B-Catenin Activates the Growth Factor Endothelin-1 in Colon Cancer Cells

Tae Hoon Kim1, Hui Xiong2, Zhuohua Zhang*,2 and Bing Ren*,1,3

1Laboratory of Gene Regulation, Ludwig Institute for Cancer Research, UCSD School of Medicine, 9500 Gilman Drive, La Jolla, CA 92093, USA; 2The Burnham Institute, 10901 N Torrey Pines Road, La Jolla, CA 92037, USA; 3Department of Cellular and Molecular Medicine, UCSD School of Medicine, 9500 Gilman Drive, La Jolla, CA 92093, USA

Endothelin-1 (EDN1) is a growth factor that is frequently plays an important role in maintaining proper vascular produced by cancer cells and plays a critical role in tone (Yanagisawa et al., 1988; Kedzierski and Yanagi- tumorigenesis. However, the molecular mechanism con- sawa, 2001). trolling the expression of EDN1 in cancers is unknown. EDN1 is frequently secreted by many solid tumors, Constitutive activation of b-catenin pathway is responsible including prostate, colorectal, liver, breast and ovarian for the initiation of the vast majority of colon cancers. cancers (Kusuhara et al., 1990; Nelson et al., 2003). In Here we show that the EDN1 gene is directly regulated by these tumor cells, EDN1 has been shown to promote cell b-catenin in colon cancer cells. A specific DNA element proliferation, suppress apoptosis, promote metastasis, within the EDN1 promoter is required for activation, and and facilitate angiogenesis (Bagnato and Spinella, 2003; is associated with b-catenin’s cognate DNA binding Grant et al., 2003; Nelson et al., 2003; Yin et al., 2003). partner, TCF4, in vivo. Inhibition of b-catenin signaling EDN1 exerts its diverse functions through two cell- results in lowered expression of EDN1, while enhance- surface receptors, EDNRA and EDNRB (Kedzierski ment of b-catenin signaling leads to further activation of and Yanagisawa, 2001).In animal tumor models, the gene. Significantly elevated EDN1 expression occurs endothelin receptor antagonists have demonstrated in 80% of primary human colon cancers, consistent with it remarkable effects in suppressing tumor growth (Bag- being a direct target of b-catenin. Furthermore, EDN1 is nato et al., 2002; Rosano et al., 2003). In clinical trials, able to rescue colon cancer cells from growth arrest and these antagonists significantly delay the progression of apoptosis resulting from inhibition of b-catenin signaling, metastasis in hormonal-refractory prostate cancers implicating a key role of EDN1 in promoting the (Nelson, 2003).These observations implicate abnormal oncogenic function of b-catenin. These results indicate expression of EDN1 to be a key step in tumorigenesis of EDN1 overexpression as a major cause in colon cancers many solid tumors. and reveal further details of the genetic programs Understanding the mechanisms of EDN1 expression responsible for tumorigenesis of colon cancers. in tumor cells may provide novel therapeutic strategies Oncogene (2005) 24, 597–604. doi:10.1038/sj.onc.1208237 to inhibit initiation and progression of cancers.Cur- Published online 22 November 2004 rently, the molecular mechanisms responsible for pathological synthesis of EDN1 in tumor cells are not Keywords: b-catenin; endothelin-1; colon cancer; ChIP; known.Unlike many oncogenes, no chromosomal promoter arrays translocations or amplifications of the EDN1 gene have been reported.It has been suggested that elevated levels of EDN1 are correlated with decreased levels of neural endopeptidase, the enzyme responsible for degradation and clearance of EDN1, in advanced prostate cancers Introduction (Papandreou et al., 1998). However, both the EDN1 peptide and mRNA levels are shown to be elevated in EDN1 plays a key role in cell growth and differentia- colon cancers, indicating that transcriptional activation tion, as well as in vascular homeostasis in mammals of the EDN1 gene is the likely mechanism of EDN1 (Kedzierski and Yanagisawa, 2001; Nelson et al., 2003). overexpression in this cancer (Egidy et al., 2000). We During embryonic development, EDN1 is produced by therefore hypothesized that EDN1 overexpression in cephalic and cardiac neural crest stem cells to mediate colon cancers is a result of transcriptional activation by proper differentiation of these cells into cranial facial an oncogenic pathway. and cardiac outflow structures (Kurihara et al., 1994; Constitutive activation of b-catenin pathway is Kedzierski and Yanagisawa, 2001).In adults, EDN1 is responsible for initiation of a majority of colon cancers primarily expressed in vascular epithelial cells where it (Kinzler and Vogelstein, 1996; Giles et al., 2003). This pathway translates the growth-promoting signals of Wnt ligands to transcriptional activation of a diverse set *Correspondence: B Ren or Z Zhang; E-mail: [email protected] or [email protected] of target genes.In normal cells, b-catenin is associated Received 17 April 2004; revised 21 September 2004; accepted 23 with E-cadherin and a-catenin to participate in cell–cell September 2004; published online 22 November 2004 interaction.Free b-catenin is phosphorylated by the b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 598 APC/Axin/GSK3b complex and targeted to protea- a some-dependent degradation (Maniatis, 1999).In re- TBE 1 TBE 2 TBE 3 TBE 4 sponse to Wnt signaling, free cytoplasmic b-catenin −827 CTTTGCT −308 AGCAAAG +97 CTTTGAG +123 CTTTGAG becomes stabilized, and is translocated into the nucleus where it complexes with the lymphoid enhancer factor/ T-cell factor (LEF/TCF) family of DNA binding ChIP PCR fragments proteins to activate the transcription of target genes -7K -1K +1 (Bienz and Clevers, 2000; He, 2003).In over 85% of b c colon cancer patients, the APC gene is inactivated TCF4 IP INPUT APC β no Ab − + − + -cat IP INPUT by mutations resulting in constitutive activation of APC no Ab − + − + b-catenin pathway (Cohen, 2003).Moreover, in APC +1 +1 positive colon cancer cases, the b-catenin gene is −1K frequently mutated at the residues required for its − targeting to proteasomal degradation (Kinzler and 7K Vogelstein, 1996; Bienz and Clevers, 2000).The CTRL activated b-catenin pathway is believed to initiate a tumorigenic program by promoting cell proliferation via e 14 d 1.6 APC− 12 activation of a target gene, MYC, and at the same time, 1.4 APC+ protecting cells from apoptosis through regulation of 1.2 10 unknown target genes (Giles et al., 2003). Here, we 1 8 0.8 demonstrate that b-catenin activates EDN1 expression 6 0.6 in colon cancers, and that EDN1 expression contributes 0.4 4 significantly to the survival of colon cancer cells. 0.2 Reporter Activity 2 EDN1 Reporter Activity 0 0 WT TCF4 DN TCF4 WT muTBE1 muTBE2 muTBE3 muTBE4 Expression Constructs EDN1 Reporters

