Gastric adenocarcinoma is one of the most common cancers in Asia. Although its incidence in other regions is lower, it is still a major health problem worldwide. In a very recent report, in the United States alone among 22,280 patients diagnosed with gastric cancer in 2006, 11,430 are expected to die (Jemal et al., 2006). This tumor also remains a serious problem in the elderly from low-socioeconomic classes. The geographical areas with more cancer deaths (per 100,000) in 2002 were Japan, China, Latin America, parts of Eastern Europe, and Portugal (Jemal et al., 2006). The peak incidence is estimated to occur at 50–70 years, as this tumor is rare before 30 years of age (Theuer et al., 1996; Nakamura et al., 1999). Males are affected more often than females (Levi et al., 2004) and African-Americans, Hispanics and Native Americans more than Caucasians (Howe et al., 2006; Jemal et al., 2006).
The causes of the decline of gastric cancer may include improvements in diet, food storage and lower rates of Helicobacter pylori infection, thanks to better overall sanitary conditions and methods of intervention such as eradication or immunization (Crew and Neugut, 2006).
The 5-year relative survival rate for all cases in US from 1995 to 2001 was only 23% (Ferlay et al., 2001; Jemal et al., 2006), in Europe was 21% in 1991–1994 (Roazzi et al., 2003; Sant et al., 2003). Anyway, as the corresponding survival rate in Japan is reported to be approximately 60% (Sasako, 2003), the crucial question is whether and how survival in countries other than Japan could be improved.
Unfortunately, diagnosis is often delayed because most patients with gastric cancer at an early stage report vague and non-specific symptoms or no symptoms at all. The classical triade of symptoms (anemia, weight loss, and refusal of meat-based foods) is related to advanced stages. Indeed, the disease is usually diagnosed after invasion of the muscularis propria. In such patients the 5-year survival rate is less than 20% (Correa, 2004).
The vast majority of gastric malignancies are adenocarcinomas (el-Rifai and Powell, 2002). Several attempts to classify gastric cancer have been made over the past decades. Most successful, and widely used, is the classification by Lauren, which distinguishes, by microscopical morphology alone, two main cancer histological types which appear clearly as dissimilar clinical and epidemiological entities. These are the intestinal-type, a well differentiated tumor characterized by cohesive neoplastic cells forming gland-like tubular structures and the diffuse-type, a poorly differentiated tumor resulting in individual cells infiltrating and thickening the stomach wall (so-called “linitis plastica” or “leather bottle appearance”) (Lauren, 1965). There are, however, many gastric cancers which do not fit into either histological type and present a mixed pattern (intestinal and diffuse). Other histological classifications are used such as the Ming classification, based on the growth pattern (Ming, 1977), the Japan Research Society for Gastric Cancer (JRSGC) classification which individuates five common types (JRSGC, 1999) similarly to that of the World Health Organization (World Health Organization Classification of Tumours, 2000). In this paper we will use the classification according to Lauren. To note that the intestinal-type of gastric cancer is frequently accompanied by liver metastasis (Marrelli et al., 2004), while diffuse-type gastric cancer is characterized by peritoneal dissemination (Roviello et al., 2003). Moreover, the intestinal adenocarcinomas have a better prognosis than the diffuse variant.
Distal gastric cancer predominates in developing countries, among blacks, and in lower socio-economic groups, whereas proximal tumors are more common in developed countries, among whites, and in higher socio-economic classes. While distal gastric cancers are often intestinal-type cancers, proximal cancers are usually diffuse-type adenocarcinomas. Epidemiological studies have shown a decrease in the incidence of the intestinal-type (more common in males and African-Americans) and an increase in the diffuse-type, especially the signet ring cell subtype, which has become more prevalent since 1973 in the USA (Henson et al., 2004). Main risk factors for distal gastric cancer include Helicobacter pylori (Hp) infection and dietary factors, whereas gastroesophageal reflux disease and obesity play important roles in the development of proximal stomach cancer (Crew and Neugut, 2006).
A multifactorial etiological model of human gastric carcinogenesis is generally accepted, according to environmental (Palli et al., 2001; Raj et al., 2003; Sjodahl et al., 2006), occupational, and social factors (Murray and Lopez, 1997; Raj et al., 2003), associated pathological conditions (Leoncini et al., 1993; IARC Working Group, 1994; Ruggiero et al., 2004; Ando et al., 2006; Kusters et al., 2006), and genetic susceptibility (Fitzgerald and Caldas, 2004), which are involved at different stages in the cancer process.
Multiple genetic and epigenetic alterations in oncogenes, tumor-suppressor genes, cell-cycle regulators, cell adhesion molecules, DNA repair genes, and genetic instability as well as telomerase activation are implicated in the multistep process of human stomach carcinogenesis (Vindigni et al., 1997; Spina et al., 1998; Yasui et al., 2000; Tahara, 2004).
The present review is intended to focus on genetic and epigenetic alterations of stomach cancer, which are key elements for gastric oncogenesis, and examine them in the light of the recent developments. In addition, recent data about some approaches to the genetic and epigenetic therapy of gastric cancer will be outlined.
