Article authored by Qichuan (Brent) Chen, Genetics Program Graduate Ph.D. Student
On September 18th, Dr. Daniel Weinberger, director and CEO of the Lieber Institute for Brain Development and Professor of Psychiatry, Neurology, and Neuroscience at the Johns Hopkins School of Medicine, presented a seminar entitled “Genomic Insights into Developmental Origins of Schizophrenia.” This seminar was jointly hosted by the W. M. Keck Center for Behavioral Biology and the Program in Genetics. Dr. Weinberger’s group at the Lieber Institute is interested in understanding the early developmental origins of
Dr. Weinberger’s group at the Lieber Institute is interested in understanding the early developmental origins of adult-onset disorders such as schizophrenia. He postulated that perhaps most major psychiatric disorders are due to early developmental issues and that these conditions are potentially states of brain development instead of illnesses. His group asked whether genes related to neurodevelopmental disorders are preferentially expressed during fetal life. To test this hypothesis, his group compared expression of genes during fetal life and postnatal life in patients with many developmental disorders, such as syndromic neurodevelopmental disorders, autism, and intellectual disability. Specifically, they looked for correlations between genes that are more highly expressed during fetal life than in postnatal life in 2,200 brains collected and characterized by an oligonucleotide-based gene expression array called BrainCloud. They found that, in the prefrontal cortex, genes associated with syndromic neurodevelopmental disorders, autism, and intellectual disability all were preferentially enriched in the prenatal transcriptome. Genes associated with schizophrenia, in general, were not enriched in the prenatal transcriptome, but genes associated with schizophrenia located in copy number variation regions were enriched in the prenatal transcriptome. Conversely, late-age neurodegenerative disorders showed enrichment in the postnatal transcriptome suggesting that this approach plausibly allows for the determination of which genes are involved in various psychiatric disorders in early developmental and late-onset disorders even at a low resolution.
To further refine this analysis, Weinberger and his group performed deep RNA-seq at 100 million paired-end reads per brain sample and quantified junction reads to ensure full coverage of transcript isoforms. They found about 50,000 differentially expressed regions (DER) that were relevant for development and aging. Of these DERs, 40% were mapped to non-exonic sequences. For schizophrenia, autism, and intellectual disability, these genes were more highly expressed in fetal life. Further, developmental DERs were enriched with positive regions from genome-wide association studies. Weinberger also identified isoform switches during brain development, such that one isoform with a specific junction would be highly expressed in the fetal stage while another isoform that did not contain the junction was highly expressed at the postnatal stage. Gene-level analysis without quantifying junction reads would not have identified a developmental pattern in these isoform switch genes.
Next, they mapped DNA methylation across development to understand changes in the epigenetic landscape. They looked at about 230,000 differentially methylated regions and found significant enrichment of CpGs that were more highly methylated in fetal life compared to postnatal life within risk regions of the genome associated with schizophrenia. They also examined the DNA methylation landscape in the brains of schizophrenia patients compared to normal brains to determine if the enriched epigenetic landscape changed due to the marks made in early development or due to marks made near the time of diagnosis. They found that 2,100 differentially methylated CpG regions associated with schizophrenia were enriched for the fetal/postnatal transition and were depleted of genomic risk variants identified in clinical studies. This suggests that environmental events during early brain development play an enduring role in the development of neurodevelopmental disorders and that some variation in risk is due to environmental sensitivity. With that idea, Weinberger focused on understanding the relationship between environmental sensitivity and risk for schizophrenia. Current research has shown that schizophrenia risk has a polygenic architecture, is significantly associated with early life environment, and that there is a higher incidence of schizophrenia in males than in females. Weinberger explained that even though both dizygotic twins and siblings share 50% of alleles, the relative risk of schizophrenia is twice as high in dizygotic twins. The only real difference between siblings and dizygotic twins is the intrauterine environment. Weinberger also stated that 97% of schizophrenia cases are explained by common variants and thus prediction is poor at the individual variant level. However, he added up all the weighted risk-associated loci from a genome-wide association study to develop a Polygenic Risk Score (PRS). This showed to be more informative and a more meaningful predictor of risk. In a dataset from the United States, Weinberger and his colleagues showed that patients with schizophrenia have much higher PRS scores. Remarkably, there was an interaction between patients that had experienced complicated pregnancies and the PRS score. Among loci associated with risk for schizophrenia, 108 were implicated in environmental stress pathways. The only real significance in PRS score differences between normal patients and schizophrenia patients was related to complications during pregnancy.
