Transgenic Cdx2 induces endogenous Cdx1 in intestinal metaplasia of Cdx2-transgenic mouse stomach

Hiroyuki Mutoh, Hiroko Hayakawa, Hirotsugu Sakamoto, Miho Sashikawa and Kentaro Sugano
Department of Medicine, Division of Gastroenterology, Jichi Medical University, Tochigi, Japan

Keywords chromatin immunoprecipitation; luciferase reporter assay; methylation; RT-PCR; siRNA Correspondence H. Mutoh, Department of Medicine, Division of Gastroenterology, Jichi Medical University, Yakushiji 3311-1, Shimotsuke, Tochigi 329-0498, Japan Fax: +81 285 44 8297 Tel: +81 285 58 7348 E-mail: muto@jichi.ac.jp (Received 12 April 2009, revised 1 August 2009, accepted 6 August 2009) doi:10.1111/j.1742-4658.2009.07263.x

Cdx1 and Cdx2, which are transcription factors regulating normal intestinal development, have been studied as potential key molecules in the pathogenesis of the precancerous intestinal metaplasia of the human stomach. However, the regulation of Cdx1 expression in the intestinal metaplasia is poorly understood. Cdx2-expressing gastric mucosa of Cdx2transgenic mouse stomach was replaced by intestinal metaplastic mucosa. The aim of this study was to investigate the following: (a) Cdx1 expression in the intestinal metaplastic mucosa of the Cdx2-transgenic mouse stomach; and (b) the relationship between Cdx1 and Cdx2. A mouse model of intestinal metaplasia, the Cdx2-transgenic mouse, was used to investigate Cdx1 gene expression by RT-PCR. DNA methylation prole analysis was performed by bisulte sequencing, and the interaction of Cdx2 with the Cdx1 promoter was examined by chromatin immunoprecipitation assay, electrophoretic mobility shift assay, and luciferase reporter assays. Cdx2 mRNA was expressed in the Cdx2-transgenic mouse stomach. However, endogenous Cdx2 mRNA was not expressed in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. On the other hand, endogenous Cdx1 mRNA and protein were expressed in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. The Cdx1 promoter was unmethylated in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. Chromatin immunoprecipitation assay and electrophoretic mobility shift assay showed that Cdx2 was bound to the Cdx1 promoter region in the intestinal metaplasia and the normal intestine. Cdx2 upregulated and siRNA-Cdx2 downregulated the transcriptional activity of the Cdx1 gene in the human gastric carcinoma cell lines AGS, MKN45, and MKN74. In conclusion, transgenic Cdx2 induced endogenous Cdx1 through the binding of Cdx2 to the unmethylated Cdx1 promoter region in the intestinal metaplasia of the Cdx2-transgenic mouse stomach.

Introduction
In intestinal metaplasia of the human stomach, normal gastric mucosa is replaced by an intestinalized epithelium, and is mainly induced together with the progression of Helicobacter pylori-infected chronic gastritis. Intestinal metaplasia of the human stomach has been extensively studied as a premalignant condition of gastric carcinoma [1]. The intestine-specic homeobox genes Cdx1 and Cdx2 have been shown to be

Abbreviations ChIP, chromatin immunoprecipitation; EMSA, electrophoretic mobility shift assay; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RA, retinoic acid; si, small interfering.

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

5821

Cdx1 expression in intestinal metaplasia

H. Mutoh et al.

aberrantly expressed in human intestinal metaplasia. Cdx1 and Cdx2 are mammalian members of the caudal-related homeobox gene family. In adult mice and humans, expression is strictly conned to the gut, from the duodenum to the rectum. Normal stomach does not express the transcription factors Cdx1 and Cdx2. We and others have reported the presence of Cdx1 and Cdx2 in the intestinal metaplasia of the H. pyloriinfected human stomach [24]. We have previously generated Cdx2-transgenic mice as model mice for intestinal metaplasia [5,6]. Cdx2transgenic mice specically express Cdx2 in the gastric mucosa, and develop intestinal metaplasia in the stomach [5,6]. Gastric carcinoma spontaneously developed from intestinal metaplasia in all stomachs of Cdx2transgenic mice examined [7]. In Barretts esophagus, normal squamous esophageal mucosa is also replaced by an intestinalized columnar epithelium in which Cdx2 is expressed [8]. Exposure to acid and or bile acids has been reported to activate Cdx2 expression in human esophageal epithelial cells through promoter demethylation [911]. However, it is still unclear how Cdx1 is induced in intestinal metaplasia. Furthermore, the relationship between Cdx1 and Cdx2 in intestinal metaplasia has not been claried as yet. To investigate these questions, we focused on the induction of endogenous Cdx1 in Cdx2-induced intestinal metaplasia using Cdx2-transgenic mice.

Endogenous Cdx2 was expressed in the normal intestine, but in none of the intestinal metaplasia of the Cx2-transgenic mouse stomachs (Fig. 1C). Transgenic Cdx2 did not induce endogenous Cdx2 expression, indicating that Cdx2 is not autoregulated in intestinal metaplasia. Next, whether endogenous Cdx1 was expressed in Cdx2-induced intestinal metaplasia was investigated. Endogenous Cdx1 was detected in the normal intestine and in all of the Cdx2-induced intestinal metaplasia, but not in the normal stomach (Fig. 2A). Cdx1 gene expression was characterized by quantitative real-time RT-PCR (Fig. 2B). The Cdx1 mRNA level in the Cdx2-transgenic mouse stomach was almost same as that in the normal mouse small intestine (Fig. 2B). Cdx1 expression in the intestinal metaplasia of the Cdx2-transgenic mouse stomach was also investigated, using immunohistochemistry. Cdx1 was expressed in the intestinal metaplasia of the Cdx2-transgenic mouse stomach (Fig. 2E) and normal intestine (Fig. 2D), but not in the normal stomach (Fig. 2C). The expression of Cdx1 mRNA and protein in the intestinal metaplasia of the Cdx2-transgenic mouse stomach indicates that Cdx1 might be induced by Cdx2 in intestinal metaplasia. Cdx1 promoter methylation status We focused on epigenetic regulation of Cdx1 gene expression as a possible cause of Cdx1 activation in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. To investigate whether the differences in Cdx1 expression were under promoter methylation control, bisulte sequencing was performed on DNA extracted from ve normal stomachs, ve normal intestines, and ve intestinal metaplasias of Cdx2-transgenic mouse stomachs. All of the CpGs including the CpGs (located around the TATA box and indicated by the box in Fig. 3A) that appear to be critical for the control of Cdx1 expression in colorectal carcinoma [12] were unmethylated in the Cdx1 promoter sequences from the ve intestinal metaplasias, the ve normal intestines and the ve normal stomachs (Fig. 3A). These results made it clear that Cdx1 promoter methylation status does not determine the expression of Cdx1 in the normal intestine and in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. Next, the methylation status of the Cdx2 promoter region was examined. All of the CpGs (shown in red in Fig. 3B) in the Cdx2 promoter sequences from ve intestinal metaplasias, ve normal intestines and ve normal stomachs were unmethylated, except for one

