Cell Host & Microbe

Review Helicobacter pylori CagA and Gastric Cancer: A Paradigm for Hit-and-Run Carcinogenesis Masanori Hatakeyama1,* 1Division of Microbiology, Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-Ku, Tokyo 113-0033, Japan *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chom.2014.02.008

Helicobacter pylori is a gastric bacterial pathogen that is etiologically linked to human gastric cancer. The cytotoxin-associated gene A (CagA) protein of H. pylori, which is delivered into gastric epithelial cells via bacterial type IV secretion, is an oncoprotein that can induce malignant neoplasms in mammals. Upon delivery, CagA perturbs multiple host signaling pathways by acting as an extrinsic scaffold or hub protein. On one hand, signals aberrantly raised by CagA are integrated into a direct oncogenic insult, whereas on the other hand, they engender genetic instability. Despite its decisive role in the development of gastric cancer, CagA is not required for the maintenance of a neoplastic phenotype in established cancer cells. Therefore, CagA-conducted gastric carcinogenesis progresses through a hit-and-run mechanism in which pro-oncogenic actions of CagA are successively taken over by a series of genetic and/or epigenetic alterations compiled in cancer-predisposing cells during long-standing infection with cagA-positive H. pylori. Introduction Gastric cancer is the fourth most common malignancy and the second leading cause of cancer-related deaths, accounting for 10% of total cancer deaths worldwide (Parkin, 2004). The vast majority of gastric cancers are adenocarcinomas, which are pathohistologically classified into intestinal and diffuse types. Gastric cancer is also characterized by large geographical variations in its incidence and, indeed, more than half of the total gastric cancers are in East Asian countries such as Japan, South Korea, and China. Like other cancers, gastric cancer develops through a multistep process, the progression of which is variably influenced by both host genetic and environmental factors (Correa, 1988; Shanks and El-Omar, 2009). Among those, the gastric pathogen Helicobacter pylori is the strongest risk factor, playing virtually indispensable roles in the development of both intestinal and diffuse gastric noncardia adenocarcinomas (Peek and Blaser, 2002). The gastric pathogen has coevolved with and adapted well to human populations and has currently infected more than half of the entire human population. Given that most members of the genus Helicobacter thrive in the intestinal and/or liver tract of mammals, the ancestor of H. pylori could have been one of such enterohepatic Helicobacter species that successfully migrated to the stomach upon the acquisition of urease during evolution. H. pylori urease converts urea from gastric juices into bicarbonate and ammonia, thereby neutralizing stomach acid around H. pylori for its survival. Subsequently, a fraction of H. pylori strains became more virulent by acquiring the ability to produce and secrete a protein called cytotoxin-associated gene A (CagA). CagA is encoded by the cagA gene, one of 30 genes present in a 40 kp DNA segment termed the cag pathogenicity island (cag PAI), which is thought to have been introduced into the H. pylori genome via horizontal transfer from an unknown organism. Many of the cag PAI genes encode components of a syringe-like pilus structure known as a type IV secretion system (TFSS) through which CagA is delivered into the cytoplasm of gastric epithelial cells during bacterial attachment 306 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

(Fischer, 2011). The worldwide infection ratio of cagA-positive versus -negative strain is approximately 6:4, except in East Asian countries, where almost all of the isolated H. pylori strains are cagA-positive. Infection with cagA-positive strains increases the risk of gastric cancer by at least one order of magnitude more than that infection with cagA-negative strains (Ekstro¨m et al., 2001). Considering technical problems in detecting cagA or CagA in earlier studies, it is highly likely that the cagA-positive strain, and not the cagA-negative strain, is the H. pylori that directs gastric carcinogenesis. This notion is experimentally supported by the infection of Mongolian gerbils with H. pylori demonstrating that intact cag PAI robustly facilitates bacterial colonization in the stomach body, resulting in corpus-predominant atrophic gastritis, a precancerous condition (Rieder et al., 2005). The oncogenic potential of CagA was directly demonstrated by the observation that transgenic mice systemically expressing cagA spontaneously developed gastrointestinal carcinomas and hematopoietic malignancies (Ohnishi et al., 2008). In a Zebrafish model, transgenic expression of CagA enhanced intestinal carcinogenesis caused by a mutation in the tumorsuppressor gene p53 (Neal et al., 2013), providing additional in vivo evidence for the oncogenic capacity of CagA. Although the contribution of other H. pylori proteins to disease development, such as vacuolating cytotoxin A (VacA) and outer inflammatory protein A (OipA), has also been shown (Yamaoka, 2010), accumulating evidence clearly points to a central role of CagA in gastric carcinogenesis. CagA as a Bacterium-Derived Scaffold or Hub Protein CagA is a 120–145 KDa H. pylori protein that shows no significant homology with known proteins. The size variation is due to structural diversity in its C-terminal region (Hatakeyama, 2004). CagA is the only known protein substrate of the H. pylori TFSS. Upon delivery into gastric epithelial cells, CagA is localized to the inner leaflet of the plasma membrane, where it undergoes tyrosine phosphorylation. Initial mechanistic insight into the pathogenic action of CagA was provided upon the discovery that

Cell Host & Microbe

Review Figure 1. Structural Diversity of H. pylori CagA (A and B) East Asian CagA (A) and Western CagA (B) are respectively characterized by the presence of an EPIYA-D segment and an EPIYA-C segment, to which SHP2 binds in a tyrosine-phosphorylation-dependent manner. In Western CagA species, the EPIYA-C segment duplicates with a variable number, usually one to three times, in tandem. The EPIYA-A and EPIYA-B segments serve as sites for Csk binding upon tyrosine phosphorylation. The CM motif interacts with PAR1. In general, Western CagA contains a single Western CM motif, whereas East Asian CagA possesses at least two Western CM motifs. (C) There are a few CagA variants containing atypical combinations of Western and East Asian CM motifs.

host-cell-delivered CagA specifically interacts with the SH2domain-containing protein tyrosine phosphatase (SHP2) in a tyrosine-phosphorylation-dependent manner (Higashi et al., 2002). The finding draws an analogy between CagA and the mammalian Gab family of scaffold proteins, which also form a complex with SHP2 in a tyrosine phosphorylation-dependent manner beneath the plasma membrane. This parallel gives rise to the possibility that, despite a lack of sequence homology, the bacterial protein behaves as a pathogenic scaffold or hub in mammalian cells (Hatakeyama, 2003). The idea that CagA is a Gab mimic is appealing and supported by Drosophila genetics demonstrating that CagA can rescue a deficiency in Drosophila Gab (DOS). Moreover, the CagA-mediated rescue of DOS deficiency requires Drosophila SHP2 (Corkscrew), providing additional evidence for the genetic interaction between CagA and SHP2 (Botham et al., 2008). CagA has been reported to interact with a variety of human proteins in both phosphorylation-dependent and -independent manners (Hatakeyama, 2011). Although many of those interactors require additional confirmational studies, it is certain that CagA acts as a promiscuous scaffold or hub protein that simultaneously disturbs multiple host proteins. EPIYA-Dependent CagA-Host Protein Interactions The CagA C-terminal region is noted for the presence of a variable number of Glu-Pro-Ile-Tyr-Ala (EPIYA) motifs, which serve as tyrosine phosphorylation sites. With respect to the amino acid sequence surrounding each of the EPIYA motifs, four distinct EPIYA segments (termed EPIYA-A–EPIYA-D) have been classified (Hatakeyama, 2004, 2011). H. pylori strains circulating in East Asian countries carry East Asian CagA, which contains EPIYA-A, EPIYA-B, and EPIYA-D segments, whereas H. pylori strains in the rest of the world carry Western CagA, which contains EPIYA-A, EPIYA-B, and multiple EPIYA-C segments (usually repeated one to three times in tandem) (Figure 1). When delivered into host cells, CagA undergoes tyrosine phosphorylation on these EPIYA segments initially by Src family kinases (SFKs) and then c-Abl kinase. A recent study showed

