Oncogenic KRAS Activates Hedgehog Signaling Pathway in Pancreatic Cancer Cells*

Hedgehog (Hh) signaling is deregulated in multiple human cancers including pancreatic ductal adenocarcinoma (PDA). Because KRAS mutation represents one of the earliest genetic alterations and occurs almost universally in PDA, we hypothesized that oncogenic KRAS promotes pancreatic tumorigenesis in part through activation of the Hh pathway. Here, we report that oncogenic KRAS activates hedgehog signaling in PDA cells, utilizing a downstream effector pathway mediated by RAF/MEK/MAPK but not phosphatidylinositol 3-kinase (PI3K)/AKT. Oncogenic KRAS transformation of human pancreatic ductal epithelial cells increases GLI transcriptional activity, an effect that is inhibited by the MEK-specific inhibitors U0126 and PD98059, but not by the PI3K-specific inhibitor wortmannin. Inactivation of KRAS activity by a small interfering RNA specific for oncogenic KRAS inhibits GLI activity and GLI1 expression in PDA cell lines with activating KRAS mutation; the MEK inhibitors U0126 and PD98059 elicit a similar response. In addition, expression of the constitutively active form of BRAFE600, but not myr-AKT, blocks the inhibitory effects of KRAS knockdown on Hh signaling. Finally, suppressing GLI activity leads to a selective attenuation of the oncogenic transformation activity of mutant KRAS-expressing PDA cells. These results demonstrate that oncogenic KRAS, through RAF/MEK/MAPK signaling, is directly involved in the activation of the hedgehog pathway in PDA cells and that collaboration between these two signaling pathways may play an important role in PDA progression.

Pancreatic ductal adenocarcinoma (PDA) 2 is the fourth leading cause of cancer-related death for both men and women in the United States. It was estimated that in 2006 33,730 new cases would be diagnosed, and 32,300 would die from the disease (American Cancer Society Cancer Facts and Figures 2006). Therefore, PDA is one of the most lethal human diseases, with a 5-year survival rate of less than 4% and a median survival of less than 6 months.
PDA is one of the better-characterized neoplasms at the genetic level. There are now sufficient clinical, genetic, and pathological data to support a tumor progression model for PDA in which the pancreatic ductal epithelium progresses from normal to increased grades of pancreatic intraepithelial neoplasia to invasive cancer (1,2). Accompanying the progressive morphological changes is the sequential accumulation of genetic alterations in the KRAS oncogene and the tumor suppressors INK4A, p53, and SMAD4/DPC4, although these alterations have not been linked to the acquisition of specific histopathological attributes (3)(4)(5). In addition to these frequent genetic abnormalities, mutations in the tumor suppressors BRCA2, TGFBR1, and TGFBR2, the serine-threonine kinases AKT2 and LKB1/STK11, and certain DNA mismatch-repair genes represent other less common genetic events in PDA (6 -11).
Aberrant RAS activation plays a critical role in tumorigenesis; activating RAS mutations are found in 30% of all human cancers (12). Of all human cancers, PDA has the highest incidence of activating KRAS mutations (13). Activating KRAS mutations, representing the earliest genetic changes associated with the transformation of normal ductal epithelium and PDA development, have been detected in pancreatic duct lesions with minimal cytological and architectural atypia and occasionally in histologically normal pancreas (3, 14 -17). The frequency of KRAS mutations correlates with disease progression, reaching almost 100% in pancreatic adenocarcinomas. Targeted endogenous expression of an oncogenic KRAS allele in the mouse pancreas is sufficient to drive the development of pancreatic intraepithelial neoplasia and subsequently at low frequency the progression to both locally invasive adenocarcinoma and metastatic disease with sites of spread exactly as found in human pancreatic cancer (18 -20). These observations suggest that KRAS plays an essential role in the initiation, development, and maintenance of PDA.
Recently, the hedgehog (Hh) signaling pathway has been implicated as playing an important role in the progression and maintenance of PDA (21)(22)(23). Hh signaling is essential for morphogenesis, tissue patterning, and stem cell maintenance in metazoan embryos (24). Hh binds to its membrane receptor Patched (PTC), releasing PTC inhibition of a seven-transmembrane protein, Smoothened (SMO), which in turn activates downstream cytoplasmic transcription factors; that is, the CI protein in Drosophila or the mammalian homologue GLI proteins (25). Components of the Hh signaling pathway, including the ligand and the receptors, are overexpressed in human PDA tissues and cell lines. Suppressing Hh activity using cyclopamine, a steroidal alkaloid that inhibits Hh signaling through direct interaction with SMO (26), in some PDA cells with activated Hh signaling, can inhibit cell growth in vitro and reduce tumor growth in vivo in the xenograft and orthotopic mouse model (21)(22)(23)27).
