|Home | About | Journals | Submit | Contact Us | Français|
Pancreatic cancer is a deadly disease characterized by late diagnosis and resistance to therapy. Much progress has been made in defining gene defects in pancreatic cancer, but a full accounting of its molecular pathogenesis awaits. Here, we show that expression of Ataxia-Telangiectasia Group D Complementing gene (ATDC), also called TRIM29, is elevated in most invasive pancreatic cancers and pancreatic cancer precursor lesions. ATDC promoted cancer cell proliferation in vitro and enhanced tumor growth and metastasis in vivo. ATDC expression correlated with elevated β-catenin levels in pancreatic cancer, and β-catenin function was required for ATDC’s oncogenic effects. ATDC was found to stabilize β-catenin via ATDC-induced effects on the disheveled-2 protein, a negative regulator of GSK3β in the Wnt/β-catenin signaling pathway.
Pancreatic cancer is an aggressive malignancy, and an improved understanding of the molecular mechanisms governing its highly aggressive behavior is needed for more effective treatment, early detection and prevention. Defects in Wnt/β-catenin signaling are common in certain cancers, such as colorectal carcinoma, and recent evidence suggests Wnt/β-catenin signaling may contribute to pancreatic cancer. In this report, we show that ATDC is over-expressed in the majority of invasive pancreatic cancers and pancreatic cancer precursor lesions. ATDC contributes to pancreatic cancer via its ability to interact with and stabilize expression of disheveled-2, with resultant stabilization of β-catenin. Besides highlighting ATDC as a potential therapeutic target in pancreatic cancer, our studies have defined a novel mechanism for activating Wnt/β-catenin signaling in cancer.
Pancreatic cancer is a highly lethal disease that is often diagnosed in an advanced state. In fact, though it is the fourth most common cause of cancer death per year in the United States, affecting 37,000 people (Jemal et al., 2007), it has the worst prognosis of any major malignancy (<5% 5-year survival rate). Recent advances in surgical and medical therapy have had only a modest impact on pancreatic cancer mortality. A major hallmark of pancreatic cancer is extensive local tumor invasion and early systemic dissemination. Pancreatic cancer is also notoriously resistant to chemotherapy and ionizing radiation.
To further understand the molecular pathogenesis of pancreatic cancer, genomic and proteomic profiling has been performed to identify differentially expressed genes and proteins that might represent novel therapeutic targets (Cao et al., 2004; Chen et al., 2005; Crnogorac-Jurcevic et al., 2003; Logsdon et al., 2003; Lowe et al., 2007; Mutter et al., 2001). Using Affymetrix gene expression profiling, we recently found that pancreatic cancer cells over-express the Ataxia-Telangiectasia Group D Complementing gene (ATDC) at an average level 20-fold higher than epithelial cells from normal pancreas or chronic pancreatitis-derived tissues (Logsdon et al., 2003). ATDC was initially described in the hunt for the gene responsible for the genetic disorder ataxia telangiectasia (AT) (Kapp et al., 1992). Both alleles of the ATDC gene in an AT patient-derived cell line were found to contain early stop codon mutations, leading to a truncated and nonfunctional ATDC protein (Tauchi et al., 2000). No germline ATDC mutations were found in AT patients, and subsequently the gene responsible for AT was identified as ATM (ataxia telangiectasia mutated) (Savitsky et al., 1995).
The ATDC gene, located at chromosome 11q23, encodes a 588 amino acid protein with multiple zinc finger motifs and an adjacent leucine zipper motif, that may allow the ATDC protein to form homo- or hetero-dimers. (Savitsky et al., 1995). Northern blot analysis revealed that ATDC is normally expressed in placenta, lung, thymus, prostate, testis, and colon, while no expression is observed in heart, brain, skeletal muscle, pancreas, spleen, ovary, or small intestine (Hosoi and Kapp, 1994). In addition to our observation of high ATDC levels in primary pancreatic cancers and pancreatic cancer cell lines, a search of the Oncomine database (www.oncomine.org) revealed that ATDC has been reported to be overexpressed in lung (Hawthorn et al., 2006), bladder (Dyrskjot et al., 2004), colorectal (Glebov et al., 2006; Ohmachi et al., 2006), ovarian (Santin et al., 2004), endometrial cancers (Mutter et al., 2001) and multiple myeloma (Zhan et al., 2002), with apparent reduced expressionin melanoma (Smith et al., 2005), breast (Nacht et al., 1999), head and neck (Zhang et al., 2006) and prostate cancer (Dyrskjot et al., 2004; LaTulippe et al., 2002; Luo et al., 2001; Mutter et al., 2001; Yu et al., 2004; Zhan et al., 2002). A recent report identified a correlation between ATDC expression in gastric cancer and poor histological grade, large tumor size, extent of tumor invasion, and lymph node metastasis (Kosaka et al., 2007).
The ATDC protein, also known as TRIM29, is a member of the tripartite motif (TRIM) family. TRIM proteins have a series of conserved domains, which include a RING (R), a B-box type 1 (B1) and B-box type 2 (B2), followed by a coiled-coil (CC) region. While some of the domains may be absent or present in the different TRIM proteins (ATDC contains the B1-B2-CC domains but lacks the R domain), their order is always maintained (R-B1-B2-CC) (Reymond et al., 2001). Proteins belonging to the TRIM family have been implicated in a variety of cellular processes, such as development and growth, and in several human diseases, including HIV infection (Stremlau et al., 2004) and leukemia (Goddard et al., 1991). The function of ATDC has not been studied previously in any physiologic or pathologic process, though the ATDC protein appears to be localized primarily to the cytoplasm (Reymond et al., 2001). In this report, we describe our data examining the expression and functional role of ATDC in pancreatic cancer.