Gastric Carcinogenesis is a Multi-Step Process
Many studies have attempted to “decode” the mechanism of gastric carcinogenesis. Correa et al. (1975) described a model in which the basic, sequential steps of stomach carcinogenesis are the following: atrophy (loss of glands), intestinal metaplasia, dysplasia, and carcinoma. As in a “cascade,” each step can be divided in smaller sub-steps in terms of extension of the mucosal surface involved, as well as in phenotypic and genotypic characteristics. Helicobacter pylori may also act as a promoter in the progression from normal to neoplastic epithelium, possibly by inducing a hyperproliferative state in the inflamed gastric mucosa (Figura et al., 1998). This model was supported by a recent study in which a cohort of 4.655 healthy subjects were monitored for 7.7 years by measuring blood pepsinogen levels (markers of atrophy) and anti-H. pylori antibodies (Ohata et al., 2004).
The metaplasia/dysplasia/carcinoma sequence fits better the intestinal-type gastric cancer. Tahara proposed a model in which intestinal and diffuse gastric carcinomas can arise either de novo (Tahara, 2004) or from precursor lesions, such as dysplasia or adenoma (Fig. 1), which are actually considered as similar lesions by the Vienna classification (Schlemper et al., 2000). At the molecular level, the intestinal-type of gastric cancer seems to have different genetic origins and biological behavior with respect to the diffuse one. In fact, the intestinal-type develops by a cumulative series of genetic alterations similar to those in colorectal cancer, which are not found in the diffuse-type (Tahara, 2004). Mixed gastric carcinomas composed of intestinal and diffuse components exhibit some but not all of the molecular events for each of the two types of stomach cancer (Yasui et al., 2000; Tahara, 2004). It has been suggested that human telomerase reverse trascriptase (hTERT)-positive epithelial cells in normal gastric mucosa, intestinal metaplasia, and gastric adenoma may be viewed as epithelial “stem cells.” Hyperplasia of hTERT-positive epithelial cells in intestinal metaplasia caused by H. pylori (Hp) may induce “chronic mitogenesis” which can facilitate especially the progression of the intestinal-type gastric cancer (Tahara, 2004).
Other scientists have recently shown that although acute injury, acute inflammation, or transient parietal cell loss within the stomach do not lead to bone marrow-derived stem cell (BMDC) recruitment, chronic infection of C57BL/6 mice with H. pylori, a known carcinogen, induced repopulation of the stomach with BMDCs. Their findings suggested that stomach cancer may originate from BMDCs rather than from stomach stem cells and thus have broad implications for the multistep model of gastric cancer progression (Houghton et al., 2004).
Relevance of Genetic Events in Gastric Oncogenesis
Genetic predisposition to gastric cancer has been suggested. The first striking example of dominantly familial predisposition to gastric cancer has been described for Napoleon Bonaparte's family (OMIM *192090). Napoleon, his father, Charles Bonaparte, his grandfather, Joseph Bonaparte, his brother and his three sisters, all died of stomach cancer, most of them at an early age (Sokoloff, 1938). At present, in OMIM database, 90% of gastric cancers are considered sporadic, 10% hereditary, which only 2% of total gastric cancers present an autosomal dominant pattern of inheritance (OMIM # 137215).
Gastric cancer is a known manifestation of inherited cancer predisposition syndromes, including hereditary nonpolyposis colon cancer (HNPCC1; OMIM #120435), Li-Fraumeni Syndrome (LFS1; OMIM #151623), familial adenomatous polyposis (FAP; OMIM +175100), Peutz–Jeghers syndrome (PJS; OMIM #175200) and Cowden Disease (CD; OMIM #158350), suggesting the presence of predisposing genes with pleiotropic effects.
Genetic predisposition to gastric cancer has also been correlated with ethnic differences. New Zealand Maori families, Hawaiians, and African-Americans have shown a high frequency of gastric cancer (Goldstein and Hirschhorn, 2004). The genetic differentiation between the races is basically associated with Alu polymorphisms and short tandem-repeat (STR) sequences; most of the variations were already present in our shared African ancestors, not more than ∼100,000 years ago (Watkins et al., 2003).
The catalogue of gene alterations in gastric cancer is growing rapidly, adding further complexity in this disease. Multifactorial models fit significantly better than single major gene models in the genesis of gastric cancer.
Multiple genetic and epigenetic alterations of oncogenes, tumor suppressor genes, DNA repair genes, cell-cycle regulators, cell adhesion molecules, and growth factor/receptor systems are involved over the course of the multi-step conversion of normal epithelial cells to gastric cancer (Yasui et al., 2000; Tahara, 2004). Identification of specific genetic pathways in gastric cancer may have an impact on prognosis and selection of treatment strategies.
Two genetic pathways are present in gastric carcinoma. The first is microsatellite instability, which targets mononucleotide tracts with coding regions of cancer-related genes and is more associated with the intestinal-type. The second pathway is chromosomal deletion involving tumor suppressor genes and correlated with the growth pattern of diffuse-type carcinoma.