Results
Expression of Cdx1 and Cdx2 in the intestinal metaplasia of the Cdx2-transgenic mouse stomach Cdx2-transgenic mice we generated showed intestinal metaplasia in the stomach [5,6]. First, Cdx2 expression in the intestinal metaplasia of Cdx2-transgenic mouse stomachs was examined, using RT-PCR. Cdx2 mRNA was detected in normal intestine and in all of the intestinal metaplasia of the Cdx2-transgenic mouse stomach, but not in the normal mouse stomach (Fig. 1B). Cdx2 expression was detected using a primer pair for the Cdx2 coding region (Cdx2 coding-fw and Cdx2 coding-rv; Fig. 1A and Table 1). When Cdx2-transgenic mice were generated, only the Cdx2 coding region, without the noncoding region, was used. To investigate whether endogenous Cdx2 was expressed in Cdx2-induced intestinal metaplasia, endogenous Cdx2 expression was detected using a primer pair for the coding region and the 3-noncoding region (Cdx2 coding-fw and Cdx2 non-coding-rv; Fig. 1A and Table 1).
5822

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

H. Mutoh et al.

Cdx1 expression in intestinal metaplasia

A
Intron

Terminal codon

Stomach

B
Marker

Intestine

Cdx2 stomach
3 4 5 6 7 8

Fig. 1. RT-PCR analysis of Cdx2 expression. (A) Scheme of a part of the mouse Cdx2 mRNA, including the stop codon tga, which is shown in red. The primers used for detecting Cdx2 transcript are indicated by underlining and yellow shading. The exon 2exon 3 boundary site is indicated by an arrow. (B) RT-PCR analysis of Cdx2 mRNA transcripts (primer pair; Cdx2 coding-fw and Cdx2 coding-rv) in normal mouse stomach (lane 1), normal mouse small intestine (lane 2), and Cdx2-transgenic mouse stomach (lanes 38). (C) RT-PCR analyses of endogenous Cdx2 mRNA transcripts (primer pair; Cdx2 coding-fw and Cdx2 non-codingrv) in normal mouse stomach (lane 1), normal mouse small intestine (lane 2), and Cdx2-transgenic mouse stomach (lanes 37). The lower panels in (B) and (C) show standard RT-PCR conducted with primers designed to detect b-actin mRNA.

-actin
Stomach

Intestine

Cdx2 stomach
3 4 5 6 7

Marker

-actin

CpG, indicated by the box in Fig. 3B, that was methylated in ve normal intestines and ve intestinal metaplasias. These results indicate that Cdx2 promoter methylation status does not determine the expression of endogenous Cdx2 in the normal intestine and in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. Cdx2 binds directly to the Cdx1 promoter region in vivo The putative TATA-box (TATAAA) sequence at positions )51 to )46 (relating to the transcription start site; GenBank number NM_009880) exhibits obvious sequence similarity with the consensus Cdx-binding site

(C TATAAAG T) (Fig. 4A), whereas no additional putative Cdx-binding site could be found elsewhere in the Cdx1 promoter (at position )2000 from the transcription start site). To examine whether the expression of Cdx1 mRNA in the intestinal metaplasia of the Cdx2-transgenic mouse stomach is associated with the binding of Cdx2 to this TATAAA region, we performed chromatin immunoprecipitation (ChIP) assays, using an antibody against Cdx2. We cross-linked the protein and DNA in the intestinal metaplasia of Cdx2transgenic mouse stomach as well as in the stomach and intestine of normal mice. The Cdx1 promoter region encompassing the TATAAA sequence at )51 to )46 was amplied by PCR with two sets of primers (Fig. 4B, Cdx1 promoter fw1 and Cdx1 promoter rv1;
5823

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

Cdx1 expression in intestinal metaplasia

H. Mutoh et al.

Table 1. The sequences of oligonucleotide primers used in this study. Primers Sequence (5- to 3)

