Hereditary diffuse gastric cancer can be caused by epithelial cadherin mutations for which genetic testing is available. Inherited cancer predisposition syndromes including Lynch, Li-Fraumeni, and Peutz-Jeghers syndromes, can be associated with gastric cancer. Chromosomal and microsatellite instability occur in gastric cancers. Several consistent genetic and molecular alterations including chromosomal instability, microsatellite instability, and epigenetic alterations have been identified in gastric cancers. Biomarkers and molecular profiles are being discovered with potential for diagnostic, prognostic, and treatment guidance implications.
Key points
- •
Hereditary diffuse gastric cancer can be caused by epithelial cadherin mutations for which genetic testing is available.
- •
Inherited cancer predisposition syndromes can be associated with gastric cancer.
- •
Chromosomal and microsatellite instability occur in gastric cancers.
- •
Several consistent genetic and molecular alterations have been identified in gastric cancers.
- •
Biomarkers and molecular profiles are being discovered with potential for diagnostic, prognostic, and treatment guidance implications.
Introduction
Gastric adenocarcinomas comprise the vast majority of malignant tumors arising from the stomach. Although other histopathologic forms of stomach tumors exist, they are rare in occurrence. Gastric cancer (GC) remains a significant worldwide health burden. GCs exhibit heterogeneity in clinical, biologic, and genetic aspects. The complexity of the genetics involved in gastric adenocarcinoma is reflected in the temporal, regional, and gender variation in GC incidence rates. A better understanding of these phenomena through molecular and genetic studies of gastric tumor genesis is anticipated to provide important insights into cancer development in general and lead to earlier diagnosis and better management options.
Introduction
Gastric adenocarcinomas comprise the vast majority of malignant tumors arising from the stomach. Although other histopathologic forms of stomach tumors exist, they are rare in occurrence. Gastric cancer (GC) remains a significant worldwide health burden. GCs exhibit heterogeneity in clinical, biologic, and genetic aspects. The complexity of the genetics involved in gastric adenocarcinoma is reflected in the temporal, regional, and gender variation in GC incidence rates. A better understanding of these phenomena through molecular and genetic studies of gastric tumor genesis is anticipated to provide important insights into cancer development in general and lead to earlier diagnosis and better management options.
Inherited susceptibility
Familial Clustering
Most cases of GC appear to occur sporadically, without an obvious hereditary component. Familial clustering has been observed in approximately 12% of gastric carcinoma cases, with a dominant inheritance pattern. Notably, Napoleon Bonaparte apparently suffered from GC involving most of his stomach and may have had other family members (ie, his father and sister) afflicted as well. In the Swedish Family Cancer Database, when a parent presented with gastric carcinoma, offspring showed an increased risk of the concordant carcinoma, with a Standardized Incidence Ratio (SIR) of 1.59, only at ages older than 50 years. The increased risk from sibling GC probands (SIR of 5.75) was noted for those diagnosed before age 50 years. Taken together, these findings suggest that some of the familial risk factors are likely to be environmental, siblings being at a higher risk than offspring-parent pairs, consistent with the transmission patterns of Helicobacter pylori infection.
Case-control studies have observed consistent (up to threefold) increases in risk for GC among relatives of patients with GC. A population-based control study found an increased risk of developing GC among first-degree relatives of affected patients (odds ratio [OR] = 1.7 with an affected parent, OR = 2.6 with an affected sibling), with the risk increasing (OR up to 8.5) if more than one first-degree relative was affected. Interestingly, a higher risk was noted in individuals with an affected mother versus an affected father. Studies have shown a slight trend toward increased concordance of GCs in monozygotic twins compared with dizygotic twins. A genomic analysis of 170 affected sib-pairs from 142 Japanese families with GC yielded several chromosomal regions, with the strongest linkage at 2q33-35, harboring potential susceptibility genes.
Inherited Predisposition Syndromes
Inherited predisposition cancer syndromes are thought to comprise only 1% to 3% of all GCs. Several genetic susceptibility traits with an inherited predisposition to GC development exist. Some are well-characterized clinically with their underlying genetic alterations being unveiled and are described in this article.
