Separation of neuronal and astrocytic signals for drug toxicity evaluation in the central nervous system

Abstract

In vitro assessment of drug responses within the central nervous system (CNS) is critical for evaluating drug toxicity in neurological disorders. Microelectrode array (MEA)-based evaluation systems commonly employ co-cultures of neurons and glial cells, leading to enhanced neural network maturation. Given the distinctive roles of astrocytes and neurons in CNS drug response, there arises a necessity to distinguish between neuronal and astrocytic functions in co-culture setups. This study aims to address this challenge by employing independent component analysis (ICA) to segregate neural signals originating from neuron and astrocyte co-culture systems on MEA. Rat primary neurons from 17-day pregnant rats and human iPS cell-derived neurons were cultured on MEA plates. Both co-culture and monoculture were established independently for each plate. Following data acquisition, ICA was employed to disentangle the neural signals acquired from the co-culture conditions into their independent components. Each component was categorized based on its source and subsequently separated and reconstructed. Notably, each cell-derived component exhibited characteristics in the low-frequency band. The application of ICA enabled the extraction of individual cell-derived components from co-culture data. Moreover, the extracted data post-ICA exhibited similar trends to monoculture data concerning drug response. This study successfully extracted neuronal and astrocytic signals from co-culture conditions using ICA, offering potential for evaluating drug responses of distinct neuronal cell types.