To identify genes with potential roles in the development and progression of pancreatic adenocarcinoma, we assessed gene expression in microdissected samples of human pancreatic carcinoma using Affymetrix arrays. We compared the expression patterns in cancer tissues to those seen in normal pancreas and chronic pancreatitis, the latter which served as an important control for the extensive fibrosis typically observed with pancreatic cancer (Logsdon et al., 2003). One of the most highly up-regulated genes was the ATDC gene, which showed elevated expression in 10 out of 10 samples of pancreatic cancers relative to normal pancreas or chronic pancreatitis samples (Figure 1A). On average, ATDC expression was roughly 20-fold higher in pancreatic adenocarcinomas. The gene expression data were further confirmed in an analysis of ATDC mRNA levels of pancreatic cancer using quantitative real time PCR (qRT-PCR) (Figure 1B). Immunohistochemical staining confirmed that ATDC protein expression was present in the neoplastic epithelium of pancreatic cancer (Figure 1C).
A progression model of pancreatic cancer is now widely accepted in which normal pancreatic ductal epithelium progresses to infiltrating cancer through a series of morphologically defined pancreatic precursors called PanINs (Hruban et al., 2000). This progression is associated with accumulation of specific genetic changes, such as K-ras mutations and inactivation of p16, that are observed in invasive pancreatic cancer. We found that ATDC was not expressed in PanIN 1 (0/4) lesions, but was occasionally expressed in PanIN 2 (1/7) lesions and was more often expressed in PanIN 3 lesions (3/6) (Figure 1D). These data suggest that up-regulation of ATDC occurs prior to the development of invasive pancreatic cancer.
To understand the function of ATDC in pancreatic cancer, we first explored the effect of ATDC ectopic expression on cellular growth in vitro in multiple cell lines with differing levels of endogenous ATDC expression. Following transfection with an ATDC cDNA expression construct, HEK 293 cells, which normally do not express ATDC, and MiaPaCa2 pancreatic cancer cells, which express low endogenous levels of ATDC, demonstrated a significant increase in cellular proliferation (Figure 2A and 2B). Similar changes were observed in monoclonal and polyclonal HEK 293 cells lines stably overexpressing ATDC (Supplemental Figure S1). Conversely, cellular proliferation was attenuated when endogenous ATDC expression was silenced by stable transfection with 2 different shRNA vectors targeting distinct regions of ATDC in Panc1 and BxPC3 pancreatic cancer cell lines, both of which have high endogenous levels of ATDC (Figure 2C and 2D). Expression of ATDC shRNA1 and 2 did not alter basal cell proliferation rates in HEK 293 cells (Supplemental Figure S2), verifying the specificity of the inhibitory function of the ATDC shRNAs on ATDC’s function.
To examine the effects of ATDC silencing on pancreatic tumor growth in vivo, we infected Panc1 cells (expressing a control shRNA or ATDC shRNA1) with a luciferase-expressing lentivirus. Following injection of 5 × 105 cells into the tail of the pancreas, tumor growth was assessed using bioluminescent imaging (n= 8 animals per group). All of the animals injected with Panc1 cells expressing control shRNA demonstrated tumor formation 14 days post-injection, while tumors were not detected in the animals injected with Panc1 ATDCshRNA cells (Figure 2E, left panels). At 60 days post-injection, the control shRNA animals tumors grew significantly larger, with evidence of metastatic spread, while only 25% (2/8) of the ATDCshRNA animals demonstrated evidence of macroscopic tumors (Figure 2E, right panels and 2F). The mean tumor volume was significantly larger in the tumors grown in mice injected with Panc1 cells expressing control shRNA compared to mice injected with Panc1 cells expressing ATDC shRNA (203.2 ± 68.8 vs. 2.01 ± 1.0mm3, *p<0.001) (Figure 2G). These data support the role of ATDC in promoting growth of pancreatic cancer cells.
In exploring possible downstream mediators of ATDC’s growth-promoting effects, we noted that over-expression of ATDC resulted in a significant increase in β-catenin levels in HEK 293 cells (Figure 3A). In addition, the over-expression of ATDC increased expression of the under- or non-phosphorylated forms of β-catenin (“active” β-catenin) presumed to be the forms of β-catenin responsible for mediating Wnt signaling in cells (van Noort et al., 2002) (Fig. 3A). Similar results were obtained in Mia PaCa2 pancreatic cancer cells ectopically expressing ATDC (Figure 3B). Conversely, ATDC shRNA1- or 2-mediated silencing of endogenous ATDC expression significantly decreased total β-catenin and active β-catenin levels in Panc1 cells (Figure 3C, 3E) and BxPC3 cells (Figure 3D, 3F).
