The Thyroid Axis Is Regulated by NCoR1 via Its Actions in the Pituitary

Abstract

TSH is the most important biomarker in the interpretation of thyroid function in man. Its levels are determined by circulating thyroid hormone (TH) levels that feed back centrally to regulate the expression of the subunits that comprise TSH from the pituitary. The nuclear corepressor 1 (NCoR1), is a critical coregulator of the TH receptor (TR) isoforms. It has been established to play a major role in the control of TSH secretion, because mice that express a mutant NCoR1 allele (NCoRΔID) that cannot interact with the TR have normal TSH levels despite low circulating TH levels. To determine how NCoR1 controls TSH secretion, we first developed a mouse model that allowed for induction of NCoRΔID expression postnatally to rule out a developmental effect of NCoR1. Expression of NCoRΔID postnatally led to a drop in TH levels without a compensatory rise in TSH production, indicating that NCoR1 acutely controls both TH production and feedback regulation of TSH. To demonstrate that this was a cell autonomous function of NCoR1, we expressed NCoRΔID in the pituitary using a Cre driven by the glycoprotein α-subunit promoter (P-ΔID mice). Importantly, P-ΔID mice have low TH levels with decreased TSH production. Additionally, the rise in TSH during hypothyroidism is blunted in P-ΔID mice. Thus, NCoR1 plays a critical role in TH-mediated regulation of TSH in the pituitary by regulating the repressive function of the TR. Furthermore, these studies suggest that endogenous NCoR1 levels in the pituitary could establish the set point of TSH secretion.

Thyroid hormone (TH) plays a crucial role during development and also has important effects in many target tissues in humans throughout life. To keep TH levels within normal limits, an elegant feedback system has developed such that circulating TH feeds back at the level of hypophysiotrophic neurons in the paraventricular nucleus of the hypothalamus that synthesize TRH and thyrotrophs in the pituitary that produce TSH (1). Indeed, the measurement of circulating TSH levels is the most common test used in humans to determine whether circulating TH levels are normal. Although TH, and in particular T₃, potently repress both TRH gene expression in the paraventricular nucleus (PVN) and TSHα and -β expression in the pituitary via the TH receptor (TR) isoforms, the molecular mechanisms remain unclear. Deciphering this mechanism is critical given the importance of TSH as a biomarker in humans (2–7).

The genomic actions of T₃ require both the TR isoforms and groups of coregulators that interact with the TR either in the presence or absence of ligand. On TR response elements (TRE) of positively regulated genes, the presence of T₃ stimulates the recruitment of coactivators, whereas in its absence the nuclear corepressors NCoR1 and silencing mediator of retinoid and thyroid hormone receptor (SMRT or NCoR2) are recruited (8, 9). Furthermore, it is now clear from a number of in vivo models that NCoR1 and SMRT control T₃ sensitivity on positive TRE but their role in negatively regulated T₃ targets, such as the TSH-subunit genes, remains unclear (10–13).

To determine this, we recently developed a mouse model that globally expresses an NCoR1 allele that cannot interact with TR isoforms (11). Remarkably, these mice have a reset thyroid axis with low circulating TH levels but inappropriately normal serum TSH levels. Although these mice exhibited enhanced sensitivity to TH on a variety of positively regulated target genes, both TSH subunit gene expression in the pituitary and TRH gene expression in the PVN were not more sensitive to TH but appeared to be reset to recognize the low TH levels as normal. These findings raised the possibility that NCoR1 plays a specific developmental role in the set point of the thyroid axis rather than in dynamic negative regulation.

To address this issue and determine where NCoR1 acts to determine the set point of the thyroid axis, we have used two novel mouse models. Importantly, we now demonstrate that NCoR1 acts independently of development to reset the thyroid axis and that NCoR1 acts within the thyrotroph to control TSH secretion. Thus, levels of available NCoR1 within the thyrotroph could be responsible for determining the unique set point of the thyroid axis in individuals.

