1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

To identify novel methylation-silenced genes in gastric cancers, we carried out a chemical genomic screening, a genome-wide search for genes upregulated by treatment with a demethylating agent, 5-aza-2′-deoxycytidine (5-aza-dC). After 5-aza-dC treatment of a gastric cancer cell line (AGS) 579 genes were upregulated 16-fold or more, using an oligonucleotide microarray with 39 000 genes. From these genes, we selected 44 known genes on autosomes whose silencing in gastric cancer has not been reported. Thirty-two of these had CpG islands (CGI) in their putative promoter regions, and all of the CGI were methylated in AGS, giving an estimated number of 421 ± 75 (95% confidence interval) methylation-silenced genes. Additionally, we analyzed the methylation status of 16 potential tumor-related genes with promoter CGI that were upregulated four-fold or more, and 14 of these were methylated in AGS. Methylation status of the 32 randomly selected and 16 potential tumor-related genes was analyzed in 10 primary gastric cancers, and 42 genes (ABHD9, ADFP, ALDH1A3, ANXA5, AREG, BDNF, BMP7, CAV1, CDH2, CLDN3, CTSL, EEF1A2, F2R, FADS1, FSD1, FST, FYN, GPR54, GREM1, IGFBP3, IGFBP7, IRS2, KISS1, MARK1, MLF1, MSX1, MTSS1, NT5E, PAX6, PLAGL1, PLAU, PPIC, RBP4, RORA, SCRN1, TBX3, TFAP2C, TNFSF9, ULBP2, WIF1, ZNF177 and ZNF559) were methylated in at least one primary gastric cancer. A metastasis suppressor gene, MTSS1, was located in a genomic region with frequent loss of heterozygosity (8q22), and was expressed abundantly in the normal gastric mucosa, suggesting its role in gastric carcinogenesis. (Cancer Sci 2006; 97: 64 –71)




CpG island


loss of heterozygosity


methylation sensitive-representational analysis


methylation-specific PCR


polymerase chain reaction.

Epigenetic alterations are involved in cancer development and progression, and methylation of promoter CGI leads to transcriptional silencing of their downstream genes.(1) In various human cancers, silencing of tumor-suppressor genes, such as CDKN2A (p16), CDH1 (E-cadherin) and MLH1, is known to be one of the major mechanisms for their inactivation, along with mutations and LOH. To identify genes silenced by promoter methylation by genome-wide screenings, various techniques have been developed.(2) Most techniques are based on the methylation status of genomic DNA, including MS-RDA and restriction landmark genomic scanning. In contrast, Suzuki et al. developed a technique that screens genes re-expressed after treatment with a demethylating agent, 5-aza-dC, using a microarray.(3) The chemical genomic screening technique is simple and is effective in identifying genes silenced in cell lines. It has been applied to colon, bladder, esophageal, pancreatic and prostate cancers.(3–7)

Gastric cancer is the second most common cause of cancer death in the world.(8) As its molecular basis, deep involvement of aberrant DNA methylation has been indicated by the higher incidences of aberrant DNA methylation of known tumor-suppressor genes than of mutations.(9) We previously searched for genes silenced in MKN28 and MKN74 cell lines using MS-RDA,(10) and identified lysyl oxidase as a novel tumor-suppressor gene.(11) However, the entire picture of methylation-silenced genes in gastric cancers is still unclear, and further searches for methylation-silenced genes are necessary.

In the present study, we carried out a chemical genomic screening of methylation-silenced genes in the human gastric cancer cell line AGS.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Tissue samples, cell lines and 5-aza-dC treatment

Ten primary gastric cancer samples (male/female = 7/3, aged 38–81 years) and two normal gastric mucosae were obtained from 10 patients undergoing gastrectomy at Aichi Cancer Center (Nagoya, Japan) with informed consent. These samples were frozen and stored at −80°C until extraction of DNA or RNA. Gastric cancer cell lines AGS, MKN28, MKN45 and KATOIII were obtained from the Japanese Collection of Research Bioresources (Tokyo, Japan) and American Type Culture Collection (Manassas, VA, USA). Two gastric cancer cell lines, HSC44 and HSC57, were gifted by Dr Kazuyoshi Yanagihara at the National Cancer Center Research Institute (Tokyo, Japan). AGS cells were seeded at a density of 3 × 105 cells/10 cm dish on day 0 and treated with freshly prepared 1 µM 5-aza-dC (Sigma) for 24 h on days 1, 3 and 5. After each treatment, the cells were placed in fresh medium and harvested on day 6. Genomic DNA was extracted by standard phenol/chloroform procedures. Total RNA was extracted using ISOGEN (Nippon Gene, Tokyo, Japan) and purified using an RNeasy Mini kit (Qiagen, Valencia, CA, USA).