The active form of β-catenin exerts its growth promoting effects by translocating from the cytoplasm to the nucleus, where it binds to transcription factors such as the TCF/lymphoid enhancer binding factor and thereby stimulates the transcription of Wnt target genes (Clevers, 2006). Over-expression of ATDC in HEK 293 and MiaPaCa2 cells increased expression of the Wnt/β-catenin target genes c-Myc and DKK1 (Figure 3A, 3B), while c-Myc and DKK1 levels were significantly reduced in Panc1 and BxPC3 cells expressing ATDC-targeting shRNA1 or 2 (Figure 3C, 3D). Consistent with the ability of ATDC to increase β-catenin levels in HEK 293 cells, we found that ATDC strongly activated the TCF-dependent TOP-FLASH reporter construct in HEK 293 cells (Figure 3G). Together, these results strongly suggest that ATDC enhances β-catenin levels and activates β-catenin/TCF target gene expression in pancreatic cancer cells.
To address in more detail the means by which ATDC increased β-catenin/TCF-regulated gene expression, we assessed the abundance of the free pool of β-catenin in ATDC-expressing cells compared to control cells. To measure the free pool of β-catenin, we utilized a recombinant GST fusion protein containing the cytoplasmic tail of E-cadherin (GST-Ecad). It has been previously shown that GST-E-cadherin can readily be used to monitor the abundance of the free pool of β-catenin that is stabilized in response to activation of the Wnt signaling pathway (Winer et al., 2006). As show in Figure 3H, following ectopic expression of ATDC or the S33Y mutant form of β-catenin in HEK 293 cells, significant increases in the levels of the free β-catenin pool were seen, as demonstrated by the recovery of β-catenin with GST-E-cadherin. In contrast, no significant β-catenin was recovered from control cell lysates following incubation with the GST-E-cadherin.
To confirm that the increase in the free pool of β-catenin seen with ATDC overexpression was associated with increased cytoplasmic and nuclear levels of β-catenin, extracts from control and ATDC-overexpressing HEK 293 cells were separated into membrane, cytoplasmic, and nuclear fractions and the relative abundance of β-catenin in these fractions analyzed. ATDC overexpression increased both the nuclear and cytoplasmic fractions of β-catenin, while having no effect on the membrane bound fraction of ATDC (Figure 3I). Previous reports have demonstrated an important role for β-catenin in cell adhesion as a part of a protein complex that includes E-cadherin. The E-cadherin expression pattern in pancreatic cancer cells was unaltered with ATDC shRNA (Supplemental Figure S3), suggesting that increased levels of ATDC affect β-catenin levels and β-catenin/TCF-dependent transcription in a fashion similar to that seen following activation of the canonical Wnt signaling pathway.
We next sought to determine if the growth-promoting effects of ATDC were mediated by activation of the β-catenin signaling pathway. Constitutive activation of Wnt/β-catenin signaling by mutations in known Wnt pathway components, such as inactivating mutations in the APC (adenomatous polyposis coli) or Axin1 tumor suppressor genes, or activating mutations in β-catenin are commonly seen in certain cancers, such as colorectal or hepatocellular carcinomas, but are rarely seen in pancreatic adenocarcinoma (Gregorieff and Clevers, 2005; Lustig and Behrens, 2003). However, constitutive activation of β-catenin/TCF–dependent transcription, independent of mutations, has been suggested to play an important role in the development of certain human breast and ovarian cancers (Bafico et al., 2004) and in mouse models of basal cell carcinoma (Hoseong Yang et al., 2008) and pancreatic cancer (Pasca di Magliano et al., 2007). To address the contribution of β-catenin/TCF transcription in the growth-promoting effects of ATDC, we ectopically expressed a β-catenin shRNA or a dominant negative TCF (dnTCF) protein in either control vector-transfected or ATDC-expressing HEK 293 and MiaPaCa2 cells. Transfection of either the β-catenin shRNA or dnTCF constructs significantly inhibited ATDC-induced TOPFLASH reporter activity (Figures 4A and 4B, upper panels) and cell proliferation (Figures 4A and 4B, lower panels) in HEK 293 and MiaPaCa2 cells. Similarly, transfection of β-catenin-targeting shRNA or dnTCF in Panc1 cells inhibited TOPFLASH reporter activity (Figure 4C, upper panel) and inhibited the enhanced cellular proliferation seen in Panc1 cells in the setting of elevated endogenous ATDC levels (Figure 4C, lower panel).
To examine the effects of β-catenin silencing on pancreatic tumor growth in vivo, we infected control or β-catenin shRNA-transduced Panc1 cells with a luciferase-expressing lentivirus. Following injection of 5 × 105 cells into the tail of the pancreas, tumor growth was assessed using bioluminescent imaging (n= 8 animals per group). All of the animals injected with Panc1 cells expressing control shRNA demonstrated tumor formation 14 days post-injection, while tumors were not detected in the animals injected with Panc1 cells expressing the β-catenin shRNA (Figure 4D, left panels, with three representative animals shown). At 60 days post-injection, the control tumors grew significantly larger, with evidence of metastatic spread, while the extent of primary tumor size and metastasis was markedly diminished in animals injected with Panc1 cells expressing the β-catenin shRNA (Figure 4D, right panels), similar to the effects observed with silencing ATDC in Panc1 cells (Figure 2E). Western blot analysis of tumors derived from β-catenin shRNA-transfected Panc1 cells harvested at 60 days after injection (Figure 4E) demonstrated effective silencing of β-catenin (Figure 4E). The mean tumor volume was significantly larger in the mice injected with Panc1-Luc cells expressing control shRNA compared to mice injected with Panc1-Luc cells expressing β-catenin shRNA (251.5±79.2 vs. 35.2±7.8 mm3, respectively, *p<0.05) (Figure 4F). Overall, these data indicate that ATDC plays a key role in cell proliferation and tumor growth, and ATDC’s growth promoting effects are dependent, at least in part, on the β-catenin/TCF signaling pathway.