Materials and Methods

Generation of UBC-ΔID and P-ΔID mice

To obtain a mouse strain with global inducible expression of NCoRΔID protein (UBC-ΔID) we have crossed NCoR^lox/lox animals maintained on a mixed B6;129S background with UBC-Cre transgenic mouse line that expresses modified tamoxifen-inducible Cre-ERT2 recombinase under the control of human ubiquitin C promoter in all tissues [B6.Cg-Tg(UBC-Cre/ERT2)1Ejb/J; The Jackson Laboratory, Bar Harbor, ME] (11). The NCoR^{lox/lox UBC-Cre} (UBC-ΔID) and NCoR^lox/lox (control) animals were given tamoxifen dissolved in corn oil for 5 consecutive days through oral gavage at a dose of 20 mg/100 g of body weight. The experiments were performed at the indicated time points, with the first day of treatment counted as d 1. The level of recombination was assessed by quantitative PCR (Q-PCR) after the mice were euthanized, and only animals with expression level of NCoRΔID over 50% (six of eight animals) were included in the gene expression analysis. To determine the expression of NCoRΔID we used two distinct TaqMan assays (Applied Biosystems, Foster City, CA), one spanning exons 40–41 (deleted in NCoRΔID) and other directed to exons 23–24 [common to both wild-type (WT) and ΔID] (10). Therefore, the first assay detects exclusively WT NCoR, whereas the second assay detects both isoforms (Total NCoR). The expression of NCoRΔID was determined by subtracting the expression of WT NCoR from total NCoR.

The mice with pituitary expression of NCoRΔID (P-ΔID) were generated by crossing the NCoR^lox/lox animals with mice that express Cre driven by the promoter of glycoprotein hormone α-subunit (α-GSU) gene (14).

Animal experiments and sample collection

All experiments were approved by Beth Israel Deaconess Medical Center (BIDMC) Institutional Animal Care and Use Committee.

The experiments were performed in age- and sex-matched mice between 8 and 17 wk of age unless otherwise specified. Animals were housed in the BIDMC animal facility on a 12-h light, 12-h dark cycle and given standard rodent chow (Harlan Teklad F6 Rodent Diet 8664; Harlan Teklad, Placentia, CA) and water ad libitum. At the end of experiments the mice were euthanized by asphyxiation with CO₂. Blood samples were taken by cardiac puncture and collected in EDTA-treated tubes. Plasma was separated by centrifugation and stored at −80 C. Tissues were rapidly collected, flash frozen in liquid nitrogen, and stored at −80 C.

Hypothyroidism was induced by feeding the animals with a low iodine diet supplemented with 0.15% propylthiouracil (PTU) (LoI/PTU, Harlan Teklad TD.95125) for indicated periods of time. The T₃ (Sigma Chemical Co., St. Louis, MO) was given as ip injections once a day at specified doses and periods of time.

TRH response test

Control and P-ΔID mice were given a single injection of TRH peptide (Ferring GmbH, Kiel, Germany) at a dose of 10 μg/kg ip. Blood samples for TSH measurement were obtained before (time 0), 15 min, and 2 h after the injection.

Hormonal analysis

Total T₄ (TT₄), total T₃, and free T₄ levels were measured by solid-phase RIA (Coat-a-Count; Diagnostic Products Corp., Los Angeles, CA) in 25 and 50 μl of plasma, respectively. Circulating TSH level was also measured in plasma, using Millipore TSH Multiplex kit (Millipore Corp., Bedford, MA).

Real-time Q-PCR

Total RNA was extracted from frozen tissues with STAT-60 reagent (Teltest) according to the manufacture's instructions. Pituitaries (0.2 μg) or 0.5 μg (livers) of total RNA were reverse-transcribed using Advantage RT-for-PCR kit (CLONTECH Laboratories, Inc., Palo Alto, CA) with random hexamer primers. TaqMan Gene Expression Assays were purchased from Applied Biosystems. Each reaction contained 15–30 ng cDNA, 5 μl of TaqMan Universal PCR Master Mix, No AmpErase UNG (Applied Biosystems), and 0.25 μl of TaqMan Gene Expression Assay in a total volume of 10 μl. Q-PCR were performed in duplicates using the 7900HT Fast Real-Time PCR System (Applied Biosystems). Relative mRNA levels were calculated using standard curve method and normalized to cyclophilin mRNA.

