Risperidone-induced changes in DNA methylation from peripheral blood in first-episode schizophrenia parallel neuroimaging and cognitive phenotype
Maolin Hu,Yan Xia,Xiaofen Zong,John A. Sweeney,Jeffrey R Bishop,Gina Giase,Bingshan Li,Leah H. Rubin,Yunpeng Wang,Liz Kunney,Zongchang Li,Ying He,Chunyu Liu,Chao Chen,Xiaogang Chen,Jinsong Tang
DOI: https://doi.org/10.1101/2020.03.31.018283
2020-01-01
Abstract:Today, second generation anti-psychotics such as clozapine and risperidone are the favored treatment for schizophrenia. Yet, the absence of relevant biomarkers that can decode their neurobiological effect shackles our ability to accurately predict and track response to treatment. While researchers have investigated DNA methylation as a biomarker for schizophrenia risk, none have performed a systematic analysis of the effect of antipsychotics upon DNA methylation. We hypothesize that disease-related methylation changes occur before treatment, and that acute antipsychotic treatment may affect DNA methylation. We designed a longitudinal DNA methylation study to estimate risperidone’s effect on DNA methylation and how changes in DNA methylation might influence risperidone’s therapeutic effect on behavioral and neuroimaging phenotypes. Thirty-eight patients with first-episode drug-naïve schizophrenia (FES) and 38 demographically-matched individuals (healthy controls) participated. We identified brain related pathways enriched in 8,204 FES-associated methylation sites. Risperidone administration altered methylation in 6,143 CpG DNA sites. Post-treatment FES associated with methylation in 6760 CpG sites. Majority of the DNA methylation changes were treatment effect in the overall CpG sites, the FES associated CpG sites, and risperidone associated CpG sites, except for the post-treatment FES associated CpG sites. There were 590 DNA methylation cites normalized by risperidone treatment. The methylation changes of these 590 CpG sites were related to alterations in symptom severity, spontaneous neurophysiological activity, and cognitive function. To our knowledge, this is the first longitudinal methylation study of drug treatment effect and side effect in psychiatric disorders to include parallel studies of neuroimaging and cognitive phenotypes. We identified FES-associated CpG sites not confounded by drug treatment as potential SCZ biomarkers. The normalization effect of risperidone monotherapy suggests that DNA methylation changes may serve as a predictive biomarker for treatment effect. The constructed methylation-phenotype network revealed a relationship between methylation and a wide range of biological and psychological variables.