Chromosomal abnormalities are good guideposts when hunting for cancer-related genes. We analyzed copy number alterations of 163 primary gastric cancers using array-based comparative genomic hybridization and simultaneously performed a genome-wide integrated analysis of copy number and gene expression using microarray data for 58 tumors. We showed that chromosome 6p21 amplification frequently occurred secondary to ERBB2 amplification, was associated with poorer prognosis and caused overexpression of half of the genes mapped. A comprehensive small interfering RNA knockdown of 58 genes overexpressed in tumors identified 32 genes that reduced gastric cancer cell growth. Enforced expression of 16 of these genes promoted cell growth in vitro, and six genes showing more than two-fold activity conferred tumor-forming ability in vivo. Among these six candidates, GLO1, encoding a detoxifying enzyme glyoxalase I (GLO1), exhibited the strongest tumor-forming activity. Coexpression of other genes with GLO1 enhanced growth-stimulating activity. A GLO1 inhibitor, S-p-bromobenzyl glutathione cyclopentyl diester, inhibited the growth of two-thirds of 24 gastric cancer cell lines examined. The efficacy was found to be associated with the mRNA expression ratio of GLO1 to GLO2, encoding glyoxalase II (GLO2), another constituent of the glyoxalase system. GLO1 downregulation affected cell growth through inactivating central carbon metabolism and reduced the transcriptional activities of nuclear factor kappa B and activator protein-1. Our study demonstrates that GLO1 is a novel metabolic oncogene of the 6p21 amplicon, which promotes tumor growth and aberrant transcriptional signals via regulating cellular metabolic activities for energy production and could be a potential therapeutic target in gastric cancer.
Gastric cancer is the second most common cause of cancer-related death worldwide and shows the highest incidence in Eastern Asia.1 Previous investigations have identified multiple genetic/epigenetic alterations that result in activation of the proto-oncogenes MET, FGFR2, ERBB2 and KRAS and inactivation or silencing of the tumor-suppressor genes p53, p73, APC, p16, RUNX3 and CDH1 in gastric cancer.2,3 In primary tumors and gastric cancer cell lines, many copy number alterations (CNAs) have been identified using array-based comparative genomic hybridization (aCGH) techniques.4, 5, 6 However, only a few studies have analyzed the genome-wide correlations between CNA and transcriptional changes.7, 8, 9 Molecular characterization of CNA that is directly related to changes in gene expression or gene structure is essential for understanding the genetic basis of gastric cancer, which may eventually facilitate in identification of critical genes in cancer development.
In the present study, to comprehensively identify novel candidate oncogenes, we performed a genome-wide integrated analysis of a total of 163 CNA profiles consisting of 79 differentiated and 84 undifferentiated adenocarcinomas and an oligonucleotide expression microarray analysis of 58 tumors and 16 non-cancerous tissues. The analysis revealed many target genes whose expression was significantly associated with changes in copy number, including candidate genes of potential clinicopathological value in gastric cancer.
Among them, we focused on the chromosome 6p21 amplicon and identified multiple candidate oncogenes with the ability to enhance cell growth in vitro and in vivo by systematic functional screening using RNA interference and enforced gene expression. We find that GLO1 gene encoding glyoxalase I (GLO1) exibits the strongest oncogenic activity in the amplicon and also other six genes cooperatively function with GLO1 to stimulate cell growth. Inhibition of the enzymatic activity and knockdown of mRNA expression resulted in defect in gastric cancer cell growth. Metabolome analysis indicated that GLO1 knockdown affected key pathways for energy production, such as glycolysis, the pentose phosphate pathway and the tricarboxylic acid cycle. Furthermore, we find that GLO1 downregulation caused suppression of the transcriptional activities of nuclear factor kappa B (NF-κB) and activator protein-1 (AP-1).
Genomic CNAs and correlation between DNA copy number and mRNA expression in gastric cancer
To identify a comprehensive pattern of genomic CNA in gastric cancer, we performed aCGH analysis of 163 primary gastric adenocarcinomas, consisting of 79 well and moderately differentiated tubular adenocarcinomas and 84 poorly differentiated adenocarcinomas, using bacterial artificial chromosome arrays covering the whole human genome with 0.6-megabase resolution (Supplementary Figure S1). The regions with recurrent genomic amplification (signal ratio⩾2.5) included 133 loci, and recurrent homozygous deletions (signal ratio <0.4) were detected at 30 loci within the hemizygously deleted regions (signal ratio <0.75) (Supplementary Tables S5 and S6). Frequent amplification was observed at the ERBB2 locus at 17q12 (12.9% of cases), the VEGFA locus at 6p21.1 (8.6%), the CCND1 locus at 11q13.3 (7.4%), the CCNE1 locus at 19q12 (7.4%), the BCAS1 locus at 20q13.2 (7.4%) and the MYC locus at 8q24.21 (5.5%).
To determine the correlation between gene copy number and mRNA expression in tumors, we performed genome-wide gene expression analysis of 58 tumors (tub 1, 24; tub 2, 18; por 1, 16) that had been analyzed by aCGH. By calculating the Pearson correlation coefficient (r) between the aCGH signal ratio detected using the nearest neighbor bacterial artificial chromosome clone and the expression signal value for each gene in the same samples, we identified genes whose expression was correlated with CNA. All the genes showing r>0.5 are listed (Supplementary Table S5) and well-known oncogenes, such as EGFR (r=0.96), FGFR2 (r=0.94), CCNE1 (r=0.86), CCND1 (r=0.76) and ERBB2 (r=0.76), exhibited significant concordance with increased gene copy number. These analyses help us to identify bona fide pathogenetic oncogene in gastric cancer.