Western blot analysis

Western blots were performed using whole-cell protein extracts as previously described (11). Briefly, 30–50 mg of frozen livers or two pooled pituitaries from genotype and sex-matched animals were used for protein isolation. total protein (10–30 μg) was resolved on 3–8% gradient Tris-Acetate or 10% Bis-Tris Novex gels (Invitrogen, Carlsbad, CA) and probed with antibodies for C-terminal portion of NCoR, and RNA-Polymerase II (05–952; Upstate Biotechnology Inc., Lake Placid, NY).

Immunohistochemistry

Control and P-ΔID mice were euthanized via transcardiac perfusion under ketamine/xylazine anesthesia. Mice were then perfused with saline followed by 10% neutral buffered formalin, and pituitaries were dissected, postfixed in 10% neutral buffered formalin for 4–6 h, and stored in 20% sucrose in PBS at 4 C overnight. After that the pituitaries were frozen in TBS Tissue Freezing Medium (American Master Tech Scientific, Lodi, CA), sectioned on a sliding cryostat (Leica CM 3050; Leica Microsystems, Wetzlar, Germany) at −18 C, and mounted on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA). Fluorescein isothiocyanate (FITC) conjugated anti-green fluorescent protein (GFP) goat (Abcam, Cambridge, MA) was used at 1:500 dilution, and a rabbit anti-TSH (NHPP, NIDDK) was diluted 1:1000. Alexa Fluor 594 antirabbit antibody (Invitrogen Molecular Probes) was applied as a secondary antibody at a dilution of 1:250, and the slides were coverslipped using Vectashield Mounting Medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA). Digital images were acquired using Zeiss Axioimager. Z1 with Axiovision 4.5 software (Carl Zeiss, Oberkochen, Germany).

Statistical analysis

Statistical analysis was performed using Prism 5 program. The differences between genotypes were tested using the unpaired Student's t test. Effect of tamoxifen treatment on different genotypes was tested using repeated measures two-way ANOVA with the Bonferoni post hoc test.

Results

NCoR1 regulates the hypothalamus-pituitary-thyroid (HPT) axis function in adult mice

We have previously shown that the global expression of the NCoRΔID results in reduced TH levels and inappropriately normal TSH levels (11). To determine whether the inappropriate TSH levels seen were developmental in NCoRΔID mice, we crossed mice homozygous for a floxed NCoR1 (NCoR^lox/lox) allele that generates NCoRΔID in the presence of the Cre-recombinase with mice expressing a tamoxifen-inducible Cre under control of the ubiquitin C promoter (UBC-Cre) (15). In the presence of tamoxifen, the NCoR1 allele would thus be altered in UBC-Cre NCoR^lox/lox mice (UBC-ΔID), generating the NCoRΔID allele in virtually every tissue, but avoiding any effect this may have during development. As shown in Fig. 1, the UBC promoter was able to drive Cre expression in the liver and pituitary, in addition to the tissues previously described (15). We determined NCoR1 recombination by measuring the exact amount of NCoRΔID and NCoR1 mRNA in both the liver and pituitary after the administration of tamoxifen. Although the level of recombination was different across mice, most had more than 50% of recombination (Fig. 1A). Those with less than 50% of recombination (more than 50% of WT NCoR expression) were excluded from further analysis. It is worth noting that the recombination level is very similar between different tissues at the mRNA level (Fig 1A) and that NCoRΔID protein expression corresponds to the mRNA level found in each mouse (Fig. 1B).

Fig. 1.

