The Future is The Past: Methylation QTLs in Schizophrenia

G C A T T A C G G C A T genes Review The Future is The Past: Methylation QTLs in Schizophrenia Anke Hoffmann, Michael Ziller and Dietmar Spengler * Max Planck Institute of Psychiatry, Translational Psychiatry, Munich, 80805, Germany; (A.H.); (M.Z.) * Correspondence: ; Tel.: +49-089-3062-2546 Academic Editor: Xiangning (Sam) Chen Received: 30 September 2016; Accepted: 16 November 2016; Published: 24 November 2016 Abstract: Genome-wide association studies (GWAS) have remarkably advanced insight into the genetic basis of schizophrenia (SCZ). Still, most of the functional variance in disease risk remains unexplained. Hence, there is a growing need to map genetic variability-to-genes-to-functions for understanding the pathophysiology of SCZ and the development of better treatments. Genetic variation can regulate various cellular functions including DNA methylation, an epigenetic mark with important roles in transcription and the mediation of environmental influences. Methylation quantitative trait loci (meQTLs) are derived by mapping levels of DNA methylation in genetically different, genotyped individuals and define loci at which DNA methylation is influenced by genetic variation. Recent evidence points to an abundance of meQTLs in brain tissues whose functional contributions to development and mental diseases are still poorly understood. Interestingly, fetal meQTLs reside in regulatory domains affecting methylome reconfiguration during early brain development and are enriched in loci identified by GWAS for SCZ. Moreover, fetal meQTLs are preserved in the adult brain and could trace early epigenomic deregulation during vulnerable periods. Overall, these findings highlight the role of fetal meQTLs in the genetic risk for and in the possible neurodevelopmental origin of SCZ. Keywords: schizophrenia; methylation quantitative trait loci; fetal brain; genome-wide association studies; non-coding variants; induced pluripotent stem cells; DNA memory 1. Introduction Schizophrenia (SCZ) is a chronic, debilitating disease characterized by the presence of positive, negative, and cognitive symptoms that affect multiple aspects of mental activity, including perception, thought, attention, memory, and emotion. The age of onset is typically adolescence or early adulthood, with a median lifetime prevalence of 4.0 per 1000 and a morbid risk of 7.2 per 1000 [1]. In this review, we will firstly discuss the current status of SCZ genetics and the urgent need to map genetic variation-to-genes-to-function. Recent progress on genome-wide functional annotation of DNA sequences opens up the perspective to prioritize genetic variation to define causal variants. Here, we will explore new insights into the role of dynamic DNA methylation and ask how genetic influences on DNA methylation could contribute to the molecular etiology of SCZ. Next, we will analyze current evidence for the presence of methylation quantitative trait loci (meQTLs) in peripheral and, particularly, in healthy and diseased brain tissues, and their potential role in transcription and RNA splicing. Most interestingly, we will discuss recent findings on the role of CpG methylation and meQTLs in fetal brains and how this integrated knowledge could inform about epigenomic deregulation during vulnerable periods of brain development and the early origin of SCZ. Genes 2016, 7, 104; doi:10.3390/genes7120104 www.mdpi.com/journal/genes Genes 2016, 7, 104 2 of 18 2. The Genetic Architecture of SCZ Some 40 years of epidemiological and genetic studies have shown that SCZ is a complex disorder of combined genetic and environmental causation [2]. Family history is a strong and widely replicated risk factor, and the heritability of SCZ exceeds 60% in two national family studies [3,4] and 80% in twin studies [5]. At the same time, increasing evidence has accumulated for an important role of environmental risk factors in SCZ including aversive perinatal and early childhood events (e.g., maternal stress or illness, infection, severe socioeconomic disparity), belonging to an immigrant group or growing up in an urbanized area [3–6]. A broad spectrum of approaches (genetic epidemiology, segregation analysis, cytogenetics, genome-wide linkage, candidate gene associations, genome-wide association studies (GWAS), copy number variants (CNVs), and resequencing) have been applied to elucidate the genetics of SCZ [7]. Overall, these data support a role for both common and rare risk variants in SCZ, which occur with varying degrees of frequency and confer different risks for disease development. The polygenicity of SCZ is diverse and involves many common variants of subtle effects, rare but highly penetrant CNVs, and possibly exome variants [2]. In any case, loci identified so far represent probably only a fraction of the existing variants. The authors of a recently published GWAS [8] hypothesized that approximately 6300–10,200 independent and mostly common single nucleotide polymorphisms (SNPs) could underlie the risk for SCZ, an estimate that agrees well with the prediction that over 12,000 SNPs confer an effect on SCZ and bipolar disorder (BIP) [9]. Although each single SNP confers only a tiny increase in risk, together, they incrementally account for around 50% of the total variance in liability to SCZ [8]. In view of SCZ heritability around 60% [3,4], these estimates indicate that common genetic variation underlies major parts of SCZ heritability and emphasize the ongoing need for adequately powered studies. Still, each genetic finding can provide a potential etiological clue and point to a discrete number of biological and developmental pathways whose dysfunction typifies SCZ. Dissection of numerous relevant loci at the network scale [10] might help to unify the polygenic nature of SCZ and offer new opportunities for therapeutic interventions [11]. 3. The GWAS Era The advent of high-throughput genotyping technologies has provided important insights into the genetic architecture of SCZ, bipolar disorder (BIP), and major depression (MDD) [2]. The HapMap and 1000 Genome projects originally identified ≈40 million genetic variants across the human genome consisting of insertion/deletions (indels), CNVs, inversions, and SNPs. The most frequent form of genetic variation are SNPs, which account for 95% of all known sequence variants [12]. Up to now more than 85 million SNPs have been identified in the human population [13]. Array systems for GWAS are designed to incorporate only tag SNPs that represent all SNPs in the same linkage disequilibrium (LD) block. These tag SNPs capture most human genome variation through haplotype-based SNP imputation [14,15]. SNPs that are statistically over-represented in disease populations are termed risk-associated SNPs, whereby multiple associations are likely to all tag a single causal variant. Since independently-associated SNPs do not refer to well-bounded chromosomal regions, it is convenient to define physical (...truncated)