TR Alpha 2 Exerts Dominant Negative Effects on Hypothalamic Trh Transcription In Vivo

Mammalian thyroid hormone receptors (TRs) have multiple isoforms, including the bona fide receptors that bind T3 (TRα1, TRβ1 and TRβ2) and a non-hormone-binding variant, TRα2. Intriguingly, TRα2 is strongly expressed in the brain, where its mRNA levels exceed those of functional TRs. Ablation of TRα2 in mice results in over-expression of TRα1, and a complex phenotype with low levels of free T3 and T4, without elevated TSH levels, suggesting an alteration in the negative feedback at the hypothalamic-pituitary level. As the hypothesis of a potential TRH response defect has never been tested, we explored the functional role of TRα2 in negative feedback on transcription of hypothalamic thyrotropin, Trh. The in vivo transcriptional effects of TRα2 on hypothalamic Trh were analysed using an in vivo reporter gene approach. Effects on Trh-luc expression were examined to that of two, T3 positively regulated genes used as controls. Applying in vivo gene transfer showed that TRα2 over-expression in the mouse hypothαlamus abrogates T3-dependent repression of Trh and T3 activation of positively regulated promoters, blocking their physiological regulation. Surprisingly, loss of function studies carried out by introducing a shTRα2 construct in the hypothalamus also blocked physiological T3 dependent regulation. Thus, modulating hypothalamic TRα2 expression by either gain or loss of function abrogated T3 dependent regulation of Trh transcription, producing constant transcriptional levels insensitive to feedback. This loss of physiological regulation was reflected at the level of the endogenous Trh gene, were gain or loss of function held mRNA levels constant. These results reveal the as yet undescribed dominant negative role of TRα2 over TRα1 effect on hypothalamic Trh transcription.


Introduction
Thyroid hormone (TH) production is controlled by the hypothalamic peptide Thyrotropin Releasing Hormone (TRH). T 3 exerts negative feedback on Trh transcription mainly through the beta forms of the thyroid receptors (TRb1 and TRb2) [1,2]. TRs are ligand-dependent transcription factors [3], produced from two genes: NR1A1 and NR1A2 [4,5]. Each gene gives rise to two major isoforms, respectively TRa1 and a2, and TRb1 and b2, by alternative splicing. Both RNA [6,7] and protein [8,9] for each isoform are found in the hypothalamic paraventricular nucleus (PVN), site of TRH regulation.
In mammals, TRa2 is identical to TRa1 in its N-terminus, but the C-terminus is entirely different rendering TRa2 unable to bind T 3 [10,11] and altering the ability of TRa2 to interact with coactivators and co-repressors [12,13]. As TRa2 can bind DNA, but not activate transcription, it has been suggested that TRa2 may act as a dominant-negative receptor. In vitro, TRa2 blocks the activity of other TRs by competing for TR binding to thyroid hormone response elements (TREs) on DNA [14][15][16] or via mechanisms that do not require TRE binding [17]. TRa2 is widely expressed and in brain, its RNA levels greatly exceed those of the functional TRs, especially in perinatal period [18]. Moreover, TRa2 is highly conserved in human, rat and mouse, but is absent in non mammalian vertebrates [19], suggesting an important function for this protein in mammals.
Generation of mutant mice lacking TRa2 has contributed to understand the roles of TRa2 on T 3 -dependent regulation of target genes in the brain [20]. In these TRa2 -/mice, TRa2 ablation results in TRa1 over-expression in brain tissue, and lower levels of free T 3 and T 4 but normal levels of TSH. This failure of TSH to adjust to the lower circulating T 3 and T 4 levels can be accounted for either by an effect at the level of the thyroid gland reducing hormone production, and/or an alteration in the negative feedback at the hypothalamic-pituitary level, which may also include a defect in TRH response. However, this latter hypothesis has never been tested. Previous studies on TRa2 function in brain have attributed a general dominant negative effect of TRa2 but never addressed its transcriptional effects on target genes in vivo, because of the technical challenge it represents. Thus we employed a synthetic gene transfer method in which our laboratory has a great expertise to follow the effects of TRa2 gain or loss of function on Trh gene transcription using positively regulated T 3 genes (Malic and Tyrosine hydroxylase Enzymes; respectively, ME and TyrH) as controls. This in vivo transfection assay provides for tissue specific physiological regulation of transcription in integrated contexts [1,2]. We used the newborn mouse brain as a model system as it was successfully used to analyse the molecular basis of thyroid hormone dependent effects of Trh transcription in vivo [1,2,21] and mainly because that every transcriptional regulation we have identified by this method has later been ratified by experiments in adult transgenic mice.
In vivo over-expression experiments show that in the hypothalamus, TRa2 acts as a dominant-negative receptor, blocking transcription of both positively and negatively T 3 regulated target genes. Moreover, transient TRa2 knockdown seems to reveal TRa1 effect on Trh promoter, the regulation of which being equivalent to the one observed when TRa1 is over-expressed. This hypothesis was emphasised by a decrease in circulating T 4 following TRa1 gain of function. Interestingly, both gain or loss of TRa2 function seems to block Trh transcription at an intermediate level between activated and repressed control levels. Indeed, an average TRH activity remains, whereas fine physiological T 3 regulation is lost. Taken together, these results reveal the physiological importance of TRa2, naturally acting as dominantnegative receptor on hypothalamic Trh transcription in vivo.