Results fgWT EDN1 mu EDN1 WT MYC mu MYC

TCF4 TCF4 TCF4 TCF4 We employed a previously established genome-wide location analysis approach (Ren et al., 2000, 2002; Li et al., 2003) to examine the in vivo binding of TCF4 in a colorectal tumor cell line HT29-APC (Morin et al., 1996).Our method involved crosslinking of cells with formaldehyde, extraction of chromatin and immunoprecipitation of protein/DNA complexes with a specific Figure 1 EDN1 promoter is directly regulated by TCF4.( a) Annotation of the EDN1 promoter and gene: TBE ¼ TCF binding antibody against TCF4.The resulting TCF4-bound sites element (TBE1–4), À7K, À1K, and þ 1 represents PCR fragments were identified by competitive hybridization of the amplified in (b).( b) HT29-APC chromatin was immunoprecipi- fluorescently labeled DNA to a microarray containing tated with anti-TCF4 antibody (TCF4 IP).The À7K, À1K, and B5000 known human gene promoters (Li et al., 2003). þ 1 fragments upstream of the EDN1 gene were PCR amplified from the immunoprecipitated and input chromatin to determine We found the EDN1 promoter to be one of the most TCF4 binding before and after APC induction.( c) HT29-APC significantly enriched genomic regions by the TCF4 chromatin was immunoprecipitated with anti-b-catenin antibody antibody (Po10À5). (b-cat IP).The þ 1 fragments upstream of the EDN1 gene were Sequence analysis of the EDN1 promoter revealed PCR amplified from the immunoprecipitated and input chromatin to determine b-catenin binding before and after APC induction.( d) four potential TCF4 binding elements (TBE) (van de Transient co-transfection experiment with the wild-type EDN1 Wetering et al., 1991) within the 1 kb upstream of the luciferase reporter with either wild-type TCF4 (WT TCF4) or transcriptional initiation site (Figure 1a).We designed dominant-negative TCF4 (DN TCF4).( e) Transient transfection specific oligonucleotide primers to amplify each TBE experiment with the wild type (WT) or mutant EDN1 luciferase containing DNA fragment and performed polymerase reporters (muTBE1, muTBE2, muTBE3, and muTBE4).Trans- fected cells were treated with or without 100 mM ZnCl2 and the chain reactions (PCR) with these primers to test luciferase activity was measured 12 h post-Zn2 þ induction.( f) enrichment of each DNA fragment by the TCF4 EMSA (electrophoretic mobility shift assay) of double stranded 45- antibody (aTCF4) in a conventional chromatin immu- mer oligo containing either WT or mutant TBE4 from EDN1 noprecipitation assay (ChIP) (Figure 1a).This analysis promoter using increasing amounts (0.125, 0.25, 0.5 and 1.0 mg) of purified recombinant GST-TCF4.The EMSA gel was stained with indicated that the þ 1 fragment containing TBE3 and SYBR green I and visualized by UV fluorescence.( g)EMSA TBE4 was specifically precipitated with the aTCF4 (electrophoretic mobility shift assay) of double stranded 36-mer antibody from crosslinked chromatin extracts, while oligo containing either WT or mutant TBE1 element from MYC fragments containing the upstream TBE1 and TBE2 promoter using increasing amounts (0.125, 0.25, 0.5 and 1.0 mg) of were not (Figure 1b).In addition, the association of purified recombinant GST-TCF4.The EMSA gel was stained with SYBR green I and visualized by UV fluorescence TCF4 with the þ 1 fragment was not affected by APC induction (Figure 1b), but the association of b-catenin with the þ 1 fragment was inhibited upon APC induction (Figure 1c), consistent with constitutive

