexpression significantly decreased and increased cell proliferation, migration and invasion, respectively (P<0.05). Knockdown of SETDB2 in MKN74 and MKN45 cells reduced global H3K9 tri-methylation (me3) levels. Microarray analysis indicated that expression of WWOX and CADM1, tumor suppressor genes, was significantly enhanced in MKN74 cells after SETDB2 knockdown. Chromatin immunoprecipitation assays showed that the H3K9me3 levels at the promoter regions of these two genes corresponded to the SETDB2 expression levels in GC cells. Moreover, ectopic SETDB2 protein was recruited to their promoter regions. Our data suggest that SETDB2 is associated with transcriptional repression of WWOX and CADM1, and hence overexpression of SETDB2 may contribute to GC progression.


Oncogenic roles of the SETDB2 histone methyltransferase in gastric cancer | Nishikawaji | Oncotarget

INTRODUCTION

Gastric cancer (GC) is the third leading cause of death from cancer in the world, and the incidence is still high [1]. Genetic and epigenetic alterations are involved in gastric carcinogenesis; for example, mutations of the CDH1 and p53 tumor suppressor genes (TSGs), and hypermethylation at the CpG island promoter region of TSGs have been detected in GCs [2, 3].

Tri-methylation mark of histone H3 lysine 4 (H3K4me3) is involved in transcriptional activation of genes, whereas tri-methylation marks of H3K9 (H3K9me3) and K27 (H3K27me3) are associated with gene silencing in the transcriptional process. These histone methylations are catalyzed by histone methyltranseferases (HMTs) containing an evolutionally served SET (SuVar3-9, enhancer of Zeste, Trithorax) domain [4]. Among them, SETDB1, SUV39H1 and EZH2 are reported to be frequently over-expressed in various cancers, such as GCs, those of which functions are shown as oncogenic activities [58]. We have recently reported that expression of SET7/9, a histone H3K4 mono-methyltransferase, was reduced in advanced GCs [9]. Thus, expression changes of histone modification genes are known to play an important role in tumor development.

SETDB2 have been characterized as H3K9 tri-methyltrasferases that contain a bifurcated SET domain, a pre-SET domain and a Methyl-CpG binding domain [10, 11]. Setdb2 restricts dorsal organizer territory and regulates structural left-right asymmetry of visceral organs and nervous system by suppressing the activity of fibroblast growth factor 8 in zebrafish embryo [10]. Moreover, it was reported that knockdown of Setdb2 in zebrafish embryo induced abnormal convergence and extension movements during gastrulation, resulting in anterior-posterior shortening [12], indicating an essential role in embryonic development. Although SETDB2 is known to be associated with chromosomal condensation and segregation through H3K9 tri-methylation (me3) regulation during mitosis [11], the molecular mechanism by which SETDB2 mediates transcriptional gene silencing remains unclear. Furthermore, there has not been reported the evidence presented of SETDB2 alterations in cancers. In this study, we aimed to clarify whether SETDB2 is linked to carcinogenesis. Since we found that SETDB2 was often overexpressed in both primary GC tissues and GC cell lines, we further analyzed its biological functions in GC cells.

RESULTS

Overexpression of SETDB2 in primary GC tissues

We analyzed the expression of SETDB2 protein in a set of 72 primary GC tissues by immunohistochemical staining (IHC). SETDB2 expression was not detected in any non-cancerous stomach tissues, whereas SETDB2 was strongly expressed in the nuclei of 30 GC tissues (41.7%, Figure 1A). Western blot analysis showed that SETDB2 protein was expressed in four of 12 primary GC tissues (33.3 %) but not in any of the non-cancerous stomach tissues tested (Figure 1B). To evaluate the significance of SETDB2 alteration in primary GCs, we studied the relationship between SETDB2 overexpression and clinicopathological factors of GCs. SETDB2 was significantly overexpressed in advanced GC tissues compared with early GC tissues (P<0.05, Table 1). A significant difference was also found between SETDB2 expression and pTNM stages (Stage I vs. II-IV, P<0.05). However, there were no significant correlations between SETDB2 expression and age, gender, histological classification or lymph node metastasis. Furthermore, SETDB2 expression was significantly associated with worse survival by Kaplan-Meier analysis (P<0.05, Figure 1C). Thus, SETDB2 overexpression might be potentially associated with clinical progression of GCs.

Analyses of SETDB2 protein expression in primary GC tissues.

Figure 1: Analyses of SETDB2 protein expression in primary GC tissues. A. Immunohistochemical staining of SETDB2 protein expression. The staining is representative of 72 primary GC tissues. The images at bottom are magnifications of the boxed regions of non-cancerous (a) and tumor (b) tissues in the top image. B. Western blot analysis of the expression levels of SETDB2 protein in primary GC tissues and non-cancerous stomach tissues. Representative data are shown. Actin expression was analyzed as an internal control. Expression in the MKN74 cell lines was used as a positive control (Figure 2A). N; non-cancerous stomach tissue, C; human gastric cancer. C. Kaplan Meier analysis of overall survival of gastric cancer patients according to SETDB2 positive or negative expression in the tumor tissue (*: P<0.05).