Hereditary diffuse gastric cancer
Three large Maori families with an obvious autosomal dominant, highly penetrant inherited predisposition to the development of GC, having sufficient power with which to perform productive linkage studies, revealed linkage to the epithelial cadherin ( E-cadherin ) /CDH1 locus on 16q22.1 in 1998. Further analysis of these families demonstrated association of GC development with germline mutations in the E-cadherin ( CDH1 ) gene. Since then, multiple germline E-cadherin mutations have been reported in more than 100 families throughout the world. The CDH1 gene mutations have been scattered across the 16 exons this gene encompasses, with approximately 70% being truncating and 30% missense in nature. Moreover, there have even been large deletions of the E-cadherin gene identified in a small percentage (4%) of hereditary diffuse gastric cancer (HDGC) families, likely involving nonallelic homologous recombination in Alu repeat regions.
The first definition of the hereditary diffuse GC trait, which is the only inherited cancer syndrome dominated by GC, was made in 1999. The ages of onset for diffuse GC in subjects harboring germline E-cadherin mutations ranged from 14 to older than 70 years. The cumulative risk estimate for advanced GC by 80 years of age was estimated to be 67% for men and 83% in women with wide confidence intervals, as these were based on 11 HDGC families. The mean age at GC diagnosis was 40 years in this study. Additionally, female CDH1 mutation carriers had a high risk of developing lobular breast cancers with a lifetime risk of 60% by the age of 80 years. The incomplete penetrance of germline E-cadherin mutations was seen in several obligate carriers, who remained unaffected even in their eighth and ninth decades of life. Variable penetrance is suggested in the larger HDGC families and a later onset of the youngest cases in more recent HDGC families, most in their late 20s to early 30s. A founder (common ancestry) mutation was observed in 4 Newfoundland families with penetrance of symptomatic GC by 75 years of age, being 40% for men and 63% for women. Whether this incomplete penetrance is due to the stochastic nature of the second allele alteration or perhaps to the presence of phenotype-altering alleles at other genetic loci, remains to be determined. Indeed, allele-specific monoclonal CDH1 promoter hypermethylation or loss of heterozygosity was found in early GC lesions from mutation carriers.
Importantly, 12 of 18 asymptomatic gene carriers who underwent prophylactic gastrectomy had occult cancer, often seen as tiny clusters of signet ring cells within and underlying the mucosa. Only one of these individuals had a diffuse GC detected endoscopically with random biopsies before gastrectomy, indicating the ineffectiveness of this type of surveillance for these lesions. Indeed, no lesions were identified endoscopically or radiologically in 6 CDH1 gene carriers, yet all 6 had early-stage (TMN stage T1a) multifocal signet ring cell GC found on pathology examination after prophylactic gastrectomy.
A study of 42 families diagnosed with HDGC trait by having at least 2 members affected had an E-cadherin mutation identified in 40% of cases. If the clinical criteria were more stringent to include 1 GC occurring before 50 years of age, then more than half the cases had E-cadherin mutations detected. When large deletions were screened in addition to point mutations and small frameshift mutations, 46% of 160 high-risk families were found to have a germline E-cadherin gene alteration. Furthermore, a study of 25 “sporadic” diffuse GCs identified 1 case with a germline E-cadherin mutation and none in 14 intestinal-type GCs. Germline CDH1 mutations were found in approximately 7% of sporadic diffuse GC cases diagnosed before the age of 50 years. No germline mutations of this gene were detected in apparent “sporadic” diffuse GC cases with a mean age of 62 years in Great Britain.
The first consensus guidelines for the clinical diagnosis and management of familial GC were developed in 1999, then updated in 2010. Management algorithms have been formulated, highlighting genetic testing and counseling based on collective experience and the literature to date. Genetic counseling for a kindred manifesting a strong predisposition toward development of diffuse GCs is imperative to ensure appropriate management. Once a gene carrier of an E-cadherin mutation reaches his or her mid to late 20s, prophylactic gastrectomy should be considered along with aggressive breast cancer surveillance starting at the age of 30. Diffuse GC families are one kindred subgroup to whom genetic testing can now be offered.
It is noteworthy that more than half to up to two-thirds of HDGC families reported have proven negative for the E-cadherin gene mutation. Allele expression imbalance of CDH1 was noted in a subset of these families ; however, most of these families likely have other molecular alterations underlying their cancer predisposition that are yet to be discovered.
