Genomic loss and epigenetic silencing of very-low-density lipoprotein receptor involved in gastric carcinogenesis

Article metrics

Abstract

Homozygous loss in the genomic sequence, a mechanism for inactivating tumor-suppressor genes (TSGs) in cancer, has been used as a tag for the identification of novel TSGs, and array-based comparative genomic hybridization (array-CGH) has a great potential for high-throughput identification of this change. We identified a homozygous loss of the very-low-density lipoprotein receptor (VLDLR) gene (9p24.2) from genome-wide screening for copy-number alterations in 32 gastric cancer (GC) cell lines using array-CGH. Although previous reports demonstrated mRNA or protein expression of VLDLR in various cancers including GC, the association between genomic losses or epigenetic silencing of this gene and carcinogenesis has never been reported before. Homozygous deletion of VLDLR was also seen in primary GCs, albeit infrequently, and about half of GC cell lines showed lost or reduced VLDLR expression. The VLDLR expression was restored in gene-silenced GC cells after treatment with 5-aza 2′-deoxycytidine. According to methylation analyses, hypermethylation of the VLDLR promoter region, which all of GC lines without its expression showed, occurred in some primary GCs. Restoration of VLDLR type I expression in GC cells reduced colony formation. These results suggest that not only the expression of VLDLR but also genetic or epigenetic silencing of this gene may contribute to tumor formation and be involved in gastric carcinogenesis.

Introduction

Very-low-density lipoprotein receptor (VLDLR) is a member of the low-density lipoprotein receptor (LDLR) superfamily (Strickland et al., 1995). VLDLR is highly similar to LDLR in structure, but is different from LDLR in function (Strickland et al., 1995). VLDLR contains five domains: an amino-terminal ligand-binding domain comprised of multiple cysteine-rich repeats, an epidermal growth factor (EGF) precursor homology domain, an O-linked sugar domain with a cluster of serine and threonine residues, a transmembrane domain and a cytoplasmic domain with an NPXY sequence (Chen et al., 1990). VLDLR has two isoforms: the full-length version, type I, and a version lacking an O-linked sugar region, type II (Sakai et al., 1994). VLDLR type I is mainly distributed in heart and skeletal muscles with active fatty acid metabolism, whereas VLDLR type II is predominant in non-muscle tissue, including kidney, spleen, adrenal gland, lung, brain, testis, uterus and ovary, except liver (Takahashi et al., 1992; Webb et al., 1994). Differences in ligand specificity and tissue distribution suggest that the VLDLR isoforms play distinct roles in various tissues and cells. Previous reports have demonstrated that VLDLR isoforms were also expressed in a variety of cancer tissues (Martensen et al., 1997; Nakamura et al., 2000), including gastric cancer (GC) (Chen et al., 2005), suggesting that the presence of VLDLR is likely to be associated with the pathogenesis of various cancers. VLDLR may serve as an energy source for the rapid growth of cancer cells or may contribute to tumor formation through ligands such as urokinase-plasminogen activator (uPA) and uPA-plasminogen activator inhibitor-1 complex (Webb et al., 1999). On the other hand, a recent report revealed that the antiproliferative activity of tissue factor pathway inhibitor, which is a potent inhibitor of endothelial proliferation, is mediated through interaction with VLDLR (Hembrough et al., 2001), and so the effects of VLDLR protein expression on the neoplastic process in various tissues seem to be controversial.

Recently, we had searched the GC-related genes using array-based comparative genomic hybridization (array-CGH) analyses relied on a custom-made bacterial artificial chromosome (BAC)-based array (Takada et al., 2005a, 2005b). In the study reported here, we identified a homozygous loss of VLDLR (9p24.2) from one of the homozygously deleted regions detected in the course of a program to screen a panel of GC cell lines. Homozygous deletion of VLDLR was observed in primary GC as well as GC cell lines, albeit infrequently. In addition, expression of this gene was also lost in some GC cell lines without homozygous deletion through aberrant hypermethylation of the promoter, although the gene was expressed in normal stomach. The results indicated that VLDLR appeared to be inactivated genetically or epigenetically during the development and/or progression of some GCs, suggesting that a lack of or reduction in the expression of this gene plays an important role in gastric carcinogenesis.

