Homeobox gene IRX1 is a tumor suppressor gene in gastric carcinoma

Article metrics


The IRX1 tumor suppressor gene is located on 5p15.33, a cancer susceptibility locus. Loss of heterozygosity of 5p15.33 in gastric cancer was identified in our previous work. In this study, we analyzed the molecular features and function of IRX1. We found that IRX1 expression was lost or reduced in gastric cancer. However, no mutations were identified in IRX1-encoding regions. IRX1 transcription was suppressed by hypermethylation, and the expression of IRX1 mRNA was partially restored in gastric cancer cells after 5-Aza-dC treatment. Restoring IRX1 expression in SGC-7901 and NCI-N87 gastric cancer cells inhibited growth, invasion and tumorigenesis in vitro and in vivo. We identified a number of target genes by global microarray analysis after IRX1 transfection combined with real-time PCR and chromatin immunoprecipitation assay. BDKRB2, an angiogenesis-related gene, HIST2H2BE and FGF7, cell proliferation and invasion-related genes, were identified as direct IRX1 target genes. The hypermethylation of IRX1 was not only detected in primary gastric cancer tissues but also in the peripheral blood of gastric cancer patients, suggesting IRX1 could potentially serve as a biomarker for gastric cancer.


Chromosome 5p15.33 is a cancer susceptibility locus. Significant genetic variants of 5p15.33 have been reported in lung cancer in Western and Chinese populations (McKay et al., 2008; Jin et al., 2009). Gastric cancer is one of the common malignancies characterized by genomic instability with multiple genetic alterations, including oncogene activation and tumor suppressor gene inactivation. Loss of heterozygosity (LOH) is one type of genetic variation, and tumor suppressor genes are always located on LOH loci. Our group previously characterized a high frequency of LOH at 5p15.33 in human gastric cancer. We hypothesized that genes located on this locus may contribute to the pathogenesis of gastric cancer (Lu et al., 2005). IRX1 is harbored in this chromosome locus and belongs to the Iroquois homeobox gene family, which has six members from IRX1 to IRX6. IRX1 is closely related to embryonic development, including foregut organs such as the lung (Ferguson et al., 1998; Becker et al., 2001; van Tuyl et al., 2006; Alarcon et al., 2008). Because the stomach is also of foregut origin, the high frequency of LOH at the IRX1 locus in gastric cancer suggested that the IRX1 gene may be involved in the development of gastric cancer. Asaka et al. (2006) reported that IRX3 is downregulated in breast cancer patients with shorter survival times. IRX3 was also downregulated in androgen-insensitive prostate cancer cell lines (Zhao et al., 2005). Lewis et al. (1999) reported that IRX2 is involved in epithelial cell differentiation as well as ductal and lobular proliferation of breast cells. The association of the IRX1 gene with gastric cancer has not previously been studied.

A typical tumor suppressor gene should have the following features: the genetic alterations occur in both alleles (for example, one deleted and another mutated) or both alleles are deleted or mutated in cancer (Knudson, 1996, 2001). In the absence of an allelic mutation, reduced gene expression in cancer may be caused by an epigenetic abnormality such as hypermethylation of the promoter region (Garinis et al., 2002; Esteller, 2007). To date, significant effort has been directed at identifying tumor suppressor genes related to gastric cancer; however, these efforts have been unsuccessful thus far. In this study, we systematically analyzed the genetic structure, promoter activity and methylation status of the IRX1 gene. We also analyzed its functional features after reconstitution of IRX1 expression in vitro and in vivo. IRX1 functions as a transcription factor, but the exact role IRX1 has in the development of gastric cancer is currently unknown. The target genes and related pathways of IRX1 have not yet been identified.

cDNA microarrays provide a powerful tool for exploring complex gene expression profiles. Microarray analysis of experimental samples, such as gene-transfected cells, has led to identification of valuable molecular markers involved in tumor proliferation, angiogenesis, prognosis and therapeutic response (Duggan et al., 1999; Quackenbush, 2006; Perez-Diez et al., 2007; Iorns et al., 2009). Thus, we used a global cDNA microarray to identify downstream target genes of IRX1. We identified a number of target genes by global microarray analysis after IRX1 transfection combined with real-time PCR and chromatin immunoprecipitation assay. This work revealed that IRX1 is downregulated in gastric cancer. Downregulation of IRX1 due to LOH of 5p15.33, combined with epigenetic suppression of the retained allele, is the mechanism of IRX1 inactivation in gastric cancer.


Analysis of IRX1 expression, mutation, gene copy numbers and promoter region

Compared with the GES-1 gastric mucosa cell line, IRX1 mRNA was barely detectable in four gastric cancer cell lines and was decreased in three gastric cancer cell lines (P<0.001, Figure 1a, top). We examined the IRX1 copy numbers on all cell lines by real-time PCR, and four out of seven gastric cancer cell lines showed IRX1 copy number reduction compared with the GES-1 cell line (Figure 1a, bottom). We sequenced PCR products for four exons of the IRX1 gene from seven gastric cancer cell lines and the GES-1 control. No mutations were found except one single nucleotide polymorphism within exon 2 in four out of seven cancer cell lines and the GES-1 cell line, which does not alter the encoded amino acid (rs844154, Supplementary Figure 1). (Sequencing results for other reported single nucleotide polymorphisms are shown in Supplementary Table 1) We used online-accessible platforms to analyze the 5′-franking from the transcriptional start site of IRX1. The fragment −600 bp from the transcriptional start site was predicted as a probable promoter region (Figure 1b). We amplified a 775-bp fragment and inserted the fragment into a luciferase reporter vector (pGL3-Basic) for promoter activity analysis. A series of promoter/reporter fusion plasmids containing progressive 5′ deletions was constructed. The different constructs were transiently transfected into 293T or GES-1 cell lines and luciferase activities were measured. The fragment in pGL3-476 showed the highest activity compared with empty vector. The activity of the fragment in pGL3-248 was markedly decreased compared with pGL3-380 (P<0.001), which suggested that the core promoter region of IRX1 is situated on the inter-region of pGL3-248 and pGL3-380, accordant with the bioinformatics prediction (Figure 1c). The fold-changes of the different constructs compared with pGL3-Basic in 293T and GES-1 cells are summarized in Supplementary Table 2.

