The risk of prostate cancer is known to be elevated in carriers of germline mutations in BRCA2, and possibly also in carriers of BRCA1 and CHEK2 mutations. These genes are components of the ATM-dependent DNA damage signalling pathways. To evaluate the hypothesis that variants in ATM itself might be associated with prostate cancer risk, we genotyped five ATM variants in DNA from 637 prostate cancer patients and 445 controls with no family history of cancer. No significant differences in the frequency of the variant alleles at 5557G4A (D1853N), 5558A4T (D1853V), ivs38-8t4c and ivs38-15g4c were found between the cases and controls. The 3161G (P1054R) variant allele was, however, significantly associated with an increased risk of developing prostate cancer (any G vs CC OR 2.13, 95% CI 1.17 -3.87, P ¼ 0.016). A lymphoblastoid cell line carrying both the 3161G and the 2572C (858L) variant in the homozygote state shows a cell cycle progression profile after exposure to ionising radiation that is significantly different to that seen in cell lines carrying a wild-type ATM gene. These results provide evidence that the presence of common variants in the ATM gene, may confer an altered cellular phenotype, and that the ATM 3161C4G variant might be associated with prostate cancer risk. Prostate cancer is the second most common malignancy and the second commonest cause of cancer deaths in men in the European Union, with 143 000 new cases and 60 000 deaths year À1 (GLOB-CAN 2000, www-dep.iarc.fr). The aetiology of prostate cancer is poorly understood. Prostate cancer is known to aggregate in families, indicating that genetic susceptibility may be important, but the genes involved are largely unknown. Linkage studies in multiple case families have suggested susceptibility loci on chromosomes 1q24, 1q42, 1p36, 8p22 -23, 17p, 20q13 and Xq (see recent reviews by DeMarzo et al, 2003;Gronberg, 2003) but none have been definitively replicated. As a consequence of these linkage studies, variants in prostate cancer families have been identified in several genes including Macrophage Scavenger Receptor 1(MSR1), 2 0 -5 0 -oligoadenylate-dependent ribonuclease L (RNASEL) and ELAC2 (chromosome 17p11/HPC2 region) (reviewed in Simard et al, 2003), but again none have been reliably associated with risk.Several independent studies have demonstrated that individuals with germline mutations in BRCA2 are at increased risk of prostate cancer (The Breast Cancer Linkage Consortium, 1999;Edwards et al, 2003;Giusti et al, 2003). There is also some evidence for an increased risk in carriers of BRCA1 mutations (Thompson et al, 2002). More recently, Seppala et al (2003) have found that the CHEK2 variant 1100delC, known to be associated with an increased risk of breast cancer, is also associated with an increased risk of prostate cancer, and Dong et al (2003) found that this and other missense variants in CHEK2 occurred at increased frequency in prostate cancer cases. The proteins encoded by the BRCA1 and BRCA2 genes participate in the maintenance of geno...