Based on our findings that ATDC is highly expressed in human pancreatic cancers and ATDC-mediated cancer cell growth in pancreatic cancer cells is dependent on β-catenin signaling, we predicted that ATDC and β-catenin expression levels would be well-correlated in pancreatic cancer cell lines and primary pancreatic cancers. As shown in Figure 5A, the levels of ATDC and β-catenin in pancreatic cancer cell lines correlated well, with the highest levels of β-catenin observed in BxPC3 cells, which have high levels of endogenous ATDC. Intermediate levels of β-catenin and ATDC proteins were found in Panc1 cells, and both β-catenin and ATDC were minimally expressed in MiaPaCa2 cells. To study further the relationship of ATDC and β-catenin expression in pancreatic cancer, immunohistochemical analysis of ATDC and β-catenin in human pancreatic normal and adenocarcinoma tissues was performed. We did not observe ATDC staining in normal human pancreas, but did observe expression of β-catenin localized to the cell membrane (Figure 5B, left panels), as has been previously described (Pasca di Magliano et al., 2007). It is commonly thought that both cytoplasmic and nuclear localization of β-catenin is an indicator of active β-catenin signaling in the Wnt pathway (Fodde and Brabletz, 2007). To assess the correlation between ATDC expression and β-catenin, we analyzed a pancreatic cancer tissue microarray (TMA) containing 47 pancreatic carcinoma samples. We observed focal nuclear staining for β-catenin in 5 (11%) tumors and elevated cytoplasmic levels of β-catenin in 24 (51%) tumors. These results correlate well with those previously published on β-catenin expression in pancreatic cancer (Pasca di Magliano et al., 2007). We observed a strong correlation between ATDC and β-catenin expression in the 47 pancreatic carcinoma cases, with moderate to high expression of ATDC in the cancer samples that showed elevated cytoplasmic and/or nuclear β-catenin expression (Supplemental Table 1). Representative samples of pancreatic cancers shown in Figure 5B (middle and right panels) demonstrate the correlation in ATDC and β-catenin expression in cancer samples with moderate and high levels of ATDC. Importantly, we did not find any evidence of pancreatic cancers expressing elevated levels of β–catenin without expressing high levels of ATDC, suggesting that elevated cytoplasmic and nuclear levels of β-catenin expression may in fact be dependent on elevated expression of ATDC.
We were interested in defining the mechanism by which elevated levels of ATDC resulted in increased β-catenin expression. In the absence of Wnt ligands, cytoplasmic levels of β-catenin are regulated by a multi-protein complex, termed the destruction complex and which contains the Axin, APC, and glycogen synthase kinase 3-β (GSK3β). Axin and APC are believed to facilitate efficient phosphorylation of β-catenin by GSK3β at multiple serine and threonine residues in its N-terminus. Phosphorylated β-catenin is then ubiquinated, leading to its rapid proteosomal degradation (Gordon et al., 2006). To determine if ATDC increased β-catenin levels by stabilizing β-catenin, Panc1 cells (with or without ATDC silencing) were incubated with cycloheximide (CHX) (10 μg/ml) to prevent new β-catenin synthesis and β-catenin levels were then measured, with levels reflective of the rate of β-catenin protein degradation. Silencing of ATDC expression significantly increased the β-catenin degradation rate, resulting in a marked reduction in the levels of remaining β-catenin (Supplemental Figure S4A, B). This effect of ATDC on β-catenin stability was validated by performing a [35S]-methionine pulse-chase assay. We found that β-catenin was rapidly degraded in Panc1 cells where ATDC expression had been silenced compared to control Panc1 cells (Figure 6A, B).
To ascertain the mechanism by which ATDC stabilizes β-catenin, we first examined whether ATDC interacted with the components of the multi-protein complex that regulates β-catenin stability. In co-immunoprecipitation experiments using antibodies directed to Axin, GSK-3β and β-catenin, evidence of complexes containing ATDC and Axin and complexes containing ATDC and GSK-3β were observed in HEK 293 cells transfected with a Flag-tagged ATDC construct (Figure 6C). To verify that this interaction was physiologically relevant, we tested the ability of endogenous ATDC in BxPC3 cells to interact with Axin and GSK3β and similar results were obtained (Figure 6D). ATDC did not interact with β-catenin in either HEK 293 cells over-expressing ATDC or in BxPC3 cells (Figures 6C, 6D). These results are consistent with the possibility that ATDC may interfere with GSK-3β– dependent phosphorylation of β-catenin by the destruction complex.
