- Email: [email protected]
β-Catenin Pathway in Ovarian Cancer
Gynecologic Oncology 77, 97–104 (2000) doi:10.1006/gyno.1999.5718, available online at http://www.idealibrary.com on
Rare Activation of the TCF/-Catenin Pathway in Ovarian Cancer 1 Michael T. Furlong, Ph.D., and Patrice J. Morin, Ph.D. 2 Laboratory of Biological Chemistry, Gerontology Research Center, National Institute on Aging, Baltimore, Maryland 21224 Received August 11, 1999
adenomatous polyposis coli (APC) gene [1]. Inactivation of the APC gene leads to the development of benign polyps, which can progress to frank malignancies after they incur additional mutations in genes such as K-ras, p53, and others [2, 3]. Although the causal relationship between APC mutation and colorectal cancer progression has been known for several years, the exact mechanism by which the APC protein suppresses tumor development has only recently begun to be elucidated. A wealth of evidence has accumulated indicating that the APC protein negatively regulates the activity of a transcriptional complex of -catenin and the T cell factor (TCF) family of transcription factors [4]. Specifically, APC binds to and promotes the degradation of the cytoplasmic pool of -catenin, thereby preventing this pool from being translocated to the nucleus to cooperate with the TCF proteins in the activation of TCF-responsive genes [5, 6]. When APC is mutationally inactivated, the cytoplasmic pool of -catenin becomes stabilized, resulting in its accumulation in the cytoplasm as well as in the nucleus, where it forms a transcriptionally competent complex with a TCF family member. Likely transcriptional targets for this complex include c-myc, cyclin D1, and others [7–10]. Although most colorectal cancers harbor APC mutations, there are some colorectal cancers that exhibit constitutive -catenin/TCF-mediated transcriptional activity despite the absence of an APC mutation. This activity was shown to arise from activating mutations in exon 3 of the -catenin gene [11, 12]. These mutations almost always involve serine and threonine residues that normally serve as phosphorylation sites for glycogen synthase kinase-3  (GSK3) [13]. Phosphorylation of these residues is believed to target wild-type -catenin for degradation by the ubiquitin-proteosome pathway [14 –17]. Such mutations have been shown to stabilize -catenin, resulting in its accumulation in the cytoplasm and in the nucleus, where it mediates transcription of target genes as a consequence of its association with a TCF family member. As a first approximation, an activating -catenin mutation is thus believed to be functionally equivalent to an inactivating APC mutation: either lesion results in the induction of constitutive -catenin/TCF-mediated transcription. Subsequent to the discovery of activating -catenin mutations in colorectal cancers, similar mutations have been de-
Objective. The activation of the T cell factor/-catenin pathway is a crucial event in colon cancer initiation. A recent report describing the presence of -catenin mutations in endometrioid ovarian cancer suggested that the TCF/-catenin pathway may be generally activated in ovarian cancer. We therefore undertook to determine the frequency of activation of this pathway in ovarian cancer cell lines using a functional screen. Methods. We functionally screened a series of ovarian cancer cell lines for the presence of constitutive TCF/-catenin-mediated transcriptional activity using a reporter assay. Lines possessing such activity were subjected to mutational and gel-shift analysis, as well as sensitivity to the introduction of dominant-negative TCF or APC alleles. A cDNA harboring a -catenin point mutation found in an ovarian cancer line was incorporated into an expression plasmid for functional analysis. Results. Constitutive TCF/-catenin transcriptional activity was detected in 21% (4 of 19) of ovarian lines studied, while 32% (6 of 19) exhibited greater than twofold repression. One of the constitutively active lines, UCI107, harbored an activating -catenin point mutation, which was shown to be capable of inducing TCF/-catenin transcriptional activity in transiently transfected 293 cells. A second active line, SW626, was shown to harbor an inactivating APC mutation and may in fact be of colonic origin. The third and fourth lines harbored neither an APC nor a -catenin mutation. Gel-shift analysis, together with the absence of sensitivity to dominant-negative TCF, indicated that the reporter activity exhibited by the latter two cell lines may not be due to a TCF/-catenin transcriptional complex. Conclusions. These results indicate that genuine constitutive activation of the TCF/-catenin pathway is infrequent in ovarian cancer, but that constitutive transcriptional repression from TCF sites is more common in this tumor type. Key Words: -catenin; TCF; APC; ovarian cancer; transcription.
INTRODUCTION One of the earliest events in the development of the majority of colorectal cancers is the mutational inactivation of the 1
Part of this work was presented at the 90th Annual Meeting of the American Association for Cancer Research, Philadelphia, PA, April 10 –14, 1999. 2 To whom correspondence should be addressed at Laboratory of Biological Chemistry, Gerontology Research Center, National Institute on Aging, 5600 Nathan Shock Drive, Baltimore, MD 21224. Fax: (410) 558-8386. E-mail: [email protected]. 97
0090-8258/00
98
FURLONG AND MORIN
tected with varying frequencies in a variety of tumor types, including endometrioid ovarian, hepatocellular, prostate, and endometrial cancers, as well as melanomas, pilomatricomas, and medulloblastomas [4, 18]. The presence of such mutations suggested that these tumors might exhibit constitutive elevation of -catenin/TCF-regulated transcriptional activity (CRT), but most of them were not assayed functionally. -catenin mutations have been found in endometrioid ovarian cancer [19]. Because the TCF/-catenin pathway can be activated through multiple genetic alterations, we decided to use a reporter assay to functionally screen a series of ovarian cancer cell lines for the presence of endogenous -catenin/ TCF-regulated transcriptional activity [6].
MATERIALS AND METHODS Several cell lines were used in this study. OV1063 (derived from an ovarian papillary adenocarcinoma), ES-2 (clear cell adenocarcinoma), MDAH 2774 (endometrioid adenocarcinoma), SK-OV-3 (ovarian cancer), CAOV-3 (ovarian papillary adenocarcinoma), HS571 (ovarian carcinoma), OVCAR-3 (ovarian papillary adenocarcinoma), HCT116 (colorectal cancer), SW480 (colorectal cancer), and SW626 (possibly colon cancer, see [20]) were obtained from the American Type Culture Collection (Manassas, VA). A2780 cell line (ovarian cancer) was obtained from Dr. Vilhelm Bohr (Baltimore, MD) and cell line BG-1 (poorly differentiated papillary ovarian cancer) was a gift from Dr. Carl Barrett (Durham, NC). Ovarian cancer cell lines AD10 (an adriamycin-resistant derivative of A2780), UCI101 (papillary ovarian adenocarcinoma), UCI107 (papillary ovarian adenocarcinoma), A222 (ovarian carcinoma), and A224 (ovarian carcinoma) were kindly provided by Dr. Michael Birrer (Rockville, MD). OVCA420, OVCA429, OVCA 432, OVCA433, all derived from ovarian serous cystadenocarcinomas [21] and HEY (papillary cystadenocarcinoma [22]), were obtained from Dr. Robert Bast (Houston, TX). OVC-2 was established in our laboratory from ascites fluid collected from a patient with a serous cystadenocarcinoma (unpublished) and was passaged continuously in vitro for approximately 6 months prior to these studies. -catenin/TCF transcriptional reporter assays were performed essentially as described [11]. Briefly, Cells were transfected in 6-well plates with either wild-type or mutant TCF reporter plasmid (0.5 g) along with a -galactosidase expression plasmid (0.5 g) to control for transfection efficiency. Transfections were performed using the Fugene-6 reagent (Roche Molecular Biochemicals). Cell extracts were prepared 30 h after transfection for determination of luciferase and -galactosidase activities. In some experiments, reporter assays were carried out in the presence of cotransfected expression vectors containing APC, -catenin, or dominant-negative TCF4 cDNAs.
