![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
REVIEW
1 Department of Medical Oncology, Dana-Farber Cancer Institute; 2 Department of Medicine, Brigham and Women’s Hospital; 3 Harvard Radiation Oncology Program; 4 Department of Genetics; 5 Gastrointestinal Unit, 6 Massachusetts General Hospital Cancer Center, Massachusettes General Hospital; 7 Center for Applied Cancer Science and Belfer Institute for Innovative Cancer Science, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115, USA
 |
Abstract |
|---|
 Top Abstract Pancreas anatomy and physiologyEpidemiology of PDACMorphological characteristics of...Molecular genetics of PDACThe cellular basis of...Genetically engineered mouse...Major challenges and...AcknowledgmentsReferences |
|---|
[Keywords: Pancreatic cancer; genetics; pancreas; mouse models; genomic stem cell]
|
Pancreas anatomy and physiology |
|---|
|
TopAbstractPancreas anatomy and physiologyEpidemiology of PDACMorphological characteristics of...Molecular genetics of PDACThe cellular basis of...Genetically engineered mouse...Major challenges and...AcknowledgmentsReferences |
|---|
|
; D. Li et al. 2004). PDAC is the focus of this review, and the reader is directed to the following excellent review covering other pancreas cancer types (Hruban et al. 2006b).
|
|
Epidemiology of PDAC |
|---|
|
TopAbstractPancreas anatomy and physiologyEpidemiology of PDACMorphological characteristics of...Molecular genetics of PDACThe cellular basis of...Genetically engineered mouse...Major challenges and...AcknowledgmentsReferences |
|---|
; Fuchs et al. 1996; Gapstur et al. 2000; Michaud et al. 2001; Berrington de Gonzalez et al. 2003; Stolzenberg-Solomon et al. 2005). Increased risk has also been documented in relatives of PDAC patients, and it is estimated that 10% of PDAC cases are associated with an inherited predisposition based on familial clustering (Schenk et al. 2001; Petersen and Hruban 2003). Correspondingly, germline mutations have been linked to familial PDAC, including those targeting the tumor suppressor genes INK4A, BRCA2, and LKB1, the DNA mismatch repair gene MLH1 and the cationic trypsinogen gene PRSS1 (Whitcomb et al. 1996; Jaffee et al. 2002). BRCA1 mutation appears to confer increased susceptibility to PDAC, albeit with a lower associated risk than BRCA2 (Thompson and Easton 2002). Given the low penetrance of PDAC and the typical age of onset associated with the above germline mutations, these genetic lesions appear to impact malignant progression of precursor lesions rather than cancer initiation. Supporting this hypothesis, INK4A and BRCA2 mutations are not detected in the earliest sporadic PDAC premalignant lesions but are only found in the later intermediate or advanced pancreatic intraepithelial neoplasm (PanIN) lesions (Wilentz et al. 1998; Goggins et al. 2000). Additionally, mice engineered with germline INK4A mutations do not develop PDAC unless combined with activated K-RAS mutations (see below).
The germline mutations listed above are estimated to account for <20% of PDAC-prone familial cases. It is clear that additional novel disease predisposition genes exist as evidenced by rare families in which PDAC is inherited as an autosomal dominant trait with high penetrance (Lynch et al. 1996). In one family, the 4q32-34 locus has been linked to the development of diabetes, pancreatic exocrine insufficiency, and PDAC with a penetrance approaching 100% (Eberle et al. 2002). The gene associated with this syndrome has yet to be identified. Furthermore, genes that predispose to pancreatitis are associated with increased occurrence of PDAC. In patients with hereditary pancreatitis caused by germline mutations in the cationic trypsinogen gene PRSS1, there is a 53-fold increased incidence of PDAC (Lowenfels et al. 1997). Another link has also been forged between mutations in the cystic fibrosis gene (CFTR) and PDAC. Heterozygous CFTR mutations are associated with chronic pancreatitis, a known risk factor for pancreatic cancer. Recently, a direct link between CFTR mutation and cancer has been posited with the detection of a mutant allele in early-onset PDAC cases. Previous studies had not conclusively identified such a link, but may have been limited by smaller numbers of cases and more limited mutational analysis (Neglia et al. 1995; Sharer et al. 1998; Malats et al. 2001; Matsubayashi et al. 2003; McWilliams et al. 2005).
While the question as to how these separate genetic conditions lead to PDAC remains to be fully understood, the clinical observation of exocrine insufficiency and pancreatitis as a common patho-physiologic process leading to PDAC is compelling. Exocrine organ dysfunction and pancreatitis could promote tumorigenesis in part by promoting the local release of growth factors, cytokines, and reactive oxygen species (ROS), thereby inducing cell proliferation, disrupting cell differentiation states, and selecting for oncogenic mutations. The observation that activating K-RAS mutations are detectable in up to a third of patients with chronic pancreatitis is consistent with this hypothesis (Lohr et al. 2000). Further evidence from mouse models also suggests the presence of a ductal precursor cell population that undergoes expansion in response to organ damage (see below in section Origins of Pancreatic Cancer). In states of pancreatic inflammation or damage, an expanded "stem cell"-like compartment could represent a subpopulation of cells susceptible to oncogenic transformation upon somatic mutation of key proto-oncogenes and tumor suppressor genes (Beachy et al. 2004).
|
Morphological characteristics of PDAC and evolving pancreatic neoplasms |
|---|
|
TopAbstractPancreas anatomy and physiologyEpidemiology of PDACMorphological characteristics of...Molecular genetics of PDACThe cellular basis of...Genetically engineered mouse...Major challenges and...AcknowledgmentsReferences |
|---|
). PDAC primarily exhibits a glandular pattern with duct-like structures and varying degrees of cellular atypia and differentiation (Fig. 2). Less common subtypes of PDAC include colloid, adenosquamous, or sarcomatoid histology. Often within an individual tumor, there are regional differences in histology, tumor grade, and degree of differentiation. Even the smallest primary lesions commonly exhibit perineural and lympho-vascular invasion, suggesting a propensity for early distant spread.
