Gastric cancer remains the 2nd leading cause of cancer mortality worldwide, with an overall 5-year survival rate that is less than 25% [1,2]. Critical to understanding the mechanisms involved in gastric cancer, and devising preventive and therapeutic interventions, is the need to develop an authentic animal model. Mice have a different gastric anatomy compared to humans. In mice the gastric fundus equivalent is lined by squamous, rather than oxyntic glandular epithelium. Thus, in mice, the squamocolumnar junction does not universally approximate the gastroesophageal junction as it does in normal human anatomy. In addition, rodents rarely develop spontaneous gastric cancer, although some reports have described spontaneous gastric adenocarcinomas in cotton rats (Sigmodon hispidus), and in the Z strain of the African rodent Mastomys natalensis [3,4,5,6,7]. However, these animals when they develop gastric tumors more frequently exhibit enterochromaffin-like cell carcinoids. Thus, experimental efforts have focused on identifying chemical, infectious or genetic means to induce gastric cancer in animals.

Prior to the discovery of Helicobacter pylori infection, major etiological factors were thought to be a diet rich in salt and nitrates/nitrites, along with a low intake of ascorbic acid and carotenoids [8], and thus early animal models utilized nitrosamines, such MNNG in rats [9,10] and MNU in mice to induce gastric tumors [11].

With the discovery of H. pylori by Marshall and Warren [12], and subsequently studies showing a strong association with gastric cancer [13,14,15,16,17], the focus shifted to the development of animal models of Helicobacter-associated gastric cancer. A variety of animals, including mice, rats, Mongolian gerbils, cats, guinea, pigs, ferrets, pigs, and macaques, have been experimentally infected with a variety of different Helicobacter species. Because of the ability to manipulate the mouse genome, mice have become the animal model of choice for cancer research. While the greatest interest has been in mouse models, only a limited number of H. pylori strains have been identified that successfully colonize the mouse stomach. The most robust and useful models to date have been H. pylori SS1-infected mice, and H. felis originally isolated from the stomach of cats and dogs [18]. Both of these gastric Helicobacters are capable of long-term colonization and have the ability to induce chronic gastritis and precancerous lesions in mice. However, the SS1 strain is not able to induce gastric cancer in most inbred strains of mice, while it does cause gastric carcinoma in C57BL/129 mice [19,20]. Chronic H. felis infection has been shown to induce severe inflammation, atrophy, metaplasia, dysplasia and gastric cancer in C57BL/6 mice [21].

While genetically engineered mouse models of cancer were developed in the 1980’s, transgenic models of gastric cancer were slow to emerge. Initial models included some containing a variety of oncogenes that were known to transform but had no known association with human gastric cancer, such as the SV40 T antigen, which binds to pRb and disrupts its function [22]. Human carcino-embryonic antigen (CEA) promoter/SV40 T antigen transgenic mice were reported to develop antral hyperplasia or gastric cancer [23,24]. Stomach-specific SV40 T antigen transgenic mice using H/K-ATPase-β subunit gene promoter [25] developed hyperplasia and abnormal cell distribution within the gastric glandular unit, and rarely developed dysplasia. Other early models included transgenic mice carrying the human adenovirus type 12 (Ad12) early region 1 under control of the mouse mammary tumor virus (MMTV) long terminal repeat (LTR), which developed adenocarcinoma or adenosquamous carcinoma [26]. In addition, transgenic mice expressing HPV-16 early region of the bovine keratin 6 gene promoter developed glandular stomach tumors [27]. However, these models did not progress to cancer through the atrophy-metaplasia-dysplasia sequence, as described by Correa [28], nor were they associated with Helicobacter infection or chronic inflammation.

