The estimated heritability of schizophrenia is 80–85%, and if one of monozygotic twins has the illness, the risk for the other twin is about 50%. It has remained unclear which factors are associated just with the increased risk, and which neurobiological mechanisms are related to the cascade leading to actual onset of illness. It has also remained a mystery why this neurodevelopmental disorder manifests itself after adolescence. We used iPSC-derived neurons from 5 monozygotic twin pairs discordant for schizophrenia (3 female and 2 male pairs; all 5 affected individuals were chronically symptomatic and treated with antipsychotics, 3 with clozapine) and 6 healthy controls matched by age and sex to enhance signal by minimizing genetic heterogeneity. Diagnoses were assigned using DSM-IV criteria and SCID-I interviews, and severity of symptoms was assessed with PANSS. Skin biopsy-derived fibroblasts were reprogrammed into integration-free induced pluripotent stem cell (iPSCs) lines by Sendai virus technology. The cells were further differentiated into cortical-like neurons expressing markers of GABAergic and glutamatergic neurons. The top genes from RNAseq analysis were validated by quantitative polymerase chain reaction (qPCR). In the proteomic analysis, the 16 samples were processed through the SysQuant workflow using Tandem Mass Tag (TMT) reagents within two TMT 10plexes. A reference pool containing all samples was also included. To identify statistically significant regulated features (peptide, phosphopeptides or proteins), a LIMMA-based modified t-test was performed at the peptide, phosphopeptide and protein levels in pairwise comparisons using distribution-dependent fold change (FC) and p-value (p) thresholds. Transcriptomic and proteomic analyses showed very large effect sizes (differences up to 7.3-fold in gene expression and 2.7-fold in protein levels in the comparison between affected twin versus healthy twin). Among the most robust findings related to disease status were down-regulation of COL6A3, SSTR2, and LHX1 genes, and altered glycosaminoglycan, CAMK2G-related AMPA/NMDA/glutamate, GABAergic synapse, and purine metabolism pathways. While only 12% of genes were differentially expressed between healthy males and females, up to 61% of the illness-related genes were sex-specific (p-values down to 2.3 x 10E-295 for the difference). Most of the genes with the largest effect sizes were different between males and females. At gestational age of 90 days, neurons of affected twins showed altered calcium responses for NMDA-specific glutamate (p < 0.001) and for GABA exposure compared with their healthy twins, and these alterations were normalized when the neurons were treated with clozapine. Results imply that somatostatin 2 receptor defect may contribute to dysfunction of NMDA receptors in somatostatin/calreticulin GABAergic interneurons, and that clozapine might have a beneficial effect on this cascade. Consistent findings on down-regulation of N-glygan and calnexin/calreticulin pathways, affecting protein folding in endoplastic reticulum, indicate also their major role in schizophrenia. Our results imply that although both sexes share many of the final common pathways involving the same proteins, the underlying primary pathophysiology of schizophrenia may differ between males and females. This may explain why the disease typically manifests after adolescence, when the expression levels of many sex-specific genes change.
Read full abstract