Other hereditary cancer predisposition syndromes
Several other hereditary cancer predisposition syndromes exist that, in addition to other predominant cancers, also appear to have an increased risk of developing GC, as listed later in this article. The incidence of GC differs somewhat, but is generally low compared with other cancers in these syndromes and both intestinal and diffuse histologic subtypes of GC have been described. A subset of families with these syndromes; however, can exhibit a striking cluster of GC.
A now well-characterized inherited predisposition syndrome that may include GC development is Lynch syndrome. GCs occurring in the setting of germline mismatch repair gene mutations underlying Lynch syndrome were diagnosed at a mean age of 56 years, were predominantly intestinal-type, tended to lack evidence of H pylori infection, and exhibited microsatellite instability in a Finnish Lynch syndrome registry study.
Li-Fraumeni syndrome is a rare inherited cancer syndrome defined by a clustering of malignancies, including sarcoma, breast cancer, brain tumor, leukemia, adrenocortical, and GC, attributable to germline mutation of the tumor suppressor gene p53. An extended family affected by this syndrome demonstrated strong evidence of linkage to p53, and 3 of 4 gastric tumors analyzed showed loss of the wild-type allele. However, GCs account for fewer than 4% of all neoplasms in this rare syndrome. GCs have also been associated with the hereditary breast and ovarian cancer syndrome.
Polyposis syndromes
The Peutz-Jeghers syndrome is a rare autosomal dominant disorder characterized by germline mutations of STK11, hamartomatous polyposis, and pigmentation of the lips, buccal mucosa, and digits that carries an increased risk of many types of cancer, including those of the gastrointestinal tract, with gastric carcinomas observed. An increased risk of GC associated with familial adenomatous polyposis has been reported in high-risk regions, such as Asia, whereas no significant increased risk was exhibited in other populations. Overall, GC is rare in this setting, and the exact contribution of the polyposis and underlying germline alterations of the APC and MYH genes is unclear. Inherited allelic defects of the MYH base excision repair gene have been linked to 2.1% of sporadic cases of GC, and even to as many as 9.9% of patients with a familial pattern of GC.
GCs have been noted to occur in patients with other gastrointestinal polyposis disease entities, such as juvenile polyposis, who harbor germline mutations in SMAD4 or BMPR1A. Up to 24% of those with generalized juvenile gastrointestinal polyposis developed GC, which was similar to a subset of patients who presented with predominantly gastric polyposis and was found to have neoplastic tissue arising in 25% of the resected gastric specimens. Cowden syndrome is a gastrointestinal hamartosis polyposis syndrome that is associated with multiple cancers, including rare GC.
Other inherited syndromes
Rare kindreds exhibiting site-specific GC predilection have been reported, occasionally associated with other inherited abnormalities. Indeed, a constitutional deletion of 18p inherited from her mother, with somatic loss of the remaining long arm of this chromosome, was observed in the GC of a 14-year-old girl with associated mental and cardiac abnormalities, suggesting a predisposing condition. A few patients with GC have been rarely found to harbor germline mutations in the ATM (ataxia-telangiectasia mutated) gene and the proto-oncogene MET. Familial intestinal GC not attributed to the previously mentioned well-characterized inherited cancer susceptibility syndromes has been described as well, although the underlying defect(s) remain uncertain. Moreover, a kindred manifesting the autosomal dominant inheritance of familial gastric hyperplastic polyposis and GC development was demonstrated not to have E-cadherin or other identifiable genetic alteration.
Genetic alterations
Molecular analyses of sporadic GCs for acquired changes have described many somatic alterations, but the significance of these changes in gastric tumorigenesis remains to be established in most instances. A study by Wright and colleagues demonstrated that the human gastric mucosa is composed of clonal units with several multipotent stem cells. Additionally, intestinal metaplasia crypts exhibited clonality containing multiple pluripotent stem cells that can spread by crypt fission. Taken together, these findings provide evidence for how genetic mutations spread clonally in human gastric mucosa.