Microsatellite and Chromosomal Instability
Widespread tumor-associated microsatellite instability (MSI) is believed to be caused by altered repair of spontaneous DNA replication errors after mutational inactivation or epigenetic silencing of at least one of various mismatch repair genes (MMRs), including hMLH1, hPMS1, hPMS2, hMSH2, and hMSH6/GTBP. The term “microsatellite” refers to short repetitive nucleotide sequences whose detection is proof of an enhanced mutation rate. Yeast and mammalian cell studies suggest that MMR genes act not only on base/base mismatches or small insertion/deletion loops, which escape proofreading by the replicating DNA polymerase, but also on chemically altered nucleotide base pairs. MMR gene modifications underlie the development of several types of cancer, including gastric cancer (Halling et al., 1999).
Tumors with a MSI (+) phenotype have diploid DNA and follow a distinctive pathway of molecular progression, including frame-shift mutations at mononucleotide runs within key cancer-related genes (Kuniyasu et al., 1992). Another report has demonstrated that high frequency of MSI in gastric cancer seems to be associated with female sex, antral location, intestinal-type histology, advanced tumor stage, vascular invasion, positive family history and blood group type A (Ottini et al., 1997). Leung et al. suggested that high-frequency MSI in sporadic gastric cancer is mostly due to epigenetic inactivation of hMLH1 gene, and the loss of hMLH1 protein is a significant event in the development of invasive tumor (Leung et al., 1999). Frameshift mutations of human gastrin receptor gene (hGARE) are observed in gastrointestinal tumors with MSI, although the exact role of hGARE mutations in tumorigenesis remains to be elucidated (Laghi et al., 2002). Mutations in TGF-beta RII, IGFII R, and BAX genes in sporadic gastric tumors with MSI show a decreased tendency for nodal metastasis and wall invasiveness (Oliveira et al., 1998).
In a high proportion of gastric cancer cases is observed loss of heterozygosity (LOH) at chromosomes 1p, 5q, 7q, 11p, 13q, 17p, and 18p, which are possible sites of tumor suppressor genes (Motomura et al., 1988; Kim et al., 1995). Usually loss of heterozygosity is required to inactivate a MMR in stomach tumor.
Semba et al. (1998) suggested that in young patients with diffuse gastric cancer, the presence of high MSI, which might be due to defect of DNA repair system rather than hMLH1 and hMSH2, was frequently associated with LOH on chromosome 17q21 including the BRCA1 gene.
Many proto-oncogenes are activated in gastric malignancy. The c-met gene, a transmembrane tyrosine kinase receptor of hepatocyte growth factor (HGF), is found amplified in 19% of intestinal-type and 39% of diffuse-type gastric cancers (Kuniyasu et al., 1992). The majority of gastric carcinomas express two forms of the transcript, sized 7.0 and 6.0 kb. The 6.0-kb transcript of the c-met gene was expressed at considerable levels in 52% of the gastric carcinoma tissues and was closely correlated with tumor staging, lymph-node metastasis, and depth of tumor invasion (Kuniyasu et al., 1993).
Amplification of K-sam gene is restricted to poorly differentiated types of gastric cancer (Hattori et al., 1990; Tahara, 2004). K-sam was first gene found amplified in the gastric cancer cell line KATO-III. It encodes a receptor tyrosine kinase that belongs to the heparin-binding growth factor receptor, or fibroblast growth factor receptor, gene family. K-sam has at least four transrciptional variants (Katoh et al., 1992). One of these, type II, encodes a receptor for keratinocyte growth factor. Amplification of K-sam was found preferentially in advanced diffuse- or scirrhous-type gastric cancers (33% of all) but not in intestinal-type carcinomas (Hattori et al., 1990). Overexpression of this oncogene is associated with worse prognosis in gastric malignancy.
The c-erbB-2 gene is another potential cell surface receptor of the tyrosine kinase gene family. Indeed, the c-erbB-2 gene is a v-erbB-related proto-oncogene which encodes a protein similar to, but distinct from, the epidermal growth factor (EGF) receptor. It is commonly amplified in the intestinal-type of gastric adenocarcinoma (Yokota et al., 1988). c-erbB-2 protein expression is enhanced in advanced stages during the progression of gastric carcinoma and is an indicator of poor short-term prognosis (Yonemura et al., 1991).
Mutations of K-ras oncogene can be found in intestinal-type cancer and the precursor lesions, intestinal metaplasia, and adenoma. However, K-ras point mutations are uncommon in stomach cancer and are not present in diffuse gastric tumor histology (Lee et al., 1995).
Finally, a very recent immunohistochemical study demonstrated high correlations between EZH2, the human nuclear protein which shows sequence homology to the “Enhancer of Zeste” protein of Droshophila, with the intestinal-type and the risk of distant metastasis, demonstrating that this protein predicts the aggressiveness in this histotype of gastric malignancy (Mattioli et al., 2007).