|
; Maitra et al. 2005). Of these precursor lesions, the most common and extensively studied is PanIN, which is found in the smaller-caliber pancreatic ducts. Surveys of pancreas specimens from autopsy studies and surgical resection cases have suggested that PanINs are a common finding in older adults, occurring in as many as 30% of specimens. A documented increased incidence of PanINs in patients with PDAC initially suggested their biologic relationship. PanINs show a spectrum of divergent morphological alterations relative to normal ducts that seem to represent graded stages of increasingly dysplastic growth (Fig. 2; Hruban et al. 2000a, 2001; Maitra et al. 2005). PanINs are graded from stages I to III, with the earliest stage characterized by the appearance of a columnar, mucinous epithelium and with increasing architectural disorganization and nuclear atypia through stages II and III (Figs. 2, 4 [below]). The high-grade PanINs ultimately transform into frank PDAC with evidence of areas of invasion beyond the basement membrane. Several molecular profiling studies have subsequently reinforced the PanIN-to-PDAC progression model through documentation of an increasing number of gene alterations in higher grade PanINs (Fig. 2; Klimstra and Longnecker 1994; Moskaluk et al. 1997; Heinmoller et al. 2000; Hruban et al. 2000b, 2001; Wilentz et al. 2000; Yamano et al. 2000; Luttges et al. 2001; Maitra et al. 2003).
Less common precursor lesions are MCNs and IPMNs. MCNs are large mucin-producing epithelial cystic lesions that harbor a distinctive ovarian-type stroma with a variable degree of epithelial dysplasia and focal regions of invasion. IPMNs resemble PanINs at the cellular level but grow into larger cystic structures. Of interest, two subtypes of invasive cancer have been found in association with IPMNs; typical PDAC, and a colloid type characterized by copious mucin production. Among IPMNs and MCNs, both common and distinct molecular events in comparison with PanINs have been described, suggesting that each precursor lesion may reflect variations on a common theme of malignant transformation of the duct (Fig. 2; Sato et al. 2001b, 2004a; Adsay et al. 2004). Along these lines, expression profiling has revealed several up-regulated genes commonly associated with PanINs, IPMNs, and PDAC, pointing toward key shared molecular events (Crnogorac-Jurcevic et al. 2002; Sato et al. 2004b; Prasad et al. 2005). How these PDAC precursor lesions relate to each other, their common or separate cellular origins, and whether each type leads to distinct molecular and biological PDAC subtypes remain to be fully explored.
|
Molecular genetics of PDAC |
|---|
|
TopAbstractPancreas anatomy and physiologyEpidemiology of PDACMorphological characteristics of...Molecular genetics of PDACThe cellular basis of...Genetically engineered mouse...Major challenges and...AcknowledgmentsReferences |
|---|
The K-RAS oncogene and its signaling pathways
K-RAS is a member of the RAS family of GTP-binding proteins that mediate a wide variety of cellular functions including proliferation, differentiation, and survival (Campbell et al. 1998; Malumbres and Barbacid 2003). Although RAS is a GTPase, its intrinsic activity is inefficient and requires GTPase activating proteins (GAPs) to promote GTP hydrolysis and attenuate downstream signaling. Activating K-RAS point mutations at codon 12 (from GGT to GAT or GTT, and more rarely CGT) results in substitution of glycine with aspartate, valine, or arginine. These mutations are the first known genetic alterations, occurring sporadically in normal pancreas tissue, and are detected in 
30% of early neoplasms with the frequency rising to nearly 100% in advanced PDAC (Klimstra and Longnecker 1994; Rozenblum et al. 1997). Consistent with a central pathogenic role of the K-RASG12D mutation, mice engineered with pancreas-specific expression of this activated K-RAS allele sustain classical PanIN lesions that can progress to PDAC in the appropriate tumor suppressor background (as discussed in depth below).
Although RAS is considered to be an attractive therapeutic target given its prominent role in the genesis of PDAC and many other human malignancies, specific biochemical properties of the protein have made this an elusive goal. Importantly, the hotspots of RAS mutations in human cancer are located near the bound nucleotide and decrease the intrinsic rate of GTP hydrolysis and make the molecule insensitive to GAPs (for review, see Wittinghofer et al. 1997; McCormick 1998). This results in a constitutively activated molecule that is essentially independent of growth factor stimulation. In contrast to the activating mutations of other oncogenes such as kinases, which increase their catalytic activity, the oncogenic mutations of RAS inhibit its enzymatic activity. Thus, rather than using the traditional paradigm of inhibiting an oncogene’s enzymatic function (e.g., c-Kit, EGFR, HER2/Neu), an effective RAS antagonist would increase the GTPase activity of RAS or make it more sensitive to GAPs. There have been attempts to inhibit K-RAS in this malignancy, mainly through inhibition of essential post-translational modifications. Despite showing promise in vitro and in PDAC xenografts (Omer and Kohl 1997), farnysltransferase inhibitors (FTIs), which inhibit a lipid modification of the C terminus of RAS proteins, have not been clinically effective (Van Cutsem et al. 2004). Among the explanations for the clinical failure of FTIs, compensatory geranyltransferase activity preserving RAS function has been suggested (Lebowitz et al. 1995; Lerner et al. 1997).