Chromosomal Instability
Most GCs exhibit significant gross chromosomal aneuploidy. One study found that 72% of differentiated tumors and 43% of undifferentiated gastric tumors were aneuploid. Variability in the classification of instability or histopathologic subtype and in the number of loci examined account for some of this variation in aneuploidy, with a trend toward more frequent occurrence in intestinal-type cancers at more advanced stages. Cytogenetic studies of GCs are few in number and have failed to identify any consistent chromosomal abnormalities. A variable number of numerical or structural aberrations have been reported in GC cells, such as those involving chromosomes 3 (rearrangements), 6 (deletion distal to 6q21), 8 (trisomy), 11 (11p13-p15 aberrations), and 13 (monosomy and translocations).
Comparative genomic hybridization (CGH) analyses have revealed several regions of consensus change in DNA copy number, indicating the possible location of candidate gastric oncogenes or tumor suppressor genes involved in gastric tumorigenesis. A knowledge of these alterations is important, as GCs of different histopathologic features have been shown to be associated with distinct patterns of genetic alterations, supporting the notion that these cancers evolve via distinct genetic pathways. Chromosomal arms 4q, 5q, 9p, 17p, and 18q exhibited frequent decreases in DNA copy number, whereas chromosomes 8q, 17q, and 20q showed frequent increases in DNA copy number. CGH analysis showed an increased frequency of 20q gains and 18q losses in tumors that metastasized to lymph nodes. As described later in this article, several known or candidate tumor suppressor genes have been isolated within some of these frequently lost regions.
Comprehensive loss of heterozygosity (LOH) analysis of these tumors has identified 3p, 4p, 4q, 5p, 8q, 13p, 17p, and 18q to be frequently lost. In LOH analysis of more than 100 archived GCs, allelic loss was most frequently noted on chromosome 3p. Moreover, 3 distinct regions of chromosome 4q were found to be frequently lost in gastroesophageal junctional adenocarcinomas, indicating the potential for multiple tumor suppressor genes to be located on this chromosomal arm.
Microsatellite Instability
Microsatellite instability (MSI) has been found in approximately 14% to 44% of sporadic GCs. The degree of genome-wide instability also varies, with more significant instability (eg, with MSI-high [MSI-H] tumors exhibiting instability in >33% of loci tested) occurring in only 10% to 16% of GCs.
Abnormal loss of protein expression of either MLH1 or MSH2 was demonstrated in all cases exhibiting MSI-H. Altered expression of MLH1 was associated with increased methylation of the promoter region of MLH1 in MSI-H cases, suggesting a silencing role of hypermethylation. Interestingly, no difference in the frequency of hMLH1 hypermethylation and MSI phenotype was noted between familial (non-Lynch and non-HDGC) and sporadic cases of GC, indicating epigenetic regulation in both these settings.
MSI-H gastric tumors exhibit distinct clinicopathologic characteristics and a unique set of genetic alterations. Consistent associations of the MSI-H phenotype with intestinal subtype, distal location (eg, antral), and more favorable prognosis have been observed. Several tumor suppressor genes have been shown to be critical targets of defective mismatch repair (MMR) in MSI-H tumors. At least one important target of MSI appears to be the transforming growth factor beta (TGF-β) type II receptor ( TGFβR2 ) at a polyadenine tract within its gene. Another gene involved in this signaling pathway, ACVR2, was found to be similarly mutated, even in a biallelic fashion, in gastric tumors exhibiting MSI. Thus, alteration of TGF-β receptors and other members of this signaling path appears to be a critical event in the development of at least a subset of GCs, allowing escape from the growth control signal of TGF-β.
The proapoptotic BAX gene and additional MMR genes have also been demonstrated to be altered in MSI-H GCs. Somatic nonframeshift mutations have been reported in BAX . Moreover, the relatively frequent missense mutations at codon 169 of BAX was shown to impair its proapoptotic activity. These same genes are observed to be infrequently mutated or altered in MSI-low or MSS tumors. Additional genes with simple tandem repeat sequences within their coding regions found to be specifically altered in GCs displaying MSI include IGFRII , hMSH3 , hMSH6 , and E2F-4, which are known to be involved in regulation of cell-cycle progression and apoptotic signaling.