Results

Identification of target gene(s) involved in the homozygous deletion at 9p24.2–24.3 by array-CGH analysis

We had already performed genome-wide screening for DNA copy-number aberrations in a panel of 32 GC cell lines using an in-house BAC-based array (MCG Cancer Array-800 and MCG Whole Genome Array-4500; Inazawa et al., 2004), and had reported genes involved in gastric carcinogenesis (Takada et al., 2005a, 2005b). In the present study, we focused on the homozygous deletion at 9p24.2–24.3 that had been identified in HSC58 cells by the previous array-CGH analyses (Figure 1a). Genome profiling in the HSC58 cell line revealed two independent homozygous deletions: at 9p21.2–21.3, which harbors representative tumor-suppressor genes such as CDKN2A/p16, MTAP and TEK, and at 9p24.2–24.3, which contains 12 possible target genes according to information archived by human genome databases (http://www.ncbi.nlm.nih.gov/ and http://genome.ucsc.edu/) within an approximately 3-Mb region estimated from array-CGH data (Figure 1b). Genomic polymerase chain reaction (PCR) analysis validated the complete loss of DNA from HSC58 in 10 of 12 genes, refining the extent of the homozygous deletion expected (Figure 1c and d): CBWD1 and RFX3 were retained (gray bars in Figure 1d). Further genomic PCR analysis using all 32 GC cell lines revealed that HSC58 is the only cell line having a homozygous deletion among those genes including VLDLR (1/32, 3.1%; Figure 2a and data not shown).

Figure 1
figure1

(a) Representative duplicate array-CGH image of the HSC58 cell line. A remarkable decrease in the copy-number ratio (log 2 ratio) of RP11-48M17 at 9p24.2–24.3 was detected. (b) Representative copy-number profiles of chromosome 9 in HSC58 cells. Vertical gray and black bars indicate candidate spots showing patterns of homozygous deletion at 9p24.2–24.3 and 9p21.2–21.3, respectively. (c) Images from genomic PCR experiments showing GAPDH, the functional control and representative genes that were harbored around the homozygous deletion at 9p24.2–24.3 using some GC cell lines including HSC58. PLC, normal peripheral lymphocytes. (d) Map of 9p24.2–24.3 covering the region homozygously deleted in the HSC58 cell line. BACs spotted on the array are shown by vertical white bars. The homozygously deleted region in HSC58 cells, as determined by genomic PCR, is indicated as a vertical white closed arrow. Thirteen genes located around this region, which are homozygously deleted or retained in the HSC58 cell line, are indicated as black or gray bars, respectively, that show positions and directions of transcription.

Figure 2
figure2

(a) Homozygous deletions of VLDLR in one GC cell line (arrowhead), detected by genomic PCR analysis. One: MKN1; two: MKN7; three: MKN28; four: MKN45; five: MKN74; six: KATO-III; seven: OKAJIMA; eight: OCUM-1; nine: HSC39; 10: HSC40A; 11: HSC41; 12: HSC42; 13: HSC43; 14: HSC44PE; 15: HSC45; 16: HSC57; 17: HSC58; 18: HSC60; 19: HSC64; 20: NUGC-2; 21: NUGC-3; 22: NUGC-4; 23: RERF-GC-1B; 24: AZ-521; 25: SNU216; 26: SNU484; 27: SNU601; 28: SNU638; 29: SNU668; 30: SNU719; 31: SH101P4 and 32: Takigawa. (b) Expression of VLDLR in GC cell lines and normal stomach, detected by RT–PCR analysis. An arrowhead indicates the cell line with the homozygous deletion indicated in (a). Note that nine of the 31 cell lines without a homozygous deletion of VLDLR (29.0%) showed almost a complete silencing of this gene and seven (22.6%) showed an apparently decreased expression compared with normal stomach. (c) Homozygous deletion of VLDLR, detected by genomic PCR analysis in one of the 39 LCM-treated primary gastric tumors examined (arrowhead). (d) Validation of genomic losses at the 9p24.2–24.3 region in 195T. Note that ANKRD15 and VLDLR were homozygously deleted but CBWD1 and RFX3 were retained in 195T. (e) Schematic map of a total of four CpG islands identified using the genome database (http://www.ebi.ac.uk/emboss/cpgplot/). Four vertical closed arrows indicate 551-, 994-, 297- and 274-bp CpG islands, respectively (−471 to +80; CpG-I, +269 to IVS +793; CpG-2, IVS +826 to IVS +1122; CpG-3, IVS +1172 to IVS +1445; CpG-4). Exon 1 is indicated by an open box, and the transcription start site is marked at +1. The fragments examined in a promoter assay (Fragments 1–3) are indicated by heavy black lines. The regions examined in the bisulfite–PCR analysis and bisulfite-sequencing (BS-1, and BS-2) are indicated by horizontal gray bars. For the bisulfite–PCR analysis, restriction sites for HhaI are indicated by black downward arrowheads. Arrows at the bottom show the positions of primers for MSP. CpG sites are indicated by vertical ticks on the expanded axis.