Efficiency of tamoxifen-induced UBC-Cre-mediated recombination in the pituitary and liver of UBC-ΔID female mice. A, NCoR1 mRNA levels were measured to assess the recombination efficiency in the pituitary and liver of UBC-ΔID mice 36 d after tamoxifen treatment using Q-PCR to estimate relative amounts of NCoR1 and NCoRΔID. Shown are data for all animals in the group including those that were not included in the final analysis (with expression levels of NCoRΔID < 50%). Note that NCoRΔID expression is estimated by subtracting WT from total NCoR expression (n = 6–8 animals per group). B, Total protein lysates were isolated from livers of female mice with indicated genotypes 36 d after tamoxifen treatment. The UBC-NCoRlox/lox mice are representative of those with different amounts of recombination. Protein (30 μg) was loaded per lane and subjected to Western analysis using anti-NCoR1 antibody and anti-RNA Polymerase II antibody (loading control). ***, P < 0.001 as determined by unpaired t test.

We next evaluated TT₄ levels in UBC-Cre NCoR^lox/lox (UBC-NCoRΔID) mice and NCoR^lox/lox control mice at 6 wk of age (before tamoxifen treatment) to ensure that they were similar (Fig. 2A). Subsequent to this, both groups received tamoxifen orally for 5 d and TT₄ levels were monitored over the ensuing 36 d. Remarkably, by d 28, UBC-NCoRΔID mice that received tamoxifen had TT₄ levels that were 40% lower than controls (Fig. 2A). Similar differences were seen at the end of the experiment at 36 d. Furthermore, the TT₄ positively correlated with the pituitary expression of the full-length NCoR1, suggesting that the higher recombination rate leads to an accentuated drop in TT₄, in the UBC-NCoRΔID group (Fig. 2B). Interestingly, there was no significant difference in total T₃ levels between the groups but free T₄ levels were also lower in tamoxifen-treated UBC-NCoRΔID mice (Fig. 2, C and D). Despite reduced circulating T₄ levels, the pituitary TSH subunit gene expression did not increase in UBC-ΔID mice (Fig. 2E). Thus, the thyrotroph in UBC-ΔID mice recognizes the lower T₄ levels as normal and appears to reset acutely and is not the result of a developmental effect. To determine whether the expression of NCoRΔID in UBC-ΔID mice affected TH signaling in other tissues, we investigated the pattern of gene expression in the liver. As we have seen previously, in NCoRΔID mice, the hepatic TH target genes were up-regulated in UBC-NCoRΔID mice consistent with increased sensitivity to TH despite the lower circulating T₄ levels (Fig. 2F).

Fig. 2.

Circulating TH levels and pituitary mRNA expression in UBC-ΔID mice after tamoxifen treatment. A, Serum TT₄ levels were measured in control and UBC-ΔID at the indicated time points after administration of tamoxifen. B, TT₄ levels positively correlate with the expression of WT NCoR1 mRNA 36 d after tamoxifen treatment. C–E, Concentrations of circulating free T₄ (FT₄) (panel C) and total T₃ (D), as well as mRNA expression levels of TSH subunits, Trhr and Dio2 (E) in the pituitary of control and UBC-ΔID mice were measured on d 36 after tamoxifen treatment. mRNA levels were quantified by Q-PCR and calculated relative to the control group. F, Expression of known TH targets Thrsp, Dio1, and Fasn in the liver of UBC-ΔID mice 36 d after tamoxifen treatment. mRNA expression levels were quantified by Q-PCR and presented relative to control group. All data presented are from the same cohort of mice (n = 6 animals per group). Data are presented as mean ± sem; **, P < 0.01; ***, P < 0.001 as determined by two-way repeated measures ANOVA (panel A) or unpaired t test (panels C–F).