Oligonucleotide microarray analysis

Oligonucleotide microarray analysis was carried out using GeneChip Human Genome 133 Plus 2.0 (Affymetrix, Santa Clara, CA, USA) with 54 000 probe sets and 47 400 transcripts from 39 000 genes. From 8 µg of total RNA, the first-strand cDNA was synthesized with SuperScript III reverse transcriptase (Invitrogen, Groningen, the Netherlands) and a T7-(dT)24 primer (Amersham Bioscience, Buckinghamshire, UK), and the double-stranded cDNA was then synthesized. From the double-stranded cDNA, biotin-labeled cRNA was prepared using a BioArray HighYield RNA transcript labeling kit (Enzo, Farmingdale, NY, USA). Labeled cRNA (20 µg) was fragmented, and the GeneChips were hybridized. The arrays were stained and scanned according to the protocol from Affymetrix. The data were processed using GeneChip Operating Software. The signal intensities were normalized so that the average of all of the genes on a GeneChip would be 500. The P-values for different expression (change P-value) were calculated in each probe by statistical algorithms based on the Wilcoxon's signed rank test. The change P-values of 0.003 and 0.997 were used as thresholds to define genes with increased and decreased expression, respectively. Expression data for the normal tissues using GeneChip were obtained from the database RefEXA (,(12) with kind permission from Dr H. Aburatani.

Methylation-specific polymerase chain reaction

DNA (1 µg) digested with BamHI was denatured in 0.3 M NaOH at 37°C for 15 min. Then, 3.6 M sodium bisulfite (pH 5.0) and 0.6 mM hydroquinone were added, and the sample underwent 15 cycles of 30-s denaturation at 95°C and a 15-min incubation at 50°C. The sample was desalted with the Wizard DNA Clean-Up system (Promega, Madison, WI, USA) and desulfonated in 0.3 M NaOH. DNA was ethanol-precipitated and dissolved in 40 µL of Tris-EDTA buffer. MSP was carried out with a primer set specific to the methylated or unmethylated sequence (M or U set), using 0.5 µL of the sodium-bisulfite-treated DNA. A region 200 bp or less upstream of a putative transcriptional start site was analyzed, except for BDNF (−401 to −214). Primer sequences and PCR conditions are shown in Table 1. DNA methylated with SssI methylase was used to determine specific conditions of PCR for M sets.

Table 1. Primers for methylation-specific polymerase chain reaction
GenesM/UForward primerReverse primerAnnealing(°C)No.cycles
  • Transcription start site = 0. All primers were designed on the top strand sequences. M, specific to methylated DNA; U, specific to unmethylated DNA.



  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Oligonucleotide microarray analysis

AGS cells were treated with 1 µM of 5-aza-dC, which caused growth suppression at 49%, and upregulated genes were searched for using an oligonucleotide microarray. Among the 39 000 genes (54 000 probe sets) analyzed, 1430 genes (1747 probes) were upregulated four-fold or more (signal log ratio > 2) and 579 genes (678 probes) were upregulated 16-fold or more (signal log ratio > 4). To identify silenced genes with known functions from the 579 genes, we excluded genes on chromosome X (95 probes, 70 genes) and genes without known functions (i.e. FLJ genes, KIAA genes, LOC genes, MG genes and Orf genes [149 probes, 141 genes]).

Among the remaining 368 genes (434 probes), we found eight genes (14 probes) whose methylation-silencing had been reported in gastric cancers (BNIP3,(13) CDKN2A (p16),(14) CHFR,(15) ID4,(16) RBP1,(17) RUNX3,(18) THBD,(10) TIMP(19)). The remaining 360 genes (420 probes) were considered as candidates for novel methylation-silenced genes in gastric cancers (Table 1).

Methylation analysis of genes upregulated by 5-aza-dC treatment

From the 360 genes upregulated 16-fold or more, we selected 44 genes randomly (Table 2). Among these 44 genes, 32 genes (73%) had CGI in their 5′ regions, which were considered as promoter regions (Table 2). To examine whether the induction of these genes by 5-aza-dC treatment was really due to promoter demethylation, the methylation status of these 5′-CGIs were analyzed by MSP. For all the 32 genes, only methylated molecules were detected before 5-aza-dC treatment, and unmethylated DNA molecules were detected after the treatment in AGS, suggesting silencing of the 32 genes by methylation of their 5′-CpG islands (representative results in Fig. 1).