that SFKs specifically phosphorylate the EPIYA-C or EPIYA-D segments, whereas c-Abl phosphorylates all four EPIYA segments without overt specificity (Mueller et al., 2012). A critical role for EPIYA phosphorylation in in vivo tumorigenesis has been indicated by the observation that, in contrast to transgenic mice expressing wild-type CagA, transgenic mice expressing a phosphorylation-resistant CagA failed to induce neoplasms (Ohnishi et al., 2008). Upon tyrosine phosphorylation, the EPIYA-C or EPIYA-D segment acquires the ability to bind to one of the two mutually related SH2 domains of SHP2. SHP2, a ubiquitously expressed protein tyrosine phosphatase, is required for the full activation of the Ras-Erk mitogenic pathway. Through the complex formation, CagA deregulates the phosphatase activity of SHP2, which in turn potentiates Ras-Erk signaling (Saito et al., 2010). CagAactivated SHP2 also dephosphorylates focal adhesion kinase, a kinase-regulating cell shape and motility, and thereby inhibits the its kinase activity, causing impaired focal adhesions that are associated with an elongated cell shape known as the hummingbird phenotype and elevated cell motility (Segal et al., 1999; Tsutsumi et al., 2006). The CagA-SHP2 interaction has drawn much attention because SHP2 is a potential oncoprotein (Matozaki et al., 2009). Somatic gain-of-function mutations in PTPN11, the gene encoding SHP2, are found in a variety of human cancers, and thus SHP2 is recognized as a proto-oncogenic phosphatase (Chan and Feng, 2007). Remarkably, the EPIYAD segment binds to SHP2 more strongly than the EPIYA-C segment does. Additionally, transgenic mice expressing East Asian CagA develop tumors more frequently than those expressing Western CagA (Miura et al., 2009). Clinically, infection with H. pylori carrying East Asian CagA is associated with severe gastric atrophy and gastric cancer (Satomi et al., 2006). Altogether, these results indicate that endemic circulation of H. pylori carrying biologically more active CagA may be one major factor underpinning the high incidence of gastric cancer in East Asian countries. In Western countries, infection with H. pylori cagA-positive strains with larger numbers of CagA EPIYA-C segments was associated with atrophic gastritis and increased the risk of gastric cancer (Basso et al., 2008; Ferreira et al., 2012). Apart from its contribution in carcinogenesis, a Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc. 307

Cell Host & Microbe

Review potential role for the CagA-SHP2 interaction in bacterial colonization has also been described. During infection with cagA-positive H. pylori, CagA-deregulated SHP2 inhibits the expression of b-defensin 3, an antimicrobial peptide active against H. pylori, by tyrosine-dephosphorylating EGF receptor and thereby terminating EGF-dependent intracellular signaling events (Bauer et al., 2012). Along with the report that CagA induces REG3g, a secreted C-type lectin that exerts direct bactericidal activity (Lee et al., 2012), CagA delivery may allow H. pylori to manipulate host innate immunity in order to favor its own survival in the stomach mucosa. The tyrosine-phosphorylated EPIYA-A or EPIYA-B segment also binds to the SH2 domain of the protein tyrosine kinase Csk (Hatakeyama, 2004). The CagA-Csk interaction recruits Csk to the plasma membrane, where Csk phosphorylates membrane-bound SFKs. Given that phosphorylation inhibits SFK kinase activity, Csk activation results in reduced EPIYA phosphorylation of CagA. Therefore, the CagA-Csk interaction may establish a negative feedback regulatory loop that prevents excess activation of tyrosine phosphorylation-dependent CagA activity, which could be harmful to the long-term colonization of H. pylori in the stomach. Inhibition of SFK activity by CagA also alters the phosphorylation status of actin binding proteins such as cortactin and vinculin, resulting in actin cytoskeletal rearrangements that promote cell motility (Tegtmeyer and Backert, 2011). In addition to SHP2 and Csk, CagA associates with Crk adaptor proteins (Crk-I, Crk-II, and Crk-L) in an EPIYA-phosphorylation-dependent manner (Suzuki et al., 2005), although details of these interactions are unknown. The resulting CagACrk complex exerts pleiotropic effects on signaling pathways, such as Sos1/H-Ras/Raf1, C3G/Rap1/B-Raf, and Dock180/ Rac1/WASP, that control cell spreading, motility, and proliferation and thus further enhance the motogenetic activity of CagA. Systemic proteomic screening with high resolution mass-spectrometry revealed that EPIYA-containing peptides derived from CagA are capable of interacting with a variety of SH2-containing proteins in a tyrosine-phosphorylation-dependent manner (Selbach et al., 2009), suggesting that SH2 domain selectivity of the bacterial EPIYA segments is unexpectedly low. CagA has also been reported to interact with the adaptor protein Grb2 and thereby activates Ras-Erk signaling in an EPIYAdependent manner (Mimuro et al., 2002). However, the CagA-Grb2 interaction occurs independently of EPIYA tyrosine phosphorylation, arguing against the idea that the interaction involves the SH2 domains of Grb2. Altogether, the bacterial protein targets pro-oncogenic Ras signaling through multiple distinct mechanisms. Given that recently identified bacterial tyrosine kinases (BY kinases) share no homology with eukaryotic tyrosine kinases (Lee and Jia, 2009) and that there is no evidence indicating that CagA is tyrosine phosphorylated inside the bacteria, the origin of the EPIYA tyrosine-phosphorylation motif in CagA remains enigmatic. In mammalian proteomes, EPIYA- or EPIYA-likemotif-containing proteins have been reportedly underrepresented (Selbach et al., 2009), whereas several bacterial effectors other than CagA also contain EPIYA or EPIYA-like motifs that can undergo tyrosine phosphorylation upon delivery into mammalian cells (Hayashi et al., 2013). Along with the seemingly low SH2 308 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

domain selectivity, these findings led to the proposal of a ‘‘master key’’ hypothesis that bacterial EPIYA effectors promiscuously perturb a variety of mammalian SH2-containing proteins, activity that may be dangerous for the sophisticated regulation of the SH2-domain-mediated signaling network in mammalian cells (Selbach et al., 2009). This interesting possibility obviously warrants further investigation. Be that as it may, there exists a mammalian protein (termed Pragmin or Sgk223) that undergoes tyrosine phosphorylation on the EPIYA motif by SFKs (Safari et al., 2011). Pragmin was originally identified as a cytoplasmic pseudokinase that specifically stimulates RhoA GTPase activity upon interacting with Rnd2 GTPase (Tanaka et al., 2006). In stark contrast to the unfaithful bacterial EPIYA motif, the Pragmin EPIYA motif specifically binds to the SH2 domain of Csk, and, unlike the CagA-Csk interaction, the Pragmin-Csk interaction sequesters Csk in the cytoplasm, thereby preventing Csk-mediated SFK phosphorylation and inactivation at the membrane. Thus, Pragmin negatively regulates the Csk-SFK signaling axis in an EPIYA-dependent manner, giving rise to sustained activation of SFKs. Since many bacterial EPIYA effectors commonly target Csk (Hayashi et al., 2013), it is possible that the Csk-SFK axis is an Achilles’ heel overlooked in mammalian cells, the malfunction of which may greatly increase the chance of successful bacterial infection. CM-Motif-Mediated CagA-Host Protein Interactions The CagA C-terminal region also contains a 16-amino-acid sequence termed the CagA-multimerization (CM) motif, the number of which varies among distinct CagA species (Figure 1) (Ren et al., 2006). East Asian CagA contains a single CM motif (termed East Asian CM) that is located immediately distal to the EPIYA-D segment. Western CagA also has a CM motif after the last repeat of the EPIYA-C segment, which is either the same as East Asian CM or closely related to it (termed Western CM). Interestingly, the N-terminal 16-amino-acid residues of the EPIYA-C segment are completely identical to those of Western CM. Thus, almost all Western CagA species contain two or more CM motifs (Hatakeyama, 2011). Although the CM motif was originally identified as a sequence that directs CagA dimerization, it acts as a binding site for a polarity-regulating serine/ threonine kinase, partitioning defective-1 (PAR1; also known as microtubule affinity-regulating kinase in mammals). Through complex formation, the CM motif occupies the substrate binding cleft of PAR1 and thereby inhibits its kinase activity (Saadat et al.,  et al., 2010). Of four PAR1 isoforms (PAR1a–PAR1d) 2007; Nesic in mammals, CagA binds most strongly to PAR1b, a predominant PAR1 member in epithelial cells (Hatakeyama, 2011). Because PAR1b forms a homodimer in the cell (Nagase et al., 2011), the resulting PAR1b dimer is capable of interacting with two CagA molecules simultaneously. This was the reason why the PAR1 binding sequence was originally defined as the motif through which CagA multimerizes (dimerizes). The ‘‘passive’’ dimerization of CagA is important for the increased biological activity of CagA carrying a single EPIYA-C or EPIYA-D segment, whereas monovalent interaction of such a type of CagA with one of the two SH2 domains of SHP2 is relatively weak, PAR1-mediated CagA dimerization enables divalent interaction of the two SH2 domains of a single SHP2 with two CagA proteins in trans,

