Human iPSC glial mouse chimeras reveal glial contributions to schizophrenia

Genetic studies have suggested a role for glial pathology in the genesis of schizophrenia (SCZ).  To assess the nature of SCZ-associated human glial dysfunction in vivo, we established human glial chimeric mice using glial progenitor cells (GPCs) produced from induced pluripotential cells (hiPSCs), derived from patients with juvenile-onset schizophrenia or healthy controls. To this end, hiPSC GPCs were implanted neonatally into either immunodeficient myelin wild-type mice, in which donor GPCs remained as progenitors or became astrocytes, or into myelin-deficient shiverer mice, in which the GPCs also gave rise to oligodendrocytes. When implanted into shiverers, the SCZ-derived GPCs exhibited less expansion in the white matter than did control GPCs, instead migrating prematurely into the cortex. The SCZ GPC-transplanted shiverers were consequently hypomyelinated relative to control GPC-engrafted mice. When established instead in myelin wild-type hosts, the SCZ hiPSC glial chimeras manifested markedly delayed and diminished astrocytic differentiation, which was associated with diminished prepulse inhibition and an aberrant behavioral phenotype across multiple modalities. Accordingly, RNA-seq revealed significant differences in both glial differentiation-associated and synaptic gene expression by SCZ GPCs. These data suggest a potent contribution of cell-autonomous glial dysfunction to the development of schizophrenia, and provide a model for the in vivo assessment of human glial pathology in this disorder. Overall design: mRNA was isolated by polyA-selection protocol from FACS-sorted PDGFRa-positive GPC cell lines produced from iPS cells derived from 4 schizophrenic patients (designated to SCZ lines 8 [N = 4 independent cell preparations], 29 [N = 3], 51 [N = 7], and 164 [N = 8]) and 3 healthy controls (designated to CTR lines 22 [N = 3], 37 [N = 4], and 205 [N = 7]). FASTQ files in the submission were edited by Trimmomatic. The software was used to trim adapter sequences and low-quality bases (see processing steps). The overall file structure and format remained unchanged. The count matrix in the submission is as obtained directly from featureCounts tool, that is, no manipulation was performed to the count data. The data normalization was performed internally in the analysis and the raw count data was used in differential expression analysis. Please refer to the Materials and Methods section in supplementary files. The complete reproducible workflow, including R scripts, sample information sheet, and count matrix, was deposited to https://github.com/cbtncph/GoldmanetalSCZ2016