Expression of NCoRΔID selectively in the anterior pituitary

To determine whether NCoR1 has a cell-autonomous effect in the thyrotroph in mediating the inappropriately normal expression of TSH subunit genes seen both in NCoRΔID and UBC-NCoRΔID mice, we next interrogated its direct role in the regulation of the TSH subunits. To do this, we generated a mouse model that expresses NCoRΔID selectively in cells of the anterior pituitary (P-ΔID) including the thyrotroph using a transgenic line that expresses the Cre-recombinase under the control of the αGSU promoter (14). To further validate the aGSU-Cre line we crossed it with an enhanced GFP reporter strain. As shown in Fig 3A and Supplemental Fig. 1 (published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org) many cells in the anterior pituitary are targeted by αGSU-Cre including a significant subset of TSH expressing thyrotrophs. Importantly, ectopic expression of αGSU-Cre is not seen in the PVN of the hypothalamus or in follicular cells in the thyroid (Supplemental Fig. 2), two other cell types that play a key role in the regulation of TH levels. Finally, to ensure that the aGSU-Cre transgene could lead to the expression of NCoRΔID in the pituitary, we analyzed protein and mRNA from P-ΔID and control mice. Western analysis shows the presence of the NCoRΔID protein in pituitaries of P-ΔID mice (Fig. 3B). In addition, the expression of WT NCoR1 is reduced in pituitaries of P-ΔID mice, but the total expression of NCoR1 (WT + NCoRΔID) has not changed (Supplemental Fig. 1C). Although Western analysis of pituitary NCoR1 expression shows only the presence of one major isoform, we also determined whether a recently described NCoR1 isoform that lacks the most N terminal of the receptor-interacting domains is expressed at the mRNA level (16). As shown in Supplemental Fig. 3, the majority of NCoR1 present in the pituitary of wild-type mice likely contains all three NCoR1 receptor-interacting domains.

Fig. 3.

Assessment of localization and efficiency of Cga-Cre-mediated recombination in the pituitary. A, Cga-Cre mice were crossed to a reporter GFP strain that expresses GFP protein in all cells in which Cre activity has been present. Horizontal pituitary sections from resulting Cga-GFP animals were stained with anti-GPF and anti-TSHβ antibodies. Obtained images were merged to assess the colocalization of the Cre-recombinase and TSHβ expression. Original magnification ×40. B, Whole-cell lysates were isolated from pituitaries of female and male control and P-ΔID mice. Protein (20 μg) isolated from two pooled pituitaries were subjected to Western analysis using anti-NCoR1 antibody.

P-ΔID mice have suppressed HPT axis activity

To initially evaluate the role of NCoR1 in the pituitary we analyzed thyroid function in P-ΔID mice and controls. P-ΔID females had a reduction in circulating TT₄ and T₃ but by only 21% and 16%, respectively, compared with the 40% in global NCoRΔID mice (11). Interestingly, no statistical difference was seen in males (Fig. 4A). TSH subunits mRNAs were down-regulated in female P-ΔID mice despite lower TH levels and followed the same trend in males (Fig. 4C). However, there was no detectable change in circulating TSH levels (Fig. 4B). Thus, the expression of NCoRΔID in the pituitary alone causes reduction in circulating TH levels, similar to the phenotype seen in mice with global expression of NCoRΔID (11). Because this phenotype was more evident in females, further experiments were conducted using only this gender.

Fig. 4.

Circulating TH and TSH levels and pituitary expression of TSH subunit mRNA in P-ΔID mice. A, TT₄ and total T₃ (TT₃) levels in control and P-ΔID female (n = 9–11) and male (n = 7–8) mice. B, Plasma TSH levels in female control and P-ΔID mice (n = 9–11). C, Expression of TSH subunit mRNA was measured in the pituitaries of mice of indicated sex and genotype by Q-PCR. The mRNA levels are calculated relative to the control group. Data are presented as mean ± sem; *, P < 0.05 as determined by unpaired t test.