Cdx2 binds to the TATAAA sequence We investigated Cdx2 binding to the TATAAA sequence, using the nuclear fractions extracted from Cdx2-expressing AGS cells (Fig. 4D). We found that nuclear extracts from AGS cells formed the Cdx2 DNA complex (Fig. 4D). The presence of Cdx2 in DNAprotein complexes was eliminated by using monoclonal antibody specic to Cdx2 (Fig. 4D, lane 3). With the use of a mutant probe, DNAprotein complexes were not formed (Fig. 4D, lane 1). These results indicate that Cdx2 binds to the TATAAA sequence. The Cdx1 promoter was activated in Cdx2-expressing human gastric carcinoma AGS, MKN45 and MKN74 cells The expression of Cdx1 in the intestinal metaplasia of the Cdx2-transgenic mouse stomach supports the hypothesis that Cdx2 could regulate Cdx1 transcription. Supporting this, the ChIP assay indicated that Cdx2 is bound to the region between )191 and +112 (relating to the transcription start site). The region between )191 and +112 contains the Cdx consensus sequence TATAAA ()51 to )46) (Fig. 4A). Furthermore, electrophoretic mobility shift assay (EMSA) indicated that Cdx2 binds to the TATAAA sequence. We examined the Cdx1 transcriptional activity in Cdx2-expressing AGS, MKN45 and MKN74 cells (Fig. 5A), using pGL4.10[luc2]Cdx1 deletion and mutation constructs. These cell lines (AGS, MKN45, and MKN74) also expressed Cdx1, which was detected by RT-PCR (Fig. 5A). The Cdx1 promoter reporter gene containing the region between )365 and +12 was activated, whereas the Cdx1 promoter reporter gene containing the region between )365 and )78 was not activated, in Cdx2-expressing AGS, MKN45 and MKN74 cells (Fig. 5B). This result suggests that the element between )77 and +12 in the Cdx1 promoter may be critical for Cdx1 gene transcriptional activity in Cdx2-expressing AGS, MKN45 and MKN74 cells. The sequence between )77 and +12 contains a potential Cdx2-binding site (TATAAA, )51 and )46). Analysis of a reporter construct with mutation of the Cdx2 consensus-binding element at )51 and )46 revealed that the element was critical for transcriptional activity of the Cdx1 reporter gene construct in AGS, MKN45 and MKN74 cells (Fig. 5B). Furthermore, we examined the effects of the transfection of the Cdx2 expression plasmid or small interfering RNA targeting Cdx2 (siRNA-Cdx2) on the transcriptional activities of the Cdx1 promoter luciferase

Primers used for mouse Cdx2 detection Cdx2-fw CGGCTGGAGCTGGAGAAGG Cdx2 coding-rv GACAGTGGAGTTTAAAACCC Cdx2 noncoding-rv GCCTGGGATTGCTGTGCCG Primers used for mouse Cdx1 detection Cdx1cDNAfw CCGAACCAAGGACAAGTACC Cdx1cDNArv GTTTACTTTGCGCTCCTTGG Primers used for mouse b-actin detection b-Actin-fw ATCTACGAGGGCTATGCTCT b-Actin-rv TACTCCTGCTTGCTGATCCA Primers used for human Cdx2 detection Cdx2-human-fw AGCCAAGTGAAAACCAGGAC Cdx2-human-rv ATTTCTTGAGGCCCCAAATC Primers used for human Cdx1 detection Cdx1-human-fw TCGGACCAAGGACAAGTACC Cdx1-human-rv TGTTGCTGCTGCTGTTTCTT Primers used for human GAPDH detection GAPDH-fw ACGGATTTGGTCGTATTGGG GAPDH-rv TGATTTTGGAGGGATCTCGC Primers used for Cdx1 methylation CpG-Cdx1-fw1 GAGTTAGTTTTTTTATTTGT [)331 )305] AATTTAG CpG-Cdx1-fw2 TAATTTAGGGGTGGGTGGTG [)312 )293] CpG-Cdx1-rv AAAAAATCCTTATCCAACAC [+114 +89] ATAACC Primers used for Cdx2 methylation CpG-Cdx2-fw1 AGTGTATTTAGGTTGGAAGGAG [)234 )212] CpG-Cdx2-fw2 GTAGTTAGTAAGAAGGGTTTGA [)206 )185] CpG-Cdx2-rv TAACTAACTACACCTCAACCCA [+194 +173] Primers used for ChIP assay Cdx1 promoter-fw1 CTAGGGTCATGCCACCACTC Cdx1 promoter-fw2 ATCCACCTCCCGCTTAGG Cdx1 promoter-rv2 GGAGTCCTTGTCCAGCACAT Primers used for Cdx1 promoter Cdx1 promoter-fw1-XhoI CTCGAGCTAGGGTCATGCCACCACTC [)365 )345] Cdx1 promoter-rv1-HindIII AAGCTTACCAGCGACTGCTCACCT [+12 )7] Cdx1 promoter-rv2-HindIII AAGCTTAAGCTTGGGCGGCTTTGC [)78 )95] ATTTCA Cdx1-Csp45I-fw TTCGAAAGGCCGGGGTGGGGC Cdx1-Csp45I-rv TTCGAAGCCGCGGGCCGTCCGC

Fig. 4C, Cdx1 promoter fw2 and Cdx1 promoter rv1). Binding of Cdx2 to the promoter region of the Cdx1 gene, including the TATAAA sequence, was detected in the intestinal metaplasia of the Cdx2-transgenic mouse stomach and the normal intestine, but not in the normal stomach (Fig. 4B,C).

5824

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

H. Mutoh et al.

Cdx1 expression in intestinal metaplasia

Marker

Normal Normal stomach intestine

1 2 3 4 5

Cdx2 stomach
6 7 8 9

-actin
B
1

0.8

Fig. 2. RT-PCR and immunohistochemical analysis of Cdx1 expression. (A) RT-PCR analysis of Cdx1 expression. RT-PCR analyses of Cdx1 mRNA transcripts in normal mouse stomach (lanes 1 and 2), normal mouse intestine (lanes 3 and 4) and Cdx2transgenic mouse stomach (lanes 59) are shown. The lower panel in (A) shows standard RT-PCR conducted with primers designed to detect b-actin mRNA. (B) Cdx1 gene expression characterized by quantitative real-time RT-PCR. The Cdx1 mRNA level in Cdx2-transgenic mouse stomach was almost the same as that in normal mouse small intestine (B). (CE) Immunohistochemical staining for Cdx1 in the normal stomach (C), the normal intestine (D) and the intestinal metaplasia of the Cdx2-transgenic mouse stomach (E).

