The activity of β-catenin (β-cat), a key component of the Wnt signaling pathway, is deregulated in about 40% of ovarian endometrioid adenocarcinomas (OEAs), usually as a result of CTNNB1 gene mutations. The function of β-cat in neoplastic transformation is dependent on T-cell factor (TCF) transcription factors, but specific genes activated by the interaction of β-cat with TCFs in OEAs and other cancers with Wnt pathway defects are largely unclear. As a strategy to identify β-cat/TCF transcriptional targets likely to contribute to OEA pathogenesis, we used oligonucleotide microarrays to compare gene expression in primary OEAs with mutational defects in β-cat regulation (n = 11) to OEAs with intact regulation of β-cat activity (n = 17). Both hierarchical clustering and principal component analysis based on global gene expression distinguished β-cat-defective tumors from those with intact β-cat regulation. We identified 81 potential β-cat/TCF targets by selecting genes with at least 2-fold increased expression in β-cat-defective versus β-cat regulation-intact tumors and significance in a t test (P < 0.05). Seven of the 81 genes have been previously reported as Wnt/β-cat pathway targets (i.e., BMP4, CCND1, CD44, FGF9, EPHB3, MMP7, and MSX2). Differential expression of several known and candidate target genes in the OEAs was confirmed. For the candidate target genes CST1 and EDN3, reporter and chromatin immunoprecipitation assays directly implicated β-cat and TCF in their regulation. Analysis of presumptive regulatory elements in 67 of the 81 candidate genes for which complete genomic sequence data were available revealed an apparent difference in the location and abundance of consensus TCF-binding sites compared with the patterns seen in control genes. Our findings imply that analysis of gene expression profiling data from primary tumor samples annotated with detailed molecular information may be a powerful approach to identify key downstream targets of signaling pathways defective in cancer cells.

The vast majority of ovarian cancers are derived from epithelial cells. These malignant epithelial tumors (carcinomas) are typically gland-forming and can be divided into four major morphological types, serous, endometrioid, clear cell, and mucinous adenocarcinomas. Recently, we used oligonucleotide microarrays to characterize global gene expression in 113 ovarian adenocarcinomas (1). Our results and those of other studies offer support for the concept that different histological types of ovarian carcinoma likely represent distinct, albeit overlapping, disease entities, with each type exhibiting specific molecular and pathobiological features [reviewed by Feeley and Wells (2) and Aunoble et al.(3)].

OEAs4 are characterized by frequent (16–54%) mutations of CTNNB1, the gene encoding β-cat, a critical component of the Wnt signaling pathway (4, 5, 6, 7, 8, 9). Wnts are a highly conserved family of secreted growth factors that bind members of the frizzled family of transmembrane receptors and, through downstream signaling, modulate many developmental and adult tissue processes including cell fate specification, proliferation, and differentiation (10). β-cat plays a critical role in both cell adhesion and Wnt signaling. The cytoplasmic/nuclear pool of β-cat involved in Wnt signaling is largely regulated by a multiprotein complex consisting of the APC tumor suppressor, AXIN, and GSK3β proteins (11, 12, 13, 14, 15, 16, 17, 18). In the absence of Wnt signals, this protein complex promotes degradation of free cytosolic β-cat via GSK3β-mediated phosphorylation of NH2-terminal β-cat residues and subsequent ubiquitination and degradation of β-cat by the proteasome. Wnt ligands, upon binding to a Frizzled-LRP (lipoprotein-receptor-related protein) transmembrane receptor complex, activate a pathway that inhibits GSK3β activity, with resultant stabilization and nuclear localization of β-cat. Nuclear β-cat cooperates with members of the TCF/lymphoid enhancer factor transcription regulator proteins (hereafter referred to collectively as TCFs) to activate transcription of specific target genes. The Wnt pathway is deregulated in many types of human cancers (17), including melanomas (19), hepatoblastomas (20), medulloblastomas (21) and carcinomas of the colon (22), prostate (23, 24), uterine endometrium (25, 26, 27), and ovary (4, 5, 6, 7, 8, 9). Presumably, many of the proteins encoded by β-cat/TCF transcriptional targets play important roles in effecting neoplastic transformation.

Many of the previous studies aimed at identifying novel Wnt pathway target genes have been carried out in developmental systems or using various in vitro or animal tumor models (28, 29, 30, 31, 32, 33). The relatively few studies based on analysis of primary human cancers have focused mainly on colorectal carcinomas, the majority of which manifest Wnt pathway defects (34, 35, 36). We recently found diverse mechanisms of β-cat deregulation in approximately one-third of primary OEAs, including frequent mutations of CTNNB1 and, less commonly, mutations of APC, AXIN1, and AXIN2(9). Expression of six previously reported β-cat/TCF-regulated genes (CCND1, c-MYC, MMP-7, CX43, ITF2, and PPAR-δ) was subsequently evaluated in OEAs with and without documented Wnt pathway defects (37). All of these genes except c-MYC were significantly increased in expression in OEAs with deregulated β-cat. These studies suggest that comparison of comprehensive gene expression data from primary OEAs with and without deregulated β-cat may provide a robust strategy for identifying as-yet-undetermined β-cat/TCF target genes with roles in the pathogenesis of OEAs and perhaps other types of human cancers with Wnt pathway defects. We applied this approach using gene expression data from oligonucleotide microarrays, and we report here on several novel candidate β-cat/TCF transcriptional targets.

Tumor Samples and RNA Isolation.

Forty-one snap-frozen primary OEAs were analyzed using oligonucleotide microarrays and/or RT-PCR (37 from the Cooperative Human Tissue Network/Gynecologic Oncology Group Tissue Bank, 2 from the University of Michigan Health System, and 2 from the Johns Hopkins Hospital). Gene expression in 28 of the 41 tumors evaluated with Affymetrix HuGeneFL oligonucleotide microarrays has been reported previously (1, 38). Of these 28 OEAs, 11 have a documented Wnt pathway defect and nuclear localization of β-cat protein (9). The subset of 11 Wnt pathway-defective tumors includes 10 tumors (OE-2, -13, -17, -18, -37, -44, -47, -48, -55, and -71) with stabilizing mutation of CTNNB1, and one tumor (OE-32) with biallelic inactivation of APC. The remaining 17 OEAs lacked Wnt pathway defects based on immunohistochemical studies showing lack of nuclear β-cat protein and mutational analyses of CTNNB1, APC, AXIN1, and/or AXIN2(9). Primary tumor tissues were manually microdissected before RNA extraction to ensure each tumor sample contained at least 70% neoplastic cells. Total RNA was extracted from frozen tissue biopsies with Trizol (Life Technologies, Inc., Carlsbad, CA) and then further purified using RNeasy spin columns (Qiagen, Valencia, CA) according to the manufacturers’ protocols. Analysis of tissues from human subjects was approved by the University of Michigan’s Institutional Review Board (IRB-MED #2001-0568).

Data Processing.

Acquisition and processing of oligonucleotide microarray data from the 28 OEAs described above has been reported previously (1). A detailed description of the methods, as well as the freely available code, can be found online.5 The analyses for the present study included data from 6955 noncontrol probe sets (of 7069 available on the chip), after elimination of probe sets that contained fewer than 8 probe pairs (n = 57) and those that were invariant (SD <10−6; n = 57). These probe sets represent approximately 5700 unique genes.

Statistical Analysis.

Gene expression values were log-transformed by logarithm10(max[X +100,0],+100). Two-sample t tests of the log-transformed data were used to compare Wnt pathway-defective with pathway-intact samples. Fold change in gene expression is the ratio (Wnt defect samples:Wnt intact samples) of the relative gene expression values, which were computed as the antilogarithm of the mean log-transformed data. We selected candidate Wnt pathway target genes by demanding that their relative gene expression levels yield P < 0.05 for the t test, as well as giving at least a 2-fold increase in the gene expression ratio.

To view relationships of the samples based on global gene expression, we performed both HC (39) and PCA (40) on the log-transformed data from all 6955 probe sets. Both HC and PCA were performed using S-PLUS 2000 software (MathSoft, Inc., Cambridge, MA). HC is a method that finds relationships between samples based on correlation of their gene expression and then displays the sample relationships in a taxonomy-like dendrogram. PCA contracts multidimensional gene expression data into ranked statistically independent projections, called components. The first PC captures the greatest fraction of variance from the expression data compared to any other projection, whereas the second PC captures the greatest variance remaining in the data, independent of the first projection, and so on. Any two PCs can be used to plot a two-dimensional view of the multidimensional gene expression data such that samples located close to each other have more similar gene expression than samples that are further apart.

Search for Consensus TCF Sites in Candidate Wnt Pathway Target Genes.

We searched for consensus TCF sites (5′-WWCAAWG-3′), as defined by Roose and Clevers (41), in selected regions of putative Wnt pathway target genes in silico. Similar studies were carried out on control genes insensitive to Wnt pathway status, i.e., genes with a geometric mean gene expression ratio of 1 ± 0.01 when Wnt pathway-deregulated samples were compared with Wnt pathway-intact samples. Genomic sequences of only fully mapped genes were downloaded from the June 2002 version of the University of California at Santa Cruz working draft of the human genome sequence,6 with highly repetitive sequences masked. Pointers to the transcription start and stop sites provided within the RefSeq portion of the University of California at Santa Cruz human genome browser database were used to extract specific gene sequences for analysis (42). We obtained sequence starting from 5000 bp upstream (5′) of the start of transcription to 1000 bp downstream (3′) from the stop of translation. For genes with multiple transcripts, only the longest transcript was used for TCF site annotation. In-house software was used to extract genomic DNA sequences and to identify consensus TCF sites within them. We compared the proportion of sites (the number of sites within a gene region divided by the total number of possible sites within the same gene region) within candidate Wnt pathway genes and control genes using a one-sided standard normal Z test. The ratio of the proportions of sites in candidate target genes relative to control genes was estimated, and 95% confidence intervals were computed using the delta method.

PCR-based Gene Expression Expression Analyses.