Figure 1

Gene expression, gene copy numbers and promoter analysis of IRX1. (a) IRX1 mRNA detection in gastric cancer cell lines and immortalized gastric mucosa cell line GES-1 by real-time PCR (top) and gene copy numbers (bottom). IRX1 mRNA expression was significantly reduced or nearly absent in cancer cells but retained in GES-1 cells, and four out of seven gastric cancer cell lines showed IRX1 copy number reduction compared with the GES-1 cell line. (b) Putative motifs of the 5′-flanking region of the IRX1 gene. Important putative motifs are identified and underlined. The transcriptional start site (TSS) is indicated as +1. (c) Progressive deletions and activity analysis of the IRX1 promoter by luciferase reporter. Progressive 5′-deletions were constructed (left side) and transiently transfected into 293 T or GES-1 cells. The promoter fragments were linked to luciferase (black box). Bars on the right side denote luciferase activities, represented as mean±s.d. The data are representative examples taken from one of three experiments.

Methylation analysis and the effect of demethylation treatment on cancer cell lines

We examined the methylation status of the promoter region by means of methylation-specific PCR (MSP) and bisulfate sequencing (BSP), which covered the regions of −236 to −388 and −519 to −679, respectively (Figure 2a). The core promoter region was methylated in all cancer cell lines according to MSP analysis. We treated cancer cell lines with 10 μM 5-Aza-2′-deoxycytidine (5-Aza-dC) for 96 or 120 h. The expression of IRX1 mRNA was restored or partially restored in some gastric cancer cell lines (Figure 2b). We examined the methylation status of 12 CpG sites within the CpG island of IRX1 by sequencing the PCR products in gastric cancer cell lines and the GES-1 control after bisulfate treatment. The 12 CpG sites of the IRX1 promoter were methylated in gastric cancer cell lines compared with GES-1 (P<0.001, Figures 2c and d), and methylation correlated inversely with IRX1 mRNA expression levels.

Figure 2

Methylation status of IRX1 CpG islands and the effects of 5-Aza-dC on IRX1 expression. (a) Schematic locations of CpG sites within the CpG islands that surround the IRX1 transcription start site by MSP and BSP analysis. (b) Upper: The core promoter region was methylated in all gastric cancer cell lines (marked as M) but unmethylated in the GES-1 cell line (marked as U). Lower: expression of IRX1 mRNA was restored in gastric cancer cell lines after 5-Aza-dC treatment (marked as E). (c) Methylated CpG site analysis by sequencing the PCR products. Compared with wild-type sequence (top letters), cytosines of methylated CpGs in gastric cancer cell lines are not converted, whereas unmethylated CpGs in GES-1 control cells are converted (indicated by arrow). (d) Schematic summary of 12 CpG sites in the promoter region from −519 to −679. Methylation analysis was performed in 10 clones for each cell line. Each row of circles represents a single clone, and each circle represents a single CpG site. Open circle represents unmethylated cytosine; filled circle represents methylated cytosine. All gastric cancer cell lines showed higher methylation levels compared with the GES-1 control. The methylation ratios of 10 clones for each cell lines are summarized in the lowest bar chart.

Functional analysis after restoring IRX1 expression in vitro

We introduced IRX1 expression in SGC-7901 and NCI-N87 gastric cancer cell lines by means of the pEGFP-N1 vector in which a reporter gene encoding enhanced green fluorescence protein (GFP) was under the control of a constitutively active upstream promoter. Successful transfection was determined by GFP fluorescence tracing in nuclei (Figure 3a) and confirmed by qPCR (P<0.001), western blot and immunocytochemical analysis based on IRX1 re-expression (Figure 3b). IRX1-transfected gastric cancer cells showed reduced colony formation and cell proliferation in vitro compared with control-transfected cells (P=0.003 and P=0.007, respectively; Figure 3c). Vector-transfected GFP+ control cells proliferated rapidly in the same culture conditions. We also examined cell migration and invasive ability by scratch healing assay (Figure 3d), transwell migration and transwell invasion assay (Figure 3e). The parental cells and empty vector-transfected cells nearly closed the wound 48 h after scratch, whereas IRX1-transfected cells were unable to heal the wound. The mean wound distances of the experimental sample and the control at 48 h were significantly different (326.844±64.08 μm vs 143.14±8.41 μm; P<0.001). The amount of migrated cells in the IRX1-transfected sample was significantly reduced compared with the control (60.50±6.72 vs 134.80±8.56, P<0.001). The amount of invasive cells in the SGC-7901/IRX1 group was also significantly reduced compared with the control (74.90±7.11 vs 169.20±6.96, P<0.001). All experimental results in NCI-N87 gastric cancer cells by IRX1 transfection showed the same tendency observed in the SGC-7901 cell line and are presented in Supplementary Figure 2.

Figure 3

Effect of IRX1 gene transfection on SGC-7901 gastric cancer cells. (a) Schematic figure for pEGFP-N1/IRX1 construction (top) and the effect of successful transfection of IRX1 (bottom). IRX1 transfection was confirmed by GFP uorescence in nuclei by fluorescence microscopy. (b) IRX1 mRNA and protein expression were induced after gene transfection and confirmed by qPCR, western blot and immunocytochemistry. The bar indicates 20 μm. (c) Colony formation (left) and cell growth activity (right) were suppressed after gene transfection based on soft agar and CCK-8 detection. (d) Cell migratory ability was suppressed by scratch wound healing assays. By wound distance analysis, cell migration ability was suppressed by IRX1 gene transfection. (e) Cell migration and invasive ability were suppressed in the transwell assay. These data are representative examples taken from one of three experiments.

Tumorigenicity after overexpressing IRX1 in nude mice

We next examined the effect of IRX1 overexpression by inoculating parental SGC-7901 or NCI-N87 cells, SGC-7901 or NCI-N87/vector and SGC-7901 or NCI-N87/IRX1 cells subcutaneously into the right flank regions of nude mice. Tumorigenicity was significantly reduced in IRX1-transfected cells. Rapid tumor growth was observed in the control groups after 1 month (Figure 4a). The tumor-inhibiting role of IRX1 transfection was apparent with both cancer cell lines (P=0.001, Figure 4b), with less mitosis and rare necrosis observed under the microscope (P<0.001, Figure 4c). These findings imply that IRX1 suppresses the growth of cancer cells.