Table 2. Genes upregulated after 5-aza-dC treatment in the AGS cell line
Probe setGene titleSymbolFold changeCGI
The 44 genes picked randomly from the genes showing greater than 16-fold upregulation after 5-aza-dC treatment
209122_atAdipose differentiation-related proteinADFP 36.0Yes
203180_atAldehyde dehydrogenase 1 family, member A3ALDH1A3 46.2Yes
200782_atAnnexin A5ANXA5 32.5Yes
205239_atAmphiregulin (schwannoma-derived growth factor)AREG 59.3Yes
206382_s_atBrain-derived neurotrophic factorBDNF 29.2Yes
203065_s_atCaveolin 1, caveolae protein, 22 kDaCAV1 37.2Yes
210140_atCystatin F (leukocystatin)CST7 22.1No
219424_atEpstein–Barr virus induced gene 3EBI3 18.5No
204540_atEukaryotic translation elongation factor 1 alpha 2EEF1A2 25.0Yes
203989_x_atCoagulation factor II (thrombin) receptorF2R 59.3Yes
208962_s_atFatty acid desaturase 1FADS1 23.0Yes
203240_atFc fragment of IgG binding proteinFCGBP  16.8No
1570515_a_atFilamin A interacting protein 1FILIP1 34.8No
219170_atFibronectin type 3 and SPRY domain containing 1FSD1 16.0Yes
226847_atFollistatinFST 29.2Yes
218469_atGremlin 1 homolog, cysteine knot superfamilyGREM1 59.3Yes
213620_s_atIntercellular adhesion molecule 2ICAM2 18.5No
201162_atInsulin-like growth factor binding protein 7IGFBP7 38.4Yes
205945_atInterleukin 6 receptorIL6R 26.0Yes
209185_s_atInsulin receptor substrate 2IRS2 24.0Yes
205563_atKiSS-1 metastasis-suppressorKISS1 54.8Yes
221047_s_atMAP/microtubule affinity-regulating kinase 1MARK1 17.6Yes
1552456_a_atMethyl-CpG binding domain protein 3-like 2MBD3L2 38.4No
206560_s_atMelanoma inhibitory activityMIA 23.0No
204784_s_atMyeloid leukemia factor 1MLF1 27.0Yes
205932_s_atMsh homeo box homolog 1 (Drosophila)MSX1 16.0Yes
202086_atMyxovirus (influenza virus) resistance 1MX1 18.5Yes
205581_s_atNitric oxide synthase 3 (endothelial cell)NOS3 18.5No
203939_at5′-nucleotidase, ecto (CD73)NT5E 53.3Yes
207002_s_atPleiomorphic adenoma gene-like 1PLAGL1 46.2Yes
204517_atPeptidylprolyl isomerase C (cyclophilin C)PPIC 23.0Yes
221666_s_atPYD and CARD domain containingPYCARD 25.0Yes
219140_s_atRetinol binding protein 4, plasmaRBP4 43.6Yes
202388_atRegulator of G-protein signaling 2, 24 kDaRGS2 33.6Yes
201462_atSecernin 1SCRN1 31.4Yes
204614_atSerine (or cysteine) proteinase inhibitor, clade B, member 2SERPINB2 36.0No
202627_s_atSerine (or cysteine) proteinase inhibitor, clade E, member 1SERPINE1 26.0No
208539_x_atSmall proline-rich protein 2 ASPRR2A 22.1No
224167_atLikely ortholog of mouse spermatogenic Zip 1SPZ1 18.5No
219682_s_atT-box 3TBX3 36.0Yes
205286_atTranscription factor AP-2 gammaTFAP2C 32.5Yes
221291_atUL16 binding protein 2ULBP2 29.2Yes
204712_atWNT inhibitory factor 1WIF1 31.4Yes
224518_s_atZinc finger protein 559ZNF559 30.3Yes
Genes showing greater than four-fold upregulation after 5-aza-dC treatment, having CpG islands, and having cancer related function or having chromosomal location in the region of frequent loss in gastric cancer.
220013_atAbhydrolase domain containing 9ABHD9 8.0Yes
209591_s_atBone morphogenetic protein 7 (osteogenic protein 1)BMP7 13.9Yes
203440_atCadherin 2, type 1, N-cadherin (neuronal)CDH2 26.0Yes
210240_s_atCyclin-dependent kinase inhibitor 2D (p19)CDKN2D 4.4Yes
203953_s_atClaudin 3CLDN3 45.3Yes
202087_s_atCathepsin LCTSL 13.0Yes
216033_s_atFYN oncogene related to SRC, FGR, YESFYN 78.8Yes
242517_atG protein-coupled receptor 54GPR54 6.3Yes
210095_s_atInsulin-like growth factor binding protein 3IGFBP3 9.8Yes
203037_s_atMetastasis suppressor 1MTSS1 6.5Yes
205646_s_atPaired box gene 6 (aniridia, keratitis)PAX6 18.4Yes
205479_s_atPlasminogen activator, urokinasePLAU 137.2Yes
210479_s_atRAR-related orphan receptor ARORA 55.7Yes
219480_atSnail homolog 1 (Drosophila)SNAI1 4.6Yes
206907_atTumor necrosis factor superfamily, member 9TNFSF9 4.8Yes
207417_s_atZinc finger protein 177ZNF177 4.4Yes
Genes reported as silenced genes in gastric cancer and showing greater than 16-fold upregulation after 5-aza-dC treatment
201848_s_atBCL2/adenovirus E1B 19 kDa interacting protein 3BNIP3 20.3Yes
207039_atCyclin-dependent kinase inhibitor 2 A (p16)CDKN2A 33.6Yes
223931_s_atCheckpoint with forkhead and ring finger domainsCHFR 16.8Yes
209291_atInhibitor of DNA binding 4ID4 57.8Yes
203423_atRetinol binding protein 1, cellularRBP1 54.8Yes
204198_s_atRunt-related transcription factor 3RUNX3 27.0Yes
203888_atThrombomodulinTHBD 29.2Yes
201147_s_atTissue inhibitor of metalloproteinase 3TIMP3 65.6Yes