K-RAS is mutated in nearly all human PDAC specimens (Almoguera et al. 1988). Mouse models have convincingly shown that K-RAS mutations are an initiating step in PDAC pathogenesis (see below on mouse models), and detailed pathological studies have demonstrated that K-RAS mutation is one of the earliest genetic events seen in human PanIN progression (Moskaluk et al. 1997). The essentiality of K-RAS in the maintenance of advanced PDAC is suggested by dominant-negative mutant studies (Hirano et al. 2002). More recently, RNA interference (RNAi) knockdown studies have provided complementary evidence that K-RAS plays a vital role in PDAC maintenance (Brummelkamp et al. 2002; Fleming et al. 2005). Additional genetic evidence will be needed to address the specific biological role of K-RAS across the progressive stages of this cancer.
Activated K-RAS engages multiple effecter pathways, notably the RAF-mitogen-activated kinase (MAPK), phosphoinositide-3-kinase, and Ral GDS pathways (Fig. 3; for review, see Campbell et al. 1998). Given the aforementioned difficulties in K-RAS inhibition, these downstream targets may provide alternative effective points of therapeutic intervention and thus are the focus of ongoing studies in pancreatic specific systems.
|
). The importance of RAF in cancer has come to light with the identification of activating B-RAF mutations in many malignancies,including melanoma, papillary thyroid, colorectal, and serous ovarian cancers. Interestingly, B-RAF and RAS mutations appear to be mutually exclusive in these cancers (Davies et al. 2002; Garnett and Marais 2004). B-RAF mutations are rare in PDAC but are present in 33% of the histologically distinct pancreatic medullary carcinomas, which are characterized by wild-type KRAS and DNA mismatch repair defects (Calhoun et al. 2003; Ishimura et al. 2003).
Inhibition of MAPK, either through the use of dominant negatives or pharmacological inhibition of the upstream activator MEK, results in decreased proliferation of PDAC cell lines and cell cycle arrest (Hirano et al. 2002; Gysin et al. 2005). This arrest may be mediated through increased expression of p27KIP1, as RNAi-mediated knockdown of p27KIP1 partially abrogates this growth arrest effect. Kinase suppressor of RAS (KSR) is a scaffolding molecule important in RAS signaling that mediates activation of the MAPK pathway (for review, see Ory and Morrison 2004). Antisense neutralization of KSR can inhibit proliferation, soft-agar growth, and invasion of cultured PDAC cells. Additionally, infusion of KSR antisense oligonucleotides causes PDAC xenograft regression and, in some cases, pretreatment with the oligos resulted in complete regression of tumors even after treatment was discontinued (Xing et al. 2003). These inhibitory effects were likely through the MAPK pathway, as transduction of activated RAF was able to restore invasion and transformation.
Phosphoinositide 3-kinase (PI3K) pathway.
The PI3K signaling pathway, which can be activated by RAS (Rodriguez-Viciana et al. 1996), regulates cell survival, size, and proliferation via several downstream effectors including AKT, p70-S6K, and the small GTPase, RAC (for review, see Cantley 2002; Vivanco and Sawyers 2002). Mutations in the PI3K pathway that are common in other cancer types, including activating mutations of the catalytic subunit of PI3K and loss-of-function mutants of the PTEN tumor suppressor, have not been commonly observed in human PDAC (Okami et al. 1998; Samuels and Velculescu 2004). Mounting evidence, however, has pointed to the general importance of this pathway and its downstream signaling elements in PDAC. There are reports of decreased PTEN expression in PDAC, possibly due to promoter hypermethylation (Asano et al. 2004), and conditional knockout studies of PTEN in the pancreas have produced PDAC in a small percentage of mice (Stanger et al. 2005). In addition, microinjection experiments of mutant K-RAS into cultured primary pancreatic ductal cells elicit an increase in proliferation and cell size that is mediated through PI3K and mTOR (Agbunag and Bar-Sagi 2004). Lastly, activation of the PI3K pathway appears necessary and sufficient to maintain oncogenic RAS-transformed xenograft tumors after the elimination of RAS expression (Lim and Counter 2005).
The PI3K downstream effector AKT2 is amplified in 10%–20% of PDAC, providing genetic evidence supporting the pathway’s importance in this tumor type (Cheng et al. 1996; Ruggeri et al. 1998; Altomare et al. 2003; Schlieman et al. 2003). Functionally, the significance of this amplification has been suggested in PDAC cell lines through the use of antisense oligonucleotides, which inhibit the growth of PDAC lines in xenograft assays (Cheng et al. 1996). Additionally, pharmacological inhibition of PI3K appears to increase the sensitivity of PDAC cell lines to chemotherapy as well as TNF-
-induced apoptosis, and diminishes serum-induced proliferation (Ng et al. 2000; Perugini et al. 2000; Shah et al. 2001). The mammalian target of rapamycin (mTOR), which acts downstream of AKT2, has been shown to be activated in PDAC. Treatment with an mTOR inhibitor impedes growth of several PDAC cell lines (Asano et al. 2005). Additionally, rapamycin inhibits PDAC xenograft growth and metastasis possibly through induction of endothelial cell death and tumor vessel thrombosis (Bruns et al. 2004).
Nuclear factor 
B (NFB).
The NFB transcription factor may be another important downstream mediator of mutated K-RAS signaling in PDAC (Sclabas et al. 2003). Activation of this pathway occurs in response to a variety of cell stresses through stimulation by proinflammatory cytokines and growth factors, and is known to regulate the immune response, apoptosis, and many other processes (Ghosh et al. 1998; Karin and Ben-Neriah 2000; Hayden and Ghosh 2004). Constitutive NFB activity is observed in many cancers, where it is thought to contribute to cell survival, angiogenesis, and invasion (Orlowski and Baldwin 2002).