Table 1: Clinicopathological correlations of SETDB2 immunohistochemistry in 72 GC tissues

Clinicopathol group

No. of cases

SETDB2

P-value1)

Positive
n=30 (%)

Negative
n=42 (%)

Age

 ≤ 60

28

12 (42.9)

16 (57.1)

0.870

 > 60

44

18 (40.9)

26 (59.1)

Gender

 Female

21

7 (33.3)

14 (66.7)

0.350

 Male

51

23 (45.1)

28 (54.9)

Lauren’s classification

 Intestinal

36

17 (47.2)

19 (52.7)

0.339

 Diffuse

36

13 (36.1)

23 (63.9)

Tumor invasion

 EGC (pT1)

12

1 (8.3)

11 (91.7)

0.010*

 AGC (pT2-4)

60

29 (48.3)

31 (51.7)

pT stage

 pT1a

2

0 (0)

2 (100.0)

0.157

 pT1b

10

1 (10.0)

9 (90.0)

 pT2

8

3 (37.5)

5 (62.5)

 pT3

24

13 (54.2)

11 (45.8)

 pT4a

25

11 (44.0)

14 (56.0)

 pT4b

3

2 (66.7)

1 (33.3)

Lymph node metastasis

 Absent (N0)

30

10 (33.3)

20 (66.7)

0.187

 Present

42

20 (47.6)

22 (52.4)

pTNM stage

 I

17

3 (17.6)

14 (82.4)

0.260

 II

23

11 (47.8)

12 (52.2)

 III

22

10 (45.5)

12 (54.5)

 IV

10

6 (60.0)

4(40.0)

pTNM stage

 I

17

3 (17.6)

14 (82.4)

0.026*

 II+III+IV

55

27 (49.1)

28 (50.9)

 I

17

3 (17.6)

14 (82.4)

0.044*

 II+III

45

21 (46.7)

24 (53.3)

1) P-values were determined by Pearson’s chi-square test. The Spearman rank correlation analysis was performed for pT and pTMN stages.

*Statistically significant difference.

Analyses of SETDB2 expression in GC cell lines

We next examined SETDB2 protein expression levels in 13 GC cell lines using Western blot analysis. By densitometric quantification of the SETDB2 bands, we subsequently classified these cell lines into SETDB2 high- (HSC57, MKN7, MKN45 and MKN74) and low- (AGS, GCIY, TGBC11TKB, HSC39, HSC44PE, HSC59 NUGC3, NUGC4 and KATOIII) expression groups, which were determined by a cut-off value of 0.50 (Figure 2A). To evaluate the biological significance of SETDB2 expression, the effect of knockdown of SETDB2 expressing using RNA-interference was analyzed. SETDB2 protein expression was efficiently suppressed by transfection of SETDB2 siRNAs (siSETDB2) into MKN74 and MKN45 cells compared with its expression in negative control siRNA-transfected cells (Figure 2B). These two cell lines in which SETDB2 was knocked down showed significant inhibition of cell proliferation compared with the cells transfected with the negative control siRNA (P<0.05, Figure 2C). Moreover, knockdown of SETDB2 markedly decreased the cell migration and invasion rates of MKN45 cells (P<0.05, Figure 2D and 2E). In addition, we analyzed the effect of SETDB2 overexpression in AGS and NUGC3 cells by transfection with the SETDB2/pcDNA3 expression vector (Figure 2F). Cell growth, migration and invasion were enhanced in the GC cells with SETDB2 overexpression compared with the GC cells transfected with empty vector (P<0.05, Figure 2G-2I). Scratch assay also demonstrated that cell migration was significantly faster in AGS and NUGC3 cells with SETDB2 overexpression than those with empty vector transfection (P<0.05, Supplementary Figure S1A).

Functional analyses of SETDB2 in cultured GC cell lines.

Figure 2: Functional analyses of SETDB2 in cultured GC cell lines. A. Western blot analysis of the expression levels of SETDB2 protein in 13 GC cell lines. Actin expression was analyzed as an internal control. SETDB2 protein expression was quantified using Image J software, and then was normalized to Actin expression. GC cell lines were classified into SETDB2-high and -low expressing groups, which were determined by a cut-off value of 0.50. Accordingly, four GC cell lines highlighted in blue (high: HSC57, MKN7, MKN45 and MKN74) strongly expressed the SETDB2 protein, while the remaining nine cell lines exhibited no or weak SETDB2 expression (low). B. Western blot analysis of SETDB2 expression in GC cells with its knockdown. SETDB2 expression was suppressed in MKN74 and MKN45 cells by transfection with SETDB2 siRNA (siSETDB2) compared with ones with negative control siRNA (NC). Actin expression was used as an internal control. C. Cell proliferation was quantified with a Cell Counting Kit-8 after transfection of SETDB2 siRNA into MKN74 and MKN45 cells. ◆, (blue), negative control siRNA; ▲, (green), SETDB2 siRNA transfection. Data are from 3 replicate experiments and error bars indicate standard deviation. Student’s t-test; **: P<0.01. D and E. The cell migration and invasion were assayed using a control insert (D) and a Matrigel invasion chamber (E), respectively, in MKN45 cells transfected with siSETDB2 or control siRNA. Original magnification: x100. Student’s t-test; **: P<0.01. F. Western blot analysis of SETDB2 overexpression in GC cells. AGS and NUGC3 were transfected with the empty vector (Empty) or the SETDB2 overexpressing vector (pSETDB2). Actin expression was analyzed as an internal control. G. Analysis of the cell proliferation of AGS and NUGC3 cells transfected with the SETDB2 overexpressing (▲, red), or the empty (◆, blue) vector. Data are from 3 replicate experiments and the error bars indicate standard deviation. Student’s t-test; **: P<0.01. H and I. Analyses of cell migration and invasion abilities in AGS and NUGC3 with SETDB2 overexpression using a control insert (H) and a Matrigel invasion chamber (I). Original magnification: x100. Student’s t-test; *: P<0.05.