Because TSH subunit gene expression appeared to respond inappropriately to low TH levels in P-ΔID mice, we stressed the axis further by inducing hypothyroidism with a low iodine diet supplemented with PTU in P-ΔID mice and controls. It is known that by using this diet TT₄ levels can be lowered below detection limits within 21 d (data not shown). Although circulating TSH was able to rise in over a 63-d period in P-ΔID mice, its levels were markedly reduced compared with controls by up to 1 order of magnitude (Fig. 5A). Furthermore, activation of both αGsu and Tshβ expression was also clearly impaired at both 21 and 63 d in P-ΔID mice (Fig. 5, B and C). We also looked at an earlier time point in another cohort and indeed the rise in circulating TSH appeared to be blunted as early as 10 d after the induction of hypothyroidism (Supplemental Fig. 4A). To ensure that the effects seen were specific for the thyrotroph, we analyzed the expression of other pituitary hormones including the β-subunits of LH and FSH and found no distinguishable differences between P-ΔID and control animals except for the activation of GH expression in the hypothyroid state in P-ΔID mice. This has been seen previously in NCoRΔID mice and is consistent with its being a positive TH target and thus being activated in the absence of functional NCoR (Supplemental Fig. 4B). Finally, to rule out any role of the Cre transgene on the thyroid axis, we compared the phenotype of αGSU-Cre and WT mice on the same genetic background as P-ΔID mice and found no effect of Cre on TSH subunit expression in the euthyroid or hypothyroid state (Supplemental Fig. 5).

Fig. 5.

Plasma TSH levels and pituitary expression of TSH subunit mRNA in P-ΔID mice during induction of hypothyroidism and T₃ replacement. A, Female control and P-ΔID mice were put on low iodine/PTU diet to induce hypothyroidism. Circulating TSH was measured at indicated time points after the start of the diet (n = 6). B and C, Pituitary expression of αGSU (B) and TSHβ (C) mRNA after 21 d and 63 d of low-iodine/PTU diet in female control and P-ΔID animals was measured by Q-PCR. The mRNA levels are calculated relative to the control group at baseline (chow diet). D, Plasma TSH concentration was measured in the indicated groups of female mice after 21 d of low-iodine/PTU diet, and 2- to 4-d periods of treatment with increasing concentrations of T₃ as shown (n = 7–8). All values are presented as mean ± sem; *, P < 0.05; **, P < 0.01; ***, P < 0.001 as determined by two-way repeated measures ANOVA (A and D) or unpaired t test (B and C).

Because P-ΔID mice have an impaired response to hypothyroidism, we next evaluated pituitary sensitivity to T₃. After 21 d on a PTU diet, the mice were treated with two increasing doses of T₃, for 4 d each (Fig. 5D). As expected, the circulating TSH concentration in the P-ΔID mice was only 50% of that of the control group after 21 d of the PTU diet. Low-dose T₃ treatment had limited effects on TSH levels in both groups of mice but P-ΔID mice continued to have lower TSH levels. With a higher dose, which usually results in normal circulating T₃ levels, circulating TSH was suppressed in a similar way in both groups such that there was no significant difference between them (Fig. 5D).

TRH receptor 1 (Trhr) is negatively regulated by TH and represents a putative mediator of NCoR1 effect on HPT axis function

Given the multifactorial regulation of thyrotroph function, we looked further to understand how NCoR1 action is required for TSH synthesis and release. An important component of the HPT axis is TRH action, which stimulates TSH synthesis and release through TRHR1 (Trhr gene) receptors on thyrotrophs (17, 18). As shown in Fig. 6, Trhr expression is decreased by 40% in euthyroid P-ΔID mice and whereas its levels rise by approximately 40-fold in hypothyroidism the rise is defective in P-ΔID mice (Fig. 6A). To further determine whether the decrease in Trhr plays a role in the impaired TSH response of the thyrotroph in P-ΔID mice we treated euthyroid and P-ΔID mice with a single dose of TRH (10 μg/kg, ip) and determined serum TSH levels at subsequent time points. Within 15 min of TRH injection, plasma TSH increased by approximately 60-fold in the control group, but this response was reduced by almost 50% in P-ΔID mice. Circulating TSH concentration approached the preinjection values 60 min after injection and was similar in both groups (Fig. 6B).

Fig. 6.