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 599 association of TCF4 and regulated association of ab − 3 b-catenin with their target genes. 1.2 APC NONE APC+ Li+ To determine whether the EDN1 promoter can be 2.5 regulated by b-catenin signaling, we constructed a 1 2 luciferase reporter with the EDN1 promoter sequence 0.8 spanning À1362 to þ 165 bps surrounding the transcrip- 1.5 tional initiation site.The activity of this reporter gene, 0.6 when transfected into HT29-APC cells, was inhibited by 0.4 1 Relative mRNA Levels co-transfecting a dorminant-negative inhibitor of TCF4 Relative mRNA Levels 0.2 0.5 (Figure 1d), indicating that the EDN1 promoter fragment is sufficient to confer responsiveness to 0 0 b-catenin signaling to a heterologous gene.To determine B2M EDN1 MYC EDN1 MYC Endogenous Genes Endogenous Genes which of the four TBE sites were actually responsible for TCF4 binding and activation, we introduced point cd mutations into each TBE (muTBE1–4, Figure 1a), and 120 100 transfected the mutant reporter constructs into HT29- 100 80 APC cells.As shown in Figure 1e, transcriptional 80 activity of the wild-type EDN1 promoter decreased 60 more than three-fold upon APC induction.The muta- 60 tions in TBE1–3 sites, while causing a two-fold 40 reduction of EDN1 promoter activity, did not affect 40 EDN1 Protein (pg/mL) its response to b-catenin signaling.In contrast, mutation 20 EDN1 Protein (pg/mL) 20 of TBE4 completely abolished the EDN1 promoter 0 0 control β-cat β-cat DN TCF4 TCF4 DN activity as well as its response to b-catenin signaling control β-cat/siRNA TCF4/siRNA (Figure 1e), highly consistent with the ChIP results Expression Constructs siRNA (Figure 1b).This result indicated that TBE4 regulates Figure 2 Endogenous EDN1 mRNA and protein levels are EDN1 expression in response to b-catenin signaling. positively regulated by b-catenin and TCF4.( a) The endogenous To confirm that the mutations in TBE4 specifically EDN1, MYC, and beta-microglobulin (B2M, control) mRNA affected the binding of TCF4 to this element, electro- levels in HT29-APC cells were determined by quantitative real time PCR before and after Zn2 þ induction of the APC gene.( b) The phoretic mobility shift assay (EMSA) was performed endogenous EDN1 and MYC mRNA levels in HCT116 cells were using recombinant TCF4 protein (Morin et al., 1997). determined by quantitative real time PCR before and after Li þ Consistent with the results from our transfection treatment.( c) HT29 cells were transfected with expression plasmids experiments, the recombinant TCF4 protein bound encoding either wild-type b-catenin (b-cat), dominant-negative specifically to WT TBE4 DNA in a concentration- b-catenin (b-cat DN), wild-type TCF4 (TCF4), or dominant- negative TCF4 (TCF4 DN).Endogenous mature EDN1 peptide in dependent manner.In contrast, TCF4 protein failed to the transfected cell culture supernatant was quantified using bind to the mutated TBE4 (Figure 1f).We also ELISA.( d) HT29 cells were transiently transfected with an empty performed the same experiment with the known TCF4 vector (control) or vectors expressing short hairpin siRNA against binding site within the MYC promoter (TBE1) as a b-catenin (b-cat/siRNA) or against TCF4 (TCF4/siRNA) control and observed a similar binding profile as the TBE4 element within the EDN1 promoter (Figure 1g). observed a two-fold increase in EDN1 mRNA level Together, these results suggest that b-catenin directly upon Li þ treatment, comparable to the increased regulates the EDN1 promoter activity via TBE4 element mRNA level observed for our positive control, MYC in the EDN1 promoter. (Po0.01, Figure 2b). Next, we examined whether endogenous EDN1 To determine if b-catenin regulated transcription of expression is indeed transcriptionally activated by EDN1 actually resulted in regulated production of the b-catenin signaling, leading to elevated EDN1 protein matured EDN1 peptide, we performed enzyme-linked levels in cancer cells.To test this hypothesis, we immunosorbent assays (ELISA).Consistent with tran- suppressed b-catenin signaling in HT29-APC cells by scriptional regulation of the EDN1 gene by b-catenin inducing expression of the wild-type APC transgene. signaling, the level of secreted EDN1 peptide in HT29 The endogenous EDN1 mRNA level was reduced by cells was significantly reduced upon introduction of a more than 50% (Po0.01) upon APC induction. The dominant-negative mutant of TCF4 or b-catenin, while extent of this reduction was similar to that of a known overexpression of wild-type b-catenin or TCF4 resulted b-catenin target gene, MYC, under the same conditions in augmented EDN1 peptide production (Figure 2c). (Figure 2a).Using HCT116 cells that are heterozygous Furthermore, siRNAs against either b-catenin or TCF4 for the wild-type b-catenin gene (Chan et al., 2002), we inhibited production of EDN1 peptide in these cells tested whether activation of b-catenin signaling would (Po0.001, Figure 2d). These results demonstrate that induce transcription of the EDN1 gene by lithium (Li þ ) activity of b-catenin pathway is indeed required for high treatment.Li þ treatment is known to inhibit GSK3b, levels of EDN1 expression and production in colorectal resulting in stabilization and activation of the endogen- cancer cells. ous wild-type b-catenin protein (Klein and Melton, Since mutations in b-catenin pathway are one of the 1996; Stambolic et al., 1996) in HCT116 cells. We most frequent oncogenic events in human cancers and

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 600 are detected in over 85% of all human colon cancers MYC gene, a well-known target of b-catenin signaling, (Kinzler and Vogelstein, 1996), we asked whether EDN1 would be also highly expressed in these tumor samples mRNA levels are elevated at a high frequency in colon over normal tissues.Indeed, we found that MYC is cancer patients.We examined EDN1 mRNA levels in overexpressed in 87.5% (14 out of 16) of the tumor tissues colon cancer samples and matched normal colon tissues with elevated EDN1 expression (Po0.05, Figure 3b and from human patients and found that signiﬁcantly c).These results provide a strong evidence for the direct elevated levels (two- to 86-fold increase) of EDN1 role of b-catenin in EDN1 activation in human colon mRNA were observed in 16 out of 20 tumor samples cancers. (Po0.05, Figure 3a and c). Importantly, the reported To evaluate the functional signiﬁcance of EDN1 frequency of EDN1 overexpression in human colon expression in colon cancer cells, we investigated the cancers is highly consistent with the combined, reported effect of exogenous EDN1 peptide on growth of colon frequency of mutations that activate b-catenin signaling cancer cells upon inhibition of b-catenin signaling. (Kinzler and Vogelstein, 1996).We reasoned that the Previous studies have shown that induction of the

