Reduced expression of SET7/9, a histone mono-methyltransferase, is associated with gastric cancer progression

Abstract

SET7/9, a histone methyltransferase, has two distinct functions for lysine methylation. SET7/9 methylates non-histone proteins, such as p53, and participates in their posttranslational modifications. Although SET7/9 transcriptionally activate the genes via H3K4 mono-methylation, its target genes are poorly understood. To clarify whether or not SET7/9 is related to carcinogenesis, we studied alterations of SET7/9 in gastric cancers (GCs). Among the 376 primary GCs, 129 cases (34.3%) showed loss or weak expression of SET7/9 protein compared to matched non-cancerous tissues by immunohistochemistry. Reduced SET7/9 expression was significantly correlated with clinical aggressiveness and worse prognosis. Knockdown of SET7/9 in GC cells markedly increased cell proliferation, migration and invasion. Expression of SREK1IP1, PGC and CCDC28B were inhibited in GC cells with SET7/9 knockdown, while matrix metalloproteinase genes (MMP1, MMP7 and MMP9) were activated. SET7/9 bound and mono-methylated H3K4 at the region of the approximately 4-6 kb upstream from the SREK1IP1 transcriptional start site and the promoters of PGC and CDC28B. Cell proliferation, migration and invasion, and expression of three MMPs were increased in GC cells with SREK1IP knockdown, which were similar to those of SET7/9 knockdown. These data suggest that SET7/9 has tumor suppressor functions, and loss of SET7/9 may contribute to gastric cancer progression.

Keywords: H3K4me1; SET7/9; SREK1IP1; gastric cancer; histone methyltransferase.

Figure 1. SET7/9 expression in primary GCs

A. Representative IHC photographs of SET7/9 in primary tissues. Non-cancerous gastric mucosa including intestinal metaplasia exhibited strong SET7/9 expression (i). Positive (ii) and negative (iii) staining of SET7/9 was detected in GCs. Original magnification x100. B. Kaplan-Meier curve of overall survival for GC patients with SET7/9 protein expression. The GC patients with loss or weak SET7/9 expression (red, N = 129) had a significantly poorer outcome than those with retained SET7/9 (blue, N = 247) (P = 0.038, logrank test). C. qRT-PCR analysis of SET7/9 expression in 25 primary GC tissues. IHC of SET7/9 was performed in these tissues, which were divided into two groups, SET7/9 protein expression-high (retained, N = 15) and -low (loss or weak, N = 10). Relative expression was calculated using GAPDH expression as an internal control. Mann—Whitney U-test, *P = 0.017.

Figure 2. SET7/9 expression in GC cell lines and non-cancerous stomach tissues

A. SET7/9 mRNA expression in 12 GC cell lines and three non-cancerous stomach tissues from GC patients. Quantitative RT-PCR was performed using a LightCycler system. The 2^nd Derivative Maximum method was utilized for determination of the concentrations, and relative expression was calculated using GAPDH expression as an internal control. The averages (blue columns) of two independent experiments for GC cell lines are indicated. The average (white column) and standard deviation (S.D) of the SET7/9 expression levels in the three non-cancerous stomach tissues (N.St) were also calculated. B. WB analysis of SET7/9 protein. a-Tubulin was used as an internal control. The Image-J 1.47v software software (http://imagej.nih.gov/ij/index.html) was used to calculate the SET7/9 protein expression levels. Red, GC cases with weak expression of both SET7/9 mRNA and protein; blue, GC cases with strong SET7/9 mRNA but weak SET7/9 protein expression. C. qRT-PCR analysis of SET7/9 expression in MKN7 and HSC43 cells after epigenetic drug treatment. MKN7 and HSC43 cells were treated with 0.3 mM trichostatin A (TSA; WAKO, Osaka, Japan) for 24 hrs. M, mock; T, TSA. *P < 0.01. D. qChIP assay of the SET7/9 promoter region in AGS and MKN7 cells exhibiting strong and weak SET7/9 expression, respectively. ChIP was quantitatively performed with anti-histone H3, and anti-acetylated histone H3 and H4 polyclonal antibodies. Input DNA samples were used as internal controls. E. Effects on histone H3 and H4 acetylation in MKN7 cells after TSA treatment. F. Longer RT-PCR of SET7/9 exons 1-8 in TGBC11TKB and KATO-III cells (left). Sequencing analysis of the PCR product of SET7/9 in TGBC11TKB cells. After subcloning of the RT-PCR product showing an abnormal SET7/9 transcript in TGBC11TKB cells, sequencing was performed (right). Deletion of exon 3 predicts to encode a truncated SET7/9 protein lacking the SET domain (5,6). We could not detect this predicted small SET7/9 protein (62 aa) in TGBC11TKB cells on WB using anti-SET7/9 antibodies, although wild-type SET7/9 (366 aa) was detected (see Figure 2B).