We used real-time q-RT-PCR analysis to validate differential expression of selected putative Wnt pathway target genes in RNA samples from 31 primary OEAs, including 18 of the 28 OEAs analyzed by oligonucleotide microarrays. Of these 31 tumors, 14 had documented β-cat deregulation, and 17 had intact β-cat/TCF signaling. The detailed methods and sequences of forward primer (f), reverse primer (r), and probe (p) for the genes validated by q-RT-PCR [CNND1 (cyclin D1), FGF9 (fibroblast growth factor 9), EDN3 (endothelin 3), SFN (stratifin), and HPRT1, which served as an internal control] are available online.5 Differences between tumors with intact versus deregulated TCF/β-cat were tested using the Student’s t test. Pearson product-moment correlations were used to estimate the degree of association between the microarray and q-RT-PCR data. Semiquantitative multiplex RT-PCR for CST1 was performed as follows. Each PCR reaction contained 2 μCi of [α-32P]dCTP (ICN, Costa Mesa, CA); 25 μm dCTP; 200 μm dATP, dGTP, and dDTP; 1.5 μm of each primer (both target gene and internal control; see website5 for primer sequences); 1 unit of platinum Taq polymerase (Invitrogen, Carlsbad, CA); and 20 ng of first-strand cDNA. HPRT1 was used as the reference gene. Reactions were performed at 95°C for 3 min; denaturing at 95°C for 30 s, annealing at 58°C for 30 s, and elongation at 72°C for 45 s for 28 cycles; followed by 7 min of extension. PCR products were resolved on 6% denaturing polyacrylamide gels. After vacuum drying, gels were exposed to phosphorimager screens (Amersham Biosciences Corp., Piscataway, NJ). To determine the ratio of target gene amplification in tumors, the values from the target gene (CST1) were normalized with the values of internal control gene (HPRT1) from each sample. All reactions were performed at least twice.

Plasmids.

Expression constructs for the mutant form of β-cat (codon 33 substitution of tyrosine for serine, S33Y) and dn-TCF-4 (TCF-4r N31) have been described previously (43). The reporter constructs pTOPFLASH, which contains three copies of an optimal TCF/lymphoid enhancer factor binding motif (CCTTTGATC), and pFOPFLASH, which contains three copies of a mutant motif (CCTTTGGCC), were generously provided by Bert Vogelstein (Johns Hopkins University) and have been described previously (44). pCH110 (Amersham, Arlington Heights, IL) contains a functional LacZ gene cloned downstream of a CMV early region promoter-enhancer element. DNA fragments containing human CST1 and EDN3 promoter sequences were obtained by PCR amplification of genomic DNA, using primers for CST1 (accession number AL591074) and EDN3 (accession number AL035250) from sequences in GenBank. PCR fragments were cloned into pCR-Blunt II-TOPO (Invitrogen) and then subcloned upstream of the luciferase reporter gene in the pGL3Basic vector (Promega, Madison, WI). Mutations of presumptive TCF-binding sites in the CST1 and EDN3 promoters were obtained in vitro by PCR-based mutagenesis using primers containing the desired mutations. All construct sequences were confirmed by automated sequencing of double-stranded DNA templates.

Cell Culture.

The cells lines 293 (transformed human kidney epithelium), DLD-1 (colon cancer), SW480 (colon cancer), and TOV112D (OEA) were obtained from the American Type Culture Collection (Manassas, VA). Cell lines with stable expression of dn-TCF-4 (TCF-4ΔN31/pPGS-CMV-CITE-neo), DLD-1/dn-TCF, SW480/dn-TCF, and TOV112D/dn-TCF or empty vector (pPGS-CMV-CITE-neo) were generated as described previously (45). All cell lines were grown in DMEM containing 10% fetal bovine serum and penicillin/streptomycin (Life Technologies, Inc., Gaithersburg, MD).

Luciferase Reporter Gene Assays.

Reporter assays for β-cat/TCF-mediated transcriptional activity were performed in triplicate as described previously (9, 46). Briefly, cells were plated into 6-well dishes and then transfected with various constructs using FuGENE 6 reagent (Roche, Indianapolis, IN). Cells were co-transfected with 0.5 μg of pCH110 (Amersham) to determine transfection efficiency based on β-galactosidase expression. After 48 h of incubation at 37°C, transfected cells were washed with PBS and lysed with reporter lysis buffer (Promega). Luciferase assay reagent (Promega) was added to the cell lysates, and then luciferase activity was measured with a luminometer (model TD-20E; Turner Corp., Mountain View, CA). To correct for variation in transfection efficiency, luciferase activity was normalized to β-galactosidase activity, which was measured using the β-galactosidase enzyme assay system (Promega) and a microplate reader. To assess the response of human CST1 and EDN3 promoter sequences to β-cat, 293 cells were cotransfected with pCH110, 0.2 μg of pcDNA3/S33Y, and 0.5 μg of various CST1 or EDN3 reporter constructs (see Figs. 3,A and 4 A) or 0.5 μg of the pGL3Basic vector as a negative control. As a positive control for β-cat activation of a TCF responsive promoter, pTOPFlash or pFOPFlash was cotransfected into 293 cells. Reporter gene assays using dn-TCF-mediated transcriptional repression were performed in triplicate as described previously (45). Cell lines stably expressing dn-TCF (pPGS-CMV-CITE-TCF-4r N31) or empty vector (pPGS-CMV-CITE-neo) were cotransfected (0.5 μg) with reporter constructs CST1-A, EDN3-A, or pGL3Basic empty vector as a negative control. Transfection efficiency was assessed as described above. The total mass of transfected DNA in each well was kept constant by adding empty vector plasmid DNA. Fold activation of TCF transcriptional activity was calculated by normalizing the luciferase activity to the empty vector (pGL3Basic) control after adjusting for transfection efficiency.

ChIP Assay.

ChIP assays were performed according to the manufacturer’s protocol (kit 17-925; Upstate Biotechnology, Lake Placid, NY), with some modifications as reported previously (47). Briefly, 1% formaldehyde was added to the culture medium of 2 × 106 adherent cells for 10 min at 37°C. Cells were washed twice with ice-cold PBS (Life Technologies, Inc.) containing proteinase inhibitors. Cells were collected and suspended in 200 μl of lysis buffer containing proteinase inhibitors, and lysed cells were sonicated to yield 200-1000-bp DNA fragments. After centrifugation to eliminate cell debris, the samples were diluted 1:10 in ChIP dilution buffer containing proteinase inhibitors and 0.1 volume of the diluted sample was removed and saved to assess input DNA. Salmon sperm DNA/protein A-agarose (80 μl) was added to 1.8 ml of the diluted sample and gently mixed for 30 min at 4°C to reduce nonspecific binding. Then, 4 μl of polyclonal anti-β-cat antibody (06-734; Upstate Biotechnology) was added to the precleared ChIP solution and incubated overnight at 4°C with gentle agitation, followed by the addition of the salmon sperm DNA/protein A-agarose slurry for 1 h. Immunoprecipitates were washed, and the DNA-protein complexes were eluted as reported previously (47). NaCl (final concentration, 200 mm) was added, and the samples were incubated at 65°C for 4 h to reverse the formaldehyde cross-link. After treatment with proteinase K, DNA was extracted from the samples with phenol/chloroform and precipitated with ethanol. DNA pellets were suspended in 40 μl of H2O. The primer pairs used to amplify EDN3 promoter sequences and “irrelevant” DNA downstream and upstream of the EDN3 promoter from the immunoprecipitated chromatin are specified on our website.5

Wnt Pathway Status Is a Major Determinant of Global Gene Expression in OEAs.

We generated gene expression profiles of 28 primary OEAs using Affymetrix HuGeneFL oligonucleotide microarrays. To visualize relationships of tumor samples based on global gene expression, both HC and PCA of the microarray data were performed. The analyses included data from nearly all of the probe sets on the chip (6955 of 7069 noncontrol probe sets as described in “Materials and Methods”) to obtain views of gene expression unbiased by preselection of genes. Fig. 1,A shows a HC dendrogram. Notably, all of the OEAs with deregulated β-cat are found on one branch of the primary division between the two main clusters. PCA provides an alternative method for viewing relationships between tumors based on multidimensional gene expression data. We have plotted the first two PCs, which represented 23% of the variation in gene expression, in Fig. 1 B. Tumor samples with deregulated β-cat occupy a space that minimally overlaps the area defined by OEAs without β-cat deregulation. The observation that the first PC for this large gene collection is strongly associated with Wnt pathway status was unexpected and suggests that this feature is a critical, but not sole, determinant of gene expression in OEAs.

Identification of Candidate Wnt Pathway Target Genes in OEAs.

We required that candidate Wnt/β-cat pathway target genes meet both of the following criteria: (a) expression value increased at least 2-fold in OEAs with deregulated β-cat compared with OEAs with apparently intact β-cat regulation; and (b) significant Student’s t test (P < 0.05). Eighty-one genes satisfying these criteria were identified (Table 1). We performed a randomization procedure to demonstrate that 81 genes is more than could be expected by chance alone. Specifically, the Wnt pathway status labels were randomly permuted across the 28 OEA samples 2000 times, and in each case, the number of genes in the randomized data that satisfied our gene selection criteria was determined. In the randomized data, the maximum number of genes satisfying our criteria was 65, the minimum number was 0, and the average number was 6.9. Compared with the number of genes found by chance alone, the observed number (n = 81) of candidate genes was highly significant (P < 0.0005).

To assess whether the 81 genes are more highly associated with each other in the tumors with deregulated β-cat than the other group, we performed a permutation procedure. We randomly selected 81 genes from the set of 6954 genes and calculated the pairwise correlation (Spearman’s rho) coefficient between every pair of 81 genes. We counted the proportion of these correlation coefficients that had an absolute value greater than 0.3. This procedure was repeated 400 times. In the observed data, this proportion is 0.47 for the Wnt pathway-defective group and 0.29 for the Wnt pathway-intact group. In the permutations, the range of proportions was 0.35–0.45 in the Wnt pathway-defective group and 0.25–0.32 in the Wnt pathway-intact group. This demonstrates that these 81 genes are more highly associated with each other in tumors with deregulated β-cat than in tumors with intact β-cat regulation.

Importantly, 7 (Table 1, bold text) of the 81 candidate Wnt pathway target genes have been reported previously as potential β-cat/TCF transcriptional targets in other systems, namely, BMP4(48), CCND1(29, 30),, CD44(49), EPHB3(50), FGF9(51), MMP7(28, 52), and MSX2(53). To the best of our knowledge, the oligonucleotide microarray used for this study includes probe sets for 34 previously implicated Wnt pathway target genes (represented by 55 probe sets) reported to date (for a list of these genes and references, see our website5 specified in “Materials and Methods”). These previously reported target genes were identified through our review of the literature and from those listed on the Nusse laboratory website (Stanford University).7 Because 7 of the 34 putative Wnt pathway target genes were found on our list of 81 candidate target genes, our list is highly enriched with presumptive Wnt pathway targets (P < 0.00001, Fisher’s exact test). Notably absent from the list, among others, is c-MYC, which has been reported to be a Wnt pathway target gene in another system (31, 54), but not in OEAs (37).

Localization of Consensus TCF-binding Sites in Candidate β-Cat/TCF Target Genes.