Cell Host & Microbe

Review

Figure 2. Coordinated Deregulation of Ras and Wnt Pathways by CagA Membrane-tethered CagA recruits a fraction of cytoplasmic SHP2 to the membrane. CagA-deregulated SHP2 hyperstimulates Ras-Erk signaling. Activated Ras, while inducing a mitogenic signal, promotes nuclear translocalization of cytoplasmic SHP2, which is not bound to CagA. At the membrane, CagA also dissociates b-catenin from E-cadherin and accelerates its nuclear translocalization. In the nucleus, where both SHP2 and b-catenin are forced to accumulate by the action of CagA, SHP2 dephosphorylates parafibromin (PF) and thereby enables the formation of the PF/b-catenin complex, which potentiates the activation of Wnt target genes.

which markedly increases the stability of the CagA-SHP2 complex (Nagase et al., 2011). Meanwhile, a CagA protein possessing more than two EPIYA-C segments can stably associate with a single SHP2 in cis, and the amount of the CagA-SHP2 complex exponentially increases as the repeat number of EPIYA-C segments in a single CagA protein increases. PAR1 is required for the development and maintenance of tight junctions (Suzuki and Ohno, 2006). Therefore, the inhibition of PAR1 activity by CagA leads to the disruption of tight junctions (Saadat et al., 2007). Given that tight junctions not only act as an intercellular barrier but also determine the apical and lateral membrane border, their disruption also causes a loss of epithelial apical-basal polarity. As a consequence, CagA-expressing cells extrude from the surrounding polarized epithelial monolayer and initiate multiple rounds of cell divisions, whereas undergoing morphological changes resembling the epithelial-mesenchymal transition (EMT) (Bagnoli et al., 2005; Saito et al., 2010). PAR1 also regulates the stability of microtubules by phosphorylating microtubule-associated proteins (Timm et al., 2008). Because the mitotic spindle is a microtubule-based structure, impaired functioning of microtubules by CagA-mediated PAR1 inhibition also perturbs mitosis, especially the segregation of sister chromatids (Umeda et al., 2009). Thus, CagA-expressing cells display chromosomal instability (CIN), a phenotype that plays an important role in the neoplastic transformation of cells. While promoting epithelial carcinogenesis, the CagA-PAR1 interaction has a unique role in the establishment of H. pylori infection. In a polarized epithelial monolayer, cagA-positive H. pylori is able to replicate and grow directly on the cell surface. This infection process requires the disruption of epithelial cell polarity, which enables adhered H. pylori to convert the apical membrane into a

replicative niche via CagA-mediated PAR1 inhibition (Tan et al., 2009). Although the precise mechanisms remain unknown, CagA-mediated polarity loss also stimulates the mislocalization of transferrin receptor apically to sites of bacterial microcolonies, ensuring the acquisition of iron from the host cell for bacterial growth and proliferation (Tan et al., 2011). CagA interacts with E-cadherin independently of tyrosine phosphorylation (Murata-Kamiya et al., 2007). Given that the interaction requires the CM motif, CagA may indirectly associate with E-cadherin via PAR1. The CagA-E-cadherin interaction destabilizes the E-cadherin/b-catenin complex to release b-catenin to the cytoplasm and/or nucleus. b-catenin is a key target of the canonical Wnt signaling pathway that plays critical roles in embryonic development and tissue homeostasis. In the canonical Wnt signaling pathway, binding of Wnt ligands to the Frizzled receptors leads to the stabilization of b-catenin, which then enters the nucleus and transactivates Wnt target genes. Consistent with the notion that Wnt target genes are critically involved in the regulation of cell proliferation and differentiation, deregulated b-catenin is associated with various human cancers (Saito-Diaz et al., 2013). It was recently reported that activated Ras-Erk signaling enhances nuclear translocalization of cytoplasmic SHP2 that is bound to YAP or its homolog TAZ, a target of the tumor-suppressive Hippo signaling pathway that controls cell proliferation and apoptosis (Takahashi et al., 2011; Tsutsumi et al., 2013). In the nucleus, SHP2 mediates tyrosine dephosphorylation of parafibromin (also known as CDC73), enabling complex formation of parafibromin with b-catenin, which enhances Wnt target gene activation. Altogether, membrane-tethered CagA recruits a fraction of cytoplasmic SHP2 molecules to the membrane, where SHP2 potentiates the Ras-Erk signaling, which in turn promotes nuclear translocalization of CagA-unbound SHP2. Accordingly, in cells expressing CagA, both b-catenin and SHP2 are forced to accumulate in the nucleus, where they collaborate in aberrant Wnt activation (Figure 2). In addition to PAR1, CagA associates with the c-Met receptor tyrosine kinase through the CM motif. The interaction deregulates c-Met kinase activity and thus aberrantly stimulates PI3-kinase/Akt signaling, which further potentiates Wnt signaling by preventing b-catenin degradation (Suzuki et al., 2009). Transgenic expression of CagA in the intestinal tract of Zebrafish resulted in abnormal proliferation of epithelial cells via Wnt activation, providing in vivo evidence for the pathophysiological interaction between CagA and the Wnt pathway (Neal et al., 2013). The importance of b-catenin activation in gastric carcinogenesis is also supported by the observation that malignant lesions in the stomach of Mongolian gerbils infected with a carcinogenic H. pylori strain are characterized by the nuclear accumulation of b-catenin and activation of Wnt target genes (Franco et al., 2005). Considering the role of the CM motif in the interaction of CagA with cellular targets such as PAR1 and c-Met, variations in its number and sequence may also influence the degree of CagA virulence and disease outcome as with the cases of CagA EPIYA segments. Tertiary Structural Basis of CagA Action 3D structural analysis of CagA revealed that the bacterial oncoprotein has a unique tertiary structure consisting of a solid N-terminal region (70% of the entire CagA) and an intrinsically Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc. 309

Cell Host & Microbe

Review Figure 3. Flip-Flop Regulation of CagA Activity CagA is delivered into gastric epithelial cells via the H. pylori type IV secretion system, where it acts as an oncogenic scaffold or hub. The CagA protein consists of structured N-terminal and disordered C-terminal regions. Folded N-terminal CagA displays a heretofore unidentified structure with three distinct domains (domains I–III). Domain I (blue) constitutes the N terminus, whereas domain II (yellow) tethers CagA to the inner plasma membrane via the basic patch (green)-phosphatidylserine (pink) interaction. Intramolecular interaction with domain III (red) potentiates binding capability of the disordered C-terminal tail with PAR1 and SHP2, thereby enhancing the oncogenic scaffold or hub function of CagA.

disordered C-terminal tail (30% of the entire CagA) that contains the EPIYA segments and the CM motifs (Figure 3) (Hayashi et al., 2012b; Kaplan-Tu¨rko¨z et al., 2012). The N-terminal core predicts a square plate-like shape with approximate dimensions of 110 3 80 3 55 A˚3 comprising three discrete domains, termed domains I–III. Domain I, the most N-terminal domain, has a small interacting surface area with domain II but not domain III. Because of this, domain I is quite mobile and flexible. Domains II and III comprise a protease-resistant structural CagA core. Domain II also contains a large antiparallel b sheet, with which CagA binds to the host membrane b1-integrin for its delivery. The disordered C-terminal tail loops back onto domain III to form a lariat loop that strengthens interaction with PAR1 and SHP2. Therefore, the degree of the scaffold or hub function of CagA is fine-tuned by the intramolecular interaction (Figure 3). Delivered CagA is tethered to the inner face of the plasma membrane. The mechanism underlying CagA-host membrane interaction is rather complicated, given that it involves at least two distinct CagA regions, the basic patch and the EPIYA motif, in a cell-context-dependent manner (Higashi et al., 2005; Murata-Kamiya et al., 2010; Hayashi et al., 2012b). The basic patch is a cluster of basic residues on the surface of CagA domain II, which adheres like Velcro to phosphatidylserine (PS), an acidic phospholipid specifically concentrated to the inner leaflet of the plasma membrane. The EPIYA motif also contributes to the association of CagA with the membrane independently of phosphorylation. The CagA-PS interaction plays a major role in the membrane tethering of CagA in polarized epithelial cells, whereas the EPIYA motif is crucial for the membrane association of CagA in nonpolarized cells. Although the crystal structure of Gab proteins has yet to be determined, the structural architecture of CagA displays a remarkable similarity to the presumed structure of mammalian scaffold or hub proteins: the N-terminal solid domain is followed by a long, disordered C-terminal tail, which interacts with its target proteins in a tyrosine phosphorylation-dependent manner (Simister and Feller, 2012). The particular structural arrangement 310 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