For Western blot analysis of endogenous APC protein expression, cell lines were grown to approximately 80% confluency in 6-well plates and harvested in 200 l of 1⫻ Laemmli sample buffer. For each lysate, 20 l was separated on a 3% agarose/TBE/SDS gel and transferred to Immobilon-P as described [23]. Membranes were probed with the FE9 monoclonal antibody to APC (Oncogene Research Products, Cambridge, MA), followed by ECL detection (Amersham Corp). FLAG epitope-tagged -catenin constructs were detected using the M2 monoclonal antibody (Sigma) following transient transfection into 293 cells. For genomic sequencing, the -catenin gene, which included exons 2 and 3 and a portion of exon 4, was PCR-amplified from genomic DNA derived from ovarian cancer cell lines using the following primers: 5⬘-dTACAACTGTTTTGAAAATCCAGCGTGGAC-3⬘ and 5⬘-dTCGAGTCATTGCATACTGTCC-3⬘. The resulting PCR products were gel-purified and sequenced directly (Thermosequenase Kit, Amersham) using an internal primer (5⬘-dTTGATGGAGTTGGACATGGC-3⬘). Nuclear extracts were prepared essentially as described [24]. Cells were grown to near confluency in 10-cm dishes, washed with PBS, and lysed in STKM buffer (30% sucrose (w/v); 40 mM Tris (pH 7.5), 37 mM KCl, 12 mM MgCl 2, 0.8% Triton X-100). The nuclei were pelleted by centrifugation at 1850g and washed twice with STKM buffer. Nuclear proteins were extracted on ice for 30 min with extraction buffer (10 mM Hepes (pH 7.9), 400 mM NaCl, 1.5 mM MgCl 2, 0.2 mM EGTA, 20% glycerol). Nuclear debris was pelleted by centrifugation (15,000 rpm; 5 min). The clarified nuclear extract was quantitated for protein (Bio-Rad) and stored in aliquots at ⫺80°C until use. Gel-shift analyses of TCF probe-binding activity were carried out as described [24]. The wild-type TCF probe was prepared by annealing the following deoxyoligonucleotides (Operon Technologies, Alameda, CA): dCCCTTTGATCCTTACC and dGGTAAGATCAAAGGG. The mutant TCF probe was prepared by annealing dCCCTTTGGCCTTACC and dGGTAAGGCCAAAGGG. Double-stranded probes were end-labeled using [␥- 32P]ATP and T4 polynucleotide kinase (Life Technologies, Inc), followed by purification on a G-25 Sephadex spin column (5 Prime to 3 Prime, Boulder, CO). Binding reactions contained 5–10 g nuclear protein, 50 ng/l poly(dI-dC), 10 mM Hepes (pH 7.9), 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, and 12% glycerol. Binding reactions involving non-radiolabeled competitor probes also contained ⬃100 fold molar excess of the competitor. Reactions (excluding radiolabeled probe) were assembled on ice and incubated for 5 min at room temperature. The radiolabeled probe (30,000 cpm) was then added and the reactions were incubated at room temperature for another 20 min. Samples were then applied to a 4% polyacrylamide gel in 0.5⫻ TBE at room temperature (3 h, 125 V). Gels were transferred to Whatman paper followed by autoradiography. Supershift experiments were performed using a -catenin monoclonal anti-
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
FIG. 1. Survey of ovarian cancer cell lines for constitutive -catenin/TCFregulated transcriptional activity (CRT). Reporter assays were performed as described under Materials and Methods. Cell lines surveyed (lanes 1–21): HCT116 (colorectal cancer cell line harboring an activating -catenin mutation), SW480 (colorectal cancer cell line harboring an inactivating APC mutation), AD10, A222, UCI101, UCI107, OV1063, A224, ES-2, MDAH 2774, A2780, OVC-2, OVCAR-3, OVCA433, OVCA429, OVCA420, OVCA432, HEY, BG-1, SK-OV-3, and SW626. Fold activation is expressed as the ratio of wild-type to mutant reporter activity.
body (Transduction Laboratories, Lexington, KY). The M2 anti-FLAG epitope monoclonal antibody was used as a negative control. Binding reactions contained 50 ng/l of antibody. RESULTS As shown in Fig. 1, 4 out of 19 ovarian cancer cell lines (UCI101, UCI107, MDAH 2774, and SW626) showed significant TCF/-catenin-mediated reporter activity. Interestingly, of the 15 ovarian cancer cell lines that we considered CRTnegative, 6 lines actually exhibited more than 2-fold higher luciferase activities when transfected with the mutant TCF reporter plasmid compared to the wild-type reporter plasmid (Fig. 1). This suggests that these lines can actively repress transcription from TCF sites. For example, cell lines OV1063 and A2780 (Fig. 1, cell line numbers 7 and 11) show levels of repression of 5- and 10-fold, respectively. Two ovarian lines included in this study (CAOV-3 and HS571) were not successfully transfected and therefore were not included in this reporter assay. However, the -catenin gene was sequenced in these lines (see below). In order to determine the molecular mechanisms underlying the observed CRT in UCI101, UCI107, MDAH 2774, and SW626 cells, we determined the mutational status of these lines using a combination of Western blot analysis and genomic sequencing. We first checked the mutational status of the APC gene in the four CRT-positive ovarian lines. Since almost all APC mutations result in truncated APC proteins [25], we screened for APC mutations by Western blot analysis (Fig. 2). Three of the four cell lines expressed full-length APC (lanes 4, 5, and 6). However, a truncated APC protein was detected in SW626 cells (lane 3), explaining the observed CRT
99
in this line. The CRT exhibited by SW626 cells (308-fold activation, Fig. 1) was approximately 4-fold lower than the CRT exhibited by the colorectal cancer cell line SW480 (1300fold activation, Fig. 1), which also harbors a truncating APC mutation. In contrast, the CRT of SW626 cells was more than an order of magnitude larger than that exhibited by the colorectal cell line HCT116 (27-fold, Fig. 1), which harbors an activating -catenin mutation in exon 3 (⌬S45, [11]). In an attempt to understand the mechanism of elevated CRT in UCI101, UCI107, and MDAH 2774 cells, we determined the mutational status of their -catenin alleles. Since essentially all activating -catenin mutations found to date are located in exon 3 of the -catenin gene [4, 18], we limited our mutational analysis to this exon. Genomic sequencing of exon 3 revealed a missense mutation (S37F) in the UCI107 cell line (Fig. 3A). The CRT exhibited by UCI107 (41-fold, Fig. 1) was similar to the 27-fold activation found in the colorectal cancer cell line HCT116, which, as mentioned above, also harbors an activating -catenin mutation [11]. Based on these data, it appears that the magnitude of the CRT of a given cell line may depend on the nature of the mutation that precipitates this activity rather than the tissue of origin of the cell line itself. Functional analysis of a -catenin cDNA that harbored this mutation indicated that the mutant protein exhibited increased steady-state levels in transient transfections compared to the wild-type protein (Fig. 3B). Furthermore, the S37F mutant was able to induce substantial levels of CRT in the CRT-negative human kidney epithelial cell line 293 (Fig. 3C). Although UCI101 and MDAH 2774 cells contained no activating -catenin point mutations, we still had to consider the possibility that the amino terminus of the protein might be deleted in these cell lines as a consequence of genomic deletion or aberrant splicing of the -catenin mRNA. Such deletions have been shown to stabilize -catenin [26, 27]. The presence of genomic deletions affecting exon 3 in UCI101 and MDAH 2774 cells seemed unlikely, as we observed only full-length PCR products upon
FIG. 2. Immunodetection of APC protein expression in CRT-positive ovarian cancer cell lines. Western blot analysis was carried out as described under Materials and Methods. HCT116 is a colon cancer cell line that expresses wild-type APC protein. SW480 is a colon cancer cell line that expresses truncated (⬃140 kDa) APC protein. SW626 was originally described as an ovarian cancer cell line, but may be of colonic origin [20].
100
FURLONG AND MORIN
FIG. 3. Detection and characterization of an activating -catenin point mutation in the CRT-positive cell line UCI107. (A) Genomic sequencing of -catenin gene in CRT-positive and negative ovarian cancer cell lines. For each cell line, a portion of the -catenin gene was PCR-amplified and sequenced as described under Materials and Methods. Cell lines sequenced (lanes 1–7): SW626, UCI101, UCI107, MDAH-2774, A2780, CAOV-3, and HS571. Arrow indicates the C to T (serine to phenylalanine) point mutation in exon 3 of the -catenin gene in UCI107 cells. (B) Enhanced steady-state levels of transiently expressed -catenin (S37F) point mutant. A total of 293 cells was transfected with 2 g of pCINeo expression vector, pCINeo-wild-type (FLAG epitope-tagged) -catenin, or pCINeo-point mutated (FLAG epitopetagged) -catenin. Laemmli lysates were prepared 30 h later for Western blot analysis as described under Materials and Methods. Arrow denotes FLAGtagged -catenin proteins. (C) Induction of CRT in 293 human kidney epithelial cells by transfection with S37F -catenin cDNA. A total of 293 cells was transfected with a wild-type TCF reporter plasmid (0.5 g), a pCINeo expression plasmid containing the indicated cDNA (1 g), and a -galactosidase plasmid (0.5 g). Cells were harvested for reporter assays after 30 h.
amplification of exons 2 through 4 of their respective -catenin genes (data not shown). To address the possibility of aberrant splicing, we performed reverse transcription-polymerase chain reaction (RT-PCR) assays on total RNA isolated from UCI101 and MDAH 2774 cells, using the same primers that we used for the aforementioned genomic PCR experiments. These experiments demonstrated that there were no deletions in the -catenin transcript in either of these cell lines (data not shown). Given that UCI101 and MDAH 2774 cells were CRTpositive, and yet appeared to harbor neither an inactivating APC mutation or an activating -catenin mutation, we decided to pursue the mechanism(s) of CRT activation in these lines. We considered several possibilities. First, it is possible that these two lines harbor inactivating missense point mutations in their APC genes, which would not be detected by Western blot
analysis. Alternatively, these cells could harbor mutations in other genes encoding proteins that are components of the TCF/-catenin pathway. There are several proteins that are believed to cooperate with APC in downregulating TCF/catenin signaling by mediating the rapid destruction of cytoplasmic pools of -catenin. Such proteins include GSK3, axin, and conductin [28 –30]. It is also possible that the CRT observed in these two lines is due to a lesion in a heretoforeunknown gene or genes that are involved in this pathway. In order to determine which of these possibilities might explain the observed CRT in UCI101 and MDAH 2774 cells, we performed additional CRT assays on the three CRT-positive ovarian cancer lines. First, we examined the effect of ectopically expressed APC protein on CRT levels in these cells (Fig. 4A). In all three cases, only a modest reduction in CRT was observed in the presence of a cotransfected wild-type APC cDNA. Furthermore, an equivalent reduction was observed in the presence of a cDNA that encoded a tumor-derived mutant APC protein (1309⌬, [11]), suggesting that this reduction was not a wild-type APC-specific effect. The lack of APC-mediated CRT attenuation in UCI107 cells was not surprising, since cells harboring mutationally activated -catenin alleles have previously been shown to exhibit CRT that is refractory to the introduction of exogenous wild-type APC [11]. The lack of attenuation by APC in UCI101 and MDAH 2774 cells was informative in that it suggested that neither of these cell lines contained an inactivating APC missense point mutation, which would have been corrected by wild-type APC. One cannot formally exclude the possibility that a dominant-negative APC missense mutation was indeed present in one or both of these cell lines, blunting the effect of the ectopically expressed APC protein. In addition to arguing against the presence of an APC missense mutation, these results suggested that the activating lesion in these two cell lines must be downstream of APC in the CRT-activating pathway. To further explore the mechanism underlying the observed CRT in UCI101 and MDAH 2774 cells, we performed an additional set of CRT assays, this time in the presence of a cotransfected dominant-negative TCF4 cDNA (Fig. 4B). This construct, which contains the DNA-binding domain of TCF4, but lacks its amino terminal -catenin interaction domain, has been previously shown to inhibit CRT in colorectal cell lines harboring either an APC or a -catenin mutation [11]. As expected, endogenous CRT activity was markedly inhibited in UCI107 cells, which, as described earlier, harbor an activating -catenin mutation. Surprisingly, however, the endogenous reporter activity in the other two CRT-positive cell lines (UCI101 and MDAH 2774) was not significantly inhibited by the dominant-negative allele. These data suggested that the observed CRT in these 2 cell lines may not be due to genuine endogenous TCF/-catenin activity, but rather due to the activity of a different transcription factor, which could presumably bind to the reporter construct, competing effectively with the ectopically expressed dominant-negative TCF4 protein.