Tumor Suppressor Genes
The p53 gene is probably the most famous tumor suppressor gene and could not be absent from the list of genes involved in gastric carcinogenesis (Roviello et al., 1999; Vogiatzi et al., 2006a). It is frequently inactivated in gastric carcinomas by loss of heterozygosity (LOH), missense mutations, or frameshift deletions. Taken together, these genetic alterations are present in more than 60% of gastric carcinomas and are also found in intestinal metaplasia, dysplasia, and adenomas (Sano et al., 1991; Tamura et al., 1991; Ochiai et al., 1996). The p53 gene (locus 17p13.1) frequently shows GC-AT transitions in diffuse-type gastric cancer, due to carcinogenic N-nitrosamines produced from dietary amines and nitrates in the acid gastric environment (Sugimura et al., 1970; Mirvish, 1971; Yokozaki et al., 1992).
LOH and abnormal expression of the p73 gene, another p53 family member mapping at 1p36, a minimal region frequently mutated in gastric cancer, preferentially occur in the de novo pathway for well-differentiated adenocarcinomas of foveolar type expressing pS2 (TFF1), a gastric-specific trefoil factor (Yokozaki et al., 1999; Tahara, 2004). The pS2 protein is normally expressed in gastric foveolar epithelial cells. Inactivation of the pS2 gene is observed in dysplasia, adenoma, and adenocarcinoma in mice (Lefebvre et al., 1996), suggesting its role at early steps of gastric carcinogenesis (Tahara, 2004).
Germline mutations in the adenomatous polyposis coli (APC) gene cause FAP, which is an autosomal-dominant colorectal cancer syndrome (Kinzler et al., 1991; Rowley, 2005). Loss of heterozygosity of the two closely spaced APC/MCC genes, which are involved in colon tumorigenesis, has been also shown to be associated with the development of gastric carcinomas (Kinzler et al., 1991). Inactivation of APC (locus 5q21-q22) as well as of DCC (locus 18q21.3) and Rb1/p105 (locus 13q14.1-q14.2), found in gastric cancers, seems to be involved in the development and progression of some human gastric cancers, regardless of histologic type (Cho et al., 1996). Notably, APC gene missense mutations are present in more than 50% of the intestinal-type gastric cancer, while they are not involved in diffuse-type cancers. Somatic mutations of the APC gene are observed in precursor lesions of the stomach, such as in 20%–40% of gastric adenomas and in 6% of intestinal metaplasias, demonstrating its role in early steps of gastric carcinogenesis (Nakatsuru et al., 1992, 1993).
Loss of heterozygosity on chromosome 10q23.31 of tumor suppressor gene PTEN appears in precancerous lesions. PTEN mutations are restricted to advanced gastric cancer. In fact, LOH and mutation of PTEN are closely related to infiltrating and metastatic gastric cancers (Li et al., 2005). In a more recent paper based on immunohistochemical analysis in a large number of patients, it is shown that SURVIVIN (BIRC5), an inhibitor of apoptosis, is positively correlated with PTEN expression in gastric cancer and is a molecular marker of lymph node metastasis, while PTEN expression is reconfirmed as a molecular marker of advanced gastric cancer (Deng et al., 2006).
RUNX3 gene is a relatively recently discovered tumor suppressor, also involved in the complex process of gastric oncogenesis (Vogiatzi et al., 2006c). Loss of RUNX3 by hypermethylation of its promoter results in many tumors, including gastric malignancy. RUNX3 methylation is observed in chronic gastritis, intestinal metaplasia, and gastric adenomas, suggesting this gene as a target for epigenetic gene silencing in stomach cancer (Li et al., 2002).
Nuclear retinoic acid receptor β, RARβ, is another tumor suppressor gene found hypermethylated in 64% of the intestinal-type gastric cancers, while alterations of this gene are not observed in the diffuse-type (Hayashi et al., 2001).
Cell-Adhesion and Metastasis-Related Molecules
Mutations in genes encoding for cell-adhesion molecules have been described in gastric cancer as well. Inactivation or downregulation of E-cadherin protein, which belongs to the functionally related trans-membrane glycoprotein family, is found in gastric cancer and contributes to an increase in cell motility, the first step of cancer invasion and metastasis. This protein, a product of the CDH1 gene (locus 16q22.1), is responsible for the Ca (2+)-dependent cell–cell adhesion mechanism; therefore, its inactivation has been suggested to play an important role in the growth and invasion either in hereditary gastric carcinoma (HGC) or in hereditary diffuse gastric cancer (HDGC) (Guilford et al., 1998; Caldas et al., 1999; Brooks-Wilson et al., 2004). It has been shown that intragenic deletion or somatic mutations of the CDH1 gene and promoter methylation synergistically induce CDH1 downregulation in HDGC patients (Oliveira et al., 2004). In recent screenings, CDH1 somatic mutations found in sporadic diffuse and in a diffuse component of mixed gastric cancer were especially in-frame deletions and missense, while the major germline mutations of CDH1 gene found in familiar gastric cancer were missense (30% of all germline mutations reported to date), frameshift, and nonsense (Gayther et al., 1998; Guilford et al., 1998; Suriano et al., 2003; Brooks-Wilson et al., 2004). In particular the mutations of CDH1 in exons 8 or 9 induced the scattered morphology, decreased cellular adhesion and increased cellular motility of diffuse-type gastric cancers (Handschuh et al., 1999). More papers also demonstrated that E-cadherin mutations together with those of β-catenin and γ-catenin are involved in the development and progression of diffuse- and schirrhous-type cancers (Kawanishi et al., 1995; Shibata et al., 1996; Caca et al., 1999).