INS-GAS mice as a model of spontaneous gastric cancer were first described in 2000. These mice were reported to develop atrophic gastritis and intestinal metaplasia, followed by corpus cancer with a high incidence rate, and tumor development was accelerated by H. felis infection, suggesting that this model closely mimicked the clinical course of human gastric carcinogenesis [29]. This initial mouse model was followed by the H/K-ATPase-IL-1β transgenic mice, which progressed through the atrophy-metaplasia-dysplasia sequence and validated human genetic data that implicated the IL-1β gene locus as a major risk factor for gastric cancer [30]. Other studies using genetic mouse models of gastric cancer have been reported, and are detailed below. These genetic mouse models have provided considerable insight on the role of the stroma, also known as the tumor microenvironment, which has been recognized as a critical factor in various types of cancer. Modulation of cytokines, chemokines or the signaling pathways upstream has also demonstrated unequivocally the importance of inflammatory responses in gastric cancer development [31,32,33,34].

Nevertheless, despite the significant advances made utilizing diverse mouse models, these models have all shown some limitations, including modest gastric pathology, slow time course, and the absence of invasive or metastatic tumors. In addition, it is important to keep in mind that the response to infection or genetic manipulation is highly dependent on the mouse genetic background, gender, diet and housing conditions. Animal stress associated with overcrowding, inadequate sanitation, and variations in temperature, humidity, and light cycles may predispose resistant animals to adverse disease outcomes. Especially in the case of the enterohepatic Helicobacter species, differences in study outcomes may be attributed to persistent colonization by these murine Helicobacters [35,36,37].

Here we review these genetic or chemical models of gastric carcinogenesis, and compare their pathological features, limitations and contribution to our understanding of gastric carcinogenesis. We also emphasize the impact of comprehensive genomic analysis on new or emerging transgenic mouse models.

To explore the mechanisms of gastric cancer development and establish a useful animal model of gastric tumorigenesis, investigators examined the utility of a variety of chemical carcinogens. In particular, researchers focused on N-nitroso compounds, which are generated in the stomach by anaerobic bacteria following ingestion of nitrates and nitrites, which were thought to be an important inducer of human cancer. N-methyl-N-nitro-N-nitrosoguanidine (MNNG) was the first nitrosamine shown by researchers to induce stomach tumors in rats. In 1966, Schoental et al. administered MNNG to rats using a stomach tube, resulting in squamous cell carcinoma in the rat forestomach [38]. In 1967, Sugimura et al. modified their method, administering MNNG orally to rats continuously in the drinking water, and achieved for the first time a high incidence of antropyloric adenocarcinoma [10]. MNNG was later found to be a very potent gastric carcinogen in Mongolian gerbils [39,40]. Exposure to 400 ppm MNNG in drinking water for 50 weeks resulted in gastric adenocarcinomas in 63.6% of gerbils [40]. Using MNNG-induced gastric cancer model, it has been reported that administration of high-salt diet [41,42], calcium-deficient diet [43], catechol [44], or IL-1β [45] promotes gastric cancer development.

Rats and gerbils are limited as model systems, given the absence of genetic models, and thus investigators explored the effect of oral administration of nitrosamines in inbred strains of mice. However, mice proved to be remarkably resistant to MNNG-induced gastric carcinogenesis. Danon et al. infected female Balb/c mice with H. heilmannii and administered 150 ppm MNNG in drinking water for 38 weeks and found that the treated mice developed squamous cell carcinomas in the mouth and forestomach, but not adenocarcinoma in the glandular stomach [46].