Expression of several possible cancer-related target genes in the homozygous deletion at 9p24.2–24.3

The homozygously deleted region identified in the HSC58 cell line at 9p24.2–24.3 contains some genes, which have been reported as cancer-related genes, including VLDLR (Martensen et al., 1997; Nakamura et al., 2000), ANKRD15 (Sarkar et al., 2002) and SMARCA2 (Reisman et al., 2003). Therefore, we next determined expression levels of VLDLR, ANKRD15 and SMARCA2 with other genes within the homozygously deleted region by means of reverse transcription (RT)-PCR in all 32 GC cell lines and in normal stomach. The HSC58 cell line as well as nine lines without homozygous loss within 9p24.2–24.3 (29.0%) completely lacked the expression of VLDLR mRNA, which comprises two isoforms divided into a PCR product of 369 bp showing type I (full length) and that of 285 bp showing type II (O-linked sugar domain lacked), and seven lines (22.6%) showed an apparently reduced expression compared with the normal stomach (Figure 2b). On the other hand, the RT–PCR product of SMARCA2 was absent less frequently, and that of KIAA0020 was not affected at all (Figure 2b). Although the frequency with which the expression of ANKRD15 was lost or reduced was similar to that for VLDLR (Figure 2b), the previous study had already reported that epigenetic methylation at CpG sites in the ANKRD15 gene led to the downregulation of its expression, resulting in loss of function. Therefore, in this study, we noted the silenced expression of VLDLR in GC and tried to solve the association between genetic or epigenetic silencing of this gene and gastric carcinogenesis.

Homozygous deletion of VLDLR in laser-capture microdissection -treated primary tumors

To confirm that the homozygous loss of VLDLR was not an artifact that arose during establishment of the cell lines, we performed genomic PCR using laser-capture microdissection (LCM)-treated primary gastric tumors, and detected a homozygous deletion of VLDLR in one of the 39 tumors examined (2.6%, Figure 2c and d).

Effect of demethylation by 5-aza 2′-deoxycytidine on VLDLR expression

Aberrant methylation of DNA in 5′ regulatory regions harboring a greater than expected number of CpG dinucleotides is a key mechanism by which genes relevant to cancer initiation and progression can be silenced (Baylin et al., 1998), and the genomic sequence of VLDLR harbors CpG-rich regions (CpG islands) around exon 1 (Figure 2e). To investigate whether demethylation could restore the expression of VLDLR mRNA, we treated GC cells lacking VLDLR expression (SNU601 and MKN74) with 10 μ M of 5-aza 2′-deoxycytidine (5-aza-dCyd), a methyltransferase inhibitor, for 5 days. The expression of type I and type II VLDLR mRNA was restored in both cell lines after treatment with 5-aza-dCyd, although mRNA of type II showed higher expression compared to that of type I (Figure 3a). However, we did not observe elevated levels of mRNA in the cells after treatment with trichostatin A (TSA), a histone deacetylase inhibitor, alone, and enhancement of the expression by 5-aza-dCyd given along with TSA, indicating that histone deacetylation plays less of a role in the transcriptional silencing of VLDLR among methylated GC cells.

Figure 3
figure3

(a) Representative results of RT–PCR analysis to reveal VLDLR expression in SNU601 and MKN74 cells with or without treatment with 5-aza-dCyd (10 μ M) for 5 days and/or TSA (10 ng/ml) for 24 h. (b) Promoter activity of the region with and/or without the VLDLR CpG-1. pGL3 basic empty vectors (mock) and constructs containing one of three different sequences upstream or around exon 1 of VLDLR (Fragments 1–3) were transfected into SNU601 and MKN74 cells. Luciferase activities were normalized versus an internal control. The data presented are the means±s.e. of three separate experiments, each performed in triplicate. (c) Representative results of COBRA of VLDLR BS-2 in GC cell lines after digestion with HhaI. Arrows show fragments specifically restricted in the sites recognized as methylated CpGs. (d) Results of bisulfite-sequencing of the VLDLR CpG-1, examined in VLDLR-expressing cell lines (+) and VLDLR-non-expressing cell lines (−). Using two fragments (BS-1 and -2), all of the 71 CpG sites were sequenced. Each square indicates a CpG site: open squares, unmethylated; solid squares, methylated.

Promoter activity of the VLDLR CpG islands

Hypermethylation in the CpG-rich promoter or exonic regions is strongly associated with the transcriptional silencing of TSGs. A total of four regions (CpG-1 to -4 in Figure 2e) around exon 1 of VLDLR were identified as CpG islands by using the genome database (http://www.ebi.ac.uk/emboss/cpgplot/). In order to determine whether typical promoter activity is observed in the CpG island located in the 5′ region of this gene, we first tested the promoter activity of sites with and/or without CpG-1 using 1005-, 655- and 360-bp fragments around exon 1 (−514 to IVS +22, −514 to +141 and +132 to IVS +22, respectively; Fragments 1–3; Figure 2e) in VLDLR-non-expressing (SNU601 and MKN74) and -expressing cells (SH101P4). A remarkable increase in transcriptional activity was observed in constructs containing Fragment 2 covering CpG-1 regardless of the expression status for VLDLR (114.6-, 57.0- and 185.0-fold increases in SNU601, MKN74 and SH101P4 cells, respectively). On the other hand, constructs containing Fragment 3, which does not contain CpG-1, showed much less transcriptional activity (9.1-, 6.4- and 31.0-fold increases, respectively), whereas constructs containing Fragment 1, which includes Fragments 2 and 3, revealed intermediate activity (Figure 3b). Therefore, the region around Fragment 2 (CpG-1) seems to be the most important sequence for promoting expression of VLDLR.