Rare genetic alterations of IQ motif-containing GTPase-activating protein 1 gene (IQGAP1), also called p195 (locus 15q26), a negative regulator of cell–cell adhesion at adherens junctions, have been found especially in diffuse gastric cancers (Morris et al., 2005). Previous study demonstrated that mutant mice exhibited a significant increase in late-onset gastric hyperplasia relative to wild-type animals of the same genetic background (Li et al., 2000).
Abnormal CD44 transcripts containing the intron 9 sequence are found in both types of gastric cancers and metastasis (Higashikawa et al., 1996). This event is also found in 60% of intestinal metaplasias, and does not take place in normal gastric mucosa (Yoshida et al., 1995). Osteopontin (OPN), a protein ligand of CD44, is upregulated in gastric cancers and together with abnormal CD44 result in lymphatic invasion and metastasis (Weber et al., 1996; Ue et al., 1998). Galectin-3, a galactoside-binding protein is another molecule implicated in gastric tumor metastasis (Lotan et al., 1994).
Cyclin E overexpression is a common event in gastric cancer and is associated with increased aggressivity in the presence of aberrant p53. The combination of cyclin E overexpression with aberrant p53 expression in gastric cancer further distinguished a subgroup of patients with poor prognosis (Bani-Hani et al., 2005).
Moreover, reduced p27 expression is a negative prognostic factor for patients with cyclin E positive gastric tumors (Xiangming et al., 2000).
The E2F family of transcription factors plays a key role in the control of cell-cycle progression. Some family members may act as oncogenes, others as tumor-suppressor genes. Suzuki et al. (1999) reported increased expression of E2F-1 mRNA in 40% of the gastric carcinomas.
Growth Factors and Cytokines
A broad range of growth factors and cytokines are produced in the gastric tumor environment by different cells accounting for complex cell interactions and for regulation of differentiation, activation, and survival of multiple cell types. Besides the role of intratumoral cytokine network, we will discuss the role of growth factors and their effects in diverse histotypes. EGF, TGFα members of the EGF family are overexpressed in the intestinal-type of gastric carcinomas (Tahara, 2004). TGFβ growth factor is more prevalent in diffuse-type carcinomas with diffusely productive fibrosis (Yoshida et al., 1989; Tahara, 2004). IGF II and bFGF growth factors are overexpressed in both histotypes of gastric cancer (Tahara, 2004).
Gastric cancer cells express neuropilin-1 (NRP-1), which is a membrane-bound coreceptor for both VEGF-165 and VEGF receptor 2 (VEGFR-2) in endothelial cells. It is known that NRP1 plays versatile roles in angiogenesis, axon guidance, cell survival, migration, and invasion. In the case of human gastric cancer, regulation of NRP-1 expression is intimately associated with the EGF/EGF-R system. It was shown that activation of EGF-R may contribute to gastric cancer angiogenesis by a mechanism that involves upregulation of VEGF and NRP-1 expression via multiple signaling pathways (Akagi et al., 2003).
Angiogenic factors, such as vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), and inteleukin-8 (IL-8) promote neovascularisation of gastric cancer. Moreover, VEGF promotes in particular the malignant progression of the intestinal-type and amphiregulin (AR), another member of EGF family, is overexpressed in both types of gastric cancer (Tahara, 2004).
IL-8 is a member of the CXC family of chemokines, which plays a pivotal role in gastric oncogenesis; more than 80% of stomach tumors express both this cytokine and its receptor (Kitadai et al., 1998, 2000). Among its activities, IL-8 enhances expression of EGF receptor, type IV collagenase (metalloproteinase (MMP)-9), VEGF, and IL-8 mRNA itself by gastric cancer cells, while also reducing CDH1 mRNA expression.
Epigenetic Events in Gastric Cancer Development and Progression
The field of cancer epigenetics is evolving rapidly on several fronts. Advances in our understanding of chromatin structure, histone modification, transcriptional activity, and DNA methylation have resulted in an increasingly integrated view of epigenetics. In response to these insights, our knowledge in gastric carcinogenesis will be fragmentary, unless we consider the link between genetics and epigenetics. The following are examples of epigenetic alterations found in gastric malignancy.
Transcriptional silencing of tumor suppressor genes by hypermethylation plays a crucial role in the progression of gastric cancer. It has been shown that there are at least two types of CpG islands in the intestinal-type and diffuse-type of gastric cancer with diverse methylation phenotypes; both are attractive for research (Ushijima et al., 2005).