Figure 4

Tumorigenesis analysis in nude mice. (a) Re-expressing IRX1 in SGC-7901 and NCI-N87 cells inhibited tumorigenesis in nude mice. Top represents the SGC-7901 group; bottom represents the NCI-N87 group. (b) Tumor nodules resected (left) from the IRX1-transfection group were smaller than those from controls. Tumorigenesis curves for the SGC-7901 or NCI-N87/IRX1 group and controls are shown on the right. Rapid tumor growth was observed in the control groups (data represented as the mean±s.d.; P=0.001). (c) Histological examination found that mitosis of tumor cells was markedly reduced in the IRX1-transfection group compared with the control groups (top, indicated by arrows, the bar indicates 20 μm). Necrosis is apparent in controls due to their rapid growth (down, the bar indicates 50 μm).

Identification of target genes after IRX1 gene transfection

We analyzed the genome-wide transcriptome profile of SGC-7901, SGC-7901/vector and SGC-7901/IRX1 cells by Agilent oligo microarray (41 000+). The microarray data set has been deposited in the GEO database (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE17399). According to fold-change (2.0) screening between SGC-7901/vector and SGC-7901/IRX1 cells, we found 224 upregulated genes and 324 downregulated genes (Supplementary Table 3). We searched genes that overlapped with cancer-associated and molecular function-related gene sets in MSigDB (C4 and C5 gene sets; http://www.broad.mit.edu/gsea/msigdb/index.jsp). We selected 42 cancer-associated genes for cluster mapping on the MeV microarray analysis platform (www.tm4.org/mev.html, Figure 5a). We also looked for genes related to molecular function and picked up the top 20 gene sets that overlapped with different function-clusters for exhibition (Figure 5b).

Figure 5

Target gene identification by global microarray analysis. (a) Clustering map of differentially expressed genes overlapped with cancer-associated genes set in the Molecular Signatures Database. Row represents gene, column represents experimental cells. Left column is IRX1-transfected cell; Middle column is SGC-7901 parental cell; Right is empty vector-transfected cell. Upregulated genes are shown in red and downregulated genes in green. (b) Fold-change map overlapped with molecular function-related genes set in the Molecular Signatures Database. Upregulated genes are listed on the left; and downregulated genes are listed on the right.

We chose 17 upregulated genes and 19 downregulated genes for real-time PCR verification and confirmed our microarray findings for 8 upregulated genes and 11 downregulated genes (Figures 6a and b). The gene names and functional annotations are listed in Table 1. To determine whether these are direct target genes of IRX1, we conducted chromatin immunoprecipitation assays using IRX1 antibody and then analyzed the pulled-down DNA. We identified three downregulated genes, BDKRB2 (NM_000623), FGF7 (NM_002009) and HIST2H2BE (NM_003528) as direct IRX1 targets (Figure 6c).

Figure 6

Verification of IRX1 target genes by real-time PCR. (a) Upregulation of PTGS2, ANPEP, KDR, UGT8, INHBA, ERMAP, RALGPS1 and SPON1 was confirmed. (b) Downregulation of BDKRB2, PHYHIPL, HIST2H2BE, FGF7, PTGER1, NPTX1, EGR1, COL9A3, CUGBP2, DKK3 and BPI was confirmed. (c) Chromatin immunoprecipitation analysis was done with IRX1 antibody using lysates from GES-1 cells. Promoter pulldowns were assessed through quantitative PCR, which revealed amplified BDKRB2, FGF7, and HIST2H2BE compared with the negative IgG control.

Table 1 Verified genes and their functions by real-time PCR

Methylation status of IRX1 in primary cancer tissues and peripheral blood of gastric carcinoma patients

We examined the methylation status of primary tumor tissues and associated plasma DNA from 15 gastric carcinoma patients (Table 2) and 10 age-matched healthy controls. The mean methylated levels of IRX1 in 15 primary tumor tissues and in plasma DNA were 51.90%±24.32 and 47.23%±35.93, respectively. The mean methylated level of plasma DNA in healthy control cases was 6.66%±11.71. (P=0.002; Figures 7a–d). Thus, significant methylation was observed in the cancer group compared with healthy controls. We analyzed the methylation level using multiple clinicopathological parameters. Higher methylation levels were closely associated with increased age (P=0.003) and TNM staging (P=0.051). The Receiver Operating Characteristics curve and the area under curve (AUC) for methylated blood DNA detection were calculated. The AUC was 0.827±0.083 and was significantly higher than that of the null hypothesis (true area was 0.5, P=0.007, Figure 7e). Thus, the level of methylated IRX1 in plasma could serve as a molecular marker for gastric cancer. To identify an optimal cut-off point to detect gastric cancer, Youden's index was used in this study. According to Youden's index, the optimal operating point of the blood methylation level of IRX1 was 26.4%. At this cut-off point, the sensitivity was 73.3% and specificity was 90%.

Table 2 Clinical information and blood methylation status of gastric cancer patients
Figure 7

Methylation analysis of primary tumor tissues and peripheral blood from gastric carcinoma patients and healthy controls. (a) Methylation status of 12 CpG sites of the IRX1 promoter from primary tumor tissues from 15 gastric carcinoma patients. (b) Methylation status of 12 CpG sites of the IRX1 promoter from plasma DNA from 15 gastric carcinoma patients. Each row of circles represents an integrated methylation ratio from three clones, and each circle represents a single CpG site. Open circle represents unmethylated cytosine; filled circle or partially filled circle represents methylated ratio. (c) Methylation status of 12 CpG sites of the IRX1 promoter from plasma DNA from 10 age-matched healthy controls. (d) The mean methylated level of cancer tissues and plasma in 15 patients was 51.90%±24.32 and 47.23%±35.93 (upper and middle). The mean methylation level of healthy controls was 6.66%±11.71 (lower). (e) The Receiver Operating Characteristics curve of blood-methylated DNA detection for IRX1. The AUC was 0.827±0.083 (P=0.007), with sensitivity and specificity of 73.3 and 90%, respectively.


