BioProject

Developing platforms for ex vivo studies of human brain development and formulation of therapeutic strategies for a variety of conditions and diseases has been hampered by limited access to relevant human tissue and the necessity of using rodent models that imperfectly recapitulate human brain physiology. A promising advance is brain organoids that, to a greater extent than monolayer or spheroid cultures, recapitulate to varying extents the patterns of cell differentiation and morphological development characteristic of human brain. Here, we present an organoid-based research platform initiated using L-MYC immortalized human fetal neural stem cells (NSCs) (LMNSC01), and) and grown in a physiological 4% oxygen environment similar that in developing brain. Expanded LMNSC01 cells maintain genomic stability and multipotency, and are non-tumorigenic. Using authentic human NSCs as starting material eliminates the ex vivo programming and reprogramming required to achieve directed differentiation of human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs). LMNSC008 brain organoids were characterized using NanoString technology to profile gene expression for over 120 days in vitro and comparing these patterns to these same genes during normal brain development (BrainSpan database). We also visualized by immunofluorescence changes in expression and distribution of genes representative of these developmental processes along with morphological changes occurring over this same period. We observe parallel developmental patterns of gene expression between organoids and developing cortex for pathways involved in neuronal cytoskeleton, neuron-glia interaction, neural connectivity, neurotransmission, metabolism, axon and dendrite structure, tissue integrity, angiogenesis, and myelination. We suggest that LMNSC01 organoids offer an alternative to the challenges of iPSC programming and neural induction. Notable properties of this platform are its initiation with a line of authentic neural stem cells (LMNSC01), the consistency of the organoids produced, and favorable comparison of their gene expression patterns with those observed during normal fetal development. Overall design: RNA expression was analyzed using the NanoString nCounter platform (NanoString Research) by digitally detecting and counting in a single reaction without amplification. We utilized the NanoString Neuropathology Profiling panel, consisting of genes targeting six fundamental themes: neurotransmission, neuron-glia interactions, neuroplasticity, cell structure integrity, neuroinflammation, and metabolism (Supplemental Table 1). Each assay also includes six positive and eight negative mRNA assay controls, plus ten housekeeping mRNA controls. RNA was hybridized with the gene panel Codeset at 65 oC for 16 hr. The post-hybridization probe-target mixture was quantified using the nCounter Digital Analyzer, with subsequent data analysis performed using nSolver. Raw data was first normalized to internal positive and negative controls to eliminate variability unrelated to the samples, then normalized to the selected housekeeping genes using Geometric Means methods. Gene expression variability of less than 20% within a single organoid was the criterion for acceptance. Replicates of 4-10 organoids were used for each timepoint.   mRNA extraction and preparation for NanoString was done at the COH Pathology Core. Briefly, mRNA was extracted using the miRNeasy mini kit (HB-1277-066, Qiagen). Concentration was assessed by Nanodrop spectrophotometer ND-1000 and Qubit 3.0 Fluorometer (Thermo Scientific), with mRNA fragmentation and quality assessed by 2100 Bioanalyzer (Agilent). NanoString data were compared against a reference database for expression of these same genes during normal fetal development drawn from the BrainSpan Atlas of the Developing Human Brain (https://www.brainspan.org/). To facilitate comparisons, organoid and fetal brain datasets were both normalized to expression of the same NanoString internal reference genes (Supplementary Table 1) using Singscore(Foroutan, Bhuva et al. 2018) (PMID: 30400809), with the score indicating deviation from these genes. Singscore ranks gene expression, and calculates a non-parametric score for individual genes in each gene set (NanoString or BrainSpan). This rank-based approach is robust to differences in the underlying data distribution, enabling scoring of gene sets across different transcriptomic datasets and for assessment of gene set activity in individual samples. In the single-sample presentation employed here, scores range between 0.0 and 1.0 and represent the extent to which each gene in the gene set is expressed compared to the distribution of reference genes, and scored as higher (score > 0.5) or lower (score < 0.5).