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
101
gel-shift assays on the colorectal cancer cell line HCT116, which exhibits constitutive -catenin/TCF activity as a consequence of an in-frame deletion of a GSK3 phosphorylation site (serine 45) in one allele [11]. As shown in Fig. 5A, we were able to detect optimal TCF probe-specific complexes. Additionally, we were able to demonstrate the presence of -catenin in these complexes (see supershifted band in lane 2). Next, we performed gel-shift experiments on the three CRTpositive ovarian lines. Based on the presence of TCF reporter activity in these lines, we anticipated that nuclear extracts derived from them would contain TCF probe retardation activity. Indeed, this was the case (Fig. 5B, compare lanes 1, 4, and 7). Significantly, the addition of a -catenin monoclonal antibody resulted in the appearance of a supershift in the UCI107 complex, but not in the complexes derived from the other two cell lines (compare lanes 2, 5, and 8). Addition of a control antibody (M2 anti-FLAG) did not induce a supershift in any of the extracts (lanes 3, 6, and 9). These results indicated that the transcription factor complex responsible for binding to the TCF probe contained -catenin in the case of UCI107derived extracts, but not in the case of extracts derived from UCI101 and MDAH 2774 cells. DISCUSSION
FIG. 4. Investigation of the mechanism of CRT induction in UCI-101 and MDAH 2774 ovarian cancer cell lines. (A) Ectopic expression of wild-type APC does not significantly downregulate CRT in ovarian cancer cell lines. Cells were transfected in 6-well plates with a wild-type TCF reporter plasmid (0.5 g) along with 2 g of pCINeo empty vector (treatment 1), wild-type APC expression plasmid (treatment 2), or 1309⌬ mutant APC expression plasmid (treatment 3). SW480 is a colorectal cancer cell line that expresses truncated APC protein. A -galactosidase expression plasmid (0.5 g) was included to control for transfection efficiency. Cell extracts were prepared 30 h posttransfection and assayed for reporter activity. Relative CRT is expressed as the ratio of reporter activity in the APC-transfected cells to the reporter activity of cells transfected with pCINeo expression vector alone. (B) Ectopic expression of dominant-negative TCF4 does not downregulate CRT in UCI101 or MDAH 2774 cell lines. Cells were transfected in 6-well plates with wild-type TCF reporter plasmid (0.5 g) together with 0, 0.5, or 3.0 g of a dominantnegative TCF4 expression plasmid and a -galactosidase expression plasmid (0.5 g). Cell extracts were prepared 30 h posttransfection and assayed for reporter activity. Relative CRT is expressed as the ratio of reporter activity observed in cells transfected with dominant-negative TCF4 plasmid to the reporter activity observed in cells transfected with pCINeo vector alone.
Additionally, these results provided further evidence that the observed CRT in UCI101 and MDAH 2774 cells was not due to a missense APC mutation, since the CRT resulting from such a lesion would be expected to be downregulated by the presence of a dominant-negative TCF4 allele. In order to pursue this possibility further, we performed gel-shift assays with nuclear extracts derived from UCI107, UCI101, and MDAH 2774 cells. We first conducted control
Although it was originally discovered as a membrane-associated protein complexed to E-cadherin, -catenin has recently
FIG. 5. Gel-shift analysis of TCF probe-binding activity in nuclear extracts derived from CRT-positive ovarian cancer cell lines. Nuclear extract preparations and gel-shift analyses were carried out as described under Materials and Methods. (A) Control gel-shift experiment using nuclear extracts derived from HCT116 colorectal cancer cells, which harbor an activating in-frame deletion of serine 45 in one -catenin allele. Arrow denotes supershift induced by addition of anti--catenin antibody to binding reaction. (B) Differential detection of -catenin in TCF probe-binding complexes derived from CRT-positive ovarian cancer lines. Arrow denotes anti--catenin antibodymediated supershift. The M2 anti-FLAG epitope monoclonal antibody was used as a negative control. Comp. denotes nonradiolabeled probe competitor (wt, wild-type; mu, mutant).
102
FURLONG AND MORIN
emerged as an important player in human neoplasia [4, 18, 31]. -catenin can become oncogenic by cooperating with a TCF family member in the activation of key genes such as cyclin D1, c-myc, and others. In order to do so, the cytoplasmic fraction of -catenin must become stabilized in order to facilitate its interaction with a TCF molecule and subsequent shuttling to the nucleus. Cytoplasmic -catenin can be stabilized in at least two ways. The first is the mutational inactivation of the APC gene, whose product has been convincingly shown to be at least in part responsible for the degradation of cytoplasmic pools of -catenin [5, 6]. The second is by point mutation of the -catenin gene, almost always in key amino terminal phosphorylatable serine or threonine residues [11, 32]. These mutations have been shown to render the encoded protein refractory to APC-mediated degradation. While APC mutations are found almost exclusively in colorectal cancer, -catenin mutations have been found in a variety of human cancers, including endometrioid ovarian cancer [18, 19]. We undertook to determine the extent to which the TCF/catenin complex is activated in ovarian cancer. Because this complex could be activated by mutations in genes other than -catenin or APC, we decided to forego a mutational analysis in favor of a reporter assay to functionally screen a series of ovarian cancer cell lines for the presence of endogenous -catenin/TCF-regulated transcriptional activity. While this approach has the advantage of not making any assumptions about which gene or genes may or may not be mutated in the -catenin/TCF pathway, it has the disadvantage of relying upon established cell lines for analysis. One can never be certain that the transcriptional activity and the mutation(s) that cause it were actually present in the original tumor from which the CRT-positive cell lines were derived. Unfortunately, reporter assays, which require transfection of plasmid DNA, cannot be conveniently carried out with primary tumor samples. It is possible to attempt short-term in vitro cultivation of primary tumors prior to transfection, but such cultures are frequently contaminated with nontransformed cells, which may interfere with accurate analysis. The contaminating cells usually die out with prolonged passage in tissue culture, but genetic alterations in the surviving tumor cells may accumulate during this period. It is encouraging to note, however, that -catenin mutations have been found in primary tumors of many tissue origins [18]. In addition, in one case in which matched archival tumor DNA was available for a CRT-positive colorectal cell line (HCT116), a heterozygous -catenin mutation was found in both the cell line and the corresponding primary tumor [11]. This confirmed that at least in this instance, a CRT-inducing mutation had occurred during tumor progression, and not during cell culture propagation. Our reporter screen detected CRT in 21% (4/19) of the ovarian cancer cell lines studied. One of these CRT-positive lines, SW626, harbored a frameshift APC mutation. This was somewhat surprising, since APC mutations are rarely found in
noncolorectal tumors and have never been found in ovarian cancer [33]. Based on the presence of an APC mutation, and upon further molecular characterization of this cell line, we have concluded that SW626 cells may in fact be of colonic origin [20]. Of the remaining three CRT-positive ovarian cancer cell lines, only one (UCI107) had an activating -catenin point mutation. The other two lines (UCI101 and MDAH 2774) did not, suggesting that they may have lesions in another gene or genes involved in this pathway. However, based on further characterization of the observed reporter activity in these lines, we have concluded that the activity observed in the UCI107 cell line is genuine -catenin/TCF-mediated transcriptional activity, whereas the activity observed in the other two ovarian cancer cell lines is likely due to the activity of an unrelated transcription factor complex. This unrelated factor may contain a TCF-like protein, but our gel-shift data suggest that it does not contain -catenin. These results indicate that caution must be used when interpreting the results of -catenin/TCF reporter assays, since they may not always accurately reflect the presence or absence of this transcriptional complex. The presence of apparent TCF-mediated transcriptional repression in six ovarian cell lines was somewhat unexpected, but not without precedent. For example, it has been shown that, in the absence of -catenin, XTcf-3 can act as a transcriptional repressor on the ventral side of Xenopus embryos [34]. This TCF-mediated repression during development is likely to be mediated through the recruitment of the Groucho corepressor [35, 36]. Groucho homologues known as TLE proteins exist in vertebrates and can also repress TCF-mediated transcription [37]. The matrylisin promoter was recently reported to be regulated by TCF/-catenin complexes. Interestingly, mutation of the TCF binding site increased basal activity, suggesting that -catenin binding to TCF may act to overcome its repressing effect on the promoter [10]. However, the relevance of TCF-mediated transcriptional repression in cancer remains unclear. The frequent observation of TCF transcriptional repression in ovarian cancer cell lines may be relevant for ovarian tumorigenesis or may simply reflect low levels of -catenin signaling and/or the presence of Groucho factors in these cells. In summary, constitutive TCF/-catenin activity is found in ovarian cancer cell lines at low frequency. The TCF reporter activity observed in one of the TCF/-catenin-active ovarian lines (UCI107) was shown to be due to an activating mutation in one of its -catenin alleles. Preliminary evidence suggests that the activity observed in two additional lines may in fact not be due to activation of the TCF/-catenin pathway, but rather due to the activity of an unrelated transcription factor complex that is devoid of -catenin. The UCI107 cell line, which was derived from an ovarian papillary adenocarcinoma [38], represents the first ovarian cancer cell line shown to harbor genuine constitutive -catenin/TCF-mediated transcriptional ac-
