Search

Background Emerging evidence implicates altered DNA methylation in mental illness including autism, ADHD, bipolar disorder, major depressive disorder, anorexia and schizophrenia. However, it is unclear whether the DNA methylation changes observed to date are causative or reflect disease progression or treatment. The neonatal period is a time of rapid neurodevelopment during which alterations in DNA methylation may contribute to the risk of mental illness later in life. Hence, we explored whether differences in DNA methylation in neonatal blood were associated with twin discordance for mental illnesses including autism, ADHD, affective disorder, anorexia, schizophrenia or bipolar disorder. Methods A total of 597 pairs of twins (220 monozygotic) discordant for mental illness born between 1981 and 2005 were identified for methylomic comparison. Blood samples obtained from neonatal Guthrie cards were used for DNA extraction and genome-wide profiling of DNA methylation with the use of Infinium HumanMethylation450 BeadChip or EPIC array from Illumina. Quality control, data pre-processing and statistical analysis was performed using the minfi package in R. Data were normalized using the ssNoob method and adjusted for batch effects using the Combat tool. Blood cell composition was estimated using FlowSorted.CordBlood.450k. Using linear regression models and including potential confounders such as sex, blood cell composition and zygosity, we observed differentially methylated positions (DMPs) associated with mental illness. We used VarElect (http://varelect.genecards.org/) to identify which genes are known to directly associate with mental illness. The GeneMania tool (http://genemania.org/) was used to visualize gene interactions in a network. Results We observed significant DMPs (P Discussion Our data indicate that DNA methylation differences are quantifiable in neonatal blood from twins discordant for mental illness later in life and suggest that susceptibility to mental illness is conferred by dysregulated neurodevelopmental genes.

Background Neonate developmental state at birth has previously been associated with later occurring psychiatric disease. As a novel possible measure of neonate epigenetic developmental state, we investigate the association between Accelerated Gestational Age (AccGA) and later occurring psychiatric disorder in six neuropsychiatric illnesses: Schizophrenia (SZ), Bipolar Disorder (BD), Major Depressive Disorder (MDD), Autism Spectrum Disorder (ASD), Attention Deficit Hyperactivity Disorder (ADHD) and Anorexia (ANX), in a cohort of dis/concordant Monozygotic (MZ) twins. Methods Neonatal dried blood spots, of 548 MZ twins (274 pairs) born between 1981 and 2005 were obtained from the Danish Neonatal Screening Biobank. The samples were analysed on the Illumina Infinium HumanMethylation 450k BeadChip. Estimated DNA methylation gestational age (EstGA) was calculated from 148 CpG probes across the array (described by Knight et al. 2016). Residual values of the linear fit between EstGA and actual gestational age, were used as measures of the in-cohort AccGA. We applied a linear mixed model for the twin pedigree structure - regressing AccGA against the six neuropsychiatric disorders and sex, while Twin Pair ID served as random effect in the twin model. Statistical significance was reached at p-value Results The EstGA correlated well with the actual gestational age (pearson r=0.86), and the AccGA correlated between twin-cotwin (Pearson r=0.67). Within the full cohort BD (N=7) and ANX (N=44) patients had significantly (p-val=0.007 and p-val=0.024) elevated AccGAs over healthy cotwins and cohort (0.967+/-0.356 and 0.310+/-0.137 weeks, respectively). For SZ (N=18) the opposite was observed (pval=0.005) with lowered AccGA (-0.585+/-0.206 weeks). Neither MDD, ASD nor ADHD were significantly associated with AccGA. Within Twin pair analysis of BD (N=6 pairs) and SZ (N=17 pairs), revealed the same significant (p-val=0.016, p-val=0.022, respectively) difference as observed above (1.185+/-0.341 and -0.697+/-0.275, respectively), while ANX (N=40 pairs) showed the same tendency although not significantly (p-val=0.053, 0.235+/-0.118). This association was independent of twin-cotwin birth weight percentile differences (Pearson r=0.0002, Not significant). Discussion We found significant AccGA differences within con/discordant MZ twin pairs and cohort of at least three main psychiatric disorders - detectable already at birth. AccGA has previously been correlated with birth weight percentile. Twin-cotwin differences of AccGA was however not explained by birth weight percentile differences, suggesting that the within pair differences caught by the estimator describes other aspects of neonate development, independent of weight. Our findings suggest that gestational age estimation, as assessed by the 148 CpG probes, in MZ twin pairs, can indicate a not previously reported epigenetic link to later psychiatric illness, this should be further investigated.

Background Quantitative psychiatric genetics has long established not only high heritability, but also appreciable levels of co-heritability, or genetic correlations, among psychiatric disorders. More recently, it is becoming increasingly demonstrated that the effects of common Single Nucleotide Polymorphisms (SNPs) are likewise rarely confined to one psychiatric diagnosis. As such, it may be reasonable to combine psychiatric patients into a single cross-disorder cohort with the aims of identifying SNPs that predispose to multiple disorders. In this study, we examine cross-disorder genetic effects in the iPsych cohort, a case-cohort collection of more than 50,000 psychiatric patients selected by treatment records for ADHD, affective disorders, anorexia, autism spectrum disorders, bipolar disorder, or schizophrenia and 30,000 healthy controls. Methods To motivate single locus cross-disorder analysis we estimated SNP-based (co)heritability among the six ascertained psychiatric cohorts using GREML implemented in GCTA. We followed that analysis with a typical case-control GWAS, where the case group were subjects who had received treatment for any psychiatric disorder and the controls had received treatment for no psychiatric disorder. Genome-wide significant loci were described using functional genomics data from the roadmap epigenetics project and publicly available GWAS results. Results In this study, we report moderate SNP-based (co)heritabilities for and among the six psychiatric cohorts with ascertainment in the iPsych Cohort, confirming a likely presence of cross-disorder loci. Our subsequent cross-disorder GWAS revealed four genome-wide significant loci with plausible biological mechanisms. Discussion In this analysis, we demonstrate moderate SNP-heritability and genetic sharing among the psychiatric treatment record based phenotypes in the Danish National Health Register, suggesting the presence of single loci with effects across multiple disorders. These analyses build upon and extend the emerging relaxation of rigid psychiatric diagnostic boundaries, suggesting at least some significant portion of risk may emerge from shared biological disruptions.