How reliable are in vitro IC50 values? Values vary with cytotoxicity assays in human glioblastoma cells.

Increasing evidence shows that discrepancies exist among in vitro cytotoxicity methods resulting in unreliable drug toxicity profiles. This is particularly criticial for cell lines such as gliomas which are histologically and genetically heterogeneous. The high level of variation in these cells makes comparative analysis difficult and is a severe limitation for the usefulness of high-throughput screening methods. Here we examine variations between four conventional in vitro cytotoxicity assays (MTT, Alamar Blue, Acid Phosphatase and Trypan Blue) for assessing the viable cell number following treatment of two human glioblastoma cell lines (U87MG and U373MG) with different chemical agents (carboplatin, etoposide, paraquat). The variations in IC50 values between the four assays suggest that even when combining several endpoints such as mitochondrial function, lysosomal activity, and membrane integrity, a reliable and uniform toxicity profile was not achieved. Because of these variations between cytotoxicity assays using compounds with varying mechanisms of cytotoxicity, then it is possible that the true IC50 value of valuable and beneficial compounds for glioblastoma may have been missed through over/underestimation. This highlights the importance of reliability and accuracy in pre-animal models such as in vitro models of cytotoxicity for better predictive in vivo responses.