Figure 1. A representative result of methylation analysis. (A) MLF1; (B) MSX1; and (C) TBX3. The left sides of each panel represent the 5′ CpG islands and regions analyzed by methylation-specific polymerase chain reaction (MSP). Vertical marks, individual GpC and CpG sites; Open boxes, non-coding and coding exons; and arrowheads, positions of MSP primers (M sets). The right sides show the results of MSP in gastric cancer cell lines, normal gastric mucosa and primary gastric cancers. 5-aza-dC, AGS cells after treatment with 5-aza-2′-deoxycytidine; SssI, genomic DNA methylated with SssI methylase.

Download figure to PowerPoint

Analysis of five additional gastric cancer cell lines (MKN28, MKN45, HSC44, HSC57, KATOIII) showed that five genes (ANXA5, AREG, CAV1, IL6R, TBX3) were methylated only in AGS, and 27 genes were methylated in multiple gastric cancer cell lines (Table 3). The microarray analysis of KATOIII and HSC57 showed that none of the 32 genes were expressed when unmethylated DNA molecules were not present.

Table 3.  Methylation profiles in gastric cancer
Gene SymbolFunctionChromo- somal locationMethylation status in gastric cancer cell lineExpression (signal intensity of GeneChip)Methylation status in gastric cancer
  1. From genes upregulated 16-fold or more, 44 genes were selected randomly, and the methylation status of 32 genes with CGI were analyzed. In addition, 16 genes with potential tumor-related functions and CGI were selected for methylation analysis. In cancer cell lines, M: only methylated molecules detected, U: only unmethylated molecules detected, M/U: both methylated and unmethylated molecules detected. In primary cancer samples, ‘M’ and ‘U’ indicate detection and absence, respectively, of methylated molecules. ‘M–’ indicates slight detection of methylated molecules. §Obtained from the RefEXA database. Underlined chromosomal locations, regions of frequent loss in gastric cancer.

