You are here

Tumor Genetics

Although atrophy and intestinal metaplasia correlate with gastric cancer risk, direct cell progression through these stages has not been conclusively shown. Indeed, gastric cancer most likely arises from stem or progenitor cells present within the gastric mucosa rather than directly from terminally differentiated metaplastic cells. Investigators have for several decades sought to unravel the mutations responsible for gastric cancer initiation and progression in an attempt to uncover a logical progression of acquired mutations akin to what is seen in colorectal cancer. However, gastric cancer does not follow a pattern like colorectal carcinoma progression, there is no clearcut linear sequence of mutations in gastric cancers, and there is as an even greater heterogeneity in genetic alterations.17 While initial work has been performed, there remains a need for genome-wide analyses of somatic mutations in gastric cancer, because the precise role, if any, that identified mutations play in initiating malignant transformation, rather than cancer progression, is still not clear.

Aneuploidy is common in gastric cancer (60% to 75%), but cytogenetic studies have failed to identify any consistent chromosomal abnormality. Comparative genomic hybridization studies have shown that chromosome arms 4q, 5q, 9p, 17p, and 18q exhibit frequent decreases in DNA copy number, while chromosomes 8q, 17q, and 20q often have increased DNA copy number.139

There is a general consensus that TP53 is the most commonly mutated gene in gastric cancer (60% to 70% of gastric cancers) and that mutations in Ras, APC, and Myc are rare.140,141 Loss of heterozygosity (LOH) at the APC locus occurs more commonly. Another genetic abnormality found at high frequency (≈60%) is the deletion or suppression of the fragile histidine triad gene (FHIT), a tumor suppressor locus on chromosome 3p. Genes that inhibit entry into the cell cycle, such as p16 and p27, show diminished expression in nearly one half of gastric cancers.142-147 Absence of p27 expression is associated with a poorer prognosis.142,144 Absence of p16 expression is seen most commonly in poorly differentiated carcinomas but has no measurable impact on prognosis.148 Diminished expression of p16 and p27 occurs in the absence of detectable mutations and is believed to be secondary to hypermethylation.146 Many of these cancers show hypermethylation of a number of promoter regions, including the MLH1 promoter region, and show the high-level microsatellite instability (MSI-H) phenotype (see Chapter 1). Multiple tumor suppressor genes have been shown to be methylated in gastric cancers. Emerging evidence suggests that these epigenetic changes, including global hypomethylation and promoter hypermethylation, occurs quite early in gastric carcinogenesis. In addition, it appears that DNA methylation changes also occur in the tumor-associated stromal fibroblasts, suggesting an important role for the tumor microenvironment.149

Overexpressions or amplifications of a number of growth factor pathways has been described, including COX-2 (70%), hepatocyte growth factor/scatter factor (HGF/SF) (60%), vascular endothelial growth factor (VEGF) (50%), c-met (50%), amplified in breast cancer-1 (AIB-1) (40%), and β-catenin (25%) (Table 54-2).150 Approximately 15% of gastric cancers have been reported to overexpress both EGF and EGF receptor (EGFR), consistent with an autocrine mechanism. Mutations in PIC3A, a gene that codes for a catalytic subunit of phosphatidylinositol 3-kinase (PI3K), has been found in up to 25% of gastric cancers analyzed.151 In addition, mutations in genes encoding human protein tyrosine phosphatases (PTPs) were found by the same laboratory in 17% of gastric cancers, with the protein tyrosinase phosphatase receptor type (PTPRT) the most frequently altered.152

Gastric-specific tumor suppressor genes TFF1 (Trefoil factor 1) and RUNX3 (Runt-related transcription factor 3), which have now been identified and may represent “gatekeepers” of the gastric cancer pathway, are logical targets for further study.153,154 Loss of TFF1 has been described in around 50% of gastric carcinomas, and TFF1 knockout mice develop spontaneous gastric antral tumors. Mutations of TFF1 have also been described, and these enhance gastric cancer cell invasion through signaling pathways that include PI3-kinase and phospholipase-C.155 TFF1 expression is repressed by STAT-3, and activation of STAT-3 is also emerging as a key pathway that leads to gastric cancer.39 RUNX3 most likely suppresses gastric epithelial growth by inducing p21 and Bim, attenuates Wnt signaling, and is altered in 82% of gastric cancers.156 Investigations into these genes and their contributions to the gastric cancer phenotype will prove valuable to our understanding of disease progression.

