Mutations in the adenomatous polyposis coli (APC) tumor suppressor gene seem to underlie the initiation of many colorectal carcinomas. Loss of APC function results in accumulation of β-catenin and activation of β-catenin/TCF–dependent transcription. Recent studies have implicated APC in controlling retinoic acid biosynthesis during normal intestinal development through a WNT-independent mechanism. Paradoxically, however, previous studies found that dietary supplementation of ApcMIN mice with retinoic acid failed to abrogate adenoma formation. While investigating the above finding, we found that expression of CYP26A1, a major retinoic acid catabolic enzyme, was up-regulated in ApcMIN mouse adenomas, human FAP adenomas, human sporadic colon carcinomas, and in the intestine of apcmcr mutant zebrafish embryos. Mechanistically, cyp26a1 induction following apc mutation is dependent on WNT signaling as antisense morpholino knockdown of tcf4 or injection of a dnLEF construct into apcmcr mutant zebrafish suppressed expression of cyp26a1 along with known WNT target genes. In addition, injection of stabilized β-catenin or dnGSK3β into wild-type embryos induced cyp26a1 expression. Genetic knockdown or pharmacologic inhibition of cyp26a1 in apcmcr mutant zebrafish embryos rescued gut differentiation defects such as expression of intestinal fatty acid–binding protein and pancreatic trypsin. These findings support a novel role for APC in balancing retinoic acid biosynthesis and catabolism through WNT-independent and WNT-dependent mechanisms. (Cancer Res 2006; 66(15): 7571-7)

Retinoic acid is derived from dietary retinol (vitamin A) and plays well-characterized roles in determining cell fates and promoting epithelial differentiation (1, 2). The actions of retinoic acid are, therefore, tightly regulated in both temporal and tissue-specific manners. To achieve this tight regulation, the production of retinoic acid is often carefully balanced by its degradation. Alcohol dehydrogenases and short-chain alcohol dehydrogenases initiate the biosynthesis of retinoic acid by converting retinol into retinaldehyde. Further oxidation of retinaldehyde by retinaldehyde dehydrogenases completes the synthesis of retinoic acid (3). The cellular effects of retinoic acid are mediated through the binding of retinoic acid to specific nuclear receptors (retinoic acid receptors and retinoid X receptors) that serve to activate or repress specific genes required for cell fate and differentiation decisions (4).

The production of retinoic acid in tissues is opposed by its degradation through the actions of specific cytochrome P450 enzymes. CYP26A1 has been implicated in restricting the diffusion of retinoic acid in several model systems, including chick, mouse, and zebrafish (58). For example, Cyp26 expression seems to permit cell proliferation and expansion in the leading edge of the developing mouse limb bud by counteracting the production of retinoic acid by the proximal trunk. This allows for coordinately regulated differentiation and proliferation (9). Similarly, Cyp26a1 is necessary for establishment of the anterior-posterior axis. Mice homozygous for mutant Cyp26a1 exhibit truncations of the posterior axis, including the hindgut, and die in utero (10). Genetic suppression of this lethality and posterior axis truncation was recently shown by Niederreither et al. (8, 11) through a cross of Cyp26a1−/− and Raldh2+/− mice, which led to a reduction of retinoic acid production. The regulation of CYP26A1 expression is not fully understood but is likely due to both basal level regulation of the CYP26A1 promoter, as well as an inducible aspect in the form of a retinoic acid response element (12).

The differentiation properties of retinoic acid have made it attractive as a potential therapeutic for the treatment of a variety of cancers. Its usefulness, however, seems limited in this context because a number of studies suggest that CYP26-mediated destruction of retinoic acid may reduce its efficacy in vivo. For example, Lopez-Berestein et al. (13, 14) have found that retinoic acid treatment of acute promyelocytic leukemia patients leads to increased metabolism and clearing of retinoic acid by CYP26A1 in intestinal, endothelial, and liver tissues. In addition, colon carcinomas and colon carcinoma cell lines have proven unresponsive to retinoic acid treatment (1517). Finally, recent literature has emerged suggesting that mice bearing mutations in the Apc tumor suppressor gene do not show a suppression of adenoma formation when fed with a daily dietary supplement of all-trans retinoic acid (ATRA; ref. 18).

The lack of responsiveness of ApcMIN mice to retinoic acid is surprising given recent evidence showing that adenomatous polyposis coli (APC) controls the production of retinoic acid and that this control is essential for normal intestinal differentiation in zebrafish. Specifically, Jette et al. (19) showed that a human retinol dehydrogenase (RDH), DHRS9, is lost in colon tumors compared with normal colon tissue. Moreover, this enzyme was induced upon reintroduction of wild-type APC into colon cancer cell lines, thereby linking APC function with retinoic acid biosynthesis for the first time. Extending these findings, Nadauld et al. (20) showed that knockdown of rdh1, rdh1l, or APC by antisense morpholino oligonucleotide injection led to a loss of terminal differentiation in the developing zebrafish gut, as well as other retinoic acid–deficient phenotypes, such as pericardial edema, lack of fin and jaw development, and loss of pancreatic differentiation. Treatment of apc-deficient zebrafish embryos (apcmcr) with ATRA partially rescued the expression of the gut differentiation markers, such as intestinal fatty acid–binding protein (i-fabp) and pancreatic trypsin.

Interestingly, APC control of retinoic acid production seems independent of the canonical model for APC function (19). In the canonical model, APC negatively regulates WNT signaling by helping to target the transcriptional coactivator β-catenin for ubiquitin-mediated proteolysis. Mutations in APC, therefore, lead to accumulation of β-catenin and subsequent activation of β-catenin/LEF transcriptional complexes. These complexes are thought to stimulate intestinal cell proliferation through the direct targeting of proliferation genes like c-myc (21) and BMP4 (22). Although our previous work showed significant rescue of i-fabp in apcmcr mutant fish by retinoic acid treatment, this rescue was incomplete. This observation, combined with the lack of effect of retinoic acid in the ApcMIN mouse, prompted us to examine the responsiveness of APC-deficient tissues to exogenous retinoic acid. Here, we show that murine, human, and zebrafish tissues harboring mutated APC exhibit increased levels of CYP26A1 and that this increased expression results from activation of WNT signaling. Inhibition of cyp26a1 enzymatic activity or morpholino knockdown of cyp26a1 restored intestinal differentiation in apcmcr mutant zebrafish. Based on these findings, we propose that APC coordinately regulates retinoic acid production and catabolism through independent pathways.