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
tivity. Mutations of the -catenin gene in ovarian cancer have previously been associated with endometrioid differentiation [19, 39]. In this report, we show that activation of the TCF/ -catenin pathway is not an obligatory step in the tumorigenesis of even this subtype of ovarian cancer, since the MDAH 2774 cell line, derived from endometrioid ovarian cancer, was found to be CRT-negative. In any event, the UCI107 line should provide a useful system for the study of genes involved in the APC/-catenin pathway as well as the relationship of this pathway to ovarian tumorigenesis. Finally, a substantial fraction of the ovarian cancer cell lines in this study exhibited constitutive TCF-mediated repression as measured by luciferase assays. The relevance of this repression in ovarian tumorigenesis is unclear but is currently under investigation. REFERENCES 1. Kinzler KW, Vogelstein B: Lessons from hereditary colorectal cancer. Cell 87:159 –170, 1996 2. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL: Genetic alterations during colorectal-tumor development. N Engl J Med 319:525–532, 1988 3. Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 61:759 –767, 1990 4. Polakis P: The oncogenic activation of -catenin. Curr Opin Genet Dev 9:15–21, 1999 5. Munemitsu S, Albert I, Souza B, Rubinfeld B, Polakis P: Regulation of intracellular -catenin levels by the adenomatous polyposis coli (APC) tumor-suppressor protein. Proc Natl Acad Sci USA 92:3046 –3050, 1995 6. Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, Kinzler KW, Vogelstein B, Clevers H: Constitutive transcriptional activation by a -catenin-Tcf complex in APC ⫺/⫺ colon carcinoma. Science 275:1784 – 1787, 1997 7. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW: Identification of c-MYC as a target of the APC pathway. Science 281:1509 –1512, 1998 8. Tetsu O, McCormick F: -catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 398:422– 426, 1999 9. Mann B, Gelos M, Siedow A, Hanski ML, Gratchev A, Ilyas M, Bodmer WF, Moyer MP, Riecken EO, Buhr HJ, Hanski C: Target genes of -catenin-T cell-factor lymphoid-enhancer-factor signaling in human colorectal carcinomas. Proc Natl Acad Sci USA 96:1603–1608, 1999 10. Crawford HC, Fingleton BM, Rudolph-Owen LA, Heppner Goss KJ, Rubinfeld B, Polakis P, Matrisian LM: The metalloproteinase matrilysin is a target of -catenin transactivation in intestinal tumors. Oncogene 18: 2883–2891, 1999 11. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, Kinzler KW: Activation of -catenin-Tcf signaling in colon cancer by mutations in -catenin or APC. Science 275:1787–1790, 1997 12. Sparks AB, Morin PJ, Vogelstein B, Kinzler KW: Mutational analysis of the APC/-catenin/Tcf pathway in colorectal cancer. Cancer Res 58: 1130 –1134, 1998 13. Yost C, Torres M, Miller JR, Huang E, Kimelman D, Moon RT: The axis-inducing activity, stability, and subcellular distribution of -catenin is regulated in Xenopus embryos by glycogen synthase kinase 3. Genes Dev 10:1443–1454, 1996 14. Aberle H, Bauer A, Stappert J, Kispert A, Kemler R: -catenin is a target for the ubiquitin-proteasome pathway. EMBO J 16:3797–3804, 1997
103
15. Orford K, Crockett C, Jensen JP, Weissman AM, Byers SW: Serine phosphorylation-regulated ubiquitination and degradation of -catenin. J Biol Chem 272:24735–24738, 1997 16. Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, Harper JW: The SCF -TRCP-ubiquitin ligase complex associates specifically with phosphorylated destruction motifs in IB␣ and -catenin and stimulates IB␣ ubiquitination in vitro. Genes Dev 13:270 –283, 1999 17. Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H, Perret C, Rubinfeld B, Margottin F, Benarous R, Polakis P: The F-box protein -TrCP associates with phosphorylated -catenin and regulates its activity in the cell. Curr Biol 9:207–210, 1999 18. Morin PJ: -catenin signaling and cancer. Bioessays 21:1021–1030, 1999 19. Palacios J, Gamallo C: Mutations in the -catenin gene (CTNNB1) in endometrioid ovarian carcinomas. Cancer Res 58:1344 –1347, 1998 20. Furlong MT, Hough CD, Sherman-Baust CA, Pizer ES, Morin PJ: Evidence for the Colonic Origin of Ovarian Cancer Cell Line SW626. J Natl Cancer Inst 91:1327–1328, 1999 21. Bast RC, Feeney M, Lazarus H, Nadler LM, Colvin RB, Knapp RC: Reactivity of a monoclonal antibody with human ovarian carcinoma. J Clin Invest 68:1331–1337, 1981 22. Buick RN, Pullano R, Trent JM: Comparative properties of five human ovarian adenocarcinoma cell lines. Cancer Res 45:3668 –3676, 1985 23. Smith KJ, Johnson KA, Bryan TM, Hill DE, Markowitz S, Willson JK, Paraskeva C, Petersen GM, Hamilton SR, Vogelstein B, Kinzler KW: The APC gene product in normal and tumor cells. Proc Natl Acad Sci USA 90:2846 –2850, 1993 24. van de Wetering M, Oosterwegel M, Dooijes D, Clevers H: Identification and cloning of TCF-1, a T lymphocyte-specific transcription factor containing a sequence-specific HMG box. EMBO J 10:123–132, 1991 25. Nagase H, Nakamura Y: Mutations of the APC (adenomatous polyposis coli) gene. Hum Mutat 2:425–34, 1993 26. Iwao K, Nakamori S, Kameyama M, Imaoka S, Kinoshita M, Fukui T, Ishiguro S, Nakamura Y, Miyoshi Y: Activation of the -catenin gene by interstitial deletions involving exon 3 in primary colorectal carcinomas without adenomatous polyposis coli mutations. Cancer Res 58:1021– 1026, 1998 27. Whitehead I, Kirk H, Kay R: Expression cloning of oncogenes by retroviral transfer of cDNA libraries. Mol Cell Biol 15:704 –710, 1995 28. Rubinfeld B, Albert I, Porfiri E, Fiol C, Munemitsu S, Polakis P: Binding of GSK3- to the APC--catenin complex and regulation of complex assembly. Science 272:1023–1025, 1996 29. Ikeda S, Kishida S, Yamamoto H, Murai H, Koyama S, Kikuchi A: Axin, a negative regulator of the wnt signaling pathway, forms a complex with GSK-3 and -catenin and promotes GSK-3-dependent phosphorylation of -catenin. EMBO J 17:1371–1384, 1998 30. Behrens J, Jerchow B, Wurtele M, Grimm J, Asbrand C, Wirtz R, Kuhl M, Wedlich D, Birchmeier W: Functional interaction of an axin homolog, conductin, with -catenin, APC, and GSK3 . Science 280:596 –599, 1998 31. Peifer M: -catenin as oncogene: The smoking gun. Science 275:1752– 1753, 1997 32. Rubinfeld B, Robbins P, El-Gamil M, Albert I, Porfiri E, Polakis P: Stabilization of -catenin by genetic defects in melanoma cell lines. Science 275:1790 –1792, 1997 33. Allan GJ, Cottrell S, Trowsdale J, Foulkes WD: Loss of heterozygosity on chromosome 5 in sporadic ovarian carcinoma is a late event and is not associated with mutations in APC at 5q21-22. Hum Mutat 3:283–291, 1994 34. Brannon M, Gomperts M, Sumoy L, Moon RT, Kimelman D: A -catenin/ XTcf-3 complex binds to the siamois promoter to regulate dorsal axis specification in Xenopus. Genes Dev 11:2359 –2370, 1997
104
FURLONG AND MORIN
35. Cavallo RA, Cox RT, Moline MM, Roose J, Polevoy GA, Clevers H, Peifer M, Bejsovec A: Drosophila Tcf and Groucho interact to repress Wingless signalling activity. Nature 395:604 – 608, 1998 36. Roose J, Molenaar M, Peterson J, Hurenkamp J, Brantjes H, Moerer P, van de Wetering M, Destree O, Clevers H: The Xenopus Wnt effector XTcf-3 interacts with Groucho-related transcriptional repressors. Nature 395:608 – 612, 1998 37. Levanon D, Goldstein RE, Bernstein Y, Tang H, Goldenberg D, Stifani S, Paroush Z, Groner Y: Transcriptional repression by AML1 and LEF-1 is
mediated by the TLE/Groucho corepressors. Proc Natl Acad Sci USA 95:11590 –11595, 1998 38. Gamboa G, Carpenter PM, Podnos YD, Dorion G, Iravani L, Bolton D, Mascarello JT, Manetta A: Characterization and development of UCI 107, a primary human ovarian carcinoma cell line. Gynecol Oncol 58:336–343, 1995 39. Wright K, Wilson P, Morland S, Campbell I, Walsh M, Hurst T, Ward B, Cummings M, Chenevix-Trench G: -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
Rare Activation of the TCF/-Catenin Pathway in Ovarian Cancer 1 Michael T. Furlong, Ph.D., and Patrice J. Morin, Ph.D. 2 Laboratory of Biological Chemistry, Gerontology Research Center, National Institute on Aging, Baltimore, Maryland 21224 Received August 11, 1999
adenomatous polyposis coli (APC) gene [1]. Inactivation of the APC gene leads to the development of benign polyps, which can progress to frank malignancies after they incur additional mutations in genes such as K-ras, p53, and others [2, 3]. Although the causal relationship between APC mutation and colorectal cancer progression has been known for several years, the exact mechanism by which the APC protein suppresses tumor development has only recently begun to be elucidated. A wealth of evidence has accumulated indicating that the APC protein negatively regulates the activity of a transcriptional complex of -catenin and the T cell factor (TCF) family of transcription factors [4]. Specifically, APC binds to and promotes the degradation of the cytoplasmic pool of -catenin, thereby preventing this pool from being translocated to the nucleus to cooperate with the TCF proteins in the activation of TCF-responsive genes [5, 6]. When APC is mutationally inactivated, the cytoplasmic pool of -catenin becomes stabilized, resulting in its accumulation in the cytoplasm as well as in the nucleus, where it forms a transcriptionally competent complex with a TCF family member. Likely transcriptional targets for this complex include c-myc, cyclin D1, and others [7–10]. Although most colorectal cancers harbor APC mutations, there are some colorectal cancers that exhibit constitutive -catenin/TCF-mediated transcriptional activity despite the absence of an APC mutation. This activity was shown to arise from activating mutations in exon 3 of the -catenin gene [11, 12]. These mutations almost always involve serine and threonine residues that normally serve as phosphorylation sites for glycogen synthase kinase-3  (GSK3) [13]. Phosphorylation of these residues is believed to target wild-type -catenin for degradation by the ubiquitin-proteosome pathway [14 –17]. Such mutations have been shown to stabilize -catenin, resulting in its accumulation in the cytoplasm and in the nucleus, where it mediates transcription of target genes as a consequence of its association with a TCF family member. As a first approximation, an activating -catenin mutation is thus believed to be functionally equivalent to an inactivating APC mutation: either lesion results in the induction of constitutive -catenin/TCF-mediated transcription. Subsequent to the discovery of activating -catenin mutations in colorectal cancers, similar mutations have been de-