32 genes selected randomly and with GGI
ADFPFatty acid transport9p22.1MUM/UUMM2191  74 331186154UUUUUMUUUU
ALDH1A3Vitamin A metabolism15q26.3UUUUMM 322  69 161435 65UUUUUMMUM–U
AREGCell proliferation4q13-q21UUUUUM95689103  72677137UUUUUMUUUU
BDNFGrowth factor activity11p13MMMM/UUM  50  20  5 447 32MUMMUMMMMM
CAV1Cell aging7q31.1UUUUUM  24 167 201411102UUUUUMUUUU
EEF1A2Translational elongation20q13.3UUMUMM 240  601127234 58UUUUUMUUUU
F2RG-protein signaling5q13UUM/UUMM 140   7  4 820 26UUUUUMMUUU
FADS1Fatty acid desaturation11q12.2-q13.1UUM/UUMM 375  19 22 542 40UUUUUMM–MMU
FSD1Microtubule depolymerization19p13.3UM/UM/UUMM  49  25 31 362  5UUUUUMUUMUU
FSTTranscription factor5q11.2UUMUMM   6   4  7 278UUUUUMMUMU
GREM1Signal transduction15q13–15q15MMMUMM  77  33  52128184MMMMMMMMMM
IGFBP7Angiogenesis4q12MUMUMM  13  11  9 934483UUM–UUMMUMM–
IL6RImmune response1q21UUUUUM 124 227 12 574 72UUUUUUUUUU
IRS2Signal transduction13q34UUUUMM6602  57 301245274UUUUUUUUM–U
KISS1Metastasis suppressor1q32M/UMMUUM  19  26 213697  7UM–M–M–UM–M–M–M–M–U
MARK1Protein amino acid phosphorylation1q41UUUUMM  79  85 45 845 43UUUUUMUMUU
MLF1Cell differentiation3q25.1MUUUM/UM 9361078 22 709 42MMMMUMMMMMU
MSX1Transcription factor4p16.3-p16.1UUM/UUMM 524  13 432675 32UUUUUMM–UMUU
MX1Response to virus21q22.3MUMUUM 8561533 48 996223UUUUUUUUUU
NT5EDNA metabolism6q14–6q21UUUUM/UM 115 394  4 821 31UUUUUUM–UUU
PLAGL1Induction of apoptosis6q24–6q25MMMM/UMM 456  16  4 668 81M–MMMMMMMMMU
PPICSignal transduction5q23.2UUUUMM2724  59 491621186UUUUUMM–UUU
PYCARDSignal transduction16p12–16p11.2MMMUUM2739 557  9 420 39UUUUUUUUUU
RBP4Vitamin A metabolism10q23–10q24MUMUUM  86 643 10 880 11UUUM–M–MMM–UM–U
RGS2G-protein signaling1q31UUUUM/UM/U  27  90 271727575UUUUUUUUUU
SCRN1Exocytosis7p14.3-p14.1M/UUUUMM1595  57 191306127UUUUUMUUMU
TBX3Morphogenesis12q24.1UUUUUM1280 881 151714 38UUUMMMMMM–MU
TFAP2CTranscription factor20q13.2UUM/UUMM1486  26 272446 51UM–MUUMMUMU
ULBP2T cell proliferation6q25UUUUM/UM 365 510 291286  8UUUUUMUUUUU
WIF1Wnt signaling12q14.3MMMM/UMM   9  41  7 241  2UUMMM–MMMMM–U
ZNF559Transcription factor19p13.2MUM/UUMM 609   7  7 532M–MM–M–UUUMM–M–U
16 genes with potential tumor-related functions
ABHD9Response to chemical substance19p13.12MMMUMM  19  18 13 140  3UMUM–MMMMMM–
BMP7Cell proliferation20q13UUMUM/UM4166  19 611185 68UUUM–UMMM–MM–
CDH2Cell adhesion18q11.2MMMUMM  13  13  9 487 57MUMMM–MMMMM
CDKN2DCdk inhibitor19p13UUUUUU  47 339119 489 28UUUUUUUUUU
CLDN3Cell adhesion7q11.23UUM/UUUM7584  69  7 359  1UUUUUMMM–MUU
CTSLProtein processing9q21–9q22UUMUM/UM1017 102 501753118UUUUM–MMMUM
FYNProto-oncogene6q21UUUUUM 3411046  5 797 64UUUUUMUUUU
GPR54G-protein signaling19p13.3UUMUMM  89  85 35 210UUUUUMUUUU
IGFBP3Induction of apoptosis7p13–7p12M/UM/UM/UUMM 106  15 31 277422MUMMM–MMMMM
MTSS1Cytoskeletal organization8p22M/UUMUUM  17 694 73 625400MUUUUMUMUU
PAX6Transcription factor11p13MUM/UUMM  17  21 29 417  6UUMMUMMMMM–U
PLAUAngiogenesis10q24MUUUMM1060 168 112950 80UUUUUM–UUUU
RORASignal transduction15q21-q22UUMUMM 195  49  2 340 20MMUUMM–MMMMU
SNAI1Transcriptional repressor20q13.1-q13.2UUUUUU 235  40 52 319  8UUUUUUUUUU
TNFSF9Apoptosis19p13.3UUUUUM 6151133 84 916 23UUUUUMUUUU
ZNF177Transcription factor19p13.2MMMMUM  76  59 47 222 14MMMMMMMMMMU

We next selected 16 potential tumor-related genes with promoter CGI and four-fold or greater upregulation as the above analysis suggested that a considerable number of silenced genes were still present among the genes with upregulation of 16-fold or less (Table 2). The potential tumor-related genes were selected based on their tumor-related function and location in genomic regions with frequent LOH (5q21-23,(20,21) 8p22,(20,21) 9p12-24(20–22)) or with DNA loss by comparative genomic hybridization (19p13.12-p13.3(23)) in gastric cancers. MSP showed that 14 of these 16 genes were methylated in AGS before 5-aza-dC treatment (Table 3). CDKN2D and SNAI1 were not methylated even before 5-aza-dC treatment, suggesting that they were induced as a stress response by 5-aza-dC treatment.