0.6

0.4

0.2

Normal Normal Cdx2 stomach intestine stomach Normal stomach

Normal intestine

Cdx2 stomach

construct containing the region between )365 and +12 or the mutant Cdx1 reporter luciferase construct. Cotransfection with the Cdx2 expression plasmid increased the transcriptional activities of the intact Cdx1 reporter gene, but did not affect the transcriptional activities of the mutant Cdx1 reporter gene, in AGS, MKN45 and MKN74 cells (Fig. 5B). Cotransfection with siRNA-Cdx2 decreased the transcriptional activities of the intact Cdx1 reporter gene, but did not affect the transcriptional activities of the mutant Cdx1 reporter gene, in AGS, MKN45 and MKN74 cells (Fig. 5B). Next, after transfection of Cdx2 expression plasmid or siRNA-Cdx2 into AGS, MKN45 and MKN74 cells, Cdx1 mRNA levels were measured using quantitative real-time RT-PCR. As compared with the transfection of a negative control, the transfection of the Cdx2 expression plasmid resulted in an increase in Cdx1 mRNA (Fig. 5C). As compared with the transfection of a negative control, the transfection of

siRNA-Cdx2 resulted in a decrease in Cdx1 mRNA (Fig. 5C).

Discussion
Intestinal metaplasia has been extensively studied as a putative preneoplastic lesion in the human stomach [1]. In the present study, endogenous Cdx1, but not Cdx2, was induced by transgenic Cdx2 in the intestinal metaplasia of the Cdx2-transgenic mouse stomach. Cdx1 is essential for anteriorposterior vertebral patterning of the body axis in the early embryonic period [13], and its expression persists selectively in the intestinal epithelium from the later embryonic period to the adult [14]. In addition to its physiological expression, Cdx1 is ectopically expressed in the precancerous intestinal metaplasia of the stomach and Barretts esophagus. The regulatory mechanisms that modulate Cdx1 gene expression during development
5825

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

Cdx1 expression in intestinal metaplasia

H. Mutoh et al.

A Cdx1 promoter

B Cdx2 promoter

Fig. 3. Cdx1 (A) and Cdx2 (B) promoter bisulte sequencing of the stomach and intestine of normal mice and the intestinal metaplasia of the Cdx2-transgenic mouse stomach. (A) A sequence of the 5-anking region for the mouse Cdx1 gene, including the TATA box, transcription start site and initiation codon (ATG). The TATA box is highlighted in green, the transcription start site in blue, and the initiation codon (ATG) in red. Cdx1 promoter CpGs are shown in red. Base positions relative to the Cdx1 transcription start site are shown on the left of each line of sequence. All CpGs were unmethylated. CpGs enclosed by the box ()54 to )68) represent those suggested to be crucial for transcriptional control [12]. (B) A sequence of the 5-anking region for the mouse Cdx2 gene, including the TATA box, transcription start site, and initiation codon (ATG). The TATA box and another AT-rich motif, designated DBS (downstream binding site) [28], are highlighted in green, the transcription start site in blue, and the initiation codon (ATG) in red. Cdx2 promoter CpGs are shown in red. Base positions relative to the Cdx2 transcription start site are shown on the left of each line of sequence. All CpGs were unmethylated, except for the CpG enclosed by the box, which was methylated in the normal intestine and the intestinal metaplasia.

and in the normal intestinal epithelium have been gradually claried. Cdx1 is a direct transcriptional target of both retinoic acid (RA) and the Wnt b-catenin signaling pathway during early embryogenesis [15,16]. The Wnt b-catenin signaling pathway is also active in the crypt compartment [17]. Cdx1 regulation by RA and Wnt3a is mediated, respectively, through the RA response element and two LEF TCF response elements present on the Cdx1 promoter [17]. However, very little is known about the molecular mechanisms for induction of the ectopic expression of the Cdx1 gene in the intestinal metaplasia of the H. pyloriinfected human stomach. In the present study, we focused on the initiation of Cdx1 gene transcription in the intestinal metaplasia through Cdx2-transgenic mouse studies. Unlike in normal regulation, ectopic expression of Cdx1 was upregulated by Cdx2. Cdx2 mRNA and protein were absent in the gastric-like heteroplasias arising spontaneously in the pericecal region and proximal colon of Cdx2 + ) mice, and, in common with that of Cdx2, Cdx1 expression was also
5826

absent in the gastric-like heteroplasias [18]. The nding that the gastric-like heteroplasia, which does not express Cdx2, also shows a lack of Cdx1 expression is consistent with our present data showing that the stomach expressing Cdx2 generated endogenous Cdx1. Epigenetic inactivation, in particular aberrant DNA hypermethylation, is an important mechanism for gene silencing. In the majority of human colon cancer specimens and colorectal cancer cell lines, Cdx1 expression is lost due to active Cdx1 gene silencing by promoter hypermethylation [12,19]. However, in this study, we demonstrated that the Cdx1 promoter is unmethylated in the normal stomach, the normal intestine, and the intestinal metaplasia, indicating that loss of Cdx1 expression in the normal stomach is not associated with promoter hypermethylation. Cdx1 and Cdx2 proteins bind to a binding site in an AT-rich motif whose consensus sequence is C TATAAAT G in direct or reverse orientation [20]. In some instances, the Cdx-binding site presents high homology with the

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

H. Mutoh et al.

Cdx1 expression in intestinal metaplasia

400

D
Cdx1 promoter-fw1

341 281 221

Cdx1 promoter-fw2
161 101

TATA box
41

Transcription start site (+1)