At least some of the genes with increased expression in OEAs with deregulated β-cat might be expected to be direct targets of β-cat/TCF-mediated transactivation. Presumably, this occurs through sequence-specific binding of the β-cat/TCF complex to TCF recognition sites in key regulatory elements, perhaps including those at or near the genes’ proximal promoters. Based on these considerations, we hypothesized that if a significant proportion of the genes on our list were direct β-cat/TCF transcriptional targets, then these genes might show differences in TCF-binding site number and/or distribution when compared with control genes (i.e., genes with invariant expression in OEAs with respect to β-cat regulation status). Consequently, we obtained publicly available genomic sequence spanning 5000 bp upstream of the inferred start of transcription to 1000 bp downstream of the translation stop codon for genes in Table 1 as well as 100 randomly selected control genes. We were able to acquire full genomic sequences for 67 of the 81 candidate target genes. We then compared the proportion of consensus TCF-binding sites between the candidate target genes and control genes in various regions, based on annotation in the RefSeq database (Table 2). In candidate β-cat/TCF target genes, we found a significantly higher proportion of consensus TCF-binding sites in the region from −501 to −1000 bp with respect to the transcriptional start site and in the 5′-UTR. Notably, a recent study has shown the presence of functionally relevant TCF-binding sites outside the promoter (i.e., in the first intron) of the β-cat/TCF target gene, AXIN2(55). Our results show that certain presumptive regulatory regions of our candidate Wnt/β-cat target genes have more consensus TCF-binding sites compared with control genes. As such, the data offer further evidence that our list of candidate target genes contains many genes likely to be directly regulated by β-cat/TCF.

Validation of Microarray Data.

Five of the most highly up-regulated candidate genes shown in Table 1 [namely, cyclinD1 (CCND1), cystatin 1/cystatinSN (CST1),8 endothelin 3 (EDN3), fibroblast growth factor 9 (FGF9), and stratifin (SFN, “14-3-3 sigma”)] were selected for analysis. Semiquantitative multiplex RT-PCR (CST1) and real-time q-RT-PCR (CCND1, EDN3, FGF9, and SFN) were used to validate the microarray data (Fig. 2). The microarray data (Fig. 2, A, C, E, G, and I) reveal, not unexpectedly, some heterogeneity of candidate gene expression within the two groups of tumors. Nonetheless, it can be seen that the majority of samples with deregulated β-cat express these genes at much higher levels than the majority of the samples with an intact Wnt pathway and, importantly, that the same conclusions can be reached from the RT-PCR data (Fig. 2, B, D, F, H, and J, respectively). Moreover, for all of the genes, the RT-PCR data were highly correlated (P < 0.01) with the microarray data (r = 0.9, 0.77, 0.87, 0.65, and 0.58, respectively) as estimated by the 18 samples included in both the PCR and microarray experiments. Validation experiments were conducted only on the five genes specified above. Hence, we found no examples of discordance between the microarray and q-RT-PCR data.

The CST1 and EDN3 Proximal Promoters Are Responsive to β-Cat/TCF.

We sought to demonstrate that some of the novel genes on our target gene list were responsive to β-cat/TCF. Two genes, CST1 and EDN3, were selected for further evaluation. Analysis of the sequence of the CST1 (NT_011387) and EDN3 (NT_011362) promoter regions obtained via NCBI Entrez Nucleotide view revealed several potential TCF-binding sites within 2000 bp upstream of the genes’ purported transcriptional start sites. We used PCR to clone promoter sequences from both CST1 and EDN3, and sequence analysis of the cloned fragments confirmed the presence of the potential TCF-binding sites, which are shown schematically in Figs. 3,A and 4,A. To determine whether these minimal promoter sequences were responsive to β-cat, we generated luciferase reporter gene constructs containing the CST1 and EDN3 promoter fragments. For both genes, constructs containing mutations or deletions of the putative TCF recognition sites were also generated (Figs. 3,A and 4,A). We found that a mutant constitutively active form of β-cat (S33Y mutant), when coexpressed with a CST1 or EDN3 reporter construct in 293 cells, strongly stimulated reporter activity (Figs. 3,B and 4,B). Mutation of all four potential TCF-binding sites in the CST1 promoter (CST1-Am1–4) completely abrogated the response to β-cat (Fig. 3,B). The other mutation and deletion constructs all showed reduced luciferase reporter activity compared with the construct containing the intact CST1 promoter fragment. Similar studies using mutated and deleted versions of the EDN3 promoter fragment suggest that the TCF-binding site at −1614 is most critical for imparting TCF responsiveness to the EDN3 promoter (Fig. 4 B). These data indicate that the proximal promoters of CST1 and EDN3 are responsive to β-cat/TCF and support the notion that both genes are directly activated by deregulated β-cat upon its interaction with TCF in OEAs with Wnt pathway defects.

The DLD-1 and SW480 colorectal cancer-derived and TOV112D OEA-derived cell lines have deregulated β-cat, as a result of biallelic inactivation of APC (DLD-1 and SW480) or activating mutation of CTNNB1 (TOV112D). To further demonstrate that the CST1 as well as EDN3 promoters are responsive to TCF/β-cat, we established cell lines that stably express a dn-mutant form of TCF (dn-TCF), which should diminish transcriptional activity of TCF-responsive promoters. Transcriptional activity, as measured by luciferase activation of CST1-A (Fig. 3,C) and EDN3-A (Fig. 4 C) reporter constructs, was reduced in the presence of dn-TCF in all three cell lines. These experiments provide additional evidence that proximal promoter sequences of CST1 and EDN3 are indeed TCF/β-cat responsive, consistent with the view that CST1 and EDN3 are likely to be directly regulated by β-cat and TCF in OEAs.

ChIP Assays Demonstrate Interaction of β-Cat with an EDN3 Promoter Region Containing Presumptive TCF-binding Sites.

To document direct interaction of β-cat with sequences in the 5′-flanking region of EDN3, we performed ChIP assays using an antibody directed against β-cat and chromatin obtained from SW480/neo or SW480/dn-TCF cells. The positions of the various primers used to amplify and detect specific regions of immunoprecipitated EDN3 promoter fragments are shown in Fig. 4,A. TCF-binding site-containing DNA fragments from the EDN3 promoter were readily recovered using primers flanking the two putative TCF-binding sites at −1693 and −1614 and, as expected, were reduced in the presence of dn-TCF (Fig. 4,D). Moreover, irrelevant upstream and downstream DNA fragments lacking TCF-binding sites were not recovered after ChIP with β-cat (Fig. 4 C). These findings support results from the reporter assays suggesting a role for the TCF-binding site at −1614 in mediating TCF responsiveness of the EDN3 promoter fragment and further bolster the case that EDN3 is a direct transcriptional target of β-cat/TCF.

The Wnt signaling pathway is frequently deregulated in certain types of human cancer. In tumors with Wnt pathway defects, stabilized β-cat interacts with TCF transcription factors to mediate increased expression of specific genes, at least some which likely play critical roles in cancer pathogenesis. Not surprisingly, identification of downstream targets of activated β-cat has become a subject of intense interest. Recent studies have identified several putative targets of β-cat signaling, primarily through analysis of cultured cells or animal model systems. We have used a novel strategy to discover potential β-cat/TCF target genes, comparing gene expression in otherwise similar primary tumors with and without Wnt/β-cat pathway defects. Our results suggest that the approach is a robust one for identifying novel candidate genes worthy of further investigation as direct β-cat/TCF transcriptional targets with functional roles in cancer development or progression. This type of approach may be more likely to yield useful insights into cancer pathogenesis than simple comparisons of gene expression in tumor cells with their normal cell counterparts, given the substantial variation in many phenotypic features between neoplastic and nonneoplastic cells. Moreover, as is the case for epithelial ovarian cancer (56), the cell of origin for many cancers may be unclear or difficult to obtain in sufficient quantities for global gene expression profiling. By comparing a sufficiently large number of otherwise similar tumors with and without defects in a particular signaling pathway, nonspecific variations in gene expression are minimized, and a significant proportion of the variation under study may be due to the consequences of the disrupted signaling pathway under scrutiny.

Other signaling pathways, such as those mediated by PTEN and, to a lesser degree, KRAS, are likely to be aberrant in some of our OEAs, based on the reported frequency of PTEN and K-RAS mutations in this tumor type (57, 58). Clearly, defects in signaling pathways other than Wnt will contribute to variation of gene expression in individual tumors. Nonetheless, our global gene expression analysis provides strong evidence that Wnt pathway status, in and of itself, is a major determinant of global gene expression in OEAs. In the future, it will be interesting to survey genes in other pathways that may modulate the one mediated by APC/β-cat/TCF. For example, the PI3K/Akt pathway has been shown to modulate β-cat/TCF-mediated gene expression in prostate cancer cells through phosphorylation and inhibition of GSK3β, a downstream substrate of PI3K/Akt, resulting in stabilization of β-cat (59). PTEN, an inhibitor of the PI3K/Akt pathway, may inhibit nuclear accumulation of β-cat in certain settings (60).

We recognize that some of the candidate Wnt/β-cat pathway target genes shown in Table 1 may not be direct targets. However, 2 of the top 10 genes on our candidate list not previously suggested as Wnt/β-cat pathway targets, namely, CST1 and EDN3, were shown to have β-cat/TCF-responsive upstream regulatory sequences. The protein product encoded by CST1, cystatin 1 (Cst1), is one of several salivary cysteine protease inhibitors and a member of a cysteine protease inhibitor superfamily (61). Very little is known about any role Cst1 may have in cancer. Based on serial analysis of gene expression (SAGE), CST1 was determined to be up-regulated in a metastatic colorectal cancer-derived cell line compared with a cell line derived from the matched primary tumor (62). Members of the cystatin superfamily have also been associated with metastasis in other studies (63). However, in B-16 melanoma cells, overexpression of cystatin C significantly reduced metastatic ability (64). Although present in saliva, in vitro, Cst1 was not found to inhibit the activities of the ubiquitously expressed lysosomal cathepsins B, H, and L (65), which are host cysteine proteases that have been shown to play a significant role in the proteolytic events causing periodontal tissue destruction (66, 67) and are also involved in tissue remodeling and turnover of the extracellular matrix, immune system modulation, and cancer (68, 69). What role, if any, CST1 plays in the pathogenesis of OEA will need to be further explored.

EDN3 (endothelin 3) is one of six genes that, when mutated, are known to be associated with Hirschsprung’s disease, a genetic disorder of neural crest development characterized by the absence of intramural ganglion cells in the hindgut (70). Edn3 is a member of the endothelin/sarafotoxin family of proteins and is a secreted vasoactive peptide that binds to the endothelin B receptor and is essential in the development of neural crest-derived cell lineages (71). No evidence currently exists that links Edn3 with cancer.