Presence of methylation in primary gastric cancers

The methylation status of the above 48 genes (32 selected randomly and 16 tumor-related genes) were examined in 10 primary gastric cancers. It was shown that 42 genes (ABHD9, ADFP, ALDH1A3, ANXA5, AREG, BDNF, BMP7, CAV1, CDH2, CLDN3, CTSL, EEF1A2, F2R, FADS1, FSD1, FST, FYN, GPR54, GREM1, IGFBP3, IGFBP7, IRS2, KISS1, MARK1, MLF1, MSX1, MTSS1, NT5E, PAX6, PLAGL1, PLAU, PPIC, RBP4, RORA, SCRN1, TBX3, TFAP2C, TNFSF9, ULBP2, WIF1, ZNF177 and ZNF559) were methylated in at least one gastric cancer (Table 3). The numbers of methylated genes in each case ranged from one to 10. Case 6 had a large number of methylated genes, which was similar to AGS (Table 3). The expression levels of the 48 genes in the normal gastric mucosae were obtained from the RefEXA database (Table 3).

There remained a possibility that these silenced genes were normally methylated or were methylated tissue-specifically. Therefore, we selected 11 genes with relatively high chances of having methylated CGI, based on their low expression in the normal gastric mucosae (CLDN3, FADS1, KISS1, PAX6, PLAGL1, RBP4, RORA, ULBP2, WIF1, ZNF177 and ZNF559). Along with three additional genes (MLF1, MSX1 and TBX3), their methylation status was examined in the normal gastric mucosae. However, none of the 14 genes were methylated.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Chemical genomic screening revealed that a considerable number of genes were methylation-silenced in the AGS gastric cancer cell line. After 5-aza-dC treatment of AGS, 579 genes were upregulated 16-fold or more. When we analyzed 44 selected genes, 32 of them had CGI in their promoter regions, and all of the 32 genes turned out to be methylation-silenced. Because 32 of the 44 genes selected from 579 genes were silenced, it was estimated that 421 ± 75 (95% confidence interval) genes were silenced in AGS. To avoid overestimation, we randomly selected 44 genes from the 360 genes after excluding: (i) genes on chromosome X, which harbors many normally methylated genes like MAGE; (ii) genes that have not been characterized yet; and (iii) genes whose methylation-associated silencing was already known in gastric cancers. Among the 16 potential tumor-related genes, 10 were upregulated 16-fold or less, and eight of the 10 genes were found to be methylation-silenced. If genes with relatively small upregulation were analyzed, the number of silenced genes in AGS was expected to be larger.

As for the number of methylation-silenced genes in a cancer, Costello et al. estimated that an average of 600 CGI in the whole genome were methylated aberrantly in the tumors.(24) However, the number was calculated by analyzing CGI in any location against a gene, and the number of genes silenced, for which methylation of promoter CGI is necessary, was not determined. Using chemical genomic screening, Sato et al. estimated that an average of 140 genes would be methylated aberrantly in pancreatic cancers.(6) Compared with this number, the number of genes silenced in the AGS cell line was considered to be much larger. We recently found that AGS had an increased rate of de novo methylation,(25) and this could be one of the mechanisms.

By methylation analysis of 48 genes (Table 3), 46 genes were found to be methylated in AGS, and 42 genes were methylated in at least one primary gastric cancer. Among the 42 genes, eight genes (CAV1,(26,27) IGFBP3,(28) IGFBP7[MAC25/IGFBP-rP1],(29) PAX6,(30) PLAGL1[ZAC/LOT1],(31,32) PLAU[uPA],(33,34) RBP4(35) and WIF1(36)) were reported to be silenced with functional relevance in cancers other than gastric cancers. In addition, two genes (CDH2(37) and FYN(38)) were reported to be methylated in some cancers, but their functional significance needs clarification.

Also among the 32 genes whose silencing was novel, we were able to find potential tumor-related genes. To achieve this, some genes were selected based on (i) antioncogenic cellular functions or (ii) location in genomic regions with frequent LOH in gastric cancers. Candidate tumor-related genes were further selected based on (iii) the presence of methylation of promoter CGI in primary gastric cancers, and (iv) expression in normal gastric mucosae when various tissues were compared. MTSS1/MIM/BEG4 met all of these criteria, and was a good candidate for a novel tumor-related gene. It mediates Sonic hedgehog signaling by potentiating Gli-dependent transcription,(39) and is known as a metastasis suppressor gene in bladder cancers.(40) Although LOH was not frequent in their locations, ANXA5, AREG, GREM, IGFBP7, IRS2, BMP7, CTSL and IGFBP3 were expressed in the normal gastric mucosae and had potential antioncogenic functions, such as mediation of SMAD signaling (BMP7)(41) and induction of apoptosis (IGFBP3).(42) There is a possibility that silencing of these genes is causally related to development and progression of gastric cancers. However, considering the large number of methylation-silenced genes, it was likely that the majority of the genes silenced in AGS did not have causal roles in gastric carcinogenesis.

