Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Sep 17.
Published in final edited form as: Cell. 2010 Sep 17;142(6):930–942. doi: 10.1016/j.cell.2010.08.030

DNA Demethylase Activity Maintains Intestinal Cells in an Undifferentiated State Following Loss of APC

Kunal Rai 1,2,3,*, Sharmistha Sarkar 1,3,*, Talmage J Broadbent 1,3,*, Matthew Voas 1,3, Kenneth F Grossmann 1,2,3, Lincoln D Nadauld 1, Somaye Dehghanizadeh 1,3, Fanuel T Hagos 1,3, Yumei Li 1,6, Rachel K Toth 1,3, Stephanie Chidester 1, Timothy M Bahr 7, W Evan Johnson 1,6, Brad Sklow 1,6, Randall Burt 1,5, Bradley R Cairns 1,2,3, David A Jones 1,3,4
PMCID: PMC2943938  NIHMSID: NIHMS234412  PMID: 20850014

Summary

Although genome-wide hypomethylation is a hallmark of many cancers, roles for active DNA demethylation during tumorigenesis are unknown. Here, loss of the APC tumor suppressor gene causes upregulation of a DNA demethylase system and the concomitant hypomethylation of key intestinal cell fating genes. Notably, this hypomethylation maintained zebrafish intestinal cells in an undifferentiated state which was released upon knock down of demethylase components. Mechanistically, the demethylase genes are directly activated by Pou5f1 and Cebpβ, and indirectly repressed by retinoic acid, which antagonizes Pou5f1 and Cebpβ. Apc mutants lack retinoic acid, due to the transcriptional repression of retinol dehydrogenase l1 via a complex that includes Lef1, Groucho2, Ctbp1, Lsd1 and Corest. Our findings imply a model wherein APC controls intestinal cell fating through a switch in DNA methylation dynamics. Wildtype APC and retinoic acid downregulate demethylase components, thereby promoting DNA methylation of key genes and helping progenitors commit to differentiation.

Introduction

Loss of adenomatous polyposis coli (APC) tumor suppressor gene is the key initiating step in a model of genetic and epigenetic events that lead to colorectal cancer. Germline mutations in APC result in the development of Familial Adenomatous Polyposis (FAP), a syndrome characterized by early onset of colon cancer (Bienz and Clevers, 2000). In addition, APC is mutated in more than 85% of sporadic colon tumors (Sparks et al., 1998). The primary function of APC is attributed to its role in negatively regulating WNT signaling. Mechanistically, APC controls WNT signaling by targeting the transcriptional coactivator β-catenin for proteasomal degradation, thereby, preventing its association with the nuclear transcription factor TCF/LEF (Bienz and Clevers, 2000). The mechanisms underlying APC suppression of tumor formation remain an area of active investigation with several lines of evidence suggesting that APC serves roles in addition to regulating WNT signaling. Recent reports suggest that APC controls retinoic acid production which in part relies upon the transcriptional regulator CtBP1 (Phelps et al., 2009b) (Nadauld et al., 2006).

Among the early events proposed to contribute to the development of a variety of cancers including colorectal cancer is an apparent genome-wide loss of DNA methylation (Goelz et al., 1985). A model explaining this hypomethylation includes that DNA methyltransferases fail to maintain normal patterns in highly proliferative cells. Consistent with this possibility, Dnmt1 hypomorph mice show genomic instability and develop T-cell lymphomas (Gaudet et al., 2003). Also, suppression of Dnmt1 accelerates loss of heterozygosity of Nf1 in Nf1+/− p53+/− (NPcis) mice (Eden et al., 2003). These studies argue strongly that Dnmt1-mediated loss of methylation results in chromosomal instability and oncogenesis. Paradoxically, however, there are no reports of DNMT mutation or loss in tumor development. In addition, loss or pharmacologic inhibition of DNMT1 appears protective in some circumstances. Notably, adenoma formation following loss of APC is reduced upon pharmacologic or genetic suppression of DNMT1 (Laird et al., 1995) (Eads et al., 2002). This suggests additional regulatory mechanisms linking loss of APC with changes in DNA methylation.

Here, we utilize zebrafish in an effort to understand the relationship between APC-dependent aberrant cell fating and a recently described DNA demethylase system (Rai et al., 2008). We show that homozygous loss of APC causes robust upregulation of the components of the demethylase and that this impairs the proper differentiation of intestinal cells. Mechanistically, upregulation of the demethylase results from an absence of retinoic acid, a factor which normally suppresses the expression of Cebpβ and Pou5f1 (previously known as Oct4), which are direct activators of demethylase components. Genome-wide methylation analyses using methylated DNA immunoprecipitation (MeDIP) identified a panel of key intestinal cell fating genes showing hypomethylation in apc homozygous mutant embryos, a hypomethylation that depended upon demethylase components. Parallel experiments using tissues taken for human subjects showed the demethylation of these same key developmental genes in human colon adenomas. Further, we demonstrate the direct repression of the retinol dehydrogenase (RDH) gene following APC mutation by multiple factors including LEF1, Groucho2/TLE3, LSD1, CoREST and CtBP1. These findings support interplay between APC, retinoic acid and DNA demethylase components in regulating intestinal cell fating and suggest misregulation of this interplay as an early event in colon adenoma development following loss of APC.

Results

Loss of APC Upregulates Demethylase Components and Correlates with Demethylation of Key Target Genes

Recently, we proposed a mechanism for active DNA demethylation in zebrafish which involves the cooperative actions of proteins from the cytidine deaminase family (activation induced deaminase (Aid), and Apobec2), the G:T mismatch specific glycosylase family (MBD4), and a DNA repair protein family (Gadd45) (Rai et al., 2008). We have also observed a partially redundant role in demethylation for the Mbd4 paralog Tdg (thymine glycosylase, data not shown) and, therefore, include Tdg in our analyses. Given the potential of DNA hypomethylation as an early event following loss of APC, we examined the demethylase components in zebrafish embryos harboring homozygous apc mutations (apcmcr) that result in a truncated protein similar to those observed in human tumors (Hurlstone et al., 2003). Remarkably, whole mount in situ hybridization revealed the robust upregulation of components of the DNA demethylase system including aid, apobec2a, apobec2b, mbd4, gadd45α, gadd45αlike, gadd45β and gadd45γ, and slight upregulation of tdg (Figure 1A-C and Table S1) at 72hpf. We observed no differences in expression of these genes between apc mutants and wildtype prior to 48hpf (Figure S1A and data not shown). In addition, homozygous apc mutant embryos show slightly reduced dnmt1 mRNA expression (Figure 1D and Figure S1B-D). By comparison, dnmt4 and dnmt8 appeared increased whereas expression of other dnmts did not change (Figure S1B).

Figure 1. Upregulation of DNA demethylase components, and hypomethylated genes in apc mutant embryos and FAP adenomas.

