Thyroid hormone-regulated gene expression in juvenile mouse liver: identification of thyroid response elements using microarray profiling and in silico analyses

Abstract

Background: Disruption of thyroid hormone signalling can alter growth, development and energy metabolism. Thyroid hormones exert their effects through interactions with thyroid receptors that directly bind thyroid response elements and can alter transcriptional activity of target genes. The effects of short-term thyroid hormone perturbation on hepatic mRNA transcription in juvenile mice were evaluated, with the goal of identifying genes containing active thyroid response elements. Thyroid hormone disruption was induced from postnatal day 12 to 15 by adding goitrogens to dams' drinking water (hypothyroid). A subgroup of thyroid hormone-disrupted pups received intraperitoneal injections of replacement thyroid hormones four hours prior to sacrifice (replacement). An additional group received only thyroid hormones four hours prior to sacrifice (hyperthyroid). Hepatic mRNA was extracted and hybridized to Agilent mouse microarrays.

Results: Transcriptional profiling enabled the identification of 28 genes that appeared to be under direct thyroid hormone-regulation. The regulatory regions of the genome adjacent to these genes were examined for half-site sequences that resemble known thyroid response elements. A bioinformatics search identified 33 thyroid response elements in the promoter regions of 13 different genes thought to be directly regulated by thyroid hormones. Thyroid response elements found in the promoter regions of Tor1a, 2310003H01Rik, Hect3d and Slc25a45 were further validated by confirming that the thyroid receptor is associated with these sequences in vivo and that it can bind directly to these sequences in vitro. Three different arrangements of thyroid response elements were identified. Some of these thyroid response elements were located far up-stream (> 7 kb) of the transcription start site of the regulated gene.

Conclusions: Transcriptional profiling of thyroid hormone disrupted animals coupled with a novel bioinformatics search revealed new thyroid response elements associated with genes previously unknown to be responsive to thyroid hormone. The work provides insight into thyroid response element sequence motif characteristics.

Figure 1

PND 15 pup serum T4 levels after short-term TH perturbation. Means are presented as well as ± standard deviations. n = 10, 5 from each sex, for control, hyper and hypo; n = 8, 4 from each sex for hypo+; n = 5, males only for hypo++. All treatment groups when compared to controls were significantly different; p ≤ 0.05, determined by student t-test.

Figure 2

Principal component analysis of microarray data carried out with the full set of significant genes (FDR-adjusted p ≤ 0.05) for all groups (control, hyper, hypo and hypo+). Data are coloured by sex (A) and by treatment group (B). In panel (A) male data points are in red and female data points are in yellow, and in panel (B) control data points are in yellow, hypothyroid data points are in red, hyperthyroid data points are in cyan and hypothyroid/replacement (hypo+) data points are in fuchsia.

Figure 3

Relative enrichment of newly identified TREs determined by ChIP-PCR. The top half of each section shows the amplicons run on an agarose gel, and the bottom half shows relative enrichment of the immunoprecipitated (IP) samples and the total input (TI) samples when compared to β-actin enrichment. ChIP-PCR validation of negative (β-actin) and positive (Mlxipl) controls are presented in section (A). Sections (B) to (E) show enrichment of Tor1a, H01Rik (2310003H01Rik), Hectd3 and Slc25a45. Asterisks (*) denote a significant difference, p ≤ 0.05, determined by student t-test. The mean enrichments (± standard deviations) are also presented for each section. All presented immunoprecipitated enrichments were significant (p ≤ 0.05 determined by student t-test) when compared to total input, except for β-actin.

Figure 4

Examining the potential binding of candidate TREs by EMSA. The left panel shows results using a classic DR4 TRE with two "AGGTCA" half-sites. The next three panels show gel shifts when using a probe targeting candidate TREs in the promoter regions of Slc25a45, Hectd3 and Tor1a.