Gastric cancer is the second most common cancer in the world. It is now apparent that multiple genetic alterations, including Helicobacter pylori infection, oncogene activation and tumor suppressor gene inactivation, are necessary steps in gastric cancer development (Telford et al., 1994; Ushijima et al., 2006). In a previous study, we revealed a relevant genetic variant at a high frequency of LOH locus at 5p15.33 in gastric carcinoma. The candidate gene IRX1 harbored on 5p15.33 was suspected as a tumor suppressor gene (Lu et al., 2005; Yu et al., 2006).

In this study, we analyzed the expression level, regulating mechanism and biological function of IRX1 in gastric carcinoma. To examine the second hit on IRX1 inactivation in addition to gene copy number deletion, sequence mutations and promoter methylation were analyzed. In addition, we reconstituted IRX1 expression in gastric cancer cells. Our results suggest that expression of IRX1 is significantly decreased in gastric cancer. The inactivation of IRX1 genes via a combination of allelic loss and promoter methylation is a common event in the development of gastric carcinogenesis. No sequence alterations (mutations) of IRX1 were found in gastric carcinoma. Our findings are consistent with the Knudson two-hit hypothesis of tumorigenesis. Our study also revealed that transcriptional silencing is the mechanism of IRX1downregulation in human gastric cancer. The high level of methylation on CpG sites of the IRX1 promoter was associated with a transcriptional block. Epigenetic events, especially promoter methylation, are now widely accepted causes of gene silencing (Kawakami et al., 2000; Robertson and Jones, 2000; Garinis et al., 2002; Momparler, 2003; Choi and Wu, 2005; Toyota and Issa, 2005; Esteller, 2007; Jee et al., 2009). 5-Aza-dC is a methyltransferase inhibitor that is widely used to restore gene expression (Bergman and Mostoslavsky, 1998; Tycko, 2000; Yamashita et al., 2002; Shaker et al., 2004; Hellebrekers et al., 2007; Schneider-Stock and Ocker, 2007; Hurtubise et al., 2008; Lemaire et al., 2008; Cairns, 2009). In the case of the IRX1 gene, introduction of 5-Aza-dC restored IRX1 expression in most of the cancer cells.

Hitherto, there has been very limited data on the functional association between the IRX1 gene and human cancer. Introducing a tumor suppressor gene to cancer cells is an important method for cancer research (Tamura et al., 1998, 1999). Therefore, we constructed an IRX1-expressing vector and restored IRX1 expression in SGC-7901 and NCI-N87 gastric cancer cells. These experiments revealed that restoring IRX1 can reduce growth and invasion of gastric carcinoma both in vitro and in vivo. The tumor suppressing function of the IRX1 gene was also confirmed by Bennett et al. (2008) recently. They found that introducing IRX1 expression in squamous cell carcinoma of the head and neck disclosed the gene's tumor suppressor potential. To clarify the precise mechanism of IRX1's tumor-inhibiting function, we screened the target genes regulated by IRX1 overexpression.

Microarray is a power technology that is able to perform genome-wide analysis in one experiment. Gene-expressing profiles can characterize genes that are differentially regulated in different experimental conditions. In this study, we compared changes in gene expression profiles in gastric cancer cells with or without IRX1 gene transfection and identified a set of IRX1 target genes. Using real-time PCR analysis, we confirmed 8 upregulated genes and 11 downregulated genes. Subsequently, we identified three direct IRX1 target genes by chromatin immunoprecipitation assay: BDKRB2, FGF7 and HIST2H2BE. HIST2H2BE encodes H2B histone, which serves as a cell proliferation marker and has been found overexpressed in serum and gastric cancer tissues by our group and others (Kamei et al., 1992; Hao et al., 2008). BDKRB2 was reported as an enhancer of angiogenesis in cancer (Plendl et al., 2000; Ishihara et al., 2001, 2002; Ikeda et al., 2004). FGF7 encoded a member of the fibroblast growth factor (FGF) family. FGF family members possess broad mitogenic and cell survival activities, and are involved in a variety of biological processes, including tumor growth and invasion. FGF7 is a potent epithelial cell-specific growth factor, and is involved in tumorigenesis through stimulating proliferation and invasion (Ropiquet et al., 1999; Shaoul et al., 2006). These findings raised the possibility that IRX1 overexpression downregulated expression of BDKRB2, FGF7 and HIST2H2BE, consequently inhibiting angiogenesis, cell proliferation and invasion in gastric cancer. Of course, other indirect target genes, such as the transgelin-encoding gene TAGLN, have been reported to have a growth-promoting effect on gastric carcinogenesis (Huang et al., 2008).

In clinical practice, gastric cancer patients are often not symptomatic in the early stage. Endoscope biopsy is still the golden standard for gastric cancer diagnosis. However, endoscope examination is uncomfortable and invasive. Serum specimen screening is a promising diagnostic tool because it is easy and inexpensive. Some serum tumor markers have been used in gastrointestinal cancer screening and monitoring. Most of them are glycoprotein antigens such as carcinoembryonic antigen (CEA), carbohydrate antigen 19–9 (CA19-9) and carbohydrate antigen 72–4 (CA72-4). However, the early diagnostic role of CEA, CA19-9 and CA72-4 for gastric cancer remains controversial. The sensitivity of a single biomarker in tumor diagnosis is low (usually less than 40%) with a higher false-positive rate. Some authors proposed that these biomarkers can be useful for therapeutic response monitoring and cancer relapse detection but not for reliable diagnostics (Tocchi et al., 1998; Ychou et al., 2000; Lai et al., 2002). Identifying novel serological biomarkers with higher specificity and sensitivity is eagerly desired for gastric cancer. Recent data showed that the presence of tumor-associated alterations (such as promoter hypermethylation) have been described in the plasma DNA from patients with various kinds of cancer (Kawakami et al., 2000; Duffy et al., 2009). DNA methylation may provide a new generation of cancer biomarkers in the next decade (Duffy et al., 2009; Tommasi et al., 2009). However, no studies have looked into the relationship between the IRX1 molecular marker and gastric cancer. We are the first to investigate the correlation between methylated IRX1 and gastric cancer using peripheral blood samples. The methylation levels in gastric cancer patients were significantly increased compared with controls (P=0.002). In addition, we found that the IRX1 methylation level in blood increased with advanced TNM staging and aging. This finding implies that circulating methylated IRX1 is a promising molecular marker for gastric cancer. However, as a methylated marker its specificity and sensitivity for early diagnosis of gastric cancer require further verification.