Methylation status of the VLDLR CpG island in GC cell lines and primary tumors

To explore the potential role that the methylation of CpG islands plays in the silencing of the transcription of VLDLR, we first assessed the methylation status of CpG-1 in GC-derived cell lines with or without expression of this gene, by means of combined bisulfite restriction analysis (COBRA) (Xiong and Laird, 1997) using HhaI for each of the PCR fragments. GC cells lacking VLDLR expression without a homozygous deletion of this gene (MKN28, MKN74, KATO-III, SNU601, SNU638 and SNU719) were found to be almost aberrantly hypermethylated, whereas VLDLR-expressing GC cells (MKN45, HSC41, NUGC-3, AZ-521, SNU216, SNU484 and SH101P4) were mainly hypomethylated (Figure 3c). Bisulfite-sequencing was performed to assess the methylation status of each CpG dinucleotide for BS-1 and BS-2 containing VLDLR CpG-1 in more detail (Figure 2e). CpG sites on the CpG-1 tended to be extensively methylated in VLDLR-nonexpressing cells without homozygous deletions (SNU601 and MKN74), whereas almost all CpG sites were unmethylated in VLDLR-expressing cells (AZ-521, Figure 3d).

As hypermethylation of the promoter region around exon1 would likely be associated with silencing of VLDLR expression in GCs, we performed methylation-specific PCR (MSP) with primer sets targeting the most frequently methylated sequence in primary GCs as well as the rest of our GC cell lines without homozygous loss of VLDLR (Figure 2e). We confirmed that only methylated products were observed in two representative cell lines without VLDLR expression (No. 5; MKN74 and No. 27; SNU601, Figure 4a), whereas one hypomethylated VLDLR-expressing cell line produced unmethylated products (No. 24; AZ-521, Figure 4a). The target sequence was methylated in 11 (35.5%) of the 31 GC cell lines (Figure 4a). Moreover, we detected its methylation in seven of 20 primary GC tissues (35.0%); five of the seven samples were not methylated in the corresponding non-cancerous stomach tissues (Figure 4b), although we found no significant relationship between the methylation status of the VLDLR promoter and clinicopathological features such as age, gender, histological subtype, stage and prognosis (data not shown). To validate the methylation status of primary tumors in a more quantitative manner, we also performed COBRA in seven cases with methylated products in tumor tissues by MSP. As expected, all cases were more or less methylated in GC tissues (Figure 4c). In addition, bisulfite-sequencing of BS-2 in cases 17 and 21 confirmed the aberrant hypermethylation in GC tissues, although hypomethylated alleles were also sequenced probably owing to the normal tissue components included in tumor tissues, whereas the corresponding non-cancerous tissues were hypomethylated in this region (Figure 4d). These findings indicate that the methylation of the VLDLR promoter region is not an artifact of passages of GC cell lines in vitro, but rather may be a cancer-related event during the pathogenesis of GC. To further establish tumor-suppressive activity of VLDLR, we performed mutation analysis for the VLDLR coding region using two representative GC cell lines (MKN7 and MKN45) and 13 primary gastric tumors that showed no methylation in MSP analysis. However, we detected no mutations, although two single-nucleotide polymorphisms were found in primary tumors: both are silent substitution from C to T in the coding region of exon 10 at base 1458 and exon 5 at base 468 in 11T and 12T, respectively.

Figure 4
figure4

(a) Representative results of MSP analysis of the VLDLR CpG-1 in GC cell lines. Number shows each GC line, as described in Figure 2a. Parallel amplification reactions were performed using primers specific for unmethylated (U) or methylated (M) DNA. (b) Representative results of MSP analysis of the VLDLR CpG-1 in primary GC tissues (T) and corresponding non-cancerous stomach tissues (N). (c) Representative results of COBRA of VLDLR BS-2 after digestion with HhaI in seven primary cases of GC, which revealed methylation in tumor tissues by mean of MSP analysis. Arrows show fragments specifically restricted at the sites recognized as methylated CpGs. (d) the methylation status of VLDLR BS-2 was determined by bisulfite sequencing in two representative paired GC samples. N, normal tissue; T, tumor tissue.

Suppression of cell growth after restoration of VLDLR expression

To gain insight into the potential role of VLDLR protein in gastric carcinogenesis, we finally investigated whether restoration of VLDLR expression would suppress growth of GC cells in which the gene had been silenced. We performed colony formation assays using each of a full coding sequence (type I) or a sequence lacking O-linked sugar domain (type II) of VLDLR cloned into a mammalian expression vector. As shown in Figure 5, 3 weeks after transfection and subsequent selection of drug-resistant colonies, the numbers of large colonies produced by VLDLR type I-transfected SNU601 cells significantly decreased compared to cells containing empty vector (P<0.05, the Mann–Whitney U-test), whereas those by VLDLR-type II-transfected SNU601 cells were less decreased.