Zebrafish stocks and embryo culture. Wild-type and apcmcr mutant Danio rerio (zebrafish) were maintained on a 14:10-hour light/dark cycle. Fertilized embryos were collected following natural spawning and allowed to develop at 28.5°C. Control and experimental embryos were raised in 0.003% phenylthiourea to inhibit pigment formation. Drs. Anna-Pavlina Haramis and Hans Clevers (Hubrecht Laboratorium, NIOB/KNAW, Center for Biomedical Genetics, CT, Utrecht, The Netherlands) kindly provided apcmcr mutant zebrafish used in this study.

Whole-mount in situ hybridizations. Zebrafish embryos were dechorionated, fixed in sucrose-buffered 4% paraformaldehyde, rinsed in PBS, dehydrated in methanol, and stored at −20°C. Digoxigenin-labeled riboprobes for cyp26a1 and i-fabp were generated as described previously (23). Both riboprobes were generated by linearization of pCRII (Invitrogen, Carlsbad, CA) containing cyp26a1 or i-fabp cDNA followed by in vitro transcription with SP6 RNA polymerase (Roche Applied Science, Indianapolis, IN). Whole-mount in situ hybridizations were carried out as described previously (23). Embryos were fixed in 4% paraformaldehyde for 1 hour at room temperature, cleared in methanol overnight, and then cleared in 70% glycerol and PBS and photographed using an Olympus DP12 digital camera.

Quantitative reverse transcription-PCR. Single-stranded cDNA was synthesized from 2 to 3 μg of total RNA using Superscript III (Invitrogen). PCR was done with 30 ng cDNA using the Roche LightCycler instrument and software, version 3.5 (Roche Diagnostics, Indianapolis, IN). Primers were as follows: z.f.cyp26a1, forward, 5′-ACGAGCAAGAACTGGTGGAA-3′; reverse, 5′-GGGCGTCTTTGTATTTCTGCT-3′; z.f.mmp13, forward, 5′-CTTCTGGCGCAGTTATCCTC-3′; reverse, 5′-TGTACTGGAAGGCTGCAGTG-3′, m.m.Cyp26a1, forward, 5′-GATAAAGAGCAAGGGCTTACT-3′, reverse, 5′-AAATAACATTCCAGCCCTTGG-3′, h.s.CYP26A1, forward, 5′-GCAGGAAATACGGCTTCATC-3′, reverse, 5′-CCCGCATAATCACCTTCTTG-3′. PCR was done in triplicate using the LightCycler FastStart DNA Master SYBR Green I kit (Roche Applied Science). PCR conditions were as follows: 40 cycles of amplification with 20 seconds of denaturation at 95°C and 20 seconds of annealing at 60°C for cyp26a1 and 35 cycles of amplification with 20 seconds of denaturation at 95°C and 20 seconds of annealing at 55°C for all others. A template-free negative control was included in each experiment. Statistical analyses were done using GraphPad Prism (version 4.03).

Morpholino microinjections. Morpholino oligonucleotides were obtained from Gene Tools LLC (Philomath, OR). The apc splice-blocking morpholino (5′-TAGCATACTCTACCTGTGCTCTTCG-3′), z.f.tcf4 splice-blocking morpholino (5′-CTTATTTGTCACTTACCTCGGAATC-3′), and z.f.cyp26a1 splice-blocking morpholino (5′-CCCTCAAACCCTGCCGATCAAAAAT-3′) were solubilized to 1 mmol/L in 1× Danieau buffer. For microinjections, 1.5 nL of 0.5 mmol/L apc antisense morpholino, 1 nL of 0.25 mmol/L tcf4 morpholino, or 1 nL of 0.0625 mmol/L cyp26a1 morpholino was injected into wild-type or apcmcr mutant embryos at the one- to two-cell stages (23).

RNA microinjections. For RNA injection experiments, full-length human β-catenin or dnGSK3β mRNA was synthesized from a linearized pCRII construct using mMessage mMachine (Ambion, Austin, TX) according to the protocol of the manufacturer. Full-length β-catenin or dnGSK3β mRNAs were injected into embryos at the one-cell stage.

Retinoic acid and drug treatments. To rescue apcmcr mutants by application of retinoic acid, the embryos were incubated in 900 nmol/L ATRA in Me2SO at 70% epiboly for 45 minutes. in the dark. The embryos were then washed in embryo water thrice. Retinoic acid treatments (100 nmol/L) were repeated every 24 hours for 45 minutes. Control embryos were treated over these periods with an equal volume of Me2SO (24). Retinoic acid metabolism blocking agent (RAMBA) treatments were the same, with the exception of 2- and 1-day applications, where RAMBA (vn/14-1) was applied at 900 nmol/L at 48 hpf and 100 nmol/L at 72 hpf for 2-day application, or 900 nmol/L at 72 hpf for 1-day application. RAMBA was synthesized as previously reported (23, 25).

Histologic analyses. Embryos were mounted in 3% low-melting-point agarose and then processed and embedded using the Polysciences Immuno-Bed kit. Five-micrometer sections were cut using a Leica 2055 microtome and counterstained with eosin. Sections were analyzed using a Zeiss Axiovert100 microscope and pictures were taken using an Olympus Magnafire color camera.

Note: Per convention, we refer to human genes in using all upper case letters (CYP26A1), murine genes with only the first letter in upper case (Cyp26a1), and zebrafish gene in small case letters (cyp26a1).