Researchers then explored the utility of N-methyl-N-nitrosourea (MNU) as a gastric carcinogen in mouse models. Tatematsu et al. treated Balb/c mice with 0.5 mg MNU by weekly intragastric intubation, but found that most of the mice died due to squamous cell carcinoma in the forestomach. Interestingly, when the mouse forestomach was removed surgically prior to MNU treatment, well-differentiated adenocarcinoma developed in the glandular stomach with 100% incidence rate by 40 weeks [11]. Thus, the glandular stomach was indeed sensitive to the carcinogenic effects of MNU, but the phenotype was obscured by the greater sensitivity of the forestomach, at least at that specific dose and route of administration. Tatematsu et al. went on to demonstrate that 30–120 ppm MNU given in drinking water was preferable to oral gavage, without the induction of tumors of the forestomach [47]. The efficiency of tumor induction by MNU was found to depend on its concentration rather than total intake [48], and MNU in the drinking water at 240 ppm on alternate weeks (total exposure; 5 weeks) was effective in inducing gastric cancer in 6 strains of mice that were studied [49]. Consequently, the protocol of 240 ppm MNU treatment in the drinking water for 5 weeks (every other week) is currently a standard and widely used murine method of gastric carcinogenesis. MNU-induced tumors in mice are located mainly in the gastric antrum, and pathologically are uniformly well- or moderately-differentiated adenocarcinomas (Figure 1). The tumors are rich in stromal cells, and occasionally invade into the submucosa although signet ring-cell carcinoma and metastatic tumors are rarely, if ever seen.

The MNU mouse model of gastric cancer has been extensively used for investigating the role of various signaling pathways or transcription factors in gastric carcinogenesis, including the roles of p53 [50], NF-κB [51], MAPK pathway [52,53], Cox-2 [54,55], β-catenin [54], E-cadherin [56], and KLF4 [57]. Although MNU is an alkylating agent that can potentially induce the formation of DNA adducts and GC > TA transversion mutations, only rare mutations have been observed in N-nitroso-compounds (NOC)-induced gastric tumors of rodents [58], raising questions about the precise mechanism of carcinogenesis. MNU is also known to modify amino acids in histone proteins, especially histone H3 lysine residues, leading to chromatin remodeling [59]. Indeed, MNU treatment in mice affects the expression of TFF1, an important gastric-specific tumor suppressor gene, through epigenetic modifications. DNA and histone methylation, and promoter methylation of the TFF1 gene have also been observed in human gastric cancer [60]. This suggests that epigenetic effects are likely to constitute a key mechanism of NOC-induced carcinogenesis.

One potential criticism of the MNU mouse model of gastric cancer is the absence of Helicobacter spp.-associated chronic inflammation. While the MNU model does not proceed through a classical atrophy-metaplasia-dysplasia sequence, this latter H. pylori-dependent pathway results in achlorhydria with subsequent bacterial overgrowth; NOCs may be generated from nitrates and nitrites in this setting, and thus the argument can be made that the generation of N-nitroso compounds may play a role in Helicobacter-associated carcinogenesis. A recent study used a combination of MNU and H. felis infection [60], and achieved a very rapid induction of antral gastric cancer. This same combination was reported to induce a high frequency of gastric cancer in H. pylori-infected Mongolian gerbils compared to gerbils receiving MNU only [61,62]. Thus, the combination of Helicobacter spp. infection followed by MNU treatment mimics in some ways the proposed pathogenesis of human antral carcinogenesis.

Given the key role of Helicobacter pylori infection in the etiology and pathogenesis of gastric cancer, researchers pursued the development of animal models of gastric Helicobacter infection. The first studies supporting a potent carcinogenic role for Helicobacter species in the gastric mucosa was the ferret model of Helicobacter mustelae infection [63,64,65]. Ferrets naturally infected with H. mustelae exposed to one dose of 100 mg/kg MNNG developed gastric cancer, while H. mustalae infected ferrets did not [63]. Unfortunately, SPF ferrets not infected with H. mustelae were not available to ascertain whether these animals also would develop MNNG-induced gastric cancer. Subsequently, aged H. mustelae-infected ferrets have been reported to develop gastritis, dysplasia, and gastric adenocarcinoma [66].

With respect to H. pylori, several H. pylori strains, such as the G1.1 strain [67], the TN2 strain [68], or the B128 strain [69], have the ability to colonize Mongolian gerbils, and induce gastric adenocarcinoma. This has allowed researchers to infect Mongolian gerbils with a variety strains, including their isogenic mutants, in order to investigate the importance of bacterial virulence factors in gastric carcinogenesis [67,68,70].