Assessment of possible indirect mechanisms of the regulation of TSH levels in the P-ΔID animals. A, Trhr mRNA levels in euthyroid and hypothyroid (21 d and 63 d of low-iodine/PTU diet) control and P-ΔID mice. B, Animals with indicated genotypes were given 10 μg/kg of recombinant TRH ip. Plasma TSH levels were measured before and 15 and 60 min after the injection (n = 9). C, Dio2 and Dio3 mRNA expression in euthyroid and hypothyroid control and P-ΔID mice. All mRNA expression levels (A, C, and D) were quantified by Q-PCR and are presented relative to euthyroid control group (set to 1). All values are presented as mean ± sem; **, P < 0.01; ***, P < 0.001 as determined by two-way repeated measures ANOVA followed by Bonferroni posttest.

In addition to TRH signaling, the deiodinases are potential modulators of TSH expression, because of the tight regulation of intracellular T₃. Thus, it was important to determine whether changes in the expression of those genes (Dio2 and Dio3) could contribute to the phenotype of P-ΔID mice. However, there was no difference in either type 2 or type 3 deiodinase mRNA expression in the euthyroid state in P-ΔID mice (Fig. 6C). Although Dio2 mRNA expression was lower in P-ΔID than in control group during hypothyroidism (Fig. 6C), this cannot explain the suppressed TSH because decreased Dio2 would lead to decreased intracellular T₃. In addition, T₃ action depends on TR isoforms, and changes in their expression can potentially modify thyrotroph function. Nevertheless, there was no difference in the pituitary expression of Thra or Thrb isoforms (Supplemental Fig. 4C). In addition to TR, other transcription factors are known to control Tshβ expression, such as Gata2 and Pou1f1, but their expression was similar in both groups studied (Supplemental Fig. 4D) (19).

Discussion

The mechanism underlying negative regulation by TH has long been controversial. The importance of understanding this mechanism is underscored by the almost universal use of the circulating TSH as the biomarker to interpret TH action in humans. However, if the mechanism underlying negative regulation is different than positive regulation, circulating TSH levels may not be the best indicator of TH action. Previous in vivo studies have established an absolute requirement of the TRβ isoform interacting with still to be defined negative TRE to mediate negative regulation of TSH subunit genes and TRH (3–6). Additionally, the steroid receptor coactivator (SRC)-1 has been shown to play a paradoxical role in T₃-mediated repression because both SRC-1 knockout mice and mice that express a mutant TRβ that cannot recruit SRC-1 are unable to repress TSH-subunit gene expression normally in the presence of T₃ (20, 21). A role for NCoR1 in negative regulation has been suggested by our recent work in which global expression of NCoRΔID in lieu of NCoR1 in mice leads to inappropriately normal TSH levels despite low thyroid normal levels, suggesting that NCoR1 may play a role in the activation of TSH-subunit expression in the presence of low levels of TH (11). This role is also supported by our recent findings demonstrating that the global expression of NCoRΔID lowers TSH levels in mice that express a mutant TRβ that causes resistance to TH (22). However, neither of these studies addressed whether the role of NCoR1 was developmental or acute or, importantly, whether NCoR1 could act in a cell-autonomous fashion to control TSH synthesis and secretion.

To address the functional and developmental role of NCoR1 in establishing the set point of the thyroid axis, we used a Cre/lox system that allowed us globally to activate the expression of NCoRΔID beginning at 6 wk of age. Herein, we demonstrate that production of NCoRΔID in lieu of NCoR1 leads to a fall in T₄ levels within 3 wk. Remarkably, TSH subunit mRNA levels do not rise. Importantly, the phenotype of these mice correlated directly with fidelity of Cre activation and the production of NCoRΔID. Because there was no detectable change in TSH subunits expression in UBC-ΔID mice, it is possible that the drop in circulating T₄ concentration also results from a primary thyroid effect. Indeed we have demonstrated that the global expression of NCoRΔID impairs thyroidal TH production, despite normal thyroid morphology (11). The fact that circulating T₃ concentration did not significantly decrease in the UBC-ΔID mice may indicate the animals were able to keep producing this hormone in lieu of T₄ throughout the time course of the experimental protocol. Also, it is possible that expression of NCoRΔID during development promotes a more profound effect on thyroid function than seen in UBC-ΔID mice. Still, even if this were the case, it does not discount the fact that the pituitary fails to respond to the lower TH levels in UBC-ΔID mice.

