Six human gastric cancer cell lines (SGC7901, HGC27, BGC823, MKN45, MGC803 and AGS) and human gastric epithelial cell line GES-1 were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). SGC7901, MKN45, MGC803, and AGS were maintained in the RPMI-1640 medium (Hyclone, South Logan, UT, USA) containing 10% fetal bovine serum (FBS, Invitrogen, Carisbad, CA, USA). HGC27, BGC823 and GES-1 were cultured in the DMEM medium (Invitrogen) supplemented with 10% FBS. All cells were incubated at 37°C in a humid atmosphere of 5% CO₂.

Fifty-two paired gastric cancer tissues and adjacent non-cancerous gastric tissues were obtained after informed consents from patients at the First Affiliated Hospital of Soochow University. Tissue samples were collected, immediately snap-frozen in liquid nitrogen and stored in −80°C until analysis. None of the patients had received chemotherapy or radiotherapy prior to surgery. This research was approved by the Academic Advisory Board of Soochow University.

5-Aza was purchased from Sigma-Aldrich Corp. (St. Louis, MO, USA) and dissolved in sterile double distilled water. Cells were plated at a density of 1×10⁵ cells per well in the 6-well plates and incubated overnight. On days 1, 3 and 5, the medium was changed to serum-free RPMI-1640 containing 10 μM 5-Aza. On days 2 and 4, the medium was changed to 5-Aza free RPMI-1640 with 3% FBS. The cells were harvested for further study on day 5.

Total RNA was extracted from cells using RNAiso Plus (Takara Biotech Co., Ltd., Dalian, China) according to the manufacturer’s protocol. Total mRNA (1 μg) was reverse-transcribed by the M-MLV reverse transcriptase (Invitrogen). Subsequently, aliquots of cDNA were subjected to PCR by using FastStart Universal SYBR Green Master (ROX) (Roche, Basel, Switzerland) on a LightCycler^® 96 Real-Time PCR System (Roche). Primers for mRNA detection of BLU, Sp1 and β-actin (internal control) were as follows: for BLU, 5′-AAC CAGCAGCATGAGAACCT-3′ (forward) and 5′-AGTTTGCG GTGGCAATAGTC-3′ (reverse); for Sp1, 5′-CCTCCAGACC ATTAACCTCAG-3′ (forward) and 5′-TCCACCTGCTGTGT CATCAT-3′ (reverse); for β-actin, 5′-CACAGAGCCTCGCCTT TGCC-3′ (forward) and 5′-CACATGCCGGAGCCGTTGTC-3′ (reverse). The amplification conditions were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 30 sec. Relative fold-change in gene expression was determined using the ΔΔCt method normalized to β-actin.

Genomic DNA was extracted from cells using DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) and then bisulfited using EpiTect Bisulfite kit (Qiagen) according to the manufacturer’s instructions. The sodium bisulfite-modified genomic DNA was subjected to PCR amplification with primers which specially designed to analysis the methylation status of 22 CpG sites within a 244-bp region of the BLU promoter (−181 to +63). The primers were: 5′-ATAAG GATTTGGAGTTTAGGAGAG-3′ (forward) and 5′-CCAAAA TCTAAAACAAAACAATTAC-3′ (reverse). PCR products were purified by Gel Extraction kit (Omega, USA) and cloned into pMD19-T vector (Takara) for DNA sequencing. At least ten clones from each sample were randomly selected and then sequenced.

Cells were lysed in RIPA buffer (Cell Signaling Technology) with the protease inhibitor (Sigma-Aldrich Corp.). Equal amounts of total protein were separated on SDS-PAGE and transferred to the nitrocellulose membranes (Millipore, Billerica, MA, USA). The membranes were blocked with 3% BSA-TBST buffer for 1 h at room temperature and then incubated with appropriate primary antibodies overnight at 4°C. Next, the membranes were washed with TBST buffer three times and incubated with the corresponding HRP-conjugated secondary antibodies for 2 h at room temperature. The specific protein band detection was performed using a commercial ECL kit (Pierce, Rockford, IL, USA). β-actin was used as internal control. The following antibodies were used for western blot analysis: rabbit anti-BLU polyclonal antibody (ab96700, 1:2,000, Abcam, Cambridge, UK); rabbit anti-Sp1 polyclonal antibody (#9389S, 1:1,000, Cell Signaling Technology, Danvers, MA, USA); rabbit anti-β-actin polyclonal antibody (sc-130656,1:1,000, Santa Cruz Biotechnology, Santa Cruz, CA, USA).