Figure 5
figure5

Effect of restoration of VLDLR on growth of GC cells. COOH-terminally 3 × Myc-tagged constructs containing VLDLR type I/II (pCMV-3Tag-4-VLDLR-type I/II) or empty vector (pCMV-3Tag-4-mock) as a control were transfected into SNU601 cells, which lack expression of the VLDLR gene because the CpG island is methylated. Western blot analysis using anti-Myc antibody demonstrated that cells were transiently transfected with pCMV-3Tag-4-VLDLR-type I/II expressed Myc-tagged VLDLR protein (arrow in upper panel). Three weeks after transfection and subsequent selection of drug-resistant colonies in six-well plates, the colonies formed by VLDLR-type I/II-transfected cells were less numerous than those formed by mock-transfected cells (middle panel). Lower panel, quantitative analysis of colony formation. Colonies larger than 1.5 mm were counted, and results are presented as the means±s.e. (bars) of three separate experiments, each performed in triplicate. The relative number of colonies of VLDLR-type I-expressing cells were decreased compared to that of the mock control (P<0.05, the Mann–Whitney U-test).

Discussion

Array-CGH analysis allows high-throughput and quantitative analysis of copy-number changes at high resolution throughout the genome, providing many advantages over conventional methods. We had already performed a genome-wide screening for DNA copy-number aberrations in a panel of 32 GC cell lines as well as various cancer lines using a custom-made BAC-based array and had successfully reported genes involved in carcinogenesis (Sonoda et al., 2004; Tanami et al., 2005; Takada et al., 2005a, 2005b). In the present study, we analysed a homozygously deleted region at 9p24.2–24.3 identified in previous array-CGH analyses (Takada et al., 2005b), and identified VLDLR as a possible target for silencing in GC through either a genetic or epigenetic mechanism among genes located within this region.

Homozygous loss and loss of heterozygosity at 9p occur at various frequencies in a number of cancers including GC (Sakata et al., 1995; Kim et al., 1997; Yamano et al., 2000). The short arm of chromosome 9 contains CDKN2A/p16, which is the TSG inactivated by homozygous loss, hemizygous loss with point mutation or an epigenetic mechanism such as aberrant methylation of the promoter in human cancers including GC (Shim et al., 2000; Vo et al., 2002). Recently, in addition to this gene, homozygous deletions at 9p24 were reported in lung cancer, breast cancer and pancreatic carcinoma cell lines (An et al., 1999; Heidenblad et al., 2004; Sato et al., 2005). Several candidate TSGs, such as ANKRD15 in renal cell carcinoma (Sarkar et al., 2002) and SMARCA2 in lung cancer (Reisman et al., 2003), were identified as targets for this 9p24.2–24.3 region. Among other genes located within this region, VLDLR has been also reported to be involved in the pathogenesis of GC and other cancers (Martensen et al., 1997; Nakamura et al., 2000; Chen et al., 2005). However, the association between genomic loss or epigenetic silencing of the VLDLR gene and gastric carcinogenesis has not been reported before. In this study, homozygous deletion of VLDLR was detected in HSC58 cells and RT–RCR showed that the expression of VLDLR mRNA was lost or reduced in several GC cell lines without a homozygous deletion of this gene (16/31; 51.6%). GC cell lines without VLDLR expression mostly showed aberrant methylation of the promoter. In addition, several primary gastric samples were methylated in cancerous tissue, but not in non-cancerous stomach tissue. These results suggest that aberrant hypermethylation of the VLDLR promoter may not be a rare event in GC and is likely to play a critical role in gastric carcinogenesis.

