Fig. 4: Developmental programming in TRα1 + m and wild-type animals by thyroid hormone in the second half of pregnancy.

a Treatment scheme of TRα1 + m mutants (red) or wild-type controls (black) that were untreated (solid color) or treated with T3 (light color) from embryonal day 12.5 (E12.5) to E17.5 by oral T3 treatment of the pregnant mother and analysed as adults. b Three-dimensional principal component analysis of the transcriptome microarray data of these animals. c Heat map for genes involved in heart rate regulation or potassium channels that are normalized in TRa1+m animals by maternal T3 treatment (upper part) or significantly affected by the maternal T3 treatment irrespective of genotype (lower part). d Three-dimensional principal component analysis of the methylome (CpG DNA methylation) data of these animals. e Average global DNA methylation in these groups. f Distribution of the relative DNA methylation ranging from fully demethylated to fully methylated CpG. g Unbiased clustering of differentially methylated CpG sites in the four groups. Arrowheads highlight clusters with noticeably changed methylation in TRα1 + m mice that are reversed upon maternal treatment. h Increased DNA methylation at a specific CpG site within the body of the Ryr2 gene in TRα1 + m mice as compared to wild-types, which is reversed by maternal T3 treatment. i Increased DNA methylation at a specific CpG site within the body of the Kcnh2 gene in both wild-type and TRα1 + m mice upon maternal T3 treatment. Values are mean ± SEM for n = 6 per group. c *P < 0.05; **P < 0.01; ***P < 0.001, and ****P < 0.0001 with two-way ANOVA for T3 effect; ^#P < 0.05, and ^####P < 0.0001 with two-way ANOVA for TRα1 effect; ^$P < 0.05, and ^$$P < 0.01 with two-way ANOVA for TRα1 and T3 interaction effect. e Data are shown as box plots (box represents median with interquartile range, ie. 25th to 75th centiles, whiskers depict range of values from minimum to maximum); ***P < 0.001, and ****P < 0.0001, Kruskal–Wallis with Wilcoxon rank-sum test with continuity correction. h *P < 0.05 for TRα1 effect, two-way ANOVA with post hoc test corrected for multiple comparisons by controlling the false discovery rate using the Benjamini, Krieger and Yekutieli method. Exact P values are provided in Supplementary Table 1.