In summary, we found a considerable number of methylation-silenced genes in a gastric cancer cell line AGS. Potential tumor-related genes were selected based on their known functions, chromosomal locations, methylation in primary samples and expression in normal gastric mucosae. The usefulness of chemical genomic screening was confirmed.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

This work was supported by a Grant-in-Aid on Priority Area from the Ministry of Education, Sciences, Culture and Sports (MEXT), Japan.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  • 1
    Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002; 3: 41528.
  • 2
    Ushijima T. Detection and interpretation of altered methylation patterns in cancer cells. Nat Rev Cancer 2005; 5: 22331.
  • 3
    Suzuki H, Gabrielson E, Chen W et al. A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer. Nat Genet 2002; 31: 14119.
  • 4
    Liang G, Gonzales FA, Jones PA, Orntoft TF, Thykjaer T. Analysis of gene induction in human fibroblasts and bladder cancer cells exposed to the methylation inhibitor 5-aza-2′-deoxycytidine. Cancer Res 2002; 62: 9616.
  • 5
    Yamashita K, Upadhyay S, Osada M et al. Pharmacologic unmasking of epigenetically silenced tumor suppressor genes in esophageal squamous cell carcinoma. Cancer Cell 2002; 2: 48595.
  • 6
    Sato N, Fukushima N, Maitra A et al. Discovery of novel targets for aberrant methylation in pancreatic carcinoma using high-throughput microarrays. Cancer Res 2003; 63: 373542.
  • 7
    Lodygin D, Epanchintsev A, Menssen A, Diebold J, Hermeking H. Functional epigenomics identifies genes frequently silenced in prostate cancer. Cancer Res 2005; 65: 421827.
  • 8
    Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin 2005; 55: 74108.
  • 9
    Ushijima T, Sasako M. Focus on gastric cancer. Cancer Cell 2004; 5: 1215.
  • 10
    Kaneda A, Kaminishi M, Yanagihara K, Sugimura T, Ushijima T. Identification of silencing of nine genes in human gastric cancers. Cancer Res 2002; 62: 664550.
  • 11
    Kaneda A, Wakazono K, Tsukamoto T et al. Lysyl oxidase is a tumor suppressor gene inactivated by methylation and loss of heterozygosity in human gastric cancers. Cancer Res 2004; 64: 641015.
  • 12
    Ge X, Yamamoto S, Tsutsumi S et al. Interpreting expression profiles of cancers by genome-wide survey of breadth of expression in normal tissues. Genomics 2005; 86: 12741.
  • 13
    Murai M, Toyota M, Suzuki H et al. Aberrant methylation and silencing of the BNIP3 gene in colorectal and gastric cancer. Clin Cancer Res 2005; 11: 10217.
  • 14
    Lee YY, Kang SH, Seo JY et al. Alterations of p16INK4A and p15INK4B genes in gastric carcinomas. Cancer 1997; 80: 188996.
  • 15
    Satoh A, Toyota M, Itoh F et al. Epigenetic inactivation of CHFR and sensitivity to microtubule inhibitors in gastric cancer. Cancer Res 2003; 63: 860613.
  • 16
    Chan AS, Tsui WY, Chen X et al. Downregulation of ID4 by promoter hypermethylation in gastric adenocarcinoma. Oncogene 2003; 22: 694653.
  • 17
    Esteller M, Guo M, Moreno V et al. Hypermethylation-associated inactivation of the cellular retinol-binding-protein 1 gene in human cancer. Cancer Res 2002; 62: 59025.
  • 18
    Li QL, Ito K, Sakakura C et al. Causal relationship between the loss of RUNX3 expression and gastric cancer. Cell 2002; 109: 11324.
  • 19
    Kang GH, Lee S, Kim JS, Jung HY. Profile of aberrant CpG island methylation along the multistep pathway of gastric carcinogenesis. Lab Invest 2003; 83: 63541.
  • 20
    Yustein AS, Harper JC, Petroni GR, Cummings OW, Moskaluk CA, Powell SM. Allelotype of gastric adenocarcinoma. Cancer Res 1999; 59: 143741.
  • 21
    Kim KM, Kwon MS, Hong SJ et al. Genetic classification of intestinal-type and diffuse-type gastric cancers based on chromosomal loss and microsatellite instability. Virchows Arch 2003; 443: 491500.
  • 22