Based on the findings, we examined if ATDC might activate disheveled (Dvl), a cytoplasmic protein that is activated by binding of Wnt ligands to the frizzled/LRP co-receptor at the cell surface. Activated Dvl binds to the Axin/GSK3β complex and antagonizes GSK-3β dependent phosphorylation of β-catenin in a manner not dissimilar to what we observed with ATDC. Indeed, we found increased Dvl-2 levels (total and phosphorylated forms) in HEK 293 cells over-expressing ATDC as well as in Panc1 and BXPC3 cells (Figures 6E–G). Furthermore, silencing of endogenous ATDC expression in Panc1 and BxPC3 cells reduced the expression of Dvl-2 (Figures 6F, 6G). Increased levels of Dvl-2 were present in primary pancreatic cancer samples that had elevated expression of ATDC (Figure 6H). Finally, immunohistochemical analysis revealed co-localization of ATDC and Dvl-2 in ATDC-transfected HEK 293 cells and Panc1 cells (Supplemental Figure S5). Co-immunoprecipation experiments in HEK 293 cells and BxPC3 cells demonstrated that ATDC formed a complex with Dvl-2 (Figures 6I, 6J). Complex formation between ATDC and Dvl-1 or Dvl-3 in either ATDC-transfected HEK 293 or BxPC3 cells was not seen (data not shown). To determine which region of the ATDC molecule interacted with Dvl-2, we created a series of Flag-tagged ATDC truncation mutants and performed co-immunoprecipitation experiments. We found that the ATDCΔ260 deletion mutant was able to interact with Dvl-2 while the ATDCΔ348 mutant did not, suggesting that amino acids 260–348 of ATDC, in, a region which contains a coiled-coil domain, interacted with Dvl-2 (Supplemental Figure S6).
We then tested whether modulating the levels of ATDC in cells affected the abundance of Dvl-2 transcripts. Dvl-2 gene expression, as measured by quantitative real time RT-PCR, was not altered in either HEK 293 cells overexpressing ATDC or in Panc1 cells with ATDC knockdown compared to control cells (Supplemental Figure S7). To examine if ATDC might be altering Dvl-2 levels by affecting protein stability. [35S]-methionine pulse-chase assays were performed. Dvl-2 was rapidly degraded in Panc1 cells with ATDC knockdown compared to control Panc1 cells (6K,6L), suggesting that ATDC forms a complex with Dvl-2 and increases Dvl-2 levels by regulating Dvl-2 post-transcriptionally, perhaps via direct effects on Dvl-2 protein stability.
To verify that the ability of ATDC to increase TCF activity and cellular proliferation is mediated through Dvl-2, we transfected two Dvl-2 shRNA constructs targeting different regions of Dvl-2 into HEK 293 and MiaPaCa2 cells and examined the effects of Dvl-2 knockdown on ATDC function. Both Dvl-2 shRNA1 and 2 were effective in knocking down levels of Dvl-2 in HEK 293 and MiaPaCa2 cells (Supplemental Figure S8). Knockdown of Dvl-2 in both HEK 293 and MiaPaCa2 cells overexpressing ATDC inhibited increases in β-catenin levels (Figures 7A,7B), TCF activity (Figures 7C,7D) and cell proliferation (Figure 7E,7F) induced by ATDC.
If ATDC stabilizes β-catenin levels by acting on upstream signaling events mediated by Dvl-2, then modulating ATDC levels should not influence β-catenin levels in cell lines with either APC or β-catenin mutations. Indeed, over-expression of ATDC in the APC mutant SW480 and DLD-1 cell lines or the β-catenin mutant HCT-116 cell line did not alter TCF activity or cell proliferation as compared to control cells (7G and 7H). Our findings also predicted that the decrease in TOPFLASH reporter caused by silencing of ATDC in Panc1 cells would be reversed by increasing the levels of β-catenin in the cells. Silencing of ATDC in Panc1 cells significantly inhibited the increase in TOPFLASH reporter activity seen in control Panc1 cells and transfection with wild type or a mutant oncogenic form of β-catenin (S33Y β-catenin) was able to reverse the inhibitory effect of silencing of ATDC on TOPFLASH reporter activity (Figure 7I) and cell proliferation (Figure 7J) in a dose-dependent fashion.
Our studies have identified ATDC as a protein highly expressed in the majority of human pancreatic adenocarcinomas and pancreatic cancer precursor lesions. We also demonstrated that expression of ATDC in pancreatic cancer cells promoted cellular proliferation and enhanced tumor growth and metastasis. Additionally, we provided evidence that elevated levels of ATDC expression correlated with elevated β-catenin levels in pancreatic cancer cell lines and primary pancreatic cancers, and that silencing of ATDC via shRNA approaches antagonized β-catenin/TCF-mediated reporter activation and activation of TCF target genes. β-catenin was implicated in the oncogenic effects of ATDC in vitro and in vivo, and the ability of ATDC to increase β-catenin levels appeared to be attributable to ATDC’s effects on disheveled-2 protein expression. In summary, our findings implicate ATDC as an important positive regulator of β-catenin-dependent signaling in pancreatic cancer.
ATDC has been reported to be up-regulated in a number of different cancer types, including lung, bladder, colorectal, ovarian and endometrial cancers and multiple myeloma (Dyrskjot et al., 2004; Glebov et al., 2006; Hawthorn et al., 2006; Mutter et al., 2001; Ohmachi et al., 2006; Santin et al., 2004; Zhan et al., 2002). A recent report correlated ATDC expression in gastric cancer and poor histological grade, large tumor size, extent of tumor invasion, and lymph node metastasis (Kosaka et al., 2007). ATDC has also been reported to be down-regulated in some cancer types (Smith et al., 2005, Nacht et al., 1999; Zhang et al., 2006; Ernst et al., 2002) suggesting the function of ATDC may be depend on cellular context. In none of these reports was the role of ATDC in tumorigenesis examined in functional studies.