The coincidence of uncontrolled activation of the RAS and Hh pathways in the early stages of PDA suggests that cross-talk between these two pathways may be a very important mechanism for the initiation and development of PDA. However, the causal effects between KRAS and Hh signaling in pancreatic tumorigenesis are not clear. Earlier results from Pdx-Shh mice had suggested that ectopic expression of Hedgehog ligands is sufficient to activate the Ras signaling pathway by inducing a mutation in the Kras gene (21), and a recent study indicates that cell-autonomous activation of the Hedgehog pathway is not sufficient to induce mutations in the Kras gene or to activate MAPK downstream of Ras (28). In addition, although expression of endogenous level of oncogenic Kras, Kras G12D , leads to pancreatic intraepithelial neoplasia identical to all three stages found in the cognate human condition and eventually PDA in mice (19), activation of Hh signaling alone is not sufficient to induce pancreatic intraepithelial neoplasia and PDA in a mouse model in which Hh signaling is activated specifically in the pancreatic epithelium (28). Because KRAS mutation represents one of the earliest genetic alterations and occurs almost universally in pancreatic adenocarcinomas, we hypothesized that oncogenic KRAS promotes pancreatic tumorigenesis in part through activation of the Hh signaling pathway in PDA. Our study shows that oncogenic transformation of human pancreatic ductal epithelial (HPDE) cells by oncogenic KRAS is indeed accompanied by enhanced GLI activation and that specific down-regulation of oncogenic KRAS activity inhibits Hh signaling in PDA cell lines with KRAS mutations. These results demonstrate that oncogenic KRAS is involved in activation of the Hh/GLI pathway in PDA cells and that cross-talk between the oncogenic KRAS and Hh pathways may play an important role in promoting cancer development during pancreatic tumorigenesis.
Cell Culture and Transfection-HPDE-c7, an immortalized pancreatic ductal epithelial cell line, was provided by Dr. Ming-Sound Tsao (University of Toronto, Canada) and cultured in keratinocyte serum-free (KSF) medium supplemented with bovine pituitary extract and epidermal growth factor (Invitrogen). KRAS V12 -transformed HPDE-c7 cells were produced by retroviral infection of HPDE-c7 cells with a KRAS V12 construct. Panc-1 and AsPC-1 cells were purchased from the American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 medium (AsPC-1) or Dulbecco's modified Essential medium (Panc-1) supplemented with 10% fetal bovine serum (Invitrogen). Transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions (the ratio of plasmid (g) to lipid (l) was 1:3).
Retroviral Infection-Retroviruses were generated by transfecting amphotropic Phoenix packaging cells with the retroviral vector pBabepuro-KRAS4B G12V and the corresponding empty vector pBabepuro using Lipofectamine Plus reagent (Invitrogen). Retroviral supernatants were collected, filtered, and incubated with the target cells in the presence of 4 g/ml Polybrene (Sigma). After 48 h, cells were subjected to selection using 0.5 g/ml puromycin (ICN Biomedicals, Irvine, CA) until all the untransduced cells had died.
RAS Activity Assay-The GTP loading status of RAS was assessed using a glutathione S-transferase (GST) fusion of the RAS binding domain (RBD) of RAF (GST-RAF-RBD) as described earlier (32).
Anchorage-independent Cell Growth Assay-Cells (1 ϫ 10 3 -5 ϫ 10 5 ) were suspended in 2 ml of HPDE cell medium with 0.35% agarose (Invitrogen), and the suspension was placed on top of 5 ml of solidified 0.7% agarose. Triplicate cultures at three different dilutions for each cell type were maintained at 37°C in a 5% CO 2 atmosphere, and fresh medium was added after 1 week. Colonies were photographed between 14 and 24 days under a phase contrast microscope. The number of colonies was counted from each experiment, and the experiment was reproduced twice. Student's t test for two samples was used to determine the statistical significance between the two groups. A p value less than 0.05 was regarded as statistically significant.