may be a common topological architecture acting as a signal integration platform. Another surprising coincidence between CagA and Gab is that the disordered C-terminal tail can interact intramolecularly with the structured N-terminal region. This intramolecular interaction observed in CagA, which acts as a fine tuner of the protein function, could also apply to the functional regulation of mammalian scaffold or hub proteins. The structural diversity in the C-terminal tail of CagA may well be explained by the intrinsically disordered nature of the EPIYA segments, which can be arranged in variable numbers and orders through homologous recombination because there are minimal structural constraints that restrict the combinatorial diversification at the protein level. As proposed recently (Hayashi et al., 2013), the duplication and/or depletion of genetic segments encoding such disordered regions may drive rapid evolution of bacterial EPIYA effectors both quantitatively and qualitatively. Recent studies have shown that the structured N-terminal CagA interacts with tumor suppressors. Interaction of N-terminal CagA with RUNX3 induces the ubiquitination and degradation of RUNX3 that abolishes its ability to function as a tumor suppressive transcription factor (Tsang et al., 2010). N-terminal CagA also associates with another tumor suppressor, the apoptosisstimulating protein of p53-2 (ASPP2). In response to DNA damage, ASPP2 binds and activates p53 in order to induce apoptosis. However, CagA promotes the proteasomal degradation of p53 by forming a CagA-ASPP2-p53 heterotrimeric complex, thereby inhibiting p53-mediated apoptosis (Buti et al., 2011). Thus, in the host cells to which it is delivered, the structured N-terminal CagA primarily acts to subvert tumor suppressors, whereas the disordered C-terminal tail serves to deregulate pro-oncogenic signaling. CagA-Mediated Cell Reprogramming Pathological transdifferentiation of cells increases cellular sensitivity to oncogenic insults. Intestinal metaplasia of the stomach, which is associated with long-term H. pylori infection, is thought to be a precancerous change of the gastric mucosa (Correa, 1988). Upon ectopic expression in gastric epithelial cells, CagA deregulates Wnt signaling and thereby aberrantly induces Wnt targets, including CDX1, a master transcription factor

Cell Host & Microbe

Review regulating intestinal differentiation and maintenance (Murata-Kamiya et al., 2007). Then, ectopically expressed CDX1 transactivates stemness-associated reprogramming factors SALL4 and KLF5 that are involved in the conversion of the CagA-expressing gastric cells into LGR5-positive intestinal stem-like cells, which can transdifferentiate into those with various intestinal traits (Fujii et al., 2012). Because CDX1 binds to its own gene promoter and generates a positive feedback loop for its own expression, once established, the CDX1-SALL4/KLF5 transcriptional network no longer requires CagA. This may explain sustained intestinal metaplasia even long after eradication of H. pylori from the stomach. Considering the key biological properties shared by tissue stem cells and cancer stem cells (Mani et al., 2008), cells enforced to dedifferentiate by CagA may have acquired cancer stem-cell-like traits, making them excellent candidates from which gastric cancer arises. Such differentiation plasticity also suggests that the cognate gastric stem cell is not necessarily the obligate target of CagA in terms of neoplastic transformation. The EMT-like change induced by the expression of CagA in polarized epithelial cells may also reflect cell reprogramming (Amieva et al., 2003; Saito et al., 2010). Those CagA-expressing cells display a fibroblast-like morphology, exhibiting elevated motility, with upregulated expression of mesenchymal markers such as ZEB1 and Snail (Bagnoli et al., 2005, Saito et al., 2010). They also exhibit high levels of the CD44 gastric cancer stem cell marker (Besse`de et al., 2013), reinforcing the reprogramming potential of CagA. Delivered CagA undergoes autophagy-dependent degradation, which is triggered by reactive oxygen species (ROS). However, CagA can escape from the ROS-induced autophagy in gastric cancer stem cells expressing a variant 9 form of CD44 (CD44v9) because the variant CD44 increases the level of intracellular glutathione that neutralizes ROS. Consequently, the oncogenic action of CagA lasts substantially longer in CD44v9-positive cells (Tsugawa et al., 2012). Therefore, CagA may play a special role in generating and/or maintaining gastric cancer stem cells. Hit-and-Run Carcinogenesis by CagA Despite its critical role in the development of gastric cancer, established gastric cancer cells maintain their transformed phenotypes in the absence of CagA. This leads to the idea that CagA-mediated gastric carcinogenesis progresses through a hit-and-run mechanism. Hit-and-run carcinogenesis was originally proposed by Skinner (1976) to explain certain types of virus-induced cancers. The conventional concept of viral carcinogenesis requires the presence of virus-specific oncogenes to maintain a malignant phenotype. In contrast, according to the hit-and-run model, a virus can initiate cell transformation while it (or its product) is no longer present in transformed cells. The model has been implicated in human neoplasms such as polyomavirus in brain tumors and hepatitis viruses in hepatocellular carcinoma and has been supported by results of a number of studies, including animal studies in which the expression of a driver oncogene was conditionally turned on and off (Niller et al., 2011; Furth, 2012). CagA-mediated hit-and-run carcinogenesis postulates multiple genomic and epigenetic alterations in the established cancer cells that ensure their neoplastic phenotype independent of CagA (Figure 4). To do so, CagA may need to create some sort of

mechanisms that accelerate genetic and epigenetic changes in host cells. In fact, this bacterial oncoprotein induces CIN by inhibiting PAR1 (Umeda et al., 2009). Aberrant activation of nuclear factor kB (NF-kB) upon infection with cagA-positive H. pylori induces ectopic expression of activation-induced cytidine deaminase (AID) in gastric epithelial cells, which confers a nucleotide mutator phenotype (Marusawa and Chiba, 2010). Additionally, the sustained activation of NF-kB in H. pyloriinfected stomach mucosa increases the level of DNA methyltransferases (DNMTs) (Huang et al., 2012), which mediate the hypermethylation of a large number of genes, including those involved in DNA mismatch repair, such as MLH1. Reduced MLH1 expression may underlie microsatellite instability observed in a fraction of gastric cancer cases, which further deteriorates genetic instability (Leung et al., 1999). These mutator phenotypes induced by chronic infection with cagA-positive H. pylori should cooperate with the direct oncogenic action of CagA in proceeding hit-and-run carcinogenesis (for a more detailed review discussing the impact of pathogens on mechanisms that maintain host genome integrity, please see the review by Weitzman and Weitzman [2014] in this issue of Cell Host & Microbe). Consistent with the elevated genetic instability as a background of hit-and-run carcinogenesis, mutations found in gastric cancers are extremely divergent and have no characteristic driver gene mutations (Figure 4). Having said this, genes involved in the Ras-Erk signaling pathway, such as FGFR2, KRAS, EGFR, ERBB2, and MET, are amplified in a mutually exclusive manner in approximately 40% of gastric cancer cases (Deng et al., 2012). The absence of mutations in the PTPN11 gene encoding SHP2 in gastric cancer may be due to the presence of the CagASHP2 complex, which dampens positive selection of cells harboring mutant PTPN11. Remarkably, KRAS mutation, the most common oncogene mutation in solid cancers, is extremely rare in gastric cancer. This raises the possibility that quantitative alteration, rather than qualitative change, in KRAS activity is a favorite choice in complementing CagA-mediated deregulation of Ras signaling. Nuclear accumulation of b-catenin is observed in approximately 30% of gastric cancers (Clements et al., 2002). In contrast to colorectal cancers, in which the APC tumor suppressor gene is most commonly mutated, activating the mutation of the CTNNB1 gene, which encodes b-catenin, is the major mechanism that deregulates the Wnt pathway in gastric cancers. Again, in this instance, CagA-mediated Wnt deregulation could be more preferentially taken over by CTNNB1 mutation than APC mutation. Germline mutation in the CDH1 gene encoding E-cadherin causes hereditary diffuse gastric cancer, which comprises approximately 1%–3% of all gastric cancers (Humar and Guilford, 2009). In sporadic cases, CDH1 mutation is found in more than half of diffuse-type cancers, whereas it is rare in intestinal-type cancers. Promoter hypermethylation of CDH1 is also frequently observed in diffuse-type gastric cancers (Yamashita et al., 2011). Thus, sustained malfunctioning of E-cadherin, which is initiated at the early stage by CagA and later by the mutation or hypermethylation of CDH1, might be a molecular precondition for the development of diffuse-type gastric cancer. The incidence of p53 mutations in gastric cancer is estimated to be approximately 30% on the basis of the 2008 International Agency for Research on Cancer TP53 Database (Olivier et al., Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc. 311

Cell Host & Microbe

Review

Figure 4. Hit-and-Run Model of CagA-Mediated Gastric Carcinogenesis Direct pro-oncogenic actions of CagA include (1) aberrant activation of promitogenic signaling pathways, (2) disruption of cell-cell interaction and loss of cell polarity, and (3) inactivation of tumor suppressors such as p53 and RUNX3. In conjunction with chronic inflammation, CagA also induces genetic and epigenetic alterations in the host cells, which, in the long run, compensate pro-oncogenic actions of CagA via secondary alterations, thereby making cells independent of CagA for neoplastic transformation. Loss-of-function mutation of the FAT tumor suppressor homolog 4 (FAT4) gene or the AT-rich interactive domain-containing protein 1A (ARID1A) gene observed in 10% of gastric cancer genomes (Zang et al., 2012; Wu and Roberts, 2013) may additionally contribute to gastric carcinogenesis by perturbing tumor suppressor pathways such as pRB or Hippo.