In conclusion, we outlined the functions of the IRX1 tumor suppressor gene during gastric carcinogenesis in Figure 8. IRX1 expression is markedly downregulated in gastric cancer cells. The expression level of IRX1 in gastric cancer correlates with promoter methylation in addition to gene copy number deletion. Restoring IRX1 expression impairs cell proliferation, migration, invasion and tumorigenesis both in vitro and in vivo. The mechanism of the tumor suppressor effect of IRX1 overexpression could be explained by a set of target genes identified by global microarray analysis and confirmed by real-time PCR and chromatin immunoprecipitation assay. Downregulation of the direct target genes BDKRB2, FGF7 and HIST2H2BE contributed to decreased angiogenesis, cell proliferation and invasion. Hypermethylation of IRX1 detected in the peripheral blood of gastric cancer patients suggests a biomarker potential for IRX1.

Figure 8

Schematic summary of the IRX1 tumor suppressor gene on carcinogenesis of gastric cancer. IRX1 is inactivated via a combination of LOH and hypermethylation, which may be induced by radiation stress or an inflammatory reaction to pathogens. Restoring IRX1 expression impairs cell proliferation, migration, invasion and tumorigenesis both in vitro and in vivo, which could be explained by the downregulation of specific target genes identified by global microarray analysis.

Materials and methods

Cell lines

Gastric cancer cell lines NCI-N87, SNU-1, SNU-16, AGS and KATO-III were obtained from American Type Culture Collection (Manassas, VA, USA). Gastric cancer cell lines MKN-45 and SGC-7901, immortalized human gastric mucosa cell line GES-1 and embryonic kidney cell line 293T were preserved in our institute.


Gastric cancer tissues and peripheral blood were collected from patients and healthy controls at Shanghai Ruijin Hospital, Shanghai Jiao Tong University, School of Medicine, after obtaining the subjects’ informed consent and with institutional review board approval of the hospital. All patients obtained a confirmed diagnosis of gastric carcinoma after resection.

Mutation, copy number and promoter analysis

Primers covering exons and splicing regions were designed. A total of seven fragments overlapping four exons were amplified. Mutations were analyzed by sequencing the PCR products (Sangong, Shanghai, China). Gene copy numbers of IRX1 for all cell lines were analyzed by real-time PCR. Three pairs of primers targeting different regions were used (P1 for exon one and intron one junction, P2 for exon two and intron two junction and P3 for intron two and exon three junction). We choose the single-copy gene RPP30 (ribonuclease P/MRP 30 kDa subunit) as internal control. The reactions were run with the following conditions: one cycle at 50 °C for 2 min, 95 °C for 10 min and then 40 cycles (95 °C for 15 s and 56 °C for 30 s). Target gene copy number was determined by the relative quantitative comparative threshold cycle (ΔΔCt) method. In this method, the relative copy-number difference of the target gene in tumor sample against reference sample is given by 2 - ΔΔ C t described by Kuga et al. (2008).

Promoter fragments were amplified from human genomic DNA by two-step nested PCR. The fragments of the 5′-flanking region were inserted into the MluI and BglII sites of the promoterless firefly luciferase reporter vector pGL3-Basic to construct luciferase reporter pGL3-775 bp. By progressive 5′deletion, we obtained PCR products for seven additional plasmids: pGL3-747, pGL3-644, pGL3-476, pGL3-410, pGL3-380, pGL3-248 and pGL3-152. The primers used are listed in Supplementary Table 4. One hundred nanograms of each human IRX1-luciferase reporter gene construct was transfected into cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). The pRL-TK vector (5 ng) (Promega, Madison, WI, USA) harboring the Renilla luciferase gene was co-transfected as an internal control for transfection efficiency in mammalian cells. Each construct was transfected in triplicate. Luciferase activity was determined with an analytical luminometer (TD-20/20, Turner Designs, Sunnyvale, CA, USA).

Methylation analysis and 5-Aza-2′-deoxycytidine treatment

Genomic DNA from cell lines, gastric cancer tissues and plasma (0.2 ml) was purified using DNAzol (Invitrogen), treated with sodium bisulfate (Sigma, Phoenix, AZ, USA), and then analyzed by MSP or bisulfate-sequencing PCR (BSP). The PCR products were confirmed by agarose gel electrophoresis and visualized using ethidium bromide staining for MSP. A negative control (without DNA) was included in every set of PCR experiments. Amplified bisulfate PCR products were subcloned into the TA vector system (Promega) according to the manufacturer’s instructions. DNA sequencing was performed on three to ten individual clones (Sangong). Cell lines were treated with 10 μM of 5-Aza-2′-deoxycytidine (5-Aza-dC; Sigma) for 96 h or 120 h; the control group was treated with 1 × PBS. Total RNA was extracted for assessment of IRX1 mRNA levels. The primers used are summarized in Supplementary Table 4.

mRNA expression analysis

Total RNA was extracted using Trizol solution. Reverse transcription (RT) was performed in a 20-μl reaction system according to the manufacturer's recommendation. RT–PCR primers for IRX1 (260 bp) were 5′-IndexTermGCTCTTCGGCAGCGACAC-3′ (F) and 5′-IndexTermGCTCTGGGGCCTCCTTTG-3′ (R). Primers for β-actin (428 bp) were 5′-IndexTermCCACTGGCATCGTGATGGAC-3′ (F) and 5′-IndexTermGCGGATGTCCACGTCACACT-3′ (R). Each PCR was performed for 30 thermal cycles and then the PCR products were observed by electrophoresis on a 1.5% agarose gel and visualized after staining with ethidium bromide. QRT–PCR was carried out with the IRX1 primers 5′-IndexTermCGCGGATCTCAGCCTCTTC-3′ (F) and 5′-IndexTermCCCCAGGGTTGTCCTTCAGT-3′ (R). Primers for GAPDH were 5′-IndexTermGGACCTGACCTGCCGTCTAG-3′ (F) and 5′-IndexTermGTAGCCCAGGATGCCCTTGA-3′(R). Quantitative measurement of IRX1 mRNA levels was performed by ABI Prism 7000 (Applied Biosystems, Foster city, CA, USA). These data were analyzed by the comparative Ct method.