Immunoblotting Analysis-Protein concentration of cell lysates was assayed with the Bio-Rad protein assay reagent. Equal amounts of protein (5-30 g) were loaded onto 10% SDS-polyacrylamide mini-gels (Bio-Rad) and transferred to polyvinylidene difluoride membranes after electrophoresis. After being blocked overnight in 5% milk in Tris-buffered saline-Tween, blots were incubated with primary antibodies for 1.5 h followed by horseradish peroxidase-conjugated secondary antibody (1:4000) for 45 min. Antigen-antibody complexes were detected by enhanced chemiluminescence.
Luciferase Reporter Gene Assay-For luciferase assay, cells were transfected with the Gli or other luciferase reporter construct (0.5 g/well) and the SV40-Renilla control plasmid (2 ng/well) using Lipofectamine 2000. After 6 h the medium was replaced with fresh growth medium, and the cells were incubated in 5% CO 2 at 37°C overnight. After treatment with U0126 or other compounds, the cells were harvested, and luciferase activity was measured with the dual luciferase reporter assay system (Promega) according to the manufacturer's instruction. Briefly, the transfected cells were lysed in the 12-well plates with 100 l of reporter lysis buffer, and the lysate was transferred into Eppendorf tubes. Cell debris was removed by centrifugation at top speed for 30 s in a refrigerated microcentrifuge. 20 l of the supernatant was mixed with 100 l of LAR II buffer, and the luminescence was immediately measured (first reading). After 20 s, 100 l of Stop & Glo reagent was added to measure the Renilla luciferase activity (second reading). The value from the first reading was divided by the value from the second reading of each sample to obtain normalized luciferase activity. Each experiment was repeated at least three times with similar results. Gli luciferase reporter constructs 8 ϫ 3ЈGli BSwt-luc and 8 ϫ 3ЈGli BSmut-luc were provided by Dr. H. Sasaki (29).
Quantitative Real-time PCR-Total RNA was isolated from cultured cells as previously described (33). Real-time PCR analysis was conducted using an Applied Biosystems Prism 7000 sequence detection system and TaqMan gene expression assays for relative quantification of GLI1 (Hs00171790_ m1) and hedgehog-interacting protein (Hs00368450_m1) mRNA. Duplicate C T values were analyzed in Microsoft Excel using the comparative C T (⌬⌬C T ) method as described by the manufacturer (Applied Biosystems). The amount of target (2 Ϫ⌬⌬CT ) was obtained by normalized to endogenous reference (18 S) and relative to a calibrator.

RESULTS
Oncogenic Transformation of HPDE by KRAS V12 -To probe the mechanism of oncogene KRASmediated pancreatic tumorigenesis, we established a KRAS oncogenebased human PDA model using an immortalized primary HPDE cell line, HPDE-c7. This well characterized cell line is a near-diploid human pancreatic duct epithelial cell line originally derived from normal pancreas. Although immortalized by E6/E7 genes of human papilloma virus-16, HPDE-c7 is non-tumorigenic and incapable of inducing tumor growth in nude mice (34,35). Stable expression of KRAS V12 in HPDE-c7 cells using a retroviral expression vector led to transformation of the cell line. For all experiments performed in this study, we used early passages of the same puromycin-selected pool of KRAS V12 -transformed HPDE-c7 cells to eliminate potential clonal and passage variations. The resultant cells, HPDE-c7-KRAS V12 , expressed increased levels of total RAS protein, showed high RAS-GTP activity, and grew anchorageindependently in soft agar. In addition, expression of KRAS V12 in HPDE-c7 cells also led to the activation of its downstream effectors, such as MAPK and AKT. The basal phospho-MAPK and phospho-AKT levels were enhanced in the HPDE-c7-KRAS V12 cells compared with the parental cells. These observations are in complete agreement with results obtained from an independently established KRAS human PDA model using the same HPDE-c7 parental cell line (36).
