AI-driven image analysis for label-free quantification of chemotherapeutic cytotoxicity in glial cells

Glioblastoma multiform is a malignant brain tumour associated with poor prognosis. To progress effective chemotherapies more advanced quantitative methods are required to provide insight into cytotoxicity in a non-perturbing manner. Here we demonstrate a robust live-cell imaging assay with label-free analysis to kinetically assess chemotherapeutic cytotoxicity in glial cells. We utilized the Incucyte® AI Cell Health Analysis Software Module which uses trained neural networks enabling label-free analysis of phase contrast images through accurate segmentation of individual cells and Live/Dead classification.To validate AI-driven classification, A172 glioblastoma and BV2 microglial cells were treated with compounds in the presence of Incucyte® Cytotox Green Dye and images were acquired using the Incucyte® Live-Cell Analysis System. Cell death was quantified using both AI Cell Health Live/Dead classification and fluorescence classification of Cytotox positive cells. Time courses and concentration-response curves showed analogous compound efficacies between AI-driven and fluorescence analyses.Three glioblastoma cell lines were treated with a chemotherapeutic panel and efficacy determined for four active compounds. U87-MG exhibited resistance compared to T98G and A172 cells for all compounds. For example, Taxol induced comparable efficacy in A172 and T98G cells with maximal cell death of 73.8 % and 76.6 %, respectively, and identical EC₅₀ values of 3 nM. However, efficacy was reduced in U87-MG cells inducing maximal death of 27.1 % and an EC₅₀ of 14 nM.These data exemplify that AI Cell Health Analysis is a powerful, unbiased approach for quantifying cytotoxicity in glia and is amenable to the screening of therapeutic candidates.