Mutation analysis and characterization of ATR sequence variants in breast cancer cases from high-risk French Canadian breast/ovarian cancer families

Abstract

Background: Ataxia telangiectasia-mutated and Rad3-related (ATR) is a member of the PIK-related family which plays, along with ATM, a central role in cell-cycle regulation. ATR has been shown to phosphorylate several tumor suppressors like BRCA1, CHEK1 and TP53. ATR appears as a good candidate breast cancer susceptibility gene and the current study was designed to screen for ATR germline mutations potentially involved in breast cancer predisposition.

Methods: ATR direct sequencing was performed using a fluorescent method while widely available programs were used for linkage disequilibrium (LD), haplotype analyses, and tagging SNP (tSNP) identification. Expression analyses were carried out using real-time PCR.

Results: The complete sequence of all exons and flanking intronic sequences were analyzed in DNA samples from 54 individuals affected with breast cancer from non-BRCA1/2 high-risk French Canadian breast/ovarian families. Although no germline mutation has been identified in the coding region, we identified 41 sequence variants, including 16 coding variants, 3 of which are not reported in public databases. SNP haplotypes were established and tSNPs were identified in 73 healthy unrelated French Canadians, providing a valuable tool for further association studies involving the ATR gene, using large cohorts. Our analyses led to the identification of two novel alternative splice transcripts. In contrast to the transcript generated by an alternative splicing site in the intron 41, the one resulting from a deletion of 121 nucleotides in exon 33 is widely expressed, at significant but relatively low levels, in both normal and tumoral cells including normal breast and ovarian tissue.

Conclusion: Although no deleterious mutations were identified in the ATR gene, the current study provides an haplotype analysis of the ATR gene polymorphisms, which allowed the identification of a set of SNPs that could be used as tSNPs for large-scale association studies. In addition, our study led to the characterization of a novel Delta33 splice form, which could generate a putative truncated protein lacking several functional domains. Additional studies in large cohorts and other populations will be needed to further evaluate if common and/or rare ATR sequence variants can be associated with a modest or intermediate breast cancer risk.

Figure 1

Genomic structure of the human ATR gene. The sizes of exons and introns are represented by proportionaly sized dark boxes and solid lines, respectively. The 5' and 3' untranslated regions are indicated by light boxes in the gene and mRNA structure. The GenBank accession number corresponding to the ATR gene contig, the mRNA and the protein are indicated below the nucleotide and protein sequence. The domains involved in the ATR protein activity are illustrated in the protein structure. All polymorphisms are indicated in open boxes while amino acid nomenclature is also represented.

Figure 2

Pairwise linkage disequilibrium (LD) measures of |D'| for the 41 SNPs identified in our breast cancer cases series. All SNPs are denoted numerically with reference to Table 1.

Figure 3

Panel A: The Table denotes the frequencies, using PHASE, of haplotypes using SNPs having a MAF higher than 5% estimated in controls. Panel B: Haplotype blocks predicted using SNPs identified in control series showing a MAF higher than 5% (17 SNPs). tSNPs identified on a block-by-block basis are denoted with an asterisk (*) above the SNP number. Population haplotype frequencies are displayed on the right of each haplotype combination while the level of recombination is displayed above the connections between two blocks. Thick connections represent haplotypes with frequencies higher than 10% while frequencies below 10% are represented by thin lines.

Figure 4

Panel A: Comparative analysis of splicing site sequences in ATR exon 32, 33, Δ33 and 34 flanking exon-intron junctions as well as prediction of impact of SNP32 (c.5739-4del9+T) on exon 34 acceptor site. Panel B: Comparative analysis of splicing site sequences in ATR exon 41, insΔInt41 and exon 42 flanking exon-intron junctions as well as prediction of impact of SNP37 (c.7041+8G/A) on exon 41 donor site. Splice Site Prediction Program using Neural Network (SSPNN) score values are indicated in parenthesis below each sequence. Exonic nucleotides are represented by uppercase letters while intronic sequences are represented by lowercase letters. m = c or a, r = a or g, y = t or c and n = any nucleotide. N.A. = value not available. Consensus sequences of acceptor and donor sites were described by Burge et al. 1999 [84].

Figure 5

Panel A: Alternative splicing of ATR Δ33 exon. Schematic representation of the design used to assess the existence of mRNA splice transcript encompassing exons 33 and 34 in immortalized cell lines of individuals affected with breast cancer. The putative truncated ATR protein of 1889 amino acids lacks the functional domains identified in the wild type ATR protein. Panel B: Alternative splicing of ATR insΔInt41 exon. Schematic representation of the design used to assess the existence of mRNA splice transcript encompassing exons 41 and 42 in immortalized cell lines of individuals affected with breast cancer. The putative truncated ATR protein of 2350 amino acids lacks a part of the kinase domains identified in the wild type ATR protein.

Figure 6

Expression levels of ATR Δ33 splice form in cell lines and human tissues as measured by quantitative real-time PCR experiments. Panel A. Relative expression levels of Δ33 splice form were calculated as Δ33 splice form/(Δ33 splice form + wild-type allele) in various human tissues and cell lines. Panel B. Standardized expression levels of WT exon 33 were calculated as WT exon 33/(Δ33 splice form + wild-type allele) in various human tissues and cell lines.

Figure 7

Expression levels of Δ33 splice form in homozygote, heterozygote and wild-type individuals for c.5739-4del9+T variant as measured by quantitative real-time PCR experiments. For the 35 wild type individuals, the confidence interval is represented by an open box with the median indicated.