Friedel Nollet

We have analysed a series of 49 human breast cancers for mutations in the entire coding region plus flanking intron sequences of the E-cadherin gene. The tumours included 41 infiltrating lobular carcinomas, two infiltrating ducto-lobular carcinomas and six infiltrative ductal carcinomas. In the lobular carcinomas 23 different somatic mutations were detected, of which seven were insertions, 11 deletions, two nonsense mutations and three splice site mutations. The other tumours showed no detectable E-cadherin mutations. All the frameshift and nonsense mutations are expected to generate a secreted E-cadherin fragment instead of a transmembrane protein with cell adhesion activity. The majority of the mutations (21 of 23) were found in combination with loss of heterozygosity of the wild type E-cadherin locus (16q22.1), a hallmark of classical tumour suppressor genes. The mutations were scattered over the whole coding region and no hot spots could be identified. All mutations described here were previously unreported. In conclusion, we have identified up to now E-cadherin mutations in 27 of 48 (56%) infiltrating lobular breast carcinomas and in 0 of 50 breast cancers of other histopathological subtypes. These data provide strong evidence that frequent E-cadherin mutations are involved in the particular etiology of sporadic lobular breast cancers.

Loss of serine or threonine phosphorylation sites from exon 3 of beta-catenin has been identified in approximately half of colorectal tumors which lack adenomatous polyposis coli (APC) mutations, but the overall contribution of beta-catenin mutations to sporadic colorectal tumorigenesis is unclear. We therefore used PCR to amplify and sequence exon 3 of beta-catenin from 202 sporadic colorectal tumors. Exon 3 beta-catenin mutations were identified in 6 of 48 small (< 1 cm) adenomas, 2 of 82 large (> or =1 cm) adenomas, and 1 of 72 invasive carcinomas. Eight of the nine mutations, including all of those in the small adenomas and the invasive cancer, involved loss of serine or threonine phosphorylation sites. The percentage of beta-catenin mutations in small adenomas (12.5%) was significantly greater than that in large adenomas (2.4%) and invasive cancers (1.4%; P = 0.05 and P = 0.02, respectively). We conclude that mutation of beta-catenin can be an early, perhaps initiating, event in colorectal tumorigenesis. Small adenomas with beta-catenin mutations do not appear to be as likely to progress to larger adenomas and invasive carcinomas as other adenomas, however, with the result that beta-catenin mutations are only rarely seen in invasive cancers. This suggests that APC and beta-catenin mutations are not functionally equivalent, and that the APC gene may have other tumor suppressor functions besides the degradation of beta-catenin.