Objective. The activation of the T cell factor/-catenin pathway is a crucial event in colon cancer initiation. A recent report describing the presence of -catenin mutations in endometrioid ovarian cancer suggested that the TCF/-catenin pathway may be generally activated in ovarian cancer. We therefore undertook to determine the frequency of activation of this pathway in ovarian cancer cell lines using a functional screen. Methods. We functionally screened a series of ovarian cancer cell lines for the presence of constitutive TCF/-catenin-mediated transcriptional activity using a reporter assay. Lines possessing such activity were subjected to mutational and gel-shift analysis, as well as sensitivity to the introduction of dominant-negative TCF or APC alleles. A cDNA harboring a -catenin point mutation found in an ovarian cancer line was incorporated into an expression plasmid for functional analysis. Results. Constitutive TCF/-catenin transcriptional activity was detected in 21% (4 of 19) of ovarian lines studied, while 32% (6 of 19) exhibited greater than twofold repression. One of the constitutively active lines, UCI107, harbored an activating -catenin point mutation, which was shown to be capable of inducing TCF/-catenin transcriptional activity in transiently transfected 293 cells. A second active line, SW626, was shown to harbor an inactivating APC mutation and may in fact be of colonic origin. The third and fourth lines harbored neither an APC nor a -catenin mutation. Gel-shift analysis, together with the absence of sensitivity to dominant-negative TCF, indicated that the reporter activity exhibited by the latter two cell lines may not be due to a TCF/-catenin transcriptional complex. Conclusions. These results indicate that genuine constitutive activation of the TCF/-catenin pathway is infrequent in ovarian cancer, but that constitutive transcriptional repression from TCF sites is more common in this tumor type. Key Words: -catenin; TCF; APC; ovarian cancer; transcription.
INTRODUCTION One of the earliest events in the development of the majority of colorectal cancers is the mutational inactivation of the 1
Part of this work was presented at the 90th Annual Meeting of the American Association for Cancer Research, Philadelphia, PA, April 10 –14, 1999. 2 To whom correspondence should be addressed at Laboratory of Biological Chemistry, Gerontology Research Center, National Institute on Aging, 5600 Nathan Shock Drive, Baltimore, MD 21224. Fax: (410) 558-8386. E-mail: [email protected]. 97
0090-8258/00
98
FURLONG AND MORIN
tected with varying frequencies in a variety of tumor types, including endometrioid ovarian, hepatocellular, prostate, and endometrial cancers, as well as melanomas, pilomatricomas, and medulloblastomas [4, 18]. The presence of such mutations suggested that these tumors might exhibit constitutive elevation of -catenin/TCF-regulated transcriptional activity (CRT), but most of them were not assayed functionally. -catenin mutations have been found in endometrioid ovarian cancer [19]. Because the TCF/-catenin pathway can be activated through multiple genetic alterations, we decided to use a reporter assay to functionally screen a series of ovarian cancer cell lines for the presence of endogenous -catenin/ TCF-regulated transcriptional activity [6].
MATERIALS AND METHODS Several cell lines were used in this study. OV1063 (derived from an ovarian papillary adenocarcinoma), ES-2 (clear cell adenocarcinoma), MDAH 2774 (endometrioid adenocarcinoma), SK-OV-3 (ovarian cancer), CAOV-3 (ovarian papillary adenocarcinoma), HS571 (ovarian carcinoma), OVCAR-3 (ovarian papillary adenocarcinoma), HCT116 (colorectal cancer), SW480 (colorectal cancer), and SW626 (possibly colon cancer, see [20]) were obtained from the American Type Culture Collection (Manassas, VA). A2780 cell line (ovarian cancer) was obtained from Dr. Vilhelm Bohr (Baltimore, MD) and cell line BG-1 (poorly differentiated papillary ovarian cancer) was a gift from Dr. Carl Barrett (Durham, NC). Ovarian cancer cell lines AD10 (an adriamycin-resistant derivative of A2780), UCI101 (papillary ovarian adenocarcinoma), UCI107 (papillary ovarian adenocarcinoma), A222 (ovarian carcinoma), and A224 (ovarian carcinoma) were kindly provided by Dr. Michael Birrer (Rockville, MD). OVCA420, OVCA429, OVCA 432, OVCA433, all derived from ovarian serous cystadenocarcinomas [21] and HEY (papillary cystadenocarcinoma [22]), were obtained from Dr. Robert Bast (Houston, TX). OVC-2 was established in our laboratory from ascites fluid collected from a patient with a serous cystadenocarcinoma (unpublished) and was passaged continuously in vitro for approximately 6 months prior to these studies. -catenin/TCF transcriptional reporter assays were performed essentially as described [11]. Briefly, Cells were transfected in 6-well plates with either wild-type or mutant TCF reporter plasmid (0.5 g) along with a -galactosidase expression plasmid (0.5 g) to control for transfection efficiency. Transfections were performed using the Fugene-6 reagent (Roche Molecular Biochemicals). Cell extracts were prepared 30 h after transfection for determination of luciferase and -galactosidase activities. In some experiments, reporter assays were carried out in the presence of cotransfected expression vectors containing APC, -catenin, or dominant-negative TCF4 cDNAs.
For Western blot analysis of endogenous APC protein expression, cell lines were grown to approximately 80% confluency in 6-well plates and harvested in 200 l of 1⫻ Laemmli sample buffer. For each lysate, 20 l was separated on a 3% agarose/TBE/SDS gel and transferred to Immobilon-P as described [23]. Membranes were probed with the FE9 monoclonal antibody to APC (Oncogene Research Products, Cambridge, MA), followed by ECL detection (Amersham Corp). FLAG epitope-tagged -catenin constructs were detected using the M2 monoclonal antibody (Sigma) following transient transfection into 293 cells. For genomic sequencing, the -catenin gene, which included exons 2 and 3 and a portion of exon 4, was PCR-amplified from genomic DNA derived from ovarian cancer cell lines using the following primers: 5⬘-dTACAACTGTTTTGAAAATCCAGCGTGGAC-3⬘ and 5⬘-dTCGAGTCATTGCATACTGTCC-3⬘. The resulting PCR products were gel-purified and sequenced directly (Thermosequenase Kit, Amersham) using an internal primer (5⬘-dTTGATGGAGTTGGACATGGC-3⬘). Nuclear extracts were prepared essentially as described [24]. Cells were grown to near confluency in 10-cm dishes, washed with PBS, and lysed in STKM buffer (30% sucrose (w/v); 40 mM Tris (pH 7.5), 37 mM KCl, 12 mM MgCl 2, 0.8% Triton X-100). The nuclei were pelleted by centrifugation at 1850g and washed twice with STKM buffer. Nuclear proteins were extracted on ice for 30 min with extraction buffer (10 mM Hepes (pH 7.9), 400 mM NaCl, 1.5 mM MgCl 2, 0.2 mM EGTA, 20% glycerol). Nuclear debris was pelleted by centrifugation (15,000 rpm; 5 min). The clarified nuclear extract was quantitated for protein (Bio-Rad) and stored in aliquots at ⫺80°C until use. Gel-shift analyses of TCF probe-binding activity were carried out as described [24]. The wild-type TCF probe was prepared by annealing the following deoxyoligonucleotides (Operon Technologies, Alameda, CA): dCCCTTTGATCCTTACC and dGGTAAGATCAAAGGG. The mutant TCF probe was prepared by annealing dCCCTTTGGCCTTACC and dGGTAAGGCCAAAGGG. Double-stranded probes were end-labeled using [␥- 32P]ATP and T4 polynucleotide kinase (Life Technologies, Inc), followed by purification on a G-25 Sephadex spin column (5 Prime to 3 Prime, Boulder, CO). Binding reactions contained 5–10 g nuclear protein, 50 ng/l poly(dI-dC), 10 mM Hepes (pH 7.9), 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, and 12% glycerol. Binding reactions involving non-radiolabeled competitor probes also contained ⬃100 fold molar excess of the competitor. Reactions (excluding radiolabeled probe) were assembled on ice and incubated for 5 min at room temperature. The radiolabeled probe (30,000 cpm) was then added and the reactions were incubated at room temperature for another 20 min. Samples were then applied to a 4% polyacrylamide gel in 0.5⫻ TBE at room temperature (3 h, 125 V). Gels were transferred to Whatman paper followed by autoradiography. Supershift experiments were performed using a -catenin monoclonal anti-
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
FIG. 1. Survey of ovarian cancer cell lines for constitutive -catenin/TCFregulated transcriptional activity (CRT). Reporter assays were performed as described under Materials and Methods. Cell lines surveyed (lanes 1–21): HCT116 (colorectal cancer cell line harboring an activating -catenin mutation), SW480 (colorectal cancer cell line harboring an inactivating APC mutation), AD10, A222, UCI101, UCI107, OV1063, A224, ES-2, MDAH 2774, A2780, OVC-2, OVCAR-3, OVCA433, OVCA429, OVCA420, OVCA432, HEY, BG-1, SK-OV-3, and SW626. Fold activation is expressed as the ratio of wild-type to mutant reporter activity.