The relationship between SETDB2 expression and global H3K9 methylation in GC cells

To investigate whether or not altered SETDB2 expression evokes H3K9 methylation change in GC cells, we compared global H3K9 methylation levels in MKN74 and MKN45 cells between SETDB2 knockdown and control cells. Global H3K9me3 at the protein levels were decreased in both of these cell lines with knockdown of SETDB2 by Western blotting, although global di- (H3K9me2, Figure 3A) and mono-methylation (H3K9me1, data not shown) levels were not changed. To further clarify whether decrease global H3K9me3 levels in our study were independently effects by SETDB2 knockdown or the secondary ones by other H3K9 tri-methyltransferase, we examined expression of SETDB1 and SUV39H1 in MKN74 and MKN45 cells with SETDB2 knockdown. The mRNA expression of SETDB1 and SUV39H1 was not changed in MKN74 and MKN45 cells after SETDB2 knockdown (Supplementary Figure S2), suggesting that SETDB2 itself is correlated with global H3K9me3 levels in these two GC cell lines.

Analyses of the SETDB2 target genes in GC cells.

Figure 3: Analyses of the SETDB2 target genes in GC cells. A. Western blot analysis of the global H3K9 methylation levels of MKN74 and MKN45 cells transfected with SETDB2 (siSETDB2) or control (NC) siRNA. B. After knockdown of SETDB2 in MKN74 cells, expressionaly altered six genes were selected as its downstream targets detected by microarray analysis. The mRNA expression levels of these genes were validated by real-time RT-PCR in MKN74 and MKN45 after knockdown of SETDB2. The white bars are cell lines transfected negative control siRNA and the black bars are cell lines transfected SETDB2 siRNA.

Analyses of SETDB2 target genes in GC cells

To determine SETDB2 target genes, microarray analysis of MKN74 cells with SETDB2 knockdown was performed. There were 520 genes that showed a greater than 1.5-fold change in expression level in MKN74 cells after SETDB2 knockdown, of which 185 genes were up-regulated and 335 genes were down-regulated. Because SETDB2 knockdown led to significant inhibition of cell growth, migration and invasion, we further selected six tumor- or differentiation-related genes (WWOX, CADM1/TSLC1, CCDC80, NDRG1, CA9 and FZD9) that have been reported to show altered expression in several cancers including GCs [1318]. Expression of WWOX, CADM1/TSLC1 and CCDC80 genes was up-regulated, while that of NDRG1, CA9 and FZD9 genes was down-regulated in MKN74 cells transfected with SETDB2 siRNA compared with control cells by real-time and conventional RT-PCR analyses (Figure 3B and Supplementary Figure S3A). These results were consistent with the microarray data. SETDB2 knockdown in MKN45 cells resulted in similar gene expression patterns except for expression of FZD9 (Figure 3B and Supplementary Figure S3A).

SETDB2 down-regulated the target genes through H3K9me3 in GCs

SETDB2 had an H3K9me3 activity in our study (Figure 3A), and H3K9me3 is known to be associated with gene silencing [4]. We therefore focused on the three genes (WWOX, CADM1 and CCDC80) whose expression was elevated by SETDB2 knockdown, and we examined how SETDB2 expression inhibited the expression of these genes in GC cells. ChIP assays showed that H3K9me3 levels at the promoter regions of WWOX and CADM1 were greatly decreased in MKN74 and MKN45 cells with SETDB2 siRNA transfection compared with negative control (Figures 4A-4D). In contrast, the H3K9me3 levels at a region approximately +3 kb downstream of the transcriptional start site (TSS) were not changed by SETDB2 siRNA transfection compared with negative control transfection (Figures 4B and 4D, right panels). Conversely, AGS cells transfected with the SETDB2 expression vector showed increased H3K9me3 levels at the promoter region compared with cells transfected with the empty vector (Figure 4E, left panels). Furthermore, ChIP assays exhibited that SETDB2 was recruited to the promoter region of WWOX and CADM1 in AGS cells after overexpression of SETDB2 (Figure 4E, right panels), suggesting that these two genes may be direct targets of SETDB2 in GC cells. Similar interaction of CCDC80 with SETDB2, however, was not detected (data not shown).

ChIP analysis of the promoter region of WWOX and CADM1 expression in GC cells.

Figure 4: ChIP analysis of the promoter region of WWOX and CADM1 expression in GC cells. A. Schematic representation of the promoter region of the WWOX gene. ChIP assay was performed at the regions of the promoter (-493 to -325) and gene body (+3040 to +3206). B. ChIP assay was performed with anti-histone H3 and anti-H3K9me3 polyclonal antibodies, and normal rabbit IgG. The H3K9me3 levels were enriched at the WWOX promoter region (-493 to -325) in MKN74 and MKN45 cells with negative control transfection (NC) compared with ones with siSETDB2 transfection. The H3K9me3 was not detected at the gene body region (+3040 to +3206). Input DNA sample was used as an internal control. C. Schematic representation of the promoter region of the CADM1 gene. D. ChIP assay was performed at the regions of the CADM1 promoter (-359 to -207) and gene body (+3085 to +3289). E) H3K9me3 levels and SETDB2 binding at the promoter regions of WWOX and CADM1 in AGS cells with SETDB2 overexpression.

We next compared the expression levels of endogenous SETDB2 and its three target genes (WWOX, CADM1 and CCDC80) in 13 GC cell lines using real-time and conventional RT-PCR analyses. The expression levels of WWOX in GC cell lines with high SETDB2 expression were lower than those with low SETDB2 expression, thereby displaying a tendency of WWOX expression to inversely correlate with SETDB2 expression (P=0.058), (Figure 5A and Supplementary Figure S3B). However, the expression levels of CADM1 and CCDC80 were not correlated with SETDB2 expression levels as assessed by either conventional RT-PCR (Figure 5A and Supplementary Figure S3B).

The relationship between SETDB2 and target gene expression in GC cell lines.