We are currently undertaking additional studies to determine the functional significance of CST1, EDN3, and selected other novel Wnt pathway target genes in OEA pathogenesis. Such studies are necessary to help distinguish those genes that play critical versus nominal roles in cellular transformation. Moreover, we recognize that Wnt pathway activation may result in the transcriptional deregulation of different genes in different tumor types with β-cat regulation defects (e.g., ovarian endometrioid versus colorectal carcinomas). Clearly, much remains to be learned about the identity and role of various downstream targets of the β-cat/TCF signaling pathway in human cancer. Nevertheless, the studies presented here represent a novel line of attack for addressing some of the many unanswered questions in the field.

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.

1

Supported by funds from the Department of Defense (DAMD 17-1-1-0727), the National Cancer Institute (U19 CA84953, RO1 CA94172, and RO1 CA85463), and in part by the Tissue Core of the University of Michigan Comprehensive Cancer Center (NIH P30 CA46952).

4

The abbreviations used are: OEA, ovarian endometrioid adenocarcinoma; TCF, T-cell factor; APC, adenomatous polyposis coli; GSK3β, glycogen synthase kinase 3β; RT-PCR, reverse transcription-PCR; q-RT-PCR, quantitative RT-PCR; HC, hierarchical clustering; PCA, principal component analysis; PC, principal component; β-cat, β-catenin; CMV, cytomegalovirus; dn-, dominant negative; ChIP, chromatin immunoprecipitation; PI3K, phosphatidylinositol 3′-kinase; UTR, untranslated region.

5

http://dot.ped.med.umich.edu:2000/pub/Ovary/index.html.

6

http://genome.ucsc.edu.

7

http://www.stanford.edu/∼rnusse/wntwindow.html.

8

The HuGeneFL microarray does not contain a probe set specific for CST1. We determined that the probe set for CST4 (“X54667_s_at”) was primarily measuring CST1, and not CST4 gene expression in our OEAs [see the website specified in “Materials and Methods” (footnote 5)].

Fig. 1.

Tumor relationships based on global gene expression data from HuGeneFL oligonucleotide microarrays. A, dendrogram of a HC analysis using average linkage. B, PCA showing the first two PCs. The Wnt/β-catenin pathway status of each tumor is annotated as indicated at the bottom of the figure.

Fig. 1.

Tumor relationships based on global gene expression data from HuGeneFL oligonucleotide microarrays. A, dendrogram of a HC analysis using average linkage. B, PCA showing the first two PCs. The Wnt/β-catenin pathway status of each tumor is annotated as indicated at the bottom of the figure.

Close modal
Fig. 2.

Comparison of CCND1, FGF9, CST1, EDN3, and SFN gene expression from microarray and RT-PCR analyses. For each gene, relative expression based on the microarray data (A, C, E, G, and I) is shown, and mean fold gene expression (normalized to HPRT) based on the q-RT-PCR (B, D, H, and J) or multiplex-RT-PCR (F) is shown. For q-RT-PCR, 17 and 14 OEAs with an intact or deregulated Wnt pathway, respectively, were used in the analysis, and for multiplex-RT-PCR, 9 and 10 OEAs with an intact or deregulated Wnt pathway, respectively, were analyzed. Expression differences for CCND1 (P < 0.001), FGF9 (P < 0.026), CST1 (P < 0.00003), EDN3 (P < 0.016), and SFN (P < 0.0004), validated by RT-PCR between tumors with intact versus deregulated β-cat/TCF signaling, were readily apparent.

Fig. 2.

Comparison of CCND1, FGF9, CST1, EDN3, and SFN gene expression from microarray and RT-PCR analyses. For each gene, relative expression based on the microarray data (A, C, E, G, and I) is shown, and mean fold gene expression (normalized to HPRT) based on the q-RT-PCR (B, D, H, and J) or multiplex-RT-PCR (F) is shown. For q-RT-PCR, 17 and 14 OEAs with an intact or deregulated Wnt pathway, respectively, were used in the analysis, and for multiplex-RT-PCR, 9 and 10 OEAs with an intact or deregulated Wnt pathway, respectively, were analyzed. Expression differences for CCND1 (P < 0.001), FGF9 (P < 0.026), CST1 (P < 0.00003), EDN3 (P < 0.016), and SFN (P < 0.0004), validated by RT-PCR between tumors with intact versus deregulated β-cat/TCF signaling, were readily apparent.

Close modal
Fig. 3.

Upstream regulatory sequences of CST1 are TCF responsive. A, schematic diagram of the human CST1 minimal promoter indicating the locations of the transcriptional start, canonical TCF-binding sites, and extent of sequences cloned into pGL3Basic luciferase reporter vector. Designations of the nine CST1 reporter constructs and the status [wild-type or mutant (m)] of candidate TCF-binding sites in the constructs are indicated. The bold/underlined nucleotides within candidate TCF-binding sites indicate the position of the mutations in the mutant reporter constructs. B, effects of β-cat on human CST1 promoter reporter constructs in 293 cells. The relative activity of the CST1 constructs shown in A was assessed after transient transfection of the cells with either pcDNA3 (empty vector) or pcDNA3 construct containing constitutively active β-cat(S33Y). C, cell lines expressing dn-TCF were transiently cotransfected with CST1-A promoter construct (or empty vector) and pCH110. Luciferase activity was normalized to β-galactosidase activity and reported as the relative fold activation in transcriptional activity compared with empty vector (pGL3Basic) control.

Fig. 3.

Upstream regulatory sequences of CST1 are TCF responsive. A, schematic diagram of the human CST1 minimal promoter indicating the locations of the transcriptional start, canonical TCF-binding sites, and extent of sequences cloned into pGL3Basic luciferase reporter vector. Designations of the nine CST1 reporter constructs and the status [wild-type or mutant (m)] of candidate TCF-binding sites in the constructs are indicated. The bold/underlined nucleotides within candidate TCF-binding sites indicate the position of the mutations in the mutant reporter constructs. B, effects of β-cat on human CST1 promoter reporter constructs in 293 cells. The relative activity of the CST1 constructs shown in A was assessed after transient transfection of the cells with either pcDNA3 (empty vector) or pcDNA3 construct containing constitutively active β-cat(S33Y). C, cell lines expressing dn-TCF were transiently cotransfected with CST1-A promoter construct (or empty vector) and pCH110. Luciferase activity was normalized to β-galactosidase activity and reported as the relative fold activation in transcriptional activity compared with empty vector (pGL3Basic) control.

Close modal
Fig. 4.

TCF elements in the promoter of human EDN3 help regulate transcriptional activity and directly interact with β-cat. A, schematic diagram of the human EDN3 minimal promoter indicating locations of the transcriptional start, canonical TCF-binding sites, and extent of sequences cloned into pGL3Basic luciferase reporter vector. Designations of the six EDN3 reporter constructs and the status [wild-type or mutant (m)] of candidate TCF-binding sites in the constructs are indicated. The bold/underlined nucleotides within candidate TCF-binding sites indicate the position of the mutations in the mutant reporter constructs. B, effects of β-cat on human EDN3 promoter reporter constructs in 293 cells. The relative activity of the EDN3 constructs shown in A was assessed after transient transfection of the cells with either pcDNA3 (empty vector) or pcDNA3 construct containing constitutively active β-cat(S33Y). C, cell lines expressing dn-TCF were transiently cotransfected with EDN3-A pGL3Basic luciferase reporter construct (or empty vector) and pCH110. Luciferase activity was normalized to β-galactosidase activity and reported as the relative fold activation in transcriptional activity compared with empty vector (pGL3Basic) control. D, ChIP assay. Cross-linked DNA from SW480/neo and SW480/dn-TCF cell lysates were combined with antibody against β-cat (or no antibody), and the immunoprecipitate was subjected to PCR to amplify various regions of the EDN3 promoter.

Fig. 4.

TCF elements in the promoter of human EDN3 help regulate transcriptional activity and directly interact with β-cat. A, schematic diagram of the human EDN3 minimal promoter indicating locations of the transcriptional start, canonical TCF-binding sites, and extent of sequences cloned into pGL3Basic luciferase reporter vector. Designations of the six EDN3 reporter constructs and the status [wild-type or mutant (m)] of candidate TCF-binding sites in the constructs are indicated. The bold/underlined nucleotides within candidate TCF-binding sites indicate the position of the mutations in the mutant reporter constructs. B, effects of β-cat on human EDN3 promoter reporter constructs in 293 cells. The relative activity of the EDN3 constructs shown in A was assessed after transient transfection of the cells with either pcDNA3 (empty vector) or pcDNA3 construct containing constitutively active β-cat(S33Y). C, cell lines expressing dn-TCF were transiently cotransfected with EDN3-A pGL3Basic luciferase reporter construct (or empty vector) and pCH110. Luciferase activity was normalized to β-galactosidase activity and reported as the relative fold activation in transcriptional activity compared with empty vector (pGL3Basic) control. D, ChIP assay. Cross-linked DNA from SW480/neo and SW480/dn-TCF cell lysates were combined with antibody against β-cat (or no antibody), and the immunoprecipitate was subjected to PCR to amplify various regions of the EDN3 promoter.