In a study of methylation profiles of p16, hMLH1 genes and four CpG islands (MINT1, MINT2, MINT25, and MINT31) in gastric carcinomas, the authors distinguished them in concordant methylation of multiple genes/loci (CIMP-high) (31% of tumors), CIMP-low (55% of tumors), and CIMP-negative (13% of all tumors). The CpG island methylator phenotype (CIMP) and microsatellite instability (MSI) status in resected gastric cancers were compared with clinicopathologic features and overall survival (An et al., 2005). MSI status of the tumor was not a significant predictor of prognosis, while CpG island methylator phenotype (CIMP) status was a good but not independent prognostic factor of gastric cancer.
Diverse studies have been done to classify the methylation behavior of the genes in gastric carcinoma. In one study it has been suggested that genes could be divided in five different classes: (1) GSTP1 and RASSF1A genes, which are methylated only in carcinoma; (2) COX-2, hMLH1, and p16 genes, presenting low methylation frequency in chronic gastritis, intestinal metaplasia, and gastric adenoma but significantly higher methylation frequency in carcinoma; (3) MGMT gene, showing low and similar methylation frequency in all cancerous steps; (4) APC and E-cadherin genes, presenting high and similar methylation frequency in all cancerous lesions; and (5) DAPkinase, p14, THBS1, and TIMP-3 genes, showing an increasing methylation frequency with the progression of the disease (Kang et al., 2003).
De novo methylations have been found in a large variety of neoplastic diseases and involve different pathways. In stomach cancer these are cell-cycle regulation, apoptosis, and cell signaling.
Among tumor suppressor cell-cycle genes, hypermethylation of the p16INK4A (locus 9p21) and p15INK4B (locus 9p21) cell-cycle inhibitors has been found in gastric cancer (Lee et al., 2002). In particular, aberrant methylation of p16INK4A promoter may predict the malignant potential of dysplasia, leading to early tumor identification (Herman et al., 1995; Esteller, 2003).
Other tumor suppressor genes related to an apoptotic response are also inactivated by de novo hypermethylation in gastric cancer. One of these is the tumor suppressor gene p14ARF (locus 9p21) which becomes unable to inhibit the MDM2 oncogenic protein, an inhibitor of p53 pro-apoptotic function (Esteller, 2003).
Inactivation of APC/E-cadherin pathways also seems to be influenced by epigenetic events in gastric cancer. Recently, it has been recognized that aberrant methylation of APC is common in many neoplasms of the aerodigestive tract and that E-cadherin, a member of the APC pathway, is also hypermethylated in gastric tumors (Esteller et al., 2001). Another member of the cadherin family, CDH4 (encoding for R-cadherin), containing CpG islands located at the 5′ first exon, is methylated in 78% of colorectal and 95% of gastric carcinomas. CDH4 methylation was also detected in histologically normal tissues located in proximity of the neoplasms, indicating that CDH4 methylation is an early event in gastrointestinal tumor progression. It has been proven that CDH4 methylation can be also revealed in the peripheral blood of cancer patients, thus suggesting that CDH4 may act as a tumor suppressor gene in human gastrointestinal tumors and can potentially be used as an early diagnostic marker for gastrointestinal tumorigenesis (Miotto et al., 2004).
Several studies have described the DNA methylation of many DNA mismatch repair genes such as hMLH1, whose promoter region was found to be methylated in 100% of MSI-H sporadic gastric cancers (Leung et al., 1999).
Transcriptional silencing of the nuclear retinoid acid receptor β (RARβ), a hormonal responsive gene, has been frequently associated with gastric carcinoma (Hayashi et al., 2001).
Another gene found to undergo aberrant methylation in gastric cancer is the RAS-related gene, RASSF1 (locus 3p21.3). At least two forms of RASSF1 are expressed in normal human cells; during carcinogenesis, however, the RASSF1A isoform is highly epigenetically inactivated in a variety of tumors such as lung, breast, ovarian, kidney, prostate, thyroid, and gastric carcinomas. RASSF1A inactivation and K-ras activation are mutually exclusive events in the development of certain carcinomas. This observation could further pinpoint the function of RASSF1A as a negative effector of K-ras in a pro-apoptotic signaling pathway. Re-expression of RASSF1A reduced the growth of human cancer cells, supporting a role for RASSF1A as a tumor suppressor gene. Loss or abnormal downregulation of RASSF1A correlate with tumor stage and grade but not with histological types of gastric tumors. No somatic mutations have been detected in RASSF1 transcripts expressed in unmethylated tumors; therefore, RASSF1A methylation could serve as a useful marker for the early diagnosis and prognosis of cancer patients and could become an important target for the pharmacological therapy of cancer (Byun et al., 2001).
It has been reported that methylation of “methylguanine-DNA methyltransferase” (MGMT) gene (locus 10q26) is associated with advanced stages and poor prognosis in gastric carcinoma (Esteller, 2003).