body (Transduction Laboratories, Lexington, KY). The M2 anti-FLAG epitope monoclonal antibody was used as a negative control. Binding reactions contained 50 ng/l of antibody. RESULTS As shown in Fig. 1, 4 out of 19 ovarian cancer cell lines (UCI101, UCI107, MDAH 2774, and SW626) showed significant TCF/-catenin-mediated reporter activity. Interestingly, of the 15 ovarian cancer cell lines that we considered CRTnegative, 6 lines actually exhibited more than 2-fold higher luciferase activities when transfected with the mutant TCF reporter plasmid compared to the wild-type reporter plasmid (Fig. 1). This suggests that these lines can actively repress transcription from TCF sites. For example, cell lines OV1063 and A2780 (Fig. 1, cell line numbers 7 and 11) show levels of repression of 5- and 10-fold, respectively. Two ovarian lines included in this study (CAOV-3 and HS571) were not successfully transfected and therefore were not included in this reporter assay. However, the -catenin gene was sequenced in these lines (see below). In order to determine the molecular mechanisms underlying the observed CRT in UCI101, UCI107, MDAH 2774, and SW626 cells, we determined the mutational status of these lines using a combination of Western blot analysis and genomic sequencing. We first checked the mutational status of the APC gene in the four CRT-positive ovarian lines. Since almost all APC mutations result in truncated APC proteins [25], we screened for APC mutations by Western blot analysis (Fig. 2). Three of the four cell lines expressed full-length APC (lanes 4, 5, and 6). However, a truncated APC protein was detected in SW626 cells (lane 3), explaining the observed CRT
99
in this line. The CRT exhibited by SW626 cells (308-fold activation, Fig. 1) was approximately 4-fold lower than the CRT exhibited by the colorectal cancer cell line SW480 (1300fold activation, Fig. 1), which also harbors a truncating APC mutation. In contrast, the CRT of SW626 cells was more than an order of magnitude larger than that exhibited by the colorectal cell line HCT116 (27-fold, Fig. 1), which harbors an activating -catenin mutation in exon 3 (⌬S45, [11]). In an attempt to understand the mechanism of elevated CRT in UCI101, UCI107, and MDAH 2774 cells, we determined the mutational status of their -catenin alleles. Since essentially all activating -catenin mutations found to date are located in exon 3 of the -catenin gene [4, 18], we limited our mutational analysis to this exon. Genomic sequencing of exon 3 revealed a missense mutation (S37F) in the UCI107 cell line (Fig. 3A). The CRT exhibited by UCI107 (41-fold, Fig. 1) was similar to the 27-fold activation found in the colorectal cancer cell line HCT116, which, as mentioned above, also harbors an activating -catenin mutation [11]. Based on these data, it appears that the magnitude of the CRT of a given cell line may depend on the nature of the mutation that precipitates this activity rather than the tissue of origin of the cell line itself. Functional analysis of a -catenin cDNA that harbored this mutation indicated that the mutant protein exhibited increased steady-state levels in transient transfections compared to the wild-type protein (Fig. 3B). Furthermore, the S37F mutant was able to induce substantial levels of CRT in the CRT-negative human kidney epithelial cell line 293 (Fig. 3C). Although UCI101 and MDAH 2774 cells contained no activating -catenin point mutations, we still had to consider the possibility that the amino terminus of the protein might be deleted in these cell lines as a consequence of genomic deletion or aberrant splicing of the -catenin mRNA. Such deletions have been shown to stabilize -catenin [26, 27]. The presence of genomic deletions affecting exon 3 in UCI101 and MDAH 2774 cells seemed unlikely, as we observed only full-length PCR products upon
FIG. 2. Immunodetection of APC protein expression in CRT-positive ovarian cancer cell lines. Western blot analysis was carried out as described under Materials and Methods. HCT116 is a colon cancer cell line that expresses wild-type APC protein. SW480 is a colon cancer cell line that expresses truncated (⬃140 kDa) APC protein. SW626 was originally described as an ovarian cancer cell line, but may be of colonic origin [20].
100
FURLONG AND MORIN
FIG. 3. Detection and characterization of an activating -catenin point mutation in the CRT-positive cell line UCI107. (A) Genomic sequencing of -catenin gene in CRT-positive and negative ovarian cancer cell lines. For each cell line, a portion of the -catenin gene was PCR-amplified and sequenced as described under Materials and Methods. Cell lines sequenced (lanes 1–7): SW626, UCI101, UCI107, MDAH-2774, A2780, CAOV-3, and HS571. Arrow indicates the C to T (serine to phenylalanine) point mutation in exon 3 of the -catenin gene in UCI107 cells. (B) Enhanced steady-state levels of transiently expressed -catenin (S37F) point mutant. A total of 293 cells was transfected with 2 g of pCINeo expression vector, pCINeo-wild-type (FLAG epitope-tagged) -catenin, or pCINeo-point mutated (FLAG epitopetagged) -catenin. Laemmli lysates were prepared 30 h later for Western blot analysis as described under Materials and Methods. Arrow denotes FLAGtagged -catenin proteins. (C) Induction of CRT in 293 human kidney epithelial cells by transfection with S37F -catenin cDNA. A total of 293 cells was transfected with a wild-type TCF reporter plasmid (0.5 g), a pCINeo expression plasmid containing the indicated cDNA (1 g), and a -galactosidase plasmid (0.5 g). Cells were harvested for reporter assays after 30 h.
amplification of exons 2 through 4 of their respective -catenin genes (data not shown). To address the possibility of aberrant splicing, we performed reverse transcription-polymerase chain reaction (RT-PCR) assays on total RNA isolated from UCI101 and MDAH 2774 cells, using the same primers that we used for the aforementioned genomic PCR experiments. These experiments demonstrated that there were no deletions in the -catenin transcript in either of these cell lines (data not shown). Given that UCI101 and MDAH 2774 cells were CRTpositive, and yet appeared to harbor neither an inactivating APC mutation or an activating -catenin mutation, we decided to pursue the mechanism(s) of CRT activation in these lines. We considered several possibilities. First, it is possible that these two lines harbor inactivating missense point mutations in their APC genes, which would not be detected by Western blot
analysis. Alternatively, these cells could harbor mutations in other genes encoding proteins that are components of the TCF/-catenin pathway. There are several proteins that are believed to cooperate with APC in downregulating TCF/catenin signaling by mediating the rapid destruction of cytoplasmic pools of -catenin. Such proteins include GSK3, axin, and conductin [28 –30]. It is also possible that the CRT observed in these two lines is due to a lesion in a heretoforeunknown gene or genes that are involved in this pathway. In order to determine which of these possibilities might explain the observed CRT in UCI101 and MDAH 2774 cells, we performed additional CRT assays on the three CRT-positive ovarian cancer lines. First, we examined the effect of ectopically expressed APC protein on CRT levels in these cells (Fig. 4A). In all three cases, only a modest reduction in CRT was observed in the presence of a cotransfected wild-type APC cDNA. Furthermore, an equivalent reduction was observed in the presence of a cDNA that encoded a tumor-derived mutant APC protein (1309⌬, [11]), suggesting that this reduction was not a wild-type APC-specific effect. The lack of APC-mediated CRT attenuation in UCI107 cells was not surprising, since cells harboring mutationally activated -catenin alleles have previously been shown to exhibit CRT that is refractory to the introduction of exogenous wild-type APC [11]. The lack of attenuation by APC in UCI101 and MDAH 2774 cells was informative in that it suggested that neither of these cell lines contained an inactivating APC missense point mutation, which would have been corrected by wild-type APC. One cannot formally exclude the possibility that a dominant-negative APC missense mutation was indeed present in one or both of these cell lines, blunting the effect of the ectopically expressed APC protein. In addition to arguing against the presence of an APC missense mutation, these results suggested that the activating lesion in these two cell lines must be downstream of APC in the CRT-activating pathway. To further explore the mechanism underlying the observed CRT in UCI101 and MDAH 2774 cells, we performed an additional set of CRT assays, this time in the presence of a cotransfected dominant-negative TCF4 cDNA (Fig. 4B). This construct, which contains the DNA-binding domain of TCF4, but lacks its amino terminal -catenin interaction domain, has been previously shown to inhibit CRT in colorectal cell lines harboring either an APC or a -catenin mutation [11]. As expected, endogenous CRT activity was markedly inhibited in UCI107 cells, which, as described earlier, harbor an activating -catenin mutation. Surprisingly, however, the endogenous reporter activity in the other two CRT-positive cell lines (UCI101 and MDAH 2774) was not significantly inhibited by the dominant-negative allele. These data suggested that the observed CRT in these 2 cell lines may not be due to genuine endogenous TCF/-catenin activity, but rather due to the activity of a different transcription factor, which could presumably bind to the reporter construct, competing effectively with the ectopically expressed dominant-negative TCF4 protein.