IRX1 overexpression

IRX1 overexpression vector pEGFP-N1–IRX1 was constructed using primers 5′-IndexTermATACTCGAGTCCTTCCGGCAGCTGGGC-3′ (F) and 5′-IndexTermATAGAATTCCGGACGGGAGGGCTGCTA-3′ (R). SGC-7901 and NCI-N87 gastric cancer cell lines were used for the overexpression studies. We obtained stably transfected clones by G418 selection (Promega). A stable transfectant of the pEGFP-N1 empty vector was used as a control. Transfection was verified by fluorescence microscopy (Olympus, Japan).

Western blot analysis

Protein from treated cell lines was extracted by mammalian protein extraction reagent (Pierce, Appleton, WI, USA) supplemented with protease inhibitors cocktail (Sigma). Fifty micrograms protein samples were resolved by 10% SDS–PAGE and then transferred to PVDF membranes. The membranes were blocked with TBST buffer (TBS plus 0.1% Tween-20) containing 5% w/v non-fat milk and hybridized with primary antibody, followed by incubation with specific HRP-conjugated secondary antibody. Protein bands were visualized by the ECL detection system (Amersham Biosciences, Uppsala, Sweden). Autoradiograms were quantified by densitometry (Quantity One software; Bio-Rad, Hercules, CA, USA). GAPDH-specific antibody was used for a loading control. Relative protein levels were calculated by reference to the amount of GAPDH protein. Mouse anti-GFP (1:800, Sigma) and mouse anti-GAPDH (1:1000, Santa Cruz, CA, USA) were used as the primary antibodies.

Immunocytochemical analysis

SGC-7901 cells were cultured in RPMI-1640 medium supplemented with 10% fetal calf serum on cover slides and maintained at 37 °C in an atmosphere containing 5% CO2 and 95% air for 24 h. Then the slides were fixed in 4% paraformaldehyde for 30 min and treated with 1% Triton–PBS solution for 15 min at room temperature. Next, protein expression of IRX1 in the SGC-7901 cell line with or without gene transfection was examined using polyclonal anti-human IRX1 (Abnova, H00079192-A01, Taiwan, working dilution: 1:600) by the EnVision two-step procedure.

Cell growth and colony formation assay

Cell growth was assayed using Cell Counting Kit-8 (Dojindo, Kumamoto, Japan). Soft agar colony formation was performed using the cell transformation detection kit (Chemicon, Temecula, CA, USA) according to the manufacturers instructions. After 21 days of culture, the colonies were stained and counted under an inverted phase-contrast microscope. A dot with cells >50 was counted as a colony. These data were reported as the means±s.d. by counting 10 fields randomly.

Invasion, migration and scratch healing assays

Cell invasion and cell migration assays were performed. For the scratch assay, cells were treated with 10 μg/ml mitomycin C (Sigma) for 3 h and then wounded with a pipette tip. Fresh, full medium was added, and the wound closing procedure was observed for 48 h. Photographs were taken every 6 h. For migration and invasion assays, cell culture was performed in Transwell chambers (8 μm, 24-well format; Corning, NY, USA). For the invasion assay, the insert membranes were coated with diluted Matrigel (BD Biosciences, San Jose, CA, USA). Cells (1 × 105) were added to the upper chamber and cultured for 48 h. For the migration assay, the insert membranes were not coated with Matrigel but were cultured under the same conditions. Finally, the insert membranes were cut and stained with Crystal violet (0.04% in water; 100 μl) and the permeating cells were counted under the inverted microscope and photographed. At least three independent experiments were performed for all conditions. The data are shown as the means±s.d.

Tumor growth in nude mice

Cells (100 μl, 1 × 106 cells) from parental cell lines SGC-7901 or NCI-N87, or from transfected lines cancer cells/vector or cancer cells/IRX1 were collected and inoculated subcutaneously into the right flank regions of 4-week-old male BALB/c nude mice (Institute of Zoology, Chinese Academy of Sciences, Shanghai, China). Experimental and control groups had five mice each. Tumor nodules were measured every 4 days with calipers. Mice were killed after 1 month. Tumor growth curves and inhibiting rates were calculated. After tumor excision, the tissues were fixed in 10% buffered formalin. All formalin-fixed and paraffin-embedded samples were carefully examined after staining with hematoxylin–eosin (HE) and photographed.

Global cDNA microarray analysis and target gene verification

The whole human genome oligo microarray (Agilent, Santa Clara, CA, USA) was used. After hybridization and washing, the microarray slides were scanned with an Agilent DNA microarray scanner. The resulting text files extracted from Agilent Feature Extraction Software (version 9.5.3) were imported into the Agilent GeneSpring GX software (version 7.3) for further analysis. Differentially expressed genes were identified through fold-change screening. For target gene verification we used real-time PCR and chromatin immunoprecipitation assay. The primers for the target genes are listed in Supplementary Table 4.


Data from cell growth, luciferase assays and mice tumorigenesis experiments were analyzed with one-way ANOVA. The data from plasma methylation detection of IRX1 were analyzed by independent-sample T-test. Receiver Operating CharacteristicsROC graph was performed. Quantitative values were expressed as the means±s.d. SPSS 15.0 software (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses, and a P-value <0.05 was considered significant. Gene expression profiling of cell lines was analyzed by fold-change. The threshold used to screen up- or downregulated genes was fold-change 2.

Accession number

Microarray data reported herein have been deposited at the NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE17399) with the accession number GSE17399.