We found that ATDC upregulated β-catenin levels in pancreatic cancer cell lines and primary pancreatic cancers. A large body of data supports the contribution of activation of the canonical (β-catenin-dependent) Wnt signaling pathway in the development of colorectal cancer. Sustained β-catenin pathway activation independent of APC, Axin1 or β-catenin mutations has been demonstrated in a subset of breast and ovarian cancer (Bafico et al., 2004). Mutations in APC or β-catenin appear to be rare in pancreatic adenocarcinoma (Zeng et al., 2006). While robust activation of the pathway due to signature mutations in components of the β-catenin signaling cascade that are commonly observed in other gastrointestinal cancers are not present in pancreatic adenocarcinoma, immunohistochemical analysis of β-catenin suggests a possible contribution of β-catenin signaling during PanIN progression and the development of invasive pancreatic cancer. Increased levels of both cytoplasmic and nuclear β–catenin, indicative of β-catenin signaling activity, have been reported in a substantial group of pancreatic adenocarcinomas and PanIN s (Zeng et al., 2006; Pasca di Magliano et al., 2007). Pasca di Magliano and colleagues reported that 65% of pancreatic cancers have an increase in either cytoplasmic and/or nuclear β-catenin. Similar results were obtained in Pdx-Cre;KrasG12D and Pdx-cre;KrasG12D;p53f/+ transgenic mice that developed PanIN lesions and subsequent invasive pancreatic cancers that were phenotypically indistinguishable from human pancreatic adenocarcinomas. Moreover, the authors showed that cancer cell survival and proliferation, depended on β-catenin signaling activity in multiple human pancreatic cancer cell lines.
The mechanisms by which ATDC levels are upregulated in pancreatic adenocarcinomas remain unclear. Interestingly, Pasca di Magliano and colleagues demonstrated that increased hedgehog signaling, one of the earliest changes in PanIN lesions, activated β-catenin signaling in transgenic mice and untransformed pancreatic duct cells, suggesting that hedgehog may play a role in upregulating β-catenin activity in some pancreatic adenocarcinomas. A possible connection between hedgehog upregulation and ATDC expression in human pancreatic cancer cell lines or primary tumors remains to be explored.
The levels of the free cytoplasmic pool of β-catenin are known to be regulated by Wnt ligands. In the absence of an activating Wnt signal, mediated via Wnt binding to the frizzled-low density lipoprotein-related (LRP) 5/6 co-receptor complex, cytoplasmic β-catenin is destabilized by a multiprotein complex containing axin, GSK3β, and APC. Axin acts as the scaffold of this complex and interacts with the other components- β catenin, APC, and GSK3β. Interaction of GSK3β with Axin in the complex facilitates efficient phosphorylation of β-catenin by GSK3β. Phosphorylated β-catenin is then ubiquinated, leading to its rapid proteosomal degradation. We found that ATDC bound to Axin and GSK-3β in pancreatic cancer cells, suggesting that ATDC interacted with the destruction complex to prevent phosphorylation and subsequent ubiquination of β-catenin.
When Wnt binds to the frizzled/LRP co-receptors at the cell surface, a cytoplasmic protein, Dvl, antagonizes GSK-3β dependent phosphorylation of β-catenin. Although it is not known if Dvl binds directly to the frizzled/LRP co-receptor or whether intermediary proteins are involved in the signal transduction between frizzled and Dvl, Dvl appears to bind to axin and inhibit GSK-3β-dependent phosphorylation of β-catenin, APC and axin. Once the phosphorylation of β-catenin is reduced, β-catenin dissociates from the axin complex, resulting in its accumulation in the cytoplasm. Once stabilized, a fraction of the β-catenin is translocated to the nucleus, where it binds to transcription factors such as the TCF/lymphoid enhancer binding factor and thereby stimulates the transcription of β-catenin target genes. We noted that ATDC expression in HEK 293 cells induced expression of Dvl-2, and demonstrated that ATDC formed a complex with Dvl-2 in pancreatic cancer cells. We demonstrated that levels of ATDC in primary pancreatic cancers correlated well with Dvl-2 levels, suggesting that ATDC upregulates Dvl-2 levels in primary pancreatic cancers, and increases β-catenin levels by this mechanism. We further show that knockdown of Dvl-2 in ATDC expressing cells abrogates enhanced TCF activity and cell proliferation induced by ATDC, directly implicating Dvl-2 as an intermediary in this process.
The regulation of the disheveled protein is still poorly understood, but recent data suggest that disheveled, like β-catenin, may be controlled by ubiquination and degradation by the proteosome (Hershko et al., 1998; Simons et al., 2005; Miyazaki et al., 2004). Dvl-1 has been reported to interact with the neuronal Homologous to E6AP carboxyl terminus (HECT)-type ubiquitin ligase NEDL1 (Miyazaki et al., 2004). The proteins inversin and the interactions between PP2A phosphatase and the Wnt-induced antagonist naked cuticle have been shown to modulate the stability of Dvl-1 (Simons et al., 2005; Creyghton et al., 2005). And, in a recent manuscript published by Angers and colleagues, KLHL12-Cullin-3 ubiquitin ligase was shown to negatively regulate the Wnt-β-catenin pathway by targeting disheveled for degradation (Angers et al., 2006). We found that ATDC did not increase Dvl-2 levels by changes in Dvl-2 gene expression but rather by enhancing the stability of the Dvl-2 protein, supporting changes in Dvl stability serve as a important mechanism in regulating the Wnt/β-catenin signaling pathway.