Figure 5: The relationship between SETDB2 and target gene expression in GC cell lines. A. The expression levels of WWOX, CADM1 and CCDC80 mRNA were examined in 13 GC cell lines using real-time RT-PCR. The black bars are cell lines with high SETDB2 protein expression and the white bars are cell lines with low SETDB2 protein expression. B. WWOX protein expression levels were examined by Western blotting in GC cells with SETDB2 knockdown (left), and overexpression (right). Actin expression was analyzed as an internal control.

Functional analysis of WWOX in GC cells

Since WWOX is known to play a role as a TSG in various cancers [19, 20], we studied the relationship between SETDB2 and WWOX. SETDB2 knockdown up-regulated WWOX protein expression in MKN74 and MKN45 cells, and, conversely, SETDB2 overexpression in AGS and NUGC3 cells repressed the protein expression of WWOX (Figure 5B). To assess whether or not the effects of SETDB2 overexpression in GC cells were mediated by reduced expression of WWOX, we performed knockdown of WWOX by transfection of its siRNA into AGS and NUGC3 GC cells (Figure 6A). The abilities of cell proliferation, migration and invasion were promoted after knockdown of WWOX in AGS and NUGC3 cells (P<0.05, Figure 6B-6D, Supplementary Figure S1B). It is noted that these effects were similar to those of SETDB2 overexpression in these cells (Figure 2G-2I, Supplementary Figure S1A).

Functional analysis of WWOX in GC cells.

Figure 6: Functional analysis of WWOX in GC cells. A. Western blot analysis of WWOX protein expression in AGS and NUGC3 cells transfected with negative control (NC) or WWOX siRNA (siWWOX). B. Cell proliferation was quantified with the Cell Counting Kit-8 after transfection of WWOX siRNA (▲, red) or negative control siRNA (◆, blue) into GC cells. Data are from 3 replicate experiments and error bars indicate standard deviation. Student’s t-test; **: P<0.01. C and D. The cell migration (C) and the invasion (D) abilities of AGS and NUGC3 cells with WWOX siRNA were assayed using a control insert and a Matrigel invasion chamber, respectively. Original magnification: x100. Student’s t-test; *: P<0.05, **: P<0.01.

DISCUSSION

Previous studies have reported the expression changes of the HMT genes in various cancers [2124]. The H3K9 tri-methyltransferase genes, SETDB1 and SUV39H1, have been shown to be overexpressed in cancers, such as lung and liver cancers [2527]. Overexpression of these two genes is related to cell proliferation and invasiveness, and Setdb1 accelerated melanoma formation in a transgenic zebrafish model, indicating that they have oncogenic properties [28]. In this study, we showed that the SETDB2 protein was highly expressed in GCs (41.7%), particularly in advanced GCs. Moreover, high SETDB2 expression in GCs was correlated with worse prognosis by Kaplan-Meier analysis (P<0.05). The in vitro studies using SETDB2 knockdown and overexpressing cells of the present study demonstrated that SETDB2 promoted cell proliferation, migration and invasion of GC cells. Thus, SETDB2 expression and function detected in our study were similar to those of SETDB1 and SUV39H1 in cancers [5, 6], suggesting that SETDB2 may have oncogenic functions and overexpression of SETDB2 play roles in gastric cancer progression.

Although multiple transcriptional/translational regulatory mechanisms in SETDB1 and SUV39H1 have been identified [26, 27], the upstream mechanisms of SETDB2 are unknown. It has been reported that SETDB1 promoter region contains multiple SP1-binding sites and SETDB1 expression was enhanced by SP1, suggesting a significant relation between SETDB1 and TF bindings [26]. We observed that three putative myc-binding sites were but only one SP1 located at the SETDB2 promoter region by using UCSC Genome Bioinformatics Site (http://genome.ucsc.edu/cgi-bin/hgGateway). Knockdown of c-myc in MKN74 and MKN45 cells decreased SETDB2 at the mRNA and protein expression levels (Supplementary Figure S4A-S4C), indicating that c-myc may act as a transcriptional regulator of SETDB2 in these two GC cell lines. Besides, amplification of the SETDB1 gene was found in lung and liver cancers [5, 25, 26], and SETDB1 and SUV39H1 were negatively regulated by miR-29 and miR-125b, respectively [26, 27]. Further studies are necessary to clarify the transcriptional and translational regulatory mechanisms of SETDB2 overexpression in GC cells.

The H3K9 methyltransferase genes mediated global and promoter/enhancer region of H3K9 tri- di- and mono- methylation [2932]. Knockdown of SETDB2 in MKN74 and MKN45 cells reduced global H3K9me3 (Figure 3A) but not H3K9me2 and H3K9me1 at the protein level, indicating that SETDB2 participates as an H3K9 tri-methyltransferase [10, 11]. High levels of global H3K9me3 protein have been shown in primary cancers including GCs [33, 34]. Therefore, we speculate that elevated H3K9me3 levels in GC tissues reported may be mediated in part by SETDB2 overexpression.

In addition, we observed that knockdown of SETDB2 resulted in a decrease in H3K9me3 levels at the promoter regions of WWOX and CADM1 genes, both of whose expression were increased by SETDB2 knockdown. Ectopic SETDB2 in GC cells was recruited to the promoters of these two genes and increased H3K9me3 levels at the regions. Therefore, SETDB2 may repress target gene expression not only through increasing global H3K9me3 levels but also directly binding and inducing H3K9me3 activity on the promoter region of WWOX and CADM1 in GC cells. In contrast, it is possible that CCDC80 may be indirectly regulated by SETDB2, because SETDB2 did not bind to its promoter region.

