Development and Application of Cross-Tissue Techniques for the Analysis Of Psychiatric Epigenetic Data

Abstract

Epigenome-wide association studies in psychiatry have become increasingly common, primarily interrogating post mortem brain tissues comprised of heterogeneous mixtures of neuronal and non-neuronal cells. DNA methylation has been established as a defining feature of individual cell types, but with differing proportions of cell types in the brain, or differences in cell populations that exhibit disease features, epigenomic studies using unsorted DNA preparations must account for cellular heterogeneity as potential confounding factors. Further, with limited availability of post mortem brain cohorts, alternative tissues, such as blood, must be considered and evaluated as viable to study epigenetic changes associated with psychiatric diseases. Therefore, in this work we developed techniques that 1) utilize blood as a viable tissue for biomarker discovery and 2) are capable of correcting for cellular heterogeneity in psychiatric epigenetic studies. Using blood taken during pregnancy, we identified two biomarker loci at the HP1BP3 and TTC9B genes that predicted postpartum depression with an AUC of 0.87 in antenatally euthymic women and 0.12 in a replication sample of antenatally depressed women. Incorporation of blood count data into the model accounted for the discrepancy and produced an AUC of 0.96 across both prepartum depressed and euthymic women. To address brain heterogeneity, we performed fluorescence activated cell sorting (FACS) of neuronal nuclei and Illumina HM450 DNA methylation profiling in post mortem frontal cortex of 29 major depression subjects and 29 matched controls. Using the top cell epigenotype specific (CETS) marks, we generated a publically available R-package, "CETS," capable of quantifying neuronal proportions and generating in silico neuronal profiles capable of removing cell type heterogeneity bias from DNA methylation data. We demonstrate CETS derived neuronal proportions correlated significantly with age in the frontal cortex and cerebellum and accounted for epigenetic variation between brain regions. We demonstrate that triplex forming oligonucleoties (TFOs) can form triplex structures with genomic DNA in a sequence specific manner, preferentially binding to methylated duplex DNA, allowing for the enrichment of cell-specific DNA based on previously identified CETS marks. We have generated a model capable of predicting TFO specificity and ability to enrich for DNA methylation.