The data in the present study support a model for the mechanism by which ATDC functions to promote the oncogenesis of pancreatic cancer cells (Figure 8). In unstimulated, normal pancreatic cells lacking ATDC, Dvl-2 is in the cytoplasm and is not bound to the Axin/Gsk-3β/APC destruction complex. This allows the destruction complex to phosphorylate β-catenin and target it for ubiquitin-mediated degradation. In pancreatic cancer cells expressing high levels of ATDC, ATDC binds to and stabilizes Dvl-2, resulting in the release of β-catenin from the destruction complex, increased β-catenin levels and subsequent activation of downstream β-catenin/TCF-regulated target genes These studies define for the first time a functional role of ATDC in human tumorigenesis and besides highlighting ATDC as a potential therapeutic target in pancreatic cancer, have defined a novel mechanism for activating Wnt/β-catenin signaling in cancer.
The human pancreatic ductal adenocarcinoma cell lines Panc1, Mia PaCa2 and BxPC3 and the human embryonic kidney cell line HEK 293 were purchased from American Type Culture Collection (ATCC, Manassas, VA). Fresh frozen and paraffin-embedded human pancreatic tissues were obtained from patients undergoing surgical resection at the University of Michigan Medical Center. All human samples were obtained following approval by the Institutional Review Board of the University of Michigan Medical Center. A pancreatic cancer tissue microarray constructed by the University of Michigan Medical Center’s Tissue Procurement Facility contained 5 samples of normal pancreas and 47 samples of pancreatic adenocarcinoma and was used for immunohistochemical analysis of ATDC expression. Twenty-five paraffin-embedded human pancreatic tissue samples containing pancreatic intraductal neoplasia (PanIn) lesions of various stages were analyzed for ATDC expression using an anti-ATDC antibody. ATDC expression in PanIn lesions was evaluated by an experienced pancreatic pathologist and graded as absent or present.
The complementary DNA (cDNA) of human ATDC (kindly provided by J. Murnane, University of California, San Franscisco) was subcloned into the pcDNA3.1 expression vector (Invitrogen, San Diego, CA). Sequence analysis after cloning showed 100% homology to the published sequence of ATDC. The methods used to create the control shRNA and ATDC shRNA constructs are listed in the Supplemental Experimental Procedures Endogenous Dvl-2 was knocked down by transient transfection of specific Dvl-2 shRNA1 and 2 (OriGene Technologies, Rockville, MD) in HEK 293 and Mia PaCa2 cells with or without ATDC overexpression. A reporter plasmid carrying three TCF binding sites upstream of a minimal c-fos promoter driving the firefly luciferase gene (TOP Flash), the plasmid carrying the mutated TCF binding sites upstream of a minimal c-fos promoter driving luciferase expression (FOP Flash), and the expression constructs of wild type or constitutively active β-catenin containing a missense mutation of tyrosine for serine at codon 33 (S33Y) were used (Caca et al., 1999). The PGEX-2T vector containing C-terminal E-cadgherin/glutathione S-transferase (GST) was generated as previously described (Winer et al., 2006).
Details regarding the creation of HEK 293 and MiaPaCa2 cell lines with stable expression of ATDC and Panc1 and BxPC-3 cells with silencing of ATDC using shRNA can be found in the Supplemental Experimental Procedures.
Total RNA from the human pancreatic ductal adenocarcinoma, normal pancreas and chronic pancreatitis specimens or Panc1 cells (with or without ATDC shRNA) were isolated using TRIzol reagent (Invitrogen, Carlsbad, CA). To avoid amplification of genomic DNA, RNA was pretreated with DNase (DNA-free kit from Applied Biosystems, Foster City, CA). cDNA synthesis was performed using a High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). The cDNA served as a template in quantitative real-time PCR utilizing TaqMan Fast Universal PCR Master Mix and TaqMan Gene Expression assay probes for ATDC (Hs00232590_m1), Dvl-2 (Hs01005253_ml) or ribosomal protein S6 (RPS6)(Hs02339423_g1) (Applied Biosystems Foster City, CA) and an ABI 7500 Fast Sequence Detection System. All reactions were done in triplicate. ATDC mRNA expression of different group specimens were normalized to endogenous ribosomal protein S6. Ct is the threshold cycle value defined as the fractional cycle number at which the target fluorescent signal passes a fixed threshold above baseline. Relative ATDC mRNA levels were presented as unit values of 2−ΔCt =2−(Ct (Ribosomal protein S6) −Ct (ATDC)).
The paraffin-embedded pancreatic tissue sections (4-μm-thick) were cut, deparaffinized, and subjected to a heat-induced epitope retrieval step. Endogenous peroxidase activity was blocked with 1% (v/v) hydrogen peroxide in distilled water. Details of the immunohistochemical analysis can be found in the Supplemental Experimental Procedures.
Pancreatic cancer cells (with or without β-catenin silencing) or HEK 293 and Mia PaCa2 cells (with or without ATDC overexpression) were transfected using the LipofectAMINE 2000 Transfection kit (Invitrogen Carlsbad, CA) according to the manufacturer’s instructions with 0.2 μg of the TOPFLASH or FOPFLASH reporter constructs with or without varying amounts (0.2–2 μg) of constructs expressing either wild type β-catenin or the S33Y β-catenin (constitutively activated) mutant and 25 ng of a β-galactosidase construct as an internal control to normal luciferase activity to tranfection efficiency. The total DNA for each transfection was kept constant by adding empty pcDNA vector. Forty eight hours after transfection, luciferase activity was measured in a luminometer and normalized to β-galactosidase expression. Endogenous β-catenin or Dvl-2 was silenced by stably expression β-catenin shRNA or transient expression of Dvl-2 shRNAs (OriGene Technologies, Rockville, MD). TCF transcriptional activity was inhibited by transiently transfection of the dnTCF vector (Upstate, Temecula, CA).