Interestingly, mice in general and the C57BL/6 strain in particular are proved to be remarkably resistant to colonization with various H. pylori strains [71,72]. Thus, alternative mouse models of gastric Helicobacter infection were explored. In 1990, Helicobacter felis, a close relative of H. pylori that was isolated from the cat stomach, was shown to readily colonize the mouse stomach in large numbers [73]. Several papers reported that H. felis had the ability to induce severe gastritis and atrophy in mice [73,74,75]. Moreover, with a longer time period of observation, analysis of H. felis-infected mice showed gastric metaplasia, dysplasia and invasive cancer [21,76]. After 12–16 months of infection, extensive dysplastic lesions were evident in the gastric corpus at the squamocolumnar junction (SCJ) along the lesser curvature. After more extended periods of infection (up to 2 years), large polypoid antral tumors develop and mimic closely lesions found in humans infected with H. pylori [77,78]. In the H. felis infection model, eradication of Helicobacter infection at early time points led to a regression of inflammation, restoration of parietal cells, reestablishment of normal architecture, and prevention against development of adenocarcinoma. Bacterial eradication at 1 year was also associated with the reappearance of parietal cells and partial restoration of architecture. Thus, eradication studies in mice have revealed that inflammation, metaplasia, and dysplasia are reversible with early eradication therapy, and that progression to dysplasia can be arrested with eradication therapy at a later time point [79,80]. In humans, eradication of H. pylori in patients with gastritis but not dysplasia is linked to a decrease in the incidence of gastric adenocarcinoma epidemiologically [81,82], thus supporting findings in mouse models. Antibiotic treatment to eradicate H. pylori in gerbils and INS-GAS mice also arrests progression to gastric lesions [73,83].

Strains of H. pylori that can colonize mice—the so-called “mouse adopted strains of H. pylori”—have been developed. Among H. pylori mouse adapted strains reported to date, the Sydney strain of H. pylori (SS1) has been the best characterized and most useful in murine model systems [20]. Relatively high levels of colonization were achieved in inbred C57BL/6 mice, while colonization levels in Balb/c, DBA/2, and C3H/He strains were lower [20]. After 8 months of infection, active gastritis and severe atrophy were observed, along with detectable levels of bacteria [20]. However, even with a 2-year follow-up, lesions did not progress to gastric cancer in mice infected with SS1 or other strains, 119p and G50, although some mice developed gastric lymphoma [84]. H. pylori SS1 infection did result in development of carcinomas in situ in C57BL/129 mice 15 months after infection [19]. H. pylori infection also causes gastric cancer in genetically modified mice, for example, INS-GAS mice (see below) [85].

In H. pylori, the cag pathogenicity island (cag-PAI), a 40-kb genomic fragment containing 31 genes, encodes a type IV secretion (TFSS) apparatus used to inject bacterial proteins such as the 120-kilodalton protein CagA into host epithelial cells [78]. A series of in vitro reports have established that injection of CagA into host cells leads to phosphorylation of CagA by host cell kinases, resulting in activation of SHP-2 tyrosine phosphatase, NF-κB signaling pathways, and MAPK signaling pathways [86,87,88]. The H. pylori peptidoglycan is also injected into host cells via the type IV secretory system, leading to activation of Nod1, an intracellular pathogen recognition molecule with specificity for gram-negative peptidoglycans [89], and mice deficient in Nod1 are more susceptible than wild-type (WT) mice to infection by cag-positive strains of H. pylori [89].