The ChIP assay was performed using EZ-ChIP™ Chromatin Immunoprecipitation kit (Millipore). Immunoprecipitation was performed by the addition of anti-Sp1 antibody, and rabbit IgG antibody was used as a negative control. A fragment of the BLU promoter (−80/+108), which contains putative Sp1-binding sites (−43/−35) was amplified by semi-quantitative PCR with primers as follows: 5′-TAGAGACCCGCCCGGAT TTA-3′ (forward) and 5′-GGGTGGGGATGCTGTCACAT-3′ (reverse). PCR products were analyzed by electrophoresis on a 2% agarose gel and stained with ethidium bromide.

Nuclear extracts were isolated from cells using Nuclear Protein Extraction kit (Solarbio, Beijing, China). The EMSA assay was performed by LightShift Chmiluminescent EMSA kit (Pierce) following the manufacturer’s instructions. The 5′ biotin-labeled and double-stranded oligo probes containing putative Sp1 binding sites (−43/−35) were chemically synthesized by Invitrogen. The sequences of probes were as follows: −39C (unmethylated) 5′-CCTCCAGGGGCGGGGCCCAGTTG-3′ and −39 5mC (methylated) 5′-CCTCCAGGGG5mCGGGGCCCAGTTG-3′; unlabeled probes 5′-CCTCCAGGGGCGGGGCCCAGTTG-3′ were synthesized as cold competitor. Electrophoresis mobility gel shift assay was performed as described previously (29).

The pGL3-basic dual luciferase vector (Promega, Madison, WI, USA) was used to produce a series of constructs containing various truncated regions of BLU promoter. Two fragments of interest BLU promoter regions containing the unmethylated and methylated −39 CpG (5′-CCTCCAGGGG CGGGGCCCAG-3′ and 5′-CCTCCAGGGG5mCGGGGCC CAG-3′, respectively) were directly chemically synthesized and massively inserted into the luciferase pGL3-basic vector, and then purified by using Universal DNA purification kit (Tiangen, Beijing, China). To assess promoter activity, cells were transiently co-transfected with BLU promoter constructs and RenillaphRL-TK plasmid (Promega) using Lipofectamine 2000 (Invitrogen). Twenty-four hours later, cells were harvested, and luciferase activities were measured by the Dual-Luciferase Reporter assay kit (Promega) on a TD20/20 Luminometer (Turner Designs, Sunnyvale, CA, USA). Results are expressed as relative luciferase activities, which are normalized to Renilla luciferase activities.

Small interfering RNA (siRNA) sequences (Table I), which target Sp1 and BLU, were chemically synthesized (GenePharma, Shanghai, China). Scrambled siRNA was used as a negative control (NC). Cells were transiently transfected with 100 pmol of siRNA sequences using Lipofectamine 2000 (Invitrogen). After 72-h transfection, the cells were harvested for further experiments.

Cell viability was measured by a cell Counting Kit-8 (CCK-8) (Yeasen, Shanghai, China). After being transfected with or without siRNA as described above, 100 μl SGC7901 and AGS cell suspensions (2×10⁴ cells/ml) were plated into each well of 96-well plates. Then 10 μl CCK-8 solution was added to each well and incubated for 2 h. The absorbance was measured at 450 nm on a microplate reader (MRX, Dynex Technologies, West Sussex, UK).

Two thousand cells were plated in 60-mm culture dishes and incubated at 37°C for 2 weeks; visible cell colonies were fixed with methanol for 15 min, then stained with Giemsa for 1 h, washed with water and colonies containing ≥50 cells were counted under a microscope.