Ethics Statement
All aspects of animal care and experimentation were in accordance with the National institutes of Health Guidelines for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of the Animal Protection and Health, Veterinary Services Direction, Paris, France.
In Vivo Gene Transfer (iGT) and Luciferase Assays DNA/PEI (polyethylenimine) complexes, iGT and luciferase assays were carried out as described previously [1]. Given the highly tissue-specific nature of Trh transcription, one of the most important steps in ensuring reproducibility is careful and consistent injection, followed by precise dissection of the hypothalamic areas transfected [1]. Briefly, pups were anesthetized by hypothermia on ice and transfected on post-natal day 2. A glass micropipette was lowered 2 mm through the skull, 0.5 mm posterior to bregma on the sagittal suture, into the hypothalamic area. Two-day-old hypothyroid newborn mice were transfected in the hypothalamic region of the brain with 262 mL of Trh-luc, or ME-tk-luc or TyrH-luc (1 mg/pup) complexed with PEI. To assess the effect of TR overexpression, in addition of the reporter genes, we added pSG5-TRa1, pSG5-TRa2, or empty pSG5 expression vector in the complexes at 100 ng/pup. Luciferase activity was assayed 18 h after transfection. In shRNA experiments, we added small hairpin expression vector (see section plasmids) at a 100 ng/ mL concentration (400 ng/pup). After 48 h, pups were decapitated, and hypothalami were dissected out for luciferase assays following the manufacturer's protocol (Promega). Luciferase activity was measured 48 h later to allow for shRNA expression.
For qPCR analysis, pups were only transfected with either the overexpression vectors or small-hairpin RNA vectors. Transfections were performed in 2 days old pups and the hypothalami were dissected at 1, 3 and 5 days post-transfection for overexpression experiments, and at 36 h post transfection (3.5 days) for sh experiments.

Animal treatments
To assess T 3 effects on reporter gene expression, pups were injected subcutaneously, with 2.5 mg/g of body weight (bw) of T 3 (Sigma-Aldrich, St Quentin Fallavier, France) in 0.9% saline, immediately after transfection. This quite high dose of T 3 is necessary to observe Trh gene regulation in the hypothalamus of newborn mice, because global metabolic rate is high at this developmental stage and the brain is a resistant organ to excess of TH levels [7,24]. Controls received the same volume of 0.9% saline. In the shRNA experiments, this procedure was repeated 24 h after transfection.

Measurement of total plasma T 4
Frozen plasma was thawed and processed according to the supplier's instructions, using the AMERLEX-M T 4 RIA Kit (Trinity Biotech, Wicklow, Ireland). Results are expressed as means 6 SEM.