Regarding the three genes (WWOX, CADM1 and CCDC80) detected as SETDB2 target genes in our study, their loss or reduced expression has been reported in various cancers [3538]. Here, loss or low expression of these three genes was found in GC cell lines examined (Figure 5A), and the expression levels of WWOX in GC cell lines showed a tendency to inversely correlate with SETDB2 expression. It is known that WWOX and CADM1 but not CCDC80 have dense CpG islands in their promoter regions and hypermethylation at the promoter regions of WWOX and CADM1 has been reported in cancer cells [39, 40]. We examined the methylation status of WWOX and CADM1 genes in 13 GC cell lines. No methylation of WWOX was found in any of the GC cell lines, while severe or partial methylation at the CADM1 promoter was detected in six GC cell lines (data not shown). MKN74 and MKN45 cells exhibited no or partial CADM1 methylation. Our data further support that WWOX and CADM1 may be repressed by SETDB2 in GC cells.

In conclusion, our data indicated that SETDB2 is often over-expressed in GC tissues and cell lines. SETDB2 overexpression significantly accelerated cell proliferation, migration and invasion of GC cells. Moreover, SETDB2 restricted the expression of TSGs such as WWOX, through H3K9me3 at the their promoter regions. Since SETDB2 showed oncogenic functions in our study, it is possible that knockdown of SETDB2 might be an effective therapeutic strategy for GCs.

MATERIALS AND METHODS

Cell lines and tissues samples

We studied 13 human GC cell lines (AGS, GCIY, HSC39, HSC44PE, HSC57, HSC59, KATOIII, MKN7, MKN45, MKN74, NUGC3, NUGC4 and TGBC11TKB). MKN7, MKN45, MKN74, GCIY and TGBC11TKB were purchased from RIKEN cell bunk, NUGC3 and NUGC4 were from Japanese Collection of Research Bioresources Cell Bank, and KATOIII and AGS were from ATCC (America Type Cell Collection). HSC39, HSC44PE, HSC57 and HSC59 were obtained from Dr. Kazuyoshi Yanagihara. All the cell lines were grown in RPMI 1640 medium (Wako, Saitama, Japan) or DMEM medium (Wako) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37°C under 5% CO2 in a humidified incubator.

Cancer tissue samples were collected from 72 GC patients in an affiliated hospital of Tokyo Medical and Dental university. Informed consent was obtained from all of the patients, and the study was approved by the appropriate institutional review committee.

RNA isolation, RT-PCR and real-time RT-PCR analysis

Total RNA was isolated using the Trizol reagent (Thermo Fisher Scientific Inc., Waltham, MA) or the RNeasy plus Minikit (QIAGEN, Valencia, CA). The isolated RNA (1 μg) was reverce-transcribed by using Super Script III Reverse transcriptase (Thermo Fisher Scientific Inc.). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified as an internal control. The primer sequences used are shown in Supplementary Table S1 (Supplementary Material).

Western blot analysis

Total protein was extracted using RIPA buffer and was then electrophoresed on SDS-PAGE. Western blotting was conducted with mouse monoclonal anti-SETDB2 (1:500; M17, Abnova, Taipei, Taiwan), and anti-WWOX (1:300; sc-374449, Santa Cruz Biotechnology, Santa Cruz, CA): and rabbit polyclonal anti-H3 (1:4000; Active Motif, Carlsbad, CA), anti-H3K9me3 (1:2000; Active Motif) and anti-H3K9me2 (1:2000; Active Motif). Mouse monoclonal anti-Pan Actin (1:5000; MILLIPORE, Billerica, MA) was used as an internal control for Western blotting.

Immunohistochemical staining

Immunohistochemistry (IHC) of formalin-fixed paraffin-embedded tissues was performed as described previously [41]. SETDB2 protein expression was evaluated in a set of 72 primary GC tissues by using rabbit polyclonal anti-SETDB2 (1:100; ab5517, abcam, Cambridge, UK). Sections in which > 10% of the cancer cells showed positive reactivity with SETDB2 were scored as positive. Similary, score of the SETDB2 positivity (>10% of cells) was used in non-cancerous stomach tissues.

Plasmid construction

Full length SETDB2 was amplified from cDNA of MKN74 cells using RT-PCR. The primer sequences used were: 5’-CCGCTTAAGCTTAAGATGGGAGAAAAAAATGGCG-3’(sense) and 5’-CGAATTGATATCCACAAACAGGCGTTAGTTACAT-3’(antisense). After digestion with Hind III and EcoR V, the PCR product was subcloned into the pcDNA3 vector.

Transfection

MKN74 and MKN45 cells with high SETDB2 expression were transfected with SETDB2 siRNAs (final concentration, 50nmol; Hs_SETDB2_0144, Sigma-Aldrich Japan, Ishikari, Japan) or negative control siRNA (Mission_SIC-001, Sigma-Aldrich Japan) using electroporation (NeonTM transfection System, Thermo Fisher Scientific Inc.). AGS and NUGC3 cells with low SETDB2 expression were transfected with the SETDB2/pcDNA3 expression vector (1 μg) or with the empty pcDNA3 vector using the FuGENE HD Transfection Reagent (Promega Corporation, Madison, WI). Transfection of WWOX siRNA (Hs_WWOX_9365, Sigma-Aldrich Japan) was conducted in GC cells by electroporation. Transfected cells were harvested after 48 or 72 hours of culture, and total RNA and protein was then extracted.