a 86 35

5 Relative EDN1 mRNA Levels 0 1234567891011121314151617181920 Paired Colon Samples

b 53 78 108 42 35

5 Relative MYC mRNA Levels 0 1 2 3 4 5 6 7 8 9 1011121314151617181920 Paired Colon Samples

c SAMPLE # SITE STAGE DIFFERENTIATION P Value (EDN1) P Value (MYC) 1 RIGHT SIDE III (Dukes' C) MODERATE 0.001289 0.013136 2 RIGHT SIDE III (Dukes' C) MUCINOUS 0.134852 0.000039 3 LIVER METASTASIS IV WELL 0.000710 0.207881 4 RIGHT SIDE II (Dukes' B) - 0.026040 0.000023 5 SIGMOID, RECTUM II (Dukes' B) WELL, MUCINOUS 0.008049 0.000765 6 RIGHT SIDE, CECUM II (Dukes' B) MODERATE, MUCINOUS 0.011385 0.003183 7 RIDE SIDE III (Dukes' C) MODERATE 0.024639 0.008979 8 - - - 0.028595 0.000576 9 RIGHT SIDE - MODERATE 0.000713 0.004926 10 RIGHT SIDE - MUCINOUS 0.091752 0.000122 11 RIGHT SIDE II (Dukes' B) MODERATE 0.000215 0.001280 12 RIGHT SIDE III (Dukes' C) MODERATE 0.001798 0.000168 13 LEFT SIDE III (Dukes' C) MODERATE 0.000166 0.000728 14 RECURRENT III (Dukes' C) MODERATE 0.010400 0.000775 15 SIGMOID, RECTUM III (Dukes' C) MODERATE 0.000153 0.001856 16 RIGHT SIDE - - 0.004926 0.000317 17 RIGHT SIDE II (Dukes' B) - 0.000144 0.000311 18 RECTAL III (Dukes' C) MODERATE 0.000031 0.001112 19 - III (Dukes' C) MODERATE 0.034241 0.002937 20 - - - 0.001068 0.001196 Figure 3 EDN1 mRNA is overexpressed in human color tumor samples.( a) In total, 20 pairs of primary human colon tissues were analysed for the endogenous EDN1 mRNA levels.Quantitative real time PCR results of NORMAL light bars and TUMOR dark bars samples.The relative mRNA levels were determined using the DDCt method using the gamma actin mRNA (ACTG1) as a reference gene.( b) In total, 20 pairs of primary human colon tissues were analysed for the endogenous MYC mRNA levels.Quantitative real time PCR results of NORMAL and TUMOR samples.The relative mRNA levels were determined using the DDCt method using the gamma actin mRNA (ACTG1) as a reference gene.( c) Available pathological information for 20 paired samples is described.The corresponding P-values from the student’s t-test of the observed EDN1 and MYC mRNA levels in TUMOR vs NORMAL samples are listed in columns labeled ‘P-value (EDN1)’ and ‘P-value (MYC)’ respectively.Cells highlighted in gray indicate Po0.05