Microsatellite instability (MSI) in dinucleotide repeats secondary to defects in DNA mismatch repair genes, such as MLH1 and MLH2 (mutL homologs 1 and 2), have been implicated in the development of colorectal cancer, and in particular the HNPCC syndrome. Patients with HNPCC have an 11% incidence of gastric cancer, suggesting that MSI may also play a role in the development of gastric cancer.132 MSI is found in 15% to 50% of sporadic gastric cancers, with a higher prevalence in intestinal type of cancers.157-162 Low-level microsatellite activity (e.g., MSI-low) can be found in 40% of areas of intestinal metaplasia in patients with gastric cancer162 and in 14% to 20% of adenomatous polyps.160,162,163 MSI-H occurs in only 10% to 16% of gastric cancers. MSI is associated with the less frequent occurrence of TP53 mutations, well- to moder- ately well-differentiated histology, and distal location in the stomach. Studies that have examined the effect of MSI on patient survival have shown inconsistent results.163,164 When the findings are taken together, it would appear that MSI does play a role in the pathogenesis of gastric cancer, likely before the development of intestinal metaplasia (see Fig. 54-3), and is most commonly due to methylation of the MLH1 promoter.

The data regarding the genetics of diffuse-type gastric cancer are less complete. Mutations in the E-cadherin (CDH1) gene have been linked to the development of this cancer. Several families with hereditary diffuse gastric cancer (HDGC) have been found to carry a germline mutation in CDH1, all with diffuse-type cancer.121-123,165,166 Further evidence supporting a role for E-cadherin in the pathogenesis of gastric cancer comes from studies showing that suppression of E-cadherin expression occurs in 51% of gastric cancers, with a higher percentage found in diffuse-type cancers.167 Furthermore, E-cadherin underexpression is associated with higher rates of lymph node metastases and reduced survival.168,169 The overall rates of CDH1 mutations in gastric cancer are low. Thus, the decreased expression of E-cadherin seen in gastric cancer is likely secondary to hypermethylation of the CDH1 promoter, which occurs in 50% of gastric cancers and 83% of diffuse-type gastric cancers.170 E-cadherin is a transmembrane protein that connects to the actin cytoskeleton through α- and β-catenins to establish cell polarity and mediates homophilic cellular interactions.171,172 Decreased expression of E-cadherin is believed to promote dissociation of cancer cells from their cell matrix, enhancing the migration and invasion of gastric cancer cells. Expression of α-catenin is also decreased or absent in 68% of gastric cancers.173 Therefore, E-cadherin appears to act as a tumor suppressor gene that may be important in the pathogenesis of diffuse gastric cancer.

Perhaps as important as the genetic alterations acquired during the progression to gastric adenocarcinoma is the question, “In what target cells do these changes occur?” In order for a cell to accumulate the quantity of genetic changes necessary for autonomous growth, it must be long lived. For these reasons, the current thinking is that a resident tissue stem cell is the target of genetic mutations and becomes the “cancer stem cell”—capable of autonomous growth and with metastatic potential. Recently, several elegant genetic lineagetracing studies in mice established markers that allow the distinction of 2 different types of GI stem cells. Crypt base columnar cells (CBC) are fast-cycling stem cells expressing Lgr5 and CD133 (Prom-1).174,175 A villin transgene has allowed the identification of a multipotent progenitor located in the lower third of a subset of antral gastric glands,176 whereas multiple intestinal stem cell markers could also be identified in the antrum. Interestingly, Lgr5 shows lineage labeling in some antral gastric glands and in the gastric cardia.175 Slower cycling cells, which are usually found at the +4 position of the crypts of the antrum (i.e., the fourth epithelial cell in the crypt, counting from the bottom of the crypt upward); these lowercycling cells are characterized by a pronounced expression of Bmi1 and Tert.177,178 Although these 2 types of cells are functionally interconnected,179 their exact hierarchical relationship remains to be identified.

In the gastric oxyntic glands, the proliferative zone with the gastric stem cell has been localized to the isthmus, the middle portion of the tubule, and cells are thought to migrate bidirectionally to supply gastric surface mucus cells that coat the gastric pits, and gastric parietal and zymogenic cells that comprise the base of the gland.180 The gastric corpus stem cell has not yet been identified; none of the markers discussed earlier labels any specific cells within the gastric isthmus. Recently, progenitor cells (e.g., Krt19+ and TFF2+ cells) have been shown through lineage tracing studies to label different gastric progenitor cells.181,182 Typically, columnar metaplasia is positive for TFF2 and Krt19. Given that intestinal metaplasia arises in the gastric mucosa and in the esophagus, it is plausible that a similar stem cell gives rise to both. Regardless of their localization (CBC or +4 position) or their function, GI stem cells depend on signals from the stem cell niche, such as pericryptal myofibroblasts and neighboring differentiated epithelia.183 Important signaling pathways required for stem cell maintenance and proliferation comprise the Wnt, Notch, bone morphogenetic proteins (BMP), and Hedgehog pathways.184