Interestingly, while the H. pylori SS1 strain was initially reported to possess an intact cag-PAI, the SS1 strain used in subsequent studies does not appear to express CagA, which may be explain to some extent the limited virulence of SS1 in mice [90]. Systemic expression of CagA in transgenic mice has led to the development of gastrointestinal and hematological malignancies [91]. mice deficient in Nod1 are more susceptible than wild-type mice to infection by cag-positive strains of H. pylori [89]. Arnold et al. used CagA-positive SS-1 (PMSS1, which was original strain isolated from a patient) and showed that mice infected with PMSS1 rapidly develop gastritis, gastric atrophy, epithelial hyperplasia, and metaplasia in a type IV secretion system-dependent manner [92]. These results suggest that CagA and/or the cag-PAI may play an important role in gastric carcinogenesis. Nevertheless, in mice, cag-negative strains such as H. felis appear to be at least as carcinogenic as cag-positive H. pylori strains. Although the ability of PMSS1 to inject CagA into host cells decreases gradually after 1 month infection and disappear after 3 months, strong inflammation and phenotypic changes could be sustained for more than 6 months [92]. In addition, inactivation in a H. pylori strain of the cagE gene coding for TFSS delayed the progression to carcinoma, but neoplasia ultimately developed in all infected INS-GAS mice with the H. pylori mutant [85]. Thus, taken together, these observations might suggest that the induction of gastric preneoplasia in mice is affected by host factors, such as the inflammatory response or other genetic factors.

Support for the importance of host genetic factors modulating gastric carcinogenesis includes observations regarding Helicobacter colonization of various inbred mouse strains, which revealed markedly different responses [75,76,93]. For example, the C57BL/6 strain is more sensitive than the Balb/c strain to H. felis-induced gastric atrophy. Potential explanations include the more T-helper-1 (Th1)-dependent immune response in C57BL/6 mice, compared to a Th2-dominant immune response in Balb/c mice [76], or possibly the reduced activity of phospholipase A2 in C57BL/6 mice [76,94]. Susceptible strains such as C57BL/6 mice show much higher levels of pro-inflammatory cytokines such as IFN-γ. The use of immunodeficient mice or mice which are deficient or overexpressing cytokines such as IFN-γ, IL-10, IL-4, or IL-7, have been useful for investigating the role of the host’s immune system in development and severity of gastritis and following mucosal changes [32,95,96,97,98,99,100,101,102]. In addition, in susceptible strains such as C57BL/6 mice, H. felis-induced proliferation and apoptosis are markedly increased compared with resistant strains [76]. The role of apoptosis was further studied through the combination of Helicobacter infection and mice which lack apoptosis-related genes, such as Fas [103,104], p53 [21], or IKK-β [105]. These studies have supported the notion that increased apoptosis is critical in the development of atrophy, metaplasia, and dysplasia.

Mouse models of Helicobacter infection have been used to examine the role of other co-factors in gastric carcinogenesis, such as gender, diet, and co-infection. Gender may be important, since gastric cancer is much more prevalent in men compared to women. Helicobacter infection of some murine strains, such as H. felis or H. pylori infection of INS-GAS mice (see below), results in greater gastric carcinogenesis in male mice compared to female mice [85,106]. However, C57BL/6 mice infected with H. felis did not show significant gender differences in the incidence of gastric carcinoma [21,85], suggesting the different mechanisms of carcinogenesis in these models. Indeed, some studies of Helicobacter infection in mice indicate that female C57BL/6 mice are more susceptible to gastric disease [107,108]. High salt diets, and diets rich in nitrates and nitrites, have been associated with an increased gastric cancer risk. Treatment with N-nitroso compounds, such as MNU, prior to H. pylori infection caused more severe preneoplastic changes and increased gastric cancer as mentioned above [109,110]. C57BL/6 mice infected with SS1 and fed a high-salt diet developed more pronounced gastric atrophy and foveolar hyperplasia [111]. Concurrent parasitic infection may also alter the effects of Helicobacter infection. Co-infection of C57BL/6 mice with the helminth, Heligmosomoides polygyrus, along with H. felis, reduced the severity of gastric atrophy and preneoplastic lesions seen with H. felis alone [112]. This was associated with a shift from the usual Th1 mucosal immune response to a polarized Th2 response. Finally, gastric atrophy due to chronic H. pylori infection is associated with bacterial overgrowth, which has been postulated to be an additional risk factor for gastric cancer. Germ-free INS-GAS mice infected with H. pylori developed less severe and delayed gastric preneoplastic lesions compared with SPF-conditioned INS-GAS mice [113], suggesting that commensal bacterial flora in the stomach also influenced the development of gastric cancer.

