Details

Autor(en) / Beteiligte

• Karlbauer, Vera
• Czamara, Darina
• Martins, Jade
• Rex-Haffner, Monika
• Entringer, Sonja
• Winter, Sybille
• Buss, Claudia
• Heim, Christine
• Binder, Elisabeth

Titel

T23. TARGETED BISULFITE SEQUENCING REVEALS DRIVERS OF FKBP5 METHYLATION IN MALTREATED AND NON-MALTREATED CHILDREN

Ist Teil von

• European neuropsychopharmacology, 2023-10, Vol.75, p.S173-S174

Ort / Verlag

Elsevier B.V

Erscheinungsjahr

2023

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Childhood maltreatment (CM) has consistently been linked to increased risk for developing psychiatric disorders later in life. Changes in DNA methylation (DNAm) along stress-associated pathways have been proposed as an underlying biological mechanism, for instance in the FKBP5 gene, which regulates the glucocorticoid (GC) receptor. Genotype-dependent differential DNAm of FKBP5 has been associated with mental disorders occurring after CM. However, previous studies on FKBP5 DNAm have only analyzed very few CpGs in this gene that are available via genome-wide arrays or pyrosequencing. Targeted bisulfite sequencing (TBS) is a high-throughput method designed to assess DNAm with much higher coverage and was optimized to target regulatory regions of the FKBP5 locus containing GC-responsive elements and CTCF binding sites. We employed the TBS approach to quantify DNAm levels across FKBP5 at two time points in maltreated and non-maltreated children and to study interactions with genotype and prenatal exposures. Analyses were performed in the BerlinLCS cohort of 83 maltreated and 79 non-maltreated children aged 3-5 years using saliva samples taken at baseline and at a one-year follow-up. FKBP5 DNAm was quantified via TBS for 49 CpGs from 5 functional regions of FKBP5 (3’TAD, i.e. topologically associating domain, 5’TAD, proximal enhancers, enhancers in introns 5 and 7). Genome-wide DNAm levels (EPIC BeadChip) were used to perform cell type deconvolution and to construct proxy scores for prenatal exposure to alcohol, cigarette smoke, and GCs. The rs1360780 SNP in FKBP5 was extracted from genome-wide genotypes (Illumina GSA-24 v2.0 BeadChip). Statistical models were run over all 49 CpGs and FDR-corrected at 0.05 for multiple testing. All analyses were controlled for age, sex, the first three genetic principal components, estimated cell type proportions, prenatal smoke and alcohol exposure scores. Additionally, the percentage of explained variance for each predictor and covariate was determined in a variance partitioning analysis. FKBP5 DNAm differed significantly between maltreated and non-maltreated children for 9 CpGs at baseline and 2 CpGs at follow-up. However, these hits did not survive FDR correction and there was no overlap between timepoints. Carriers of the rs1360780 risk allele showed significant hypomethylation at 2 CpGs in the 3’TAD across time and after correction. In the variance partitioning analysis, cell type proportions and prenatal exposure scores explained the largest percentages of variance (≤ 62%) in FKBP5 DNAm, especially in the 3’TAD, compared to much smaller contributions of maltreatment (≤ 2%). Increased prenatal GC exposure was significantly associated with hypomethylation of 2 CpGs in the 3’TAD. This effect survived FDR correction, was consistent over time, and displayed additive genotype interaction. The TBS approach revealed nominal differences in FKBP5 DNAm in maltreated as compared to non-maltreated children. However, due to limited sample size, these effects did not survive correction. Previous literature reporting FKBP5 hypomethylation in risk allele carriers was replicated. Prenatal exposures and cell type proportions explained the largest proportion of variance in FKBP5 DNAm. These results illustrate the importance of genome-wide covariates even for candidate gene analyses. They also underscore the impact of GC exposure on FKBP5 DNAm and highlight the relevance of the 3’TAD, a region not usually assessed in FKBP5 studies.