Activation of Hh Signaling by KRAS V12 in HPDE-To examine the roles of the activating KRAS oncogene in controlling Hh signaling during the oncogenic transformation of HPDE, we compared the activities of Hh signaling in parental, vector control, and KRAS V12 -expressing HPDE cells. As shown in Fig. 1A, KRAS V12 expression led to a significantly increased activation of Gli-mediated luciferase activity, whereas the expression of the Myc reporter and a mutant Gli reporter that is incapable of binding GLI transcription factors were not affected. Moreover, the KRAS V12induced Gli-luciferase activity was insensitive to cyclopamine treatment (5 M), whereas the same treatment resulted in significant inhibition of Gli-luciferase activity in an SMO-expressing 10T1/2 cell line (Fig. 1B). These results suggest that KRAS V12 activates Hh signaling in HPDE cells in a ligand-independent manner. Further supporting our results based on the luciferase assay, oncogenic KRAS also up-regulated endogenous levels of GLI1 and hedgehoginteracting protein mRNAs, both of which are Hh target genes, in immortalized HPDE cells as measured by real-time PCR (Fig. 1C). To ensure that the apparent activation of Hh signaling observed in HPDE-c7-KRAS V12 cells was caused specifically by activation of the KRAS pathway rather than an indirect effect associated with the transformation of HPDE, we suppressed the oncogenic KRAS activity in HPDE-c7-KRAS V12 cells using a specific nontoxic RAS antagonist, FTS, which dislodges RAS from its membrane anchorage domains and accelerates its degradation (37,38). FTS has been shown to act as a functional KRAS inhibitor in human pancreatic cell lines that express activated KRAS (39). As shown in Fig. 1D, treatment of HPDE-c7-KRAS V12 with FTS led to an inhibition of Gli-luciferase and endogenous GLI1 mRNA. Taken together, our studies suggest that KRAS V12 specifically activates Hh signaling in HPDE cells as GLI1 is not only a downstream effector but also a direct target gene and a reliable marker of Hh signaling pathway activities (40).
Oncogenic KRAS Is Responsible for GLI1 Activation in PDA Cells-To further test whether oncogenic KRAS is essential for GLI1 activation in PDA cells, we knocked down the expression of the onco-  genic KRAS in human PDA cell lines (AsPC-1 and Panc-1) that express mutant KRAS D12 using RNA interference. A previously reported siRNA duplex that specifically knocks down KRAS D12 in AsPC-1 and Panc-1 cells was used (41). The inactivation of KRAS oncogene by the KRAS D12 -specific siRNA duplex in AsPC-1 and Panc-1 cells led to decreased cellular RAS levels ( Fig. 2A). Consistent with the reduced KRAS level, the phospho-MAPK levels were suppressed in KRAS D12knockdown AsPC-1 and Panc-1 cells compared with cells treated with the control siRNA, confirming the specificity and effectiveness of KRAS D12 -siRNA. Gene silencing of KRAS D12 in AsPC-1 and Panc-1 cells eventually led to growth inhibition and cell death 96 h after siRNA transfection. Therefore, experiments related to KRAS D12 siRNA were performed 24 h after transfection when no noticeable difference between control and KRAS D12 siRNA-transfected cells could be observed. KRAS D12 knockdown by siRNA resulted in a dose-dependent inhibition of Gli-luciferase activity and endogenous levels of GLI1 mRNA in AsPC-1 (Fig. 2B). Similar results were also observed in Panc-1 cells (Fig. 2C). To exclude the possibility that our observed inactivation of Gli-luciferase by KRAS D12specific siRNA was due to potential off-target effects, we used a PDA cell line, BxPC-3 cells (which does not contain KRAS mutation), to confirm the specificity of KRAS D12 siRNA. As expected, the Gli-luciferase activity and GLI1 mRNA level in BxPC-3 cells did not change significantly after introducing KRAS D12 -specific siRNA (Fig. 2D). These results demonstrate that effects of KRAS D12 siRNA observed in AsPC-1 and Panc-1 cells are specific to the KRAS D12 mutant allele. Taken together, our data indicate that oncogenic KRAS plays an important role in regulating Hh signaling in PDA cells.
RAF/MEK/MAPK, but Not the PI3K/AKT Pathway, Is Required for the KRAS-mediated Activation of Hh Signaling-To further determine which downstream effectors of the oncogene KRAS mediate the activation of Hh signaling, we examined the levels of Hh activity in HPDE-c7-KRAS V12 cells in response to specific inhibitors that target the RAS downstream effectors MEK and PI3K. Inhibition of MEK by U0126 (20 M) and PD98059 (40 M) led to a significant reduction of Gli-luciferase activity and endogenous GLI1 mRNA levels (Fig. 3A), whereas wortmannin (100 nM), a PI3K-specific inhibitor, had little effect. These results suggested that the RAF/MEK/MAP kinase pathway was directly responsible for the KRAS-mediated activation of Hh signaling in HPDE-c7-KRAS V12 cells. The inhibitory effect of U0126 and PD98059 was also observed in PDA cells expressing activating KRAS mutant. When AsPC-1 and Panc-1 cells were treated with the various inhibitors for 24 h, U0126 and PD98059 inhibited Gli1-luciferase expression (Fig. 3B) as well as endogenous GLI1 mRNA levels (Fig. 3C), whereas wortmannin had no effects. To ensure that the apparent non-effect of wortmannin is not due to a lack of efficacy of the compound, we determined the phosphorylation status of MAPK and AKT in Panc-1 cells treated with various pharmacological inhibitors. As shown in Fig. 3D, U0126 and PD98059 significantly suppressed the phosphorylation levels of MAPK. Wortmannin abolished the phosphorylation of Ser-473 of AKT, which is essential for its activation.
