Background
Gastric cancer (GC) is one of the leading causes of cancer mortality in the world, particularly in East Asian countries such as China, Japan and Korea, as well as other developing countries. Over the past decades, the overall survival for GC has not significantly improved in spite of improvement in surgical technique and significant advancement of chemotherapy and radiotherapy options [
1]. Therefore, it is important to understand the molecular mechanisms involved in the carcinogenesis of GC.
Loss of heterozygosity (LOH) at specific sites of the cancer genome is considered to embody tumor suppressor genes (TSGs). Frequent LOH at 7q31.1/2 has been detected in many human malignancies including GC [
2]. Recently, we found a high frequency of LOH region on 7q31 in primary GC from China, and identified D7S486 to be the most frequent LOH locus [
3]. This study was designed to explore what TSGs associated with GC were located around D7S486 in this region.
Using microarray technology, a high-throughput single nucleotide polymorphisms (SNP) genotyping system was used to evaluate the LOH status around D7S486 on 7q31 in 75 primary GC samples and to discover possible candidate genes. As a result, TESTIN (TES) showed the potential to be a TSG in GC after initial screening. To clarify its role in GC, we examined TES expression in primary GC and its relationship to clinicopathological characteristics and prognosis. We also examined the effect of TES overexpression on the proliferation of several GC cell lines. Furthermore, mutation and methylation analysis were performed to explore its possible mechanisms of inactivation in GC.
Discussion
In our previous studies [
3,
13,
14], we found that the frequency of LOH at D7S486 was as high as 30.36% in primary GC tissues. We also found that most GC patients with a high frequency of LOH at D7S486 were in their middle to advanced clinical stage (stage III and IV), and usually correlated with lymph node metastasis [
3,
13,
14]. Taken together, these results suggest that one or more tumor suppressor genes associated with gastric carcinomas might situate around D7S486.
In this study, using a high-throughput SNPs microarray fabricated in-house, we performed the SNP LOH assay in 75 paired GC samples. These SNPs are located within genes and ESTs around D7S486 and with a heterozygosity frequency greater than 0.1. As a result, 6 SNPs with a LOH frequency of > 30% were identified. After bioinformatics analysis, 5 corresponding identified genes were determined, including
ST7,
FOXP2,
MDFIC,
TES and
CAV1. Among these genes, no reports have shown any correlations between
FOXP2 and
MDFIC and human malignancy. Although
ST7 [
5] and
CAV1 [
6] have been reported to be candidate tumor suppressor in several types of cancer, no direct evidences support their relationship with gastric carcinogenesis [
5,
14‐
16]. However, many previous studies on
TES [
7‐
11,
17] have demonstrated that it may be a candidate TSG in human cancers.
Human gene encoding TES was first identified by Tatarelli et al. [
9] and localized to the fragile site FRA7G at 7q31.2. So far, down regulation of TES has been reported in many types of human malignancies [
8,
10]. In addition, a profound reduction in growth potential was detected in different cancer cell lines in which
TES was overexpressed [
9,
11]. All these findings suggest that
TES may present a candidate tumor suppressor gene. However, to date, there is no definitive report regarding whether
TES works as a TSG in primary GC.
One of the most notable features of TSG is the mutation that may lead to gene inactivation, especially mutation in exon. Missense mutation has been detected in exons 3 and 5 of
TES in the leukemic cell line CEM and in the ovarian cancer cell line CAOV3 [
10]. However, no mutation was detected in these two exons in our study. A heterozygous mutation at codon 221 in exon 4 of
TES has been reported in the breast cancer cell line MDA-MB 453 [
9] and in head and neck squamous cell carcinoma tissues [
8]. Similarly, the same mutation in exon 4 was also detected in HGC27 GC cells in this study. And a sense point mutation was also found in exon 7 of
TES. However, all the above substitutions were in accordance with the cSNP sites of the
TES gene, which had already been recorded in the NCBI SNP database, suggesting that they were not tumor-specific. Moreover, in exon 6, we found a new heterozygous missense point mutation in a GC tissue sample. Although it has not been reported, the same mutation was also observed in the matched non-tumor tissue. The above findings suggest that gene mutations may not be a mechanism for
TES inactivation in GC.