Cell growth assay

The effects of SETDB2 knockdown and overexpression on cell proliferation were quantified using the Cell Counting Kit-8 (Dojindo, Kumamoto, Japan). MKN74 (4×103 cells/well) and MKN45 (3×103 cells/well) cells were transfected with SETDB2 siRNA or negative control siRNA. AGS (1×103 cells/well), and NUGC3 (1×103 cells/well) cells were transfected with the SETDB2/pcDNA3 or empty pcDNA3 expression vector. Moreover, the cell proliferation rates of AGS and NUGC3 cells transfected with WWOX siRNA were examined. GC cells transfected with siRNA or empty vector were seeded onto 96-well plates, and their cell proliferation was then determined by WST-8 activity.

Transwell migration and invasion assays

MKN45, AGS and NUGC3 cells were transfected with SETDB2 siRNA, WWOX siRNA, full length SETDB2/pcDNA3 or the empty/pcDNA3expression vector. After 48 hours, MKN45 (4×103 cells/0.5ml), AGS (3×103 cells/0.5ml) or NUGC3 (5×103 cells/0.5ml) cells were seeded in a control insert or a Matrigel invasion chamber (BD Biosciences, Franklin Lakes, NJ) [9]. Medium containing 10% FBS in the lower chamber served as chemoattractant. After 48 hours, the non-invaded cells were gently removed with a cotton swab and the migrated and invaded cells located on the lower side of the chamber were stained with crystal violet.

Scratch wound healing assay

At 48 hours after transfection of AGS and NUGC3 cells with the SETDB2 expression vector or with WWOX siRNA, the cells were seeded in 6 well plates at a density of 4×105 cells/well and cultured to confluence. A scratch was made through the cell monolayer using a pipet tip. To monitor the migration of cells into the wounded area, photographs of the wounded area were taken immediately after the scratch was made (0 hour) and after then 18 hours later. The relative wound healing range was determined [9].

Microarray analysis

MKN74 cells were transfected with SETDB2 siRNA or negative control siRNA, and were then grown for 48 hours. Total RNA was extracted using RNEasy and cDNA microarray analysis was conducted by DNA Chip Research Inc. (Kanagawa, Japan) using the Agilent Human GE 4x44K v2 Microarray (Platform ID: GPL13497). The raw data have been deposited in NCBI Gene Expression Omnibus (GEO) under the accession number GSE76135.

Chromatin immunoprecipitation (ChIP) assay

MKN74 and MKN45 cells were transfected with SETDB2 siRNA or negative control siRNA for 48 hours. ChIP assays were performed using the ChIP-IT Express Kit (Active Motif) according to the manufacturer’s instructions [9]. AGS cells were transfected with the SETDB2/pcDNA3 or empty/pcDNA3 expression vector. Antibodies used were anti-H3 (Active Motif), anti-H3K9me3 (Active Motif), Normal Rabbit IgG (No. 2729, Cell Signaling Technology, Danvers, MA) and anti-SETDB2 (No. 5517, abcam).

Statistical analyses

The Mann-Whitney’s U-test, c2 and Fisher’s exact tests were used to compare the values for the test and control samples. Each transfection experiment was repeated at least in triplicate. Patient survival rate was calculated by the Kaplan-Meier method. A value of P<0.05 was determined as statistically significant.

Abbreviations

ChIP, Chromatin immunoprecipitation; GC, gastric cancer; HMT, histone methyltransferase; H3K9me3, histone H3 lysine 9 tri-methylation; IHC, immunohistochemistry; RT-PCR, reverse transcription-polymerase chain reaction; TSGs, tumor suppressor genes; TSS, transcriptional start site.

ACKNOWLEDGMENTS

We thank Ms. Hiromi Nagasaki and Dr. Ayuna Sakamoto (Tokyo Medical and Dental University) for technical assistance and helpful suggestion. We thank Dr. Kazuyoshi Yanagihara (National Cancer Center, Japan) for providing HSC39, HSC44PE, HSC57 and HSC59 cells.

CONFLICTS OF INTEREST

All authors declare that they have no conflict of interest to disclose.

GRANT SUPPORT

This work was supported in part by Grants-in-Aid for Scientific Research (C) (Y.A., 24590444 and 16K08647), (B) (Y.Y., 24650609), (A) (S.T., 25253081 and 16H02670) and the A3 Foresight Program (AA005), of the Japan Society for the Promotion of Science (JSPS).

Author contributions

Conception, design of the study and writing of the manuscript: T.N., Y.A, Y.Y., S.T.; analysis and/or interpretation of data: T.N., Y.A., S.S.; patient sample collection: K.K., T.K., Y.E.; drafting or revision of the manuscript: T.N., Y.A., S.S., Y.Y., S.T.

REFERENCES

1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015; 136: E359-386.

2. Yuasa Y. Control of gut differentiation and intestinal-type gastric carcinogenesis. Nat Rev Cancer. 2003; 3:592-600. Review.

3. Wanajo A, Sasaki A, Nagasaki H, Shimada S, Otsubo T, Owaki S, Shimizu Y, Eishi Y, Kojima K, Nakajima Y, Kawano T, Yuasa Y, Akiyama Y. Methylation of the calcium channel-related gene, CACNA2D3, is frequent and a poor prognostic factor in gastric cancer. Gastroenterology. 2008; 35: 580-590.

4. Albert M, Helin K. Histone methyltransferases in cancer. Semin Cell Dev Biol. 2010; 21: 209-220.

5. Rodriguez-Paredes M, Martinez de Paz A, Simó-Riudalbas L, Sayols S, Moutinho C, Moran S, Villanueva A, Vázquez-Cedeira M, Lazo PA, Carneiro F, Moura CS, Vieira J, Teixeira MR, et al. Gene amplification of the histone methyltransferase SETDB1 contributes to human lung tumorigenesis. Oncogene. 2014; 33: 2807-2813.