Accession codes




  1. Alarcon P, Rodriguez-Seguel E, Fernandez-Gonzalez A, Rubio R, Gomez-Skarmeta JL . (2008). A dual requirement for Iroquois genes during Xenopus kidney development. Development 135: 3197–3207.

  2. Asaka S, Fujimoto T, Akaishi J, Ogawa K, Onda M . (2006). Genetic prognostic index influences patient outcome for node-positive breast cancer. Surg Today 36: 793–801.

  3. Becker MB, Zulch A, Bosse A, Gruss P . (2001). Irx1 and Irx2 expression in early lung development. Mech Dev 106: 155–158.

  4. Bennett KL, Karpenko M, Lin MT, Claus R, Arab K, Dyckhoff G et al. (2008). Frequently methylated tumor suppressor genes in head and neck squamous cell carcinoma. Cancer Res 68: 4494–4499.

  5. Bergman Y, Mostoslavsky R . (1998). DNA demethylation: turning genes on. Biol Chem 379: 401–407.

  6. Cairns P . (2009). 5′-azacytidine expression arrays. Methods Mol Biol 507: 165–174.

  7. Choi IS, Wu TT . (2005). Epigenetic alterations in gastric carcinogenesis. Cell Res 15: 247–254.

  8. Duffy MJ, Napieralski R, Martens JW, Span PN, Spyratos F, Sweep FC et al. (2009). Methylated genes as new cancer biomarkers. Eur J Cancer 45: 335–346.

  9. Duggan DJ, Bittner M, Chen Y, Meltzer P, Trent JM . (1999). Expression profiling using cDNA microarrays. Nat Genet 21: 10–14.

  10. Esteller M . (2007). Epigenetic gene silencing in cancer: the DNA hypermethylome. Hum Mol Genet 16 (Spec No 1): R50–R59.

  11. Ferguson CA, Tucker AS, Christensen L, Lau AL, Matzuk MM, Sharpe PT . (1998). Activin is an essential early mesenchymal signal in tooth development that is required for patterning of the murine dentition. Genes Dev 12: 2636–2649.

  12. Garinis GA, Patrinos GP, Spanakis NE, Menounos PG . (2002). DNA hypermethylation: when tumour suppressor genes go silent. Hum Genet 111: 115–127.

  13. Hao Y, Yu Y, Wang L, Yan M, Ji J, Qu Y et al. (2008). IPO-38 is identified as a novel serum biomarker of gastric cancer based on clinical proteomics technology. J Proteome Res 7: 3668–3677.

  14. Hellebrekers DM, Melotte V, Vire E, Langenkamp E, Molema G, Fuks F et al. (2007). Identification of epigenetically silenced genes in tumor endothelial cells. Cancer Res 67: 4138–4148.

  15. Huang Q, Chen W, Wang L, Lin W, Lin J, Lin X . (2008). Identification of transgelin as a potential novel biomarker for gastric adenocarcinoma based on proteomics technology. J Cancer Res Clin Oncol 134: 1219–1227.

  16. Hurtubise A, Bernstein ML, Momparler RL . (2008). Preclinical evaluation of the antineoplastic action of 5-aza-2′-deoxycytidine and different histone deacetylase inhibitors on human Ewing’s sarcoma cells. Cancer Cell Int 8: 16.

  17. Ikeda Y, Hayashi I, Kamoshita E, Yamazaki A, Endo H, Ishihara K et al. (2004). Host stromal bradykinin B2 receptor signaling facilitates tumor-associated angiogenesis and tumor growth. Cancer Res 64: 5178–5185.

  18. Iorns E, Lord CJ, Grigoriadis A, McDonald S, Fenwick K, Mackay A et al. (2009). Integrated functional, gene expression and genomic analysis for the identification of cancer targets. PLoS One 4: e5120.

  19. Ishihara K, Hayash I, Yamashina S, Majima M . (2001). A potential role of bradykinin in angiogenesis and growth of S-180 mouse tumors. Jpn J Pharmacol 87: 318–326.

  20. Ishihara K, Kamata M, Hayashi I, Yamashina S, Majima M . (2002). Roles of bradykinin in vascular permeability and angiogenesis in solid tumor. Int Immunopharmacol 2: 499–509.

  21. Jee CD, Kim MA, Jung EJ, Kim J, Kim WH . (2009). Identification of genes epigenetically silenced by CpG methylation in human gastric carcinoma. Eur J Cancer 45: 1282–1293.

  22. Jin G, Xu L, Shu Y, Tian T, Liang J, Xu Y et al. (2009). Common genetic variants on 5p15.33 contribute to risk of lung adenocarcinoma in a Chinese population. Carcinogenesis 30: 987–990.

  23. Kamei M, Kato M, Mochizuki K, Kuroda K, Sato S, Hashizume S et al. (1992). Serodiagnosis of cancers by ELISA of anti-histone H2B antibody. Biotherapy 4: 17–22.

  24. Kawakami K, Brabender J, Lord RV, Groshen S, Greenwald BD, Krasna MJ et al. (2000). Hypermethylated APC DNA in plasma and prognosis of patients with esophageal adenocarcinoma. J Natl Cancer Inst 92: 1805–1811.

  25. Knudson AG . (1996). Hereditary cancer: two hits revisited. J Cancer Res Clin Oncol 122: 135–140.

  26. Knudson AG . (2001). Two genetic hits (more or less) to cancer. Nat Rev Cancer 1: 157–162.

  27. Kuga D, Mizoguchi M, Guan Y, Hata N, Yoshimoto K, Shono T et al. (2008). Prevalence of copy-number neutral LOH in glioblastomas revealed by genomewide analysis of laser-microdissected tissues. Neuro Oncol 10: 995–1003.

  28. Lai IR, Lee WJ, Huang MT, Lin HH . (2002). Comparison of serum CA72-4, CEA, TPA, CA19-9 and CA125 levels in gastric cancer patients and correlation with recurrence. Hepatogastroenterology 49: 1157–1160.

  29. Lemaire M, Chabot GG, Raynal NJ, Momparler LF, Hurtubise A, Bernstein ML et al. (2008). Importance of dose-schedule of 5-aza-2′-deoxycytidine for epigenetic therapy of cancer. BMC Cancer 8: 128.