Chromosome 6p21 genomic amplification in gastric cancer
Chromosomal amplification at the 6p21 locus has been detected in gastric cancer4,7, 8, 9 and in other tumors.10, 11, 12, 13, 14 In this study, we also detected frequent gains (signal ratio⩾1.3, up to 32.5% of tumors) and high-level amplifications (signal ratio⩾2.5, up to 8.6% of tumors) in the chromosomal band. Most of the 6p21 copy number gains/amplifications in primary cases were spread over a rather broad region, and high-level amplifications occurred intensively in the region of 6p21.2-p21.1 (Figure 1a). Multivariate analysis using the Cox proportional hazards model revealed that copy number gain detected by RPCI11–89L17 at 6p21.1 was an independent factor related to overall survival of gastric cancer, in addition to the depth of invasion and the status of lymph node and extranodal metastasis (Table 1). This observation strongly suggested that chromosome 6p21 gain/amplification has an important role in the pathogenesis of gastric cancer.
In the 6p21 region, most genes showing significant correlations between expression patterns and copy number gains were located within the more proximal region of 6p21.2-p21.1 (Figure 1b). Fifty-nine (49.2%) out of the 120 genes in this region showed significant correlation with increased copy number. The list included candidate genes that have been reported as targets for 6p21 amplification or as cancer-related genes in various tumors (Supplementary Table S7).
Loss-of-function assay shows that multiple genes at the 6p21 locus are associated with gastric cancer cell growth
To isolate potential oncogenes at this locus, we used functional assays. We first tested endogenous growth-promoting activity using comprehensive small interfering RNA (siRNA) knockdown on two gastric cancer cell lines HSC58 and HSC60 with moderate copy number increase (signal ratio ≈1.3) at the 6p21 locus (Supplementary Figure S2). Downregulation of 58 out of the 59 candidate genes in the 6p21 locus was successful using specific siRNAs. This screening identified 45 and 40 siRNAs, which inhibited cell growth at <70% compared with the control siRNA in HSC58 and HSC60 cells, respectively. In total, downregulation of 36 genes caused concordant reduction of cell growth in both cell lines (Figure 2a).
Multiple genes at the 6p21 locus confer colony-forming activity in vitro and tumorigenicity in vivo
To further evaluate the oncogenic activities attributable to their overexpression, we introduced the candidate genes on the 6p21 amplicon into HEK293 epithelial cells and tested whether their overexpression enhanced colony-forming activity in vitro. Sixteen of the 51 genes examined enhanced the colony-forming activity: (1) PPIL1, (3) MTCH1, (5) TBC1D22B, (10) GLO1, (11) C6orf64, (13) C6orf130, (20) CCND3, (25) TBCC, (30) GNMT, (32) MEA1, (38) PTK7, (40) C6orf108, (45) POLR1C, (48) GTPBP2, (51) VEGFA and (58) AARS2 (Figure 2b). Reproducible colony-forming activities of NIH3T3 fibroblast cells were observed at similar stimulation rates in almost all of them (Table 2).
We then attempted to examine whether the six genes showing more than two-fold enhancement of in vitro cell growth could confer in vivo tumorigenicity by transplanting polyclonal cells expressing them into mice. Overexpression of the six genes led to tumor development in vivo with frequencies ranging from 17% to 83% within 12 weeks (Supplementary Figure S3). The results indicated that at least these six genes upregulated in the 6p21 amplicon had the individual potential to stimulate cell growth and form tumors in vivo (Table 2). Among them, GLO1 showed the strongest tumor-forming ability, with the tumors developing at the earliest time (4–5 weeks after injection), growing more quickly and attaining a larger size than the others.
GLO1 exhibits a stimulated oncogenic activity in cooperation with other genes on 6p21
Integration of copy number analysis, gene expression analysis and three different kinds of functional analyses identified GLO1 as the most likely oncogene. However, each 6p21 amplicon in primary cases usually lies within a wide range and is accompanied by coamplification of multiple potential oncogenes (Supplementary Figure S1b), we hypothesized that there might be cooperative tumorigenic activity between GLO1 and other genes within the 6p21 amplicon. We double-transfected seven genes into a GLO1-expressing clone (HEK293-GLO1-Zeo) and tested their growth-stimulatory activities. Significant enhancement of colony formation was detected for all the genes examined, except for PPIL1 (Figures 2c and d), indicating that GLO1 confers a broad synergistic effect on tumor formation with other genes on 6p21.
In addition to the high-level amplification of the GLO1 gene, we searched for somatic mutation of the GLO1 gene in our gastric cancer cohort (72 tumors) and 23 cell lines. However, we detected germline variations only in the 5′-untranslated region of exon 1, in intron 1 and in the coding exon 4 (Supplementary Figure S4). GLO1 is a ubiquitous detoxifying enzyme of methylglyoxal (MG) that is a potent glycating agent and induces oxidative stress and apoptosis.15 A recent report has demonstrated that upregulation of GLO1 expression occurs in response to oxidative stress, by binding a stress-responsive transcription factor Nrf2 to the ARE (antioxidant-response element) in 5′-untranslated region of exon 1 (from−19 to−10, numbered from the start codon).16 Whether the variations found in this study relate to the regulation by Nrf2 remains unknown.