Cyp26a1 is induced in ApcMIN mouse adenomas and human FAP adenomas and human sporadic carcinomas. Mutations in APC are implicated in the progression of colon cancer. Our previous findings that human colon tumors are deficient in retinoic acid biosynthesis prompted us to examine whether CYP26A1, a major retinoic acid catabolic enzyme, was also up-regulated in mice and humans with APC mutations. To evaluate this, both wild-type littermates and ApcMIN heterozygotes were grown to 12 to 16 weeks and sacrificed. Adenomas and normal tissue of the colon were dissected out and total RNA was isolated. Quantitative reverse transcription-PCR (RT-PCR) indicated that Cyp26a1 was significantly elevated in murine colon adenomas compared with normal tissues (Fig. 1A). We then looked at CYP26A1 expression in human FAP colon adenomas. As seen in mice, RT-PCR analysis for CYP26A1 expression from patient-matched uninvolved and FAP adenomas showed that all samples tested had elevated CYP26A1 expression (Fig. 1B). Extending these findings, we also found that sporadic human colon tumor samples expressed significantly higher levels of CYP26A1 compared with normal tissues (Fig. 1C). Up-regulation of CYP26A1 in both mice and humans with APC mutations supports the idea that dysregulation of CYP26A1 plays a role in the development of colorectal carcinoma cancer.

Figure 1.

CYP26A1 is high in murine and human adenomas and human sporadic carcinoma samples. A, quantitative RT-PCR of Cyp26a1 expression in ApcMIN mouse colon adenomas. Levels of Cyp26a1 were normalized in comparison to levels of 18S ribosomal subunit RNA. B, CYP26A1 expression in matched human FAP adenomas. Values are fold-change in adenomas relative to patient-matched uninvolved tissue. P1-5 refer to individual patients, while P1a-d refer to separate samples from one patient. Error bars refer to standard deviations (SD) of three measurements. C, CYP26A1 expression in human sporadic carcinoma samples normalized to β-actin.

Figure 1.

CYP26A1 is high in murine and human adenomas and human sporadic carcinoma samples. A, quantitative RT-PCR of Cyp26a1 expression in ApcMIN mouse colon adenomas. Levels of Cyp26a1 were normalized in comparison to levels of 18S ribosomal subunit RNA. B, CYP26A1 expression in matched human FAP adenomas. Values are fold-change in adenomas relative to patient-matched uninvolved tissue. P1-5 refer to individual patients, while P1a-d refer to separate samples from one patient. Error bars refer to standard deviations (SD) of three measurements. C, CYP26A1 expression in human sporadic carcinoma samples normalized to β-actin.

Close modal

Cyp26a1 is up-regulated in apc-deficient zebrafish. The above findings in murine and human adenomas suggested dysregulation of CYP26A1 as an early event following APC mutation. Recent studies have shown that homozygous apcmcr mutant and apc antisense knockdown zebrafish embryos harbor a number of developmental defects consistent with loss of retinoic acid, including pancreatic and intestinal differentiation defects (20, 23). In addition, adult zebrafish carrying heterozygous apc mutations develop spontaneous adenomas in the liver, pancreas, and intestine (26). We, therefore, examined the relationship between apc and cyp26a1 in vivo using apcmcr mutant and morphant zebrafish. Whole-mount in situ hybridization showed that cyp26a1 is expressed at very low levels in the brachial arches and neural retina at 80 hpf of wild-type embryos. In contrast, however, cyp26a1 expression is strongly up-regulated in the brachial arches, neural retina, and gut of apcmcr mutant and morphant embryos. Quantitative RT-PCR analysis revealed a 3- to 4-fold increase in cyp26a1 expression in either apcmcr mutant or apc morphant embryos compared with wild-type embryos (Fig. 2A and B). Histologic cross-sectioning showed that the gut-specific expression of cyp26a1 was confined to the epithelium throughout the intestinal tube (Fig. 2C). To further examine this divergence in expression from wild type, we analyzed expression levels in wild-type and apcmcr mutants over a developmental time course starting at 48 hpf when the gross morphology of homozygous apcmcr mutant embryos could be confidently distinguished from wild-type siblings. Expression of cyp26a1 was 1.8-fold higher in apcmcr mutants at 48 hpf and gradually increased to 3- to 4-fold at ∼80 hpf when apcmcr mutant embryos begin to die. This increase in expression was due to increasing cyp26a1 mRNA levels in mutants relative to β-actin, as well as decreasing expression of cyp26a1 in wild-type embryos (Fig. 2D).

Figure 2.

Cyp26a1 is up-regulated in apc-deficient zebrafish. A, whole-mount in situ hybridization was performed on 80 hpf fish and stained for cyp26a1 expression in wild type, apc morpholino-injected, and apcmcr mutant fish, indicating up-regulation of cyp26a1 in the neural retina, brachial arches, and gut. B, quantitative RT-PCR for cyp26a1 in wild-type, apc morphants and mutants at 80 hpf. Error bars refer to SD of three measurements. C, H&E sections of wild-type and apcmcr mutant fish stained for cyp26a1 expression in gut epithelium. D, quantitative RT-PCR for cyp26a1 in apcmcr mutant fish over a developmental time course.

Figure 2.

Cyp26a1 is up-regulated in apc-deficient zebrafish. A, whole-mount in situ hybridization was performed on 80 hpf fish and stained for cyp26a1 expression in wild type, apc morpholino-injected, and apcmcr mutant fish, indicating up-regulation of cyp26a1 in the neural retina, brachial arches, and gut. B, quantitative RT-PCR for cyp26a1 in wild-type, apc morphants and mutants at 80 hpf. Error bars refer to SD of three measurements. C, H&E sections of wild-type and apcmcr mutant fish stained for cyp26a1 expression in gut epithelium. D, quantitative RT-PCR for cyp26a1 in apcmcr mutant fish over a developmental time course.