Most primary pancreatic cancers and cell lines, but not normal pancreas specimens, show constitutive NFB activity (Wang et al. 1999; Chandler et al. 2004). Induction of NFB may directly involve K-RAS signaling since expression of a dominant-negative RAS allele abrogates NFB activity in PDAC cell lines (Liptay et al. 2003). While NFB induction has been shown to be crucial for RAS transformation of several cell types (Mayo et al. 1997; Arsura et al. 2000), the NFB subunits RelA/p65 and c-Rel appear to be dispensable for H-RAS-induced transformation of MEFs (Hanson et al. 2004). A possible mechanism for increased NFB activation may be increased expression of components of the ubiquitin-mediated degradation pathway leading to increased degradation of IB (Muerkoster et al. 2005). In vitro, NFB appears to play a role in the regulation of cell survival genes, VEGF (vascular endotheial growth factor), urokinase, and other proinvasive or angiogenic factors (Fujioka et al. 2003; Xiong et al. 2004). The NFB pathway may also contribute to the prominent chemoresistance of PDAC (Dong et al. 2002; Arlt et al. 2003), perhaps via its capacity to up-regulate BCL-2 and BCL-XL, as well as multiple other anti-apoptotic proteins (for review, see Bharti and Aggarwal 2002). These diverse roles of NFB highlight the need to further explore this complex pathway specifically in PDAC.
Other Ras superfamily GTPases.
The RAS superfamily of GTPases is comprised of at least 150 members and subdivided into five subfamilies—RAS, RHO, RAB, ARF, and RAN. The common and distinct functions as well as the biochemical inter-relationships among the members are complex and are the subject of considerable study (for review, see Mitin et al. 2005). The RHO family of GTPases is involved in several important cellular processes including actin cytoskeleton rearrangement, cell size, proliferation, survival, polarity, and membrane trafficking (for review, see Gomez del Pulgar et al. 2005; Wennerberg et al. 2005).
RHO proteins with GTPase-inactivating mutations transform mouse fibroblasts and studies using dominant-negative RHO family mutants have shown that they are necessary for RAS transformation of rodent fibroblasts. Activating mutations of RHO family members have not been demonstrated in human cancers; however, overexpression of RHO-C has been shown to correlate with PDAC tumor metastasis and prognosis (Suwa et al. 1998). Recent work has also implicated the RHO family in PDAC through the exchange factor VAV1 (Fernandez-Zapico et al. 2005). Although normally restricted to the hematopoietic system, VAV1 is expressed in PDAC specimens likely through promoter demethylation. Importantly, this ectopic expression correlates with decreased patient survival, and RNAi-mediated knockdown of VAV1 expression suppresses tumorigenicity in xenografts.
RAL GTPases are members of the RAS subfamily and are thought to be downstream of RAS through RAS’s capacity to activate RAL exchange factors (for review, see Feig 2003). Activated RAL and RAL exchange factors enhance RAS-induced cellular transformation, and the dominant-negative mutants inhibit it (Chien and White 2003). Recently, RAL A was shown to be activated in a variety of PDAC cell lines, and knockdown of RAL A suppressed tumorigenicity of RAS-transformed human cells (Lim et al. 2005). The role of RAL proteins in tumorigenesis may relate to the requirement of these factors for maintaining the polarity of epithelial cells through the regulated transport of basolateral membrane proteins (for review, see Camonis and White 2005).
A preponderance of evidence supports the role of K-RAS and many of its downstream effectors in both the ini- tiation and likely the maintenance of PDAC. Given this important place, it is crucial to define precisely which K-RAS effector pathways are important to each aspect of tumorigenesis to guide future therapeutic strategies.
The 9p21 locus and the INK4A and ARF tumor suppressors
Loss of INK4A function—brought about by mutation, deletion, or promoter hypermethylation—occurs in 80%–95% of sporadic PDAC (Rozenblum et al. 1997; Hustinx et al. 2005). INK4A loss is generally seen in moderately advanced lesions that show features of dysplasia. Germline mutations in INK4A are associated with the Familial Atypical Mole–Malignant Melanoma (FAMMM) syndrome, which is characterized by a high incidence of melanoma, as well as a 13-fold increased risk of pancreatic cancer (Goldstein et al. 1995; Whelan et al. 1995). The appearance and age of onset of pancreatic adenocarcinoma are variable among FAMMM kindreds with INK4A mutations, indicating a modulating role for environmental factors in disease penetrance (Goldstein et al. 2000; Lynch et al. 2002). FAMMM kindreds that harbor mutant loci other than INK4A, such as cyclin-dependent kinase 4 (CDK4) alleles that abrogate INK4A binding or other, as yet uncharacterized loci, do not have an increased incidence of PDAC (Goldstein et al. 1995; Borg et al. 2000), suggesting that INK4A may participate in pancreatic carcinogenesis by additional mechanisms distinct from its regulation of G1 CDK activity.
The 9q21 locus encodes two overlapping tumor suppressors—INK4A and ARF, and their respective protein products p16INK4A and p19ARF—via distinct first exons and alternative reading frames in shared downstream exons (Sherr 2001). INK4A inhibits CDK4/6-mediated phosphorylation of RB, thereby blocking entry into the S (DNA synthesis) phase of the cell cycle; ARF stabilizes p53 by inhibiting its MDM2-dependent proteolysis. Given this physical juxtaposition and frequent homozygous deletion of 9p21 (in 40% of tumors), many pancreatic cancers sustain loss of INK4A and ARF tumor suppression pathways. INK4A clearly plays a central role as a PDAC tumor suppressor, as germline and sporadic mutations have been identified that target INK4A, but spare ARF (Rozenblum et al. 1997; Liu et al. 1999; Lal et al. 2000). Recent mouse modeling studies have begun to shed light on the specific roles of INK4A and ARF tumor suppression pathways in this disease. These studies reinforce the relevance of INK4A in the pathogenesis of PDAC, since INK4A mutations cooperate with KRAS in the development of PDAC and accelerate tumor progression in the setting of concurrent mutations in p53 (Bardeesy et al. 2006). Moreover, ARF was shown to have an independent, cooperative function in PDAC tumor suppression. These mouse studies and the functional interactions of ARF and p53 are discussed in depth below.
