Ascorbic acid, a well-known antioxidant, reduces Cu cations to Cu0 only when the cations are bound to organic substrates such as DNA in the presence of oxygen in the dark or via autocatalysis on Cu metal seeds in the absence of stabilizing organic ligands . As an example of synthetic control, pHdependent conformation of histidine-rich peptides has led to larger nanocrystals of Cu0 at pH 7–10 than at pH 4–6 . Plants produce ascorbic acid for many functions and rhizospheres often contain the breakdown products of ascorbic acid, which facilitates electron transfer during mineral weathering . Plants produce more ascorbic acid when grown in soils contaminated with heavy metals including copper . Fungi, which proliferate over plants and bacteria in metal-contaminated soils, can stabilize excess copper by extracellular cation binding or oxalate precipitation , but mechanisms probably also require enzymes, thiol-rich proteins and peptides, and antioxidants . The formation of electron-dense Cu granules within hyphae of arbuscular mycorrhizal fungi isolated from Cu- and As-contaminated soil suggests that fungi also can produce nanoparticulate copper. Some copper reduction possibly occurred in response to the European heat wave of the summer of 2003 . Elevated expression of heat shock protein HSP90 and metallothionein genes has been observed in hyphae of an arbuscular mycorrhizal fungus in the presence of 2 × 10–5 M CuSO4 in the laboratory . This suggests that a single driving force can trigger a biological defense mechanism that has multiple purposes. Thus, reduction of toxic cations to native elements may increase as rhizosphere biota fight metal stress and stresses imposed by elevated temperatures expected from global warming.Despite schizophrenia being a debilitating disorder affecting 1% of the population, there are no extant biomarkers to aid the clinician in identifying this disorder. Studies predict the genetic risk to be up to 80%,but despite strenuous research efforts the genes and polymorphisms found to be associated with schizophrenia account for very little of the genetic risk. Environmental risk such as urbancity,migrant status,childhood maltreatment,prenatal infections,cannabis use and maternal vitamin D deficiency also contribute to schizophrenia susceptibility. However, not all individuals exposed to environmental risk develop schizophrenia.This observation suggests that interaction between susceptibility genes and environmental factors may better account for schizophrenia. DNA methylation has been identified as a key mechanism for environmental regulation of gene expression.DNA methylation is an epigenetic modification that is essential for normal human development via regulation of gene function.
DNA methylation results in the addition of a methyl group on the cytosine of CpG dinucleotides,vertical growing systems which can then be inherited through cell division. These cytosine modifications can affect gene expression by altering the binding of transcription factors to promoter regions or changing mRNA processing. DNA methylation studies of the brain and peripheral tissue have previously been reported for schizophrenia. However, to our knowledge, no study has published results from an Illumina Infinium HumanMethylation450 Beadchip in the brain tissue of patients with schizophrenia. Studies to date have typically been performed in peripheral tissues and have been limited to the analysis of CpG islands in the promoter regions. A recent DNA methylation study analysed 27 578 CpG sites in peripheral blood cells from 18 patients with schizophrenia and 15 normal controls.This study revealed 603 CpG sites that had significantly different DNA methylation levels between schizophrenia and controls. Among these genes were HTR1E, COMTD1 and SLC6A3, which have previously been found to be associated with schizophrenia. An epigenetic study of monozygotic twins discordant for schizophrenia identified a number of loci differentially methylated in peripheral blood.Selected gene promoters have also been analysed for differential DNA methylation in the brain tissue from small numbers of patients with schizophrenia. Some of these genes include RELN,COMT,SOX1016 and HTR2A.An earlier study of 12 000 CpG islands in the frontal cortex of 35 schizophrenia and 35 controls revealed differential DNA methylation in genes associated with glutamatergic and GABAergic pathways.Apart from the present study, the only extant study using a 450 000 genome-wide methylation array was performed in leukocytes from patients with schizophrenia.DNA methylation analysis of schizophrenia has been more widely performed in peripheral tissue, because it can be readily obtained from living patients. The epigenetic profile differs in the brain compared with the peripheral tissue; however, some regions may have common patterns,which would make these regions ideal as potential biomarkers for schizophrenia. Some of the genes found to be differentially methylated in peripheral tissue of schizophrenia patients include HTR1A,HTR2A, BDNF,GRM2,GRM5and COMT.Brain tissue from the Human Brain and Spinal Fluid Resource Centre, CA, USA, was obtained in order to examine tissue involved in the etiology of schizophrenia. We analysed this tissue in a genome-wide methylation study of schizophrenia. We report significant differences in methylation status in brain tissue from schizophrenia patients compared with that from controls. In addition, unsupervised clustering analysis revealed two distinct groups corresponding to schizophrenia and controls.