Statistical analysis was performed using the SPSS 19.0 software (SPSS Inc, Chicago, IL, USA). All quantitative data were obtained from at least three independent experiments and represented as means ± SD or SEM. Difference between the means of two groups was compared using Student’s t-tests. Comparisons between clinicopathologic characteristics and expression levels of mRNA in gastric cancer samples were performed with non-parametric tests (Mann-Whitney U test for 2 groups, Kruskall-Wallis test for ≥3 groups). P<0.05 was considered to be statistically significant.

To verify whether inactivation of BLU contributes to gastric carcinogenesis, we first detected BLU mRNA expression in human gastric cancer cells and GES-1 cells. As shown in Fig. 1A, BLU mRNA was significantly lower in MGC803, SGC7901, BGC823, MKN45, AGS and HGC27 cells than that in GES-1 cells (P<0.05, P<0.01, P<0.001, P<0.001, P<0.001 and P<0.001, respectively). Consistently, BLU protein levels were also downregulated in gastric cancer cells (Fig. 1B). In addition, BLU mRNA expression was notably reduced in 69% (36/52) gastric cancer tissues when compared with their paired adjacent non-cancerous tissues (P<0.01, Fig. 1C and D). No significant difference in BLU mRNA level was observed between gastric cancer tissues when classified by various clinicopathologic characteristics (Table II).

To define the BLU promoter, a 2,100-bp sequence, including BLU proximal promoter (−2,000–+1)) and 5′-UTR (+1–+100) regions (RefSeq: NC_000003.12 and NM_015896), was available for analysis. A 251-bp region spanning positions −290 to −40 relative to the transcription start site was predicted to be the putative promoter of the BLU gene using the Promoter Scan program (Fig. 2A). This region contains a GC box, but lacks a TATA box. A potential CpG island spanning positions −574 to +90 was identified by Methyl Primer Express^® software v1.0 (Applied Biosystems) (Fig. 2A).

To evaluate the effects of this sequence on transcription, we generated the luciferase report constructs containing various lengths of the BLU proximal promoter and 5′-UTR region. As presented in Fig. 2B, the fragment (−181 to +63) had the highest activity, while the second largest fragment (−565 to −174) had less activity than the largest and smallest fragments (−565 to +63; −34 to +63, respectively) (Fig. 2B), suggesting that an active cis-element was putatively at position −181 to +63 and a repressive element resides was located between positions −565 to −174 in the BLU promoter. Taken together, the results indicated that the indentified BLU promoter is indeed functional.

Using Methyl Primer Express software (Applied Biosystems), we identified that the 244-bp BLU promoter region (−181 to +63) contains 22 CpG sites (Fig. 3A). Using the transcription factor search program TFSEARCH (http://mbs.cbrc.jp/research/db/TFSEARCH.html), two putative transcription factor Sp1-binding sites were predicted at position −43/−35 and −89/−81 relative to the transcription start site, each of them contains one CpG site (−39 CpG and −85 CpG, respectively). Binding sites for several other potential transcription factor, such as IK-2 (one site), CdxA (two sites), GATA-1 (one site), E2F (one site), MZF1 (one site) were also noted (Fig. 3A). Since the BLU promoter lacks TATA box and the binding sites of transcription factor Sp1 was reported to be crucial for the basal transcription of TATA-less promoters, we then investigated the putative functions of −39 CpG and −85 CpG in the regulation of BLU expression.

We performed bisulfite sequencing (BSP) analysis to investigate whether promoter CpG methylation is associated with the loss of BLU expression. As shown in Fig. 3B, the hypermethylation of BLU was detected in AGS, MKN45, HGC27, BGC823 and SGC7901 cells, while BLU was hypomethylated in GES-1. Importantly, we observed that the methylation level of −39 CpG is significantly higher in the five gastric cancer cell lines with relatively low BLU mRNA expression (Fig. 3B–D). In addition, we also found a negative correlation between −39 CpG methylation level and BLU expression in tumor samples (Fig. 3E and F). Furthermore, we treated AGS and SGC7901 cells with demethylating agent 5-Aza, and found that BLU mRNA and protein levels were restored in AGS and SGC7901 cells (Fig. 3G). These results implied that hypermethylation of BLU promoter may play an important role in transcription silencing of the BLU gene in human gastric cancer.