Acquired somatic genetic/molecular alterations
TFF-1 Loss
Loss of the trefoil peptide TFF1(pS2), a stable 3-loop molecule synthesized in mucus-secreting cells, has been described in approximately 50% of GCs. The biologic significance of this loss was reported in a knockout mouse model of gastric antral neoplasia. Additionally, expression of TFF1 was observed to be lower in some gastric intestinal metaplasia and gastric adenomatous lesions compared with adjacent normal or hyperplastic mucosa. TFF1 resides on chromosome 21q22, a region noted to be deleted in some GCs in LOH studies. Overexpression of TFF1 in the GC cell line, AGS, inhibited its growth. Furthermore, the overwhelming majority of GCs studied had absent to minimal transcript levels compared with normal gastric mucosa. Moreover, C/EBP-β was found to be overexpressed in most GCs in a corresponding manner and bind to the promoter of TFF1, suggesting a regulatory factor role. Finally, cytokine signaling through the coreceptor gp130, with increased STAT-3 expression, has also been observed to decrease TFF1 and lead to gastric lesions in mice. Additionally, characterization of gastrokine (GKN) 2, a stomach-specific tumor suppressor protein, was found to be structurally and physically related to TFF1. Noteworthy, loss of GKN1 and GKN2 expression was found to occur in gastric adenocarcinoma and portend an overall shorter survival.
E-Cadherin Alteration
In addition to germline mutations leading to HDGC described previously, several sporadic diffuse GCs have displayed altered E-cadherin . E-cadherin is a transmembrane, calcium ion–dependent adhesion molecule important in epithelial cell homotypic interactions, which, when decreased in expression, is associated with invasive properties. Reduced E-cadherin expression determined by immunohistochemical analysis was noted in most GCs, 92% of 60 cases, when compared with their adjacent normal tissue. Genetic abnormalities of the E-cadherin gene and transcripts were demonstrated in a third to one-half of diffuse GCs. E-cadherin splice site alterations producing exon deletion and skipping, large deletions including allelic loss, and point mutations mostly of the missense nature have been demonstrated in diffuse-type cancers, some even exhibiting alterations in both alleles. Methylation of the E-cadherin promoter region was found in 16 (26%) of 61 GCs studied.
Kinases/Phosphatases
Activation of protein kinases by tyrosine kinases and phosphatases has been shown to affect many cellular activities, including cell growth, differentiation, and survival. Noteworthy, a novel KGF-R2 phosphorylation inhibitor, Ki23057, decreased proliferation of 2 scirrhous GC cell lines with K-sam amplification. Phosphatidylinositol 3-kinases (PI3K) are lipid kinases that regulate pathways for neoplastic proliferation, adhesion, survival and motility. Mutations in PIC3A, which encodes a catalytic subunit of PI3K, have been characterized in several cancers including gliobastoma, colon, and up to 25% of GCs analyzed. Further attempts at identifying mutations in PIC3A have reported a lower prevalence of 4.3% in GCs; however, this study also reported increased expression of this gene product in gastric tumors and a specific expression profile associated with this upregulation. Moreover, increased PI3K/AKT signaling from nuclear factor kappaB (NF-kB) has been shown to activate transcription of HuR, which has proliferative and antiapoptotic effects on GC cells. Furthermore, a kinase involved in chromosome segregation and stability, STK15 (BTAK, Aurora2), was found to be amplified and overexpressed in GCs.
Wang and colleagues identified 83 somatic mutations in human protein tyrosine phosphatases (PTPs), most affecting colon cancer, but also found them in 17% of GCs analyzed. The most commonly altered PTP, protein-tyrosinase phosphatase-receptor type (PTPRT) was found to decrease its activity, suggesting a role as a tumor suppressor gene. However, an additional polymerase chain reaction–based study suggests a much smaller role for alterations in PTPRT in gastric tumorigenesis, being found in only 1% of cases. Overexpression of DARP32 and a novel truncated isoform was found in most GCs. Moreover, an antiapoptotic effect of this overexpression of DARP32 and t-DARP was observed in vitro.
Overexpression of the MET gene, which encodes a tyrosine kinase receptor for the hepatocyte growth factor, has been reported to have prognostic value to indicate poorer survival in multivariate analysis. There have been numerous reports in the literature indicating that the MET gene is amplified in approximately 15% and its expression elevated in up to 50% of GCs.
Methylation Silencing Alterations (Epigenetic Alteration)
A significant number (41%) of GCs exhibited CpG island methylation in a study of the promoter region of p16. Many of these cases with hypermethylation of promoter regions displayed the MSI-H phenotype with multiple sites of methylation, including the MLH1 promoter region.
