A low-frequency haplotype spanning SLX4/FANCP constitutes a new risk locus for early-onset breast cancer (<60 years) and is associated with reduced DNA repair capacity

Only a fraction of breast cancer (BC) cases can be yet explained by mutations in genes or genomic variants discovered in linkage, genome-wide association and sequencing studies. The known genes entailing medium or high risk for BC are strongly enriched for a function in DNA double strand repair. Thus, aiming at identifying low frequency variants conferring an intermediate risk, we here investigated 17 variants (MAF: 0.01-0.1) in 10 candidate genes involved in DNA repair or cell cycle control. In an exploration cohort of 437 cases and 1189 controls, we show the variant rs3810813 in the SLX4/FANCP gene to be significantly associated with both BC (60 years; OR=2.6(1.6-3.9), p=1.6E-05) and decreased DNA repair capacity (60 years; beta=37.8(17.9-57.8), p=5.3E-4). BC association was confirmed in a verification cohort (N=2441). Both associations were absent from cases diagnosed >60 years and stronger the earlier the diagnosis. By imputation we show that rs3810813 tags a haplotype with 5 additional variants with the same allele frequency (R-2>0.9), and a pattern of association very similar for both phenotypes (cases <60 years, p<0.001, the Bonferroni threshold derived from unlinked variants in the region). In young cases (60 years) carrying the risk haplotype, micronucleus test results are predictive for BC (AUC>0.9). Our findings propose a risk variant with high penetrance on the haplotype spanning SLX4/FANCP to be functionally associated to BC predisposition via decreased repair capacity and suggest this variant is carried by a fraction of these haplotypes that is enriched in early onset BC cases.