Given that NCoR1 acts acutely to influence the set point of TSH production, we next addressed whether NCoR1 function in the pituitary alone was sufficient to recapitulate the phenotype seen in mice expressing NCoRΔID in all tissues. Indeed, female P-ΔID mice have lower circulating TH levels with lower levels of TSH subunit gene expression at baseline, although serum TSH appears normal. This lower expression of TSH subunit genes in P-ΔID animals is in contrast with both constitutive and inducible global ΔID models in which TSH expression was not affected. This can potentially be explained by differences in regulation of TSH by central inputs other than TH as they may be modified in the whole-body ΔID animals. Because NCoR1 action becomes more relevant at lower concentrations of TH, we induced a hypothyroid state and investigated the changes in plasma TSH and pituitary gene expression. Remarkably, in contrast to what is seen in SRC-1 knockout mice, the hypothyroidism-induced increase in TSH levels is impaired in P-ΔID mice (20). Thus, we demonstrate in vivo that similar to the regulation of positive TH targets, the NCoR1-TR interaction is most relevant for the regulation of a negative target when TH concentrations are low.

Based on our work on positive TR targets, it would be expected that the lack of interaction between NCoR1 and TR favors the recruitment of coactivators and would thus make the thyrotrophs more sensitive to TH. Nevertheless, when hypothyroid mice were treated with increasing doses of T₃ the observed drop in plasma TSH was similar in both P-ΔID and control groups, suggesting a similar sensitivity to T₃. Therefore, it seems that the lack of functional NCoR does not increase the sensitivity of thyrotrophs, but rather resets their function to a lower level of TSH production. These data strongly demonstrate that NCoR1 is required for the full activation of TSH subunit genes in the presence of limiting amounts of T₃. Thus, NCoR1 appears to play a paradoxical role in the regulation of TSH subunit gene expression leading to their activation. Interestingly, mice that globally express NCoRΔID do not have suppressed TSH subunit gene expression at baseline or when made hypothyroid, suggesting that the global expression of NCoRΔID allows for compensatory input that elevates TSH subunit gene expression and thus serum TSH.

Our data suggest that one compensatory factor that allows for this is the TRH-R1. Previous work has established that the TRH-R1 is a negative T₃ target allowing for the full induction of TSH levels in hypothyroidism through the enhanced production of TRH in the PVN (23–25). Interestingly, TRH-R1 expression is clearly impaired in P-ΔID mice both in the basal state and when rendered hypothyroid. However, TRH-R1 expression was not impaired in mice that globally express NCoRΔID. The physiological relevance of the decreased levels of expression of TRH-R1 is made clear by the diminished TSH response to exogenous TRH in P-ΔID mice. Thus, the pituitary-specific expression of NCoRΔID prevents the activation of TRH-R1 by lower levels of T₃ leading to diminished levels of TSH subunit gene expression and likely lower amounts TSH production overtime and the lower levels of TH seen. It remains to be determined why TRH-R1 expression is not affected in mice that globally express NCoRΔID.

Although these data strongly suggest a direct role for NCoR on the control of thyrotroph function, we cannot rule out the possible effect on other pituitary cell types, such as stellate cells. Unfortunately, there is no mouse model that would allow us to target the thyrotrophs exclusively. A study using a new TSHβ-Cre mouse has been recently published and, although it targets many thyrotrophs, other pituitary cell types are also affected (26).

In summary, we have demonstrated that NCoR1 plays an active role in establishing the set point of the thyroid axis in adult animals. Furthermore, NCoR is required for the activation of TSH subunit gene expression in the thyrotroph in the presence of limiting amounts of T₃. This likely occurs both by regulating the response to TRH by controlling expression of the TRH-R1 and also by directly regulating TSH subunit gene expression. Taken together these data demonstrate, for the first time, a role for NCoR1 in vivo in negative regulation and suggest that available functional NCoR1 in the thyrotroph could explain different set points of the thyroid axis across individuals.