FKBP5 allele-specific epigenetic modification in gene by environment interaction.

Abstract

The likelihood to develop stress-related psychiatric disorders in response to childhood trauma exposure may be moderated by the individual's genetic predisposition (Manuck and McCaffery, 2014). One of the genetic variants reported to alter the risk for psychiatric disorders following childhood trauma is a functional variant in FKBP5, a gene encoding a co-chaperone of the glucocorticoid receptor (GR). FKBP5 is strongly induced following stress exposure via binding of activated GR to a number of intronic and promoter GR response elements (GREs). The protein itself then binds to the GR complex, reduces the affinity of GR to cortisol and decreases translocation of the GR to the nucleus, providing an ultrashort negative feedback for GR activation on the genomic and protein level (Zannas and Binder, 2014). We have identified a functional polymorphism in close proximity of a GRE in intron 2 of FKBP5, which alters the extent of mRNA and protein induction following GR activation, likely by an altered 3D conformation. This results in a variable interaction of the intron 2 GRE with the transcription start site, leading to increased or reduced mRNA induction, respectively (Klengel et al, 2013). Individuals carrying the allele associated with stronger FKBP5 mRNA induction show GR resistance, prolonged cortisol response following stress, altered activation of brain regions important for threat response, such as the amygdala, and increased risk to a number of psychiatric disorders including major depression and post-traumatic stress disorder when exposed to childhood trauma. Interestingly, while the genetic effects on the physiological stress response are seen in adults, no interaction of adult trauma with this genotype on psychiatric risk is observed, suggesting an additional mechanism that explains the FKBP5 × childhood trauma interaction. In fact, we could show that exposure to childhood trauma leads to allele-specific epigenetic changes with a decrease in DNA methylation in a second GRE located in intron 7 of the gene, but only in carriers of the risk allele. This demethylation further de-represses FKBP5 induction following GR exposure and is likely mediated by the genetically determined increase in cortisol response following stress (Klengel et al, 2013) (Figure 1). Indeed, direct GR activation with a selective agonist in a neuronal progenitor cell line leads to a demethylation in exactly the same CG dinucleotides (CpGs) that are shown to be less methylated in DNA from peripheral blood in trauma-exposed risk allele carriers. These CpGs are located either within or between GR consensus binding site sequences, while more proximal CpGs are unaffected. We thus speculate that the demethylation is an active demethylation, induced by GR binding. Such active demethylation at GREs has been described before (Kress et al, 2006) and, although not directly shown, hydroxymethylation could be an intermediate step in this active transcription factor binding-induced demethylation (Bhutani et al, 2011), a process that has been described for other transcription factors as well (Feldmann et al, 2013). On a more general level, any genetic variant that alters binding of stress-induced transcription factors may thus lead to local differences in subsequent epigenetic changes. Thereby, allele-specific epigenetic modifications can contribute to gene × environment interactions, leading to long-term effects of stress on endocrine levels, brain activity, and the risk to develop psychiatric disorders. Figure 1 Schematic representation of an allele-specific epigenetic modification in FKBP5. The single-nucleotide polymorphism rs1360780 close to a functional GRE in intron 2 constitutes the genetic predisposition to an increased FKBP5 transcriptional response to ...