GLO1 is a novel metabolic oncogene affecting gastric cancer cell growth by regulating energy producing pathways
We tried to decrease the expression of GLO1 stably in eight gastric cancer cell lines (HSC41, HSC43, HSC44, HSC45, HSC58, HSC60, MKN28 and NUGC3) using a lentiviral short hairpin RNA (shRNA) and were able to isolate seven of them (with the exception of HSC58) exhibiting reductions of expression ranging from 11% to 69% relative to the negative control. Remarkable reduction of cell growth was detected in HSC43, HSC44, HSC45, HSC60, MKN28 and NUGC3 (Figure 3a) but not in HSC41 (data not shown). GLO1 stably knockdown HSC60 clones exhibited the most severe growth retardation and ceased proliferating within a month after isolation. Together with the failure to establish HSC58-GLO1-knockdown clones, the data indicated that the growth of some gastric cancer cells is strongly dependent on GLO1 function, whereas the dependency is less marked in others.
Because GLO1 is an enzyme that detoxifies MG generated during glycolysis, we assumed that downregulation of GLO1 would reduce glycolytic activity by accumulation of MG.17 The intracellular metabolites of key pathways for energy production were analyzed by capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS), which enables comprehensive and quantitative analysis of charged metabolites.18 The CE-TOFMS system identified and quantified 239 and 228 candidate compounds comparable in NUGC3-shGLO1 vs NUGC3-scr and MKN28-shGLO1 vs MKN28-scr, respectively. Unexpectedly, lower concentration of almost all the intermediates that are included in glycolysis, the pentose phosphate pathway and tricarboxylic acid cycle were detected in both NUGC3-shGLO1 and MKN28-shGLO1 cells compared with their respective reference cells (Table 3 and Supplementary Figure S5). In addition, we found relative lower levels of glutamine/glutamic acid than that of the other amino acids. It might indicate glutaminolysis activation to compensate for reduced central carbon metabolism as a secondary effect, in which glutamine converts to lactic acid via tricarboxylic acid cycle.19
Suppression of gastric cancer cell growth by S-p-bromobenzyl glutathione cyclopentyl diester (BBGC), a glyoxalase I inhibitor
We further tested whether a GLO1 inhibitor, BBGC, was able to exert an anti-growth effect on gastric cancer. We treated 24 gastric cancer cell lines with a range of BBGC concentrations and found that 16 cell lines showed reproducible reduction of growth (the IC50 value=4.4–13.1 μM, Figure 3b). Previous studies have demonstrated that sensitivity to BBGC is correlated with GLO1 overexpression/amplification in cell lines.20, 21, 22 Because of the relatively low-level copy number gain at the GLO1 locus in the cell lines (signal ratios 1.20~1.34), we were unable to identify any correlations between higher sensitivity to BBGC and increased GLO1 gene dosage (Figure 3c, left). We measured GLO1 activity and mRNA expression levels of GLO1 and GLO2, encoding glyoxalase II (GLO2), another constituent of the glyoxalase system.15 No correlation between the BBGC sensitivity and GLO1 activity (Figure 3c, middle) or GLO1 mRNA quantity (data not shown) was evident. We found that the expression ratio of GLO1/GLO2 mRNA is high in BBGC-insensitive compared with BBGC-sensitive cells (1.238 vs 0.478). The difference was statistically significant at P=0.0162 (Figure 3c, right).
GLO1 regulates the transcriptional activities of NF-κB and AP-1 in gastric cancer cells
A recent study has demonstrated that overexpression of GLO1 suppresses basal and tumor necrosis factor-induced NF-κB activity in HEK293 cells.23 As activation of NF-κB has been shown to have critical oncogenic roles in a variety of solid tumors including gastric cancer,24, 25, 26 we examined the transcriptional activities of NF-κB, AP-1 and signal transducer and activator of transcription factor 3 (STAT3) using a stably GLO1-overexpressing HEK293 clone. The GLO1 activity of HEK293-GLO1 cells was increased by 1.8-fold, and the transcriptional activities of NF-κB and AP-1 were increased by 2.0-fold and 1.9-fold relative to the basal levels, respectively (Figures 4a and b). In contrast, no increase of STAT3 transcriptional activity was seen. On the other hand, cells stably expressing GLO1-shRNA showed decreased GLO1 activity (14–36% of the negative control) (Figure 4c), and the basal activities of NF-κB and AP-1 were significantly suppressed in GLO1-downregulated HSC44, NUGC3 and MKN28 cells (Figure 4d). These results are consistent with the reverse observations in HEK293 GLO1-overexpressing cells, indicating that GLO1 expression and/or GLO1 activity is positively linked to the transcriptional activities of NF-κB and AP-1.
To confirm NF-κB activation by GLO1 overexpression, we searched for NF-κB target genes upregulated in the GLO1-overexpressing tumors. This was carried out by comparing the expression level for each gene among the three tumor groups divided by the expression level of GLO1: GLO1-high, averaged expression level=2139 (n=18); GLO1-medium, averaged expression level=1080 (n=26, P=1 × 10−8); and GLO1-low, averaged expression level=785 (n=14, P=4 × 10−10). Twenty-one NF-κB target genes showed differential gene expression between GLO1 highly expressing tumors and others among the 463 NF-κB target genes (Supplementary Table S8). The upregulated gene list includes apoptosis/anti-apoptosis-related genes, tumor invasion/metastasis-related genes, oncogenes E2F3 and MCTS1, an NF-κB activator RIPK2 and a glycolytic enzyme PGK1. We confirmed that some of these NF-κB target genes were almost downregulated in gastric cancer cells stably expressing GLO1 shRNA compared with those expressing non-targeting shRNA (Figure 4e).