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
101
gel-shift assays on the colorectal cancer cell line HCT116, which exhibits constitutive -catenin/TCF activity as a consequence of an in-frame deletion of a GSK3 phosphorylation site (serine 45) in one allele [11]. As shown in Fig. 5A, we were able to detect optimal TCF probe-specific complexes. Additionally, we were able to demonstrate the presence of -catenin in these complexes (see supershifted band in lane 2). Next, we performed gel-shift experiments on the three CRTpositive ovarian lines. Based on the presence of TCF reporter activity in these lines, we anticipated that nuclear extracts derived from them would contain TCF probe retardation activity. Indeed, this was the case (Fig. 5B, compare lanes 1, 4, and 7). Significantly, the addition of a -catenin monoclonal antibody resulted in the appearance of a supershift in the UCI107 complex, but not in the complexes derived from the other two cell lines (compare lanes 2, 5, and 8). Addition of a control antibody (M2 anti-FLAG) did not induce a supershift in any of the extracts (lanes 3, 6, and 9). These results indicated that the transcription factor complex responsible for binding to the TCF probe contained -catenin in the case of UCI107derived extracts, but not in the case of extracts derived from UCI101 and MDAH 2774 cells. DISCUSSION
FIG. 4. Investigation of the mechanism of CRT induction in UCI-101 and MDAH 2774 ovarian cancer cell lines. (A) Ectopic expression of wild-type APC does not significantly downregulate CRT in ovarian cancer cell lines. Cells were transfected in 6-well plates with a wild-type TCF reporter plasmid (0.5 g) along with 2 g of pCINeo empty vector (treatment 1), wild-type APC expression plasmid (treatment 2), or 1309⌬ mutant APC expression plasmid (treatment 3). SW480 is a colorectal cancer cell line that expresses truncated APC protein. A -galactosidase expression plasmid (0.5 g) was included to control for transfection efficiency. Cell extracts were prepared 30 h posttransfection and assayed for reporter activity. Relative CRT is expressed as the ratio of reporter activity in the APC-transfected cells to the reporter activity of cells transfected with pCINeo expression vector alone. (B) Ectopic expression of dominant-negative TCF4 does not downregulate CRT in UCI101 or MDAH 2774 cell lines. Cells were transfected in 6-well plates with wild-type TCF reporter plasmid (0.5 g) together with 0, 0.5, or 3.0 g of a dominantnegative TCF4 expression plasmid and a -galactosidase expression plasmid (0.5 g). Cell extracts were prepared 30 h posttransfection and assayed for reporter activity. Relative CRT is expressed as the ratio of reporter activity observed in cells transfected with dominant-negative TCF4 plasmid to the reporter activity observed in cells transfected with pCINeo vector alone.
Additionally, these results provided further evidence that the observed CRT in UCI101 and MDAH 2774 cells was not due to a missense APC mutation, since the CRT resulting from such a lesion would be expected to be downregulated by the presence of a dominant-negative TCF4 allele. In order to pursue this possibility further, we performed gel-shift assays with nuclear extracts derived from UCI107, UCI101, and MDAH 2774 cells. We first conducted control
Although it was originally discovered as a membrane-associated protein complexed to E-cadherin, -catenin has recently
FIG. 5. Gel-shift analysis of TCF probe-binding activity in nuclear extracts derived from CRT-positive ovarian cancer cell lines. Nuclear extract preparations and gel-shift analyses were carried out as described under Materials and Methods. (A) Control gel-shift experiment using nuclear extracts derived from HCT116 colorectal cancer cells, which harbor an activating in-frame deletion of serine 45 in one -catenin allele. Arrow denotes supershift induced by addition of anti--catenin antibody to binding reaction. (B) Differential detection of -catenin in TCF probe-binding complexes derived from CRT-positive ovarian cancer lines. Arrow denotes anti--catenin antibodymediated supershift. The M2 anti-FLAG epitope monoclonal antibody was used as a negative control. Comp. denotes nonradiolabeled probe competitor (wt, wild-type; mu, mutant).
102
FURLONG AND MORIN
emerged as an important player in human neoplasia [4, 18, 31]. -catenin can become oncogenic by cooperating with a TCF family member in the activation of key genes such as cyclin D1, c-myc, and others. In order to do so, the cytoplasmic fraction of -catenin must become stabilized in order to facilitate its interaction with a TCF molecule and subsequent shuttling to the nucleus. Cytoplasmic -catenin can be stabilized in at least two ways. The first is the mutational inactivation of the APC gene, whose product has been convincingly shown to be at least in part responsible for the degradation of cytoplasmic pools of -catenin [5, 6]. The second is by point mutation of the -catenin gene, almost always in key amino terminal phosphorylatable serine or threonine residues [11, 32]. These mutations have been shown to render the encoded protein refractory to APC-mediated degradation. While APC mutations are found almost exclusively in colorectal cancer, -catenin mutations have been found in a variety of human cancers, including endometrioid ovarian cancer [18, 19]. We undertook to determine the extent to which the TCF/catenin complex is activated in ovarian cancer. Because this complex could be activated by mutations in genes other than -catenin or APC, we decided to forego a mutational analysis in favor of a reporter assay to functionally screen a series of ovarian cancer cell lines for the presence of endogenous -catenin/TCF-regulated transcriptional activity. While this approach has the advantage of not making any assumptions about which gene or genes may or may not be mutated in the -catenin/TCF pathway, it has the disadvantage of relying upon established cell lines for analysis. One can never be certain that the transcriptional activity and the mutation(s) that cause it were actually present in the original tumor from which the CRT-positive cell lines were derived. Unfortunately, reporter assays, which require transfection of plasmid DNA, cannot be conveniently carried out with primary tumor samples. It is possible to attempt short-term in vitro cultivation of primary tumors prior to transfection, but such cultures are frequently contaminated with nontransformed cells, which may interfere with accurate analysis. The contaminating cells usually die out with prolonged passage in tissue culture, but genetic alterations in the surviving tumor cells may accumulate during this period. It is encouraging to note, however, that -catenin mutations have been found in primary tumors of many tissue origins [18]. In addition, in one case in which matched archival tumor DNA was available for a CRT-positive colorectal cell line (HCT116), a heterozygous -catenin mutation was found in both the cell line and the corresponding primary tumor [11]. This confirmed that at least in this instance, a CRT-inducing mutation had occurred during tumor progression, and not during cell culture propagation. Our reporter screen detected CRT in 21% (4/19) of the ovarian cancer cell lines studied. One of these CRT-positive lines, SW626, harbored a frameshift APC mutation. This was somewhat surprising, since APC mutations are rarely found in
noncolorectal tumors and have never been found in ovarian cancer [33]. Based on the presence of an APC mutation, and upon further molecular characterization of this cell line, we have concluded that SW626 cells may in fact be of colonic origin [20]. Of the remaining three CRT-positive ovarian cancer cell lines, only one (UCI107) had an activating -catenin point mutation. The other two lines (UCI101 and MDAH 2774) did not, suggesting that they may have lesions in another gene or genes involved in this pathway. However, based on further characterization of the observed reporter activity in these lines, we have concluded that the activity observed in the UCI107 cell line is genuine -catenin/TCF-mediated transcriptional activity, whereas the activity observed in the other two ovarian cancer cell lines is likely due to the activity of an unrelated transcription factor complex. This unrelated factor may contain a TCF-like protein, but our gel-shift data suggest that it does not contain -catenin. These results indicate that caution must be used when interpreting the results of -catenin/TCF reporter assays, since they may not always accurately reflect the presence or absence of this transcriptional complex. The presence of apparent TCF-mediated transcriptional repression in six ovarian cell lines was somewhat unexpected, but not without precedent. For example, it has been shown that, in the absence of -catenin, XTcf-3 can act as a transcriptional repressor on the ventral side of Xenopus embryos [34]. This TCF-mediated repression during development is likely to be mediated through the recruitment of the Groucho corepressor [35, 36]. Groucho homologues known as TLE proteins exist in vertebrates and can also repress TCF-mediated transcription [37]. The matrylisin promoter was recently reported to be regulated by TCF/-catenin complexes. Interestingly, mutation of the TCF binding site increased basal activity, suggesting that -catenin binding to TCF may act to overcome its repressing effect on the promoter [10]. However, the relevance of TCF-mediated transcriptional repression in cancer remains unclear. The frequent observation of TCF transcriptional repression in ovarian cancer cell lines may be relevant for ovarian tumorigenesis or may simply reflect low levels of -catenin signaling and/or the presence of Groucho factors in these cells. In summary, constitutive TCF/-catenin activity is found in ovarian cancer cell lines at low frequency. The TCF reporter activity observed in one of the TCF/-catenin-active ovarian lines (UCI107) was shown to be due to an activating mutation in one of its -catenin alleles. Preliminary evidence suggests that the activity observed in two additional lines may in fact not be due to activation of the TCF/-catenin pathway, but rather due to the activity of an unrelated transcription factor complex that is devoid of -catenin. The UCI107 cell line, which was derived from an ovarian papillary adenocarcinoma [38], represents the first ovarian cancer cell line shown to harbor genuine constitutive -catenin/TCF-mediated transcriptional ac-
