We have made an evaluation of mutation detection techniques for their abilities to detect mosaic mutations. In this study, Sanger sequencing, single-strand conformation polymorphism (SSCP)/heteroduplex analysis (HD), protein truncation test (PTT), and denaturating high-performance liquid chromatography (DHPLC) were compared with parallel sequencing. In total DNA samples from nine patients were included in this study. Mosaic mutations were artificially constructed from seven of these samples, which were from heterozygote mutation carriers with the mutant allele present at 50%. The mutations analyzed were as follows: c.646C>T, c.2626C>T, c.2828C>A, c.1817_1818insA, c.2788dupA, c.416_419delAAGA, and c.607delC in the APC gene. The lowest degree of mutant alleles detected with SSCP/HD and DHPLC varied between 5% and 25%, and between 15% and 50% for Sanger sequencing. Three of the mutations were analyzed with PTT with considerable variations in detection levels (from 10 to 100%). Using parallel sequencing a detection frequency down to 1% was reached, but to achieve this high sensitivity sufficient coverage was required. Two patients with natural mosaic mutations were also included in this study. These two mutations had previously been identified with Sanger sequencing (NF2 c.1026_1027delGA) and SSCP/HD (APC c.2700_2701delTC). In conclusion, all the evaluated methods are applicable for mosaic mutation screening even though combinations of the conventional methods should be used to reach an adequate sensitivity. Sanger sequencing alone is not sensitive enough to detect low mosaic levels. Parallel sequencing seems to be the ultimate choice but the possibilities to use this technique is today limited by its complexity, economics, and availability of instruments.

Familial adenomatous polyposis (FAP) is caused by germline mutations in the adenomatous polyposis coli (APC) gene. Two promoters, 1A and 1B, have been recognized in APC, and 1B is thought to have a minor role in the regulation of the gene. We have identified a novel deletion encompassing half of this promoter in the largest family (Family 1) of the Swedish Polyposis Registry. The mutation leads to an imbalance in allele-specific expression of APC, and transcription from promoter 1B was highly impaired in both normal colorectal mucosa and blood from mutation carriers. To establish the significance of promoter 1B in normal colorectal mucosa (from controls), expression levels of specific transcripts from each of the promoters, 1A and 1B, were examined, and the expression from 1B was significantly higher compared with 1A. Significant amounts of transcripts generated from promoter 1B were also determined in a panel of 20 various normal tissues examined. In FAP-related tumors, the APC germline mutation is proposed to dictate the second hit. Mutations leaving two or three out of seven 20-amino-acid repeats in the central domain of APC intact seem to be required for tumorigenesis. We examined adenomas from mutation carriers in Family 1 for second hits in the entire gene without any findings, however, loss of the residual expression of the deleterious allele was observed. Three major conclusions of significant importance in relation to the function of APC can be drawn from this study; (i) germline inactivation of promoter 1B is disease causing in FAP; (ii) expression of transcripts from promoter 1B is generated at considerable higher levels compared with 1A, demonstrating a hitherto unknown importance of 1B; (iii) adenoma formation in FAP, caused by impaired function of promoter 1B, does not require homozygous inactivation of APC allowing for alternative genetic models as basis for adenoma formation.

Somatic mutations in the POLE gene encoding the catalytic subunit of DNA polymerase ɛ have been found in sporadic colorectal cancers (CRCs) and are most likely of importance in tumour development and/or progression. Recently, families with dominantly inherited colorectal adenomas and colorectal cancer were shown to have a causative heterozygous germline mutation in the proofreading exonuclease domain of POLE. The highly penetrant mutation was associated with predisposition to CRC only and no extra-colonic tumours were observed. We have identified a mutation in a large family in which the carriers not only developed CRC, they also demonstrate a highly penetrant predisposition to extra-intestinal tumours such as ovarian, endometrial and brain tumours. The mutation, NM_006231.2:c.1089C>A, p.Asn363Lys, also located in the proofreading exonuclease domain is directly involved in DNA binding. Theoretical prediction of the amino acid substitution suggests a profound effect of the substrate binding capability and a more severe impairment of the catalytic activity compared to the previously reported germline mutation. A possible genotype to phenotype correlation for deleterious mutations in POLE might exist that needs to be considered in the follow-up of mutation carriers.