Immunoblot analysis
Western blot analysis was made on hypothalamus protein extract from brains transfected with CMV H1-shTRa2. Briefly, the hypothalami of two distinct mice were collected for each group under a stereo-microscope. The whole experiment was repeated twice. Tissues were lysed mechanically and proteins were extracted in RIPA buffer (150 mM NaCl, 1% Triton X-100, 0.5% Sodium deoxycholate, 0.1% SDS, 50 mM Tris pH 8) according to the manufacturer's instructions. Protein content was determined by Qbit assays (Invitrogen). Total cell lysates (30 mg) were fractionated by SDS-PAGE 4-20% (Pierce) and transferred to nitrocellulose membranes (Biorad). Membranes were blocked with 5% non fat milk in Tris-buffer saline (TBS; 10 mM Tris-HCl, pH 7.5, 150 mM NaCl), followed by overnight incubation at 4uC with the indicated antibody diluted in TBS with 0.05% Tween-20 (TBS-T). After three washes with TBS-T, membranes were incubated with the appropriate secondary antibody coupled to peroxidase, and immunocomplexes visualized by enhanced chemiluminescence (ECL plus from GE Healthcare Amersham) according to manufacturer's instructions. Primary antibodies for Western-blotting, included rabbit polyclonal anti-TRa2 (1:100; Millipore), rabbit polyclonal anti-bACTIN (1:3000; Sigma). Secondary antibody was anti-rabbit IgG Peroxidase Conjugate from Sigma. Chemiluminescence was revealed by film exposure.

RNA extraction and cDNA synthesis
Hypothalami were dissected from individual newborn mice (transfected either by overexpression vector or small-hairpin RNA vector (see section in vivo gene transfer) under stereo-microscope (limits for hypothalamic dissection: posterior to the optic chiasma, anterior to the mammillary bodies, along both lateral sulcus and 1 mm in depth) and kept in ''RNAlater'' (Ambion Inc, Austin, TX, USA) until extraction. RNA extraction was performed using RNAble reagent following manufacturer's protocol (Eurobio, Les Ulis, France). Concentration (A260) of the total RNA was determined and RNA was stored in Tris 10 mM/EDTA 0.1 mM (PH 7.4) at 280uC.
Prior to qPCR, 2 mg of total RNA were reverse-transcribed using Superscript II Rnase H-reverse transcriptase (Invitrogen, Carlsbad, CA, USA) according to manufacturer's protocol. Control reactions without reverse-transcriptase were done in parallel.

Primers
18S primers and TaqMan probe were provided in the kit Eukaryotic 18S rRNA Endogenous Control (VIC/MGB Probe, Primer Limited) from Applied Biosystems, Warrington, UK. Trh primers were described in [7].

Quantitative polymerase chain reaction (QPCR)
Direct detection of the PCR product was monitored by measuring the increase in fluorescence generated by the TaqMan probe (18S) or by the binding of SYBR Green to dsDNA (Trh). For Trh, 2 ml of cDNA were added to a mix containing Trh primers (300 nM), and 2x SYBR Green Master Mix (Applied Biosystems) to a final reactional volume of 20 ml. For 18S RNA (endogenous control), samples containing 2 ml of cDNA, 1 ml of 18S probe and 10 ml of 2x TaqMan R universal PCR Master Mix (Applied Biosystems) were prepared in a final volume of 20 ml. The genespecific PCR products were measured continuously by means of ABI PRISM 7300 Sequence Detection System (Applied Biosystems) during 40 cycles. All experiments were run in duplicate, and the same thermal cycling parameters were used (95uC for 10 min (1 cycle), 95uC for 15 sec and 60uC for 1 min (40 cycles)). Nontemplate controls and dissociation curves were used to detect primer-dimer conformation and non-specific amplification. According to the widely accepted MiQE (Minimum Information for Publication of Quantitative Real-Time PCR Experiments, [25]) guidelines, we verified the efficiency of the PCR for trh set of primers by using a serial 10 times dilution of the template. The dynamic range covered four orders of magnitude. We determined amplification efficiency from the slope of the log-linear portion of the calibration curve. The resulting efficiency was close to 100% (97%). Given this efficiency, which is required to be able to use the ddCT relative quantification method, we can deduce that we can use these primers for standard quality qPCR studies. The threshold cycle (CT) of each target product was determined and DCT between target and endogenous control was calculated. The CT is the number of PCR cycles required for the fluorescence signal to exceed the detection threshold value. The detection threshold was set to the log linear range of the amplification curve and kept constant for all data analysis. The difference in DCT values of two genes (DDCT) was used to calculate the fold difference (F = 2 -DDCT ). The relative quantitative results were used to determine changes in Trh gene expression in groups where TRa1 or TRa2 was overexpressed as compared to control samples (empty pSG5 vector) at the ages shown. In the shRNA experiments, the Gapdh [23] was used as an endogenous control gene for normalisation.