Close modal

Cyp26a1 expression in apcmcr mutants is not due to retinoic acid induction. Previous studies have shown that cyp26a1 is strongly regulated by retinoic acid, its chemical substrate (12). Indeed, cyp26a1 can be induced in the gut of 80 hpf zebrafish by the addition of retinoic acid (Fig. 3A). However, previous studies have also shown that apcmcr mutants are retinoic acid deficient (20, 23). To verify that apcmcr mutants are retinoic acid deficient, we did RT-PCR on 80 hpf wild type and apcmcr mutant fish and examined the expression of known retinoic acid target genes. In this analysis, crbp1, crbp2, hnf4α, rarγ, apoD, and meis1 were all down-regulated by 2-fold compared with wild-type embryos (Fig. 3B). This indicates that cyp26a1 up-regulation was not due to the presence of retinoic acid. In addition, treatment of apcmcr mutant embryos with 10 μmol/L 4-(N,N-diethylamino) benzaldehyde (DEAB) to inhibit retinoic acid production failed to reduce levels of cyp26a1 in apcmcr mutants, compared with wild type (data not shown). Finally, raldh2 mutant embryos (nls), which are also retinoic acid deficient and share a number of phenotypes with apcmcr mutant embryos, did not exhibit increased levels of cyp26a1 (Fig. 3C), indicating that up-regulation of cyp26a1 in apcmcr mutant embryos is not simply due to retinoic acid deficiency (7). These observations indicated that cyp26a1 elevation in apcmcr mutants was specifically due to loss of apc.

Figure 3.

RA induces cyp26a1 expression but is not responsible for its induction in apcmcr mutants. A, whole mount in situ hybridization of 80 hpf embryos stained for cyp26a1 in wildtype, RA-treated wild-type and apcmcr mutant fish. B, quantitative RT-PCR of known RA-target genes in apcmcr mutants. C, quantitative RT-PCR of cyp26a1 in raldh2 mutant fish, neckless (nls). Error bars refer to SD of three measurements.

Figure 3.

RA induces cyp26a1 expression but is not responsible for its induction in apcmcr mutants. A, whole mount in situ hybridization of 80 hpf embryos stained for cyp26a1 in wildtype, RA-treated wild-type and apcmcr mutant fish. B, quantitative RT-PCR of known RA-target genes in apcmcr mutants. C, quantitative RT-PCR of cyp26a1 in raldh2 mutant fish, neckless (nls). Error bars refer to SD of three measurements.

Close modal

Cyp26a1 is a WNT/β-catenin/TCF target gene. Because the targets of the WNT/β-catenin/TCF pathway are up-regulated in the presence of APC mutations, we examined the response of cyp26a1 expression to perturbation of several components in this pathway. First, we designed a splice-blocking antisense oligonucleotide morpholino directed against tcf4, the TCF expressed in the gut epithelium (27). We injected this into an apcmcr mutant carrying an integrated TOPdGFP transgene (28) to verify interrupted WNT signaling. Full knockdown of tcf4 showed severe lethality (data not shown). As such, we injected sublethal doses of the antisense into one- to two-cell embryos and harvested embryos for whole-mount in situ and quantitative RT-PCR at 56 and 80 hpf. Expression of cyp26a1 decreased in the brachial arches, eyes, and gut of tcf4 antisense-injected mutant embryos (Fig. 4A-C). In addition, injection of dominant negative LEF (dnLEF), a form of LEF unable to bind β-catenin, reduced cyp26a1 expression. In both cases, the reduction in cyp26a1 was paralleled by decreased expression of the control green fluorescent protein (GFP) reporter construct (Fig. 4D; ref. 28).

Figure 4.

Regulation of cyp26a1 by TCF/LEF. A, whole mount in situ hybridization of 56 and 80 hpf embryos stained for cyp26a1 with or without tcf4 antisense oligonucleotides morpholino knockdown. B, quantitative RT-PCR of cyp26a1 expression in wild-type (wt), apcmcr mutants (mut), and tcf4 morpholino-injected apcmcr mutants (mo) at 56 and 80 hpf. Error bars refer to SD after 3 replicates of 3 independent experiments. C, Cyp26a1 expression in the eyes of 80 hpf fish in wild-type, apcmcr mutant, and tcf4 morpholino-injected fish. D, quantitative RT-PCR of cyp26a1 (crossed bars) and GFP (black bars) expression in wild-type (wt), apcmcr mutants (mut), and 8 and 40 pg dnLEF-injected apcmcr mutants at 56 hpf in a TopGFP background. Error bars refer to SD after 3 replicates of 2 independent experiments.

Figure 4.

Regulation of cyp26a1 by TCF/LEF. A, whole mount in situ hybridization of 56 and 80 hpf embryos stained for cyp26a1 with or without tcf4 antisense oligonucleotides morpholino knockdown. B, quantitative RT-PCR of cyp26a1 expression in wild-type (wt), apcmcr mutants (mut), and tcf4 morpholino-injected apcmcr mutants (mo) at 56 and 80 hpf. Error bars refer to SD after 3 replicates of 3 independent experiments. C, Cyp26a1 expression in the eyes of 80 hpf fish in wild-type, apcmcr mutant, and tcf4 morpholino-injected fish. D, quantitative RT-PCR of cyp26a1 (crossed bars) and GFP (black bars) expression in wild-type (wt), apcmcr mutants (mut), and 8 and 40 pg dnLEF-injected apcmcr mutants at 56 hpf in a TopGFP background. Error bars refer to SD after 3 replicates of 2 independent experiments.

Close modal

We next looked at overexpressing β-catenin by injecting mRNA of the stabilized form of β-catenin, S37A, into wild-type embryos at the one-cell stage at 0, 5, 10, 15, and 25 pg amounts. Expression of cyp26a1 rose in a dose-dependent manner, along with another WNT target gene, matrix metalloproteinase 13 (mmp13; Fig. 5A; ref. 29). Further, injection of 10 pg of a dominant-negative form of GSK3β mRNA, which stabilizes β-catenin levels, also caused cyp26a1 and mmp13 expression to increase, 2- and 4-fold, respectively (Fig. 5B). These data indicate that cyp26a1 is a target of the WNT/β-catenin/TCF pathway.