INK4A displays a highly restricted expression pattern and is dispensable for normal development and tissue homeostasis (Zindy et al. 1997; Nielsen et al. 1999; Krimpenfort et al. 2001; Sharpless et al. 2001a). When primary cells are placed into culture, INK4A expression is induced via a presumed stress response to the inappropriate growth environment associated with in vitro culture (Sherr and DePinho 2000; Ramirez et al. 2001). Similar induction is observed in vivo in association with reactive processes and aging associated with senescence (Nielsen et al. 1999; Krishnamurthy et al. 2004). This regulation of INK4A by environmental stress, age, and aberrant proliferative signals provides a plausible basis for its tumor suppression function. A role for INK4A has also been suggested in response to telomere erosion, which is observed in PanINs (van Heek et al. 2002) and is an established stimulus of senescence, although other studies suggest that p53 and INK4A are operative in separate senescent pathways (Beausejour et al. 2003; Herbig et al. 2004; Jacobs and de Lange 2004).
The genetic evidence of coincident mutations in PDAC and a wide range of other human malignancies (lung, colon, ovarian, melanocytic, etc.) have led to a significant effort in functionally defining the relationship between RAS and INK4A/ARF. In human fibroblasts, robust RAS activation can induce INK4A, which, in turn, results in premature senescence (Serrano et al. 1997; Zhu et al. 1998; Brookes et al. 2002; Drayton et al. 2003; for review, see Sharpless and DePinho 2005), a presumed defense mechanism against oncogene activation. The relevance of these findings has recently been demonstrated in vivo with mouse models in which oncogenic Kras has been expressed in the lung and pancreatic ducts. Kras-induced preneoplastic lung adenomas and PanINs exhibit markers of senescence, including INK4A expression among others, which are then lost in fully transformed invasive lesions (Collado et al. 2005). Cooperation between activated RAS alleles and the loss of INK4A has been observed in animal models, including PDAC (Chin et al. 1997; Fisher et al. 2001; Aguirre et al. 2003). This relationship had initially been thought to contribute to the coincident mutations of RAS and INK4A in cancer—namely, that RAS mutation leads to senescence directly through induction of INK4A expression. Thus, mutation in RAS leads to selective pressure for subsequent mutation in the INK4A/ARF locus.
Several lines of evidence suggest, however, that the K-RAS-induced senescent phenotype requires additional cooperating stimuli, as K-RAS activation, in itself, seems to confer a survival advantage in certain cell systems. This is corroborated by studies in mouse cells in which an activated K-RAS allele, expressed at physiological levels, provokes immortalization of fibroblasts in vitro and preneoplastic hyperplasias of the lung and GI track when activated in vivo (Jackson et al. 2001; Guerra et al. 2003; Tuveson et al. 2004). The observation that K-RAS induces a proliferative phenotype without attendant senescence and INK4A induction is echoed in studies using very early passage human fibroblasts, which have not had time to experience culture-induced changes, and mouse fibroblasts grown under reduced oxygen tensions and in serum-free media. In both of these systems, which seek to minimize known prosenescent signals such as ROS, growth factors, and culture shock (for review, see Sherr and DePinho 2000), oncogenic K-RAS alone induced proliferation/immortalization phenotypes and senescence was not observed (Benanti and Galloway 2004; Woo and Poon 2004). Pathologic/molecular data from human pancreases are also supportive of a proliferative role for K-RAS. K-RAS mutations may be found in nonneoplastic states such as chronic pancreatitis and possibly in normal pancreas (Luttges et al. 1999). Furthermore, multiple different K-RAS mutations may be detected in individual PanIN lesions, suggesting that these mutants lead to the propagation of clonal populations (Moskaluk et al. 1997; Laghi et al. 2002). Finally, the loss of INK4A generally occurs in later stages of pancreatic neoplasia, subsequent to the acquisition of K-RAS mutations. Together, these observations point toward the possibility of intermediary events, such as disrupted contacts with the extracellular matrix, elevations in the level of activated K-RAS, concurrent growth factor signaling, or genomic damage from ROS, en route to INK4A loss and the development of frank PDAC.
The p53 tumor suppressor
The p53 tumor suppressor gene is mutated, generally by missense alterations of the DNA-binding domain, in >50% of PDAC cases (Rozenblum et al. 1997). Consistent with a role in constraining malignant progression, p53 mutation appears in later-stage PanINs that have acquired significant features of dysplasia (Boschman et al. 1994; Maitra et al. 2003). In these more advanced PanINs, the selective pressure to eliminate p53 may stem in part from a collective accumulation of genetic damage, from telomere erosion and ROS, for example, resulting in the activation of p53-dependent DNA damage checkpoint responses. Thus, loss of p53 function could serve to enable the growth and survival of cells harboring procarcinogenic chromosomal aberrations. Given that human PDAC is characterized by profound aneuploidy and complex chromosomal rearrangements, as well as significant intratumoral genomic heterogeneity, a clear understanding of how p53 participates in genome stability mechanisms would provide important insights into disease pathogenesis and ultimately treatment. The rampant genomic instability in PDAC could serve to both fuel the rise of advanced disease and provide a basis for its resistance to therapeutic modalities (Gorunova et al. 1998; Harada et al. 2002).