RUNX3, a tumor suppressor gene, appears to suppress gastric epithelial cell growth by inducing p21 (WAF1/Cip1) expression in cooperation with TGF-β–activated SMAD. RUNX3 was found to be altered in 82% of GCs through either gene silencing or protein mislocalization to the cytoplasm. Multiple other tumor suppressor genes and candidates, including protocadherin 10 (PCDH10), CDKN2A, GSTP1, APC, MGMT, DAKP, XAF1, THBS-1, RUNX1, and CDH1, have been shown to be methylated in GCs.
Apoptosis Signaling Alterations
Mutations of the BAX gene, believed to be a promoter of apoptosis, have been identified in up to 33% of GCs, as well as in colorectal and endometrial cancers. Another study supporting the role of the BAX gene in tumor suppression linked low expression levels of the protein Bax-interacting factor-1 (Bif-1) in GCs, suggesting another breakdown in the pathway of regulating apoptosis. The bcl-2 homolog BAK, another promoter of apoptosis, was found to have missense mutations in 12.5% of GCs analyzed. Defects in cell surface receptors of the apoptosis pathway have also been observed in gastric tumors, such as in the death receptors DR4 and DR5.
Other Alterations
The p53 gene has consistently been demonstrated to be altered in GCs, both by allelic loss in more than 60% of cases and mutations identified in approximately 30% to 50% of cases. The spectrum of mutations in this gene is similar to other cancers, with a predominance of base transitions, especially at CpG dinucleotides.
The human epidermal growth factor receptor 2 (HER2/neu/ERBB2) has been demonstrated to be overexpressed in many types of cancer, including GC; and to be associated with metastasis. HER2 was shown to interact with CD44 and upregulated CXCR4 through epigenetic silencing of microRNA (miRNA)-139 in GC cells. Several of the vascular endothelial growth factor (VEGF) gene polymorphisms were found to be independent prognostic markers for patients with GC and the analysis of VEGF gene polymorphisms may help identify patients at greatest risk for poor outcomes.
Evidence of tumor suppressor loci on chromosome 3p has accumulated from a variety of studies, including allelic loss in primary GCs (46%) and homozygous deletion in a GC cell line (KATO III), as well as xenografted tumors. The FHIT gene was isolated from the common fragile site FRA3B region at 3p14.2 and found to have abnormal transcripts with deleted exons in 5 of 9 GCs. Furthermore, loss of FHIT protein expression was demonstrated immunohistochemically in most GCs. One somatic missense mutation was identified in exon 6 of the FHIT gene during a coding region analysis of 40 GCs. Additional studies are needed to determine the role that breakpoints in this region of 3p have in GC development.
Somatic mutations in the chromatin remodeling gene, ARIDIA, occurred in several GCs as well as other tumors. Only 1 of 35 GCs contained an intragenic mutation of SMAD4 along with allelic loss, suggesting this gene is infrequently altered in sporadic gastric tumorigenesis. Additionally, only one missense change of uncertain functional significance in the LKB1 (STK11) gene was noted in a study of 28 sporadic GCs.
Members of the Wnt signaling pathway, APC and β-Catenin , have had several somatic alterations noted in a few cases of GC. Missense somatic mutations in β-Catenin , which also has connections with the cell adhesion complex involving E-cadherin, as well as Lcf/LEF transcription regulation, were identified in a few cases of intestinal-type GC. Elevated expressions of the hedgehog target genes, human patched gene 1 (PTCH1) or Glil, were reported in 63 of 99 primary GCs. Interestingly, treatment of GC cells with KAAD-cyclopamine, a hedgehog signaling inhibitor, decreased expression of Glil and PTCH1, resulting in cell growth inhibition and apoptosis. Additionally, cyclopamine decreased the growth of GC cell lines that expressed high levels of spermine oxidase.
Global molecular profiling
Deregulation of canonical pathways, such as TFF1, HER2, p53, and Wnt/B-catenin, are known to occur with varying frequency in GC; however, studies have typically focused on single pathways measuring only one or a few targets. Recent evidence indicates that most cancers are products of complex interactions between multiple pathways. More global analyses are beginning to provide better definition of higher-order relationships between distinct oncogenic pathways. Elucidating these relationships may help to bridge the gap to clinical utility. Moreover, the advancement in our molecular understanding of tumorigenesis as a whole with these global profiling studies may result in specific targeted chemotherapy with dramatic effects in controlling tumor growth and survival.