In this study, we identified a set of 808 genes whose overexpression correlated with copy number gain from 107 gain/amplification loci (Supplementary Table S5) and a set of 83 genes whose underexpression correlated with copy number loss from 41 hemizygous/homozygous deletion loci (Supplementary Table S6) in gastric cancer. In particular, genes showing strong correlation (r⩾0.7) appeared to be potential targets for frequent gene amplification, for example, FAM84B and PVT1 at the 8q24.2 MYC locus and C20orf43 and RAB22A at the 20q13.2 BCAS1 locus, by analogy with the patterns of well-known oncogenes.
The frequency of high-level 6p21 amplification encompassing a stretch of 8 megabases was second only to the ERBB2 locus, and the copy number gains at chromosome band 6p21.1 was an independent prognostic factor of overall survival of gastric cancer patients (Table 1). We expected that there would be multiple pathogenetic genes for gastric cancer in this region, because a tight and strong correlation between CNAs and gene expression was observed in almost half of the genes mapped (Figure 1b). Gene knockdown analysis using siRNA revealed that downregulation of 36 genes inhibit the growth in both HSC58 and HSC60 cells (Figure 2a). The results clearly demonstrated that many of the upregulated 6p21 genes have a role in the viability of gastric cancer cells. These include several cancer-related genes such as PPIL1,27 PIM1,28 GLO1,21,22 TFEB,29 CCND3,11,12 CUL7,30 C6orf108,31 VEGFA14 and CDC5L.13 To focus generally on growth-promoting genes, we examined activities for in vitro colony formation and in vivo tumor formation using HEK293 and NIH3T3 cells, because no immortalized gastric epithelial cells were available. More than two-fold enhanced colony-forming activities were observed in six genes by enforced overexpression (Figure 2b). In vivo tumor-formation efficiencies were: GLO1, 83%; C6orf64, 67%; GNMT, 38%; GTPBP2, 33%; C6orf130, 17%; and TBCC, 17% (Table 2). These results demonstrated that GLO1 is the most likely oncogene in the 6p21 amplicon. Furthermore, these genes and some others such as CCND3, C6orf108 and VEGFA showed marked enhancement of growth-promoting activity in the presence of GLO1 overexpression (Figure 2c). Therefore, we concluded that the 6p21 amplicon includes at least six potential oncogenes and several cell growth-supporting genes and that all of these genes might function cooperatively in cancer formation and/or progression. Recent reports have indicated that GNMT encodes an enzyme glycine N-methyltransferase, involved in methionine metabolism, which has a role in liver regeneration,32 and that C6orf108 encodes an enzyme, deoxynucleoside 5′-monophosphate N-glycosidase, which is implicated in purine or pyrimidine salvage and stimulation of glycolysis by supplying deoxyribose 5-phosphate (P) as an energy source.33 Chromosomal amplification may be the most efficient mechanism for activating a subset of 6p21 genes favorable for cancer cell growth. Although a recent search has identified GLO1 to be the commonly amplified gene in various cancer cells,22 our study is the first to provide experimental evidence that GLO1 exhibited tumorigenic potential.
The glyoxalase system comprises two enzymes, GLO1 and GLO2, and sequentially metabolizes cytotoxic MG to D-lactate via the intermediate S-D-lactoylglutathione, using reduced glutathione as a cofactor (Figure 4f). MG is mainly produced from dihydroxyacetone-P or glyceraldehyde 3-P during glycolysis. Accumulation of MG causes oxidative stress and damage to cells, then leads to apoptosis.34 Therefore, overexpression of GLO1 in a rapidly proliferating cancer cell may confer a growth advantage by reducing MG stress, in which aerobic glycolysis (the Warburg effect) is predominantly active.35, 36, 37 Metabolome analysis has demonstrated an enhanced glycolysis and glutaminolysis, which is an alternative energy-producing pathway involving a glucose-independent tricarboxylic acid cycle, in gastric cancer tissues.18 In this study, we observed that the concentration of many metabolic intermediates involved in central carbon metabolism was significantly lowered in both gastric cancer cells NUGC3 and MKN28 when GLO1 is stably downregulated (Table 3 and Supplementary Figure S5). We also showed that the expression ratio of GLO1/GLO2 mRNA tends to be high in BBGC-insensitive relative to BBGC-sensitive cells (Figure 3c, right).
Furthermore, GLO1 was found to function as a modulator of the transcriptional activities of NF-κB and AP-1 (Figures 4a–d). The mechanism as to how GLO1 influences on the activities of NF-κB and AP-1 is not clear, but two recent reports have demonstrated that GLO1 silencing or MG administration in prostate cancer cell induces mitochondrial apoptotic pathway through inactivation of NF-κB by the decrease in serine 32-phosphorylated IκBα and the increase in total IκBα levels.38,39 A previous promoter sequence analysis of GLO1 gene has indicated a possible positive regulation by NF-κB and AP-1 under oxidative stress.40 Also, GLO1 has been shown to be a target of another stress-responsive transcription factor Nrf2 that functions for the sustained activation of phosphatidylinositol 3′-kinase-AKT signaling and the metabolic reprogramming to contribute maintenance of malignant phenotype of cancer.41 We identified 21 upregulated NF-κB target genes in a class of GLO1 highly expressing gastric tumors, including an NF-κB activator RIPK2 and a glycolytic enzyme PGK1 (Supplementary Table S8).