Figure 1

Open in a new tab

(A-D) Whole mount in situ staining for aid (A), apobec2a (A), apobec2b (A), mbd4 (B), tdg (B), gadd45α (C), gadd45αlike(C), gadd45β (C), gadd45γ (C) and dnmt1 (D) in apc mutants (apcmcr) and siblings (apcwt) zebrafish embryos at 72hpf. (E) MeDIP-qPCR for genes shown (selected from genome-wide MeDIP-ChIP microarray analysis) in apcmcr and apcwt (72hpf) which are either uninjected or injected with aaa Mo (combination of aid, apobec2a and apobec2b morpholinos; 0.5ng each) or mbd4 and tdg morpholinos together (1ng each) or V5-Dnmt1 expressing plasmid (1pg, an amount that rescues Dnmt1 morphants with 97% knockdown of Dnmt1 levels). Dnmt1 protein levels in Figure S1C. (F) MeDIP-qPCR for genes shown (selected from genome-wide MeDIP-ChIP microarray analysis) in human adenomas (A) and matching uninvolved tissues (U) from FAP patients. P1-P10 refers to ten different patients. In both E and F, the Y-axis shows values for each promoter region normalized to a negative control region lacking CpGs, and then normalized to the values from wild type or uninvolved, valued at 1. Error bars indicate +/− SD. See also Figure S1.

Although demethylase components were upregulated in apc mutant zebrafish embryos, we did not observe significant changes in bulk DNA methylation (data not shown), which largely monitors repetitive heterochromatin and transposons. To detect genome-wide changes in DNA methylation at promoters, we applied methylated DNA immunoprecipitation (MeDIP, (Weber et al., 2007)), which uses an antibody directed against 5-methyl-cytosine (5meC), followed by hybridization to Agilent 244K oligonucleotide arrays that tile the majority of zebrafish gene promoters (−9 to +2kb). Analysis of DNA taken from whole embryos uncovered hypomethylation on a number of interesting genes that have been implicated in intestinal fating and development of colorectal cancer including aldh1a2, hoxa13a, evx1, pitx2, cyclind1, hoxd13a, junbl, frizzled8a, cdx4, sox9b, cyclinb2, and sox4 (complete list Table S2). Verification of methylation status for a subset was conducted by qPCR of MeDIP samples (Figure 1E) or bisulfite sequencing (Figure S1E). To test whether the demethylation of these promoters resulted from the upregulation of the demethylase, we knocked down the demethylase components using morpholino antisense oligonucleotides (hereafter referred to as morpholinos) developed and validated in our previous studies (Rai et al., 2008). Following injection, we noted the reversal of hypomethylation for many genes tested (Figure 1E). Interestingly, overexpression of dnmt1 (via mRNA injection at the 1-cell stage, protein abundance data in Figure S1C) largely failed to restore promoter methylation (Figure 1E).

Next, we examined the conservation of the above findings in human colon adenomas obtained from FAP patients which harbor germline APC mutations. Consistent with zebrafish embryos, AID, MBD4 and GADD45α were significantly upregulated in at least six out of eleven colon adenoma samples which harbor germline APC mutations (Figure S1F). Genome-wide changes in promoter DNA methylation in adenoma samples were assessed from two different FAP patients (compared to normal-appearing colonic tissue) using MeDIP, which revealed significant overlap with promoters affected in apc mutant zebrafish (complete gene list in Table S3). A subset of these regions were tested and verified by qPCR from MeDIP performed in FAP adenoma tissue samples from 10 different patients. The promoter regions of CYCLINB2, HOXD13, JUN, PITX2, SOX4 and SOX17 showed demethylation (Figure 1F). PITX2 promoter hypomethylation was verified by bisulfite sequencing in adenomas from five different patients (Figure S1G).

Regulation of the Demethylation System by Retinoic Acid

Previous studies indicate that APC controls both Wnt signaling and retinoic acid (RA) biosynthesis (Phelps et al., 2009a). To determine whether aberrant demethylase expression resulted from misregulated Wnt signaling or loss of RA production in apc mutant zebrafish, we treated embryos with either all-trans retinoic acid (ATRA) or knocked down levels of β-catenin by blocking Cox-2 activity using the pharmacological inhibitor (NS-398) (Eisinger et al., 2007) (Figure S2A). Notably, treatment of apc mutants with ATRA reduced the expression of aid, gadd45α and mbd4 whereas treatment with NS-398 had little effect (Figure 2A-B and Table S4). However, RA treatment did not restore dnmt1 expression (Figure S1D). Importantly, three additional models of RA deficiency support a role for RA in downregulating the demethylase: rdh1l morphants (Nadauld et al., 2005), nls mutants (Begemann et al., 2001) and DEAB-(4-(diethylamino) benzaldehyde; an inhibitor of aldehyde dehydrogenases (Russo et al., 1988)) treated wild type embryos all showed increased expression of the demethylase genes (Figure 2C). Moreover, restroration of RA levels by CtBP1 knockdown in apcmcr embryos led to demethylase downregulation (Figure 2A). Finally, RA treatment led to repression of AID, MBD4 and GADD45α levels in human colon cancer cell lines (Figure S2B). Similar to our results following morpholino knock down of demethylase components, treatment of apc mutant embryos with RA reversed the loss of DNA methylation observed on the panel of genes implicated in intestinal differentiation and proliferation (Figure 2D). Taken together, we provide multiple lines of evidence for robust downregulation of demethylase components by retinoic acid.

Figure 2. APC negatively regulates demethylase expression via retinoic acid.

Figure 2

Open in a new tab

(A) Whole mount in situ staining for aid, gadd45α and mbd4 in apcmcr and apcwt embryos treated with DMSO, or all-trans retinoic acid (ATRA, 1μM) or Cox-2 inhibitor NS-398 (10μM) or injected with control morpholino or ctbp1 morpholino (2ng). β-catenin knockdown efficiency with NS-398 treatment is depicted in Figure S2A. (B-C) Quantitative RT-PCR for aid, mbd4 and gadd45α in apcmcr and apcwt embryos treated with DMSO or ATRA (B) or in different RA deficiency models in zebrafish, as shown (C). Expression of genes are first normalized to 28S, and then to the control embryo mRNA/28S ratio, valued as 1. (D) MeDIP-qPCR for genes shown (as in Figure 1E) in apc mutants (72hpf) which are either untreated or treated with ATRA. Error bars indicate +/− SD. See also Figure S2.

Cebpβ and pou5f1 Mediate Regulation of the Demethylase Downstream of Retinoic Acid

We have shown previously that, although apc mutant embryos develop primordial intestines expressing key patterning markers such as such as gata6, foxa3, gata5, hnf1β and hnf4α (Figure S3A), these intestinal cells fail to differentiate. Differentiation failure results from loss of retinol dehydrogenase (RDH) expression which catalyzes the conversion of dietary retinol into retinaldehyde (Jette et al., 2004; Nadauld et al., 2004; Nadauld et al., 2005). To better understand the status of RA biosynthesis following apc mutation, we also examined the expression of aldh1a2 (also known as raldh2) which further coverts retinaldehyde into retinoic acid. (We note that zebrafish lack aldh1a1). ALDH1 is a commonly used marker of normal and cancer stem cells including those derived from human colon and colon carcinomas (Huang et al., 2009). Whole mount in situ hybridization revealed upregulation of aldh1a2 in apc mutants beginning at 36hpf which remained elevated through 72hpf (Figure 3A, S3B,C). This was paralleled by increased aldh1a2 protein levels (Figure S3D). The upregulation of aldh1a2 in apcmcr embryos combined with an absence of retinol dehydrogenases (rdh1 and rdh1l) suggested poising of the intestinal progenitor cells for retinoic acid biosynthesis upon production of the required substrate. Indeed, treatment of apcmcr embryos with exogenous retinaldehyde (RAL) partially restored intestinal differentiation marked by fabp2 expression (Mudumana et al., 2004) (Figure 3B), a result we also observed in rdh1l morphant embryos (Figure 3C). Consistent with studies indicating that aldh1a2 expression can be repressed by RA (Dobbs-McAuliffe et al., 2004) (Niederreither et al., 1997), treatment of apc mutant, rdh1l morphant or nls mutant embryos with RA decreased the expression of aldh1a2 (Figure S3C,E,F).