6. Spyropoulou A, Gargalionis A, Dalagiorgou G, Adamopoulos C, Papavassiliou KA, Lea RW, Piperi C, Papavassiliou AG. Role of histone lysine methyltransferases SUV39H1 and SETDB1 in gliomagenesis: modulation of cell proliferation, migration, and colony formation. Neuromolecular Med. 2014; 16: 70-82.

7. Chiba T, Saito T, Yuki K, Zen Y, Koide S, Kanogawa N, Motoyama T, Ogasawara S, Suzuki E, Ooka Y, Tawada A, Otsuka M, Miyazaki M, et al. Histone lysine methyltransferase SUV39H1 is a potent target for epigenetic therapy of hepatocellular carcinoma. Int J Cancer. 2015; 136: 289-298.

8. Matsukawa Y, Semba S, Kato H, Ito A, Yanagihara K, Yokozaki H. Expression of the enhancer of zeste homolog 2 is correlated with poor prognosis in human gastric cancer. Cancer Sci. 2006; 97:484-491.

9. Akiyama Y, Koda Y, Byeon SJ, Shimada S, Nishikawaji T, Sakamoto A, Chen Y, Kojima K, Kawano T, Eishi Y, Deng D, Ho Kim W, et al. Reduced expression of SET7/9, a histone mono-methyltransferase, is associated with gastric cancer progression. Oncotarget. 2016; 7:3966-83. doi: 10.18632/oncotarget.6681.

10. Xu PF, Zhu KY, Jin Y, Chen Y, Sun XJ, Deng M, Chen SJ, Chen Z, Liu TX. Setdb2 restricts dorsal organizer territory and regulates left-right asymmetry through suppressing fgf8 activity. Proc Natl Acad Sci U S A. 2010; 107:2521-2526.

11. Falandry C, Fourel G, Galy V, Ristriani T, Horard B, Bensimon E, Salles G, Gilson E, Magdinier F. CLLD8/KMT1F is a lysine methyltransferase that is important for chromosome segregation. J Biol chem. 2010; 285: 20234-20241.

12. Du TT, Xu PF, Dong ZW, Fan HB, Jin Y, Dong M, Chen Y, Pan WJ, Ren RB, Liu TX, Deng M, Huang QH. Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev Biol. 2014; 392: 233-244. doi: 10.1016/j.ydbio.2014.05.022.

13. Guo W, Wang G, Dong Y, Guo Y, Kuang G, Dong Z. Decreased expression of WWOX in the development of esophageal squamous cell carcinoma. Mol Carcinog 2013; 52: 265-274.

14. Wikman H, Westphal L, Schmid F, Pollari S, Kropidlowski J, Sielaff-Frimpong B, Glatzel M, Matschke J, Westphal M, Iljin K, Huhtala H, Terracciano L, Kallioniemi A, et al. Loss of CADM1 expression is associated with poor prognosis and brain metastasis in breast cancer patient. Oncotarget. 2014; 5: 3076-3087. doi: 10.18632/oncotarget.1832.

15. Bommer GT, Jäger C, Dürr EM, Baehs S, Eichhorst ST, Brabletz T, Hu G, Fröhlich T, Arnold G, Kress DC, Göke B, Fearon ER, Kolligs FT. DRO1, a gene down-regulated by oncogenes, mediates growth inhibition in colon and pancreatic cancer cells. J Biol Chem. 2005; 280: 7962-7975.

16. Ureshino H, Murakami Y, Watari K, Izumi H, Kawahara A, Kage M, Arao T, Nishio K, Yanagihara K, Kinoshita H, Kuwano M, Ono M. N-myc downstream regulated gene 1 (NDRG1) promotes metastasis of human scirrhous gastric cancer cells through epithelial mesenchymal transition. PLoS One. 2012; 7: e41312.

17. Lou Y, McDonald PC, Oloumi A, Chia S, Ostlund C, Ahmadi A, Kyle A, Auf dem Keller U, Leung S, Huntsman D, Clarke B, Sutherland BW, Waterhouse D, et al. Targeting tumor hypoxia: suppression of breast tumor growth and metastasis by novel carbonic anhydrase IX inhibitors. Cancer Res. 2011; 71: 3364-3376.

18. Zhang Z, Schittenhelm J, Guo K, Bühring HJ, Trautmann K, Meyermann R, Schluesener HJ. Upregulation of frizzled 9 in astrocytomas. Neuropathol Appl Neurobiol. 2006; 32: 615-624.

19. Del Mare S, Salah Z, Aqeilan RI. WWOX: its genomics, partners, and functions. J Cell Biochem. 2009; 108:737-745. Review.

20. Guo W, Dong Z, Dong Y, Guo Y, Kuang G, Yang Z. Genetic and epigenetic alterations of WWOX in the development of gastric cardia adenocarcinoma. Environ Mol Mutagen. 2013; 54: 112-123.

21. Greer EL, Shi Y. Histone methylation: a dynamic mark in health, disease and inheritance. Nat Rev Genet. 2012; 13:343-357. Review.

22. Varier RA, Timmers HT. Histone lysine methylation and demethylation pathways in cancer. Biochim Biophys Acta. 2011; 1815:75-89.

23. Bae WK, Yoo KH, Lee JS, Kim Y, Chung IJ, Park MH, Yoon JH, Furth PA, Hennighausen L. The methyltransferase EZH2 is not required for mammary cancer development, although high EZH2 and low H3K27me3 correlate with poor prognosis of ER-positive breast cancers. Mol Carcinog. 2015; 54: 1172-1180.

24. Ren G, Baritaki S, Marathe H, Feng J, Park S, Beach S, Bazeley PS, Beshir AB, Fenteany G, Mehra R, Daignault S, Al-Mulla F, et al. Polycomb protein EZH2 regulates tumor invasion via the transcriptional repression of the metastasis suppressor RKIP in breast and prostate cancer. Cancer Res. 2012; 72: 3091-3104.