Based on these observations, GLO1 would belong to a novel type of metabolic oncogene that regulates both cancer-specific metabolism and gene regulation, and IDH1 is a prototype of this category.42 Figure 4f schematically shows metabolomic reprogramming, especially in glycolysis, by cancer-related genes,35,36 including our findings: that chromosome 6p21 amplification causes multiple gene overexpression, including several metabolic oncogenes such as GLO1 and C6orf108; that GLO1 activates glycolysis through controlling MG/oxidative stress; that GLO1 overexpression upregulates the transcriptional activities of NF-κB and AP-1; and that the NF-κB target genes, such as an NF-κB activator RIPK2 and a glycolytic enzyme PGK1, are differentially upregulated in GLO1-overexpressing gastric tumors.
In conclusion, the present data indicate that 6p21 genomic amplification has considerable clinical significance and that the locus includes a number of possible therapeutic targets for gastric cancer. In particular, several novel enzymatic activities would be feasible molecular targets for the development of new drugs. GLO1 is one of such attractive targets for gastric cancer, and development of an appropriate indicator of its sensitivity to inhibitors is warranted to ensure effective treatment.
Materials and methods
Clinical samples and cell lines
Gastric tissue samples were collected from patients who underwent gastric surgery at the National Cancer Center Hospital, Tokyo, Japan between 1998 and 2001 as described previously.43 Clinicopathological features of the 163 cases of gastric cancer are summarized (Supplementary Table S1). This study was approved by the Ethics Committee of the National Cancer Center and written informed consent was obtained from all patients. Human gastric cancer cell lines, human HEK293 and mouse NIH3T3 cell lines were obtained from the Riken Cell Bank (Ibaraki, Japan) or the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan). The HSC series of human gastric cancer cell lines were established at our institution and have been characterized previously.44 Identification of cancer cell line was performed using Cell ID System (Promega, Madison, WI, USA; Supplementary Table S2).
aCGH and expression microarray analyses
DNA sample preparation from gastric tumor cells and matched normal gastric epithelial cells isolated by laser capture microdissection and array hybridization were carried out as described previously,43 using our custom-made bacterial artificial chromosome arrays, MCG Whole Genome Array-4500 and MCG Cancer Array-800.5,43 Total RNA was extracted from frozen gastric tissue samples by homogenizing in TRIZOL Reagent (Invitrogen, San Diego, CA, USA) and purified using RNeasy MinElute Cleanup Kit (Qiagen, Hilden, Germany). Comparative expression analysis was performed using a GeneChip Human Genome U133 Plus 2.0 Array in accordance with the manufacturer’s protocols (Affymetrix, Santa Clara, CA, USA).45
Gene knockdown analysis with siRNA or lentiviral shRNA
Pre-designed siRNA (Dharmacon siGenome SMART pool) targeting the 6p21 individual genes and a non-targeting siRNA were purchased from Thermo Scientific (Lafayette, CO, USA). The shRNA sequences targeted for GLO1 (5′-IndexTermGTTCTTGGAATGACGCTAA-3′) and for a non-targeting negative control (5′-IndexTermTAATATCGAGTATGCTCGG-3′, designed by B-Bridge International at Cupertino, CA, USA) were synthesized together with the 5′- and 3′-flanking sequences and cloned into the pSIF1-H1-Puro vector (System Biosciences, Mountain View, CA, USA). Sequences are as follows: GLO1-sense: 5′-IndexTermGATCCGTTCTTGGAATGACGCTAACTTCCTGTCAGATTAGCGTCATTCCAAGAACTTTTTG-3′; GLO1-antisense: 5′-IndexTermAATTCAAAAAGTTCTTGGAATGACGCTAATCTGACAGGAAGTTAGCGTCATTCCAAGAACG-3′; non-targeting-sense: 5′-IndexTermGATCCGTAATATCGAGTATGCTCGGCTTCCTGTCAGACCGAGCATACTCGATATTATTTTTG-3′; and non-targeting-antisense: 5′-IndexTermAATTCAAAAATAATATCGAGTATGCTCGGTCTGACAGGAAGCCGAGCATACTCGATATTACG-3′. Human gastric cancer cells were transfected with each siRNA using Lipofectamine 2000 (Invitrogen) or infected with lentivirus at a multiplicity of infection of 5. The infected cells were selected by incubation with 2 μg/ml puromycin (Sigma-Aldrich, St Louis, MO, USA), and the cell clones stably expressing a decreased amount of GLO1 mRNA were isolated. For transient assay, reverse transcription (RT) was performed on cell lysates 1 day after transfection using a FastLane Cell RT-PCR kit (Qiagen). Real-time PCR was carried out using the universal probe library (Roche, Indianapolis, IN, USA) and gene-specific PCR primers or TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA, USA) (Supplementary Table S3). Cell growth inhibition was evaluated by CellTiter-Glo Luminescent Cell Viability Assay (Promega).