When LOH occurs in a gene, the expression of the gene may be reduced. Therefore, we then investigated the expression of TES in primary GC. As expected, both transcriptional and translational level of TES were reduced in primary GC tissues, which is consistent with its expression level in other types of human tumors, indicating that the deletion or down regulation of TES might be involved in the occurrence and development of GC. Furthermore, less differentiated tumors exhibited lower TES expression, indicating a significant correlation between TES expression and tumor differentiation. In addition, patients with negative expression of TES had a shorter life span than those with positive expression, suggesting that the detection of TES expression might be helpful to assess prognosis in GC. In this study, we also found for the first time that the introduction of TES caused a significant growth delay in three GC cell lines. This result demonstrated the suppressor effect of TES on GC cells, supporting the idea that TES functions as a tumor suppressor gene in GC, at least in these three cell lines.
However, the mechanism of down regulation of TES expression in GC is not fully understood. Besides LOH causing reduction of gene expression, hypermethylation of CpG islands located in the promoter region of a gene is also frequently correlated with its transcriptional down regulation [
18], and represents an alternative mechanism for the inactivation of TSG [
19]. Several teams have reported on the methylation status of
TES promoter in different tumor types. Tobias [
11] found frequent methylation of the CpG islands at the 5'end of
TES in 7 of 10 ovarian carcinomas and in all 30 tumor-derived cell lines tested, but only discovered a frame-shift mutation in one allele of a breast cancer cell line and polymorphisms in a number of samples. Tatarelli [
10] found the
TES promoter was fully methylated in the majority of 46 tumor-derived cell lines, and only 3 of 46 tumor-derived cell lines showed partial methylation at the promoter of
TES gene. Similarly, no homozygous deletions or a high frequency of convincing somatic mutations was found within the coding region of
TES in a total of 26 cancer cell lines. All of these data demonstrate that
TES is inactivated primarily by transcriptional silencing resulting from CpG island methylation rather than gene mutation.
In this study, looking at 30 primary GC samples with reduced TES expression as determined by immunochemistry, 18 samples were found to be completely methylated at the promoter of TES, 4 partially methylated and 8 were found to be unmethylated. In three GC cell lines with an absence of TES expression, the MGC803 and HGC27 cells displayed complete methylation at the TES promoter, while no methylation was detected in SGC7901 cells. These results suggest that the methylation of CpG in the TES promoter was a frequent event in GC and might be involved in the inactivation of TES.
In this study, we further investigated the effect of DAC, a pyrimidine nucleoside analog that strongly inhibits DNA methyltransferase activity by titrating out DNA methyltransferase activity via a covalent trapping mechanism, on MGC803 and HGC27 GC cell lines, which showed complete methylation at the TES promoter. Interestingly, after DAC treatment, the complete methylation of TES promoter changed to partial methylation and nonmethylation in MGC803 and HGC27 cells, respectively. The mRNA and protein level of TES also changed from negative to positive in both MGC803 and HGC27 cells following DAC treatment. These findings further confirm that methylation of TES promoter plays an important role in TES down regulation in GC.
In conclusion, the present study provides novel and definitive data to support the idea that TES may play an important role in primary GC as a tumor suppressor, and also conforms that, besides LOH, hypermethylation in the promoter of the TES gene rather than gene mutations contributes to its inactivation and to the progression of gastric carcinogenesis.
Methods
Loss of heterozygosity analysis of the genes around D7S486 by SNP genotyping system
Patients and tissue specimens
75 paired fresh GC and matched adjacent non-tumor tissue specimens were obtained from patients underwent surgical resection but without any preoperative treatment in the Sun Yat-sen University Cancer Center between 2004 and 2005. The 75 patients included 54 males and 21 females with a median age of 60 years. After surgical resection, the fresh tissues were immediately immersed in RNAlater (Ambion, USA) and stored at 4°C overnight to allow thorough penetration of the tissues, then frozen at -80°C until RNA and DNA extraction. Both cancer and matched adjacent non-tumor tissues not less than 2 cm away from the GC patients were sampled, respectively, and confirmed by pathological examination. The study was approved by the Ethics Committee of Sun Yat-sen University Cancer Center and informed consent was obtained from each patient.
DNA samples were extracted from 75 paired GC specimens obtained by microdissection using TRIzol kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The procedures for microdissection were performed as previously described [
20].
SNP genotyping and LOH analysis by microarray system
SNP Selection
A computer program written for SNP selection was used as previously described [
4]. SNPs in a 4 Mb region around D7S486 on human chromosome 7q were selected from the dbSNP database
ftp://ftp.ncbi.nih.gov/snp/organisms/human_9606/chr_rpts/ maintained by National Center for Biotechnology Information (NCBI, Build 36.3). To ensure that the selected SNPs were real and suitable for the multiplex system, a series of criteria described before was used for selection [
21].