2010), and G:C-to-A:T transition is the most common type of p53 mutation regardless of histological type. The mutational signature is explained by (1) dietary mutagens associated with the metabolism of nitrogenous compounds, (2) nitric oxide, which is produced by H. pylori infection, and (3) aberrant expression of AID in gastric mucosa infected with cagA-positive H. pylori. Interestingly, p53 mutations were found in more than 50% of H. pylori-associated gastritis cases (Murakami et al., 1999). Therefore, the relatively low incidence of p53 mutations in established gastric cancers (30%) suggests that inactivating mutation of p53 is not necessarily a requisite for CagA-driven gastric carcinogenesis. Enhanced degradation of p53 by the CagAASPP2 interaction possibly compensates for p53 mutations during the early phase of gastric carcinogenesis. In such cases, malignant progression at the later phase may no longer require p53 inactivation because other mutations are functionally complementary. Epigenetic Alterations that Support Hit-and-Run Carcinogenesis Epigenetic modifications such as DNA methylation and histone modification influence gene expression without changing genomic sequences. Gastric cancer displays a high frequency of aberrant methylation, especially in CpG islands (CpG island methylator phenotype [CIMP]) (Jang and Kim, 2011). This is at least partly due to chronic inflammation in gastric mucosa infected with H. pylori, given that prolonged NF-kB activation in the in312 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

flamed mucosa could upregulate DNMTs (O’Gorman et al., 2010). A large number of genes are repressed through hypermethylation in gastric cancer cells. These include tumor suppressor genes such as p16, p14, E-cadherin, MGMT, RASSF1, RUNX3, and DLC1. Notably, the expression of RUNX3 is lost in about half of gastric cancers by hemizygous deletion and promoter hypermethylation (Chuang and Ito, 2010). Therefore, the RUNX3 inactivation may be attributable to CIN and CIMP phenotypes in cells that have been exposed to CagA for a long time. Although many genes are aberrantly hypermethylated, the hypomethylation of certain genes such as the Notch ligand Delta-like 1 has also been demonstrated in intestinal-type gastric cancer and in the stomach of INS-GAS mice (mice bearing a chimeric insulin-gastrin transgene) infected with H. pylori, which recapitulates intestinal-type gastric carcinogenesis (Piazzi et al., 2011). Notch signaling plays a critical role in tissue stem cell maintenance, and its perturbation is associated with the development of cancer. Although the mechanism underlying the hypomethylation of particular genes in gastric cancer cells is currently unknown, it is speculated that the activation of DNA demethylase via a protein-kinase C (PKC)-dependent pathway, which could be stimulated by the gastric hormone gastrin, is involved in the process (Tomita et al., 2011). Role of Inflammation in CagA-Mediated Carcinogenesis Much attention has recently focused on the role of inflammation in fostering carcinogenesis. Since no obvious inflammatory

Cell Host & Microbe

Review Figure 5. Directional Switching of CagA Signaling (A) In early life, injection of CagA into normal gastric epithelial cells elicits oncogenic stress, which induces premature cell senescence or apoptosis. As a result, a reduced number of acidsecreting cells and subsequent increase in gastric pH provide direct merit for H. pylori colonization. By inducing junctional and polarity defects, CagA may also cause chronic gastritis and/or peptic ulceration because of structural disintegration of gastric mucosa. (B) Chronic inflammation associated with longterm cagA-positive H. pylori infection causes genetic and epigenetic instability in the infected stomach mucosa, which markedly enhances the ability of cells to acquire resistance to oncogenic stress by secondary adaptation. Delivery of CagA into such abnormal cells switches the direction of the CagA signal from senescence or apoptosis to unconstraint proliferation, thereby expanding cancer precursor cells in elderly stomach mucosa.

response was associated with neoplastic lesions that developed in cagA-transgenic mice, the oncogenic potential of CagA on its own appears to be cell autonomous (Ohnishi et al., 2008). Nonetheless, there is no doubt about the role of chronic inflammation as an auxiliary engine of gastric carcinogenesis. NF-kB, a transcription factor that dictates inflammation, is strongly activated in response to H. pylori infection. Although its contribution to H. pylori-induced NF-kB activation remains uncertain, CagA can activate NF-kB through multiple distinct pathways, including Ras, PI3K-AKT, and TRAF6-TAK1, in host cells to which it has been delivered (Backert and Naumann, 2010). CagA has also been reported to activate STAT3, another key transcriptional factor that promotes inflammation. Although detailed mechanisms remain to be elucidated, CagA has been shown to deregulate STAT3 via the IL-6/gp130 signaling pathway in a phosphorylation-independent manner (Lee et al., 2010). Therefore, CagA phosphorylation status may determine its downstream signaling pathway, either SHP2-Erk-NF-kB or IL6-JAK-STAT signaling, that leads to the activation of transcription factors during the inflammatory process (Lee et al., 2010). Then, activated NF-kB and/or STAT3 induce the production of proinflammatory cytokines and antiapoptotic proteins that support the expansion of cancer-predisposed cells while preventing apoptosis. The inflammatory responses also induce ROS, which increases DNA damage and thereby accelerates the accumulation of mutations. Additionally, cells in the chronic inflammatory milieu undergo accelerated epigenetic alterations, which are characteristic of CIMP through elevated DNMTs, as described above. An important question is whether ‘‘simple’’ but ‘‘prolonged’’ inflammation in the stomach mucosa is per se sufficient to induce gastric cancer. A recent report on CLDN18-null mice may provide some clues to this question. Claudin-18, encoded by CLDN18, is the major Claudin component of tight junctions in the stomach mucosa. CLDN18-null mice, although viable, showed impaired tight junctions and developed chronic atrophic gastritis and precancerous metaplasia (Hayashi et al., 2012a).

Nonetheless, CLDN18-null mice developed neither dysplasia nor carcinoma. The results of that study indicate that chronic inflammation simply induced by physicochemical stimuli cannot fully support multistep gastric carcinogenesis and suggest the requirement of an additional factor, such as CagA, that can act as an oncogenic driver during the process of neoplastic transformation. Conclusions and Perspectives In most cases, H. pylori infection is established during childhood by oral transmission. In contrast, the majority of gastric cancer patients are over 50 years old. Considering that the generation time of H. pylori is at most several days, killing the host due to gastric cancer is not likely the intended outcome of this gastric pathogen. Why then do virulent H. pylori strains carry CagA? Once delivered, CagA immediately triggers an oncogenic signal. However, normal epithelial cells that have received such an oncogenic insult prevent neoplastic transformation by triggering premature senescence or apoptosis (oncogenic stress response) due, in many cases, to activation of the ARF/p53/ p21 or p16-pRB pathway (Sherr and Weber, 2000; Yaswen and Campisi, 2007). Thus, CagA injection gives rise to a decrease in the number of epithelial cells, which is concomitantly associated with reduced acid secretion (hypochloridia). Although H. pylori is able to survive under highly acidic conditions, an elevated pH would be an obvious advantage for the long-term colonization of H. pylori in the stomach (Figure 5A). However, excessive loss of acid-secreting cells facilitates the intrusion of oral or intestinal bacteria into the stomach due to the neutralized pH. Accordingly, ‘‘inefficient’’ and ‘‘timeconsuming’’ killing of host cells by CagA injection, which causes a local and moderate increase in pH at the site of H. pylori infection, would be a deliberate strategy that makes the stomach a ‘‘great sanctuary’’ for H. pylori. Such a change in gastric pH also influences the composition of the gastric microbiota in addition to H. pylori that may additionally contribute to the development of mucosal lesions (Sheh and Fox, 2013). Ironically, CagA Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc. 313