+20 +80 Initiation codon

Cdx1 promoter-rv1

Fig. 4. Cdx2 is present on the Cdx1 promoter region in vivo. (A) A sequence of the 5-anking region for the mouse Cdx1 gene, including the TATA box, transcription start site, and initiation codon. The TATA box is highlighted in green, the transcription start site in blue, and the initiation codon in red. PCR fragments corresponding to the DNA sequences including the TATA box were designed for ChIP analysis. The sequences for the primers used for ChIP assays are underlined and highlighted in yellow. The base positions relative to the transcription start site for the mouse Cdx1 gene are shown on the left of each line of sequence. (B, C) ChIP assays that were performed using a Cdx2 antibody [26] or control IgG. The region of the Cdx1 promoter encompassing the TATA box sequence was amplied by PCR with the following primer pairs: (B) Cdx1 promoter-fw1 and Cdx1 promoter-rv1; (C) Cdx1 promoter-fw2 and Cdx1 promoter-rv1. Lane 1: normal stomach. Lane 2: normal intestine. Lane 3: Cdx2-transgenic mouse stomach. Lane 4: input. Lane 5: control IgG. (D) EMSA. A radiolabeled dsDNA probe (CCCGCGGCTATAAAAGGCCGGGGTGGGG) containing the TATAAA sequence in the Cdx1 promoter was incubated with nuclear extracts from AGS cells and separated on a 5% polyacrylamide gel (lane 2). Specicity was determined by addition of antibody for supershift (lane 3) and mutant probe (CCCGCGGCTTCGAAAGGCCGGGGTGGGG) (lane 1).

canonical TATA-box sequence, and, indeed, the Cdx1 and or Cdx2 homeoproteins were found to be able to bind to the TATA-boxes of some intestinal genes, such as those of the calbindin-D9 gene [21], the clusterin gene [22], and the glucose-6-phosphatase gene [23]. In the present study, CpGs in the 5-region of the TATAAAA sequence located )51 )45 upstream of the transcription start site were also found to be unmethylated in the normal stomach, the normal intestine, and the intestinal metaplasia. The present results, including those from ChIP, EMSA and reporter gene analysis, indicate that Cdx2 is present on the Cdx1 promoter region containing the TATAAAA sequence located at )51 )45. On the other hand, endogenous Cdx2 was not expressed in the intestinal metaplasia of the Cdx2-transgenic mouse stomach, indicating that endogenous Cdx2 was not autoregulated. In the present study, we demonstrated that Cdx1 is expressed in the Cdx2-induced intestinal metaplasia of Cdx2-transgenic mice. This may coincide with our

previous clinical data why the expression of Cdx2 precedes that of Cdx1 during the progression of intestinal metaplasia [3]. These clinical data also suggest that Cdx2 might induce Cdx1 expression. In conclusion, we propose that the ectopic expression of Cdx2 in the gastric epithelium is triggered rst, and in turn Cdx1 is directly induced by Cdx2 in the intestinal metaplasia. The present results indicate that Cdx2 induces Cdx1 expression by directly binding to the Cdx2-consensus cis-regulatory element of the unmethylated Cdx1 promoter region.

Experimental procedures
Cdx2-transgenic mice
The Cdx2-transgenic mice we generated had free access to standard food and drinking water and were maintained on a 12 h light dark cycle. All experiments in this study were performed in accordance with the Jichi Medical University Guide for Laboratory Animals.

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

5827

Cdx1 expression in intestinal metaplasia

H. Mutoh et al.

Cdx2 Cdx1 GAPDH

B
0
Luciferase

AGS MKN45 MKN74

Luciferase activity (Ratio of firefly to renilla luciferase) 2 4 6 8 10 12 14 16

18

() () ()

365

+12

() +Cdx2 +siCdx2 () +Cdx2 +siCdx2 () +Cdx2 +siCdx2

365

78

() () ()

Fig. 5. Activation of the Cdx1 promoter in Cdx2-expressing AGS, MKN45 and MKN74 cells. (A) Cdx2 and Cdx1 expression determined by RT-PCR. Human gastric carcinoma AGS, MKN45 and MKN74 cells expressed both Cdx2 and Cdx1. The lower panel in (A) shows standard RT-PCR conducted with primers designed to detect GAPDH mRNA. (B) Cdx1 promoter reporter gene activities. AGS, MKN45 and MKN74 cells were transiently transfected with the different fragments of Cdx1 promoter fused to a luciferase reporter vector, pGL4.10[luc2], and pGL4.70[hRluc] vector. Luciferase activities were normalized relative to the level of Renilla luciferase activities. The lengths of the promoter fragments tested are indicated. The numbers correspond to the relative positions with respect to the transcription start site. The sequence of the presumptive Cdx2-binding site (TATAAA) was changed to TTCGAA. Cdx1 promoter reporter plasmids were added to each plate with or without Cdx2 expression vector (pRC CMVCdx2) or siRNA (Applied Biosystems, Silencer Select Pre-designed siRNA, #s2878; UUCUUGUUGAUUUUCCUCUcc). The luciferase activities of empty pGL4.10[luc2], which does not contain any Cdx1 promoter, were used as controls for AGS, MKN45 and MKN74 cells, respectively. Each bar represents the mean standard error. Transfections were performed in triplicate and repeated three times. (C) Cdx1 mRNA levels of AGS, MKN45 and MKN74 cells transfected with Cdx2 expression plasmid, Cdx2 siRNA, or negative control. At 24 h after transfection, total RNA was extracted.

365

() +Cdx2 +siCdx2 +12 () +Cdx2 TATAAA +siCdx2 51 46 () TTCGAA +Cdx2 +siCdx2

AGS MKN45 MKN74

Relative Cdx1 expression

C 2

0
Cdx2 siRNA

(normal mice), and intestinal metaplasia (Cdx2-transgenic mice), PCR amplication was performed using the primer pairs Cdx1cDNAfw and Cdx1cDNArv (for endogenous Cdx1), Cdx2-fw and Cdx2 coding-rv (for total Cdx2), and Cdx2-fw and Cdx2 noncoding-rv (for endogenous Cdx2) (Table 1), by incubation at 94 C for 2 min, followed by 35 cycles of 94 C for 30 s, 60 C for 30 s and 72 C for 30 s, and a nal extension at 72 C for 10 min. The PCR products were separated in 2% agarose gels. As an internal standard, RT-PCR was performed with primers hybridizing to the mRNA encoding b-actin or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Table 1).
+