Design of primer and probe
To select sequence frames for primers, a computer program described [
4] before was used. Within a user-defined sequence range around the polymorphic sites (150 bp in the present study), the candidate sequence frames were first selected according to a user-defined melting temperature range (55°C to 75°C in this application). To minimize primer-primer interactions, further selection was performed on qualified frames based on a series of criteria described before [
4].
Multiplex PCR and ssDNA preparation
The procedures for multiplex amplification were performed following the method of Wang
et al. [
4] with minor modifications. In brief, the first multiplex PCR reaction was performed in 25 μl of PCR mix containing 2.5 μl of 10 × PCR buffer (50 mM KCl, 100 mM Tris-HCl at pH 8.3, 1.5 mM MgCl
2, and 100 μg/ml gelatin), 0.5 μl of 10 nM dNTPs, primer mix (300 nM each) for all SNPs in the multiplex group, 5 units of HotStart
Taq DNA polymerase, and 200 ng of DNA. The PCR cycling conditions were: 95°C (15 min) for 1 cycle, 94°C (40 sec), 55°C (2 min), ramping from 55°C to 72°C (5 min) for 40 cycles and a final extension of 72°C (10 min). ssDNA was generated in both directions using the same conditions for multiplex PCR with the following modifications: (1) 1.0 μl of product from the multiplex PCR was used as templates, (2) only one primer (one of the primer-probes) for each SNP was used, and (3) 45 PCR cycles were performed.
SNP genotype determination by microarray
In the SNP genotyping system [
4] adopted in this paper, after generating single-stranded DNAs (ssDNA), they were hybridized to the probes on a microarray fabricated in-house. The probes were designed in such a way that their 3' ends were adjacent to the polymorphic sites in the hybridizing ssDNA. In this way, the probes could be labeled with the commonly used single-base extension method [
22‐
24], during which single dideoxyribonucleotides (ddNTPs) were added to the probe in an allele-specific manner that was dependent on the hybridizing allelic sequence(s). When the corresponding ddNTPs were labeled with different fluorescent chromophores (cyanine dyes, either Cy3 or C5, in our system), the allelic state of the SNPs were determined by analyzing the amount of incorporated fluorescence with the written program "AccTyping" [
25]. In addition, SNP genotypes could be determined independently with the two DNA strands as separate templates so that results from such dual-probe analysis could be compared to ensure a high degree of accuracy.
Expression of TES in primary GC samples and its relationship to clinicopathological characteristics
140 paraffin-embedded GC samples were retrieved according to the 1999-2001 surgical pathology files in the Sun Yat-sen University Cancer Center, which included the patients without pretreatment. The 140 patients included 95 males and 45 females with a median age of 56 years. All tissue blocks were cut into serial 4 μm thick sections. The histological types were assigned according to WHO classification criteria.
Total RNA was extracted from 140 patients with primary GC using TRIzol solution (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol, and RNAse-free DNase I was used to remove DNA contamination. Reverse transcription (RT) was performed with 2 μg total RNA using M-MLV Reverse transcriptase (Promega, Madison, WI, USA) to synthesize first-strand cDNA according to the manufacturer's recommendation, followed by cDNA amplification using the specific primer set for TES and the GAPDH primer set used as an internal control. Primers used in this study were as follows: 5'-ACTGTGGCAGACATTACTGTGACA-3' (sense) and 5'-GATAGCTATGGCTCGATACTTCTGGGTGC-3' (antisense) for TES; 5'-CGGGAAGCTTGTCATCAATGG-3' (sense) and 5'-G GCAGTGATGGCATGGACTG-3' (antisense) for GAPDH, and the corresponding PCR products were 440 bp and 358 bp, respectively.
Western blot and immunohistochemistry analysis was carried out as our previously described [
16]. Primary polyclonal antibody against TES (Santa Cruz, CA, USA) was used at 1:500 and 1:100 dilutions for Western blot and immunohistochemistry, respectively. As our previously described [
16], the total TES immunohistochemical staining score was calculated as the sum of the percent positivity of stained tumor cells and the staining intensity.
Overexpression of TES and its effect on proliferation of GC cells
Three GC cell lines used in this study were MGC803, SGC7901 and HGC27. All cell lines were obtained from the Committee of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China) and preserved in our lab, and grown in RPMI 1640 supplemented with 10% FBS (fetal bovine serum) and antibiotics (50 μg/ml each of penicillin, streptomycin and gentamicin) at 37°C in a humidified 5% CO2 atmosphere.
