Functional analysis of genetic variants in putative low penetrate breast cancer genes

Sammanfattning: Paper I & II CDH1 germline mutations predispose individuals to diffuse gastric cancer, but its role in breast cancer is less clear. Somatic CDH1 mutations were reported to be frequent in lobular tumours (ILC), but they have not been found in ductal carcinomas (IDC). To define the role of CDH1 in breast cancer, we used denaturing high performance liquid chromatography to screen a series of breast cancer samples for mutations. Somatic mutations were detected in 4 of 83 IDC (5%) and 3 of 25 ILC (12%). No germline mutation was found in 19 familial breast cancer patients that showed loss of heterozygosity (LOH) at 16q24, in 12 cases from 10 families with breast, gastric and colon cancer and in 13 familial lobular breast tumours. Another somatic mutation was detected in one of the familial breast cancer patient with both ductal and lobular foci. Putative breast cancer risk conferred by the promoter polymorphism -161C-A of CDH1 or 1774G-A (Ala592Thr) was also analyzed in case-control studies. No significant difference in allelic frequency was found between the breast cancer patients and controls for either polymorphism. A novel promoter polymorphism was identified at position -152 with a similar frequency of the rare C allele in both breast cancer patients and controls. Transient transfection assay using constructs containing -16IC/-152C or -161A/-152T showed only a slight decrease of the transcription activity as compared to the wild type constructs carrying -161C/152T. We conclude that CDH1 is not a prominent lowpenetrance gene in breast cancer, but CDH1 mutations contribute to the progression of both lobular and ductal tumours. Paper III BACH1, a gene located in 17q22 and encoding a protein directly interacting with BRCA1, was suggested to be a candidate gene for breast cancer susceptibility. Using PCR-SSCP, we screened for germline BACH] mutations with 29 breast cancer families linked to 17q22 and additional 95 familial breast cancer cases, which were all without detectable BRCA112 mutations. No mutation was found. A C/T polymorphism at position 517 was detected, which leads to Arg173Cys substitution in the putative nuclear localization sequence. This alteration may contribute to the development of breast cancer, but BACH1 is not a major breast cancer gene. Paper IV ATM has been suggested to act as a tumour suppression gene that requires the inactivation of both alleles in the development of a malignancy. Two mutations designated as T7271G and IVS1O+6 T-G were reported to increase breast cancer risk in multiple-case families in a dominant negative manner. We evaluated the population frequency of these two mutations in Sweden and Czech populations using PCR-RFLP. The mutation T7271G was not detected, mutation IVS 10+6 T-G was found in 2 of 768 cases and I in 557 controls, giving the allelic frequency of 0. 1%. We also tested the hypothesis that ATM mutations would be enriched in breast cancer patients with LOH at 11q22-23 if ATM acted as TSG. Forty-two selected DNA samples from breast cancer cases were screened for mutations in ATM from exons 50 to 66 using PCR-SSCP, but no mutation was detected. Paper V (manuscript) LST1 is a gene located in the TNFalpha region and producing many isoforms with a possible role in immune response. In an attempt to understand the mechanism of the 3' splicing site selection in LST1, we found that removal of a nearby AG dinucleotide repeat TNFd led to a serine-arginine protein-dependent activation of a cryptic 3' splice site. The highest yield of the novel isoform named LST1/n was induced by ASF and SRp40. This effect was dose-dependent, it was also dependent on the length of TNFd. Deletion of RNA recognition motif 2 (RRM2) of ASF/SF2 abolished the usage of the cryptic splice site, indicating that the activation is mediated by RNA binding. In addition, we showed that hnRNP Al but not hnRNP K could effectively antagonize SR-protein induced activation. Replacement of TNFd with AC or AT repeat indicated a AG-AC>AT repressing hierarchy on the SR protein induced activation. Further removal of a putative splicing silencer A3 alleviated the requirement for exogenous SR proteins in LST/n activation. We also showed that lack of this event in the wild type construct was not due to the differential efficiency in RTPCR for TNFd+ or TNFd- isoform. Insertion of TNFd in a heterologous context showed no evidence of splicing inhibition in in vitro splicing analysis. Insertion of A3 into a middle exon of a heterologousACE Alu + 136/3'ss minigene led to almost full exon skipping, suggesting that A3 contains splicing inhibitory sequences. All these data are consistent with the existence of multiple mechanisms in the repression of pseudo splice sites and with the concept that dinucleotide repeats could act as splicing regulatory elements in the vicinity of splice sites.