+ AGS

MKN45

MKN74

Real-time RT-PCR
One hundred nanograms of cDNA was used in each realtime PCR reaction. Expression levels for the Cdx1 gene were determined by real-time PCR using ready-to use Assay-on-Demand gene expression product (Applied Biosystems, Foster City, CA, USA): Mm00438172_m1 for mouse Cdx1, and Hs00156451_m1 for human Cdx1. Each Assay-on-Demand gene expression product contains target-specic primers and probes and a Taqman Gene Expression Master Mix containing AmpErase uracil-N-glycosylase (Applied Biosystems) to prevent reamplication of carryover PCR products. PCR amplication and uorescence data collection were performed with the ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems), using the following conditions: 50 C for

RNA isolation and RT-PCR

Total RNA was extracted from the stomach (normal mice), small intestine (normal mice), intestinal metaplasia (Cdx2transgenic mice), and human gastric cancer cell lines AGS, MKN45 and MKN74, using the guanidinium isothiocyanate phenol method (Isogen; Nippon Gene, Tokyo, Japan), according to the manufacturers instructions. Total RNA (1 lg) was reverse-transcribed as previously described [24]. To compare endogenous Cdx1 expression, endogenous Cdx2 expression and total (endogenous and transgenic) Cdx2 expression in the stomach (normal mice), small intestine

5828

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

H. Mutoh et al.

Cdx1 expression in intestinal metaplasia

2 min, 95 C for 10 min, and 40 cycles for amplication (95 C for 15 s, and 60 C for 1 min). PCR reactions were performed in 96-well plates, using a nal volume of 20 lL, and the Cdx1 gene was studied in triplicate. In order to normalize RNA transcript abundance for the Cdx1 gene, a housekeeping gene (the b-actin gene) (Pre-Developed Taqman Assay Reagents; Applied biosystems) was used to calculate the DCT (DCT = CT target CT actin). The Ct values for the b-actin gene for the normal stomach, the normal intestine and Cdx2-transgenic mouse stomach tissues fell in a close range, with no specic pattern of spatial or temporal variation (data not shown). A relative quantication approach was used in this study to describe the change in expression of the target gene in a test sample relative to a calibrator sample (reference). The relative RNA transcript abundance value was calculated as follows. First, the DCT for the normal stomach, normal small intestine and Cdx2transgenic mouse stomach tissues was calculated. In the second step, differences between the normal and Cdx2transgenic mouse stomach tissues were calculated as DDCT (DCT target DCT reference). The normal mouse small intestine was used as reference for Cdx1 expression. Finally, the fold difference (relative abundance) was calculated using the formula 2)DDCT [25], and was plotted as means (n = 6).

reaction was performed using the forward primer CpG-Cdx2-fw1[)234 )212] and reverse primer CpG-Cdx2rv[+194 +173] (Table 1). A second, nested, PCR was then performed on 1 lL of the amplicate, using the upstream (CpG-Cdx2-fw2[)206 )185]) and downstream (CpG-Cdx2-rv[+194 +173]) primers (Table 1). The primer pairs were designed to bind sequences lacking any CpGs, therefore avoiding any preferential amplication of methylated or unmethylated DNA strands. The PCR products were puried (GenElute agarose spin column; Sigma, St Louis, MO, USA), and the puried product was used for cloning (Topo TA Cloning kit; Invitrogen, Carlsbad, CA, USA) and sequencing by using the Big Dye Terminator Cycle Sequencing kit (Applied Biosystems).

ChIP assay
The mucosae removed from the stomach (normal mice), the small intestine (normal mice) and the intestinal metaplasia (Cdx2-transgenic mice) were incubated with xation solution (1% formaldehyde, 4.5 mm Hepes, pH 8.0, 9 mm NaCl, 0.09 mm EDTA, 0.04 mm EGTA) in NaCl Pi for 30 min at 37 C. The reaction was terminated by the addition of glycine to a nal concentration of 150 mm. After being washed in NaCl/Pi containing protease inhibitors (Protease inhibitor cocktail; Sigma), the samples were sonicated in SDS lysis buffer (50 mm Tris HCl, pH 8.0, 10 mm EDTA, pH 8.0, 1% SDS, 0.5 mm phenylmethanesulfonyl uoride), when the DNA size of samples was 200500 bp. The solubilized chromatin was incubated with anti-Cdx2 IgG (BioGenex, San Ramon, CA, USA) [26] or control IgG for 90 min at 4 C. Beads were washed ve times with IP buffer (50 mm Hepes, pH 7.5, 150 mm KCl, 5 mm MgCl2, 10 lm ZnSO4, 1% Triton X-100, 0.05% SDS), and then incubated with elution buffer (50 mm Tris HCl, pH 8.0, 1% SDS, 10 mm EDTA) for 30 min at 65 C. The supernatant was collected and coimmunoprecipitated DNA was recovered. Primer sequences used for the ChIP assays are listed in Table 1. All ChIP assays were repeated at least twice, and representative data are presented.

Immunohistochemistry
Murine tissue sections were stained with the antibody for Cdx1 (1 : 40, rabbit polyclonal; Abcam, Cambridge, UK) after antigenicity was enhanced by autoclaving the sections, as previously described [24].

Bisulte sequencing for Cdx1 and Cdx2 promoters

The methylation status of gene promoter CpGs is best analyzed by using direct sequencing after sodium bisulte modication of target DNA (bisulte sequencing). DNA (1 lg of DNA per sample) was sodium bisulte modied with the DNA modication kit (Zymo Research Intergen, Purchase, NY, USA), according to the manufacturers instructions. A 426 bp region of Cdx1 was amplied from bisulte-modied genomic DNA by nested PCR using two sets of primers. Genomic DNAs were extracted from ve stomachs and ve intestines of ve normal mice and ve stomachs of ve Cdx2-transgenic mice. The rst PCR reaction was performed using the forward primer CpGCdx1-fw1[)331 )305] and the reverse primer CpG-Cdx1rv[+114 +89] (Table 1). A second, nested, PCR was then performed on 1 lL of the amplicate, using the upstream (CpG-Cdx1-fw2[)312 )293]) and downstream (CpG-Cdx1rv[+114 +89]) primers (Table 1). A 400 bp region of Cdx2 was amplied from bisulte-modied genomic DNA by nested PCR, using two sets of primers. The rst PCR