Cell Host & Microbe

Review exploits oncogenic stress in order to achieve its mission. In the long run, continuous exposure of epithelial cells to such a bacterial effector that constantly exhibits pro-oncogenic activity should greatly elevate the chance of cells to overcome the oncogenic stress triggered by CagA through secondary adaptations. According to this scenario, gastric cancer is a contingent byproduct of long-term H. pylori infection that has become evident with increased human longevity. Until modern civilization, gastric cancer must have been quite rare, given that ancient people mostly died at much younger ages. The great success of H. pylori infection in humans may even suggest that H. pylori infection poses some advantage to infected individuals (Cover and Blaser, 2009). As long as host cells respond to CagA by triggering premature senescence or apoptosis, the chance for developing gastric cancer is low. On the other hand, failure to trigger senescence or apoptosis may convert the cellular response to CagA from an anti- to a pro-oncogenic direction. In the gastric mucosa that has long been infected with cagApositive H. pylori, CagA-induced genetic instability may promote the generation of abnormal cells that are defective in cell machineries that mediate premature senescence or apoptosis. Delivery of CagA into such abnormal cells leads to its clonal expansion (Figure 5B). The dual role of CagA, direct priming of pro-oncogenic signaling and indirect generation of genomic instability, ensures the progression of the multistep cellular process underlying neoplastic transformation. Therefore, it may be inevitable that, with time, cancer precursor cells will acquire CagA independence, making CagA-initiated gastric carcinogenesis an excellent example of hit-and-run carcinogenesis. ACKNOWLEDGMENTS I thank Dr. Takeru Hayashi for preparation of figures. This work was supported by a Grant-in-Aid for Scientific Research on Innovative Area from the Ministry of Education, Culture, Sports, and Technology (MEXT) of Japan.

Buti, L., Spooner, E., Van der Veen, A.G., Rappuoli, R., Covacci, A., and Ploegh, H.L. (2011). Helicobacter pylori cytotoxin-associated gene A (CagA) subverts the apoptosis-stimulating protein of p53 (ASPP2) tumor suppressor pathway of the host. Proc. Natl. Acad. Sci. USA 108, 9238–9243. Chan, R.J., and Feng, G.S. (2007). PTPN11 is the first identified proto-oncogene that encodes a tyrosine phosphatase. Blood 109, 862–867. Chuang, L.S., and Ito, Y. (2010). RUNX3 is multifunctional in carcinogenesis of multiple solid tumors. Oncogene 29, 2605–2615. Clements, W.M., Wang, J., Sarnaik, A., Kim, O.J., MacDonald, J., FenoglioPreiser, C., Groden, J., and Lowy, A.M. (2002). b-Catenin mutation is a frequent cause of Wnt pathway activation in gastric cancer. Cancer Res. 62, 3503–3506. Correa, P. (1988). A human model of gastric carcinogenesis. Cancer Res. 48, 3554–3560. Cover, T.L., and Blaser, M.J. (2009). Helicobacter pylori in health and disease. Gastroenterology 136, 1863–1873. Deng, N., Goh, L.K., Wang, H., Das, K., Tao, J., Tan, I.B., Zhang, S., Lee, M., Wu, J., Lim, K.H., et al. (2012). A comprehensive survey of genomic alterations in gastric cancer reveals systematic patterns of molecular exclusivity and co-occurrence among distinct therapeutic targets. Gut 61, 673–684. Ekstro¨m, A.M., Held, M., Hansson, L.-E., Engstrand, L., and Nyre´n, O. (2001). Helicobacter pylori in gastric cancer established by CagA immunoblot as a marker of past infection. Gastroenterology 121, 784–791. Ferreira, R.M., Machado, J.C., Leite, M., Carneiro, F., and Figueiredo, C. (2012). The number of Helicobacter pylori CagA EPIYA C tyrosine phosphorylation motifs influences the pattern of gastritis and the development of gastric carcinoma. Histopathology 60, 992–998. Fischer, W. (2011). Assembly and molecular mode of action of the Helicobacter pylori Cag type IV secretion apparatus. FEBS J. 278, 1203–1212. Franco, A.T., Israel, D.A., Washington, M.K., Krishna, U., Fox, J.G., Rogers, A.B., Neish, A.S., Collier-Hyams, L., Perez-Perez, G.I., Hatakeyama, M., et al. (2005). Activation of b-catenin by carcinogenic Helicobacter pylori. Proc. Natl. Acad. Sci. USA 102, 10646–10651. Fujii, Y., Yoshihashi, K., Suzuki, H., Tsutsumi, S., Mutoh, H., Maeda, S., Yamagata, Y., Seto, Y., Aburatani, H., and Hatakeyama, M. (2012). CDX1 confers intestinal phenotype on gastric epithelial cells via induction of stemness-associated reprogramming factors SALL4 and KLF5. Proc. Natl. Acad. Sci. USA 109, 20584–20589.

REFERENCES Amieva, M.R., Vogelmann, R., Covacci, A., Tompkins, L.S., Nelson, W.J., and Falkow, S. (2003). Disruption of the epithelial apical-junctional complex by Helicobacter pylori CagA. Science 300, 1430–1434. Backert, S., and Naumann, M. (2010). What a disorder: proinflammatory signaling pathways induced by Helicobacter pylori. Trends Microbiol. 18, 479–486.

Furth, P.A. (2012). Cancer prevention as biomodulation: targeting the initiating stimulus and secondary adaptations. Ann. N Y Acad. Sci. 1271, 1–9. Hatakeyama, M. (2003). Helicobacter pylori CagA—a potential bacterial oncoprotein that functionally mimics the mammalian Gab family of adaptor proteins. Microbes Infect. 5, 143–150. Hatakeyama, M. (2004). Oncogenic mechanisms of the Helicobacter pylori CagA protein. Nat. Rev. Cancer 4, 688–694.

Bagnoli, F., Buti, L., Tompkins, L., Covacci, A., and Amieva, M.R. (2005). Helicobacter pylori CagA induces a transition from polarized to invasive phenotypes in MDCK cells. Proc. Natl. Acad. Sci. USA 102, 16339–16344.

Hatakeyama, M. (2011). Anthropological and clinical implications for the structural diversity of the Helicobacter pylori CagA oncoprotein. Cancer Sci. 102, 36–43.

Basso, D., Zambon, C.F., Letley, D.P., Stranges, A., Marchet, A., Rhead, J.L., Schiavon, S., Guariso, G., Ceroti, M., Nitti, D., et al. (2008). Clinical relevance of Helicobacter pylori cagA and vacA gene polymorphisms. Gastroenterology 135, 91–99.

Hayashi, D., Tamura, A., Tanaka, H., Yamazaki, Y., Watanabe, S., Suzuki, K., Suzuki, K., Sentani, K., Yasui, W., Rakugi, H., et al. (2012a). Deficiency of claudin-18 causes paracellular H+ leakage, up-regulation of interleukin-1b, and atrophic gastritis in mice. Gastroenterology 142, 292–304.

Bauer, B., Pang, E., Holland, C., Kessler, M., Bartfeld, S., and Meyer, T.F. (2012). The Helicobacter pylori virulence effector CagA abrogates human b-defensin 3 expression via inactivation of EGFR signaling. Cell Host Microbe 11, 576–586.

Hayashi, T., Senda, M., Morohashi, H., Higashi, H., Horio, M., Kashiba, Y., Nagase, L., Sasaya, D., Shimizu, T., Venugopalan, N., et al. (2012b). Tertiary structure-function analysis reveals the pathogenic signaling potentiation mechanism of Helicobacter pylori oncogenic effector CagA. Cell Host Microbe 12, 20–33.

Besse`de, E., Staedel, C., Acun˜a Amador, L.A., Nguyen, P.H., Chambonnier, L., Hatakeyama, M., Belleanne´e, G., Me´graud, F., and Varon, C. (2013). Helicobacter pylori generates cells with cancer stem cell properties via epithelialmesenchymal transition-like changes. Oncogene. http://dx.doi.org/10.1038/ onc.2013.380. Botham, C.M., Wandler, A.M., and Guillemin, K. (2008). A transgenic Drosophila model demonstrates that the Helicobacter pylori CagA protein functions as a eukaryotic Gab adaptor. PLoS Pathog. 4, e1000064.