Figure 3. Pou5f1 and Cebpβ directly regulate aid, mbd4 and gadd45α expression in apcmcr embryos.

Figure 3

Open in a new tab

(A) Whole mount in situ staining for aldh1a2 in apcmcr and apcwt at 72hpf. (B-C) Whole mount in situ staining for fabp2 in apcmcr and apcwt embryos (B) or in control morphant and rdh1l morphant embryos (C) (at 72hpf) treated with vehicle or RAL (2μM). (D) RT-PCR for Pou5f1 (Oct4) and Cebpβ in apcmcr and apcwt treated with DMSO or ATRA (1μM). The Y-axis shows fold induction normalized to 28S and wild type DMSO treated sample. (E-F) Whole mount in situ staining for aid, gadd45α and mbd4 in apcmcr and apcwt injected with control, Pou2f1a, pou5f1 or cebpβ Mo (E) or in wild type embryos injected with control, Pou2f1, pou5f1 or cebpβ expression plasmid (F). (G) Graph showing fold enrichment near the aid or gadd45α TSS (containing overlapping Oct and Cebp binding sites) for Cebpβ and Pou5f1 in embryos injected with V5-Cebpβ (along with Pou5f1 Mo, 80pg), V5-Pou5f1 or FLAG-Pou2f1 expressing plasmids. ChIP was performed with antibodies against the tags. Normalization control primers are located ~3kb upstream (region without Cebpβ sites) of TSS of Gadd45α gene. Error bars indicate +/− SD. See also Figure S3.

As RA can directly represses transcription of pou5f1 in the zebrafish by binding to a RARE in the pou5f1 promoter (Parvin et al., 2008) and since a related POU family transcription factor activates expression of aldh1a1 (Guimond et al., 2002), we examined the expression of pou5f1 in apcmcr embryos and found it elevated (Figure 3D). Furthermore, treatment of apcmcr embryos with RA reduced the expression of pou5f1 (Figure 3D). Cebpβ is another transcription factor upregulated in apc mutants and down regulated by RA treatment (Figure 3D) (Eisinger et al., 2006). Since components of the demethylase complex contain OCT and CEBP sites in their promoter, we suppressed Cebpβ and Pou5f1 levels in apc mutants by injection of morpholinos and observed reduced expression of the demethylase to an extent similar to that elicited by treatment with RA (Figure 3E and Table S4). As a control, we also knocked down Pou2f1a (aka Oct1), Smad4 and Tcf3a/3b and saw no effect (Figure 3E, Table S4 and data not shown). To test if Pou5f1 and Cebpβ were sufficient to upregulate these genes, we overexpressed each and observed robust upregulation of all demethylase components in multiple tissues (Figure 3F and Figure S3G-H). Knock down of pou5f1 also caused a significant reduction in the aldh1a2 expression in apc mutants as measured by whole mount in situ and quantitative RT-PCR (Figure S3I-J).

Notably, gadd45α and aid promoters contain adjacent and partially overlapping OCT and CEBP recognition sequences at +400 and −1300 position from their respective transcription start sites (TSS). To assess binding, we performed chromatin immunoprecipitation (ChIP) following the expression of tagged derivatives in 56hpf embryos. Here, Pou5f1 enrichment was readily detected on both aid and gadd45α promoters when overexpressed (Figure 3G). Significant Cebpβ enrichment was detected when Cebpβ was overexpressed along with a small amount of pou5f1 morpholino (Figure 3G), however without pou5f1 morpholino Cebpβ enrichment was ~2 fold. It is possible that Pou5f1 sterically hinders binding of exogenous Cebpβ. These data suggest that both Pou5f1 and Cebpβ, whose levels are regulated by retinoic acid, are direct positive regulators of demethylase genes.

The Demethylase System Helps Maintain Intestinal Epithelial Cells in an Undifferentiated State Following Loss of Apc

As shown above apc mutant embryos expressed high levels of aldh1a2, a marker of colon crypt progenitor cells, (Figure 3A, 4A and Table S5) (Huang et al., 2009). Further, the aldh1a2 promoter was hypomethylated in apc mutant embryos (Figure 1E), and this hypomethylation relied on demethylase components (Figure 1E). In examining intestine-specific effects, we noted that cross sections of whole mount in situ stained embryos revealed clear upregulation of aid (52hpf) and mbd4 (72hpf) within the intestine of apc mutants (Figure 4B). Further, we generated an apc mutant line carrying gut-specific GFP driven by the foxa3 promoter and used these to sort and enrich gut cells from whole embryos. Sorting of GFP-positive cells resulted in approximately 11-fold enrichment based on RT-PCR analysis of fabp2 expression in parallel wild type embryos (Figure S4A-B). Analysis of DNA isolated from GFP-selected cells demonstrated specific demethylation of aldh1a2, hoxa13a, hoxd13a, cyclind1 and pitx2 in apc mutants compared to siblings (Figure 4C). Given these intestine-specific effects, we tested if the demethylase components were required for maintaining the undifferentiated progenitor population in apc mutants. Notably, knock down of demethylase components, but not of Pou2f1a, Tcf3a/3b or Smad4 (as controls), in apc mutant embryos reduced expression of aldh1a2 as determined both by whole mount in situ hybridization or RT-PCR (Figure 4A, S4C, Table S5 and data not shown), an effect that was paralleled by increased intestinal differentiation assessed by fabp2 expression (Figure 4A, S4C and Table S5). Interestingly, analysis of DLD1 and HT29 colon cancer cell lines showed increased expression of AID, MBD4 and GADD45α in ALDH positive cells (colon stem cells) in comparison to ALDH negative fractions (Figure S4D).

Figure 4. Demethylase activity is required to maintain progenitor cell populations in apc mutants.

Figure 4

Open in a new tab

(A) Whole mount in situ staining for aldh1a2, fabp2 and hoxa13a in apcmcr and apcwt embryos injected with control Mo, aaa Mo, combined mbd4 and tdg Mo or overexpression of dnmt1 as in Figure 1. (B) Cross-sections of the intestines of apcmcr and apcwt embryos whole-mount in situ stained with aid (52hpf) or mbd4 (72hpf). Asterisks show pro-nephric ducts and arrows point to the intestine. N-notochord, SC-spinal cord. (C) MeDIP-qPCR for genes shown in the GFP-positive cells from gutGFP;apcmcr and gutGFP;apcwt embryos (72hpf). FACS parameters and efficiency are shown in Figure S4A-B. (D) Immunostaining, (E) RT-PCR and (F) Me-DIP PCR for indicated genes in the intestines from adult wildtype zebrafish treated with DMSO or DEAB for 7 days. In panel E, RT-PCR was performed on epithelial cells isolated from the intestinal crypts. Error bars indicate +/− SD. See also Figure S4.

We observed a clear affect on the expression pattern of hoxa13a; wild type embryos restrict hoxa13a expression to the distal tip of the gut, whereas expression is observed throughout the gut in apc mutants. Strikingly, knock down of demethylase components restores proper spatial expression of hoxa13a in apc mutant embryos (Figure 4A and Table S5). Furthermore, hoxa13a was hypomethylated in apc mutant guts in a demethylase-dependent manner (Figure 1E, 4C). Finally, overexpression of dnmt1 did not affect expression of these markers (Figure 4A), focusing the causal affect on the upregulation of the demethylase rather than the possible downregulation of dnmt1.

