Multi-tissue gene-expression analysis in a mouse model of thyroid hormone resistance

Abstract

Background: Resistance to thyroid hormone (RTH) is caused by mutations of the thyroid hormone receptor beta (TRbeta) gene. To understand the transcriptional program underlying TRbeta mutant-induced phenotypic expression of RTH, cDNA microarrays were used to profile the expression of 11,500 genes in a mouse model of human RTH.

Results: We analyzed transcript levels in cerebellum, heart and white adipose tissue from a knock-in mouse (TRbetaPV/PV mouse) that harbors a human mutation (referred to as PV) and faithfully reproduces human RTH. Because TRbetaPV/PV mice have elevated thyroid hormone (T3), to define T3-responsive genes in the context of normal TRbeta, we also analyzed T3 effects in hyperthyroid wild-type gender-matched littermates. Microarray analysis revealed 163 genes responsive to T3 treatment and 187 genes differentially expressed between TRbetaPV/PV mice and wild-type littermates. Both the magnitude and gene make-up of the transcriptional response varied widely across tissues and conditions. We identified genes modulated in T3-dependent PV-independent, T3- and PV-dependent, and T3-independent PV-dependent pathways that illuminated the biological consequences of PV action in vivo. Most T3-responsive genes that were dysregulated in the heart and white adipose tissue of TRbetaPV/PV mice were repressed in T3-treated wild-type mice and upregulated in TRbetaPV/PV mice, suggesting the inappropriate activation of T3-suppressed genes in RTH.

Conclusions: Comprehensive multi-tissue gene-expression analysis uncovered complex multiple signaling pathways that mediate the molecular actions of TRbeta mutants in vivo. In particular, the T3-independent mutant-dependent genomic response unveiled the contribution of a novel 'change-of-function' of TRbeta mutants to the pathogenesis of RTH. Thus, the molecular actions of TRbeta mutants are more complex than previously envisioned.

Figure 1

Concordance of gene expression determined by microarrays and RT-PCR. Expression ratios of representative genes in (a) cerebellum and (b) white adipose tissue identified as outliers by microarrays were determined by RT-PCR as described in Materials and methods. The solid bars represent the data from the microarrays and the open bars are from the RT-PCR (mean ± SEM, n = 3).

Figure 2

Category analysis of transcriptional response patterns. Intra-tissue expression patterns of genes showing twofold change or more in iT3 and/or RTH (PV) mice are shown. Red indicates higher expression levels in iT3 or RTH mice; green indicates lower expression in iT3 or RTH mice. Black indicates less than 1.2-fold change. The level of color saturation reflects the magnitude of the expression ratio. The number of genes found in each category is shown.

Figure 3

Hierarchical clustering of differentially expressed genes identifies gene clusters with biological associations. Gene-expression patterns are shown in rows; tissue profiles in columns. Degree of color saturation reflects the magnitude of the expression ratio. Note that for optimal clustering, the expression data for each individual dye-swap experiment was used (see key for directionality of expression via color pairs). The black bar at the right side of the main array indicates the immunity cluster, the blue bar the lipogenesis cluster, and purple bars the cell-cycle/growth-inhibitor genes.

Figure 4

Treeview visualization of iT3- and PV-responsive immunity/lymphocyte-related genes. Higher and lower transcript levels in T3-treated animals (iT3) or RTH mice (PV) are indicated by arrows (as relative to iT3 control or wild-type mice, respectively). IMAGE clone IDs and UniGene names are given; asterisks (*) indicate immune cell-specific genes.