A Direct Comparison of Thyroid Hormone Receptor Protein Levels in Mice Provides Unexpected Insights into Thyroid Hormone Action

Abstract

Background: Thyroid hormone (TH) action is mediated by three major thyroid hormone receptor (THR) isoforms α1, β1, and β2 (THRA1, THRB1, and THRB2). These THRs and a fourth major but non-TH binding isoform, THRA2, are encoded by two genes Thra and Thrb. Reliable antibodies against all THR isoforms are not available, and THR isoform protein levels in mammalian tissues are often inferred from messenger RNA (mRNA) levels. Methods: We generated knock-in mouse models expressing endogenously and identically 2X hemagglutenin epitope (HA)-tagged THRs (THRA1/2, THRB1, and THRB2), which could then be detected by commercially available anti-HA antibodies. Using nuclear enrichment, immunoprecipitation, and Western blotting, we determined relative THR protein expression in 16 mouse organs. Results: In all peripheral organs tested except the liver, the predominant THR isoform was THRA1. Surprisingly, in metabolically active organs such as fat and muscle, THRB1 protein levels were up to 10 times lower than that of THRA1, while their mRNA levels appeared similar. In contrast to peripheral organs, the central nervous system (CNS) had a unique pattern with relatively low levels of both THRB1 and THRA1, and high levels of THRA2 expression. As expected, THRB2 was highly expressed in the pituitary, but a previously unknown sex-specific difference in THRB2 expression was found (female mice having higher pituitary expression than male mice). Higher THRB2 expression appears to make the central axis more sensitive to TH as both serum thyrotropin and Tshb mRNA levels were lower in female mice. Conclusions: Direct comparison of THR protein abundance in different organs using endogenously tagged HA-THR mouse lines shows that expression of THR isoforms is regulated at transcriptional and posttranscriptional levels, and in organ-specific manner. The prevalence of THRA1 and low abundance of THRB1 in majority of peripheral tissues suggest that peripheral actions of these isoforms should be revisited. A unique pattern of high THRA2 in CNS warrants further exploration of this non-TH binding isoform in brain development. Finally, THRB2, in addition to cell-specific control, is also regulated in a sex-specific manner, which may change the hypothalamus-pituitary-thyroid axis set point and perhaps metabolism in males and females.

Keywords: endogenous gene tagging; isoform specificity; nuclear hormone receptor; pituitary; thyroid hormone receptor (TRH).

FIG. 1.

CRISPR Cas9-mediated knock-in of 2XHA tag. (A) Schematic representation of the mouse Thra and Thrb genes with respective location of HA-coding sequences for Thra^HA, Thrb1^HA, and Thrb2^HA. Open triangles indicate position of insert, solid rectangles and vertical lines represent coding exons, and open rectangles represent untranslated portions of exons. Long introns (horizontal lines) interrupted by double slash. (B) Relative expression of Thra and Thrb in homozygous knock-in mice livers compared with WT C57BL/6. Gene expression levels were determined by qRT-PCR and normalized to β-actin (n > 4). (C) Knock-in animals showed no significant difference in serum thyroid hormone level or TSH compared with WT. Adult male mice (4–6 months old) of each genotype were used for T3 (n = 3), fT4 (n > 6), and TSH (n > 6) measurement. Error bars are SD. One-way ANOVA was used for statistical analysis. ANOVA, analysis of variance; fT4, free thyroxine; HA, hemagglutenin epitope; qRT-PCR, quantitative reverse transcription PCR; SD, standard deviations; tT3, total triiodothyronine; TSH, thyrotropin; WT, wild type.

FIG. 2.