The above studies in embryos suggested a model wherein retinoic acid mediated suppression of the DNA demethylase system is essential for intestinal differentiation. Next, we examined the impact of retinoic acid deficiency in the adult zebrafish intestine via treatment for seven days with DEAB to inhibit retinoic acid biosynthesis. This treatment caused a profound conversion of intestinal cells to a proliferative phenotype as assessed by PCNA staining and incorporation of BrdU (Figure 4D). Treatment of adults with both DEAB and ATRA restored normal proliferation and confirmed the specificity of the DEAB treatment (Figure 4D). Consistent with our findings in embryos, intestines from DEAB-treated adults showed increased expression of aid, gadd45α, mbd4, aldh1a2 and cyclinD1, with fabp2 levels coincidently reduced (Figure 4E, S4E). Consistent with these changes in expression, intestinal cells from the DEAB-treated adults showed demethylation of aldh1a2, hoxa13a, hoxd13a, cyclind1 and pitx2 (Figure 4F).

The Demethylase Governs Cell Proliferation in the Brain following Apc Loss

The impact of the demethylase system extends beyond intestinal differentiation, as clear effects on brain proliferation were observed in apc mutants. First, we observed high levels (by in situ staining) of pcna, cyclind1 and pitx2 (which activates cyclind1) within brain of apc mutants, which was greatly attenuated by knock down of demethylase components but not through ectopic dnmt1 expression (Figure S4F and Table S5). Furthermore, increased pcna and cyclind1 expression depended on Pitx2 (Figure S4F). As with the gut, primordial brain markers appeared normal in apcmcr embryos and ruled out gross patterning defects in accounting for the observed phenotypes (Figure S4G-H).

Lef1 and Groucho2 Directly Repress rdh1l and Suppress Intestinal Differentiation

Next we sought to determine the mechanism of RDH repression upon apc mutation to better understand the regulation of RA production. The RDH promoter in both humans and zebrafish contains numerous TCF/LEF binding sites, and apcmcr embryos expressed aberrantly high levels of lef1 and its co-regulator groucho2, but not of tcf4 or groucho3 (Figure 5A, S5A). Remarkably, knock down of Lef1 in apcmcr embryos improved many of the overt morphological defects present in apcmcr embryos including intestinal and pancreatic differentiation, but not eye differentiation defects (Figure 5B, S5B-F and Table S6). Knock down of either Lef1 or Groucho2 (Fig. S5G) was accompanied by the induction of rdh1l expression within the intestine (Figure 5B and Table S6). Importantly, Lef1 directly regulates rdh1l expression, as ChIP experiments revealed a 7.4-fold enrichment of Lef1 on the rdh1l promoter (second and the third Lef1 binding sites upstream of the TSS, Figure 5C) in apcmcr embryos compared to their wild type siblings (Figure 5D). Co-injection of both lef1 and groucho2 mRNA, but not either alone, reduced expression of fabp2 in the intestine of 80hpf wild type zebrafish embryos (Figure 5E and Table S5). The expression of gata6, a marker of primordial gut, was unaffected by injection of lef1 and groucho2 mRNA, thereby eliminating the possibility that the intestine failed to develop (Figure 5E).

Figure 5. Lef1 and Groucho2 suppress rdh1l in apcmcr zebrafish.

Figure 5

Open in a new tab

(A) Whole mount in situ staining for lef1, groucho2, tcf4 or groucho3 in apcwt and apcmcr embryos at 72hpf. Red arrow points to the intestine. (B) Whole mount in situ staining for fabp2, trypsin, rdh1l and irbp in apcmcr/control Mo, apcmcr /lef1 Mo and apcmcr /groucho2 Mo embryos at 72hpf. Groucho2 mo efficiency in Figure S5G. (C-D) ChIP for Lef1 on rdh1l promoter in apcwt and apcmcr embryos. (C) Schematic showing the Lef1 sites on the rdh1l promoter and location of the primer set P1 (containing Lef1 sites) and P2 (without Lef1 sites). (D) Graph showing fold enrichment of Lef1 occupancy at P1/P2 on rdh1l promoter in apcmcr embryos compared to apcwt at 72hpf. Lef1 values were normalized to ones obtained using a non-specific antibody and then expressed as fold enrichment compared to apcwt. Error bars indicate +/− SD. (E) Whole mount in situ staining for fabp2 and gata6 in 72hpf embryos injected with full length lef1 or dominant negative (deficient in β-catenin binding) lef1 mRNA alone or with groucho2 mRNA. See also Figure S5.

Retinoic acid production in APC mutant human cells is β-catenin independent (Naduald et al, 2006, Phelps et al 2009), and we reinforce this by showing that reduction in β-catenin levels by morpholino knockdown or pharmacological inhibition of Cox-2 (an upstream regulator of β-catenin) did not change levels of rdh1l or fabp2 in apcmcr zebrafish embryos (Figure S5H-J). Surprisingly, and consistent with data presented here, Lef1 levels remained unchanged upon β-catenin knockdown but were rescued upon reintroduction of APC (Figure S5J-K). Human LEF1 physically interacts with Groucho/TLE family members in the absence of WNT signaling. Notably, human LEF1 and TLE3 (a homolog of Groucho2) showed robust interaction by Co-IP (Figure 6A), and similarly zebrafish Lef1 and Groucho2 interacted (data not shown) in SW-480 cells, which harbor an APC mutation. A role for LEF1 and TLE3 overexpression may extend to human samples, as they were each significantly upregulated in 50-60% of the ten adenoma samples tested (Figure S6A-B). Next, we analyzed roles of LEF1 and Groucho2/TLE3 in repressing retinol dehydrogenase expression in three colon cancer cell lines harboring APC mutations: DLD1, HT29 and SW480. LEF1 and TLE3 were knocked down using short interfering RNAs (siRNAs), which significantly reduced their respective expression (Figure S6C-D) and restored DHRS9 (previously known as RDHL) expression in all three cell lines (Figure 6B). These data strongly support roles of LEF1 and TLE3 together in repressing RDH expression in zebrafish and humans.

Figure 6. CtBP1 and LSD1 physically interact with LEF1/Groucho2/TLE3 and repress RDHL/DHRS9 expression.

Figure 6

Open in a new tab

(A, C-E) Western blot showing co-immunoprecipitation between LEF1 and TLE3 (A), CtBP1 and TLE3 (C), LSD1 and TLE3 (D), LEF1 and LSD1 (E), in SW480 cells. Antibodies used for IP are shown on top and antibodies used for immunoblot are shown adjacent to blot. (B, H) Quantitative PCR measuring DHRS9 (RDHL) expression in DLD1, SW480, and HT29 cells which were transfected with either a Scrambled (SCR) siRNA or specific siRNAs against LEF1 or TLE3 (B) or siRNAs against LSD1 or CoREST (H) or treated with pargyline (3mM) (H). siRNA knockdown efficiency shown in Figure S6C-F. Y-axis values represent fold change in DHRS9 expression. Normalization for DHRS9 absolute values was done first to 18S rRNA values and then to DHRS9/18S ratio from Scr siRNA. Error bars indicate standard deviation. (F-G) ChIP of CtBP1 and LSD1 in untreated (F) or SCR or LEF1 siRNA treated (G) SW480 cells on the DHRS9 promoter. (F) PCR amplified product was run on agarose gel which shows enrichment on a region on DHRS9 promoter which contains LEF1 sites (P1) compared to a region devoid of LEF1 sites (P2). (G) Quantitative PCR showing enrichment of CtBP1 and LSD1. Y-axis shows fold enrichment on P1 compared to P2. Error bars indicate +/− SD. See also Figure S6.