Histone acetylation appears to play an important role in transcriptional regulation. Inactivation of chromatin by histone deacetylation is involved in the transcriptional repression of several tumor suppressor genes, including p21WAF1/CIP1. A study in gastric carcinoma investigating the status of histone acetylation found that histone H4 acetylation in both the promoter and coding regions of the p21WAF1/CIP1 gene in cells expressing dominant-negative p53 was less than half compared to that observed in cells expressing wild-type p53, whereas histone H3 acetylation in both the promoter and coding regions was slightly reduced (by approximately 20%) in cells expressing the dominant-negative p53 (Mitani et al., 2005).
“Pin2-interacting protein 1” gene (briefly PINX1) maps at 8p23 and is a potent telomerase inhibitor and a putative tumor suppressor gene (Zhou and Lu, 2001). LOH of PINX1 locus and hypoacetylation of histone H4 in the 5′ UTR of PINX1 are associated with reduced expression in this malignancy (Kondo et al., 2005).
Although hypomethylation was the originally identified epigenetic change in cancer, it was overlooked for many years compared to hypermethylation. Recently, gene activation by cancer-linked hypomethylation has been rediscovered. Recent paper proved that demethylation of specific CpG sites within the first intron of R-RAS oncogene caused activation in more than half of gastric carcinomas. Silencing R-RAS-expressing cells resulted in the disappearance of the adhering cells, suggesting that functional blocking of the R-RAS-signaling pathway can be used in gastric cancer therapy (Nishigaki et al., 2005).
Gene Therapy for Gastric Cancer
The field of gastric cancer presents an impressive flourishing of studies testing the possibilities and actual efficacy of the many different strategies employed in gene therapy. Overall results seem to be two-sided: while original ideas and innovative protocols are providing extremely interesting contributions with great potential, more advanced-phase studies concluded so far have fallen short of expectations regarding efficacy, although invariably demonstrating little or no toxicity (Vogiatzi et al., 2006b). Some examples of gene therapy applications in gastric cancer follow.
The p53 gene is critical for the suppression of tumorigenesis and this property has proven useful in gene therapy applications. Introduction of the p53 tumor suppressor gene via a recombinant adenovirus inhibits the growth of gastric cancer cells in vitro and in vivo (Ohashi et al., 1999).
The proapoptotic BCL2-associated X protein (BAX) induces cell death by acting on mitochondria. It has been shown that adenovirus mediated transfer of pro-apoptotic Bax gene was successful in vitro and in vivo, and may prove effective in gene therapy of stomach cancers (Tsunemitsu et al., 2004).
The adenovirus mediated gene transfer of caspase-8 in gastric cancer cell lines induces apoptosis in detached carcinoma cells and shows potential activity against dissemination of gastric and possibly other carcinoma cells (Nishimura et al., 2001).
The blockade by adenovirus-mediated expression of a truncated dominant negative insulin-like growth factor (IGF) I receptor sensitized the gastric cancer cells to chemotherapy and suppressed their peritoneal dissemination in vivo (Min et al., 2005).
It has also been proposed the recombinant expression of the bacterial enzyme nitroimidazole reductase gene with the prodrug CB1954 in a phase I study as a valid treatment of gastric cancer (Chung-Faye et al., 2001).
Ueda et al. (2001) showed that double transduction by the recombinant adeno vector-expressing carcinoembryonic antigen and cytosine deaminase renders gastric cancer more sensitive to 5-fluocytosine in vitro and in vivo than the single infection.
Other experimental approaches of viral therapy targeting cell-to-cell interaction molecules in gastric cancer have been recently reported. The gene transfer of CD80, a ligand of CD28, into gastric cancer cells using an adenoviral vector in vitro and in vivo has shown potential efficacy in vivo (Kosaka et al., 2004).
Adenoviral-mediated gene transduction of NK4, an antagonist of hepatocyte growth factor (HGF), inhibits both peritoneal metastasis and intra-tumor vessels in gastric cancer in vivo, regardless of the level of c-Met/HGF receptor expression in the tumor cells, and especially in the early stages of peritoneal metastasis (Ueda et al., 2004).
Tanaka et al. (2004) constructed an adenoviral vector, AdICAM-2, that encodes the full-length human ICAM-2 gene under the control of the cytomegalovirus promoter, and investigated its antitumor effects in vitro and in vivo. They concluded that their gene therapy approach might be advantageous for the cure of human scirrhous gastric carcinoma, which develops peritoneal dissemination with high frequency.
A recent study on gastric and pancreatic carcinoma cells demonstrated that adenovirus-mediated enhancement of the c-Jun NH2-terminal kinase (JNK) reduces the level of P-glycoprotein in a dose- and time-dependent manner (Zhou et al., 2006).
It has been shown that the direct suppression of Hif-1α decreased the VEGF expression inhibiting gastric tumor growth in vivo (Stoeltzing et al., 2004). Anti-angiogenic therapy can be based on transduction in the tissues surrounding the tumor to create an anti-angiogenic environment, rather than on transduction of target genes into cancer cells.
Finally, Sako et al. (2004) found that the soluble VEGF receptor sFlt-1 transduced by an adenovector in peritoneal mesothelial cells is able to inhibit the peritoneal dissemination of gastric cancer in vivo and to increase the survival of treated animals.