Close modal
Table 1

Candidate Wnt pathway target genes in OEA

Affymetrix probe setGene symbolDescriptionRelative mean gene expressionPLocusLinkRefa
Intact Wnt pathwayDeregulated Wnt pathwayRatio
X54667_s_at CST4 Cystatin S 927.1 9065.9 9.8 0.0000 1472  
U23070_at NMA Putative transmembrane protein 346.8 2835.7 8.2 0.0000 25805  
X57348_s_at SFN Stratifin 483.0 3918.3 8.1 0.0000 2810  
M81883_at GAD1 Glutamate decarboxylase 1 (brain, 67 kDa) 320.0 2480.9 7.8 0.0000 2571  
U65932_at ECM1 Extracellular matrix protein 1 463.5 2712.3 5.9 0.0000 1893  
D89377_s_at MSX2 msh (Drosophila) homeo box homologue 2              a 246.6 1425.0 5.8 0.0000 4488  53  
M31682_at INHBB Inhibin, β B (activin AB β polypeptide) 137.8 770.4 5.6 0.0001 3625  
X52001_at EDN3 Endothelin 3 209.1 1122.6 5.4 0.0000 1908  
L22524_s_at MMP7 Matrix metalloproteinase 7 (matrilysin) 1267.0 6059.6 4.8 0.0066 4316 52,28 
U14528_at SLC26A2 Solute carrier family 26 (sulfate transporter), member 2 476.7 2263.6 4.7 0.0000 1836  
S79219_s_at PCCA Propionyl coenzyme A carboxylase, α polypeptide 1225.0 5661.3 4.6 0.0000 5095  
D14838_at FGF9 Fibroblast growth factor 9 423.1 1950.8 4.6 0.0000 2254  51  
K01396_at SERPINA1 Serine (or cys) proteinase inhibitor, clade A, member 1 1661.8 7340.6 4.4 0.0007 5265  
M68516_rna1_at SERPINA5 Serine (or cys) proteinase inhibitor, clade A, member 5 705.4 2973.1 4.2 0.0060 5104  
X68742_at ITGA1 Integrin, α 1 448.4 1847.2 4.1 0.0000 3672  
M21389_at KRT5 Keratin 5 (epidermolysis bullosa simplex) 425.5 1688.1 4.0 0.0068 3852  
HG880-HT880_s_at MUC6 Mucin 6, gastric 130.3 512.2 3.9 0.0038 4588  
L38517_at IHH Indian hedgehog (Drosophila) homologue 176.0 675.2 3.8 0.0005 3549  
M55593_at MMP2 Matrix metalloproteinase 2 (gelatinase A) 1057.0 3934.7 3.7 0.0000 4313  
U33147_at MGB1 Mammaglobin 1 570.3 2117.6 3.7 0.0009 4250  
J05257_at DPEP1 Dipeptidase 1 (renal) 202.0 731.4 3.6 0.0038 1800  
X91117_rna1_at SLC6A2 Solute carrier family 6 (noradrenalin transporter), member 2 265.6 923.9 3.5 0.0000 6530  
U53506_at DIO2 Deiodinase, iodothyronine, type II 109.4 376.0 3.4 0.0004 1734  
X51441_s_at SAA1 Serum amyloid A1 302.1 1004.6 3.3 0.0311 6288  
X67697_at SPAG11 Sperm associated antigen 11 406.9 1342.0 3.3 0.0016 10407  
X61118_rna1_at LMO2 LIM domain only 2 (rhombotin-like 1) 469.9 1507.7 3.2 0.0000 4005  
HG2981-HT3127_s_at CD44 CD44 antigen (Indian blood group system) 396.9 1239.1 3.1 0.0000 960  49  
X59798_at CCND1 Cyclin D1 (PRAD1: parathyroid adenomatosis 1) 1390.7 4318.4 3.1 0.0081 893 29,30 
X16354_at CEACAM1 Carcinoembryonic antigen-related cell adhesion molecule 1 589.7 1734.8 2.9 0.0010 634  
L13286_at CYP24 Cytochrome P450, subfamily 24 (vitamin D 24-hydroxylase) 211.3 613.9 2.9 0.0001 1591  
J02611_at APOD Apolipoprotein D 787.4 2285.9 2.9 0.0013 347  
M21624_at TRD@ T cell receptor Δ locus 366.1 1057.5 2.9 0.0005 6964  
D50840_at UGCG UDP-glucose ceramide glucosyltransferase 1613.0 4520.1 2.8 0.0001 7357  
U59321_at DDX17 DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptide 17 427.9 1177.4 2.8 0.0056 10521  
X13255_at DBH Dopamine β-hydroxylase 203.2 526.3 2.6 0.0029 1621  
X68733_rna1_at SERPINA3 Serine (or cysteine) proteinase inhibitor, clade A, 3 1172.8 3010.1 2.6 0.0153 12  
X78706_at CRAT Carnitine acetyltransferase 872.0 2213.1 2.5 0.0035 1384  
AB000220_at SEMA3C Sema domain, immunoglobulin domain, (semaphorin) 3C 1517.5 3840.1 2.5 0.0038 10512  
M31994_at ALDH1 Aldehyde dehydrogenase 1, soluble 1522.8 3809.8 2.5 0.0128 216  
X07730_at KLK3 Kallikrein 3, (prostate specific antigen) 545.9 1362.6 2.5 0.0105 354  
M57710_at LGALS3 Lectin, galactoside-binding, soluble, 3 (galectin 3) 7561.3 18451.6 2.4 0.0006 3958  
L27479_at X123 Friedreich ataxia region gene X123 205.3 499.9 2.4 0.0020 9413  
D42073_at RCN1 Reticulocalbin 1, EF-hand calcium binding domain 1952.3 4737.2 2.4 0.0003 5954  
M14539_at F13A1 Coagulation factor XIII, A1 polypeptide 549.7 1327.9 2.4 0.0098 2162  
S39329_at KLK2 Kallikrein 2, prostatic 193.8 464.0 2.4 0.0204 3817  
HG1067-HT1067_r_at NULL Homo sapiens clone 24747 mRNA sequence 125.2 298.4 2.4 0.0287 NULL  
U53347_at SLC1A5 Solute carrier family 1 (neutral A.A. transporter), 5 1495.1 3556.5 2.4 0.0009 6510  
X68314_at GPX2 Glutathione peroxidase 2 (gastrointestinal) 245.7 582.3 2.4 0.0047 2877  
U83115_at AIM1 Absent in melanoma 1 1147.9 2685.3 2.3 0.0153 202  
M22490_at BMP4 Bone morphogenetic protein 4 840.0 1950.7 2.3 0.0002 652  48  
S62539_at IRS1 Insulin receptor substrate 1 525.5 1217.1 2.3 0.0003 3667  
L11708_at HSD17B2 Hydroxysteroid (17-β) dehydrogenase 2 267.9 616.5 2.3 0.0003 3294  
M27492_at IL1R1 Interleukin 1 receptor, type I 1579.7 3624.7 2.3 0.0019 3554  
X76717_at MT1L Metallothionein 1L 2966.1 6801.2 2.3 0.0020 4500  
M35252_at TM4SF3 Transmembrane 4 superfamily member 3 549.6 1254.8 2.3 0.0478 7103  
D64109_at TOB2 Transducer of ERBB2, 2 1082.9 2458.5 2.3 0.0007 10766  
M97925_rna1_at DEFA5 Defensin, α 5, Paneth cell-specific 515.6 1166.1 2.3 0.0018 1670  
U90911_at NULL H. sapiens cDNA: FLJ23260 fis, clone COL05804 1311.3 2946.8 2.2 0.0005 NULL  
HG2755-HT2862_at PLS3 Plastin 3 (T isoform) 838.7 1846.4 2.2 0.0276 5358  
J03779_at MME Membrane metallo-endopeptidase (neutral, CD10) 520.3 1129.1 2.2 0.0032 4311  
X75208_at EPHB3 EphB3 1070.5 2304.6 2.2 0.0001 2049  50  
U46689_at ALDH10 Aldehyde dehydrogenase 10 (fatty aldehyde dehydrogenase) 630.0 1353.6 2.1 0.0000 224  
U73799_at NULL H. sapiens mRNA, (from clone DKFZp434B1620) 115.0 245.0 2.1 0.0014 NULL  
X16832_at NULL H. sapiens cDNA: FLJ22499 fis, similar to Hu cathepsin H 2622.9 5520.0 2.1 0.0111 NULL  
X69111_at ID3 Inhibitor of DNA binding 3, dominant negative H-L-H protein 1670.5 3503.4 2.1 0.0014 3399  
HG2167-HT2237_at LBC Lymphoid blast crisis oncogene 1455.2 3028.9 2.1 0.0000 3928  
U07807_at NULL H. sapiens pseudogene for metallothionein and AG/CT repeat 196.8 405.5 2.1 0.0001 NULL  
M94151_at CTNNA2 Catenin (cadherin-associated protein), α 2 308.1 634.7 2.1 0.0154 1496  
U35048_at TSC22 Transforming growth factor β-stimulated protein TSC-22 3889.6 7986.1 2.1 0.0008 8848  
J04164_at IFITM1 Interferon induced transmembrane protein 1 (9-27) 16201.5 33164.5 2.0 0.0471 8519  
X81892_at GPR64 G protein-coupled receptor 64 353.8 723.9 2.0 0.0313 10149  
M14745_at BCL2 B-cell CLL/lymphoma 2 662.6 1355.5 2.0 0.0020 596  
U52828_s_at CTNND2 Catenin, Δ 2 (neural plakophilin-related arm-repeat protein) 335.2 684.4 2.0 0.0012 1501  
X71125_at QPCT Glutaminyl-peptide cyclotransferase (glutaminyl cyclase) 303.5 619.4 2.0 0.0001 25797  
D83735_at CNN2 Calponin 2 1416.7 2882.3 2.0 0.0047 1265  
U43148_at PTCH Patched (Drosophila) homologue 404.5 822.9 2.0 0.0004 5727  
U73960_at ARL4 ADP-ribosylation factor-like 4 725.2 1473.4 2.0 0.0003 10124  
M80482_at PACE4 Paired basic amino acid cleaving system 4 891.8 1800.0 2.0 0.0018 5046  
M34455_at INDO Indoleamine-pyrrole 2,3 dioxygenase 1402.3 2812.3 2.0 0.0249 3620  
D86980_at KIAA0227 KIAA0227 protein 288.0 576.6 2.0 0.0000 23508  
D87258_at PRSS11 Protease, serine, 11 (IGF binding) 2996.2 5994.7 2.0 0.0127 5654  