LEF1/TLE3 represses retinol dehydrogenase through recruitment of the CtBP1/LSD1 corepressor complex

CtBP1 appears to suppress retinol dehydrogenase (RDH) expression in colon cancer cell lines and in apcmcr zebrafish (Nadauld et al., 2006). It has been shown to exist in a multi-protein repressor complex which includes the histone demethylase LSD1 and the scaffold CoREST (Shi et al., 2003) and. Interestingly, immunoprecipitation of TLE3 from SW480 cells efficiently co-precipitated CtBP1 (Figure 6C), and LSD1 (Figure 6D). Furthermore, LEF1 co-precipitated LSD1, suggesting that CtBP1/LSD1 complex can interact with both LEF1 and TLE3 (Figure 6E). Next, we tested if these interacting proteins can occupy the DHRS9 promoter. To test this, we performed ChIP of CtBP1 and LSD1 on the DHRS9 promoter in SW480 cells, which revealed significant enrichment of CtBP1 and LSD1 on the DHRS9 promoter region containing LEF1 sites (Figure 6F). Furthermore, CtBP1 and LSD1 occupancy of the DHRS9 promoter was reduced in cell lines treated with LEF1 siRNA suggesting that LEF1 helps in recruitment of CtBP1 complex to DHRS9 promoter (Figure 6G). Also, knockdown of LSD1 by siRNA was sufficient to restore DHRS9 levels in DLD1, HT29 and SW480 cells (Figure 6H, S6E). Moreover, inhibition of LSD1 enzymatic activity by treatment with pargyline (Shi et al., 2004) also upregulated DHRS9 expression in all three cell lines (Figure 6H). Finally, knockdown of CoREST also restored DHRS9 expression in the same three cell lines (Figure 6H, S6F). Taken together, these data suggest that a LEF1/TLE3 module binds to the retinol dehyrogenase promoter and recruits the CtBP1/LSD1/CoREST complex for repression.

LSD1 and CoREST are upregulated in APC mutant tissues and repress DHRS9

CtBP1 protein levels are elevated in APC mutant cells (Nadauld et al., 2006), as they lack APC- and proteasome-mediated CtBP1 destruction. We tested LSD1 protein levels in FAP polyps. Immunostaining on cross sections of FAP colonic adenomas showed upregulation of both LSD1 and CtBP1 compared to matched normal mucosa samples (Figure S6G). Western blot further confirmed upregulation of LSD1 protein in FAP polyps compared to matched normal tissue (Figure S6H). Here, unlike CtBP1, and similar to LEF1 and TLE3, LSD1 transcript levels were upregulated in multiple adenoma tissues by RT-PCR analysis (Figure S6I).

Next, we tested if LSD1 and CoREST function in intestinal differentiation and retinol dehydrogenase regulation using zebrafish. Immunoblot analysis revealed Lsd1 and Ctbp1 protein upregulation in apcmcr zebrafish embryos (Figure S7A). Similar to adenomas, the transcript levels of Lsd1 and Corest are also upregulated in apcmcr zebrafish embryos at 80hpf as examined by whole mount in situ staining (Figure 7A) and verified by RT-PCR (Figure S5A). Importantly, knockdown of Lsd1 and Corest rescued the expression of intestinal rdh1l in apcmcr embryos (Figure 7B and Figure S7B-C). Furthermore, rescue of rdh1l expression was accompanied by restoration of intestinal differentiation (fabp2), but not retinal differentiation (irbp) (Figure 7C). Treatment of apcmcr zebrafish embryos with pargyline also rescued expression of rdh1l (Figure 7B) and fabp2 (Figure 7C), suggesting that the catalytic activity of Lsd1 is essential for this regulation in zebrafish. These data suggest that Lsd1 and Corest are required to maintain the repression of rdh1l, resulting undifferentiated status of the intestines of apcmcr zebrafish embryos.

Figure 7. Lsd1 represses rdh1l expression and intestinal differentiation, and a model for retinoic acid regulation of DNA methylation dynamics.

Figure 7

Open in a new tab

(A) Whole mount in situ staining for lsd1 and corest in the intestine (arrows) of apcmcr and apcwt embryos (72hpf). (B) Fold change in rdh1l expression compared to 28S levels in apcwt and apcmcr embryos injected with control/lsd1/corest Mo or treated with pargyline (to inhibit Lsd1 activity). Error bars indicate +/− SD. Lsd1 and Corest morpholino knockdown efficiency is shown in Figure S7A-B. (C) Whole mount in situ staining for fabp2 and irbp in apcmcr and apcwt embryos injected with control/lsd1/corest Mo or treated with pargyline (3mM). (D-E) Model of APC regulation of intestinal fating via retinoic acid and demethylase. APC promotes RA production by directly negatively regulating CtBP1 levels in a proteasome-dependent fashion. APC also inhibits the transcription of LSD1, CoREST, LEF1 and TLE3. LEF1 binds to the RDH promoter and recruits TLE3 (Groucho2)/CtBP1/LSD1 repressors which silence RDH expression. Retinoic acid negatively regulates demethylase components by inhibiting Pou5f1 and Cebpβ. Furthermore, regulation of demethylase components by APC is independent of β-catenin. Demethylase promotes the demethylation of key fate regulators (like aldh1a2, hoxa13a, evx1) and proliferative genes (like cyclind1 and pitx2). Fate regulators like aldh1a2 possibly help in maintaining a progenitor cell population. See also Figure S7.

Discussion

Genomic hypomethylation was proposed by Holliday as an oncogenic mechanism nearly 25 years ago (Holliday and Jeggo, 1985). Since then many studies have focused on DNA methyltransferases and the paradoxical hypermethylation of tumor suppressor genes that occurs in the background of genome-wide hypomethylation. Mechanisms explaining the underlying hypomethylation, its contribution to tumor initiation and progression and its relationship to genetic events are largely unknown. Here, we present a model linking APC loss - a key genetic determinant of colon adenoma development - with the misregulation of DNA methylation dynamics through activation of a DNA demethylase system. This dynamic regulation appears critical to the proper fating of intestinal cells in zebrafish and targets key regulators of cell maintenance and differentiation. Mechanistically, loss of APC results in suppression of retinoic acid biosynthesis due to repression of RDH by aberrant upregulation of a transcriptional complex including LEF1, Groucho2/TLE3, CtBP1, LSD1 and Corest. Downstream of retinoic acid loss the transcriptional regulators Pou5f1 and Cebpβ appear required for activation of the demethylase components (Figure 7D).

Recent studies have confirmed the involvement of AID in DNA demethylation during reprogramming of human cells (Bhutani et al., 2009) and in the erasure of DNA methylation from mouse primordial germ cells (Popp et al., 2010). Our data offers a new mechanistic explanation for previous studies implicating DNA hypomethylation in cancer development (Goelz et al., 1985). Other models have considered the possibility that genome-wide demethylation during tumorigenesis results from inability of DNA methyltransferases to maintain normal DNA methylation patterns in highly proliferative cells. Interestingly, we observed a loss of dnmt1 in parallel with upregulation of the demethylase system in apc mutant zebrafish embryos. However, unlike knock down of the demethylase components, injection of dnmt1 into apc mutants failed to restore intestinal differentiation and methylation on specific genes. This finding suggests a major role for the misregulation of the demethylase, rather than Dnmt1 loss, in maintaining the progenitor-like state of intestinal cell present in apc mutants. Consistent with this model, knockout of Apobec-1, a mammalian 5meC deaminase highly related to AID (Morgan et al., 2004), reduced adenoma formation in the ApcMin mouse model (Blanc et al., 2007).

