Most loci identified by GWAS contain many candidate causal variants (CCVs) with molecular pleiotropy implicating multiple target genes, not all of which may be causal. CRISPR screens using relevant phenotypes can provide an efficient way of identifying causal genes. We previously performed such screens to identify breast cancer risk genes involved in proliferation, tumorigenicity and DNA damage repair. To extend to other cancer hallmarks, we hypothesized that some risk genes may help cancer cells specifically dampen down the T cell mediated antitumor immunity. In order to identify GWAS target genes expressed by a breast cancer cell line (MCF7) which might render it more sensitive or resistant to T cell killing, we established a cytotoxicity assay using HLA-matched HER2-restricted CD8+ T cells. We have performed two CRISPR knockout (ko) and two inhibition (i) screens, for each using a custom library containing guides for predicted breast cancer risk genes, and another with a genome-wide CRISPRko/i library for which the analysis was performed on only the predicted GWAS target genes. In addition to positive controls, we identified 63 candidate breast cancer risk genes that validated (FDR <0.2) in both CRISPRko,or both CRISPRi, screens, including CASP8, CFLAR, IRF1, ATF7IP, CCDC170, CCNDI, ESR1 and GATA3. To determine whether CASP8 expression might be regulated by the single fine-mapped CCV, we cloned a putative enhancer containing rs3769821. The protective allele was significantly more active in a luciferase assay than the risk allele. Our results therefore suggest that one way in which CASP8 acts as a breast cancer risk gene is by conferring resistance to T cell killing when expression levels are reduced. We are now performing additional validations and luciferase assays. In conclusion, CRISPR screens are an efficient method to follow up GWAS findings.