If the RAF/MEK/MAPK, but not the PI3K/AKT pathway, is indeed responsible for the KRAS-mediated activation of Hh signaling, we expect that expressing a constitutively active RAF kinase, but not a constitutively active AKT, will rescue the inhibitory effect of oncogene KRAS knockdown on Hh activation. To confirm our study using pharmacological inhibitors, we co-transfected either BRAF E600 , a constitutively active BRAF mutant, or myrAKT2, a membrane-targeted, constitutively active form of AKT2, together with the KRAS D12 -specific siRNA duplex into AsPC-1 and subsequently monitored the Gli-luciferase activity. As shown in Fig. 4A, expression of BRAF E600 restored KRAS D12 -specific siRNA-mediated suppression of MEK activity and completely blocked the inhibitory effect of KRAS D12 -specific siRNA on Hh signaling (Fig. 4B). On the other hand, although myr-AKT2 brought back the AKT

. Constitutively active BRAF E600 rescues Gli activation in KRAS D12 knockdown ASPC-1 cells.
A, total cellular KRAS, phosphorylated MAPK (P-MAPK) and total MAPK levels in AsPC-1 cells co-transfected with control or KRAS D12 -specific siRNA and control or BRAF E600 expression vector. B, Gli-luciferase activities in AsPC-1 cells co-transfected with control or KRAS D12 -specific siRNA and control or BRAF E600 expression vector. C, total cellular KRAS, phosphorylated glycogen synthase kinase 3␤ (P-GSK3␤), and total GSK3␤ levels in AsPC-1 cells co-transfected with control or KRAS D12 -specific siRNA and control or myr-AKT2 expression vector. D, Gli-luciferase activities in AsPC-1 cells co-transfected with control or KRAS D12 -specific siRNA and control or myr-AKT2 expression vector. Data are represented as mean Ϯ S.E. p values are indicated by asterisks (relative to control): **, p Ͻ 0.01; ***, p Ͻ 0.001. activity inhibited by KRAS D12 -siRNA, as measured by glycogen synthase kinase 3␤ (GSK3␤) phosphorylation (Fig. 4C), expression of myr-AKT2 failed to rescue Gli-luciferase activity brought down by KRAS D12 -siRNA (Fig. 4D). Similar results were obtained using Panc-1 cells (data not shown). These results indicate that the RAF/MEK/MAPK signaling, but not the PI3K/AKT pathway, is critical for KRAS-mediated GLI activation in PDA.
KRAS, through the RAF/MEK/MAPK Pathway, Regulates GLI1 Protein Degradation-To further elucidate the mechanism by which oncogene KRAS mediates activation of Hh signaling, we determine whether oncogenic KRAS could regulate GLI1 at the protein level. Because none of the commercially available GLI-antibodies was specific or sensitive enough to monitor the endogenous GLI protein levels, we co-transfected an Myc-tagged GLI1 expression vector (42) and KRAS D12 -specific siRNA duplex into AsPC-1 and Panc-1 cells and subsequently monitored the expression levels of GLI1 24 h posttransfection using anti-Myc antibody. As shown in Fig. 5A, the expression levels of GLI1 were significantly reduced in both AsPC-1 and Panc-1 cells with oncogenic KRAS activity silenced by siRNA. After inhibition of new protein synthesis by cycloheximide (50 g/ml), we observed that inhibition of MEK by U0126 led to a significant reduction of GLI1 stability in Panc-1 cells (Fig. 5B). The apparent half-life of GLI1 protein in untreated Panc-1 cells was about 3 h, whereas the half-life of GLI1 protein in U0126-treated Panc-1 cells was estimated to be less than 1 h (Fig. 5C). Furthermore, GLI1 protein degradation was blocked by the proteasome inhibitor MG132 in the absence or presence of U0126. Taken together, these results suggest the oncogene KRAS, through the RAF/MEK/MAKP pathway, blocks proteasome-mediated GLI1 degradation and consequently leads to the activation of Hh signaling in pancreatic cancer cells.