Our work also provides insight into an apparent contradiction brought about by previous studies indicating that genetic loss or pharmacologic inhibition of DNMT1 appears protective in some circumstances (Eads et al., 2002; Laird et al., 1995). It is important to note that the demethylase system works through a process that first involves deamination of 5meC to thymine and may, therefore, enhance C to T transitions. It is possible that loss of methylation due to DNMT1 inhibition lowers the yield of deamination products (5meC>T transitions) by removing the substrate, 5meC. This could in turn slow adenoma progression by limiting this mode of mutation. Interestingly, Laird et al. speculated reduced mutation rates in the ApcMin mice treated with 5-Aza-cytidine as this treatment effectively prevented adenoma initiation but failed to affect established polyps (Laird et al., 1995). Similarly, knockout of MBD4 increases intestinal tumorigenesis in parallel with increased rates of C to T transitions (Millar et al., 2002).

Although the transcriptional effects of retinoic acid in controlling cell patterning, fate and differentiation are well documented, the regulation of RA production as a determinant of cell fate decisions is less well understood. Production of retinoic acid first requires converting dietary retinol (vitamin A) into retinoic acid, a process that occurs via two enzymatic steps (Duester, 2000): the conversion of retinol into retinal by alcohol dehydrogenases (ADH) and short chain dehydrogenases (SDR), followed by the conversion of retinal into retinoic acid via aldehyde dehydrogenases (ALDH) (Duester, 2000). Studies in mice and chickens indicated that RALDHs, rather than RDHs, represent the primary point for regulating tissue-specific production of retinoic acid (Duester et al., 2003). Recent studies in zebrafish, however, have demonstrated that Rdh1 and Rdh1l are essential for intestinal differentiation, and the development of the pectoral fin, jaw, eye, and exocrine pancreas (Nadauld et al., 2004; Nadauld et al., 2005). Our findings here provide a model suggesting a dynamic inverse relationship between RDHs and ALDHs in regulating RA production in intestinal cells. In this model, high levels of ALDH appear to poise cells for RA production upon generation of RAL by RDH. This in turn initiates a program of differentiation that relies in part on the suppression of the demethylase system (Figure 7E).

Interestingly, we show that multiple transcriptional repressors, namely LEF1, Groucho2/TLE3, CtBP1, LSD1 and CoREST work together to repress the production of retinoic acid by direct binding to, and repression of, the rdh1 promoter. Further, the activity of LSD1 is required for maintaining repression of the RDH promoter in the presence of APC mutation. Consistent with the model, knockdown of these proteins restored rdh1l expression and intestinal differentiation to apc mutants. Furthermore, our data indicate that APC regulation of RDHs and the demethylase system occurs independently of β-catenin thereby revealing a novel role for LEF1, which is known to be upregulated in sporadic colon carcinomas (Hovanes et al., 2001), but thought previously to act primarily through an interaction with β-catenin. Consistent with this idea, human FAP adenoma tissues also showed increased levels of all of these factors. Taken together, these data support a model wherein mutations in APC cause upregulation of LEF1, Groucho2/TLE3, CtBP1, LSD1 and CoREST and that this complex serves to repress intestinal differentiation by directly targeting RDHs.

Our results suggest a mechanism for a rapid and robust change in DNA methylation dynamics during development, mediated by retinoic acid. We observe a marked upregulation of demethylase components in RA deficient embryos, and a marked downregulation of demethylase components following treatment with retinoic acid. Notably, we demonstrate that retinoic acid antagonizes two key direct activators of demethylase components, Pou5f1 and Cebpβ, leading to demethylase downregulation. These results raise the possibility that part of the basis for the developmental plasticity and maintenance of progenitor cells is the DNA methylation dynamics provided by the demethylase system – a dynamic loss following exposure to retinoic acid and differentiation, possibly to favor the imposition of DNA methylation patterns at the promoters of key developmental regulators, which may help commit cells to differentiation and help maintain that commitment. Further studies will determine the potential roles for other candidate demethylase components downstream of APC.

In summary, we have shown that DNA cytosine demethylases are upregulated in APC mutant human and zebrafish tissue, and are responsible for demethylation of genes which are important for maintaining a progenitor cell population in APC mutant tissues. These findings offer a mechanistic model implicating active DNA demethylation in contributing to cell fating defects following loss of APC.

Experimental Procedures

Zebrafish care and Injections

Wild type (Tu strain), apc heterozygous adults (kind gift of Drs. Anna Pavlina-Haramis and Hans Clevers) and gutGFP adults (Tg(XlEef1a1:GFP)s854; obtained from ZIRC) maintained in Z-Mods at 28 degrees in a 14:10 hour light cycle. For all whole mount in situ experiments embryos were raised in 0.003% phelthiourea (PTU) treated embryo water to inhibit pigment formation. Morpholinos (Gene-tools), DNA and RNA were injected at 1-cell stage. Sequences are in supplementary methods.

Whole mount in situ hybridization

Whole mount in situ hybridizations were performed as described earlier (Nadauld et al., 2004). Riboprobes for the indicated genes were made using digoxigenin labeled UTPs (Roche). Genes were cloned in pCR-II-TOPO vector (Invitrogen).

Cell culture and Transfection

SW480, DLD1 and HT29 cells were purchased from ATCC and grown as per guidelines of manufacturer. Transfection experiments were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s recommendations. Following transfection, cells were incubated for 24hrs (DNA), 48hrs (IP) or 72hrs (siRNA) and then harvested for western blotting analysis, immunoprecipitation or RT-PCR.

RT-PCR

RNA from whole embryos or human adenoma tissues was isolated using Trizol (Invitrogen). cDNA library was prepared using Superscript III kit (Invitrogen). RT-PCR was performed on a Roche light cycler. Primer information is available upon request.

Western blotting

Protein extracts were made and western blotting performed as described earlier (Rai et al., 2008).

Drug Treatments

Embryos were treated with all-trans retinoic acid (1μM) or DMSO (vehicle) every day starting at 24hpf for 45min. NS-398 treatment was similar except at 10μM. For pargyline treatment, drug (3mM or water control) was left in embryo water starting at 75% epiboly and new drug added everyday. Cells were treated with pargyline at 3mM or all-trans retinoic acid (or DMSO) at 1μM each day for two days and RNA was isolated.

MeDIP (Methylated DNA immunoprecipitation) and Array hybridization

This procedure was previously described (Weber et al., 2007). Briefly, Dynabeads (conjugated with sheep anti-mouse secondary antibody, Invitrogen) were incubated with 10ug of 5MeC antibody (Eurogentec) for 2hrs, washed, and then incubated with 4μg of sonicated DNA (300-1000bp fragments) overnight. Immunoprecipitated DNA was used for qPCR or amplified using whole genome amplification kit (Sigma Aldrich). Following amplification, 2μg of amplified DNA was labeled with Cy5, and 2μg of input DNA was labeled with Cy3 (Bio labs). Samples were competitively hybridized to 244K oligonucleotide promoter arrays (Agilent Inc).