Construction of gene expression vectors, plasmid transfection and colony-formation assay
Individual full-length cDNAs of the 6p21 genes were synthesized from a commercial human stomach cDNA (Clontech, Mountain View, CA, USA) or gastric cancer cell line cDNA by nested PCR amplification using gene-specific primers (Supplementary Table S4) and cloned into the pcDNA3.1D/V5-His-TOPO vector (Invitrogen). For double transfection experiments, GLO1 cDNA (+V5-His tag) was transferred to the pcDNA3.1/Zeo (+) vector. Introduction of recombinant plasmid DNA into HEK293 or NIH3T3 cells was performed using an Amaxa Nucleofector electroporator (Lonza, Basel, Switzerland). Efficiency of colony formation was evaluated by counting Giemsa-stained colonies after 12 days of growth under drug selection.
Tumor-formation assay in nude mice
In all, 1 × 106 HEK293- or NIH3T3-derived polyclonal cells expressing the 6p21 genes in 0.2 ml of phosphate-buffered saline were injected subcutaneously on the flanks of 8-week-old female BALB/c nude mice (CLEA Japan, Tokyo, Japan) on day 0. Tumor sizes were measured in two dimensions at once a week for 3 months. The protocol of the experiment was approved by the Committee for Ethics in Animal Experimentation and conducted in accordance with the Guidelines for Animal Experiments of the National Cancer Center.
Promoter reporters and dual luciferase assays
One hundred thousand cells in a 24-well format were transfected with 0.4 μg each of pNFκB-Luc, pAP1-Luc or pSTAT3-TA-Luc (Clontech) using Lipofectamine 2000. Transfection efficiency was normalized by cotransfection with 0.04 μg of pRL-TK (Promega) containing a full-length renilla luciferase gene under control of the herpes simplex virus thymidine kinase promoter. Twenty-four hours after transfection, the cells were lysed in passive lysis buffer. Firefly luciferase and renilla luciferase activities were quantified using the Dual-Luciferase Reporter Assay System and a GloMax 96 Microplate Luminometer (Promega).
GLO1 enzyme assay and cytotoxicity assay
GLO1 enzyme activity was analyzed according to a standard spectrophotometric method monitoring the rate of formation of S-D-lactoylglutathione.46 The drug sensitivity of human gastric cancer cell lines was evaluated by cell growth inhibition. Five thousand cells in a 96-well format were incubated with various concentrations of BBGC for 24 h,20 and then the cell viability was estimated by CellTiter-Glo Luminescent Cell Viability Assay.
Intracellular metabolites in the GLO1-knockdown gastric cancer cell and its reference cell grown in a standard medium were extracted by ice-cold methanol according to the manufacturer’s instructions. All charged metabolic intermediates were quantitatively analyzed using a CE-TOFMS method by HMT (Human Metabolome Technologies, Inc., Tsuruoka, Japan).18
Statistical analyses were performed using the JMP5.1 (SAS Institute Japan, Tokyo, Japan) and Statview 5.0 (Abacus Concepts, Piscataway, NJ, USA) statistical software packages. The overall survival curve was calculated by JMP5.1 using the Kaplan–Meier method and log-rank test. The Cox proportional hazards model was used for multivariate analysis to determine independent factors related to survival and recurrence, based on the variables selected by univariate analysis in Statview 5.0.
aCGH, array-based comparative genomic hybridization; AP-1, activator protein-1; BAC, bacterial artificial chromosome; BBGC, S-p-bromobenzyl glutathione cyclopentyl diester; CE-TOFMS, capillary electrophoresis-time-of-flight mass spectrometry; CNA, copy number alteration; GLO1, glyoxalase I; GLO2, glyoxalase II; MG, methylglyoxal; NF-κB, nuclear factor kappa B; PPP, pentose phosphate pathway; P, phosphate; shRNA, short hairpin RNA; siRNA, small interfering RNA; TCA, tricarboxylic acid.
Lochhead P, El-Omar EM . Gastric cancer. Br Med Bull 2008; 85: 87–100.
Smith MG, Hold GL, Tahara E, El-Omar EM . Cellular and molecular aspects of gastric cancer. World J Gastroenterol 2006; 12: 2979–2990.
Tamura G . Alterations of tumor suppressor and tumor-related genes in the development and progression of gastric cancer. World J Gastroenterol 2006; 12: 192–198.
Gorringe KL, Boussioutas A Melbourne Gastric Cancer Group, Peter Mac Microarray Facility, Bowtell DDL . Novel regions of chromosomal amplification at 6p21, 5p13, and 12q14 in gastric cancer identified by array comparative genomic hybridization. Genes Chromosomes Cancer 2005; 42: 247–259.
Takada H, Imoto I, Tsuda H, Nakanishi Y, Ichikura T, Mochizuki H et al. ADAM23, a possible tumor suppressor gene, is frequently silenced in gastric cancers by homozygous deletion or aberrant promoter hypermethylation. Oncogene 2005; 24: 8051–8060.
Vauhkonen H, Vauhkonen M, Sajantila A, Sipponen P, Knuutila S . DNA copy number aberrations in intestinal-type gastric cancer revealed by array-based comparative genomic hybridization. Cancer Genet Cytogenet 2006; 167: 150–154.
Yang S, Jeung H-C, Jeong HJ, Choi YH, Kim JE, Jung J-J et al. Identification of genes with correlated patterns of variations in DNA copy number and gene expression level in gastric cancer. Genomics 2007; 89: 451–459.