Figure 5.

Regulation of cyp26a1 by β-catenin. A, dose-dependent up-regulation of mmp13 (black bars) and cyp26a1 (crossed bars) at 80 hpf following injection of β-catenin mRNA into wild-type embryos. B, induction of mmp13 (black bars) and cyp26a1 (crossed bars) at 80 hpf following injection of 10 pg dnGSK3β mRNA into wild-type embryos. Error bars refer to SD of 3 replicates in 3 independent experiments.

Figure 5.

Regulation of cyp26a1 by β-catenin. A, dose-dependent up-regulation of mmp13 (black bars) and cyp26a1 (crossed bars) at 80 hpf following injection of β-catenin mRNA into wild-type embryos. B, induction of mmp13 (black bars) and cyp26a1 (crossed bars) at 80 hpf following injection of 10 pg dnGSK3β mRNA into wild-type embryos. Error bars refer to SD of 3 replicates in 3 independent experiments.

Close modal

Inhibition of cyp26a1 increases intestinal differentiation. We have shown previously that apcmcr mutants lack markers of terminal differentiation in the gut, pancreas, eye, and liver and that addition of retinoic acid rescued these terminal differentiation markers (20, 23). Because cyp26a1 is largely responsible for degrading retinoic acid (30), we hypothesized that cyp26a1 overexpression contributed to the lack of terminal differentiation seen in the intestines of apcmcr embryos. To test this idea, we designed an antisense morpholino to block cyp26a1 splicing and injected it into apcmcr mutant embryos along with parallel injections of control oligonucleotides. Confirmation of cyp26a1 splice blocking was provided by quantitative RT-PCR for the unspliced product (data not shown). Following confirmation of cyp26a1 knockdown, we examined control and cyp26a1 morphant embryos for expression of the intestinal differentiation marker i-fabp by whole-mount in situ hybridization. A parallel set of embryos was treated with retinoic acid as previously described and served as a positive control. As expected, treatment of embryos with retinoic acid increased i-fabp expression with ∼30% (n = 92) of the embryos staining positive. Similarly, knockdown of cyp26a1 increased the number of i-fabp-positive guts in apcmcr mutants compared with ∼60% (n = 57). This was in contrast to control injected embryos with only 3% positive staining for i-fabp (Fig. 6A and B).

Figure 6.

Reduction of cyp26a1-mediated RA catabolism leads to increased terminal differentiation. A, whole-mount in situ hybridization of 80 hpf embryos stained for gut terminal differentiation marker i-fabp (arrows) in wild-type, apcmcr mutants, apcmcr mutants treated with retinoic acid, apcmcr mutants injected with splice-blocking morpholino to cyp26a1, and apcmcr mutants treated with RAMBA. B, embryos scored for presence of i-fabp in the gut in wild-type, apcmcr mutants, apcmcr mutants treated with retinoic acid, apcmcr mutants injected with splice-blocking morpholino targeted to cyp26a1, and apcmcr mutants treated with RAMBA.

Figure 6.

Reduction of cyp26a1-mediated RA catabolism leads to increased terminal differentiation. A, whole-mount in situ hybridization of 80 hpf embryos stained for gut terminal differentiation marker i-fabp (arrows) in wild-type, apcmcr mutants, apcmcr mutants treated with retinoic acid, apcmcr mutants injected with splice-blocking morpholino to cyp26a1, and apcmcr mutants treated with RAMBA. B, embryos scored for presence of i-fabp in the gut in wild-type, apcmcr mutants, apcmcr mutants treated with retinoic acid, apcmcr mutants injected with splice-blocking morpholino targeted to cyp26a1, and apcmcr mutants treated with RAMBA.

Close modal

To further assess the role of cyp26a1 in preventing differentiation in apcmcr mutants, we synthesized a known small-molecule inhibitor of cyp26a1, a RAMBA (25). We treated apcmcr mutants with vehicle or RAMBA at (900 nmol/L) and determined the number of embryos that stained positively for the expression of i-fabp. Apcmcr mutants treated continuously with RAMBA for 4 days starting at 70% epiboly exhibited 70% (n = 35) rescue of terminal marker i-fabp compared with the ∼30% rescue with retinoic acid alone (Fig. 6A and B). In addition to the standard treatment, treatment with RAMBA starting at 48 and 72 hpf led to some rescue, ∼70% (n = 25) and ∼30% (n = 36), respectively, whereas treatment with retinoic acid starting at 48 or 72 hpf was without effect. Therefore, inhibition of retinoic acid catabolism by interference with cyp26a1 activity, either by antisense knockdown or pharmacologic inhibition, leads to increased terminal differentiation in apcmcr mutant fish.

Impaired retinoic acid signaling has been found in a large number of cancers, including those of the head, neck, colon, breast, prostate, and kidney (14, 3032). This dysregulation can occur at multiple levels, including loss of retinoic acid receptor function, vitamin A absorption and storage defects, impaired retinoic acid biosynthesis, and enhanced retinoic acid catabolism (20, 23, 3134). Although deficiencies have been documented for each of these steps, we know little about the mechanisms underlying these alterations or whether they are causative in tumor formation. Here, we show that increased expression of CYP26A1, the major retinoic acid–metabolizing cytochrome P450 enzyme in the intestine (35), occurs as a direct consequence of mutations in the colon tumor suppressor gene, APC. In addition, we show for the first time that CYP26A1 is regulated by the WNT signaling pathway and that overexpression of CYP26A1 within the intestine impairs intestinal differentiation. Our data suggest a model wherein APC plays a role in promoting intestinal cell differentiation by suppressing the catabolism of retinoic acid.