Chromatin Immunoprecipitation

ChIP was performed as described earlier (Rai et al., 2008). ChIP for Cebpβ and Pou5f1 were performed in wild type (Tu) whole embryos. V5-zfCEBPβ (along with pou5f1 morpholino, 80pg) or V5-zfPou5f1 (both cloned in pcDNA3.1/nV5 DEST; Invitrogen) was injected at one cell stage and embryos collected at 48hpf. ChIP was performed using V5 antibody (Abcam). ChIP for Lef1 was performed using rabbit-anti-zfLef1 antibody in extracts made from apcwt and apcmcr whole embryos collected at 72hpf. ChIP in SW-480 cells were performed using antibodies α-CtBP (Santa Cruz; sc-11390), α-LSD-1 (Abcam; ab-17721) and rabbit IgG (control).

Co-Immunoprecipation

Constructs expressing tagged proteins were transfected in SW480 cells and protein harvested in 1X IPH buffer 48 hours after transfection. Immunoprecipitation was performed as described earlier (Rai et al., 2008) using rabbit-α-V5 antibody (Abcam) or α-Myc monoclonal antibody. Western blot was probed using mouse-α-V5 (Invitrogen), mouse-α-GFP (BD Biosciences), rabbit-α-myc (Santa Cruz) and rabbit-α-V5 (Abcam) antibodies.

