ABSTRACT: Human induced pluripotent stem cell (hiPSC) technology and microphysiological systems (MPS), such as organoids and spheroids, are key players in the transition toward more human relevant models, offering improved physiological relevance compared to traditional systems. However, their broader adoption critically depends on reproducibility and standardization. Here, we present a multi-site evaluation of the BrainSphere model, a three dimensional central nervous system microtissue derived from hiPSCs, conducted across three international laboratories using five hiPSC lines and two protocols to generate neural progenitor cells, which serve as the starting material for BrainSphere formation. Despite differences in cell line origin, passage number, and laboratory equipment, BrainSpheres consistently developed into mixed neural populations, including neurons, astrocytes, and oligodendrocytes. While single cell RNA sequencing revealed differences in cell type proportions and transcriptional states between lines, immunohistochemistry and proteomics consistently confirmed the presence of these key populations across all sites. Notably, transcriptomic variation across lines highlighted differences in gene expression trajectories, suggesting that cell maturation states may vary depending on passage number, small protocol differences (such as gyratory shaking speed and orbital) or genetic background. Our findings strongly support the BrainSphere model as a robust and reliable tool for neurobiological research. However, to maximize its utility and ensure reproducibility, it is essential to implement standardized practices for hiPSC characterization, BrainSphere generation, and quality management across laboratories.