A number of recent studies indicate that the up-regulation of CYP26A1 might contribute to the formation of carcinomas. For example, high levels of CYP26A1 were reported in head and neck cancers (32). Similarly, breast and colon carcinoma cell lines seem to harbor high levels of CYP26A1 (30). Further, retinoic acid resistance in acute promyelocytic leukemia patients has been attributed to increased CYP26A1 activity (13, 14). Finally, a recent finding by Osanai et al. (36) indicates that expression of Cyp26a1 protects cells from apoptosis, a key hallmark of malignant cells. The finding that CYP26A1 levels are increased in APC mutant tissues of zebrafish, mice, and humans agrees with previous reports implying a role for CYP26A1 in carcinoma formation. However, our studies extend these findings in a number of ways. First, the up-regulation of CYP26A1 is dependent on loss of APC, a major colon tumor suppressor gene. Moreover, the induction of cyp26a1 seems to precede carcinoma formation in that the intestines of apcmcr mutant zebrafish embryos at 80 hpf already show robust up-regulation. Similarly, APC-dependent adenomas from mice and humans also showed CYP26A1 induction. Interestingly, cyp26a1 induction in homozygous apcmcr mutant zebrafish embryos occurred in the absence of any evident formation of intestinal neoplasms, thereby suggesting that up-regulation of cyp26a1 is part of the initiating events following apc loss. Consistent with CYP26A1 dysregulation as an early event, we found that its induction is dependent on WNT signaling, a well described consequence of APC loss.

The control of CYP26A1 by APC and WNT provides evidence for a dual role for APC in the control of both retinoic acid production as well as retinoic acid catabolism. Recent studies indicate that loss of APC results in loss of expression of retinoic acid biosynthetic enzymes like DHRS9 in humans and rdh1 and rdh1l in the zebrafish intestine (19, 20, 23). Combined with our current findings, it seems that APC may switch intestinal cells between a retinoic acid synthetic state and a retinoic acid catabolic state, thereby helping to promote intestinal cell differentiation. This mechanism would allow for the precise control of retinoic acid levels within cells at various stages of maturation and differentiation within the intestine. For example, stem cells undergoing self-renewal in response to WNT stimulation could be protected from retinoic acid effects by up-regulation of CYP26A1, a notion supported by the fact that Cyp26a1 was originally identified in mouse embryonic stem cells (37). Indeed, this type of juxtaposition of retinoic acid production with retinoic acid catabolism occurs in a number of developmental settings. Koubova et al. (38) recently found that the expression of Cyp26b1 determines expression of Stra8, which determines sex-specific regulation of germline meiosis in stem cells. Further, Cyp26b1 has also been shown to protect the distally expanding developing mouse limb bud from retinoic acid–mediated differentiation initiated at the proximal trunk, coordinately regulating proliferation with differentiation (9). Finally, the dorsal/ventral patterning of the developing chick retina is determined by the expression of several Cyp26 isoforms and retinoic acid (39).

The regulation of CYP26A1 by WNT is significant because previous studies of CYP26A1 expression have defined retinoic acid as its only known direct regulator. Indeed CYP26A1 is a well-known, direct target of ATRA (12). It seems, however, that the up-regulation of CYP26A1 following APC loss is independent of retinoic acid. Several lines of evidence support this possibility. First, apcmcr mutant zebrafish lack RDHs and show profound retinoic acid deficiency phenotypes such as lack of pectoral fins and lack of jaw development (20, 23). In agreement with this finding, we found that a number of retinoic acid target genes, including crbpI, crbpII, hnf4α, rarγ, apoD, and meis1 were down-regulated in apcmcr mutants, whereas cyp26a1 was up-regulated (4044). In addition, treatment of apcmcr mutant embryos with DEAB, a raldh2 inhibitor failed to suppress cyp26a1 levels. As such, it seems the induction of cyp26a1 in the apcmcr mutant zebrafish is not a result of increased retinoic acid biosynthesis. Regulation of cyp26a1 through retinoic acid–independent pathways is in agreement with findings by Dobbs-McAuliffe et al. (45) showing the continued expression of cyp26a1 in the presence of DEAB in the caudal end of the developing zebrafish embryo, as well as, by Iulianella et al. (46) in mouse embryos.

A number of previous studies in rodents support the use of retinoic acid in preventing colon carcinogenesis. For example, several retinoids suppressed aberrant crypt formation in F344 rats treated with azoxymethane (47). Similarly, ATRA reduced the number of aberrant crypts in Sprague-Dawley rats exposed to the same carcinogen (48). In contrast, however, a recent study using ATRA in the ApcMIN mouse found enhanced adenoma formation within the small intestine. However, no changes in adenoma formation were observed within colon of ApcMIN mice following treatment with ATRA (22). It is unclear what accounts for the difference in response between adenomas of the small intestine and the colon to ATRA. Our studies examined CYP26A1 expression in adenomas taken from the colon of ApcMIN mice, as well as human FAP adenomas and sporadic carcinomas taken from the colon. It is possible that the small intestine and colon respond differently to retinoic acid and that up-regulation of Cyp26a1 in colon adenomas of ApcMIN mice may have protected colon adenomas from the action of ATRA. The complexities of retinoic acid responses in the intestine require further investigation to resolve these apparent paradoxical findings.

We found that genetic or pharmacologic inhibition of cyp26a1 alone reversed intestinal differentiation defects in apcmcr mutant zebrafish. Our findings are in agreement with previous studies supporting the notion that CYP26A1 represents a potential pharmacologic target for the treatment of cancer. For example, low-dose treatment of MCF7 cells with liarozole, a specific CYP26 inhibitor, and retinoic acid led to significant increases in differentiation, reduction of proliferation, cell adhesion, and metastatic potential (49). Similarly, R116010, another highly specific CYP26 inhibitor, inhibited the growth of mammary tumor cells (50). Finally, Njar et al. (23, 25) recently found that treatment with novel synthetic RAMBAs resulted in increased apoptosis, differentiation, and a reduction in tumor size in mouse xenograft models. Taken together, the above findings indicate a need for continued investigation into the possibility of improving retinoic acid responsiveness through simultaneous treatment with retinoic acid and CYP26A1 inhibitors.