Supplementary Material

01
02
03

Acknowledgements

We thank Smitha James and Adam Karpf for LC-MS assay, Kornelia Edes, Chad Zavala, Jennifer Swift, Abby Evans and Erek Winett for technical help and Richard Dorsky for reagents. This work is supported by grants from the National Cancer Institute (CA073992 and CA96934) awarded to D.A.J., the National Institute of Child Health and Development (HD058506 to B.R.C. and D.A.J), the Howard Hughes Medical Institute (to B.R.C), and the Huntsman Cancer Foundation. The work was also supported by access to technical cores supported by a Cancer Center Support Grant (CA042014).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Baek SH, Kioussi C, Briata P, Wang D, Nguyen HD, Ohgi KA, Glass CK, Wynshaw-Boris A, Rose DW, Rosenfeld MG. Regulated subset of G1 growth-control genes in response to derepression by the Wnt pathway. Proc Natl Acad Sci U S A. 2003;100:3245–3250. doi: 10.1073/pnas.0330217100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Begemann G, Schilling TF, Rauch GJ, Geisler R, Ingham PW. The zebrafish neckless mutation reveals a requirement for raldh2 in mesodermal signals that pattern the hindbrain. Development. 2001;128:3081–3094. doi: 10.1242/dev.128.16.3081. [DOI] [PubMed] [Google Scholar]
  3. Bhutani N, Brady JJ, Damian M, Sacco A, Corbel SY, Blau HM. Reprogramming towards pluripotency requires AID-dependent DNA demethylation. Nature. 2009;463:1042–1047. doi: 10.1038/nature08752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bienz M, Clevers H. Linking colorectal cancer to Wnt signaling. Cell. 2000;103:311–320. doi: 10.1016/s0092-8674(00)00122-7. [DOI] [PubMed] [Google Scholar]
  5. Blanc V, Henderson JO, Newberry RD, Xie Y, Cho SJ, Newberry EP, Kennedy S, Rubin DC, Wang HL, Luo J, et al. Deletion of the AU-rich RNA binding protein Apobec-1 reduces intestinal tumor burden in Apc(min) mice. Cancer Res. 2007;67:8565–8573. doi: 10.1158/0008-5472.CAN-07-1593. [DOI] [PubMed] [Google Scholar]
  6. Dobbs-McAuliffe B, Zhao Q, Linney E. Feedback mechanisms regulate retinoic acid production and degradation in the zebrafish embryo. Mech Dev. 2004;121:339–350. doi: 10.1016/j.mod.2004.02.008. [DOI] [PubMed] [Google Scholar]
  7. Duester G. Families of retinoid dehydrogenases regulating vitamin A function: production of visual pigment and retinoic acid. Eur J Biochem. 2000;267:4315–4324. doi: 10.1046/j.1432-1327.2000.01497.x. [DOI] [PubMed] [Google Scholar]
  8. Duester G, Mic FA, Molotkov A. Cytosolic retinoid dehydrogenases govern ubiquitous metabolism of retinol to retinaldehyde followed by tissue-specific metabolism to retinoic acid. Chem Biol Interact. 2003;143-144:201–210. doi: 10.1016/s0009-2797(02)00204-1. [DOI] [PubMed] [Google Scholar]
  9. Eads CA, Nickel AE, Laird PW. Complete genetic suppression of polyp formation and reduction of CpG-island hypermethylation in Apc(Min/+) Dnmt1-hypomorphic Mice. Cancer Res. 2002;62:1296–1299. [PubMed] [Google Scholar]
  10. Eden A, Gaudet F, Waghmare A, Jaenisch R. Chromosomal instability and tumors promoted by DNA hypomethylation. Science. 2003;300:455. doi: 10.1126/science.1083557. [DOI] [PubMed] [Google Scholar]
  11. Eisinger AL, Nadauld LD, Shelton DN, Peterson PW, Phelps RA, Chidester S, Stafforini DM, Prescott SM, Jones DA. The adenomatous polyposis coli tumor suppressor gene regulates expression of cyclooxygenase-2 by a mechanism that involves retinoic acid. J Biol Chem. 2006;281:20474–20482. doi: 10.1074/jbc.M602859200. [DOI] [PubMed] [Google Scholar]
  12. Eisinger AL, Nadauld LD, Shelton DN, Prescott SM, Stafforini DM, Jones DA. Retinoic acid inhibits beta-catenin through suppression of Cox-2: a role for truncated adenomatous polyposis coli. J Biol Chem. 2007;282:29394–29400. doi: 10.1074/jbc.M609768200. [DOI] [PubMed] [Google Scholar]
  13. Gaudet F, Hodgson JG, Eden A, Jackson-Grusby L, Dausman J, Gray JW, Leonhardt H, Jaenisch R. Induction of tumors in mice by genomic hypomethylation. Science. 2003;300:489–492. doi: 10.1126/science.1083558. [DOI] [PubMed] [Google Scholar]
  14. Goelz SE, Vogelstein B, Hamilton SR, Feinberg AP. Hypomethylation of DNA from benign and malignant human colon neoplasms. Science. 1985;228:187–190. doi: 10.1126/science.2579435. [DOI] [PubMed] [Google Scholar]
  15. Guimond J, Devost D, Brodeur H, Mader S, Bhat PV. Characterization of the rat RALDH1 promoter. A functional CCAAT and octamer motif are critical for basal promoter activity. Biochim Biophys Acta. 2002;1579:81–91. doi: 10.1016/s0167-4781(02)00510-9. [DOI] [PubMed] [Google Scholar]
  16. Holliday R, Jeggo PA. Mechanisms for changing gene expression and their possible relationship to carcinogenesis. Cancer Surv. 1985;4:557–581. [PubMed] [Google Scholar]
  17. Hovanes K, Li TW, Munguia JE, Truong T, Milovanovic T, Lawrence Marsh J, Holcombe RF, Waterman ML. Beta-catenin-sensitive isoforms of lymphoid enhancer factor-1 are selectively expressed in colon cancer. Nat Genet. 2001;28:53–57. doi: 10.1038/ng0501-53. [DOI] [PubMed] [Google Scholar]
  18. Huang EH, Hynes MJ, Zhang T, Ginestier C, Dontu G, Appelman H, Fields JZ, Wicha MS, Boman BM. Aldehyde dehydrogenase 1 is a marker for normal and malignant human colonic stem cells (SC) and tracks SC overpopulation during colon tumorigenesis. Cancer Res. 2009;69:3382–3389. doi: 10.1158/0008-5472.CAN-08-4418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hurlstone AF, Haramis AP, Wienholds E, Begthel H, Korving J, Van Eeden F, Cuppen E, Zivkovic D, Plasterk RH, Clevers H. The Wnt/beta-catenin pathway regulates cardiac valve formation. Nature. 2003;425:633–637. doi: 10.1038/nature02028. [DOI] [PubMed] [Google Scholar]
  20. Jette C, Peterson PW, Sandoval IT, Manos EJ, Hadley E, Ireland CM, Jones DA. The tumor suppressor adenomatous polyposis coli and caudal related homeodomain protein regulate expression of retinol dehydrogenase L. J Biol Chem. 2004;279:34397–34405. doi: 10.1074/jbc.M314021200. [DOI] [PubMed] [Google Scholar]
  21. Laird PW, Jackson-Grusby L, Fazeli A, Dickinson SL, Jung WE, Li E, Weinberg RA, Jaenisch R. Suppression of intestinal neoplasia by DNA hypomethylation. Cell. 1995;81:197–205. doi: 10.1016/0092-8674(95)90329-1. [DOI] [PubMed] [Google Scholar]
  22. Millar CB, Guy J, Sansom OJ, Selfridge J, MacDougall E, Hendrich B, Keightley PD, Bishop SM, Clarke AR, Bird A. Enhanced CpG mutability and tumorigenesis in MBD4-deficient mice. Science. 2002;297:403–405. doi: 10.1126/science.1073354. [DOI] [PubMed] [Google Scholar]
  23. Morgan HD, Dean W, Coker HA, Reik W, Petersen-Mahrt SK. Activation-induced cytidine deaminase deaminates 5-methylcytosine in DNA and is expressed in pluripotent tissues: implications for epigenetic reprogramming. J Biol Chem. 2004;279:52353–52360. doi: 10.1074/jbc.M407695200. [DOI] [PubMed] [Google Scholar]
  24. Mudumana SP, Wan H, Singh M, Korzh V, Gong Z. Expression analyses of zebrafish transferrin, ifabp, and elastaseB mRNAs as differentiation markers for the three major endodermal organs: liver, intestine, and exocrine pancreas. Dev Dyn. 2004;230:165–173. doi: 10.1002/dvdy.20032. [DOI] [PubMed] [Google Scholar]
  25. Nadauld LD, Phelps R, Moore BC, Eisinger A, Sandoval IT, Chidester S, Peterson PW, Manos EJ, Sklow B, Burt RW, et al. Adenomatous polyposis coli control of C-terminal binding protein-1 stability regulates expression of intestinal retinol dehydrogenases. J Biol Chem. 2006;281:37828–37835. doi: 10.1074/jbc.M602119200. [DOI] [PubMed] [Google Scholar]
  26. Nadauld LD, Sandoval IT, Chidester S, Yost HJ, Jones DA. Adenomatous polyposis coli control of retinoic acid biosynthesis is critical for zebrafish intestinal development and differentiation. J Biol Chem. 2004;279:51581–51589. doi: 10.1074/jbc.M408830200. [DOI] [PubMed] [Google Scholar]
  27. Nadauld LD, Shelton DN, Chidester S, Yost HJ, Jones DA. The zebrafish retinol dehydrogenase, rdh1l, is essential for intestinal development and is regulated by the tumor suppressor adenomatous polyposis coli. J Biol Chem. 2005;280:30490–30495. doi: 10.1074/jbc.M504973200. [DOI] [PubMed] [Google Scholar]
  28. Niederreither K, McCaffery P, Drager UC, Chambon P, Dolle P. Restricted expression and retinoic acid-induced downregulation of the retinaldehyde dehydrogenase type 2 (RALDH-2) gene during mouse development. Mech Dev. 1997;62:67–78. doi: 10.1016/s0925-4773(96)00653-3. [DOI] [PubMed] [Google Scholar]
  29. Parvin MS, Okuyama N, Inoue F, Islam ME, Kawakami A, Takeda H, Yamasu K. Autoregulatory loop and retinoic acid repression regulate pou2/pou5f1 gene expression in the zebrafish embryonic brain. Dev Dyn. 2008;237:1373–1388. doi: 10.1002/dvdy.21539. [DOI] [PubMed] [Google Scholar]
  30. Phelps RA, Broadbent TJ, Stafforini DM, Jones DA. New perspectives on APC control of cell fate and proliferation in colorectal cancer. Cell Cycle. 2009a;8:2549–2556. doi: 10.4161/cc.8.16.9278. [DOI] [PubMed] [Google Scholar]
  31. Phelps RA, Chidester S, Dehghanizadeh S, Phelps J, Sandoval IT, Rai K, Broadbent T, Sarkar S, Burt RW, Jones DA. A two-step model for colon adenoma initiation and progression caused by APC loss. Cell. 2009b;137:623–634. doi: 10.1016/j.cell.2009.02.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Popp C, Dean W, Feng S, Cokus SJ, Andrews S, Pellegrini M, Jacobsen SE, Reik W. Genome-wide erasure of DNA methylation in mouse primordial germ cells is affected by AID deficiency. Nature. 2010;463:1101–1105. doi: 10.1038/nature08829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Rai K, Huggins IJ, James SR, Karpf AR, Jones DA, Cairns BR. DNA demethylation in zebrafish involves the coupling of a deaminase, a glycosylase, and gadd45. Cell. 2008;135:1201–1212. doi: 10.1016/j.cell.2008.11.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Russo JE, Hauguitz D, Hilton J. Inhibition of mouse cytosolic aldehyde dehydrogenase by 4-(diethylamino)benzaldehyde. Biochem Pharmacol. 1988;37:1639–1642. doi: 10.1016/0006-2952(88)90030-5. [DOI] [PubMed] [Google Scholar]
  35. Shi Y, Lan F, Matson C, Mulligan P, Whetstine JR, Cole PA, Casero RA, Shi Y. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell. 2004;119:941–953. doi: 10.1016/j.cell.2004.12.012. [DOI] [PubMed] [Google Scholar]
  36. Shi Y, Sawada J, Sui G, Affar el B, Whetstine JR, Lan F, Ogawa H, Luke MP, Nakatani Y, Shi Y. Coordinated histone modifications mediated by a CtBP co-repressor complex. Nature. 2003;422:735–738. doi: 10.1038/nature01550. [DOI] [PubMed] [Google Scholar]
  37. Sparks AB, Morin PJ, Vogelstein B, Kinzler KW. Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res. 1998;58:1130–1134. [PubMed] [Google Scholar]
  38. Weber M, Hellmann I, Stadler MB, Ramos L, Paabo S, Rebhan M, Schubeler D. Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome. Nat Genet. 2007;39:457–466. doi: 10.1038/ng1990. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS234412-supplement-01.doc (10MB, doc)
02
NIHMS234412-supplement-02.xls (3.6MB, xls)
03
NIHMS234412-supplement-03.xls (9.2MB, xls)

RESOURCES