EMSA
Nuclear fractions were extracted for EMSA from AGS cells. To extract nuclear fractions for EMSA studies, AGS cells were washed in NaCl Pi, and subjected to swelling in 400 lL of hypotonic buffer A (10 mm Hepes, pH 7.9, 10 mm KCl, 0.1 mm EDTA, 0.1 mm EGTA, 1 mm dithiothreitol) supplemented with protease inhibitor cocktail (Sigma Chemical Co.), and lysed [27]. Then, 25 lL of 10% Nonidet P-40 solution were added, and nuclear fractions were collected by sedimentation for 5 min at 500 g. Supernatants were discarded, and precipitated nuclei were resuspended in 100 lL of buffer C (20 mm Hepes, pH 7.9, 400 mm NaCl, 1 mm dithiothreitol, 1 mm EDTA, 1 mm

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

5829

Cdx1 expression in intestinal metaplasia

H. Mutoh et al.

EGTA, and protease inhibitor cocktail) and centrifuged for 5 min at 14 000 g. Supernatants containing nuclear proteins were collected, and tested for their ability to bind labeled nucleotides corresponding to the Cdx1 promoter. All DNAprotein binding reaction protocols were those of the manufacturer (Promega, Madison, WI, USA). The dsDNA probes used in the gel mobility shift assays were as follows: wild-type sequence, CCCGCGGCTATAAAAGGCCGGG GTGGGG; mutant sequence, CCCGCGGCTTCGAAAG GCCGGGGTGGGG. Briey, 0.5 ng of 32P-labeled probe was incubated for 20 min at 4 C with 5 lg of nuclear extracts in the presence of 1 gel shift buffer (Promega). Subsequently, 1.5 lL of 10 loading buffer were added to the reaction, and this was followed by separation by electrophoresis on 5% nondenaturing polyacrylamide gel until free probe was close to the bottom of the gel.

Transfections of Cdx2 expression plasmid or Cdx2 siRNA

AGS, MKN45 and MKN74 cells were plated in 10 cm plates 24 h before transfection. Transfections were performed using Lipofectamine 2000, following the manufacturers protocol (Invitrogen). Six micrograms of Cdx2 expression plasmid and 25 pmol of siRNA or negative control were used for the transfection. siRNA (Applied Biosystems, Silencer Select Pre-designed siRNA, #s2878; UUCUUGUUGAUUUUCCUCUcc) was used. At 24 h after transfection, total RNA was extracted.

References
1 Correa P (1992) Human gastric carcinogenesis: a multistep and multifactorial process First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention. Cancer Res 52, 67356740. 2 Silberg DG, Furth EE, Taylor JK, Schuck T, Chiou T & Traber PG (1997) CDX1 protein expression in normal, metaplastic, and neoplastic human alimentary tract epithelium. Gastroenterology 113, 478486. 3 Eda A, Osawa H, Yanaka I, Satoh K, Mutoh H, Kihira K & Sugano K (2002) Expression of homeobox gene CDX2 precedes that of CDX1 during the progression of intestinal metaplasia. J Gastroenterol 37, 94100. 4 Satoh K, Mutoh H, Eda A, Yanaka I, Osawa H, Honda S, Kawata H, Kihira K & Sugano K (2002) Aberrant expression of CDX2 in the gastric mucosa with and without intestinal metaplasia: effect of eradication of Helicobacter pylori. Helicobacter 7, 192198. 5 Mutoh H, Hakamata Y, Sato K, Eda A, Yanaka I, Honda S, Osawa H, Kaneko Y & Sugano K (2002) Conversion of gastric mucosa to intestinal metaplasia in Cdx2-expressing transgenic mice. Biochem Biophys Res Commun 294, 470479. 6 Mutoh H, Satoh K, Kita H, Sakamoto H, Hayakawa H, Yamamoto H, Isoda N, Tamada K, Ido K & Sugano K (2005) Cdx2 species the differentiation of morphological as well as functional absorptive enterocytes of the small intestine. Int J Dev Biol 49, 867871. 7 Mutoh H, Sakurai S, Satoh K, Tamada K, Kita H, Osawa H, Tomiyama T, Sato Y, Yamamoto H, Isoda N et al. (2004) Development of gastric carcinoma from intestinal metaplasia in Cdx2-transgenic mice. Cancer Res 64, 77407747. 8 Eda A, Osawa H, Satoh K, Yanaka I, Kihira K, Ishino Y, Mutoh H & Sugano K (2003) Aberrant expression of CDX2 in Barretts epithelium and inammatory esophageal mucosa. J Gastroenterol 38, 1422. 9 Kazumori H, Ishihara S, Rumi MA, Kadowaki Y & Kinoshita Y (2006) Bile acids directly augment caudal related homeobox gene Cdx2 expression in oesophageal keratinocytes in Barretts epithelium. Gut 55, 1625.