Grant support: American Cancer Society, National Cancer Institute, and Huntsman Cancer Foundation (D.A. Jones).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Richard Dorsky (University of Utah) for providing the TOPdGFP zebrafish, and the DNA peptide resource, the centralized zebrafish animal resource, and the mouse facility at the University of Utah.

1
Baltes S, Nau H, Lampen A. All-trans retinoic acid enhances differentiation and influences permeability of intestinal Caco-2 cells under serum-free conditions.
Dev Growth Differ
2004
;
46
:
503
–14.
2
De Luca L, Wolf G. Mechanism of action of vitamin A in differentiation of mucus-secreting epithelia.
J Agric Food Chem
1972
;
20
:
474
–6.
3
Blaner WS, Piantedosi R, Sykes A, Vogel S. Retinoids: the biochemical basis of vitamin A and retinoid action. Berlin: Springer-Verlag; 1999. p. 117–44.
4
Chambon P. A decade of molecular biology of retinoic acid receptors.
FASEB J
1996
;
10
:
940
–54.
5
Hollemann T, Chen Y, Grunz H, Pieler T. Regionalized metabolic activity establishes boundaries of retinoic acid signalling.
EMBO J
1998
;
17
:
7361
–72.
6
Eichele G, Thaller C. Characterization of concentration gradients of a morphogenetically active retinoid in the chick limb bud.
J Cell Biol
1987
;
105
:
1917
–23.
7
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
–94.
8
Niederreither K, Abu-Abed S, Schuhbaur B, Petkovich M, Chambon P, Dolle P. Genetic evidence that oxidative derivatives of retinoic acid are not involved in retinoid signaling during mouse development.
Nat Genet
2002
;
31
:
84
–8.
9
Yashiro K, Zhao X, Nishino J, et al. Regulation of retinoic acid distribution is required for proximodistal patterning and outgrowth of the developing mouse limb.
Dev Cell
2004
;
6
:
411
–22.
10
Abu-Abed SS, Beckett BR, Metzger D, et al. Mouse P450RAI (CYP26) expression and retinoic acid-inducible retinoic acid metabolism in F9 cells are regulated by retinoic acid receptor γ and retinoid X receptor α.
J Biol Chem
1998
;
273
:
2409
–15.
11
Abu-Abed S, Dolle P, Chambon P, et al. Developing with lethal RA levels: genetic ablation of Rarg can restore the viability of mice lacking Cyp26a1.
Development
2003
;
130
:
1449
–59.
12
Loudig O, Babichuk C, White J, Abu-Abed S, Mueller C, Petkovich M. Cytochrome P450RAI(CYP26) promoter: a distinct composite retinoic acid response element underlies the complex regulation of retinoic acid metabolism.
Mol Endocrinol
2000
;
14
:
1483
–97.
13
Ozpolat B, Mehta K, Lopez-Berestein G. Regulation of a highly specific retinoic acid-4-hydroxylase (CYP26A1) enzyme and all-trans-retinoic acid metabolism in human intestinal, liver, endothelial, and acute promyelocytic leukemia cells.
Leuk Lymphoma
2005
;
46
:
1497
–506.
14
Ozpolat B, Mehta K, Tari AM, Lopez-Berestein G. All-trans-retinoic acid-induced expression and regulation of retinoic acid 4-hydroxylase (CYP26) in human promyelocytic leukemia.
Am J Hematol
2002
;
70
:
39
–47.
15
Decensi A, Costa A. Recent advances in cancer chemoprevention, with emphasis on breast and colorectal cancer.
Eur J Cancer
2000
;
36
:
694
–709.
16
Callari D, Sinatra F, Fiore B, et al. All trans retinoic acid sensitizes colon cancer cells to hyperthermia cytotoxic effects.
Int J Oncol
2003
;
23
:
181
–8.
17
Conney AH, Lou YR, Liu Y, et al. Some perspectives on dietary inhibition of carcinogenesis: studies with curcumin and tea.
Proc Soc Exp Biol Med
1997
;
216
:
234
–45.
18
Mollersen L, Paulsen JE, Olstorn HB, Knutsen HK, Alexander J. Dietary retinoic acid supplementation stimulates intestinal tumour formation and growth in multiple intestinal neoplasia (Min)/+ mice.
Carcinogenesis
2004
;
25
:
149
–53.
19
Jette C, Peterson PW, Jones DA, et al. The tumor suppressor adenomatous polyposis coli and caudal related homeodomain protein regulate expression of retinol dehydrogenase L.
J Biol Chem
2004
;
279
:
34397
–405.
20
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
–9.
21
He TC, Sparks AB, da Costa LT, et al. Identification of c-MYC as a target of the APC pathway.
Science
1998
;
281
:
1509
–12.
22
Shu W, Guttentag S, Lu MM, et al. Wnt/β-catenin signaling acts upstream of N-myc, BMP4, and FGF signaling to regulate proximal-distal patterning in the lung.
Dev Biol
2005
;
283
:
226
–39.
23
Patel JB, Brodie AM, Njar VC, et al. Novel retinoic acid metabolism blocking agents endowed with multiple biological activities are efficient growth inhibitors of human breast and prostate cancer cells in vitro and a breast tumor xenograft in nude mice.
J Med Chem
2004
;
47
:
6716
–29.
24
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
–5.
25
Njar VC, Nnane IP, Brodie AM. Potent inhibition of retinoic acid metabolism enzyme(s) by novel azolyl retinoids.
Bioorg Med Chem Lett
2000
;
10
:
1905
–8.
26
Haramis AP, Hurlstone A, Offerhaus GJ, et al. Adenomatous polyposis coli-deficient zebrafish are susceptible to digestive tract neoplasia.