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 601 wild-type APC transgene by 100 mM ZnCl2 in HT29- a APC cells results in growth inhibition and apoptosis NONE EDN1 (Morin et al., 1996), demonstrating the importance of b-catenin pathway in mediating cell proliferation and survival.EDN1 (100 p M) had little effect on the growth of HT29-APC cells in the absence of APC induction (Figure 4a and b).However, we found that addition of exogenous EDN1 (100 pM) to the APC-induced, growth arrested cells resulted in restora- tion of their proliferation and protection from apoptosis (Figure 4a).This effect appears to be largely mediated by EDN1 through suppressing cell death; 36% of the APC induced cells were dead, but addition of exogenous EDN1 reduced the dead cells to only APC APC + EDN1 8% of the population, similar to level observed for untreated cells (Figure 4b). b Dead G1 G2 S Previously, it was shown that MYC is a major effecter NONE EDN1 Dead - 12.88% Dead - 13.31% of cell growth from b-catenin signaling, and reactivation G1 - 42.57% G1 - 40.52% 500 of MYC by EDN1 would be a simple explanation of our 600 G2 - 6.37% G2 - 8.65% findings (van de Wetering et al., 2002). To determine if S - 38.18% S - 37.52% MYC reactivation by EDN1 may be the reason for the 300 300 protective effect of EDN1, we investigated the MYC mRNA levels in these cells.Although EDN1 alone can slightly augment MYC expression, when APC is 80 160 80 160 induced and exogenous EDN1 is added, the MYC mRNA level remained at a level observed when only APC APC + EDN1 Cell Number Dead - 35.95% Dead - 7.85% APC is induced, to a level less than 50% of the control G1 - 28.92% G1 - 33.39% G2 - 11.65% G2 - 22.32% cells (Figure 4c).Therefore, EDN1 suppression of APC 200 240 induced cell death is likely through a MYC-indepen- S - 23.47% S - 36.44% dent mechanism.From these results and previous 100 120 observations that endothelin receptor antagonists block proliferation and induce apoptosis of HT29 cells (Peduto Eberl et al., 2003), we conclude that activation 80 160 80 160 of EDN1 expression is an essential aspect of b-catenin DNA Content pathway in maintaining the growth and survival of c 1.6 colon cancer cells. 1.4 1.2 Discussion 1.0 0.8 Based on these findings, we propose that the frequent 0.6 overexpression of the EDN1 gene in human colon 0.4 0.2 cancers is a direct consequence of genetic alterations of Relative MYC mRNA Levels 0 b-catenin signaling in these tumors.Inactivating muta- NONE APC EDN-1 APC + EDN-1 tions in the APC gene or activating mutations in Treatment b-catenin lead to the formation of b-catenin/TCF4 Figure 4 EDN1 plays a key role in mediating oncogenic activity of complex on the EDN1 promoter, which in turn activates b-catenin pathway.( a) Crystal violet staining of HT29-APC cells transcription of the gene.EDN1 would contribute to that were untreated (NONE) or treated with 100 mM ZnCl2 (APC), b-catenin’s oncogenic program by providing antiapop- exogenous 100 pM EDN1 (EDN1), or 100 mM ZnCl2 and 100 pM EDN1 (APC þ EDN1) for 72 h.( b) FACS analysis of HT29-APC totic and growth-promoting functions. cells that were treated as in the panel (a).Cells were fixed at 72 h Many target genes for the b-catenin pathway have post-treatment with 100 mM ZnCl2 and EDN1 peptide, and stained been previously identified (http://www.stanford.edu/ with propidium iodide.The ModFit cell cycle analysis program ~rnusse/wntwindow.html). One such target gene, was used to model the distribution of G1, S, G2, and dead cells in MYC, appears to play a central role by suppressing each treatment.( c) Quantitative realtime PCR of MYC mRNA in HT29-APC cells that were untreated (NONE) or treated with expression of the cyclin-dependent kinase inhibitor p21 100 mM ZnCl2 (APC), exogenous 100 pM EDN1 (EDN1), or 100 mM and by promoting cell proliferation (van de Wetering ZnCl2 and 100 pM EDN1 (APC þ EDN1) for 12 h et al., 2002). In addition to its role in proliferation, MYC is known to induce apoptosis by activating expression of proapoptotic genes (Pelengaris et al., 2002). For the most likely provided by EDN1.Consistent with this cancer cell population to expand, the MYC-induced model, it has been shown previously that EDN1 can apoptotic program must be suppressed, and this role is prevent MYC-induced apoptosis in Rat fibroblast cells