  30. Lewis MT, Ross S, Strickland PA, Snyder CJ, Daniel CW . (1999). Regulated expression patterns of IRX-2, an Iroquois-class homeobox gene, in the human breast. Cell Tissue Res 296: 549–554.

  31. Lu Y, Yu Y, Zhu Z, Xu H, Ji J, Bu L et al. (2005). Identification of a new target region by loss of heterozygosity at 5p15.33 in sporadic gastric carcinomas: genotype and phenotype related. Cancer Lett 224: 329–337.

  32. McKay JD, Hung RJ, Gaborieau V, Boffetta P, Chabrier A, Byrnes G et al. (2008). Lung cancer susceptibility locus at 5p15.33. Nat Genet 40: 1404–1406.

  33. Momparler RL . (2003). Cancer epigenetics. Oncogene 22: 6479–6483.

  34. Perez-Diez A, Morgun A, Shulzhenko N . (2007). Microarrays for cancer diagnosis and classification. Adv Exp Med Biol 593: 74–85.

  35. Plendl J, Snyman C, Naidoo S, Sawant S, Mahabeer R, Bhoola KD . (2000). Expression of tissue kallikrein and kinin receptors in angiogenic microvascular endothelial cells. Biol Chem 381: 1103–1115.

  36. Quackenbush J . (2006). Microarray analysis and tumor classification. N Engl J Med 354: 2463–2472.

  37. Robertson KD, Jones PA . (2000). DNA methylation: past, present and future directions. Carcinogenesis 21: 461–467.

  38. Ropiquet F, Huguenin S, Villette JM, Ronfle V, Le Brun G, Maitland NJ et al. (1999). FGF7/KGF triggers cell transformation and invasion on immortalised human prostatic epithelial PNT1A cells. Int J Cancer 82: 237–243.

  39. Schneider-Stock R, Ocker M . (2007). Epigenetic therapy in cancer: molecular background and clinical development of histone deacetylase and DNA methyltransferase inhibitors. IDrugs 10: 557–561.

  40. Shaker S, Bernstein M, Momparler RL . (2004). Antineoplastic action of 5-aza-2′-deoxycytidine. (Dacogen) and depsipeptide on Raji lymphoma cells. Oncol Rep 11: 1253–1256.

  41. Shaoul R, Eliahu L, Sher I, Hamlet Y, Miselevich I, Goldshmidt O et al. (2006). Elevated expression of FGF7 protein is common in human gastric diseases. Biochem Biophys Res Commun 350: 825–833.

  42. Tamura M, Gu J, Matsumoto K, Aota S, Parsons R, Yamada KM . (1998). Inhibition of cell migration, spreading, and focal adhesions by tumor suppressor PTEN. Science 280: 1614–1617.

  43. Tamura M, Gu J, Takino T, Yamada KM . (1999). Tumor suppressor PTEN inhibition of cell invasion, migration, and growth: differential involvement of focal adhesion kinase and p130Cas. Cancer Res 59: 442–449.

  44. Telford JL, Covacci A, Ghiara P, Montecucco C, Rappuoli R . (1994). Unravelling the pathogenic role of Helicobacter pylori in peptic ulcer: potential new therapies and vaccines. Trends Biotechnol 12: 420–426.

  45. Tocchi A, Costa G, Lepre L, Liotta G, Mazzoni G, Cianetti A et al. (1998). The role of serum and gastric juice levels of carcinoembryonic antigen, CA19.9 and CA72.4 in patients with gastric cancer. J Cancer Res Clin Oncol 124: 450–455.

  46. Tommasi S, Karm DL, Wu X, Yen Y, Pfeifer GP . (2009). Methylation of homeobox genes is a frequent and early epigenetic event in breast cancer. Breast Cancer Res 11: R14.

  47. Toyota M, Issa JP . (2005). Epigenetic changes in solid and hematopoietic tumors. Semin Oncol 32: 521–530.

  48. Tycko B . (2000). Epigenetic gene silencing in cancer. J Clin Invest 105: 401–407.

  49. Ushijima T, Nakajima T, Maekita T . (2006). DNA methylation as a marker for the past and future. J Gastroenterol 41: 401–407.

  50. van Tuyl M, Liu J, Groenman F, Ridsdale R, Han RN, Venkatesh V et al. (2006). Iroquois genes influence proximo-distal morphogenesis during rat lung development. Am J Physiol Lung Cell Mol Physiol 290: L777–L789.

  51. Yamashita K, Upadhyay S, Osada M, Hoque MO, Xiao Y, Mori M et al. (2002). Pharmacologic unmasking of epigenetically silenced tumor suppressor genes in esophageal squamous cell carcinoma. Cancer Cell 2: 485–495.

  52. Ychou M, Duffour J, Karma A, Gourgou S, Grenier J . (2000). Clinical significance and prognostic value of CA72-4 compared with CEA and CA19-9 in patients with gastric cancer. Dis Markers 16: 105–110.

  53. Yu YY, Ji J, Lu Y, Bu L, Liu BY, Zhu ZG et al. (2006). [High-resolution analysis of chromosome 5 and identification of candidate genes in gastric cancer]. Zhonghua Zhong Liu Za Zhi 28: 84–87.

  54. Zhao H, Kim Y, Wang P, Lapointe J, Tibshirani R, Pollack JR et al. (2005). Genome-wide characterization of gene expression variations and DNA copy number changes in prostate cancer cell lines. Prostate 63: 187–197.

Download references


This work was supported, in part by grants from the National Natural Science Foundation of China (30572127, 30770961 and 30973486), the Chinese National High Tech Program (863-2006AA02A402, 2006AA02A301), the Shanghai Pu Jiang Project (PJ200700367), Key Research Project from Shanghai Science and Technology Commission (09DZ1950101 and 09JC1409600), and Shanghai Charity Foundation for Cancer Research.

Author information

Correspondence to Z Zhu or Y Yu.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article


  • epigenetic
  • gastric carcinoma
  • homeobox genes
  • IRX1
  • tumor suppressor

Further reading