It is well known that Sp1 is one of the first identified eukaryotic transcription factors (30). Hypermethylation of CpG dinucleotides within Sp1-binding sites could block Sp1 binding to promoter (22,23). Being in silico predicted to be within Sp1-binding sites in BLU promoter, the −39 CpG was hypermethylated in 5 gastric cancer cell lines with low BLU mRNA expression. Thus, we suggest the possibility that hypermethylated −39 CpG may affect the recruitment of Sp1 to BLU promoter. To test this, we performed the following experiments.

First, western blot analysis showed that Sp1 protein expression in SGC7901 cells was nearly equaled with that in AGS cells (Fig. 4A). Subsequently, we performed ChIP assay in AGS and SGC7901 cells. As shown in Fig. 4B, with the DNA samples immunoprecipitated by anti-Sp1 as templates, a 188 bp BLU promoter sequence (−80 to +108 bp) containing −39 CpG was amplified in SGC7901 cells, but not in AGS cells. Considering that the methylation level of −39 CpG in AGS cells was higher than SGC7901 cells (Fig. 3B and C), the results suggested that hypermethylation of −39 CpG significantly inhibited the interaction between Sp1 and BLU promoter.

Next, EMSA assay showed that although both the probe containing unmethylated −39 CpG and the probe containing methylated −39 CpG could form DNA-protein complexes with nuclear extracts of SGC7901 cells, binding ability of unmethylated probe is stronger than the methylated probe (Fig. 4C). This indicated that methylation of −39 CpG within Sp1 binding-sites may reduce the complex formation.

Furthermore, we constructed the luciferase reporter vectors containing the corresponding unmethylated and methylated fragments (Sp1-binding sites) and then transiently transfected them into SGC7901 cells. As shown in Fig. 4D, methylated −39 CpG significantly diminished the luciferase activity by ~70%. Taken together, our results demonstrated that Sp1 could specifically bind with the BLU promoter and hypermethylation of −39 CpG significantly reduced the recruitment of Sp1 to BLU promoter.

To explore whether Sp1 could regulate transcriptional activity of the BLU promoter, we performed the luciferase report assay in SGC7901 cells. As a result, the constructs containing the Sp1-binding site (BP5: −59/+63) showed significant higher activity than the constructs with deletion of Sp1-binding site (BP3: −34/+63) and the mutant constructs (BP6: −59/+63) (Fig. 5A). Furthermore, siRNA-mediated knockdown of endogenous Sp1 effectively decreased the mRNA and protein expression level of BLU (Fig. 5B and C). After co-transfection of the BLU promoter constructs (BP2: −181/+63 and BP5: −59/+63) containing the Sp1-binding sites with si-Sp1, knockdown of Sp1 decreased the luciferase activities of these constructs when compared with the negative control (NC) (Fig. 5D). By contrast, downregulation of Sp1 presented no significant effects on the luciferase activity of the construct (BP6: −34/+63) lacking the Sp1-binding sites. Our results demonstrated that the sequence of the BLU promoter (positions −43 to −35) is a functional Sp1-binding site and Sp1 could activate the BLU promoter.

BLU was frequently downregulated in gastric cancer cells and primary tumors, suggesting that BLU is likely a tumor suppressor in gastric cancer. To investigate the affects of BLU on the growth of gastric cancer cells, we performed CCK-8 and colony formation assays in SGC7901 cells transfected with BLU siRNAs or negative control. As expected (Fig. 6A), BLU was significantly silenced in SGC7901 cells transfected with si-BLU-2 and si-BLU-3. Subsequently, CCK-8 and colony formation experiment showed that BLU knockdown promoted cell viability and clonogenicity (Fig. 6B and C), suggesting that BLU acts as a tumor suppressor via inhibiting cell growth in gastric carcinoma.