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 602 (Shichiri et al., 1998). Similarly, EDN1 can also protect A-30; and CTRL_Left, 50-TGGGTTTGGTAGGGGACATA-30, FasL-induced apoptosis in colorectal cancer cells (Peduto and CTRL_Right, 50-CTGGGCTCTGCTGGCTTA-30. Eberl et al., 2003). In our study, EDN1 can rescue colon cancer cells from cell death resulting from inhibition of Reporter constructs b-catenin pathway (Figure 4b).A possible mechanism of The EDN1 promoter sequence spanning À1362 to þ 165 bps EDN1 mediated suppression of cell death may involve surrounding the transcriptional initiation site was amplified activation of AKT by EDN1 as shown in ovarian from human genomic DNA by PCR using the primers, 50-CTC carcinoma cells (Del Bufalo et al.,2002). GAGGTTCCATGCACCTTTTGCTT-30 and 50-AAGCTTC Our results demonstrate b-catenin as a key regulator CCAAAGGAAAACGAAGAAAA-30.The EDN1 promoter of EDN1 and provide a framework for understanding PCR fragment was restriction enzyme digested with XhoI and the molecular basis of frequent EDN1 activation in BamHI (New England Biolabs, Beverly, MA, USA) and human cancer cells.Besides colon cancers, EDN1 cloned into the XhoI and BamHI acceptor sites within the overexpression has been reported for a number of other pGL2 luciferase reporter (Promega, Madison, WI, USA). cancers including prostate, breast, liver, and skin Mutations within the potential TCF4 binding elements cancers.Coincidentally, genetic lesions in APC, Axin, (TBE1–4) were introduced by site-directed mutagenesis using the QuickChanges kit (Invitrogen, Carlsbad, CA, USA).The and b-catenin genes have also been observed in all these primers used for site-directed mutagenesis were the following solid tumors.It would be interesting to test whether pairs: muTBE1_S, 50-CTGTGTGAAAAACACTTTTACTTT overexpression of EDN1 in these other cancer tissues is TCAATCAGTTTATCAGCCTCCTCCGC-30, and muTBE1_AS, a result of activation of b-catenin pathway. 50-GCGGAGGAGGCTGATAAACTGATTGAAAAGTAA AAGTGTTTTTCACACAG-30; muTBE2_S, 50-GCACGGG- CAGGTTTAGTAAAAGTCTCTAATGGG-30, and muTB E2_AS, 50-CCCATTAGAGACTTTTACTAAACCTGCCC Materials and methods GTGC-30; muTBE3_S, 50-GCGGGTTCCTCAGATCTTAA GGCGATCCTTCAGCCC-30, and muTBE3_AS, 50-GCGGG Chromatin immunoprecipitation TTCCTCAGATCTTAAGGCGATCCTTCAGCCC-30; and Chromatin immunoprecipitation was performed as described muTBE4_S, 50-GAGGAACCCGCAGCGCCTTAAGGGAC previously (Li et al., 2003). Briefly, 2 mg of sonicated CTGAAGCTGTTTTTCTTCG-30, and muTBE4_AS, 50-CG chromatin (OD260) was incubated with 10 mg of mouse AAGAAAAACAGCTTCAGGTCCCTTAAGGCGCTGCG monoclonal TCF4 antibody (Upstate Biotechnology, Charlot- GGTTCCTC-30. tesville, VA, USA) coupled to the sheep anti-mouse IgG magnetic beads (Dynal Biotech, Lake Success, NY, USA).The Transfections and ELISA magnetic beads were washed eight times with RIPA buffer containing 50 mM HEPES (pH 8.0), 1 mM EDTA, 1% NP-40, Transient transfection experiments were performed using 0.7% DOC, and 0.5 M LiCl, supplemented with Completet Lipofectaminet 2000 transfection reagent (Invitrogen, Carls- protease inhibitors (Roche Applied Science, Indianapolis, IN, bad, CA, USA), as described by the manufacturer.Briefly, USA), and washed once with TE (10 mM Tris pH 8.0, 1 mM HT29-APC cells were seeded at 1 Â 105 cells in each well of 24- EDTA).After washing, the bound DNA was eluted by heating well plate 24 h prior to transfection.Each luciferase reporter the beads to 651C in elution buffer (10 mM Tris, pH 8.0, 1 mM DNA (0.6 mg) was combined with 0.2 mg of CMV-lacZ control EDTA, and 1% SDS).The eluted DNA was incubated at 65 1C reporter DNA and mixed with 3 ml of Lipofectaminet 2000 in for 12 more hours to reverse the crosslinks.Following Opti-MEMs media (Invitrogen, Carlsbad, CA, USA).The incubation, the immunoprecipitated DNA was desalted using DNA–Lipofectamine mixture was placed into each well.After QIAquicks PCR purification kit (Qiagen, Valencia, CA, 24 h post-transfection, media was replaced and cells were USA).The purified DNA was blunt ended using T4 induced with 100 mM ZnCl2.After 12 h postinduction, cells polymerase (New England Biolabs, Beverly, MA, USA) and were harvested, and the luciferase and b-galactosidase activity ligated to the linkers (oJW102, 50-GCGGTGACCCGGGA- were determined using Dual-Lights luciferase and b-galacto- GATCTGAATTC-30; and oJW103, 50-GAATTCAGATC-30) sidase reporter gene assay system (Tropix, Bedford, MA, (Ren et al., 2000). The ligated DNA was subjected to ligation USA) on a luminometer.Relative reporter activities were mediated PCR, labeled with Cy3 and Cy5 dCTP using determined by normalizing the luciferase activity to the BioPrimers DNA labeling kit (Invitrogen, Carlsbad, CA, corresponding b-galactosidase activity. USA), and hybridized to the DNA microarray containing the Synthesized siRNA oligos against b-catenin mRNA (target previously described B5000 unique human promoters (Li sequence: 50-AAAGCUGAUAUUGAUGGACAG-30) and et al., 2003). The immunoprecipitated DNA was subjected to TCF4 mRNA (SMARTpool) were obtained commercially an end point PCR analysis using primers designed to amplify the (Dharmacon RNA Technologies, Lafayette, CO, USA).Cells specific regions of the EDN1 promoter (À7, À1 kb, and TATA cultured in 24-well dish were transfected with double-stranded regions) and a control genomic region (À10 kb upstream of the siRNA oligos (5 nM/well) using Lipofectaminet 2000 transfec- CAD gene).The PCR was performed on a DNA Engine Tetrad tion reagent in triplicate.After 36 h post-transfection, the cell thermal cycler (MJ Research, Waltham, MA, USA) using cycling culture medium was collected for ELISA.Cell extracts were parameters of 5 min at 941Cfollowedby35cyclesof1minat analysed for the effectiveness of RNAi by immunoblotting 941C, 30 s at 601C and 1 min at 721C.The following primer pairs with b-catenin (BD Biosciences Pharmingen, San Diego, CA, were used for the PCR of the ChIP DNA: À7K_Left, 50-CCTTT USA) and TCF4 (Upstate Biotechnology, Charlottesville, VA, GCCACAGCTCCTAAA-30,andÀ7K_Right, 50-ACAATGA USA) antibodies.The expression plasmids encoding b-catenin TTGGGGTGAGGAA-30; À1K_Left, 50-CATCAGCCCCTGT and TCF4 variants were described previously (Roose et al., AGCTCTT-30, and À1K_Right, 50-CATGCCCAGTCTTTTC 1999; van de Wetering et al., 2002). These plasmids were CCTA-30; TATA_Left, 50-GGCAGAGAGCTGTCCAAGT transfected into cells using Lipofectaminet 2000 transfection C-30, and TATA_Right, 50-CGGGTTCCTCAGATCTCAA reagent.After 36 h post-transfection, the cell culture medium