Mouse models of Helicobacter-induced gastric cancer make it possible to explore the origin of cancer cells and surrounding tissues. By using a bone marrow transplantation technique, Houghton et al. reported that chronic infection of C57BL/6 mice with H. felis results in population of the stomach with bone marrow-derived cells (BMDCs), and that these cells progress through metaplasia and dysplasia to intraepithelial cancer [114]. These findings were confirmed in an independent study, which showed that H. pylori infection also recruits and accumulates BMDCs in gastric epithelium [115]. Fibroblasts around gastric cancer, which have been called cancer-associated fibroblasts (CAFs) recently, are expanded by H. felis infection, and they were reported to be partly derived from bone marrow [31,116].

Taken together, studies to date have demonstrated that mouse models of chronic Helicobacter infection are robust and reproducible models that provide insights into the molecular mechanism of gastric carcinogenesis. However, there are limitations to Helicobacter mouse models, which include: the limited strains of H. pylori, particularly cag positive strains, that are able to colonize in mice; the slow time course for the development of tumors; the low incidence rate of advanced or invasive gastric cancer; anatomical differences (e.g., forestomach) between human and mice; the predominance of high grade dysplasia/early gastric cancers; and the absence of metastatic disease in these carcinogenesis models.

Progress in genetic engineering technology leading to the development of transgenic or knockout mice has been useful for developing additional models of gastric cancer that have also shed light on the role of host genetic factors in gastric cancer development. These have included overexpression (or deficiencies) of growth factors and cytokines, as well as mutation of classical tumor suppressor genes and oncogenes. Below we have summarized six sets of animal models that have been reasonably well validated and that have provided unique insights into the pathogenesis of gastric adenocarcinoma. In addition, we mention briefly several other useful genetic models.

H. pylori-associated gastric cancer in humans is preceded by a cascade of precancerous lesions, and cancer emerges following a number of discrete stages, including chronic gastritis, gastric atrophy, intestinal metaplasia, and dysplasia. Thus, in addition to mouse models of gastric cancer, there are a number of genetically engineered models that show decreased numbers of parietal cells, or gastric atrophy, along with metaplasia. In this section, we review briefly a number of models of atrophy and metaplasia that could be considered for use in experimental studies (Table 3). However, most of these models do not appear to progress to neoplasia, and most have not been examined for susceptibility to cancer in response to carcinogens.

Numerous mouse models with interesting gastric phenotypes are now available for studies of gastric carcinogenesis. These include transgenic mice (e.g., INS-GAS, H/K-ATPase-IL-1β, K19-Wnt1/C2mE), knockout mice (GAS^−/−, Tff1^−/−, gp130^F/F), Helicobacter infection (H. felis, H. pylori) and carcinogen (MNU) models. These models have defined potential roles for gender, diet, bacterial flora, inflammatory cytokines, T helper immune response, acid secretion, virulence, and colonization properties of H. pylori strains and host genetic background. Reasonable models are now available for studies of early stage pathogenesis and cancer prevention.