Results of future epigenetic studies hold great promise of a schizophrenia biomarker and treatment, as epigenetic processes can be reversed.In order to assess differences in methylation between groups, the original n = 385 167 β-values were converted to M-values via the logit transformation as recommended by Du et al.Differentially methylated probes were detected using the limma package.The limma procedure uses linear models to assess differential methylation, whereby information is shared across probes.A major benefit of the limma procedure is that it allows the inclusion of covariates or other factors in the specification of the linear model. As such, we were able to adjust for age and PMI in the detection of differentially methylated probes by including age and autolysis covariates in the specification of the design matrix. Although most studies have found that methylation status is unaffected by PMI, we decided to adjust for PMI as a confounder.Probes were considered to be differentially methylated if the resulting adjusted P-value was o0.05. The Benjamini–Hochberg method was used to adjust the P-values and ensure that the false discovery rate was o0.05. The corresponding gene list was derived from the gene annotations associated with the probes.For cluster analysis, the top 3000 most variable probes were selected . A recursively partitioned mixture model was used to cluster the β-scores. RPMM is a model-based unsupervised clustering algorithm developed for measurements that lie between 0 and 1. This algorithm was implemented using the RPMM Bioconductor package.The implementation of RPMM was identical to Hinoue et al.who used a fuzzy clustering algorithm for initialisation and level-weighted version of Bayesian Information Criterion as a split criterion. In order to adjust for age and PMI, a series of linear models were fitted to the M-values using the function lmFit in the limma package. Coefficients for age and PMI, along with an intercept were estimated for each probe. Owing to this model specification, the residuals of the linear model represent the methylation values adjusted for the effect of age and PMI. The residuals were then back-transformed and clustered using the RPMM method implemented for the unadjusted probes. To allow visualisation of the distance between samples and to further reinforce the RPMM clustering, multidimensional scaling with a Euclidian distance metric was performed on both the adjusted M-values and the adjusted β-values. The first two coordinates, along with the RPMM clusters, are visualised in Figures 1a and b.
The results of the clustering indicate that the methylation profiles in those with schizophrenia are a heterogeneous group. There were some profiles that were consistently deemed distinct from the controls, whereas there were others that were not found to be significantly dissimilar. Twelve samples in particular tended to exhibit the former trait. When comparing these two potential subgroups of those with schizophrenia,vertical grow system we can see that the two subgroups exhibit no obvious difference in characteristics . Thus, there is potential for methylation arrays to be used to detect differences within these two potential subgroups. Differential methylation analysis between the two schizophrenia subgroups indicated that there were 73 222 probes that were differentially methylated . Of those probes, 6681 were promoter-associated and 2006 were both promoter-associated and located at a CpG island. After adjusting for age and PM1, 56 001 probes were found to be differentially methylated , 4779 being promoter-associated and 1238 both promoter-associated and located at a CpG island. The abundance of differentially methylated probes suggests significant groupings within the schizophrenia methylation profile. By contrast, a history of completed suicide or the presence of another psychiatric disorder revealed no significant differences in methylation.Differential DNA methylation in schizophrenia has been reported in several studies to date, although most of these studies involve the use of non-functional tissues such as blood. In this study, we analysed DNA methylation status in brain tissue, the primary tissue of pathology in schizophrenia, employing a genome-wide methylation array with very extensive coverage of the potential methylation sites in the human genome. After adjusting for age and PMI, 4641 probes corresponding to 2929 unique genes were found to be differentially methylated. When we compared the differentially methylated gene list with past studies using peripheral leukocyte samples, we found a high concordance rate, particularly for genes previously found to be associated with schizophrenia. Of the 589 genes Nishioka et al.11 found to be differentially methylated in peripheral blood cells from patients with schizophrenia, we were able to replicate 99 of these in the brain tissue. This shows promise for the use of non-invasive tissue such as blood or saliva to be used as a future diagnostic indicator of schizophrenia. We are aware of only one other study that used the 450 Illumina array in schizophrenia, although peripheral leukocytes were analysed rather than the brain tissue.That study identified 10 747 differential DNA methylation sites in medication free subjects.One of the genes they identified was RAI1, which has altered DNA methylation in the present study as well as an earlier schizophrenia study using the brain tissue.Other genes found to be differentially methylated in both the leukocyte study and the present brain tissue study includes HDAC4, GFRA2 and GDNF. The leukocyte study did not replicate COMTD1 and HTR1A that were found to be differentially validated from a previousstudy in the peripheral tissue.However, we report that these genes are differentially methylated in the brain. Although we were able to validate many of the previously identified CpG sites, experimental validation using an alternative method, such as pyrosequencing, would also confirm our results. Functional significance of genes found to be differentially methylated should also be tested by gene expression.
Unsupervised clustering of the top 3000 most variable probes revealed two distinct groups after adjusting for age and PMI. Cluster 1 comprised 88% patients with schizophrenia and 12% controls, whereas cluster 2 comprised 27% patients with schizophrenia and 73% controls. To our knowledge, this is the first report of DNA methylation profiling that is able to significantly differentiate between those with schizophrenia and control subjects. Although Nishioka et al.was able to identify site specific DNA methylation changes in patients with schizophrenia, they were unable to discriminate between controls and schizophrenia patients using unsupervised clustering.11 DNA methylation patterns differ in brain cells compared with peripheral tissues such as blood,and this may explain the lack of separation reported by Nishioka et al.Although some genes may have the same epigenetic profiles in peripheral and brain tissue, a more comprehensive list of tissue-specific genes may be required to differentiate controls from those with schizophrenia. Another reason may be the analysis of fewer CpG sites in the previous study. This is potentially important, as a DNA methylation signature across the whole genome is required to identify the most important differentially methylated probes. The results of our clustering analysis will need to be confirmed in an independent brain tissue cohort.It is possible that these two subgroups have specific symptomatology that warrants further investigation in a sample set with a comprehensive clinical history. After adjusting for age and PMI, DTNBP1, COMT and DRD2 were found to be differentially methylated between the two schizophrenia subgroups. Interestingly, these are genes that we have previously found to be associated with schizophrenia.A recent study has found significant DNA methylation changes in the early stages of development and suggest that aberrant DNA methylation during the transition from the fetal to the postnatal period of development could be critical for the pathogenesis of schizophrenia.