Genomics
High-throughput analyses, such as microarrays and next-generation sequencing that comprehensively determine gene sequences or DNA copy number patterns, are being explored in various diseases, including GC. Genome-wide association studies have led to identification of multigene markers that correlate with prognostic significance in GCs. One study analyzed 47,296 transcripts from GC and, through microarray analysis, identified a 10-gene panel that correlated with overall prognosis from 39 cancer samples. The panel consisted of 6 ribosomal proteins (RPLP2, RPS12, RPS8, RPS19, RPS12, and RPS15P4) and EIF3S6, GLTSCR2, TMSB10, and SEC61G, thought to be targets of tumor suppressor and oncogenes. The panel was then validated in an independent data set containing 33 cancer samples. Furthermore, a genome-wide association study in 3279 individuals identified noncardia GC susceptibility loci at 5p13.1, which contains the PTGER gene responsible for regulation of cell migration and immune response and thought to induce growth inhibition of human GC cell lines. The same study identified cancer susceptibility loci at 3q13.31, which contains the ZBtB20 gene that encodes zinc finger protein homologous to bcl6 that is involved in the immune response, hematopoiesis, and oncogenesis. Moreover, the investigators validated several other loci, including rs2294008 and rs2976392 on 8q24, rs4072037 on 1q22, and rs13042395 on 20p13 being associated with noncardia GC in a further 6897 subjects.
Using genome-wide promoter methylation analysis, ADAMTS9 was identified as a functional tumor suppressor in GC by blocking AKT/mTOR signaling. Indeed epigenetic silencing through hypermethylation was identified in 29.2% of primary GC tumors. Additionally, multivariate analysis showed that patients with ADAMTS9 methylation had a poorer overall survival.
Genome copy number aberrations (CNAs) have been extensively characterized by comparative genomic hybridization in GC, as stated previously. A more global study, by Kuroda and colleagues, demonstrated no significant association between number of genomic CNAs and GC submucosal invasion or lymph node metastasis. They did, however, demonstrate that subclones that acquire gain of 11q13, 11q14, 11q22, 14q32, or amplification of 17q21 were more frequent in submucosal-invasive GCs.
Transcriptomes
Although genomics has identified numerous genes/pathways involved in GC, transcriptome analysis elaborates on this technique, specifically identifying those genes that are silenced or upregulated. Regulation of genes, through CNAs, methylation, and other epigenetic mechanisms, ultimately affects their expression. Comprehensive serial analysis of gene expression has identified novel genetic alterations, including overexpression of calcium-binding proteins.
Noncoding RNA sequences, miRNAs, have demonstrated considerable association with tumorigenesis, including GC. Mechanisms that dysregulate miRNAs include aberrant miRNA biogenesis and transcription, DNA methylation, epigenetic alteration, and amplification or loss of genomic regions that encode miRNAs. This dysregulation can promote progression of the GC cell cycle by altering translation of mRNAs that encode cyclin-dependent kinase inhibitors (CKIs), such as p21 and p16. Furthermore, this altered expression of CKI mRNAs reduces apoptotic signaling through regulation of Bcl-2, which has previously been described as a prognostic factor associated with the stage of GC. Genome-wide studies have also shown that miRNAs are frequently found at genomic regions where LOH and amplification occur. For instance, miR-486, which has a role as a tumor suppressor, was found to be downregulated via loss of its locus in 25% to 30% of 106 gastric tissue samples. Several miRNAs expressed aberrantly in GCs affect important signaling pathways, such as TGF-β. Expression profiles of miRNAs have revealed 21 individual miRNA and 6 miRNA clusters that are consistently upregulated in GCs.
It was postulated that PRKAA2 was a target gene for miR-19a when an inverse correlation was observed. PRKAA2 is involved in the AMPK pathway, which inhibits the mTOR pathway, a major cancer growth-promoting signaling that regulates hypoxia-inducible factor-1a (HIF-1a). Furthermore, hepatocyte-NF-4a (HNF4a) and HIF-1a expression levels were demonstrated to be significantly higher in stage I-II GC explants compared with stage III-IV or normal tissue, suggesting a role as a biomarker of early GC. Interestingly, studies to amplify the AMPK pathway, and thereby increase the concentration of PRKAA2 mRNA levels using metformin, resulted in decreased expression of HNF4a and an inhibition of the expression and transactivating activity of HIF-1a.