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 603 was collected for ELISA.The secreted EDN1 levels in the The primers, 50-TGGGAAAAAGTGTATTTATCAGCA-30 culture supernatants were determined in triplicate using an and 50-TTTGACGCTGTTTCTCATGG-30 were used to endothelin-1 ELISA kit (R&D Systems, Minneapolis, MN, amplify the EDN1 mRNA.The primers, 5 0-CTGGTGC USA).ELISA was performed according to the instruction TCCATGAGGAGA-30 and 50-CTCTGACCTTTTGCCAG provided by the manufacturer. GAG-30 were used to amplify the MYC mRNA.The primers, 50-GCAAAGACCTGTACGCCAAC-30 and 50-ACACCGAG EMSA TACTTGCGCTCT-30 were used to amplify a reference gene, the gamma actin (ACTG1) mRNA.The quantitative real-time Electrophoresis mobility shift assays were performed using a PCR of each sample was performed in triplicate using iCyclert recombinant GST-TCF4 protein, and doubled stranded oligo and SYBR green iQt SYBR green supermix reagent (Bio-Rad with the following sequences.Double stranded oligos of the Laboratories, Hercules, CA, USA).The threshold cycle (Ct) 0 wild-type TBE4 (5 -GAGGAACCCGCAGCGCTTTGAGG- values were calculated automatically by the iCycle iQt Real- 0 0 GACCTGAAGCTGTTTTTCTTCG-3 ) or muTBE4 (5 -GA Time Dectection System Software (Bio-Rad Laboratories, GGAACCCGCAGCGCCTTAAGGGACCTGAAGCTGTT Hercules, CA, USA).Normalized Ct ( DCt) values for each 0 TTTCTTCG-3 ) from the EDN1 promoter or the wild-type sample were then calculated by subtracting the Ct value 0 TBE1 (5 -CTAGCGCACCTTTGATTTCTGCACCTTTGAT obtained for the ACTG1 gene from the Ct value for the EDN1 TTCTG-30) or muTBE1 (50-CTAGCGCACCTTTGGCTTC (or MYC) gene (DCt ¼ CtEDN1ÀCtACTG1).Using DDCt values TGCACCTTTGGCTTCTG-30) from MYC promoter (Morin (calculated from DDCt ¼ DCtTUMORÀDCtNORMAL) and the et al., 1996) were used for EMSA. The recombinant GST- formula, 2À(DDCt), the relative EDN1 mRNA levels in the tumor TCF4 protein was purified from Escherechia coli.An E. coli (or induced) sample compared to normal (or uninduced) DE3 strain, transformed with the pGEX-TCF4 plasmid (He sample were determined. et al., 1998), was grown to OD600 of 0.5, induced with 1 mM IPTG, and grown for additional 4 h at 301C.The induced cells were harvested by centrifugation, and the cell pellet was Proliferation and FACS analysis washed with PBS.The cells were lysed by sonication in PBS HT29-APC cells were seeded in six-well plates at 1 Â 105 cells/ with 5% glycerol and Completet protease inhibitors.The cell wells in triplicate for each treatment.Next day, the media was lysate was clarified by centrifugation and incubated with 1 ml replaced, and 100 pM EDN1 (EMD Biosciences, San Diego, CA, GST-Sepharose 4B resin (Amersham Biosciences, Piscataway, USA), 100 mM ZnCl2, or combined 100 pM EDN1 and 100 mM NJ, USA) in batch at 41C for 2 h.The resin was transferred to ZnCl2 was added directly to cells.After 72 h post-treatment, cells a disposable Econocolumnt (Bio-Rad Laboratories, Hercules, were carefully washed twice with PBS and fixed using 1% CA, USA) and washed in the column with 20 ml PBS.The paraformaldehyde in PBS for 15 min at room temperature.The bound proteins were eluted with 10 mM glutathione (Amer- fixed cells were washed twice with PBS and were stained with sham Biosciences, Piscataway, NJ, USA), and dialysed in PBS 0.5% crystal violet (w/v in ethanol) for 15 min at room with 10% glycerol.A serial dilution of the GST-TCF protein temperature.The stained cells were thoroughly but gently washed (0.125, 0.25, 0.5 and 1.0 mg) was incubated with 1.5 pmol of with dH2O and dried and visualized by digital photography. oligo in binding buffer (12.5 mM HEPES, pH 8.0, 50 mM NaCl, FACS analyses were performed with the cells at 72 h post- 5mM DTT, 5% glycerol, and 1 mg/ml BSA) for 30 min at treatment, were harvested and fixed with 80% ethanol at 41Cfor room temperature.The reaction mixture was loaded onto 6% 24 h.The fixed cells were treated with 40 mg/ml RNase A and (29 : 1 mono : bis) polyacrylamide gel containing 1 TBE Â stained with 20 mg/ml propidium iodide in dH2O.The stained (89 mM Tris, 89 mM boric acid, and 2 mM EDTA).The cells were analysed by FACS using Becton Dickson FACSort polyacrylamide gel was subjected to electric current of 5 V/ instrument (BD Biosciences Immunocytometry Systems, San cm for 1 h at room temperature, stained with SYBR green I Jose, CA, USA).G1, G2, S, and dead cells were determined from DNA dye (Molecular Probes, Eugene, OR, USA), and the generated FACS dataset for each treatment using ModFit visualized under UV illumination. software (Verity Software House, Topsham, ME, USA).

Quantitative real-time PCR Acknowledgements Total RNA from cultured colon cancer cells and primary We thank Frank Furnari for his assistance with the FACS tissue samples were prepared using Trizols reagent (Invitro- analysis, Bert Vogelstein and Ken Kinzler for providing the gen, Carlsbad, CA, USA).Total RNA from HT29-APC cells HT29-APC cell line and the GST-TCF4 expression construct, was prepared from cells treated with or without 100 mM ZnCl2 Hans Clevers for providing the TCF4 and b-catenin expression for 12 h.Total RNA from HCT116 cells was prepared after constructs, Linda Wasserman for providing the primary treating the cells with 20 mM LiCl for 36 h.Total RNA from human colon samples, Christine Cheng for her help in primary human tissues was prepared by gentle sonication to constructing the mutant EDN1 reporters, and Webster disperse the solid tissue in Trizols reagent.cDNA was Cavenee for his comments on the manuscript.THK was synthesized from 50 mg of total RNA using poly-dT16 primer supported by Ruth L Kirschstein National Research Service and Superscripts II reverse transcriptase (Invitrogen, Carls- Award #1F32CA108313.This work was supported by the bad, CA, USA).After cDNA synthesis, RNase A was added generous funds from PCRP of DOD (BR) Sidney Kimmel to reaction to hydrolyse RNA, and the remaining cDNA was Foundation (BR) Charlotte Geyer Foundation (BR) Charlotte puriﬁed using QIAquicks PCR puriﬁcation kit, and 25 ng of Gayer Foundation (BR) and Ludwig Institute for Cancer cDNA was used as templates for quantitative real-time PCR. Research (BR).