Myllykangas S, Junnila S, Kokkola A, Autio R, Scheinin I, Kiviluoto T et al. Integrated gene copy number and expression microarray analysis of gastric cancer highlights potential target genes. Int J Cancer 2008; 123: 817–825.
Tsukamoto Y, Uchida T, Kaman S, Noguchi T, Nguyen LT, Tanigawa M et al. Genome-wide analysis of DNA copy number alternations and gene expression in gastric cancer. J Pathol 2008; 216: 471–482.
Santos GC, Zielenska M, Prasad M, Squire JA . Chromosome 6p amplification and cancer progression. J Clin Pathol 2007; 60: 1–7.
Tanami H, Tsuda H, Okabe S, Iwai T, Sugihara K, Imoto I et al. Involvement of cyclin D3 in liver metastasis of colorectal cancer, revealed by genome-wide copy-number analysis. Lab Invest 2005; 85: 1118–1129.
Kasugai Y, Tagawa H, Kameoka Y, Morishima Y, Nakamura S, Seto M . Identification of CCND3 and BYSL as candidate targets for the 6p21 amplification in diffuse large B-cell lymphoma. Clin Cancer Res 2005; 11: 8265–8272.
Lu X-Y, Lu Y, Zhao Y-J, Jaeweon K, Kang J, Xiao-Nan L et al. Cell cycle regulator gene CDC5L, a potential target for 6p12-p21 amplicon in osteosarcoma. Mol Cancer Res 2008; 6: 937–946.
Chiang DY, Villanueva A, Hoshida Y, Peix J, Newell P, Minguez B et al. Focal gains of VEGFA and molecular classification of hepatocellular carcinoma. Cancer Res 2008; 68: 6779–6788.
Thornalley PJ . The glyoxalase system: new developments towards functional characterization of a metabolic pathway fundamental to biological life. Biochem J 1990; 269: 1–11.
Xue M, Rabbani N, Momiji H, Imbasi P, Anwar MM, Kitteringham N et al. Transcriptional control of glyoxalase 1 by Nrf2 provides a stress-responsive defence against dicarbonyl glycation. Biochem J 2012; 443: 213–222.
Hansen F, de Souza DF, Silveira Sda L, Hoefel AL, Fontoura JB, Tramontina AC et al. Methylglyoxal alters glucose metabolism and increases AGEs content in C6 glioma cells. Metab Brain Dis 2012; 27: 531–539.
Hirayama A, Kami K, Sugimoto M, Sugawara M, Toki N, Onozuka H et al. Quantitative metabolome profiling of colon and stomach cancer microenvironment by capillary electrophoresis time-of-flight mass spectrometry. Cancer Res 2009; 69: 4918–4925.
DeBerardinis RJ, Mancuso A, Daikhin E, Nissim I, Yudkoff M, Wehrli S et al. Beyond aerobic glycolysis: Transformed cells can engage in glutamine metabolism that exceeds the requirement for protein and nucleotide synthesis. Proc Natl Acad Sci USA 2007; 104: 19345–19350.
Ranganathan S, Walsh ES, Tew KD . Glyoxalase I in detoxification: studies using a glyoxalase I transfectant cell line. Biochem J 1995; 309: 127–131.
Sakamoto H, Mashima T, Sato S, Hashimoto Y, Yamori T, Tsuruo T . Selective activation of apoptosis program by S-p-bromobenzylglutathione cyclopentyl diester in glyoxalase I-overexpressing human lung cancer cells. Clin Cancer Res 2001; 7: 2513–2518.
Santarius T, Bignell GR, Greenman CD, Widaa S, Chen L, Mahoney CL et al. GLOI—a novel amplified gene in human cancer. Genes Chromosomes Cancer 2010; 49: 711–725.
de Hemptinne V, Rondas D, Toepoel M, Vancompernolle K . Phosphorylation on Thr-106 and NO-modification of glyoxalase I suppress the TNF-induced transcriptional activity of NF-κB. Mol Cell Biochem 2009; 325: 169–178.
Maeda S, Omata M . Inflammation and cancer: Role of nuclear factor-kappaB activation. Cancer Sci 2008; 99: 836–842.
Aggarwal BB, Gehlot P . Inflammation and cancer: How friendly is the relationship for cancer patients? Curr Opin Pharmacol 2009; 9: 351–369.
Kim DJ, Park K-S, Kim J-H, Yang S-H, Yoon JY, Han B-G et al. Helicobacter pylori proinflammatory protein up-regulates NF-κB as a cell-translocating Ser/Thr kinase. Proc Natl Acad Sci USA 2010; 107: 21418–21423.
Obama K, Kato T, Hasegawa S, Satoh S, Nakamura Y, Furukawa Y . Overexpression of peptidyl-prolyl isomerase-like 1 is associated with the growth of colon cancer cells. Clin Cancer Res 2006; 12: 70–76.
Chen J, Kobayashi M, Darmanin S, Qiao Y, Gully C, Zhao R et al. Hypoxia-mediated up-regulation of Pim-1 contributes to solid tumor formation. Am J Pathol 2009; 175: 400–411.
Kuiper RP, Schepens M, Thijssen J, van Asseldonk M, van den Berg E, Bridge J et al. Upregulation of the transcription factor TFEB in t(6;11)(p21;q13)-positive renal cell carcinomas due to promoter substitution. Hum Mol Genet 2003; 12: 1661–1669.