    Choi SW, Park SW, Lee KY, Kim KM, Chung YJ, Rhyu MG. Fractional allelic loss in gastric carcinoma correlates with growth patterns. Oncogene 1998; 17: 26559.
  • 23
    Weiss MM, Kuipers EJ, Postma C et al. Genomic alterations in primary gastric adenocarcinomas correlate with clinicopathological characteristics and survival. Cell Oncol 2004; 26: 30717.
  • 24
    Costello JF, Fruhwald MC, Smiraglia DJ et al. Aberrant CpG-island methylation has non-random and tumour-type-specific patterns. Nat Genet 2000; 24: 1328.
  • 25
    Ushijima T, Watanabe N, Shimizu K, Miyamoto K, Sugimura T, Kaneda A. Decreased fidelity in replicating CpG methylation patterns in cancer cells. Cancer Res 2005; 65: 1117.
  • 26
    Sunaga N, Miyajima K, Suzuki M et al. Different roles for caveolin-1 in the development of non-small cell lung cancer versus small cell lung cancer. Cancer Res 2004; 64: 427785.
  • 27
    Wiechen K, Diatchenko L, Agoulnik A et al. Caveolin-1 is down-regulated in human ovarian carcinoma and acts as a candidate tumor suppressor gene. Am J Pathol 2001; 159: 163543.
  • 28
    Hanafusa T, Yumoto Y, Nouso K et al. Reduced expression of insulin-like growth factor binding protein-3 and its promoter hypermethylation in human hepatocellular carcinoma. Cancer Lett 2002; 176: 14958.
  • 29
    Komatsu S, Okazaki Y, Tateno M et al. Methylation and downregulated expression of mac25/insulin-like growth factor binding protein-7 is associated with liver tumorigenesis in SV40T/t antigen transgenic mice, screened by restriction landmark genomic scanning for methylation (RLGS-M). Biochem Biophys Res Commun 2000; 267: 10917.
  • 30
    Toyota M, Ho C, Ahuja N et al. Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res 1999; 59: 230712.
  • 31
    Abdollahi A, Pisarcik D, Roberts D, Weinstein J, Cairns P, Hamilton TC. LOT1 (PLAGL1/ZAC1), the candidate tumor suppressor gene at chromosome 6q24-25 is epigenetically regulated in cancer. J Biol Chem 2003; 278: 60419.
  • 32
    Bilanges B, Varrault A, Basyuk E et al. Loss of expression of the candidate tumor suppressor gene ZAC in breast cancer cell lines and primary tumors. Oncogene 1999; 18: 397988.
  • 33
    Xing RH, Rabbani SA. Transcriptional regulation of urokinase (uPA) gene expression in breast cancer cells: role of DNA methylation. Int J Cancer 1999; 81: 44350.
  • 34
    Pakneshan P, Xing RH, Rabbani SA. Methylation status of uPA promoter as a molecular mechanism regulating prostate cancer invasion and growth in vitro and in vivo. Faseb J 2003; 17: 10818.
  • 35
    Kwong J, Lo KW, Shuk-Nga Chow L et al. Epigenetic silencing of cellular retinol-binding proteins in nasopharyngeal carcinoma. Neoplasia 2005; 7: 6774.
  • 36
    Mazieres J, He B, You L et al. Wnt inhibitory factor-1 is silenced by promoter hypermethylation in human lung cancer. Cancer Res 2004; 64: 471720.
  • 37
    Hagihara A, Miyamoto K, Furuta J et al. Identification of 27 5′ CpG islands aberrantly methylated and 13 genes silenced in human pancreatic cancers. Oncogene 2004; 23: 870510.
  • 38
    Wang W, Huper G, Guo Y, Murphy SK, Olson JA Jr, Marks JR. Analysis of methylation-sensitive transcriptome identifies GADD45a as a frequently methylated gene in breast cancer. Oncogene 2005; 24: 270514.
  • 39
    Callahan CA, Ofstad T, Horng L et al. MIM/BEG4, a sonic hedgehog-responsive gene that potentiates Gli-dependent transcription. Genes Dev 2004; 18: 27249.
  • 40
    Lee YG, Macoska JA, Korenchuk S, Pienta KJ. MIM, a potential metastasis suppressor gene in bladder cancer. Neoplasia 2002; 4: 2914.
  • 41
    Yang S, Zhong C, Frenkel B, Reddi AH, Roy-Burman P. Diverse biological effect and Smad signaling of bone morphogenetic protein 7 in prostate tumor cells. Cancer Res 2005; 65: 576977.
  • 42
    Butt AJ, Firth SM, King MA, Baxter RC. Insulin-like growth factor-binding protein-3 modulates expression of Bax and Bcl-2 and potentiates p53-independent radiation-induced apoptosis in human breast cancer cells. J Biol Chem 2000; 275: 39 174–81.