References

Bagnato A, Cirilli A, Salani D, Simeone P, Muller A, Nicotra Bagnato A and Spinella F.(2003). Trends Endocrinol. Metab., MR, Natali PG and Venuti A.(2002). Cancer Res., 62, 14, 44–50. 6381–6384. Bienz M and Clevers H.(2000). Cell, 103, 311–320.

Oncogene b-Catenin activates endothelin-1 in colon cancer cells TH Kim et al 604 Chan TA, Wang Z, Dang LH, Vogelstein B and Kinzler KW. Papandreou CN, Usmani B, Geng Y, Bogenrieder T, Freeman (2002). Proc. Natl. Acad. Sci. USA, 99, 8265–8270. R, Wilk S, Finstad CL, Reuter VE, Powell CT, Scheinberg Cohen Jr MM.(2003). Am. J. Med. Genet., 122A, 303–314. D, Magill C, Scher HI, Albino AP and Nanus DM.(1998). Del Bufalo D, Di Castro V, Biroccio A, Varmi M, Salani D, Nat. Med., 4, 50–57. Rosano L, Trisciuoglio D, Spinella F and Bagnato A. Peduto Eberl L, Bovey R and Juillerat-Jeanneret L.(2003). Br. (2002). Mol. Pharmacol., 61, 524–532. J. Cancer, 88, 788–795. Egidy G, Juillerat-Jeanneret L, Jeannin JF, Korth P, Bosman Pelengaris S, Khan M and Evan G.(2002). Nat. Rev. Cancer, FT and Pinet F.(2000). Am. J. Pathol., 157, 1863–1874. 2, 764–776. Giles RH, van Es JH and Clevers H.(2003). Biochim. Biophys. Ren B, Cam H, Takahashi Y, Volkert T, Terragni J, Young Acta, 1653, 1–24. RA and Dynlacht BD.(2002). Genes Dev., 16, 245–256. Grant K, Loizidou M and Taylor I.(2003). Br. J. Cancer, 88, Ren B, Robert F, Wyrick JJ, Aparicio O, Jennings EG, Simon 163–166. I, Zeitlinger J, Schreiber J, Hannett N, Kanin E, Volkert TL, He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa Wilson CJ, Bell SP and Young RA.(2000). Science, 290, LT, Morin PJ, Vogelstein B and Kinzler KW.(1998). 2306–2309. Science, 281, 1509–1512. Roose J, Huls G, van Beest M, Moerer P, van der Horn K, He X.(2003). Dev. Cell, 4, 791–797. Goldschmeding R, Logtenberg T and Clevers H.(1999). Kedzierski RM and Yanagisawa M.(2001). Annu. Rev. Science, 285, 1923–1926. Pharmacol. Toxicol., 41, 851–876. Rosano L, Spinella F, Salani D, Di Castro V, Venuti A, Kinzler KW and Vogelstein B.(1996). Cell, 87, 159–170. Nicotra MR, Natali PG and Bagnato A.(2003). Cancer Klein PS and Melton DA.(1996). Proc. Natl. Acad. Sci. USA, Res., 63, 2447–2453. 93, 8455–8459. Shichiri M, Sedivy JM, Marumo F and Hirata Y.(1998). Mol. Kurihara Y, Kurihara H, Suzuki H, Kodama T, Maemura K, Endocrinol., 12, 172–180. Nagai R, Oda H, Kuwaki T, Cao WH, Kamada N, Jishage Stambolic V, Ruel L and Woodgett JR.(1996). Curr. Biol., 6, K, Ouchi Y, Azuma S, Toyoda Y, Ishikawa T, Kumada M 1664–1668. and Yazaki Y.(1994). Nature, 368, 703–710. van de Wetering M, Oosterwegel M, Dooijes D and Clevers H. Kusuhara M, Yamaguchi K, Nagasaki K, Hayashi C, Suzaki (1991). EMBO J., 10, 123–132. A, Hori S, Handa S, Nakamura Y and Abe K.(1990). van de Wetering M, Sancho E, Verweij C, de Lau W, Oving I, Cancer Res., 50, 3257–3261. Hurlstone A, van der Horn K, Batlle E, Coudreuse D, Li Z, Van Calcar S, Qu C, Cavenee WK, Zhang MQ and Ren Haramis AP, Tjon-Pon-Fong M, Moerer P, van den Born B.(2003). Proc. Natl. Acad. Sci. USA, 100, 8164–8169. M, Soete G, Pals S, Eilers M, Medema R and Clevers H. Maniatis T.(1999). Genes Dev., 13, 505–510. (2002). Cell, 111, 241–250. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobaya- Vogelstein B and Kinzler KW.(1997). Science, 275, 1787–1790. shi M, Mitsui Y, Yazaki Y, Goto K and Masaki T.(1988). Morin PJ, Vogelstein B and Kinzler KW.(1996). Proc. Natl. Nature, 332, 411–415. Acad. Sci. USA, 93, 7950–7954. Yin JJ, Mohammad KS, Kakonen SM, Harris S, Wu-Wong Nelson J, Bagnato A, Battistini B and Nisen P.(2003). Nat. JR, Wessale JL, Padley RJ, Garrett IR, Chirgwin JM Rev. Cancer, 3, 110–116. and Guise TA.(2003). Proc. Natl. Acad. Sci. USA, 100, Nelson JB.(2003). J. Urol., 170, S65–S67, discussion S67–S68. 10954–10959.

Oncogene