Kim SS, Shago M, Kaustov L, Boutros PC, Clendening JW, Sheng Y et al. CUL7 is a novel antiapoptotic oncogene. Cancer Res 2007; 67: 9616–9622.
Lewis BC, Prescott JE, Campbell SE, Shim H, Orlowski RZ, Dang CV . Tumor induction by the c-Myc target genes rcl and lactate dehydrogenase A. Cancer Res 2000; 60: 6178–6783.
Varela-Rey M, Fernández-Ramos D, Martínez-López N, Embade N, Gómez-Santos L, Beraza N et al. Impaired liver regeneration in mice lacking glycine N-methyltransferase. Hepatology 2009; 50: 443–452.
Ghiorghi YK, Zeller KI, Dang CV, Kaminski PA . The c-MYC target gene Rcl (C6orf108) encodes a novel enzyme, deoxynucleoside 5’-monophosphate N-glycosidase. J Biol Chem 2007; 282: 8150–8156.
Thornalley PJ . Protein and nucleotide damage by glyoxal and methylglyoxal in physiological systems–role in aging and disease. Drug Metabol Drug Interact 2008; 23: 125–150.
Vander Heiden MG, Cantley LC, Thompson CB . Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 2009; 324: 1029–1033.
Levine AJ, Puzio-Kuter AM . The control of the metabolic switch in cancers by oncogenes and tumor suppressor genes. Science 2010; 330: 1340–1344.
Bair III WB, Cabell CM, Uchida K, Bause AS, Wondrak GT . GLO1 overexpression in human malignant melanoma. Melanoma Res 2010; 20: 85–96.
Antognelli C, Mezzasoma L, Fettucciari K, Mearini E, Talesa VN . Role of glyoxalase I in the proliferation and apoptosis control of human LNCaP and PC3 prostate cancer cells. Prostate 2013; 73: 121–132.
Antognelli C, Mezzasoma L, Fettucciari K, Talesa VN . A novel mechanism of methylglyoxal cytotoxity in prostate cancer cells. Int J Biochem Cell Biol 2013; 45: 836–844.
Ranganathan S, Ciaccio PJ, Walsh ES, Tew KD . Genomic sequence of human glyoxalase-I: analysis of promoter activity and its regulation. Gene 1999; 240: 149–155.
Mitsuishi Y, Taguchi K, Kawatani Y, Shibata T, Nukiwa T, Aburatani H et al. Nrf2 redirects glucose and glutamine into anabolic pathways in metabolic reprogramming. Cancer Cell 2012; 22: 66–79.
Figueroa ME, Abdel-Wahab O, Lu C, Ward PS, Patel J, Shih A et al. Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell 2010; 18: 553–567.
Nakamura Y, Migita T, Hosoda F, Okada N, Gotoh M, Arai Y et al. Krüppel-like factor 12 plays a significant role in poorly differentiated gastric cancer progression. Int J Cancer 2009; 125: 1859–1867.
Yanagihara K, Tanaka H, Takigahira M, Ino Y, Yamaguchi Y, Toge T et al. Establishment of two cell lines from human gastric scirrhous carcinoma that possess the potential to metastasize spontaneously in nude mice. Cancer Sci 2004; 95: 575–582.
Yagi T, Morimoto A, Eguchi M, Hibi S, Sako M, Ishii E et al. Identification of a gene expression signature associated with pediatric AML prognosis. Blood 2003; 102: 1849–1856.
Oray B, Norton SJ . Glyoxalase I from mouse liver. Methods Enzymol 1982; 90: 542–546.
We thank Setsuo Hirohashi for generous support and encouragement; Yukihiro Nakanishi for advice on histological classification of gastric tumors; Tokuki Sakiyama and Go Maeno for helping with the analysis of the aCGH data; Jun Yasuda for advice on the RNA interference techniques; Yu Nakamura, Michiyo Fukushima, Satomi Uryu and Yasuko Kuwabara for providing considerable contributions to the sample preparation of BAC DNA and patients’ DNA and array hybridization; and Sayaka Kadoguchi and Kenjiro Kami for advice on metabolome analysis. This research was supported in part by a Grant-in-Aid for the Comprehensive 10-Year-Strategy for Cancer Control and from the Ministry of Health, Labor and Welfare, Japan, a grant from the New Energy and Industrial Technology Development Organization (NEDO), Japan and a grant from the National Institute of Biomedical Innovation (NiBio), Japan. MM and NK were recipients of a Research Resident Fellowship from the Foundation for Promotion of Cancer Research in Japan.
The authors declare no conflict of interest.
Supplementary Information accompanies this paper on the Oncogene website
About this article
Cite this article
Hosoda, F., Arai, Y., Okada, N. et al. Integrated genomic and functional analyses reveal glyoxalase I as a novel metabolic oncogene in human gastric cancer. Oncogene 34, 1196–1206 (2015). https://doi.org/10.1038/onc.2014.57
Methylglyoxal, a Highly Reactive Dicarbonyl Compound, in Diabetes, Its Vascular Complications, and Other Age-Related Diseases
Physiological Reviews (2020)
International Journal of Molecular Sciences (2020)
International Journal of Molecular Sciences (2020)
Multi-Armed 1,2,3-Selenadiazole and 1,2,3-Thiadiazole Benzene Derivatives as Novel Glyoxalase-I Inhibitors
Biochimica et Biophysica Acta (BBA) - Reviews on Cancer (2019)