Although NF-kB has been shown to be activated by H pylori , a known GC carcinogen, and altered NF-kB is known to play a role in multiple inflammation-based cancers, studies to date have not demonstrated overt differences in NF-kB activation, perhaps because of the limitations of older technologies. However, pathway profiling analysis of multiple components, including posttranslational modification of gene transcripts, has demonstrated NF-kB to be constitutively activated in GC. Thus, targeted NF-kB inhibitors are currently being developed. Indeed, Sohma and colleagues demonstrated NF-kB phosphorylation to be downregulated by the inhibitor parthenolide. Furthermore, they demonstrated that the inhibitory action of parthenolide on NF-kB signaling was associated with enhanced chemosensitivity to paclitaxel, thereby providing evidence of a promising potential therapeutic target against GC.
Proteomics
Multiple and multidimensional proteomic techniques, including 2-DE (dimensional electrophoresis), iTRAQ (isobaric tags for relative and absolute quantitation), ICAT (isotope-coded affinity tag), protein chip array and liquid chromatography, have become more advanced and are now being applied to analyzing quantitative differences in proteins present in various specimens. A proteomic analysis identified 9 proteins with increased expression and 13 proteins with decreased expression in GCs, which included those involved in mitotic check points, such as MAD1L1 and EB1, as well as others, including HSP27, CYR61, and CLPP. Proteomic analysis using 2-DE profiles has revealed differential expression of annexins in H pylori –associated GCs. Additionally, protein chip array and surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) analyses have demonstrated upregulation of pepsinogen C and pepsin A, as well as downregulation of alpha-defensins in patients with GC. Fibroblast growth factor–inducible factor 14 (Fn14), a type 1 transmembrane protein that belongs to the tumor necrosis factor (TNF) superfamily that functions through the NF-kB pathway, was found to be overexpressed in GC compared with normal tissue. Moreover Fn14 expression levels were inversely correlated with patient survival in 87 patients with late-stage GC.
Although proteomic analysis in GCs has previously been focused on surgical specimens, one study using matrix-assisted laser desorption/ionizations (MALDI) mass spectrometry on endoscopic biopsies identified a signature protein profile consisting of 73 signals that included alpha-defensins that distinguished cancer samples from normal tissue. In a validation set, this panel exhibited 93.8% sensitivity and 95.5% specificity. Furthermore, the investigators were also able to identify a protein signal that distinguished stage 1a GCs from more advanced lesions.
Novel splicing variants of ALDOC were identified by combined analysis with proteomics and transcriptome analysis. Specifically, the variant 4b was found to be highly expressed in most metastasis-related GC cell lines and poorly differentiated GC tumors with overexpression of ERB2/HER2/neu. A total of 151 differentially expressed proteins, including galectin-2 and other proteins linked to TNF and p53, were upregulated in lymph node metastasis–negative GCs, providing candidate markers related to progression of tumor malignancy. Twenty-two proteins involved in signal transduction controlling cancer cell proliferation/invasion/metastasis were found to be differentially expressed between the gastric tumor and the normal tissue, including Notch4, Akt, and B-catenin, which were upregulated, and cyclin E, p27, E-cadherin, HIF-3a, and NF-kB, which were downregulated. Furthermore, changes in HIF-3a and NF-kB were associated with more invasive/metastatic tumors.
The most extensively studied biomarker of poor prognosis in GC is the HER2 protein. This relationship was first described in 1986 and validated by numerous subsequent studies; however, the association was equivocal until Allgayer and colleagues used directed antibodies against HER2 to demonstrate very high rates of membranous and cytoplasmic HER2 expression in a prospective series of 203 patients with GC. More recently, it has become evident that HER2 overexpression correlates more strongly with gastro-esophageal junction (GEJ) cancers than gastric tumors. The demonstration that HER2-negative GCs correlated with a median survival twice as long as those of HER2-positive cancers (12.9 months vs 6.6 months) highlights the fact that proteomic identification of biomarkers is critical to prognostication and directed future therapies.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