Finally, in many other cancer types, there exists a near reciprocal relationship in the loss of ARF and p53 (Rozenblum et al. 1997; Pomerantz et al. 1998; Ruas and Peters 1998; Sharpless and DePinho 1999). As mentioned above, this relationship likely reflects the fact that ARF inhibits MDM2-mediated targeting of the p53 protein for proteasomal degradation (Pomerantz et al. 1998; Zhang et al. 1998). Thus, ARF deficiency would result in marked reduction of p53 protein levels and attenuation of p53 pathway function in diverse cancer-relevant processes (Lowe and Sherr 2003). On the other hand, in human PDAC, p53 mutations and ARF deletions coexist in 40% of cases, potentially pointing to nonoverlapping functions for these factors in pancreatic cancer suppression (Heinmoller et al. 2000; Maitra et al. 2003; Hustinx et al. 2005). Mounting evidence suggests that ARF possesses p53-independent functions including the inhibition of ribosomal RNA processing (Rocha et al. 2003; Sugimoto et al. 2003; Qi et al. 2004; Paliwal et al. 2006). In addition, ARF does not appear to neutralize the DNA damage checkpoint that would be activated upon genetic damage (e.g., induced by teleomere dysfunction), thereby necessitating the additional loss of p53 function as such signals intensify during PDAC progression (Greenberg et al. 1999). Alternatively, ARF deletion in PDAC could represent a "bystander" effect associated with mutational events targeting INK4A. The resolution of these issues will require systematic genetic analysis of various mutant genotype combinations in murine PDAC mouse models as well a greater understanding of the molecular actions of ARF versus p53 in PDAC tumor biology.
The SMAD4/DPC4 tumor suppressor and complexities of transforming growth factor-
(TGF-) signaling
Another frequent event associated with PDAC progression is loss of the SMAD4 (DPC4) transcriptional regulator (Hahn et al. 1996), which serves as a central component in the TGF- signaling cascade (Massague et al. 2000). The SMAD4 gene maps to chromosome 18q21 and is targeted for deletion or intragenic point mutations in 50% of PDAC cases (Hahn et al. 1996). SMAD4 has been designated a progression allele for PDAC on the basis of its loss in later-stage PanINs (Wilentz et al. 2000; Luttges et al. 2001; Maitra et al. 2003). The impact of SMAD4 loss on PDAC prognosis is not clearly established, and different studies have reached opposite conclusions regarding the relationship between SMAD4 status and survival (Tascilar et al. 2001; Biankin et al. 2002). On the histo-pathologic level, tumors with an intact SMAD4 may have a higher propensity for showing poorly differentiated features (Biankin et al. 2002). The mechanism by which SMAD4 loss contributes to tumorigenesis is likely to involve its central role in TGF--mediated growth inhibition.
TGF-.
TGF- is the prototypic member of a superfamily of secreted proteins, whose other members include the Bone Morphogenic Proteins (BMPs) and Activins (for review, see ten Dijke and Hill 2004). These growth factors signal through serine/threonine kinase receptor complexes that, upon ligand binding, phosphorylate receptor-regulated Smad proteins (SMAD2, SMAD3, and the obligate binding partner SMAD4) regulating a variety of cellular functions including proliferation, differentiation, migration, and apoptosis. The biological role of the TGF- pathway in human malignancy is complex, exerting both growth-inhibitory and growth-promoting effects depending on the cell type and cell context (for review, see Siegel and Massague 2003). In numerous epithelial cell lines and in epithelial tissue in vivo, TGF- exerts a growth inhibitory program that involves modulation of cell cycle regulators including induction of p15INK4B and p21CIP1 expression and repression of c-Myc and ID family transcription factors, as well as induction of apoptotic machinery, and repression of telomerase (for review, see Elliott and Blobe 2005). Likewise, elevations in TGF- signaling inhibit epithelial cancer initiation in vivo, and lesions in this pathway promote intestinal, ovarian, and pancreatic tumorigenesis. On the other hand, TGF- promotes the proliferation and transformation of fibroblasts and the epithelial-to-mesenchymal transition (EMT) in breast cancer and skin cancer, a process by which advanced carcinomas lose their differentiated features and acquire a highly aggressive, invasive phenotype (Janda et al. 2002; Oft et al. 2002; Tang et al. 2003; for review, see Zavadil and Bottinger 2005). Therefore, in some carcinomas, TGF- signaling can have biphasic effects, inhibiting tumor initiation yet promoting the high-grade advancement of established tumors (Akhurst and Derynck 2001).
The importance of TGF- signaling in pancreatic cancer is illustrated by the fact that 90% of tumors show loss of heterozygosity (LOH) at the SMAD4 locus, with 50% of PDAC having either homozygous deletion or mutational inactivation of the second allele, as discussed above. The loss of SMAD4 in PDAC may have a primary role in modulating the interaction of the tumor with the microenvironment rather than in growth control of the tumor cells themselves. Along these lines, SMAD4 restoration in some pancreatic cancer cell lines has a minimal impact on cell growth in vitro, although some inhibition of anchorage-independent growth has been observed in specific cell lines. Importantly, the prominent impact of SMAD4 restoration has been observed in tumor formation in xenotransplants with documented repression of angiogenesis and extracellular matrix remodeling (Schwarte-Waldhoff et al. 2000; Peng et al. 2002; Duda et al. 2003).
There is recent evidence that SMAD4 deficiency may inhibit TGF--induced cell cycle arrest and cell migration, while not affecting EMT, thereby shifting the balance of TGF- signaling from tumor suppression to tumor promotion (Levy and Hill 2005). Consistent with these observations, it appears that elevated TGF- expression contributes to PDAC progression. TGF- family ligands are expressed at elevated levels in PDAC cells relative to normal pancreas (Friess et al. 1993) and may help to promote the characteristic desmoplastic response of this malignancy as suggested from xenograft studies (Lohr et al. 2001). TGF- signaling may also contribute to tumorigenesis in an autocrine manner since PDACs often overexpress the type II TGF- receptor relative to normal pancreas (Wagner et al. 1999; for review, see Rane et al. 2006) while experimental blockade of TGF- signaling by expression of soluble type II TGF- receptor attenuates tumorigenicity and metastasis of xenografts (Rowland-Goldsmith et al. 2001, 2002). Furthermore, antibodies to TGF- inhibit the invasion of PDAC cell lines in vitro, while exogenous addition of this cytokine enhanced invasion and promotes the EMT (Ellenrieder et al. 2001a, b).