THR isoform expression in peripheral organs. (A) Representative Western blots (4–20% acrylamide gel, anti-HA antibody) showing HA-THRA1/2 (430 and 512 aa) in the lung, and HA-THRB1 (481 aa) and HA THRA1 in the liver nuclear extracts. No HA-THRB2 (495 aa) was detected in these organs. Histones were used for a loading control. Graphs represent relative band density of THRs from individual animals. Average density of THRA1 was assumed 100%. Both sexes were used for the liver (AM = 6, AF = 4, B2M = 4, B2F = 2, B1M = 6, B1F = 3), spleen (AM = 5, AF = 3, B2M = 3, B2F = 2, B1M = 5, B1F = 3), BAT (AM = 4, AF = 2, B2M = 2, B2F = 2, B1M = 3, B1F = 2), and sWAT (AM = 4, AF = 2, B2M = 2, B2F = 1, B1M = 4, B1F = 2); male tissues were used for the lung, heart, intestine, and pancreas. (B) Representative Western blots (4–20% acrylamide gel, anti-HA antibody) showing HA-THR proteins immunoprecipitated from various peripheral organs. Thrb2^HA/HA mice served as negative controls. Data points on the graph represent relative densities of immunoprecipitated THRs from individual animals with average THRA1 density from the same experiment at 100%. Data points below the line represent failure to detect protein. AF, THRA females; AM, THRA males; B1F, THRB1 females; B1M, THRB1 males; B2F, THRB2 females; B2M, THRB2 males; sWAT, subcutaneous white adipose tissue; THR, thyroid hormone receptor.

FIG. 3.

THR transcripts in mouse organs. (A) “Electronic northern blots” for Thra and Thrb generated form RNA-seq data (34). Expression numbers on the Y-axis correspond to mean RPKM values from two to four males (M) and two to four females (F); individual values are shown as dots. Peripheral organs show no sex-specific difference in transcript levels; however, the low number of animals does not allow statistical analysis. (B) Thrb to Thra transcript ratio calculated from RNA-seq data Thrb RPKM/Thra RPKM (n > 4). (C) Thrb1/Thra1 transcript ratio assessed by qRT-PCR. Y-axis represent 2^−ΔCT. RNA was isolated from HA knock-in animals, which are the same as used for protein analysis (n > 6). Rectangles enclose organs analyzed by both methods. RPKM, reads per kilobase per million.

FIG. 4.

THR isoform expression in the central nervous system. Representative Western blots (4–20% acrylamide gel, anti-HA antibody) showing HA-THRA1/2, and HA-THRB1 in nuclear extracts from the cerebellum, cortex, hypothalamus, olfactory bulb, and spinal cord. THRA2 is a predominant isoform in all tested tissues tested. NS marks nonspecific bands present in all samples. Histones were used for a loading control. Relative band density of THRs from several animals is shown in the graph. Average density of THRA1 was 100%. Each datum point represents an individual animal. Both sexes were used for the cerebellum (AM = 6, AF = 4, B2M = 3, B2F = 2, B1M = 3, B1F = 3), cortex (AM = 3, AF = 2, B2M = 2, B2F = 1, B1M = 3, B1F = 2), and hypothalamus (AM = 4, AF = 2, B2M = 3, B2F = 1, B1M = 4, B1F = 2); males for spinal cord and olfactory bulb. Data points below line represent failure to detect protein.

FIG. 5.

THR isoform expression in mouse pituitary. (A) Western blots (4–20% acrylamide gel, anti-HA antibody) showing HA-THRB2 (495 aa), HA-THRA1/2 (430 and 512 aa), and HA-THRB1 (481 aa) in total protein extracts from individual pituitaries of male (M) and female (F) mice. Relative band density of THRs is shown in the graph. Average density of THRA1 was 100%. (B) Western blot (4–20% acrylamide gel, anti-HA antibody) showing THRB2 expression in females is significantly higher than that in males. (C, D) Relative expression of Thrb is significantly higher in female mice than in male mice of both WT and Thrb ^HA/HA genotypes, while Tshb relative expression is lower in females. Gene expression levels were determined by qRT-PCR and normalized to β-actin. (E) TSH is lower in female mice than in male mice. Younger (5 weeks old) animals show more pronounces difference than adult (5 months old) mice. Results are mean (n > 4), and error bars are SD. Student T-test or one-way ANOVA were used for statistical analysis. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.