Résumé : The objective of the present study was to develop a cellular phenotype assay for nucleotide excision repair (NER), using benzo[a]pyrene diol epoxide (BPDE) as model mutagen. Since in vitro exposure to BPDE may lead to DNA strand breaks resulting from both direct interaction with DNA and incisions introduced by the repair enzymes, we aimed to discriminate between both types of breaks using the comet assay and quantified the DNA strand breaks after in vitro challenge of peripheral blood mononucleated cells (PBMCs) with BPDE in the presence or absence of the DNA polymerase inhibitor aphidicolin (APC). The assay was performed with a low (0.5 μM) and a high (2.5 μM) BPDE concentration. The individual NER capacity was defined as the amount of DNA damage induced by BPDE in presence of APC, diminished with the damage induced by BPDE and APC alone. First, the assay was applied to a NER-deficient human fibroblast cell line (XPA-/-) to validate the methodology. Lower repair capacity and a higher amount of BPDE-induced DNA adducts were observed for the XPA-/- fibroblasts as compared to the wild-type fibroblasts. Repeated experiments on PBMCs from four donors showed low intra-individual, intra-experimental and inter-assay variation for both concentrations, indicating the reliability of the method. To assess the inter-individual variation, the assay was applied to PBMCs from 22 donors, comparing the repair capacity after exposure to 0.5 μM (N=10) and 2.5 μM (N=12) BPDE. The repair capacity showed a higher inter-individual variation as compared to the intra-individual variation. Moreover, this difference was more pronounced using the low concentration. All these results indicate the adequacy of the method using this low concentration. Further improvement, however, should be recommended by applying the study with low BPDE concentration in a larger population and taking into account the relevant genotypes for NER.