Affymetrix probe setGene symbolDescriptionRelative mean gene expressionPLocusLinkRefa
Intact Wnt pathwayDeregulated Wnt pathwayRatio
X54667_s_at CST4 Cystatin S 927.1 9065.9 9.8 0.0000 1472  
U23070_at NMA Putative transmembrane protein 346.8 2835.7 8.2 0.0000 25805  
X57348_s_at SFN Stratifin 483.0 3918.3 8.1 0.0000 2810  
M81883_at GAD1 Glutamate decarboxylase 1 (brain, 67 kDa) 320.0 2480.9 7.8 0.0000 2571  
U65932_at ECM1 Extracellular matrix protein 1 463.5 2712.3 5.9 0.0000 1893  
D89377_s_at MSX2 msh (Drosophila) homeo box homologue 2              a 246.6 1425.0 5.8 0.0000 4488  53  
M31682_at INHBB Inhibin, β B (activin AB β polypeptide) 137.8 770.4 5.6 0.0001 3625  
X52001_at EDN3 Endothelin 3 209.1 1122.6 5.4 0.0000 1908  
L22524_s_at MMP7 Matrix metalloproteinase 7 (matrilysin) 1267.0 6059.6 4.8 0.0066 4316 52,28 
U14528_at SLC26A2 Solute carrier family 26 (sulfate transporter), member 2 476.7 2263.6 4.7 0.0000 1836  
S79219_s_at PCCA Propionyl coenzyme A carboxylase, α polypeptide 1225.0 5661.3 4.6 0.0000 5095  
D14838_at FGF9 Fibroblast growth factor 9 423.1 1950.8 4.6 0.0000 2254  51  
K01396_at SERPINA1 Serine (or cys) proteinase inhibitor, clade A, member 1 1661.8 7340.6 4.4 0.0007 5265  
M68516_rna1_at SERPINA5 Serine (or cys) proteinase inhibitor, clade A, member 5 705.4 2973.1 4.2 0.0060 5104  
X68742_at ITGA1 Integrin, α 1 448.4 1847.2 4.1 0.0000 3672  
M21389_at KRT5 Keratin 5 (epidermolysis bullosa simplex) 425.5 1688.1 4.0 0.0068 3852  
HG880-HT880_s_at MUC6 Mucin 6, gastric 130.3 512.2 3.9 0.0038 4588  
L38517_at IHH Indian hedgehog (Drosophila) homologue 176.0 675.2 3.8 0.0005 3549  
M55593_at MMP2 Matrix metalloproteinase 2 (gelatinase A) 1057.0 3934.7 3.7 0.0000 4313  
U33147_at MGB1 Mammaglobin 1 570.3 2117.6 3.7 0.0009 4250  
J05257_at DPEP1 Dipeptidase 1 (renal) 202.0 731.4 3.6 0.0038 1800  
X91117_rna1_at SLC6A2 Solute carrier family 6 (noradrenalin transporter), member 2 265.6 923.9 3.5 0.0000 6530  
U53506_at DIO2 Deiodinase, iodothyronine, type II 109.4 376.0 3.4 0.0004 1734  
X51441_s_at SAA1 Serum amyloid A1 302.1 1004.6 3.3 0.0311 6288  
X67697_at SPAG11 Sperm associated antigen 11 406.9 1342.0 3.3 0.0016 10407  
X61118_rna1_at LMO2 LIM domain only 2 (rhombotin-like 1) 469.9 1507.7 3.2 0.0000 4005  
HG2981-HT3127_s_at CD44 CD44 antigen (Indian blood group system) 396.9 1239.1 3.1 0.0000 960  49  
X59798_at CCND1 Cyclin D1 (PRAD1: parathyroid adenomatosis 1) 1390.7 4318.4 3.1 0.0081 893 29,30 
X16354_at CEACAM1 Carcinoembryonic antigen-related cell adhesion molecule 1 589.7 1734.8 2.9 0.0010 634  
L13286_at CYP24 Cytochrome P450, subfamily 24 (vitamin D 24-hydroxylase) 211.3 613.9 2.9 0.0001 1591  
J02611_at APOD Apolipoprotein D 787.4 2285.9 2.9 0.0013 347  
M21624_at TRD@ T cell receptor Δ locus 366.1 1057.5 2.9 0.0005 6964  
D50840_at UGCG UDP-glucose ceramide glucosyltransferase 1613.0 4520.1 2.8 0.0001 7357  
U59321_at DDX17 DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptide 17 427.9 1177.4 2.8 0.0056 10521  
X13255_at DBH Dopamine β-hydroxylase 203.2 526.3 2.6 0.0029 1621  
X68733_rna1_at SERPINA3 Serine (or cysteine) proteinase inhibitor, clade A, 3 1172.8 3010.1 2.6 0.0153 12  
X78706_at CRAT Carnitine acetyltransferase 872.0 2213.1 2.5 0.0035 1384  
AB000220_at SEMA3C Sema domain, immunoglobulin domain, (semaphorin) 3C 1517.5 3840.1 2.5 0.0038 10512  
M31994_at ALDH1 Aldehyde dehydrogenase 1, soluble 1522.8 3809.8 2.5 0.0128 216  
X07730_at KLK3 Kallikrein 3, (prostate specific antigen) 545.9 1362.6 2.5 0.0105 354  
M57710_at LGALS3 Lectin, galactoside-binding, soluble, 3 (galectin 3) 7561.3 18451.6 2.4 0.0006 3958  
L27479_at X123 Friedreich ataxia region gene X123 205.3 499.9 2.4 0.0020 9413  
D42073_at RCN1 Reticulocalbin 1, EF-hand calcium binding domain 1952.3 4737.2 2.4 0.0003 5954  
M14539_at F13A1 Coagulation factor XIII, A1 polypeptide 549.7 1327.9 2.4 0.0098 2162  
S39329_at KLK2 Kallikrein 2, prostatic 193.8 464.0 2.4 0.0204 3817  
HG1067-HT1067_r_at NULL Homo sapiens clone 24747 mRNA sequence 125.2 298.4 2.4 0.0287 NULL  
U53347_at SLC1A5 Solute carrier family 1 (neutral A.A. transporter), 5 1495.1 3556.5 2.4 0.0009 6510  
X68314_at GPX2 Glutathione peroxidase 2 (gastrointestinal) 245.7 582.3 2.4 0.0047 2877  
U83115_at AIM1 Absent in melanoma 1 1147.9 2685.3 2.3 0.0153 202  
M22490_at BMP4 Bone morphogenetic protein 4 840.0 1950.7 2.3 0.0002 652  48  
S62539_at IRS1 Insulin receptor substrate 1 525.5 1217.1 2.3 0.0003 3667  
L11708_at HSD17B2 Hydroxysteroid (17-β) dehydrogenase 2 267.9 616.5 2.3 0.0003 3294  
M27492_at IL1R1 Interleukin 1 receptor, type I 1579.7 3624.7 2.3 0.0019 3554  
X76717_at MT1L Metallothionein 1L 2966.1 6801.2 2.3 0.0020 4500  
M35252_at TM4SF3 Transmembrane 4 superfamily member 3 549.6 1254.8 2.3 0.0478 7103  
D64109_at TOB2 Transducer of ERBB2, 2 1082.9 2458.5 2.3 0.0007 10766  
M97925_rna1_at DEFA5 Defensin, α 5, Paneth cell-specific 515.6 1166.1 2.3 0.0018 1670  
U90911_at NULL H. sapiens cDNA: FLJ23260 fis, clone COL05804 1311.3 2946.8 2.2 0.0005 NULL  
HG2755-HT2862_at PLS3 Plastin 3 (T isoform) 838.7 1846.4 2.2 0.0276 5358  
J03779_at MME Membrane metallo-endopeptidase (neutral, CD10) 520.3 1129.1 2.2 0.0032 4311  
X75208_at EPHB3 EphB3 1070.5 2304.6 2.2 0.0001 2049  50  
U46689_at ALDH10 Aldehyde dehydrogenase 10 (fatty aldehyde dehydrogenase) 630.0 1353.6 2.1 0.0000 224  
U73799_at NULL H. sapiens mRNA, (from clone DKFZp434B1620) 115.0 245.0 2.1 0.0014 NULL  
X16832_at NULL H. sapiens cDNA: FLJ22499 fis, similar to Hu cathepsin H 2622.9 5520.0 2.1 0.0111 NULL  
X69111_at ID3 Inhibitor of DNA binding 3, dominant negative H-L-H protein 1670.5 3503.4 2.1 0.0014 3399  
HG2167-HT2237_at LBC Lymphoid blast crisis oncogene 1455.2 3028.9 2.1 0.0000 3928  
U07807_at NULL H. sapiens pseudogene for metallothionein and AG/CT repeat 196.8 405.5 2.1 0.0001 NULL  
M94151_at CTNNA2 Catenin (cadherin-associated protein), α 2 308.1 634.7 2.1 0.0154 1496  
U35048_at TSC22 Transforming growth factor β-stimulated protein TSC-22 3889.6 7986.1 2.1 0.0008 8848  
J04164_at IFITM1 Interferon induced transmembrane protein 1 (9-27) 16201.5 33164.5 2.0 0.0471 8519  
X81892_at GPR64 G protein-coupled receptor 64 353.8 723.9 2.0 0.0313 10149  
M14745_at BCL2 B-cell CLL/lymphoma 2 662.6 1355.5 2.0 0.0020 596  
U52828_s_at CTNND2 Catenin, Δ 2 (neural plakophilin-related arm-repeat protein) 335.2 684.4 2.0 0.0012 1501  
X71125_at QPCT Glutaminyl-peptide cyclotransferase (glutaminyl cyclase) 303.5 619.4 2.0 0.0001 25797  
D83735_at CNN2 Calponin 2 1416.7 2882.3 2.0 0.0047 1265  
U43148_at PTCH Patched (Drosophila) homologue 404.5 822.9 2.0 0.0004 5727  
U73960_at ARL4 ADP-ribosylation factor-like 4 725.2 1473.4 2.0 0.0003 10124  
M80482_at PACE4 Paired basic amino acid cleaving system 4 891.8 1800.0 2.0 0.0018 5046  
M34455_at INDO Indoleamine-pyrrole 2,3 dioxygenase 1402.3 2812.3 2.0 0.0249 3620  
D86980_at KIAA0227 KIAA0227 protein 288.0 576.6 2.0 0.0000 23508  
D87258_at PRSS11 Protease, serine, 11 (IGF binding) 2996.2 5994.7 2.0 0.0127 5654  
a