The mouse stomach consists of four different parts: forestomach, cardia, corpus, and antrum. Some models have been described in which forestomach tumor, i.e., squamous cell carcinoma, is induced with a carcinogen or by genetic alterations [26,46,47]. These models are not useful for investigating the mechanism of human gastric cancer, as humans do not have a forestomach and gastric squamous cell carcinoma is very rare. While there has been no model that resembles human cardia cancer, a variety of corpus and antral cancer models have been reported as described above. However, in order to select appropriate mouse models of gastric cancer, investigators need to consider the histopathology, the cell type of origin and the geographic distribution of the resultant cancer. Sometimes, a given physiological stimuli will show a different phenotype between the corpus and antrum, such as seen in gastrin transgenic and knockout mice [29,60,143].

Unfortunately, genetic models of metastatic gastric cancer similar to those developed for pancreatic cancer, comprising two or three mutations targeted to specific cell lineages, are not available. These models could facilitate preclinical studies and the testing of newer therapeutics. The major limitations in the development of these models have been the weak and scattered activity of promoters used in the stomach and the lack of stomach specific promoters that target antral progenitors but are not expressed elsewhere (Table 4). K19-Cre and Foxa3-Cre constructs, for example, are expressed in the stomach and can be used as a model for the analysis of gastric cancer, but these promoter constructs are also expressed in the intestine, colon and pancreas, as well as other tissues [105,172]. Recent lineage tracing studies revealed several stem/progenitor markers of gastrointestinal tissue, including Lgr5, Bmi1, Hopx, Lrig, Sox2, and Sox9 [234,235,236,237,238,239]. Among them, Lgr5, Lrig, and Sox2 have been reported to be expressed in the mouse stomach. Lgr5 positive cells are likely to be stem cells in the gastric antrum [234,240], and can be used for mutagenic targeting to the distal stomach, but Lgr5 is also widely expressed in the intestine and elsewhere making targeted stomach mutations impractical. Similarly, Lrig and Sox2 are expressed in the intestine and other organs as well as the stomach. Villin-Cre constructs also can target antral glands, but are also expressed in whole intestine [241]. On the other hand, stomach specific promoters such as the H/K-ATPase promoter are not ideal since they target more mature parietal cells, which are unlikely to be the precursor lineage of distal gastric cancers.

One potential candidate for a gastric targeting vector might be TFF1, which shows specific expression in foveolar cells of the stomach. To date, one study has reported the use of the TFF1 promoter to disrupt TGF-β2 specifically in the stomach [194]. Further studies using TFF1-specific gene engineering should be considered. Both TFF2-CreERT and Mist1-Cre mice are available. In the corpus, TFF2 and Mist1 are known to be potent progenitor cell lineages. Mist1-expressing cells in the stomach give rise to SPEM after stimulation [226]. TFF2mRNA-expressing cells are localized to the isthmus and are progenitors for mucus neck, parietal and zymogenic cells, but not for pit or enterochromaffin-like cell lineages [242]. TFF2 is also expressed in the base of the antral glands, suggesting that TFF2 could be a potent candidate promoter vector for targeting the stomach. In order to obtain stomach-specific gene control, other stomach-specific stem/progenitor markers should be explored by using emerging lineage tracing technique in vivo and stem cell culture methods in vitro.

Recent exome sequencing or GWAS studies have clarified significant gene mutations in human gastric cancer, such as PSCA, PLCE1, FAT4 and ARID1A [244,245,246,247]. Newer mouse models which contain these gene alterations should be useful for investigating the function of these genes. In turn, comprehensive analysis of genetic and epigenetic changes of mouse gastric cancer should also be helpful for identifying additional genetic changes or epigenetic modifications which are not covered by exome sequencing or GWAS studies focusing on human gastric cancer.

In summary, several mouse models of gastric cancer are available, which have distinct mechanisms, as well as different tumor phenotypes such as time course, locations, and pathology. Researchers are thus able to use appropriate mouse models for their studies and are commensurate with their rationale and hypothesis. Newly emerging research methods, including lineage tracing or genome-wide comprehensive analysis, and also by examining non-epithelial targets, should prove helpful for understanding both the cause and ultimately the cure of gastric cancer.