Suppression of GLI Activity Attenuates the Transforming Activity of KRAS Oncogene-To investigate the physiological relevance and contribution of KRAS-mediated Hh signaling activation in pancreatic cancer development, we suppressed GLI1 expression using an established GLI1-specific siRNA (43). The specificity of this particular GLI1 siRNA has been extensively demonstrated previously (43). To test the effectiveness of the GLI1 siRNA, we co-transfected a Myctagged GLI1 construct with control or GLI1 siRNA into AsPC-1 and Panc-1 cells. GLI1 siRNA effectively suppressed GLI1 protein expression in Panc-1 cells (Fig. 6A, inset). Although GLI1 siRNA does not affect growth or apoptosis of AsPC-1 and Panc-1 cells under typical tissue culture growth conditions (supplemental Fig. S1), GLI1 gene silencing led to a significant attenuation of anchorage-independent growth of Panc-1 cells as measured by the soft agar colony formation assay (Fig.  6A). Similar results were observed using AsPC-1 cells (data not shown). To further confirm this observation, we ectopically expressed a C-terminal-truncated Gli3 construct in AsPC-1 and Panc-1 cells. This specific deletion mutant, Gli3CЈ⌬ClaI, has been demonstrated to exert strong dominant negative effects on Hh signaling both in vitro and in vivo (30). As expected, expressing of Gli3CЈ⌬ClaI led to attenuated Hh signaling as indicated by the apparent reduction of Gli-luciferase activity in AsPC-1 cells (Fig. 6B). Moreover, expression of Gli3CЈ⌬ClaI significantly suppressed anchorage-independent growth of AsPC-1 in a manner similar to that of GLI1 siRNA (Fig. 6C). We observed similar results in Panc-1 cells (data not shown). In contrast, GLI1 gene silencing had no significant effect on anchorage-independent growth of BxPC-3 cells, which contain no KRAS mutation (Fig. 6D). Taken together, these data suggest that activation of Hh signaling is important for KRAS oncogenic transformation in pancreatic cancer cells.

DISCUSSION
Constitutively active KRAS mutations are one of the earliest and most common genetic alterations in pancreatic carcinomas (13). Cross-talk between KRAS and other oncogenic pathways in PDA have not been investigated extensively. In this study we explored the cross-talk between oncogenic KRAS and the Hh signaling pathway, which has recently been shown to be up-regulated in human PDA tissues and cell lines (21)(22)(23). In the immortalized human pancreatic ductal epithelial cell line, HPDE-c7, we demonstrated that expressing KRAS V12 led to increased endogenous GLI1 levels and transcriptional activity, whereas suppressing oncogenic KRAS expression by siRNA inhibits GLI activity and GLI1 expression in PDA cell lines with activating KRAS mutation. Furthermore, we found that KRASmediated Hh activation was suppressed by pharmacological inhibitors specific for MEK, but not PI3K, in KRAS-transformed HPDE and mutant KRAS expressing PDA cell lines. Expression of constitutive active form of BRAF E600 , but not myr-AKT, blocks the inhibitory effects of KRAS knockdown on Hh signaling. Our studies suggest that oncogenic KRAS, through the RAF/MEK/MAPK pathway, suppresses GLI1 protein degradation and consequently plays an important role in activating the Hh signaling pathway in the absence of additional Hh ligand during pancreatic tumorigenesis.
Although biochemical and genetic analyses suggest that three GLI transcription factor isoforms (GLI1, GLI2, and GLi3) play distinct as well as overlapping roles in mammalian Hh signal transduction during development (44), the individual roles of GLI1, GLI2, and GLi3 in pancreatic tumorigenesis are not very well understood. Overexpression of GLI1, along with other Hh signaling molecules such as Sonic hedgehog (Shh), Indian hedgehog, SMO, and PATCH, have been reported in human pancreatic cancer tissues and pancreatic cancer cell lines (21)(22)(23). On the other hand, the status of GLI2 in human pancreatic cancer has not been extensively studied. A recent study of a transgenic mice mode based on a dominant active form of the Gli2 transcription factor suggest that although Gli2 activation is sufficient to drive pancreatic neoplasia, it does not recapitulate human pancreatic carcinogenesis (28). Considering the extensive sequence and structure homology between GLI1 and GLI2, it is conceivable that these proteins are regulated by RAS-RAF-MAPK signaling pathway in a similar manner. Indeed, our preliminary studies suggest that expression of GLI2 is also inhibited by KRAS siRNA knockdown and MEK inhibitors, just like GLI1 expression in both AsPC-1 and Panc-1 cells (supplemental Fig. S2).
