DNA methylation signatures link prenatal famine exposure to growth and metabolism

Abstract

Periconceptional diet may persistently influence DNA methylation levels with phenotypic consequences. However, a comprehensive assessment of the characteristics of prenatal malnutrition-associated differentially methylated regions (P-DMRs) is lacking in humans. Here we report on a genome-scale analysis of differential DNA methylation in whole blood after periconceptional exposure to famine during the Dutch Hunger Winter. We show that P-DMRs preferentially occur at regulatory regions, are characterized by intermediate levels of DNA methylation and map to genes enriched for differential expression during early development. Validation and further exploratory analysis of six P-DMRs highlight the critical role of gestational timing. Interestingly, differential methylation of the P-DMRs extends along pathways related to growth and metabolism. P-DMRs located in INSR and CPT1A have enhancer activity in vitro and differential methylation is associated with birth weight and serum LDL cholesterol. Epigenetic modulation of pathways by prenatal malnutrition may promote an adverse metabolic phenotype in later life.

Figure 1. The methylation level of genes in the RRBS data set.

The methylation level across a gene. A lowess has been fit across all data for all entrez genes. The width of the gene elements represents the relative amount of data for the elements in the total data set.

Figure 2. The average within-pair difference for the 181 regions associated with prenatal famine exposure after correction for multiple testing.

A histogram for the average within pair difference (%) between the exposed and unexposed siblings. A positive number reflects relative higher DNA methylation levels in the exposed.

Figure 3. The correspondence between RRBS and EpiTYPER measurements of DNA methylation at P-DMRs.

Individual regions with a Pearson correlation >0.7 are denoted in color and plotted separately along the main figure. The correlation of the other regions can be found in Supplementary Table 3.

Figure 4. Results across the famine period.

(a) A lowess curve depicting the average within-pair difference (y axis) stratified by the estimate of the start of pregnancy (LMP; x axis). Each coloured line represents an individual P-DMR. (b) the blue bars depict the official daily rations (kcal per day) per week, the black line represents a lowess curve depicting the average 24 h temperature (source KNMI; DeBilt weather station). The daily requirement of (non-pregnant) women of 2,000 kcal per day is denoted in red.

Figure 5. Revigo analysis of the significant pathways.

A sunburst graph of the non-redundant clustered FDR significant GO terms associated with prenatal famine exposure. The size of the circular boxes are proportional to the level of statistical evidence. In bold are the dominant terms of the clusters, which are denoted in different colours.

Figure 6. The INSR and CPT1A P-DMRs.

(a) Scatterplot between birth weight (x axis) and the average DNA methylation of the INSR P-DMR (y axis) in the 60 prenatally exposed individuals. (b) Scatterplot between LDL (x axis) and the average DNA methylation of the CPT1A P-DMR (y axis) in all 120 siblings. (c) Genomic annotation of INSR DMR. The P-DMR overlaps an enhancer in the HSMM and NHLF cell lines and an DNaseI hypersensitivy cluster in over 30 cell lines. (d) Genomic annotation of CPT1A DMR. The P-DMR overlaps an enhancer in the blood derived GM12878 and embryonic stem cell line H1 and a DNaseI hypersensitive cluster in over 30 cell lines. Furthermore, the BAF155 transcription factor binds in this region.

Figure 7. The difference in reporter gene expression for the INSR and CPTIA P-DMRs.

The P-DMRs were inserted in front of an EF1 promoter in the CpGL CpG-free vector. The Luciferase activity was normalized against Renilla activity. Next, the normalized fold change was calculated relative to activity of the CpGL-EF1 vector, so the same vector but devoid of an enhancer. The blue bars denote the unmethylated vector set against the CpGL-EF1 vector and the yellow bars the same vector, but now methylated, against the CpGL-EF1 vector. Error bars denote the s.e. Each experiment was performed three times in duplicate. (a) The CpGL-INSR/EF1 vector has a greater than threefold higher activity than the CpGL-EF1 vector. DNA methylation of the INSR P-DMR does not result in a significant difference in activity as compared with the unmethylated CpGL-INSR/EF1 vector. (b) The CpGL-CPT1A/EF1 vector has a greater than threefold higher activity than the CpGL-EF1 vector. DNA methylation of the CPT1A P-DMR results in a significantly lower activity of the vector as compared with the unmethylated CpGL-CPT1A/EF1 vector.