VLDLR has been considered to play roles in lipid metabolism through binding to ligands such as apolipoprotein E-rich VLDL, chylomicrons and lipoprotein lipase (Takahashi et al., 1992, 1995). VLDLR has been also shown to modulate cell migration and foam cell formation, suggesting a role for the VLDLR in vascular pathology associated with intimal thickening and atherothogenesis (Suzuki et al., 1995; Hiltunen et al., 1998). In addition, VLDLR is important for the developing human brain and binding of reelin to the VLDLR contributes to the correct regulation of neuronal migration (D'Arcangelo et al., 1999). Despite those known biological roles, the differential effects of VLDLR itself or each VLDLR isoform on carcinogenesis remain poorly understood. According to a previous report, RT–PCR analysis for VLDLR expression demonstrated that the splice variation giving rise to the two forms of VLDLR mRNA was highly cell specific, and human breast carcinomas expressed predominantly or exclusively the VLDLR type II (Martensen et al., 1997). Another report showed that VLDLR type II predominated in the expression of AGC cell line and gastric adenocarcinoma by means of RT–PCR (Nakamura et al., 2000). In both reports, immunohistochemistry demonstrated that VLDLR was mainly expressed in epithelial cancer cells, although some kinds of cancer showed no expression. Consistent with those results, our RT–PCR analysis also revealed that GC cell lines predominantly expressing VLDLR type II were observed more frequently than those expressing type I, whereas some GC cell lines did not express VLDLR at all. These data suggest that VLDLR type II expression is likely to play a role in the pathogenesis of epithelial tumors. On the other hand, another report has showed that cell growth was inhibited by overexpression of VLDLR in a ligand-independent manner, when COS-7 cells were transfected with an expression plasmid containing cDNA of this gene (Wada et al., 2000). The O-linked glycosylation region appeared to be responsible for this growth-inhibiting activity (Wada et al., 2000). Consistent with these data, our colony formation assay showed that the numbers of large colonies produced by VLDLR-type I-transfected SNU601 cells significantly decreased compared to cells containing empty vector, whereas those by VLDLR-type II-transfected SNU601 cells were less decreased. In this regard, a loss or decrease of VLDLR, especially type I form, is likely to affect cell growth, resulting in an advantage for tumor formation. Very recently, in addition, Sato et al. (2006) have reported that reelin is frequently silenced in pancreatic cancers and that siRNA-mediated knockdown of this gene as well as its downstream components, including VLDLR, ApoER2 and DAB1, resulted in increased cell motility of pancreatic cancer cells. These findings support our hypothesis that VLDLR may have potential for a tumor suppressor. Therefore, epigenetic silencing by aberrant hypermethylation of the promoter as well as genetic loss of the VLDLR gene seems to be involved in carcinogenesis including that of GC.

Materials and methods

Cell lines and primary tumors

A total of 32 GC cell lines were employed. Features of the cell lines were described in previous reports (Fukuda et al., 2000; Takada et al., 2005a). All cell lines were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum and 100 U/ml penicillin/100 μg/ml streptomycin.

Primary tumor samples were obtained from 59 patients who were undergoing surgery at Kyoto Prefectural University of Medicine in Kyoto, or the National Cancer Center Hospital in Tokyo, Japan, with prior written consent from each patient in the formal style and after approval by the local ethics committees. Samples from 39 patients with well-differentiated adenocarcinomas were embedded in paraffin for LCM after their fixation in methanol for 24 h, as described elsewhere (Noguchi et al., 1997). From the other 20 patients, we obtained paired samples from cancerous and adjacent non-cancerous tissues, which were frozen immediately in liquid nitrogen and stored at −80°C until required. None of the patients had been administered preoperative radiation, chemotherapy or immunotherapy.

Array-CGH analysis

Hybridizations were carried out as described elsewhere (Sonoda et al., 2004; Takada et al., 2005a) with minor modifications. Hybridized slides were scanned with a GenePix 4000B (Axon Instruments, Foster City, CA, USA). Acquired images were analysed with GenePix Pro 4.1 imaging software (Axon Instruments). Fluorescence ratios were normalized so that the mean of the middle third of log 2 ratios across the array was zero. Average ratios that deviated significantly (>2 s.d.) from zero were considered abnormal.

Screening of GC cell lines and primary tumors for homozygous deletions by genomic PCR

Methanol-fixed, paraffin-embedded tissues were prepared for LCM with a PixCell II LCM system (Arcturus Engineering, Mountain View, CA, USA). Each genomic DNA was isolated in lysis buffer (10 mM Tris-HCl at pH 7.5, 1 mM ethylenediaminetetraaceticacid, and 0.5% sodium dodecyl sulfate) and amplified by adaptor-ligation-mediated PCR after end-filling, as described by Tanabe et al. (2003).

We screened DNAs from GC cell lines and primary GCs for homozygous losses by genomic PCR. All the relevant primer sequences, except for analyses of the VLDLR gene (Table 1), are available on request.

Table 1 Primer sequences and PCR conditions for the VLDLR gene

RT–PCR

Single-stranded cDNAs were generated from total RNAs using the SuperScript™ First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA), and amplified with primers specific for each gene. The gene encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified at the same time to estimate the efficiency of cDNA synthesis.

Drug treatment

Cells were treated with 10 μ M of 5-aza-dCyd for 5 days and/or with 10 ng/ml of TSA for various periods. For the synergistic study, 10 μ M of 5-aza-dCyd was present in the cultures for 5 days, and 10 ng/ml TSA was added for the last 24 h.

Promoter reporter assay

We obtained by PCR three DNA fragments (Fragment 1–3 with 1005, 655 and 360 bp, respectively), which contain sequences of CpG islands around exon 1 of VLDLR predicted by the CpGPLOT program (http://www.ebi.ac.uk/emboss/cpgplot/). Each fragment was ligated into the vector pGL3-Basic (Promega, Madison, WI, USA), and an equal amount of each construct was introduced into cells along with an internal control vector (pRL-hTK, Promega), using FuGENE 6 (Roche Diagnostics, Tokyo, Japan). A pGL3-Basic vector without an insert served as a negative control. Firefly luciferase and Renilla luciferase activities were each measured 36 h after transfection, using the Dual-Luciferase Reporter Assay System (Promega); relative luciferase activities were calculated and normalized versus Renilla luciferase activity.

