Résumé: Successive unexplained shellfish toxicity events have been observed in Arcachon Bay (Atlantic coast, France) since 2005. The positive mouse bioassay (MBA) revealing atypical toxicity did not match the phytoplankton observations or the liquid chromatography-tandem mass spectrometry (LC-MS/MS) investigations used to detect some known lipophilic toxins in shellfish. The use of the three cell lines (Caco2, HepG2, and Neuro2a) allows detection of azaspiracid-1 (AZA1), okadaic acid (OA), or pectenotoxin-2 (PTX2). In this study, we proposed the cell-based assays (CBA) as complementary tools for collecting toxicity data about atypical positive MBA shellfish extracts and tracking their chromatographic fractionation in order to identify toxic compound(s). The present study was intended to investigate the responses of these cell lines to shellfish extracts, which were either control or spiked with AZA1, OA, or PTX2 used as positive controls. Digestive glands of control shellfish were extracted using the procedure of the standard MBA for lipophilic toxins and then tested for their cytotoxic effects in CBA. The same screening strategy previously used with pure lipophilic toxins was conducted for determining the intra- and inter-laboratory variabilities of the responses. Cytotoxicity was induced by control shellfish extracts whatever the cell line used and regardless of the geographical origin of the extracts. Even though the control shellfish extracts demonstrated some toxic effects on the selected cell lines, the extracts spiked with the selected lipophilic toxins were significantly more toxic than the control ones. This study is a crucial step for supporting that cell-based assays can contribute to the detection of the toxic compound(s) responsible for the atypical toxicity observed in Arcachon Bay, and which could also occur at other coastal areas.