Bold text indicates genes that have been previously reported as potential β-cat/TCF transcriptional targets in other systems.

Table 2

The distribution of consensus TCF (5′-WWCAAWG-3′) sites differs between candidate Wnt pathway target genes and control genes

Gene regionUp-regulateda (ur), n = 67Controlb (c), n = 100Pur/Pc95% Confidence intervalPf
NcTdPureNTPcLowerUpper
−5000:transcript-start 303 335000 9.044 413 500000 8.30 1.09 .93 1.26 0.115 
−500:transcript-start 32 33500 9.55 42 50000 8.40 1.14 .61 1.66 0.292 
−1000:−501 41 33500 12.239 40 50000 8.00 1.53 .86 2.20 0.027              g 
−1500:−1001 33 33500 9.851 44 50000 8.80 1.12 .61 1.62 0.312 
−2000:−1501 24 33500 7.164 53 50000 10.60 0.68 .35 1.00 0.946 
−3000:−2001 53 67000 7.91 90 100000 9.00 0.88 .58 1.18 0.772 
−4000:−3001 53 67000 7.91 65 100000 6.50 1.22 .78 1.66 0.144 
−5000:−4001 67 67000 10.00 79 100000 7.90 1.27 .85 1.68 0.077 
5′-UTR 203 162313 12.51 281 289367 9.70 1.29 1.06 1.52 0.003 
3′-UTR 112 95710 11.702 289 245374 11.80 0.99 .78 1.21 0.523 
−1000 from translation-stop 72 67000 10.746 135 100000 13.50 0.80 .57 1.02 0.942 
translation-stop:+500 43 33500 12.836 82 50000 16.40 0.78 .49 1.07 0.904 
+501:1000 29 33500 8.657 53 50000 10.60 0.82 .45 1.19 0.810 
Gene regionUp-regulateda (ur), n = 67Controlb (c), n = 100Pur/Pc95% Confidence intervalPf
NcTdPureNTPcLowerUpper
−5000:transcript-start 303 335000 9.044 413 500000 8.30 1.09 .93 1.26 0.115 
−500:transcript-start 32 33500 9.55 42 50000 8.40 1.14 .61 1.66 0.292 
−1000:−501 41 33500 12.239 40 50000 8.00 1.53 .86 2.20 0.027              g 
−1500:−1001 33 33500 9.851 44 50000 8.80 1.12 .61 1.62 0.312 
−2000:−1501 24 33500 7.164 53 50000 10.60 0.68 .35 1.00 0.946 
−3000:−2001 53 67000 7.91 90 100000 9.00 0.88 .58 1.18 0.772 
−4000:−3001 53 67000 7.91 65 100000 6.50 1.22 .78 1.66 0.144 
−5000:−4001 67 67000 10.00 79 100000 7.90 1.27 .85 1.68 0.077 
5′-UTR 203 162313 12.51 281 289367 9.70 1.29 1.06 1.52 0.003 
3′-UTR 112 95710 11.702 289 245374 11.80 0.99 .78 1.21 0.523 
−1000 from translation-stop 72 67000 10.746 135 100000 13.50 0.80 .57 1.02 0.942 
translation-stop:+500 43 33500 12.836 82 50000 16.40 0.78 .49 1.07 0.904 
+501:1000 29 33500 8.657 53 50000 10.60 0.82 .45 1.19 0.810 
a

Genes from Table 1 for which complete genomic sequence was available.

b

Control genes that have a gene expression ratio of 1 ± 0.01 (Wnt deregulated/Wnt intact).

c

N, number of consensus (5′-WWCAAWG-3′) TCF sites within a specified gene region.

d

T, total number of possible sites of a 7-mer sequence within the specified gene region.

e

P, proportion of consensus TCF sites within the specified up-regulated (Pur) or control (Pc) gene region (N/T) × 104.

f

P is from one-sided Z test.

g

Bold indicates significance.

