Genetic aetiology of diffuse gastric cancer: so near, yet so far.

H ereditary diffuse gastric cancer (MIM#137215) is an autosomal dominant disease in which gastric cancer develops at a young age. The first report in the medical literature describing familial clustering of gastric cancer was that of Aldred Warthin in 1913, although this family (‘‘family G’’) was subsequently shown to have hereditary non-polyposis colorectal cancer (HNPCC). Further reports of familial predisposition to gastric cancer were, likewise, indirect and based only on clinical and epidemiological observation. Direct proof of a clear molecular basis for diffuse gastric cancer was identified only 6 years ago, when Parry Guilford and colleagues demonstrated germline inactivating mutation of the CDH1 (E-cadherin) gene in a large New Zealand family of Maori ethnicity with early onset, diffuse gastric cancer. Shortly thereafter, it was shown that CDH1 inactivating mutation accounted for a proportion of European families with gastric cancer, a few families of Japanese and Korean ethnicity, as well as one family each of African-American and Pakistani origin. Importantly, all these families have diffuse-type gastric cancer. This specificity of tumour type and association with CDH1 germline mutation led to the designation of hereditary diffuse gastric cancer (HDGC) by the International Gastric Cancer Linkage Consortium (IGCLC). Preliminary analysis of these families has suggested that penetrance for gastric cancer is upwards to 70%. A few studies also suggest an increased risk for colon cancer and lobular breast cancer although this remains open for debate. 10 14 The full-length human CDH1 was cloned and isolated from its location on chromosome 16q22.1 in 1995. It has 16 exons spanning approximately 100 kb of genomic DNA. The gene structure is similar to that of other cadherins. The mature E-cadherin protein consists of three major domains: a large extracellular portion (exons 4–13), which mediates homophilic cellular interactions; and smaller transmembrane (exons 13–14) and cytoplasmic domains (exons 14–16), the latter providing a link to the actin cytoskeleton through an association with various catenins, such as B-catenin. To date, a total of 30 CDH1 germline mutations have been described in HDGC families; 25 have been inactivating (frameshift, nonsense, and splice-site), the remainder are missense. 16 The mutations are distributed equally throughout the gene. Interestingly and relevant to HDGC, mutations in CDH1 are found quite frequently (70%) in sporadic diffuse gastric cancer as well as in the diffuse component (83%) of mixed-type gastric carcinomas. However these are usually missense mutations, which cluster between exons 7 and 9. Somatic CDH1 mutation is also a frequent occurrence in lobular breast cancers but rarely in other tumours, such as glandular/intestinal gastric cancers and ductal breast cancers, although the latter are more common cancers. Because CDH1 acts as a tumour suppressor gene, the formation of HDGC tumours requires bi-allelic inactivation, which in at least one-half of cases is accomplished by promoter methylation of the wild-type allele. The finding of a somatic and germline CDH1 mutation exclusively in the diffuse and not in the glandular/intestinal and solid type of gastric carcinoma illustrates the importance of distinguishing these subtypes within the clinical setting. This distinction was the basis for development of Consortium criteria by the IGCLC to screen for families with diffuse gastric cancer prior to performing linkage studies and mutation analysis. The current clinical criteria are: (1) two or more documented cases of diffuse gastric cancer in first/second degree relatives, with at least one diagnosed before the age of 50, or (2) three or more cases of documented diffuse gastric cancer in first/second degree relatives, independently of age of onset. When gastric cancer families are ascertained strictly by the Consortium criteria, the CDH1 mutation frequency approximates 25–30%. To extend these observations, BrooksWilson and colleagues in this issue of the Journal of Medical Genetics sought germline mutations in the CDH1 gene in 43 new families with hereditary gastric cancer. In 42 of these families, there was at least one case of pathologically confirmed diffuse gastric cancer. The families were from Canada, the United States, and the United Kingdom. Their intention was to further develop evidence-based guidelines for performing CDH1 genetic screening and to examine other cancer risks, specifically risks for breast cancer associated with germline mutation. Importantly, some of these families were ascertained using the set IGCLC criteria, others through more relaxed criteria requiring only one documented case of diffuse gastric cancer in the family. Each proband had thorough genomic sequencing for germline CDH1 gene mutation of the 16 coding exons, including exon/intron boundaries by bi-directional sequencing. Novel mutations were found in 13/43 of these families. Most were inactivating mutations, which included two small insertions, five deletions, two splice site substitutions, and one complex deletion/insertion involving a splice site. Non-conservative missense mutations were identified in three families. For the sake of completeness, functional analysis was performed through in vitro assays on each of the missense alterations to prove pathogenicity. These findings clearly illustrate that families with diffuse gastric cancer and associated CDH1 mutation are those with strong family histories of early onset diffuse gastric cancer. The converse also appears to be true—single affected individuals, even when diagnosed under age 50, are less likely to have CDH1 mutation. This is highly relevant for clinical practice and will allow clinicians to incorporate the revised criteria to ascertain families most appropriate for CDH1 mutation screening. Thus for families with two or more cases of gastric cancer, with at least one affected familymember having documented diffuse gastric cancer occurring before age 50, 48% will have CDH1mutation. These results support the use of these revised criteria in appropriate assessment for these high risk families. However, this does not answer the question of genetic aetiology for the remainder of these HDGC families. Collectively, this and previous studies suggest other genes which predispose to HDGC. For example, a separate familial syndrome which presents with hyperplastic gastric polyps and diffuse gastric carcinoma appears to be not genetically linked to CDH1. 20 A subset of diffuse