Our findings provide direct evidence that oncogenic KRAS activates Hh signaling through up-regulation of GLI in PDA. Although deregulation of Ras signaling in mice lacking p53 function in the pancreatic epithelium has recently been shown to induce Shh expression (20), we believe that ligand-independent activation of the Hh pathway represents another important mechanism by which oncogenic KRAS promotes tumor formation. The ability of oncogenic KRAS to activate Hh signaling in the absence of Hh ligand also offers an explanation for why more than 50% of PDA cells lines with sustained Hh signaling activity are resistant to cyclopamine (21). The present findings combined with an earlier observation that sustained Hh activation activates platelet-derived growth factor receptor ␣ and the RAS pathway (45) suggest that RAS and Hh signaling pathways can potentially form a positive feedback loop to promote tumorigenesis in pancreatic cancer.
Although PI-3K and AKT activities are essential for liganddependent Shh signaling in the specification of neuronal fates in chicken neural explants and chondrogenic differentiation of 10T1/2 cells, activation of the phosphoinositide 3-kinase/AKT pathway alone is not sufficient to drive Gli activation (46). Our study indicates that activation of the RAS pathway is sufficient to drive Hh signaling activation in PDA cells through up-regulation of GLI expression without the addition of Hh ligand, and the effect of RAS is mediated exclusively by the RAF/MEK/ MAPK pathway independent of PI3K/AKT. Furthermore, suppression of GLI1 expression in oncogenic KRAS-containing PDA cells leads to a significant attenuation of anchorage-independent growth in soft agar, suggesting that activation of the Hh signaling pathway may represent an important mechanism for KRAS-mediated pancreatic tumorigenesis.
Our study is consistent with a recent report by Hebrok and co-workers (28), showing that although Hh signaling alone does not recapitulate human pancreatic carcinogenesis, cooperation of Hh and Ras signaling significantly reduces the latency of PanIN formation in mice simultaneously expressing dominant active Gli2 and Kras D12 compared with transgenic mice FIGURE 6. Down-regulation of GLI activity suppresses anchorage-independent growth of PDA cells with oncogene KRAS. A, anchorage-independent growth potential of Panc-1 cells transfected with control or GLI1-specific siRNAs tested by soft agar assay. Inset, levels of ectopically expressed GLI1-Myc proteins in Panc-1 transfected with control and GLI1-specific siRNAs. B, Gli-luciferase activities in AsPC-1 cells transfected with empty vector or Gli3CЈ⌬ClaI construct. Inset, levels of ectopically expressed dominant negative Gli3 deletion mutant, Gli3CЈ⌬ClaI, in AsPC-1 transfected with empty vector and Gli3CЈ⌬ClaI construct. C, anchorage-independent growth potential of AsPC-1 cells transfected with control or Gli3CЈ⌬ClaI vector tested by soft agar assay. D, anchorage-independent growth potential of BxPC-3 cells transfected with control or GLI1-specific siRNAs tested by soft agar assay. Data, represented as the mean Ϯ S.E. p values, are indicated by asterisks (relative to control): one asterisk, p Ͻ 0.05; two asterisks, p Ͻ 0.01.
expressing Kras D12 alone. Our results are also in agreement with an earlier observation in which phorbol ester, acting through protein kinase C␦ and the extracellular signal-regulated kinase pathway, can stimulate Gli transcriptional activity in NIH3T3 cells in a ligand-independent manner (47). Taken together, these observations and our findings suggest that activation of the RAS/MAPK pathway, induced by diverse upstream signals and converging at the level of GLI transcription factors, is an integral component of the Hh signaling. The realization of cross-talk between RAS/MAPK and Hh signaling pathways in pancreatic carcinomas also suggests that targeting the RAS and Hh pathways synergistically may represent a new therapeutic strategy for PDA.