Luciferase assays
To construct the luciferase reporter vector pGL4.10[luc2] Cdx1, 377 bp ()365 to +12) and 288 bp ()365 to )78) fragments, located at 5-region of the mouse Cdx1 coding sequence, were amplied by PCR with specic primers (Table 1) from 500 ng of mouse genomic DNA. The amplied fragments for the Cdx1 promoter were directly cloned into the TA cloning vector pCRII (Invitrogen), to yield the plasmid pCRII Cdx1 promoter. Each pCRII Cdx1 promoter was digested with XhoI and HindIII (sites underlined in the primers in Table 1), and the resulting fragments were subcloned into the XhoI and HindIII restriction sites of the pGL4.10[luc2] vector (Promega) and conrmed by sequence analysis. The sequence of the presumptive Cdx2-binding site (TATAAA) was changed to TTCGAA (underlined in the primers) by using Cdx1-Csp45I-fw and Cdx1-Csp45I-rv primers (Table 1). AGS, MKN45 and MKN74 cells were seeded at 2 105 cells per well in Nunc 24-well dishes 1824 h before transfection. Transient transfections were performed using Lipofectamine 2000 (Invitrogen). One hundred nanograms of a Cdx1 promoter reporter plasmid with or without 800 ng of Cdx2 expression vector (pRC CMVCdx2) or 2.5 pmol of siRNA (Applied Biosystems, Silencer Select Pre-designed siRNA, #s2878; UUCUUGUUGAUUUUC CUCUcc) were added to each plate, together with 50 ng of the Renilla luciferase control reporter plasmid (pGL4.70[hRluc]; Promega) as a control for the transfection efciency. At 24 h after transfection, the cells were lysed in lysis buffer (Promega), and the rey and Renilla luciferase activities were measured, using the Dual-Luciferase Reporter Assay System (Promega) in a luminometer. The relative rey luciferase activities were calculated by normalizing the transfection efciency according to the Renilla luciferase activities produced by the internal control plasmid pGL4.70[hRluc]. Three separate experiments were performed in triplicate.

5830

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

H. Mutoh et al.

Cdx1 expression in intestinal metaplasia

10 Debruyne PR, Witek M, Gong L, Birbe R, Chervoneva I, Jin T, Domon-Cell C, Palazzo JP, Freund JN, Li P et al. (2006) Bile acids induce ectopic expression of intestinal guanylyl cyclase C through nuclear factorkappaB and Cdx2 in human esophageal cells. Gastroenterology 130, 11911206. 11 Liu T, Zhang X, So CK, Wang S, Wang P, Yan L, Myers R, Chen Z, Patterson AP, Yang CS et al. (2007) Regulation of Cdx2 expression by promoter methylation, and effects of Cdx2 transfection on morphology and gene expression of human esophageal epithelial cells. Carcinogenesis 28, 488496. 12 Wong NA, Britton MP, Choi GS, Stanton TK, Bicknell DC, Wilding JL & Bodmer WF (2004) Loss of CDX1 expression in colorectal carcinoma: promoter methylation, mutation, and loss of heterozygosity analyses of 37 cell lines. Proc Natl Acad Sci USA 101, 574579. 13 Subramanian V, Meyer BI & Gruss P (1995) Disruption of the murine homeobox gene Cdx1 affects axial skeletal identities by altering the mesodermal expression domains of Hox genes. Cell 83, 641653. 14 Silberg DG, Swain GP, Suh ER & Traber PG (2000) Cdx1 and cdx2 expression during intestinal development. Gastroenterology 119, 961971. 15 Houle M, Prinos P, Iulianella A, Bouchard N & Lohnes D (2000) Retinoic acid regulation of Cdx1: an indirect mechanism for retinoids and vertebral specication. Mol Cell Biol 20, 65796586. 16 Ikeya M & Takada S (2001) Wnt-3a is required for somite specication along the anteroposterior axis of the mouse embryo and for regulation of cdx-1 expression. Mech Dev 103, 2733. 17 Lickert H, Domon C, Huls G, Wehrle C, Duluc I, Clevers H, Meyer BI, Freund JN & Kemler R (2000) Wnt (beta)-catenin signaling regulates the expression of the homeobox gene Cdx1 in embryonic intestine. Development 127, 38053813. 18 Bonhomme C, Duluc I, Martin E, Chawengsaksophak K, Chenard MP, Kedinger M, Beck F, Freund JN & Domon-Dell C (2003) The Cdx2 homeobox gene has a tumour suppressor function in the distal colon in addition to a homeotic role during gut development. Gut 52, 14651471.

19 Suh ER, Ha CS, Rankin EB, Toyota M & Traber PG (2002) DNA methylation down-regulates CDX1 gene expression in colorectal cancer cell lines. J Biol Chem 277, 3579535800. 20 Margalit Y, Yarus S, Shapira E, Gruenbaum Y & Fainsod A (1993) Isolation and characterization of target sequences of the chicken CdxA homeobox gene. Nucleic Acids Res 21, 49154922. 21 Lambert M, Colnot S, Suh E, LHorset F, Blin C, Calliot ME, Raymondjean M, Thomasset M, Traber PG & Perret C (1996) cis-Acting elements and transcription factors involved in the intestinal specic expression of the rat calbindin-D9K gene: binding of the intestinespecic transcription factor Cdx-2 to the TATA box. Eur J Biochem 236, 778788. 22 Suh E, Wang Z, Swain GP, Tenniswood M & Traber PG (2001) Clusterin gene transcription is activated by caudal-related homeobox genes in intestinal epithelium. Am J Physiol Gastrointest Liver Physiol 280, G149G156. 23 Gautier-Stein A, Domon-Dell C, Calon A, Bady I, Freund JN, Mithieux G & Rajas F (2003) Differential regulation of the glucose-6-phosphatase TATA box by intestine-specic homeodomain proteins CDX1 and CDX2. Nucleic Acids Res 31, 52385246. 24 Mutoh H, Sakurai S, Satoh K, Osawa H, Hakamata Y, Takeuchi T & Sugano K (2004) Cdx1 induced intestinal metaplasia in the transgenic mouse stomach: comparative study with Cdx2 transgenic mice. Gut 53, 1416 1423. 25 Livak KJ & Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 25, 402 408. 26 Uesaka T & Kageyama N (2004) Cdx2 homeodomain protein regulates the expression of MOK, a member of the mitogen-activated protein kinase superfamily, in the intestinal epithelial cells. FEBS Lett 573, 147154. 27 Schreiber E, Matthias P, Muller MM & Schaffner W (1989) Rapid detection of octamer binding proteins with mini-extracts, prepared from a small number of cells. Nucleic Acids Res 17, 6419. 28 Xu F, Li H & Jin T (1999) Cell type-specic autoregulation of the caudal-related homeobox gene Cdx-2 3. J Biol Chem 274, 3431034316.

FEBS Journal 276 (2009) 58215831 2009 The Authors Journal compilation 2009 FEBS

5831