314 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

Hayashi, T., Morohashi, H., and Hatakeyama, M. (2013). Bacterial EPIYA effectors—where do they come from? What are they? Where are they going? Cell. Microbiol. 15, 377–385. Higashi, H., Tsutsumi, R., Muto, S., Sugiyama, T., Azuma, T., Asaka, M., and Hatakeyama, M. (2002). SHP-2 tyrosine phosphatase as an intracellular target of Helicobacter pylori CagA protein. Science 295, 683–686.

Cell Host & Microbe

Review Higashi, H., Yokoyama, K., Fujii, Y., Ren, S., Yuasa, H., Saadat, I., MurataKamiya, N., Azuma, T., and Hatakeyama, M. (2005). EPIYA motif is a membrane-targeting signal of Helicobacter pylori virulence factor CagA in mammalian cells. J. Biol. Chem. 280, 23130–23137. Huang, F.Y., Chan, A.O., Rashid, A., Wong, D.K., Cho, C.H., and Yuen, M.F. (2012). Helicobacter pylori induces promoter methylation of E-cadherin via interleukin-1b activation of nitric oxide production in gastric cancer cells. Cancer 118, 4969–4980. Humar, B., and Guilford, P. (2009). Hereditary diffuse gastric cancer: a manifestation of lost cell polarity. Cancer Sci. 100, 1151–1157. Jang, B.G., and Kim, W.H. (2011). Molecular pathology of gastric carcinoma. Pathobiology 78, 302–310. Kaplan-Tu¨rko¨z, B., Jime´nez-Soto, L.F., Dian, C., Ertl, C., Remaut, H., Louche, A., Tosi, T., Haas, R., and Terradot, L. (2012). Structural insights into Helicobacter pylori oncoprotein CagA interaction with b1 integrin. Proc. Natl. Acad. Sci. USA 109, 14640–14645.

Nagase, L., Murata-Kamiya, N., and Hatakeyama, M. (2011). Potentiation of Helicobacter pylori CagA protein virulence through homodimerization. J. Biol. Chem. 286, 33622–33631. Neal, J.T., Peterson, T.S., Kent, M.L., and Guillemin, K. (2013). H. pylori virulence factor CagA increases intestinal cell proliferation by Wnt pathway activation in a transgenic zebrafish model. Dis. Model. Mech. 6, 802–810. , D., Miller, M.C., Quinkert, Z.T., Stein, M., Chait, B.T., and Stebbins, C.E. Nesic (2010). Helicobacter pylori CagA inhibits PAR1-MARK family kinases by mimicking host substrates. Nat. Struct. Mol. Biol. 17, 130–132. Niller, H.H., Wolf, H., and Minarovits, J. (2011). Viral hit and run-oncogenesis: genetic and epigenetic scenarios. Cancer Lett. 305, 200–217. O’Gorman, A., Colleran, A., Ryan, A., Mann, J., and Egan, L.J. (2010). Regulation of NF-kappaB responses by epigenetic suppression of IkappaBalpha expression in HCT116 intestinal epithelial cells. Am. J. Physiol. Gastrointest. Liver Physiol. 299, G96–G105.

Lee, D.C., and Jia, Z. (2009). Emerging structural insights into bacterial tyrosine kinases. Trends Biochem. Sci. 34, 351–357.

Ohnishi, N., Yuasa, H., Tanaka, S., Sawa, H., Miura, M., Matsui, A., Higashi, H., Musashi, M., Iwabuchi, K., Suzuki, M., et al. (2008). Transgenic expression of Helicobacter pylori CagA induces gastrointestinal and hematopoietic neoplasms in mouse. Proc. Natl. Acad. Sci. USA 105, 1003–1008.

Lee, I.O., Kim, J.H., Choi, Y.J., Pillinger, M.H., Kim, S.Y., Blaser, M.J., and Lee, Y.C. (2010). Helicobacter pylori CagA phosphorylation status determines the gp130-activated SHP2/ERK and JAK/STAT signal transduction pathways in gastric epithelial cells. J. Biol. Chem. 285, 16042–16050.

Olivier, M., Hollstein, M., and Hainaut, P. (2010). TP53 mutations in human cancers: origins, consequences, and clinical use. Cold Spring Harb. Perspect. Biol. 2, a001008.

Lee, K.S., Kalantzis, A., Jackson, C.B., O’Connor, L., Murata-Kamiya, N., Hatakeyama, M., Judd, L.M., Giraud, A.S., and Menheniott, T.R. (2012). Helicobacter pylori CagA triggers expression of the bactericidal lectin REG3g via gastric STAT3 activation. PLoS ONE 7, e30786. Leung, S.Y., Yuen, S.T., Chung, L.P., Chu, K.M., Chan, A.S., and Ho, J.C. (1999). hMLH1 promoter methylation and lack of hMLH1 expression in sporadic gastric carcinomas with high-frequency microsatellite instability. Cancer Res. 59, 159–164. Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks, M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelialmesenchymal transition generates cells with properties of stem cells. Cell 133, 704–715. Marusawa, H., and Chiba, T. (2010). Helicobacter pylori-induced activationinduced cytidine deaminase expression and carcinogenesis. Curr. Opin. Immunol. 22, 442–447. Matozaki, T., Murata, Y., Saito, Y., Okazawa, H., and Ohnishi, H. (2009). Protein tyrosine phosphatase SHP-2: a proto-oncogene product that promotes Ras activation. Cancer Sci. 100, 1786–1793. Mimuro, H., Suzuki, T., Tanaka, J., Asahi, M., Haas, R., and Sasakawa, C. (2002). Grb2 is a key mediator of helicobacter pylori CagA protein activities. Mol. Cell 10, 745–755. Miura, M., Ohnishi, N., Tanaka, S., Yanagiya, K., and Hatakeyama, M. (2009). Differential oncogenic potential of geographically distinct Helicobacter pylori CagA isoforms in mice. Int. J. Cancer 125, 2497–2504. Mueller, D., Tegtmeyer, N., Brandt, S., Yamaoka, Y., De Poire, E., Sgouras, D., Wessler, S., Torres, J., Smolka, A., and Backert, S. (2012). c-Src and c-Abl kinases control hierarchic phosphorylation and function of the CagA effector protein in Western and East Asian Helicobacter pylori strains. J. Clin. Invest. 122, 1553–1566. Murakami, K., Fujioka, T., Okimoto, T., Mitsuishi, Y., Oda, T., Nishizono, A., and Nasu, M. (1999). Analysis of p53 gene mutations in Helicobacter pyloriassociated gastritis mucosa in endoscopic biopsy specimens. Scand. J. Gastroenterol. 34, 474–477. Murata-Kamiya, N., Kurashima, Y., Teishikata, Y., Yamahashi, Y., Saito, Y., Higashi, H., Aburatani, H., Akiyama, T., Peek, R.M., Jr., Azuma, T., and Hatakeyama, M. (2007). Helicobacter pylori CagA interacts with E-cadherin and deregulates the b-catenin signal that promotes intestinal transdifferentiation in gastric epithelial cells. Oncogene 26, 4617–4626. Murata-Kamiya, N., Kikuchi, K., Hayashi, T., Higashi, H., and Hatakeyama, M. (2010). Helicobacter pylori exploits host membrane phosphatidylserine for delivery, localization, and pathophysiological action of the CagA oncoprotein. Cell Host Microbe 7, 399–411.