Demethylating/Deacetylating Agents in Therapy of Gastric Cancer
DNA methylating markers have been proposed for risk assessment, early detection, prognostic evaluation, and as therapeutic targets. Demethylating agents, such as 5-aza-cytidine (azacitidine) or 5-aza-2′-deoxycytidine (decitabine) are DNA methyltransferase inhibitors and have been shown to increase expression of de novo methylation-silenced genes and to induce cell differentiation, apoptosis, and growth suppression in a variety of cancers by global hypomethylation. They are active only in S-phase cells, where they serve as powerful mechanism-based inhibitors of DNA methylation. Moreover, the demethylating effect of 5-aza-2′-deoxycytidine seems to be universal, affecting all human cancer lines. On the other hand, demethylating drugs may have toxic effects on normal cells. Covalent attachment of the various DNA methyltransferases to DNA might be responsible for the citotoxicity of these agents, particularly at high doses (Egger et al., 2004).
DNA methyltransferase (DNMT) inhibitors reactivate gene expression in vitro in various gastrointestinal malignancies and it has been shown that histone deacetylase (HDAC) inhibitors reinforce this effect. In fact, many tumor suppressor genes are methylated in gastric cancer and their re-expression using the inhibitors of DNMT and HDAC could represent an innovative therapeutic approach in the treatment of this tumor. Recently growth inhibition effect of SK-7041 and SK-7068, HDAC inhibitors was documented, related with the induction of aberrant mitosis in human gastric cancer cells. Moreover, SK-7041 had a significant antitumor activity in human gastric xenograft model in vivo (Park et al., 2004).
Recent study investigated the combination effect of adenoviral vector carrying wild-type p53 (Ad-p53) with histone deacetylase inhibitors (HDACI) and sodium butyrate, on xenografted human gastric cancer cells (KATO-III) and hepatocellular carcinoma cells (HuH7) in nude mice (Takimoto et al., 2005). They confirmed an increased expression of Coxsackie adenovirus receptors with an associated increment of transgene (X-gal) expression with sodium butyrate treatment in KATO-III cells, and highlighted the role of sodium butyrate as a powerful enhancer of p53 gene therapy for cancer.
Trichostatin A (TSA), a deacetylating agent, can inhibit cell growth and induce apoptosis of gastric carcinoma cells through modulation of the expression of cell cycle regulators and apoptosis-regulating proteins. A recent study successfully used a combination of trichostatin A with demethylating agents in the treatment of gastric cancer, inducing specific apoptotic response in gastric tumor cells (Nishigaki et al., 2005).
It has been proposed that the demethylation can be one of the cancer-preventive mechanisms in stomach cancer. However, despite these evidences pointing on possible clinical applications of chemicals with demethylating activities, their use in clinical practice needs careful evaluation due to difficulties in correct targeting. Inducing DNA hypomethylation may also have short-term anticancer effects, but there is also the risk to a late increase in the speed of tumor progression (Ehrlich, 2002).
This is an up to date review of the genetic and epigenetic factors of gastric carcinoma. Of course they are only an aspect of a complex problem, such as stomach cancer, on which many scientists join forces to tailor a better management. The available scientific evidence indicates the combination of genetic and epigenetic changes as causes of this disease. It is particularly important to characterize any specific epigenetic alterations in gastric precursor lesions leading to malignant transformation since not all these lesions carry the same malignant transforming potential. Moreover, since gastric cancer is preceded by a very prolonged latency period, a better knowledge of involved factors, and an accurate selection of high-risk patients for genetic testing would certainly improve survival. Combining studies on human susceptibility polymorphisms and on exogenous factors such as Hp virulence may further identify other groups at high risk. Targeting earlier stages of the precursor lesions of stomach cancer in younger individuals may especially increase the likelihood of cancer prevention. The sensitivity and specificity of DNA methylation markers in cancer diagnosis depend on several factors including the type of cancer and the gene to be studied, the type of body fluid to be examined, and the techniques involved. Present genomic technologies such as cDNA microarrays, CGH array, tissue microarray technology (TMA) provide rapid methods which make it possible to identify gene expression profiles directly associated with the stage of gastric cancer; this may allow us to sharpen therapeutic strategies. Finally, understanding how the single genes specifically operate in a multi-gene approach inside the oncologic scene of gastric cancer on the whole, may help to obtain an appropriate personalized therapy, paying attention at the individual resistance or toxicity towards a specific drug. Further rigorous, placebo controlled, large-scale prospective clinical randomized trials with long periods of follow-up are needed.
The authors apologize for not being able to cite many interesting reports, due to space constraint. They gratefully acknowledge Dr. Caterina Cinti in the first steps of this manuscript and Dr. Marco Cassone for critical reading of the final manuscript and for providing useful suggestions. Paraskevi Vogiatzi acknowledges the Ph.D. in “Medical Genetics” coordinated with the specific program in “Oncological Genetics” of the University of Siena, Italy.