Nuclear and cytoplasmic proteins were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce, Rockford, IL). Membrane proteins were extracted using the Mem-PER Eukaryotic Membrane Protein Extraction Kit (Pierce, Rockford, IL) according to the manufacturer’s instructions.
The free pool of β-catenin in ATDC transfected and control HEK 293 cells was measured using a C-terminal E-cadherin/glutathione S-transferase fusion protein as previously described (Winer et al., 2006).
Cell proliferation was measured using a CellTiter 96 AQ nonradioactive cell proliferation assay (Promega, Madison, WI) as we have previously described (Zhang et al., 2004).
Cells were lysed by sonicating for 5 sec in 1 ml of detergent free lysis buffer (PBS, 5 mM EDTA, 0.02% sodium azide, 10 mM iodoacetamide, 1 mM PMSF and 2 μg leupeptin) at 4°C. The lysates were cleared by microcentrifuging for 15 minutes at 16,000 × g at 4°C. Antibody conjugated beads were prepared by combining 1 μg of polyclonal antibodies with 30 μl of a 50% protein A-Sepharose bead slurry in 0.5 ml of ice-cold PBS for 1 hr at 4°C in a tube rotator and then were washed two times with 1 ml of lysis buffer. The antibodies used for co-immunoprecipitation were: Dvl-2, β-catenin (Cell Signaling Technology, Beverly, MA), axin, and GSK3β (Upstate, Temecula, CA). Cell lysates (500 μg) were incubated with the prepared beads and 10 μl of 10% BSA overnight at 4°C. The beads were washed four times with washing buffer (50 mM Tris-Hcl [pH 7.4], 300 mM NaCl, 5 mM EDTA, 0.02% sodium azide, 0.1% Triton X-100) and one time with ice-cold PBS. Proteins were revealed after SDS-PAGE and Western blotting with antibodies to Flag (Sigma, St Louis, MO), Dvl-2, β-catenin, axin, GSK3β, and ATDC (Santa Cruz Biotechnology, Santa Cruz, CA). Images were visualized using an ECL detection system.
Cells were cultured in 6 well plates to 70% confluence and then starved in 2 ml methionine-deficient DMEM (Sigma; Madison, WI) for 30 min at 37°C with 5% CO2 Cells were pulse-labeled with 100 μCi/ml [35S]-methionine (Amersham Pharmacia, Pittsburgh, PA) for 30 minutes at 37°C. Labeled cells were chased in DMEM with a saturating amount of cold methionine (2 mM) for various times and lysed in RIPA buffer (10 mM Tris, pH 7.2, 158 mM NaCl, 1 mM EGTA, 1% Triton X-100, 1% deoxycholate, 0.1% SDS, 1mM PMSF, and 1x concentration of protease inhibitors (Complete Protease Inhibitor Cocktail, Roche Molecular Biochemicals)). Total cell lysates were then prepared and immunoprecipitated with a β-catenin antibody or Dvl-2 antibody and protein G-Sepharose beads (Invitrogen, San Diego, CA) for 3 hours at 4°C. Immunocomplexes were separated on SDS-PAGE (4–20% gradient gel) and transferred to nitrocellulose. The precipitates were subjected to autoradiography, and then the densities of the labeled proteins were analyzed by Kodak Gel Documentation System (model 1D 3.6).
Six-week-old male NOD/SCID (non-obese diabetic/severe combined immunodeficient) mice (Taconic, Germantown, NY) were housed under pathogen-free conditions. Animal experiments were approved by the University of Michigan Animal Care and Use Committee and were performed in accordance with their established guidelines. Mice were anesthetized with an i.p injection of xylazine (9 mg/kg) and ketamine (100 mg/kg). A median laparotomy was done and 5 × 105 Panc1 cells infected with a lentivirus encoding luciferase (control shRNA, ATDC-shRNA1 or β-Catenin shRNA stably transfected) in a volume of 30 μl were injected into the pancreatic tail using a 30-gauge needle (n= 8 per group). To prevent leak at the injection site, the needle was slowly withdrawn and a sterile cotton swab was applied to the injection site for 30 sec. Bioluminescent imaging of the mice was performed bi-weekly using a Xenogen IVIS 200 imaging system (Xenogen Biosciences, Cranbury, NJ). Sixty days following cancer cell injection, mice were euthanized with carbon dioxide inhalation, and autopsies were performed to assess the extent of primary tumor growth and metastasis.
Data are represented as mean ± SE from at least three independent experiments. The significance of the difference between groups was evaluated by Student’s t test or ANOVA test. p < 0.05 was considered significant.
This work was funded by NIH grant R01 CA131045 (DMS) and grant support from the Lustgarten Foundation (DMS). Special thanks to Dr. Marina Pasca di Magliano for her useful comments regarding β-catenin signaling in pancreatic cancer.
The Supplemental Data can be found with this article online at
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.