With that idea, Weinberger focused on understanding the relationship between environmental sensitivity and risk for schizophrenia. Current research has shown that schizophrenia risk has a polygenic architecture, is significantly associated with early life environment, and that there is a higher incidence of schizophrenia in males than in females. Weinberger explained that even though both dizygotic twins and siblings share 50% of alleles, the relative risk of schizophrenia is twice as high in dizygotic twins. The only real difference between siblings and dizygotic twins is the intrauterine environment. Weinberger also stated that 97% of schizophrenia cases are explained by common variants and thus prediction is poor at the individual variant level. However, he added up all the weighted risk-associated loci from a genome-wide association study to develop a Polygenic Risk Score (PRS). This showed to be more informative and a more meaningful predictor of risk. In a dataset from the United States, Weinberger and his colleagues showed that patients with schizophrenia have much higher PRS scores. Remarkably, there was an interaction between patients that had experienced complicated pregnancies and the PRS score. Among loci associated with risk for schizophrenia, 108 were implicated in environmental stress pathways. The only real significance in PRS score differences between normal patients and schizophrenia patients was related to complications during pregnancy. Taking into account pregnancy complications improved prediction of schizophrenia from ~1% accuracy to ~10% accuracy. They recapitulated similar findings in a Japanese dataset as well as in a German dataset. To further understand the role of complicated pregnancies on schizophrenia, Weinberger examined the role of the uterine environment and placental genes. Are schizophrenia risk genes enriched in placental samples from complicated pregnancies? His team looked for schizophrenia risk genes that were differentially expressed in the placentas of complicated pregnancies. They found that differentially expressed genes associated with schizophrenia risk were implicated in the interaction between schizophrenia and complicated pregnancy. Additionally, GOTERM analysis showed the strongest category in which those genes were represented were oxidative stress and hypoxic stress. This finding potentially suggests that stress response genes could predispose to schizophrenia through inflammation. Thus, Weinberger examined 23 canonical inflammation pathways and found them to be co-expressed with the schizophrenia risk genes that were differentially regulated in the placenta. Lastly, he wanted to look at the causes of greater prevalence of schizophrenia in males and to investigate a potential link between the placental environment and increased risk of schizophrenia in males. He found in the placenta samples higher upregulation of the differentially expressed schizophrenia genes in males compared to females. Weinberger concluded that a focus on prenatal health may be the best way to prevent schizophrenia in individuals with high genetic risk.
To further understand the role of complicated pregnancies on schizophrenia, Weinberger examined the role of the uterine environment and placental genes. Are schizophrenia risk genes enriched in placental samples from complicated pregnancies? His team looked for schizophrenia risk genes that were differentially expressed in the placentas of complicated pregnancies. They found that differentially expressed genes associated with schizophrenia risk were implicated in the interaction between schizophrenia and complicated pregnancy. Additionally, GOTERM analysis showed the strongest category in which those genes were represented were oxidative stress and hypoxic stress. This finding potentially suggests that stress response genes could predispose to schizophrenia through inflammation. Thus, Weinberger examined 23 canonical inflammation pathways and found them to be co-expressed with the schizophrenia risk genes that were differentially regulated in the placenta. Lastly, he wanted to look at the causes of greater prevalence of schizophrenia in males and to investigate a potential link between the placental environment and increased risk of schizophrenia in males. He found in the placenta samples higher upregulation of the differentially expressed schizophrenia genes in males compared to females. Weinberger concluded that a focus on prenatal health may be the best way to prevent schizophrenia in individuals with high genetic risk.