EMBO Rep
2006
;
7
:
444
–9.
27
Young RM, Reyes AE, Allende ML. Expression and splice variant analysis of the zebrafish tcf4 transcription factor.
Mech Dev
2002
;
117
:
269
–73.
28
Dorsky RI, Sheldahl LC, Moon RT. A transgenic Lef1/β-catenin-dependent reporter is expressed in spatially restricted domains throughout zebrafish development.
Dev Biol
2002
;
241
:
229
–37.
29
Martinez C, Bhattacharya S, Freeman T, Churchman M, Ilyas M. Expression profiling of murine intestinal adenomas reveals early deregulation of multiple matrix metalloproteinase (Mmp) genes.
J Pathol
2005
;
206
:
100
–10.
30
Sonneveld E, van den Brink CE, van der Burg B, et al. Human retinoic acid (RA) 4-hydroxylase (CYP26) is highly specific for all-trans-RA and can be induced through RA receptors in human breast and colon carcinoma cells.
Cell Growth Differ
1998
;
9
:
629
–37.
31
Downie D, McFadyen MC, Miller ID, et al. Profiling cytochrome P450 expression in ovarian cancer: identification of prognostic markers.
Clin Cancer Res
2005
;
11
:
7369
–75.
32
Klaassen I, Brakenhoff RH, Smeets SJ, Snow GB, Braakhuis BJ. Enhanced turnover of all-trans-retinoic acid and increased formation of polar metabolites in head and neck squamous cell carcinoma lines compared with normal oral keratinocytes.
Clin Cancer Res
2001
;
7
:
1017
–25.
33
Ren M, Pozzi S, Bistulfi G, Somenzi G, Rossetti S, Sacchi N. Impaired retinoic acid (RA) signal leads to RARβ2 epigenetic silencing and RA resistance.
Mol Cell Biol
2005
;
25
:
10591
–603.
34
Arapshian A, Bertran S, Kuppumbatti YS, Nakajo S, Mira-y-Lopez R. Epigenetic CRBP downregulation appears to be an evolutionarily conserved (human and mouse) and oncogene-specific phenomenon in breast cancer.
Mol Cancer
2004
;
3
:
13
.
35
Lampen A, Meyer S, Arnhold T, Nau H. Metabolism of vitamin A and its active metabolite all-trans-retinoic acid in small intestinal enterocytes.
J Pharmacol Exp Ther
2000
;
295
:
979
–85.
36
Osanai M, Petkovich M. Expression of the retinoic acid-metabolizing enzyme CYP26A1 limits programmed cell death.
Mol Pharmacol
2005
;
67
:
1808
–17.
37
Ray WJ, Bain G, Yao M, Gottlieb DI. CYP26, a novel mammalian cytochrome P450, is induced by retinoic acid and defines a new family.
J Biol Chem
1997
;
272
:
18702
–8.
38
Koubova J, Menke DB, Zhou Q, Capel B, Griswold MD, Page DC. Retinoic acid regulates sex-specific timing of meiotic initiation in mice.
Proc Natl Acad Sci U S A
2006
;
103
:
2474
–9.
39
Sen J, Harpavat S, Peters MA, Cepko CL. Retinoic acid regulates the expression of dorsoventral topographic guidance molecules in the chick retina.
Development
2005
;
132
:
5147
–59.
40
Balmer JE, Blomhoff R. Gene expression regulation by retinoic acid.
J Lipid Res
2002
;
43
:
1773
–808.
41
Liu T, Bohlken A, Smith S, et al. The retinoid anticancer signal: mechanisms of target gene regulation.
Br J Cancer
2005
;
93
:
310
–8.
42
Ogura Y, Suruga K, Takase S, Goda T. Developmental changes of the expression of the genes regulated by retinoic acid in the small intestine of rats.
Life Sci
2005
;
77
:
2804
–13.
43
Lopez-Boado YS, Klaus M, Dawson MI, Lopez-Otin C. Retinoic acid-induced expression of apolipoprotein D and concomitant growth arrest in human breast cancer cells are mediated through a retinoic acid receptor RARα-dependent signaling pathway.
J Biol Chem
1996
;
271
:
32105
–11.
44
Freemantle SJ, Kerley JS, Olsen SL, Gross RH, Spinella MJ. Developmentally-related candidate retinoic acid target genes regulated early during neuronal differentiation of human embryonal carcinoma.
Oncogene
2002
;
21
:
2880
–9.
45
Dobbs-McAuliffe B, Zhao Q, Linney E. Feedback mechanisms regulate retinoic acid production and degradation in the zebrafish embryo.
Mech Dev
2004
;
121
:
339
–50.
46
Iulianella A, Beckett B, Petkovich M, Lohnes D. A molecular basis for retinoic acid-induced axial truncation.
Dev Biol
1999
;
205
:
33
–48.
47
Zheng Y, Kramer PM, Lubet RA, Steele VE, Kelloff GJ, Pereira MA. Effect of retinoids on AOM-induced colon cancer in rats: modulation of cell proliferation, apoptosis and aberrant crypt foci.
Carcinogenesis
1999
;
20
:
255
–60.
48
Stopera SA, Bird RP. Effects of all-trans retinoic acid as a potential chemopreventive agent on the formation of azoxymethane-induced aberrant crypt foci: differential expression of c-myc and c-fos MRNA and protein.
Int J Cancer
1993
;
53
:
798
–803.
49
Wouters W, van Dun J, Dillen A, Coene MC, Cools W, De Coster R. Effects of liarozole, a new antitumoral compound, on retinoic acid-induced inhibition of cell growth and on retinoic acid metabolism in MCF-7 human breast cancer cells.
Cancer Res
1992
;
52
:
2841
–6.
50
Van Heusden J, Van Ginckel R, Floren W, et al. Inhibition of all-trans-retinoic acid metabolism by R116010 induces antitumour activity.
Br J Cancer
2002
;
86
:
605
–11.