25. Sun QY, Ding LW, Xiao JF, Chien W, Lim SL, Hattori N, Goodglick L, Chia D, Mah V, Alavi M, Kim SR, Doan NB, Said JW, et al. SETDB1 accelerates tumourigenesis by regulating the WNT signalling pathway. J Pathol. 2015; 235: 559-570.

26. Wong CM, Wei L, Law CT, Ho DW, Tsang FH, Au SL, Sze KM, Lee JM, Wong CC, Ng IO. Up-regulation of histone methyltransferase SETDB1 by multiple mechanisms in hepatocellular carcinoma promotes cancer metastasis. Hepatology. 2015; doi: 10.1002/hep.28304.

27. Fan DN, Tsang FH, Tam AH, Au SL, Wong CC, Wei L, Lee JM, He X, Ng IO, Wong CM. Histone lysine methyltransferase, suppressor of variegation 3-9 homolog 1, promotes hepatocellular carcinoma progression and is negatively regulated by microRNA-125b. Hepatology. 2013; 57:637-647.

28. Ceol CJ, Houvras Y, Jane-Valbuena J, Bilodeau S, Orlando DA, Battisti V, Fritsch L, Lin WM, Hollmann TJ, Ferré F, Bourque C, Burke CJ, Turner L, et al. The histone methyltransferase SETDB1 is recurrently amplified in melanoma and accelerates its onset. Nature. 2011; 471: 513-517. doi: 10.1038/nature09806.

29. Tan SL, Nishi M, Ohtsuka T, Matsui T, Takemoto K, Kamio-Miura A, Aburatani H, Shinkai Y, Kageyama R. Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development. Development. 2012; 139:3806-3816.

30. Rivera C, Saavedra F, Alvarez F, Díaz-Celis C, Ugalde V, Li J, Forné I, Gurard-Levin ZA, Almouzni G, Imhof A, Loyola A. Methylation of histone H3 lysine 9 occurs during translation. Nucleic Acids Res. 2015; 43:9097-9106. doi: 10.1093/nar/gkv929.

31. Dong C, Wu Y, Wang Y, Wang C, Kang T, Rychahou PG, Chi YI, Evers BM, Zhou BP. Interaction with Suv39H1 is critical for Snail-mediated E-cadherin repression in breast cancer. Oncogene. 2013; 32: 1351-1362. doi: 10.1038/onc.2012.169.

32. Chen MW, Hua KT, Kao HJ, Chi CC, Wei LH, Johansson G, Shiah SG, Chen PS, Jeng YM, Cheng TY, Lai TC, Chang JS, Jan YH, et al. H3K9 histone methyltransferase G9a promotes lung cancer invasion and metastasis by silencing the cell adhesion molecule Ep-CAM. Cancer Res. 2010; 70: 7830-7840. doi: 10.1158/0008-5472.CAN-10-0833.

33. Park YS, Jin MY, Kim YJ, Yook JH, Kim BS, Jang SJ. The global histone modification pattern correlates with cancer recurrence and overall survival in gastric adenocarcinoma. Ann Surg Oncol. 2008; 15:1968-1976.

34. Cai L, Ma X, Huang Y, Zou Y, Chen X. Aberrant histone methylation and the effect of Suv39H1 siRNA on gastric carcinoma. Oncol Rep. 2014; 31: 2593-2600. doi: 10.3892/or.2014.3135.

35. Park SW, Ludes-Meyers J, Zimonjic DB, Durkin ME, Popescu NC, Aldaz CM. Frequent downregulation and loss of WWOX gene expression in human hepatocellular carcinoma. Br J Cancer. 2004; 91:753-759.

36. Iliopoulos D, Fabbri M, Druck T, Qin HR, Han SY, Huebner K. Inhibition of breast cancer cell growth in vitro and in vivo: effect of restoration of Wwox expression. Clin Cancer Res. 2007; 13: 268-274.

37. Kuramochi M, Fukuhara H, Nobukuni T, Kanbe T, Maruyama T, Ghosh HP, Pletcher M, Isomura M, Onizuka M, Kitamura T, Sekiya T, Reeves RH, et al. TSLC1 is a tumor-suppressor gene in human non-small-cell lung cancer. Nat Genet. 2001; 27:427-430.

38. Ferraro A, Schepis F, Leone V, Federico A, Borbone E, Pallante P, Berlingieri MT, Chiappetta G, Monaco M, Palmieri D, Chiariotti L, Santoro M, et al. Tumor suppressor role of the CL2/DRO1/CCDC80 gene in thyroid carcinogenesis. J Clin Endocrinol Metab. 2013; 98:2834-2843.

39. Maeda N, Semba S, Nakayama S, Yanagihara K, Yokozaki H. Loss of WW domain-containing oxidoreductase expression in the progression and development of gastric carcinoma: clinical and histopathologic correlations. Virchows Arch. 2010; 457:423-432.

40. Fukami T, Fukuhara H, Kuramochi M, Maruyama T, Isogai K, Sakamoto M, Takamoto S, Murakami Y. Promoter methylation of the TSLC1 gene in advanced lung tumors and various cancer cell lines. Int J Cancer. 2003; 107: 53-59.

41. Sakamoto A, Akiyama Y, Shimada S, Zhu WG, Yuasa Y, Tanaka S. DNA Methylation in the Exon 1 Region and Complex Regulation of Twist1 Expression in Gastric Cancer Cells. PLoS One. 2015; 10: e0145630.


Creative Commons License All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 License.
PII: 11625