COBRA and bisulfite-sequencing

Genomic DNAs were treated with sodium bisulfite using an EZ DNA Methylation kit (Zymo Research, Orange, CA, USA), and subjected to PCR using primer sets designed to amplify the regions of interest, which were divided into two PCR fragments to analyse the entire sequences effectively (Table 1).

For COBRA, PCR products were digested with HhaI, which recognizes sequences unique to the methylated alleles but cannot recognize unmethylated alleles, and electrophoresed (Xiong and Laird, 1997). For bisulfite-sequencing, the PCR products were sub-cloned and then sequenced.

MSP

Genomic DNA treated with sodium bisulfite was amplified using primers specific to the methylated and unmethylated forms of DNA sequences of interest (Table 1). DNAs from the cell line AZ-521, which were recognized as unmethylated forms by bisulfite-sequencing analysis, and from peripheral blood lymphocytes of a healthy male were used as negative controls for the MSP assay; DNAs from cell lines SNU601 and MKN74, recognized as methylated, served as positive controls for methylated alleles. PCR products were visualized on 3% agarose gels stained with ethidium bromide.

Mutation analysis

Mutation detection of the VLDLR gene was carried out by exon amplification and direct sequencing using previously described primers (Near et al., 2001) in addition to those designed for exons 1 and 2 (Table 2). Base changes in tumor samples were confirmed by sequencing these tumors and corresponding normal tissues in both directions.

Table 2 Amplification primers for the coding region of the VLDLR gene

Transient transfection, Western blotting and colony formation assays

Two plasmids expressing COOH-terminally 3 × Myc-tagged VLDLR type I and type II (pCMV-3Tag-4-VLDLR-type I/II) were obtained by cloning a full coding sequence and a sequence lacking O-linked sugar domain of VLDLR, respectively, into the pCMV-3Tag-4 vector (Stratagene, La Jolla, CA, USA) in-frame along with the Myc-epitope. pCMV-3Tag-4-VLDLR-type I/II, or the empty vector (pCMV-3Tag-4-mock) control, were transfected into cells for colony formation assays as described elsewhere (Yuki et al., 2004). Expressions of VLDLR-type I/II protein in transiently transfected cells were confirmed 36 h after transfection by Western blot analysis using anti-Myc antibody (Cell Signaling Technology, Beverly, MA, USA), as described elsewhere (Yuki et al., 2004). After 3 weeks of incubation with appropriate concentrations of G418 in six-well plates, cells were fixed with 70% ethanol and stained with crystal violet.

References

  1. An HX, Claas A, Savelyeva L, Seitz S, Schlag P, Scherneck S et al. (1999). Cancer Res 59: 3941–3943.

  2. Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP . (1998). Adv Cancer Res 72: 141–196.

  3. Chen T, Wu F, Chen FM, Tian J, Qu S . (2005). World J Gastroenterol 11: 2817–2821.

  4. Chen WJ, Goldstein JL, Brown MS . (1990). J Biol Chem 265: 3116–3123.

  5. D'Arcangelo G, Homayouni R, Keshvara L, Rice DS, Sheldon M, Curran T . (1999). Neuron 24: 471–479.

  6. Fukuda Y, Kurihara N, Imoto I, Yasui K, Yoshida M, Yanagihara K et al. (2000). Genes Chromosomes Cancer 29: 315–324.

  7. Heidenblad M, Schoenmakers EF, Jonson T, Gorunova L, Veltman JA, van Kessel AG et al. (2004). Cancer Res 64: 3052–3059.

  8. Hembrough TA, Ruiz JF, Papathanassiu AE, Green SJ, Strickland DK . (2001). J Biol Chem 276: 12241–12248.

  9. Hiltunen TP, Luoma JS, Nikkari T, Yla-Herttuala S . (1998). Circulation 97: 1079–1086.

  10. Inazawa J, Inoue J, Imoto I . (2004). Cancer Sci 95: 559–563.

  11. Kim SK, Ro JY, Kemp BL, Lee JS, Kwon TJ, Fong KM et al. (1997). Cancer Res 57: 400–403.

  12. Martensen PM, Oka K, Christensen L, Rettenberger PM, Petersen HH, Christensen A et al. (1997). Eur J Biochem 248: 583–591.

