Abstract
Invasion of tumor cells through the brain remains a significant challenge for treating glioblastoma (GBM). In an effort to better target this phenomenon, we developed a technique for organotypic slice culture of human GBM, which when used in combination with GFP retroviral transduction and time-lapse confocal microscopy allows for observation of migrating tumor cells within an ex vivo three-dimensional architecture. Our prior studies with this model utilized manual tracking to calculate speed and directionality of cell movement. However, this complicated method is labor intensive, requires a skilled analyst, and leads to lack of scalability. Therefore, an automated method for migration analysis would significantly enhance the utility of the slice model platform for larger-scale experimentation. In the current study, we use our prior data to compare our manual tracking paradigm with the automated path tracking platform ImageJ running the ADAPT plugin. Allowing for optimization of parameters critical for cell identification, including GFP fluorescence intensity, object size, and minimum trajectory, ADAPT can detect both changes in membrane morphology and cell migration. Using this comparative approach, we demonstrated that on average ADAPT tracks two-fold more cells with a 97% decrease in time required for analysis. Although absolute migration speeds varied between manual and automated path-tracking methods by approximately 20%, ADAPT analysis of blinded samples reproduced previously identified statistically significant gefitinib-induced decreases in migration speed seen specifically within EGFR amplified tumors. These data suggest potent utility for ADAPT as an unbiased and robust method for analysis of cell migration in patient-derived GBM slice culture.