1
Schwartz D. R., Kardia S. L., Shedden K. A., Kuick R., Michailidis G., Taylor J. M., Misek D. E., Wu R., Zhai Y., Darrah D. M., Reed H., Ellenson L. H., Giordano T. J., Fearon E. R., Hanash S. M., Cho K. R. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas.
Cancer Res.
,
62
:
4722
-4729,  
2002
.
2
Feeley K. M., Wells M. Precursor lesions of ovarian epithelial malignancy.
Histopathology
,
38
:
87
-95,  
2001
.
3
Aunoble B., Sanches R., Didier E., Bignon Y. J. Major oncogenes and tumor suppressor genes involved in epithelial ovarian cancer.
Int. J. Oncol.
,
16
:
567
-576,  
2000
.
4
Palacios J., Gamallo C. Mutations in the β-catenin gene (CTNNB1) in endometrioid ovarian carcinomas.
Cancer Res.
,
58
:
1344
-1347,  
1998
.
5
Gamallo C., Palacios J., Moreno G., Calvo de Mora J., Suarez A., Armas A. β-Catenin expression pattern in stage I and II ovarian carcinomas: relationship with β-catenin gene mutations, clinicopathological features, and clinical outcome.
Am. J. Pathol.
,
155
:
527
-536,  
1999
.
6
Wright K., Wilson P., Morland S., Campbell I., Walsh M., Hurst T., Ward B., Cummings M., Chenevix-Trench G. β-Catenin mutation and expression analysis in ovarian cancer: exon 3 mutations and nuclear translocation in 16% of endometrioid tumours.
Int. J. Cancer
,
82
:
625
-629,  
1999
.
7
Sagae S., Kobayashi K., Nishioka Y., Sugimura M., Ishioka S., Nagata M., Terasawa K., Tokino T., Kudo R. Mutational analysis of β-catenin gene in Japanese ovarian carcinomas: frequent mutations in endometrioid carcinomas.
Jpn. J. Cancer Res.
,
90
:
510
-515,  
1999
.
8
Saegusa M., Okayasu I. Frequent nuclear β-catenin accumulation and associated mutations in endometrioid-type endometrial and ovarian carcinomas with squamous differentiation.
J. Pathol.
,
194
:
59
-67,  
2001
.
9
Wu R., Zhai Y., Fearon E. R., Cho K. R. Diverse mechanisms of β-catenin deregulation in ovarian endometrioid adenocarcinomas.
Cancer Res.
,
61
:
8247
-8255,  
2001
.
10
Cadigan K. M., Nusse R. Wnt signaling: a common theme in animal development.
Genes Dev.
,
11
:
3286
-3305,  
1997
.
11
Rubinfeld B., Souza B., Albert I., Muller O., Chamberlain S. H., Masiarz F. R., Munemitsu S., Polakis P. Association of the APC gene product with β-catenin.
Science (Wash. DC)
,
262
:
1731
-1734,  
1993
.
12
Su L-K., Vogelstein B., Kinzler K. W. Association of the APC tumor suppressor protein with catenins.
Science (Wash. DC)
,
262
:
1734
-1737,  
1993
.
13
Hart M. J., de los Santos R., Albert I. N., Rubinfeld B., Polakis P. Downregulation of β-catenin by human Axin and its association with the APC tumor suppressor, β-catenin and GSK3 β.
Curr. Biol.
,
8
:
573
-581,  
1998
.
14
Munemitsu S., Albert I., Souza B., Polakis P. Regulation of intracellular β-catenin levels by the adenomatous polyposis coli (APC) tumor-suppressor protein.
Proc. Natl. Acad. Sci. USA
,
92
:
3046
-3050,  
1995
.
15
Yamamoto H., Kishida S., Uochi T., Ikeda S., Koyama S., Asashima M., Kikuchi A. Axil, a member of the Axin family, interacts with both glycogen synthase kinase 3β and β-catenin and inhibits axis formation of Xenopus embryos.
Mol. Cell. Biol.
,
18
:
2867
-2875,  
1998
.
16
Nakamura T., Hamada F., Ishidate T., Anai K., Kawahara K., Toyoshima K., Akiyama T. Axin, an inhibitor of the Wnt signalling pathway, interacts with β-catenin, GSK-3β and APC and reduces the β-catenin level.
Genes Cells
,
3
:
395
-403,  
1998
.
17
Polakis P. Wnt signaling and cancer.
Genes Dev.
,
14
:
1837
-1851,  
2000
.
18
Peifer M., Polakis P. Wnt signaling in oncogenesis and embryogenesis: a look outside the nucleus.
Science (Wash. DC)
,
287
:
1606
-1609,  
2000
.
19
Rubinfeld B., Robbins P. D., El-Gamil M., Albert I., Porfiri E., Polakis P. Stabilization of β-catenin by genetic defects in melanoma cell lines.
Science (Wash. DC)
,
275
:
1790
-1792,  
1997
.
20
Koch A., Denkhaus D., Albrecht S., Leuschner I., von Schweinitz D., Pietsch T. Childhood hepatoblastomas frequently carry a mutated degradation targeting box of the β-catenin gene.
Cancer Res.
,
59
:
269
-273,  
1999
.
21
Zurawel R. H., Chiappa S. A., Allen C., Raffel C. Sporadic medulloblastomas contain oncogenic β-catenin mutations.
Cancer Res.
,
58
:
896
-899,  
1998
.
22
Morin P. J., Sparks A. B., Korinek V., Barker N., Clevers H., Vogelstein B., Kinzler K. W. Activation of β-catenin-Tcf signaling in colon cancer by mutations in β-catenin or APC.
Science (Wash. DC)
,
275
:
1787
-1790,  
1997
.
23
Voeller H. J., Truica C. I., Gelmann E. P. β-Catenin mutations in human prostate cancer.
Cancer Res.
,
58
:
2520
-2523,  
1998
.
24
Chesire D. R., Ewing C. M., Sauvageot J., Bova G. S., Isaacs W. B. Detection and analysis of β-catenin mutations in prostate cancer.
Prostate
,
45
:
323
-334,  
2000
.
25
Fukuchi T., Sakamoto M., Tsuda H., Maruyama K., Nozawa S., Hirohashi S. β-Catenin mutation in carcinoma of the uterine endometrium.
Cancer Res.
,
58
:
3526
-3528,  
1998
.
26
Mirabelli-Primdahl L., Gryfe R., Kim H., Millar A., Luceri C., Dale D., Holowaty E., Bapat B., Gallinger S., Redston M. β-Catenin mutations are specific for colorectal carcinomas with microsatellite instability but occur in endometrial carcinomas irrespective of mutator pathway.
Cancer Res.
,
59
:
3346
-3351,  
1999
.
27
Schlosshauer P. W., Pirog E. C., Levine R. L., Hedrick-Ellenson L. Mutational analysis of the CTNNB1 and APC genes in uterine endometrioid carcinoma.
Mod. Pathol.
,
13
:
1066
-1071,  
2000
.
28
Crawford H. C., Fingleton B. M., Rudolph-Owen L. A., Goss K. J., Rubinfeld B., Polakis P., Matrisian L. M. The metalloproteinase matrilysin is a target of β-catenin transactivation in intestinal tumors.
Oncogene
,
18
:
2883
-2891,  
1999
.
29
Shtutman M., Zhurinsky J., Simcha I., Albanese C., D’Amico M., Pestell R., Ben-Ze’ev A. The cyclin D1 gene is a target of the β-catenin/LEF-1 pathway.
Proc. Natl. Acad. Sci. USA
,
96
:
5522
-5527,  
1999
.
30
Tetsu O., McCormick F. β-Catenin regulates expression of cyclin D1 in colon carcinoma cells.
Nature (Lond.)
,
398
:
422
-426,  
1999
.
31
He T. C., Sparks A. B., Rago C., Hermeking H., Zawel L., da Costa L. T., Morin P. J., Vogelstein B., Kinzler K. W. Identification of c-MYC as a target of the APC pathway.
Science (Wash. DC)
,
281
:
1509
-1512,  
1998
.
32
Carnac G., Kodjabachian L., Gurdon J. B., Lemaire P. The homeobox gene Siamois is a target of the Wnt dorsalisation pathway and triggers organiser activity in the absence of mesoderm.
Development (Cambr.)
,
122
:
3055
-3065,  
1996
.
33
Danielian P. S., McMahon A. P. Engrailed-1 as a target of the Wnt-1 signalling pathway in vertebrate midbrain development.
Nature (Lond.)
,
383
:
332
-334,  
1996
.
34
Fujita M., Furukawa Y., Tsunoda T., Tanaka T., Ogawa M., Nakamura Y. Up-regulation of the ectodermal-neural cortex 1 (ENC1) gene, a downstream target of the β-catenin/T-cell factor complex, in colorectal carcinomas.
Cancer Res.
,
61
:
7722
-7726,  
2001
.
35
Rockman S. P., Currie S. A., Ciavarella M., Vincan E., Dow C., Thomas R. J., Phillips W. A. Id2 is a target of the β-catenin/TCF pathway in colon carcinoma.
J. Biol. Chem.
,
276
:
45113
-45119,  
2001
.
36
Pennica D., Swanson T. A., Welsh J. W., Roy M. A., Lawrence D. A., Lee J., Brush J., Taneyhill L. A., Deuel B., Lew M., Watanabe C., Cohen R. L., Melhem M. F., Finley G. G., Quirke P., Goddard A. D., Hillan K. J., Gurney A. L., Botstein D., Levine A. J. WISP genes are members of the connective tissue growth factor family that are up-regulated in wnt-1-transformed cells and aberrantly expressed in human colon tumors.
Proc. Natl. Acad. Sci. USA
,
95
:
14717
-14722,  
1998
.
37
Zhai Y., Wu R., Schwartz D. R., Darrah D., Reed H., Kolligs F. T., Nieman M. T., Fearon E. R., Cho K. R. Role of β-catenin/T-cell factor-regulated genes in ovarian endometrioid adenocarcinomas.
Am. J. Pathol.
,
160
:
1229
-1238,  
2002
.
38
Giordano T. J., Shedden K. A., Schwartz D. R., Kuick R., Taylor J. M., Lee N., Misek D. E., Greenson J. K., Kardia S. L., Beer D. G., Rennert G., Cho K. R., Gruber S. B., Fearon E. R., Hanash S. Organ-specific molecular classification of primary lung, colon, and ovarian adenocarcinomas using gene expression profiles.
Am. J. Pathol.
,
159
:
1231
-1238,  
2001
.
39
Everitt B. .
Cluster Analysis
, 2nd ed. Halsted New York  
1980
.
40
Johnson R., Wichern D. W. Principal components.
Applied multivariate statistical analysis
, 4th ed.
458
-513, Prentice Hall Upper Saddle River, NJ  
1998
.
41
Roose J., Clevers H. TCF transcription factors: molecular switches in carcinogenesis.
Biochim. Biophys. Acta
,
1424
:
M23
-M37,  
1999
.
42
Kent W. J., Sugnet C. W., Furey T. S., Roskin K. M., Pringle T. H., Zahler A. M., Haussler D. The human genome browser at UCSC.
Genome Res.
,
12
:
996
-1006,  
2002
.
43
Kolligs F. T., Hu G., Dang C. V., Fearon E. R. Neoplastic transformation of RK3E by mutant β-catenin requires deregulation of Tcf/Lef transcription but not activation of c-myc expression.
Mol. Cell. Biol.
,
19
:
5696
-5706,  
1999
.
44
Korinek V., Barker N., Morin P. J., van Wichen D., De Weger R. A., Kinzler K. W., Vogelstein B., Clevers H. Constitutive transcriptional activation by a β-catenin-Tcf complex in APC−/− colon carcinoma.
Science (Wash. DC)
,
275
:
1784
-1789,  
1998
.
45
Kolligs F. T., Nieman M. T., Winer I., Hu G., Van Mater D., Feng Y., Smith I. M., Wu R., Zhai Y., Cho K. R., Fearon E. R. ITF-2, a downstream target of the Wnt/TCF pathway, is activated in human cancers with β-catenin defects and promotes neoplastic transformation.
Cancer Cell
,
1
:
145
-155,  
2002
.
46
Kolligs F. T., Kolligs B., Hajra K. M., Hu G., Tani M., Cho K. R., Fearon E. R. γ-Catenin is regulated by the APC tumor suppressor and its oncogenic activity is distinct from that of β-catenin.
Genes Dev.
,
14
:
1319
-1331,  
2000
.
47
Hinoi T., Lucas P. C., Kuick R., Hanash S., Cho K. R., Fearon E. R. CDX2 regulates liver intestine-cadherin expression in normal and malignant colon epithelium and intestinal metaplasia.
Gastroenterology
,
123
:
1565
-1577,  
2002
.
48
Kim J. S., Crooks H., Dracheva T., Nishanian T. G., Singh B., Jen J., Waldman T. Oncogenic β-catenin is required for bone morphogenetic protein 4 expression in human cancer cells.
Cancer Res.
,
62
:
2744
-2748,  
2002
.
49
Wielenga V. J., Smits R., Korinek V., Smit L., Kielman M., Fodde R., Clevers H., Pals S. T. Expression of CD44 in Apc and Tcf mutant mice implies regulation by the WNT pathway.
Am. J. Pathol.
,
154
:
515
-523,  
1999
.
50
Batlle E., Henderson J. T., Beghtel H., van den Born M. M., Sancho E., Huls G., Meeldijk J., Robertson J., van de Wetering M., Pawson T., Clevers H. β-Catenin and TCF mediate cell positioning in the intestinal epithelium by controlling the expression of EphB/ephrinB.
Cell
,
111
:
251
-263,  
2002
.
51
Imai K. S., Satoh N., Satou Y. Early embryonic expression of FGF4/6/9 gene and its role in the induction of mesenchyme and notochord in Ciona savignyi embryos.
Development (Cambr.)
,
129
:
1729
-1738,  
2002
.
52
Brabletz T., Jung A., Dag S., Hlubek F., Kirchner T. β-Catenin regulates the expression of the matrix metalloproteinase-7 in human colorectal cancer.
Am. J. Pathol.
,
155
:
1033
-1038,  
1999
.
53
Willert J., Epping M., Pollack J. R., Brown P. O., Nusse R. A transcriptional response to Wnt protein in human embryonic carcinoma cells.
BMC Dev. Biol.
,
2
:
8
2002
.
54
Romanov S. R., Kozakiewicz B. K., Holst C. R., Stampfer M. R., Haupt L. M., Tlsty T. D. Normal human mammary epithelial cells spontaneously escape senescence and acquire genomic changes.
Nature (Lond.)
,
409
:
633
-637,  
2001
.
55
Jho E. H., Zhang T., Domon C., Joo C. K., Freund J. N., Costantini F. Wnt/β-catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of the signaling pathway.
Mol. Cell. Biol.
,
22
:
1172
-1183,  
2002
.
56
Dubeau L. The cell of origin of ovarian epithelial tumors and the ovarian surface epithelium dogma: does the emperor have no clothes?.
Gynecol. Oncol.
,
72
:
437
-442,  
1999
.
57
Obata K., Morland S. J., Watson R. H., Hitchcock A., Chenevix-Trench G., Thomas E. J., Campbell I. G. Frequent PTEN/MMAC mutations in endometrioid but not serous or mucinous epithelial ovarian tumors.
Cancer Res.
,
58
:
2095
-2097,  
1998
.
58
Cuatrecasas M., Erill N., Musulen E., Costa I., Matias-Guiu X., Prat J. K-ras mutations in nonmucinous ovarian epithelial tumors: a molecular analysis and clinicopathologic study of 144 patients.
Cancer (Phila.)
,
82
:
1088
-1095,  
1998
.
59
Sharma M., Chuang W. W., Sun Z. Phosphatidylinositol 3-kinase/Akt stimulates androgen pathway through GSK3β inhibition and nuclear β-catenin accumulation.
J. Biol. Chem.
,
277
:
30935
-30941,  
2002
.
60
Persad S., Troussard A. A., McPhee T. R., Mulholland D. J., Dedhar S. Tumor suppressor PTEN inhibits nuclear accumulation of β-catenin and T cell/lymphoid enhancer factor 1-mediated transcriptional activation.
J. Cell Biol.
,
153
:
1161
-1174,  
2001
.
61
Bobek L. A., Aguirre A., Levine M. J. Human salivary cystatin S. Cloning, sequence analysis, hybridization in situ and immunocytochemistry.
Biochem. J.
,
278
:
627
-635,  
1991
.
62
Parle-McDermott A., McWilliam P., Tighe O., Dunican D., Croke D. T. Serial analysis of gene expression identifies putative metastasis-associated transcripts in colon tumour cell lines.
Br. J. Cancer
,
83
:
725
-728,  
2000
.
63
Sotiropoulou G., Anisowicz A., Sager R. Identification, cloning, and characterization of cystatin M, a novel cysteine proteinase inhibitor, down-regulated in breast cancer.
J. Biol. Chem.
,
272
:
903
-910,  
1997
.
64
Cox J. L., Sexton P. S., Green T. J., Darmani N. A. Inhibition of B16 melanoma metastasis by overexpression of the cysteine proteinase inhibitor cystatin C.
Melanoma Res.
,
9
:
369
-374,  
1999
.
65
Baron A., DeCarlo A., Featherstone J. Functional aspects of the human salivary cystatins in the oral environment.
Oral Dis.
,
5
:
234
-240,  
1999
.
66
Kunimatsu K., Yamamoto K., Ichimaru E., Kato Y., Kato I. Cathepsins B, H and L activities in gingival crevicular fluid from chronic adult periodontitis patients and experimental gingivitis subjects.
J. Periodontal Res.
,
25
:
69
-73,  
1990
.
67
Eley B. M., Cox S. W. Cathepsin B/L-, elastase-, tryptase-, trypsin- and dipeptidyl peptidase IV-like activities in gingival crevicular fluid: a comparison of levels before and after periodontal surgery in chronic periodontitis patients.
J. Periodontol.
,
63
:
412
-417,  
1992
.
68
Dickinson D. P. Cysteine peptidases of mammals: their biological roles and potential effects in the oral cavity and other tissues in health and disease.
Crit. Rev. Oral Biol. Med.
,
13
:
238
-275,  
2002
.
69
Krepela E. Cysteine proteinases in tumor cell growth and apoptosis.
Neoplasma (Bratisl.)
,
48
:
332
-349,  
2001
.
70
Martucciello G., Ceccherini I., Lerone M., Jasonni V. Pathogenesis of Hirschsprung’s disease.
J. Pediatr. Surg
,
35
:
1017
-1025,  
2000
.
71
Inoue A., Yanagisawa M., Kimura S., Kasuya Y., Miyauchi T., Goto K., Masaki T. The human endothelin family: three structurally and pharmacologically distinct isopeptides predicted by three separate genes.
Proc. Natl. Acad. Sci. USA
,
86
:
2863
-2867,  
1989
.