Introduction:
The lack of high quality predictive models for drug-induced QT prolongation continues to be a significant problem in pharmaceutical development. While human pluripotent stem cell derived-cardiomyocytes (hPSC-CMs) hold promise to be a valuable tool for drug discovery, efforts have been frustrated by the labor-intensive nature of electrophysiological recordings and the heterogeneity of hPSC-CMs populations.
Methods:
Using lentivirus, we introduced the genetically encoded fluorescent voltage reporter, A242-Arclight, into hPSC-CM monolayers in multi-well plates. An inverted fluorescence microscope was fit with an environmentally controlled enclosure and automated stage. High speed imaging with a Photometrics Evolve 128 EMCCD camera was performed at baseline and after administration of test compounds. Optical traces were processed using a custom program and composite AP durations, APD80, were compared before and after drug application (Figures A & B).
Results:
Baseline APD80 values displayed high degree of consistency between wells: 483±59 msec. High-throughput data acquisition demonstrated dose dependent APD80 increases from all QT-prolonging agents tested as well as dose dependent APD80 decrease from pinacidil. In contrast, negative control compounds caused no significant changes in APD80. Results from a representative plate are shown (Figure C).
Conclusions:
Optical measurements provide rapid recordings of drug-induced AP duration changes, and offer a strategy to non-invasively screen hPSC-CMs in high-throughput. Recording from cell monolayers as opposed to single cells and using paired comparisons may be beneficial in addressing the heterogeneity amongst hPSC-CM preparations.
Improved method of structure-based virtual screening based on ensemble learning