  13. Nakamura Y, Yamamoto M, Kumamaru E . (2000). Arch Pathol Lab Med 124: 119–122.

  14. Near SE, Wang J, Hegele RA . (2001). J Hum Genet 46: 490–493.

  15. Noguchi M, Furuya S, Takeuchi T, Hirohashi S . (1997). Pathol Int 47: 685–691.

  16. Reisman DN, Sciarrotta J, Wang W, Funkhouser WK, Weissman BE . (2003). Cancer Res 63: 560–566.

  17. Sakai J, Hoshino A, Takahashi S, Miura Y, Ishii H, Suzuki H et al. (1994). J Biol Chem 269: 2173–2182.

  18. Sakata K, Tamura G, Maesawa C, Suzuki Y, Terashima M, Satoh K et al. (1995). Jpn J Cancer Res 86: 333–335.

  19. Sarkar S, Roy BC, Hatano N, Aoyagi T, Gohji K, Kiyama R . (2002). J Biol Chem 277: 36585–36591.

  20. Sato M, Takahashi K, Nagayama K, Arai Y, Ito N, Okada M et al. (2005). Genes Chromosomes Cancer 44: 405–414.

  21. Sato N, Fukushima N, Chang R, Matsubayashi H, Goggins M . (2006). Gastroenterology 130: 548–565.

  22. Shim YH, Kang GH, Ro JY . (2000). Lab Invest 80: 689–695.

  23. Sonoda I, Imoto I, Inoue J, Shibata T, Shimada Y, Chin K et al. (2004). Cancer Res 64: 3741–3747.

  24. Strickland DK, Kounnas MZ, Argraves WS . (1995). FASEB J 9: 890–898.

  25. Suzuki J, Takahashi S, Oida K, Shimada A, Kohno M, Tamai T et al. (1995). Biochem Biophys Res Commun 206: 835–842.

  26. Takada H, Imoto I, Tsuda H, Nakanishi Y, Ichikura T, Mochizuki H et al. (2005b). Oncogene 24: 8051–8060.

  27. Takada H, Imoto I, Tsuda H, Sonoda I, Ichikura T, Mochizuki H et al. (2005a). Cancer Sci 96: 100–110.

  28. Takahashi S, Kawarabayasi Y, Nakai T, Sakai J, Yamamoto T . (1992). Proc Natl Acad Sci USA 89: 9252–9256.

  29. Takahashi S, Suzuki J, Kohno M, Oida K, Tamai T, Miyabo S et al. (1995). J Biol Chem 270: 15747–15754.

  30. Tanabe C, Aoyagi K, Sakiyama T, Kohno T, Yanagitani N, Akimoto S et al. (2003). Genes Chromosomes Cancer 38: 168–176.

  31. Tanami H, Tsuda H, Okabe S, Iwai T, Sugihara K, Imoto I et al. (2005). Lab Invest 85: 1118–1129.

  32. Vo QN, Geradts J, Boudreau DA, Bravo JC, Schneider BG . (2002). Hum Pathol 33: 1200–1204.

  33. Wada Y, Homma Y, Nakazato K, Ishibashi T, Maruyama Y . (2000). Heart Vessels 15: 74–80.

  34. Webb DJ, Nguyen DH, Sankovic M, Gonias SL . (1999). J Biol Chem 274: 7412–7420.

  35. Webb JC, Patel DD, Jones MD, Knight BL, Soutar AK . (1994). Hum Mol Genet 3: 531–537.

  36. Xiong Z, Laird PW . (1997). Nucleic Acids Res 25: 2532–2534.

  37. Yamano M, Fujii H, Takagaki T, Kadowaki N, Watanabe H, Shirai T . (2000). Am J Pathol 156: 2123–2133.

  38. Yuki Y, Imoto I, Imaizumi M, Hibi S, Kaneko Y, Amagasa T et al. (2004). Cancer Sci 95: 503–507.

Download references

Acknowledgements

We are grateful to Professor Yusuke Nakamura (Human Genome Center, The Institute of Medical Science, The University of Tokyo) for his continuous encouragement throughout this work. We thank Professor Jae-Gahb Park (Laboratory of Cell Biology, Cancer Research Institute, Seoul National University College of Medicine) and Dr Kazuyoshi Yanagihara (Central Animal Laboratory, National Cancer Center Research Institute) for providing GC cell lines and Ai Watanabe for technical assistance.This work was supported by Grants-in-Aid for Scientific Research on priority areas (C) from the Ministry of Education, Culture, Sports, Science and Technology, Japan; by a Grant-in-Aid from Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Corporation (JST); by a Center of Excellence (COE) program for Frontier Research on Molecular Destruction and Reconstitution of Tooth and Bone; by the program for promotion of Fundamental Studies in Health Sciences of the Pharmaceuticals and Medical Devices Agency (PMDA); and by the Third Term Comprehensive Control Research for Cancer of the Ministry of Health, Labour and Welfare.

Author information

Correspondence to J Inazawa.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Takada, H., Imoto, I., Tsuda, H. et al. Genomic loss and epigenetic silencing of very-low-density lipoprotein receptor involved in gastric carcinogenesis. Oncogene 25, 6554–6562 (2006) doi:10.1038/sj.onc.1209657

Download citation

Keywords

  • array-CGH
  • gastric cancer
  • VLDLR
  • homozygous deletion
  • methylation

Further reading