Figure 3. Effects of siRNA-based SET7/9 knockdown in GC cells

A. Locations of the two SET7/9 siRNAs used in this study. B. qRT-PCR analysis of SET7/9 expression in four GC cell lines after its siRNA transfection. The average (column) ± S.D (bar) of three independent experiments is indicated. *P < 0.01. C. WB analysis of SET7/9 protein expression in GC cell lines. a-Tubulin was used as an internal control. D. Cell proliferation assays of four GC cell lines after SET7/9 knockdown. The number of viable cells was determined using a WST-8 cell proliferation assay kit. Three independent assays were conducted. Representative data are shown and error bars indicate S.D. *P < 0.05; **P < 0.01; NS, not significant. E. Cell migration and invasion assays of MKN74 and MKN45 cells after SET7/9 knockdown. GC cells with SET7/9 (si7-1) and negative control (NC) siRNA transfection were grown in 24-transwell plates coated with (invasion) or without (migration) matrigel for 18-48 hrs. The total numbers of migrating and invading cells on the membrane were counted, and then relative migration and invasion ratios were calculated. The average (column) ± S.D (bar) of three independent experiments is indicated (*P < 0.01). Photographs show representative fields of migrating and invading cancer cells on the membrane (x100).

Figure 4. Analyses of the SET7/9 target genes in GC cells

A. Venn diagrams of transcriptionally down- and up-regulated gene expression of the genes in MKN74 and MKN45 cells after SET7/9 knockdown by microarray. The cut-off values of altered expression levels were determined as 1.5-fold change in GC cells with SET7/9 knockdown compared to ones with negative control transfection. B. Heat maps of transcriptionally down- and up-regulated gene expression of the genes in MKN74 and MKN45 cells after SET7/9 knockdown. The expression levels of the genes are represented by log2 ratios, and cut-off is determined as 1.5-fold changes. Six genes showing reduced expression including SET7/9 (Green), and five (Red) ones showing increased expression were selected in this study, as shown by arrows. C. RT-PCR analyses of the predicted STE7/9 target genes in 4 GC cell lines after SET7/9 siRNA transfection. MKN74, MKN45, KATO-III and AGS cells, which exhibited strong SET7/9 expression, were used in this study. At 48 hrs after transfection, RT-PCR was performed. GAPDH was used as an internal control. PCR products showing down- and up-regulation of the SET7/9 target genes are shown by asterisks and triangles, respectively. D. qRT-PCR analyses of the selected genes showing increased and decreased expression by SET7/9 knockdown in four GC cells. The average (column) ± S.D (bar) of three independent experiments is indicated. *P < 0.01; **P < 0.05.

Figure 5. Effects of SET7/9 overexpression in GC cells

A. WB analysis of SET7/9 protein expression in MKN7 cells. Exogenous SET7/9 protein expression was detected using anti-SET7/9 and anti-FLAG antibodies. Empty: empty vector (p3X-FLAG-CMV^TM-10), and pSET7/9: wild-type SET7/9. B. and C. RT-PCR analyses of the SET7/9 target genes in MKN7 cells after SET7/9 overexpression. The expression levels of the six genes in non-cancerous stomach tissues were also analyzed by RT-PCR. Expressions of the genes showing increased and decreased expression are shown by triangles and asterisks, respectively. D. qRT-PCR analyses of SREK1IP1, PGC, CCDC28B and three MMP genes in MKN7 cells exhibiting SET7/9 overexpression. The average (column) ± S.D (bar) of three independent experiments is indicated. *P < 0.01.

Figure 6. Transcriptional regulatory mechanisms of SREK1IP1, PGC and CCDC28B via H3K4me1 in GCs

A. Schematic representation of the 5′-regions of SREK1IP1, PGC and CCDC28B. TSS and closed boxes demonstrate the transcriptional start site and exons, respectively, and horizontal arrows are the qChIP sites in this study. A dotted line indicates an H3K4me1 enrichment region that has been shown by ChIP-seq data on the UCSC database. The SREK1IP1 and CWC27 genes are organized in a head-to-head configuration. B. The H3K4me1 levels in association with SET7/9 expression. qChIP was performed using anti-H3K4me1 antibodies in MKN7 cells with exogenous SET7/9 overexpression (left) and AGC cells with knockdown of the SET7/9 (right). As SET7/9 was overexpressed in MKN7 cells on transfection of SET7/9-3x FLAG expression vector, ChIP using anti-FLAG antibodies (middle) was performed.*P < 0.01; **P < 0.05; NS, not significant.

Figure 7. Effects of SREK1IP1 siRNA transfection in GC cells

A. Endogenous SREK1IP1 protein expression levels in seven GC cell lines and three non-cancerous stomach tissues. Although the size of SREK1IP1 protein is known to be 18kDa, two different sizes (approximately 18kDa and 20kDa, arrowheads) of SREK1IP1 protein were detected on WB. B. WB analyses of SREK1IP1 protein expression in AGS and MKN45 cells after SET7/9 (left, si7-1 and si7-2) and SREK1IP1 (right, siSR-1 and siSR-2) siRNA transfection. Knockdown of each gene inhibited expression of the predicted (18kDa) and large size (20kDa) proteins in MKN45 and AGS cells, respectively. C. Cell proliferation assays after SREK1IP1 knockdown in AGS and MKN45 cells. To compare the effects on SREK1IP1 and SET7/9, we also performed knockdown of SET7/9 (si7-1) in this study. Three independent assays of each knockdown were conducted. Representative data are shown and error bars indicate S.D. *P < 0.05. D. Matrigel cell invasion and migration assays of AGS and MKN45 cells. The total numbers of migrating and invading cells on the membrane in at least three fields were counted. The average (column) ± S.D (bar) of three independent experiments is indicated (*P < 0.01, top). Photographs show representative fields of migrating and invading cells on the membrane (bottom). E. The effects of SREK1IP1 knockdown by its siRNA transfection and the resultant MMP1 expression were quantitatively analyzed by qRT-PCR. The average (column) ± S.D (bar) of three independent experiments is indicated (*P < 0.01).