Parkin, D.M. (2004). International variation. Oncogene 23, 6329–6340. Peek, R.M., Jr., and Blaser, M.J. (2002). Helicobacter pylori and gastrointestinal tract adenocarcinomas. Nat. Rev. Cancer 2, 28–37. Piazzi, G., Fini, L., Selgrad, M., Garcia, M., Daoud, Y., Wex, T., Malfertheiner, P., Gasbarrini, A., Romano, M., Meyer, R.L., et al. (2011). Epigenetic regulation of Delta-Like1 controls Notch1 activation in gastric cancer. Oncotarget 2, 1291–1301. Ren, S., Higashi, H., Lu, H., Azuma, T., and Hatakeyama, M. (2006). Structural basis and functional consequence of Helicobacter pylori CagA multimerization in cells. J. Biol. Chem. 281, 32344–32352. Rieder, G., Merchant, J.L., and Haas, R. (2005). Helicobacter pylori cag-type IV secretion system facilitates corpus colonization to induce precancerous conditions in Mongolian gerbils. Gastroenterology 128, 1229–1242. Saadat, I., Higashi, H., Obuse, C., Umeda, M., Murata-Kamiya, N., Saito, Y., Lu, H., Ohnishi, N., Azuma, T., Suzuki, A., et al. (2007). Helicobacter pylori CagA targets PAR1/MARK kinase to disrupt epithelial cell polarity. Nature 447, 330–333. Safari, F., Murata-Kamiya, N., Saito, Y., and Hatakeyama, M. (2011). Mammalian Pragmin regulates Src family kinases via the Glu-Pro-Ile-Tyr-Ala (EPIYA) motif that is exploited by bacterial effectors. Proc. Natl. Acad. Sci. USA 108, 14938–14943. Saito, Y., Murata-Kamiya, N., Hirayama, T., Ohba, Y., and Hatakeyama, M. (2010). Conversion of Helicobacter pylori CagA from senescence inducer to oncogenic driver through polarity-dependent regulation of p21. J. Exp. Med. 207, 2157–2174. Saito-Diaz, K., Chen, T.W., Wang, X., Thorne, C.A., Wallace, H.A., PageMcCaw, A., and Lee, E. (2013). The way Wnt works: components and mechanism. Growth Factors 31, 1–31. Satomi, S., Yamakawa, A., Matsunaga, S., Masaki, R., Inagaki, T., Okuda, T., Suto, H., Ito, Y., Yamazaki, Y., Kuriyama, M., et al. (2006). Relationship between the diversity of the cagA gene of Helicobacter pylori and gastric cancer in Okinawa, Japan. J. Gastroenterol. 41, 668–673. Segal, E.D., Cha, J., Lo, J., Falkow, S., and Tompkins, L.S. (1999). Altered states: involvement of phosphorylated CagA in the induction of host cellular growth changes by Helicobacter pylori. Proc. Natl. Acad. Sci. USA 96, 14559–14564. Selbach, M., Paul, F.E., Brandt, S., Guye, P., Daumke, O., Backert, S., Dehio, C., and Mann, M. (2009). Host cell interactome of tyrosine-phosphorylated bacterial proteins. Cell Host Microbe 5, 397–403. Shanks, A.M., and El-Omar, E.M. (2009). Helicobacter pylori infection, host genetics and gastric cancer. J. Dig. Dis. 10, 157–164.

Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc. 315

Cell Host & Microbe

Review Sheh, A., and Fox, J.G. (2013). The role of the gastrointestinal microbiome in Helicobacter pylori pathogenesis. Gut Microbes 4, 505–531. Sherr, C.J., and Weber, J.D. (2000). The ARF/p53 pathway. Curr. Opin. Genet. Dev. 10, 94–99. Simister, P.C., and Feller, S.M. (2012). Order and disorder in large multi-site docking proteins of the Gab family—implications for signalling complex formation and inhibitor design strategies. Mol. Biosyst. 8, 33–46. Skinner, G.R. (1976). Transformation of primary hamster embryo fibroblasts by type 2 simplex virus: evidence for a ‘‘hit and run’’ mechanism. Br. J. Exp. Pathol. 57, 361–376. Suzuki, A., and Ohno, S. (2006). The PAR-aPKC system: lessons in polarity. J. Cell Sci. 119, 979–987. Suzuki, M., Mimuro, H., Suzuki, T., Park, M., Yamamoto, T., and Sasakawa, C. (2005). Interaction of CagA with Crk plays an important role in Helicobacter pylori-induced loss of gastric epithelial cell adhesion. J. Exp. Med. 202, 1235–1247. Suzuki, M., Mimuro, H., Kiga, K., Fukumatsu, M., Ishijima, N., Morikawa, H., Nagai, S., Koyasu, S., Gilman, R.H., Kersulyte, D., et al. (2009). Helicobacter pylori CagA phosphorylation-independent function in epithelial proliferation and inflammation. Cell Host Microbe 5, 23–34. Takahashi, A., Tsutsumi, R., Kikuchi, I., Obuse, C., Saito, Y., Seidi, A., Karisch, R., Fernandez, M., Cho, T., Ohnishi, N., et al. (2011). SHP2 tyrosine phosphatase converts parafibromin/Cdc73 from a tumor suppressor to an oncogenic driver. Mol. Cell 43, 45–56. Tan, S., Tompkins, L.S., and Amieva, M.R. (2009). Helicobacter pylori usurps cell polarity to turn the cell surface into a replicative niche. PLoS Pathog. 5, e1000407. Tan, S., Noto, J.M., Romero-Gallo, J., Peek, R.M., Jr., and Amieva, M.R. (2011). Helicobacter pylori perturbs iron trafficking in the epithelium to grow on the cell surface. PLoS Pathog. 7, e1002050. Tanaka, H., Katoh, H., and Negishi, M. (2006). Pragmin, a novel effector of Rnd2 GTPase, stimulates RhoA activity. J. Biol. Chem. 281, 10355–10364. Tegtmeyer, N., and Backert, S. (2011). Role of Abl and Src family kinases in actin-cytoskeletal rearrangements induced by the Helicobacter pylori CagA protein. Eur. J. Cell Biol. 90, 880–890. Timm, T., Marx, A., Panneerselvam, S., Mandelkow, E., and Mandelkow, E.M. (2008). Structure and regulation of MARK, a kinase involved in abnormal phosphorylation of Tau protein. BMC Neurosci. 9 (Suppl 2 ), S9.

316 Cell Host & Microbe 15, March 12, 2014 ª2014 Elsevier Inc.

Tomita, H., Takaishi, S., Menheniott, T.R., Yang, X., Shibata, W., Jin, G., Betz, K.S., Kawakami, K., Minamoto, T., Tomasetto, C., et al. (2011). Inhibition of gastric carcinogenesis by the hormone gastrin is mediated by suppression of TFF1 epigenetic silencing. Gastroenterology 140, 879–891. Tsang, Y.H., Lamb, A., Romero-Gallo, J., Huang, B., Ito, K., Peek, R.M., Jr., Ito, Y., and Chen, L.F. (2010). Helicobacter pylori CagA targets gastric tumor suppressor RUNX3 for proteasome-mediated degradation. Oncogene 29, 5643–5650. Tsugawa, H., Suzuki, H., Saya, H., Hatakeyama, M., Hirayama, T., Hirata, K., Nagano, O., Matsuzaki, J., and Hibi, T. (2012). Reactive oxygen speciesinduced autophagic degradation of Helicobacter pylori CagA is specifically suppressed in cancer stem-like cells. Cell Host Microbe 12, 764–777. Tsutsumi, R., Takahashi, A., Azuma, T., Higashi, H., and Hatakeyama, M. (2006). Focal adhesion kinase is a substrate and downstream effector of SHP-2 complexed with Helicobacter pylori CagA. Mol. Cell. Biol. 26, 261–276. Tsutsumi, R., Masoudi, M., Takahashi, A., Fujii, Y., Hayashi, T., Kikuchi, I., Satou, Y., Taira, M., and Hatakeyama, M. (2013). YAP and TAZ, Hippo signaling targets, act as a rheostat for nuclear SHP2 function. Dev. Cell 26, 658–665. Umeda, M., Murata-Kamiya, N., Saito, Y., Ohba, Y., Takahashi, M., and Hatakeyama, M. (2009). Helicobacter pylori CagA causes mitotic impairment and induces chromosomal instability. J. Biol. Chem. 284, 22166–22172. Weitzman, M.D., and Weitzman, J.B. (2014). What’s the damage? The impact of pathogens on pathways that maintain host genome integrity. Cell Host Microbe 15, this issue, 283–294. Wu, J.N., and Roberts, C.W. (2013). ARID1A mutations in cancer: another epigenetic tumor suppressor? Cancer Discov 3, 35–43. Yamaoka, Y. (2010). Mechanisms of disease: Helicobacter pylori virulence factors. Nat Rev Gastroenterol Hepatol 7, 629–641. Yamashita, K., Sakuramoto, S., and Watanabe, M. (2011). Genomic and epigenetic profiles of gastric cancer: potential diagnostic and therapeutic applications. Surg. Today 41, 24–38. Yaswen, P., and Campisi, J. (2007). Oncogene-induced senescence pathways weave an intricate tapestry. Cell 128, 233–234. Zang, Z.J., Cutcutache, I., Poon, S.L., Zhang, S.L., McPherson, J.R., Tao, J., Rajasegaran, V., Heng, H.L., Deng, N., Gan, A., et al. (2012). Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes. Nat. Genet. 44, 570–574.