Plasmid pOTB7 containing full-length TES cDNA were purchased from Invitrogen (Carlsbad, CA, USA). The TES cDNA was amplified with the following primers: upstream 5'-GCAAGCTTCATGGACCTGGAAAAC-3', downstream 5'-GCGGATCCCTAAGACATCCTCTT-3'. The 1283 bp full length TES cDNA PCR product was recovered with the QIAquick Gel Extraction Kit (Qiagen Inc., Valencia, CA). The TES cDNA was first cloned into a pGEM-T and then into a pEGFP-C2 vector to create the pEGFP-C2/TES eukaryotic expressing vector in which the TES cDNA was fused with the downstream of EGFP cDNA.
Three gastric cell lines were transfected with pEGFP-C2/TES or empty pEGFP-C2 plasmid, respectively, using LipofectamineTM 2000 reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Briefly, cells were seeded into a 75 cm2 culture flask, and 12-24 μg plasmid and 36-60 μl Lipofectamine 2000 reagent were used for transfection when cells were at 90% confluence. Cells were harvested 36 h after transfection and tested in the proliferation assay or harvested for RT-PCR detection.
Measurement of proliferation was determined by the MTS assay. Briefly, GFP-positive cells were sorted 36 h after transfection with BD Aria II(BD) and then seeded into 96-well plates (1000 cells per well) in 100 μl RMPI 1640 medium for five replicates. The next day, after 20 μl CellTiter 96 Aqueous One Solution (Promega, Madison, WI, USA) were added into the medium, the cells were incubated at 37°C for 4 h, and the absorbance at 490 nm (OD490) was measured daily for 7 days.
Mutation analysis of TES gene
Each exon of TES was PCR amplified respectively. The PCR mixture contained 1 μg genomic DNA, 20 μl 2 × Pfu PCR MasterMix (TIANGEN Biotech, Beijing, China) and 0.1 pmol/μl of each primer in a total 40 μl volume. The PCR cycling conditions for exons 2 to 7 were: 94°C (2 min) for 1 cycle, 94°C (30 sec), 55°C (40 sec), 72°C (2 min) for 40 cycles and a final extension of 72°C (10 min). Exon 1 PCR conditions were as described above with the exception that the initial denaturation was performed at 97.6°C for 5 min, and the denaturation temperature in the 40 cycles was also set at 97.6°C. For identification of mutations, PCR products were purified and then sequenced on an ABI Prism 3700 automated DNA Analyzer. All mutations were confirmed by sequencing in both directions.
Methylation analysis of TES promoter in primary GC
Three GC cell lines used in this study were MGC803, SGC7901 and HGC27 with the same culture condition as described above. 10 μM of 5-aza-2'-deoxycytidine (decitabine; DAC) was added when cell density reached 75%. Culture medium was changed with fresh RPMI1640 with the same concentration of DAC on alternate days. PBS (10 μM) was used as a negative control. Cells were harvested for the experiment six days later.
The genomic DNA isolated with the DNeasy Tissue Kit (Qiagen Inc., Valencia, CA) was modified by bisulfite treatment with the CpGenome DNA Modification Kit (Chemicon) and amplified by PCR with two sets of TES promoter specific primer pairs (M-TES sense 5'-TATTGAGTTTGTTTAGTAGGGCGTC-3', M-TES antisense 5'-AATAACAACCGAACAACTCCG-3', PCR product 133 bp; U-TES sense 5'-TGAGTTTGTTTAGTAGGGTGTTG-3', U-TES antisense 5'-ATAACAACCAAACAACTCCAA-3', PCR product 129 bp) that recognize either the methylated or unmethylated CpG sequences and then analyzed by electrophoresis.
Statistical analysis
Quantitative values were expressed as the means ± SD or median (range). The Quantity One software (Bio-Rad Laboratories, Inc. Hercules, CA, USA) was used to quantify the densities of the bands in RT-PCR and Western blot. Paired-samples t-test was used to compare mRNA and protein expression of TES in GC with paired adjacent non-tumor tissue samples. The χ2 test for proportion and Spearman's correlation was used to analyze the relationship between TES expression and various clinicopathological characteristics. Survival curves were calculated using the Kaplan-Meier method and compared by the log-rank test. The SPSS 15.0 software (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses and p < 0.05 was considered significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
DW, HM, YC, WH, and KP participated in the acquisition of data. ZZ, HW, and JX were involved with the study concept and design. DW, HM, JS, HW, and QW contributed to the statistical analysis. HM, DW, JX participated in manuscript preparation. All authors participated in the interpretation of results and critical revision of the manuscript for important intellectual content. All authors have read and approved the final manuscript.