The LKB1/STK11 tumor suppressor
The Peutz-Jeghers syndrome (PJS), linked to LKB1/STK11 mutations, is another familial cancer syndrome associated with an increased incidence of PDAC (Hemminki et al. 1998; Jenne et al. 1998; Giardiello et al. 2000). PJS patients are primarily afflicted with benign intestinal polyposis at a young age (Cooper 1998), although advancing age carries increased risk of gastrointestinal malignancies including a >40-fold increase in PDAC (Giardiello et al. 2000). At the same time, somatic mutation of LKB1 in sporadic PDAC appears to be rare, detected in only 4%–6% of sporadic cases examined (Su et al. 1999), although there is some evidence that the rates of inactivation are higher in IPMNs (Sahin et al. 2003).
LKB1 encodes a serine/threonine kinase that is involved in regulation of diverse processes such as cell polarity and metabolism, and has been linked to specific signaling pathways including mTOR, the latter via its capacity to regulate AMPK (Bardeesy et al. 2002b; Ossipova et al. 2003; Baas et al. 2004; Corradetti et al. 2004; Lizcano et al. 2004; Shaw et al. 2004a, b, 2005; Hardie 2005). Exactly how LKB1 loss, and through deregulation of which of these pathways/processes, promotes tumorigenesis remains to be established. Control of mTOR signaling through AMPK links this gene to a common pathway harboring two other tumor suppressors, PTEN and TSC. The biochemical link to these well-characterized cancer signaling pathways may provide insights into the biological mechanisms through which LKB1 suppresses tumor formation. At the same time, the role of LKB1 in cell polarity, and likely regulation of several less-well-characterized kinases, leaves open several other plausible mechanisms of tumor suppression. Current efforts are now directed toward defining additional LKB1 substrates and linked biological processes.
The BRCA2 tumor suppressor.
Inherited BRCA2 mutations are typically associated with familial breast and ovarian cancer syndrome, but also carry a significant risk for the development of pancreatic cancer. One study estimates that 17% of pancreatic cancers occurring in a familial setting harbor mutations in this gene (Murphy et al. 2002); the del6174T founder mutation is particularly common in familial pancreatic cancers that arise in the Ashkenazi Jewish population. As is the case for those with germline INK4A mutations, the penetrance of PDAC in BRCA2 mutation carriers is relatively low, and the age of onset is similar to patients with the sporadic form of the disease. Loss of the wild-type BRCA2 allele seems to be a late event in those inheriting germline heterozygous mutations of BRCA2, restricted to severely dysplastic PanINs and PDACs (Goggins et al. 2000). Together, these data are consistent with the model that loss of function of BRCA2 promotes the malignant progression of pancreatic neoplasms.
BRCA2 is known to play a critical role in the maintenance of genomic stability by regulating homologous recombination-based DNA repair processes. Consequently, BRCA2 deficiency in normal cells results in the accumulation of procarcinogenic or lethal chromosomal aberrations (Venkitaraman 2002). The fact that BRCA2 is selectively mutated late in tumorigenesis likely reflects the need for DNA damage response pathways (which in a normal cell would lead to senescence or apoptosis) to be inactivated first—for example, by p53 mutation—so that the genetic damage incurred can be tolerated. Thus, as is the case for telomeres, the carcinogenic role of BRCA2 deficiency may be manifest only in the appropriate genotypic and cell-type context. BRCA2 mutational status may also have therapeutic implications, as it seems to confer susceptibility to DNA cross-linking agents such as Mitomycin C, as is seen in other related Fanconi anemia family genes, particularly FancG and FancC (Taniguchi et al. 2003; van der Heijden et al. 2004).
Additional growth factor receptor signaling circuits in PDAC
Epidermal growth factor.
PDAC shows elevated expression of EGF receptors (EGFR and ERBB3) and their ligands (TGF- and EGF), consistent with the presence of an autocrine loop (Barton et al. 1991; Korc et al. 1992; Lemoine et al. 1992; Friess et al. 1995, 1999). Importantly, EGFR inhibitors decrease PDAC cell growth and tumorigenesis in vitro (J. Li et al. 2004), as well as inhibit growth of orthotopic tumors in combination with cytotoxic chemotherapy (Bruns et al. 2000). This inhibition appears to be due to a decrease in tumor vasculature through inhibition of proangiogenic factors, resulting in endothelial apoptosis. In line with these antineoplastic activities, EGFR inhibitors have been approved for clinical use in PDAC patients.
Insulin-like growth factor (IGF).
The IGF signaling pathway regulates survival, invasion, and angiogenesis of many human cancers. PDACs show elevated expression of IGF-I in both the tumor cells and the stroma and display aberrant activation of the IGF-I receptor (IGF-IR) in tumor cells (Bergmann et al. 1995; Ouban et al. 2003; Stoeltzing et al. 2003). In vitro, autocrine IGF-I signaling promotes cell proliferation and growth-factor-independent survival (Nair et al. 2001). Inhibition of the pathway by anti-IGF-IR antibodies or expression of a dominant-negative form of IGF-IR inhibits the growth of xenografts and sensitizes tumor cells to chemotherapy (Maloney et al. 2003; Min et al. 2003).
Met and hepatocyte growth factor (HGF).