TCF-MEDIATED TRANSCRIPTION IN OVARIAN CANCER
tivity. Mutations of the -catenin gene in ovarian cancer have previously been associated with endometrioid differentiation [19, 39]. In this report, we show that activation of the TCF/ -catenin pathway is not an obligatory step in the tumorigenesis of even this subtype of ovarian cancer, since the MDAH 2774 cell line, derived from endometrioid ovarian cancer, was found to be CRT-negative. In any event, the UCI107 line should provide a useful system for the study of genes involved in the APC/-catenin pathway as well as the relationship of this pathway to ovarian tumorigenesis. Finally, a substantial fraction of the ovarian cancer cell lines in this study exhibited constitutive TCF-mediated repression as measured by luciferase assays. The relevance of this repression in ovarian tumorigenesis is unclear but is currently under investigation. REFERENCES 1. Kinzler KW, Vogelstein B: Lessons from hereditary colorectal cancer. Cell 87:159 –170, 1996 2. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL: Genetic alterations during colorectal-tumor development. N Engl J Med 319:525–532, 1988 3. Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 61:759 –767, 1990 4. Polakis P: The oncogenic activation of -catenin. Curr Opin Genet Dev 9:15–21, 1999 5. Munemitsu S, Albert I, Souza B, Rubinfeld B, Polakis P: Regulation of intracellular -catenin levels by the adenomatous polyposis coli (APC) tumor-suppressor protein. Proc Natl Acad Sci USA 92:3046 –3050, 1995 6. Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, Kinzler KW, Vogelstein B, Clevers H: Constitutive transcriptional activation by a -catenin-Tcf complex in APC ⫺/⫺ colon carcinoma. Science 275:1784 – 1787, 1997 7. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW: Identification of c-MYC as a target of the APC pathway. Science 281:1509 –1512, 1998 8. Tetsu O, McCormick F: -catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 398:422– 426, 1999 9. Mann B, Gelos M, Siedow A, Hanski ML, Gratchev A, Ilyas M, Bodmer WF, Moyer MP, Riecken EO, Buhr HJ, Hanski C: Target genes of -catenin-T cell-factor lymphoid-enhancer-factor signaling in human colorectal carcinomas. Proc Natl Acad Sci USA 96:1603–1608, 1999 10. Crawford HC, Fingleton BM, Rudolph-Owen LA, Heppner Goss KJ, Rubinfeld B, Polakis P, Matrisian LM: The metalloproteinase matrilysin is a target of -catenin transactivation in intestinal tumors. Oncogene 18: 2883–2891, 1999 11. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, Kinzler KW: Activation of -catenin-Tcf signaling in colon cancer by mutations in -catenin or APC. Science 275:1787–1790, 1997 12. Sparks AB, Morin PJ, Vogelstein B, Kinzler KW: Mutational analysis of the APC/-catenin/Tcf pathway in colorectal cancer. Cancer Res 58: 1130 –1134, 1998 13. Yost C, Torres M, Miller JR, Huang E, Kimelman D, Moon RT: The axis-inducing activity, stability, and subcellular distribution of -catenin is regulated in Xenopus embryos by glycogen synthase kinase 3. Genes Dev 10:1443–1454, 1996 14. Aberle H, Bauer A, Stappert J, Kispert A, Kemler R: -catenin is a target for the ubiquitin-proteasome pathway. EMBO J 16:3797–3804, 1997
103
15. Orford K, Crockett C, Jensen JP, Weissman AM, Byers SW: Serine phosphorylation-regulated ubiquitination and degradation of -catenin. J Biol Chem 272:24735–24738, 1997 16. Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, Harper JW: The SCF -TRCP-ubiquitin ligase complex associates specifically with phosphorylated destruction motifs in IB␣ and -catenin and stimulates IB␣ ubiquitination in vitro. Genes Dev 13:270 –283, 1999 17. Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H, Perret C, Rubinfeld B, Margottin F, Benarous R, Polakis P: The F-box protein -TrCP associates with phosphorylated -catenin and regulates its activity in the cell. Curr Biol 9:207–210, 1999 18. Morin PJ: -catenin signaling and cancer. Bioessays 21:1021–1030, 1999 19. Palacios J, Gamallo C: Mutations in the -catenin gene (CTNNB1) in endometrioid ovarian carcinomas. Cancer Res 58:1344 –1347, 1998 20. Furlong MT, Hough CD, Sherman-Baust CA, Pizer ES, Morin PJ: Evidence for the Colonic Origin of Ovarian Cancer Cell Line SW626. J Natl Cancer Inst 91:1327–1328, 1999 21. Bast RC, Feeney M, Lazarus H, Nadler LM, Colvin RB, Knapp RC: Reactivity of a monoclonal antibody with human ovarian carcinoma. J Clin Invest 68:1331–1337, 1981 22. Buick RN, Pullano R, Trent JM: Comparative properties of five human ovarian adenocarcinoma cell lines. Cancer Res 45:3668 –3676, 1985 23. Smith KJ, Johnson KA, Bryan TM, Hill DE, Markowitz S, Willson JK, Paraskeva C, Petersen GM, Hamilton SR, Vogelstein B, Kinzler KW: The APC gene product in normal and tumor cells. Proc Natl Acad Sci USA 90:2846 –2850, 1993 24. van de Wetering M, Oosterwegel M, Dooijes D, Clevers H: Identification and cloning of TCF-1, a T lymphocyte-specific transcription factor containing a sequence-specific HMG box. EMBO J 10:123–132, 1991 25. Nagase H, Nakamura Y: Mutations of the APC (adenomatous polyposis coli) gene. Hum Mutat 2:425–34, 1993 26. Iwao K, Nakamori S, Kameyama M, Imaoka S, Kinoshita M, Fukui T, Ishiguro S, Nakamura Y, Miyoshi Y: Activation of the -catenin gene by interstitial deletions involving exon 3 in primary colorectal carcinomas without adenomatous polyposis coli mutations. Cancer Res 58:1021– 1026, 1998 27. Whitehead I, Kirk H, Kay R: Expression cloning of oncogenes by retroviral transfer of cDNA libraries. Mol Cell Biol 15:704 –710, 1995 28. Rubinfeld B, Albert I, Porfiri E, Fiol C, Munemitsu S, Polakis P: Binding of GSK3- to the APC--catenin complex and regulation of complex assembly. Science 272:1023–1025, 1996 29. Ikeda S, Kishida S, Yamamoto H, Murai H, Koyama S, Kikuchi A: Axin, a negative regulator of the wnt signaling pathway, forms a complex with GSK-3 and -catenin and promotes GSK-3-dependent phosphorylation of -catenin. EMBO J 17:1371–1384, 1998 30. Behrens J, Jerchow B, Wurtele M, Grimm J, Asbrand C, Wirtz R, Kuhl M, Wedlich D, Birchmeier W: Functional interaction of an axin homolog, conductin, with -catenin, APC, and GSK3 . Science 280:596 –599, 1998 31. Peifer M: -catenin as oncogene: The smoking gun. Science 275:1752– 1753, 1997 32. Rubinfeld B, Robbins P, El-Gamil M, Albert I, Porfiri E, Polakis P: Stabilization of -catenin by genetic defects in melanoma cell lines. Science 275:1790 –1792, 1997 33. Allan GJ, Cottrell S, Trowsdale J, Foulkes WD: Loss of heterozygosity on chromosome 5 in sporadic ovarian carcinoma is a late event and is not associated with mutations in APC at 5q21-22. Hum Mutat 3:283–291, 1994 34. Brannon M, Gomperts M, Sumoy L, Moon RT, Kimelman D: A -catenin/ XTcf-3 complex binds to the siamois promoter to regulate dorsal axis specification in Xenopus. Genes Dev 11:2359 –2370, 1997
104
FURLONG AND MORIN
35. Cavallo RA, Cox RT, Moline MM, Roose J, Polevoy GA, Clevers H, Peifer M, Bejsovec A: Drosophila Tcf and Groucho interact to repress Wingless signalling activity. Nature 395:604 – 608, 1998 36. Roose J, Molenaar M, Peterson J, Hurenkamp J, Brantjes H, Moerer P, van de Wetering M, Destree O, Clevers H: The Xenopus Wnt effector XTcf-3 interacts with Groucho-related transcriptional repressors. Nature 395:608 – 612, 1998 37. Levanon D, Goldstein RE, Bernstein Y, Tang H, Goldenberg D, Stifani S, Paroush Z, Groner Y: Transcriptional repression by AML1 and LEF-1 is
mediated by the TLE/Groucho corepressors. Proc Natl Acad Sci USA 95:11590 –11595, 1998 38. Gamboa G, Carpenter PM, Podnos YD, Dorion G, Iravani L, Bolton D, Mascarello JT, Manetta A: Characterization and development of UCI 107, a primary human ovarian carcinoma cell line. Gynecol Oncol 58:336–343, 1995 39. Wright K, Wilson P, Morland S, Campbell I, Walsh M, Hurst T, Ward B, Cummings M, Chenevix-Trench G: -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