The Met receptor tyrosine kinase and its ligand, HGF/scatter factor, regulate cell motility, invasion, and proliferation and the deregulation of this signaling pathway contributes to the progression of several malignancies (for review, see Corso et al. 2005). The Met receptor is expressed at low levels in the exocrine pancreas and shows marked up-regulation in PanIN lesions and in PDACs. Additionally, HGF is induced during PDAC progression, present in the epithelium of PanIN lesions and in the stromal cells of advanced tumors (Ebert et al. 1994; Di Renzo et al. 1995; Furukawa et al. 1995; Paciucci et al. 1998). HGF promotes motility of PDAC cells in vitro, and inhibition of this pathway through the administration of blocking antibodies or the truncated HGF fragment NK4 inhibits invasive growth and angiogenesis of xenografts (Tomioka et al. 2001; Saimura et al. 2002).
Fibroblast growth factor (FGF).
FGF signaling (for review, see Cross and Claesson-Welsh 2001) appears to contribute to mitogenesis and angiogenesis of PDAC. Expression of numerous FGF receptors and glypican-1, a membrane heparin sulfate proteoglycan that facilitates FGF–FGFR interactions, have been detected in primary PDAC samples (Kobrin et al. 1993; Yamanaka et al. 1993a, b; Ohta et al. 1995; Kornmann et al. 1997; Ishiwata et al. 1998; Kleeff et al. 1998; Kornmann et al. 2002). Consistent with a role for FGF signaling in supporting PDAC growth, dominant-negative FGFR-1 mutants or antisense glypican-1 can inhibit the growth of pancreatic cancer cell lines in vitro and suppress their tumorigenic potential in xenografts (Wagner et al. 1998a; Ogawa et al. 2002; Kleeff et al. 2004). FGF signaling may also contribute to the desmoplasia associated with PDAC, since elevated bFGF levels are associated with this phenotype in primary tumors (Kuniyasu et al. 2001).
VEGF.
VEGF promotes endothelial cell proliferation and survival by binding to the VEGFR-1 and VEGFR-2 endothelial cell transmembrane receptors (for review, see Ferrara et al. 2003). VEGF is overexpressed by PDAC cells (Itakura et al. 1997; Seo et al. 2000), whereas disruption of VEGF signaling by expression of soluble VEGF receptors, VEGF high-affinity binding chimeras, anti-VEGF antibodies, or ribozymes strongly suppresses the tumorigenic growth of pancreatic cancer xenografts (von Marschall et al. 2000; Hoshida et al. 2002; Tokunaga et al. 2002; Hotz et al. 2003; Fukasawa and Korc 2004). VEGF-C, a regulator of lymphoangiogenesis, is also overexpressed in PDAC and may contribute to lymphatic spread and the lymph node metastasis common in this malignancy (Tang et al. 2001; Kurahara et al. 2004). Further study will be required to validate VEGF-C as a drug development target.
Developmental signaling pathways in PDAC
The roles of the Hedgehog and Notch signaling pathways in PDAC pathogenesis have recently been appreciated, further drawing attention to the connections between development and cancer. The relationship between these two pathways, normal pancreatic organogenesis, tissue homeostasis, disease, and the development of cancer are discussed in detail in subsequent sections. Below we briefly describe the biochemical and molecular circuitry of Hedgehog and Notch signaling and their links with PDAC.
Hedgehog.
The mammalian Hedgehog family of secreted signaling proteins comprised of Sonic, Indian, and Desert Hedgehog (SHH, IHH, and DHH, respectively) regulates the growth and patterning of many organs, including the pancreas, during embryogenesis (Ingham and McMahon 2001). The Hedgehog pathway is negatively regulated by the Patched (PTC) tumor suppressor protein, which tonically inactivates the Smoothed protein (SMO). Hedgehog ligands engage the PTC transmembrane protein and disrupt inhibition of Smo, activating the Gli family of transcriptional regulators. Alterations that activate this pathway, including loss of PTC, activating mutations in SMO, and overexpression of GLI and HH proteins, have been implicated in a variety cancers (for reviews on Hedgehog signaling and cancer, see Taipale and Beachy 2001; Pasca di Magliano and Hebrok 2003). Activation of the Hedgehog pathway has been implicated in both the initiation of pancreatic ductal neoplasia and in the maintenance of advanced cancers. SHH is absent from the normal adult pancreas, but is activated in PanINs, exhibiting a graded increase in progressively later-stage lesions and carcinomas, where signaling seems to be necessary for tumor maintenance (Berman et al. 2003; Thayer et al. 2003).
Notch.
The Notch signaling pathway, which is important in directing cell fate and cell proliferation during embryonic development, has been shown to contribute to cell transformation in vitro and to the development of human cancers when aberrantly regulated (for review, see Radtke and Raj 2003; Kadesch 2004; Lai 2004; Sjolund et al. 2005). Notch pathway activation involves the binding of membrane-bound Notch receptors (Notch 1–4) to their ligands (Delta-like and Jagged). These receptor–ligand interactions induce proteolysis of the Notch receptor and subsequent nuclear translocation of the Notch intracellular domain (NICD), which mediates the transcriptional activation of a series of target genes. Notch and its ligands are expressed at low or undetectable levels in the normal adult pancreas. However, in PanIN lesions and in pancreatic adenocarcinomas, there are prominent elevations in expression of these factors and an associated induction of transcriptional target genes such as HES-1, consistent with activation of this pathway during malignant progression of this malignancy (Miyamoto et al. 2003). Although ectopic activation of Notch signaling within rodent pancreatic progenitor cells in vivo does not result in subsequent carcinogenesis (Murtaugh et al. 2003), there is an accumulating body of literature demonstrating interactions of Notch with RAS, both in development as well as in tumorigenesis (for review, see Sundaram 2005). In particular, several different cell-based systems have shown that activated RAS cooperates with Notch to t