Statistical analysis of the results
For in vivo gene transfer, results were expressed as the mean 6 SEM from an appropriate number of experiments as indicated in the figure legends. Nonparametric test with permutations (StatXact Cytel Studio software, Cambridge, MA) was used to assess for statistical differences. For post-test comparisons, we took into account the multiple testing factor, using a non parametric solution. p,0.05 was considered significant (*, p,0.05; **, p,0.01; ***, p,0.001). Each experiment was carried out with n $10, repeated at least two times providing the same results. For qPCR experiments, data were plotted as traditional Tukey whiskers (represent 1.5 times the interquartile distance or to the highest or lowest point, whichever is shorter). Statistical analysis compared the median of DCT values using nonparametric ANOVA, followed by a permutation test (StatXact Cytel Studio software, Cambridge, MA) to compare the control and treated groups.

Results
TRa2 exerts dominant negative activity on positively and negatively regulated T 3

target genes in the mouse hypothalamus in vivo
To test whether TRa2 acts as a dominant-negative receptor on hypothalamic gene transcription in vivo, two reporter gene assays were carried out, using a positively T 3 regulated promoter (ME-tkluc) and a negatively regulated one (Trh-luc). In both cases we compared the effects of TRa2 over-expression to those of TRa1 overexpression, the action of which being already well characterised on both promoters in vivo.
When using the ME-tk-luc construct in the in vivo transfection paradigm, we found that in controls (figure 1A, left columns), T 3 significantly increased ME-tk-luc transcription by two fold (p,0.001). When TRa2 was co-transfected with ME-tk-luc, transcription was blocked at the basal level seen in controls in the presence of T 3 (figure 1A, far right columns). Thus, TRa2 over-expression blocks the stimulatory effect of endogenous receptors on ME-tk-luc transcription. In contrast, co-transfection of TRa1 activated ME-tk-luc transcription about five fold in the presence of T 3 (figure 1A, middle pair of columns, p,0.001), but did not modify transcription levels in absence of T 3 (p = 0.054). Thus, TRa2 does act as a dominant-negative receptor in vivo, blocking the regulation of transcription from a positively regulated TRE.
We next examined the effects of TRa2 on the negatively regulated Trh gene, using the same in vivo gene transfer paradigm. In control mice, expression from the Trh-luc construct (cotransfected with an empty expression vector) was reduced by 37% in animals injected with T 3 as compared with animals receiving saline (first pair of columns in figure 1B). This repression was significant with a p value ,0.001. When the TRa2 isoform was over-expressed, it abolished physiological regulation of Trh. As expected, given that TRa2 cannot bind ligand, T 3 had no effect on transcription levels induced by TRa2. However, in both cases (TRa2 with or without T 3 (figure 1B, far right columns) Trh-luc levels were raised to the maximum levels seen in controls (i.e. in activated, saline injected hypothyroid animals, figure 1B, left columns). Thus, here again TRa2 was acting as a dominantnegative receptor, blocking the effects of the functional endogenous TRs. As an internal control, we used TRa1 which is known to inhibit both T 3 -dependent and T 3 -independent regulation of Trh [21]. As expected, TRa1 blocked Trh transcription at low levels, at 55% of the T 3 -independent control level whether or not T 3 was present.
Transient TRa2 knockdown has no transcriptional effect on positively and negatively regulated T 3

target genes in vivo
To examine TRa2 transcriptional effects on T 3 target genes further, transient knockdown of TRa2 was applied. First, in order to determine if the knockdown in TRa2 expression could trigger a detectable decrease in TRa2 protein level, the TRa2 content of the transfected hypothalami was analysed by Western blotting (Figure 2A). To ensure that equivalent amounts of proteins were blotted in each lane, b-actin levels were determined. We find that the amount of TRa2 protein detected in hypothalami 48 h after shTRa2 injection (mix of two sets of shTRa2, sh1TRa2 and sh2TRa2) was strongly decreased compared to TRa2 levels detected in control group (transfected with shCt). This decrease in TRa2 level demonstrates that the knockdown was efficient 48 h after shRNA injection.
We next investigated the effects of transient TRa2 knockdown on a positively regulated T 3 target gene, Tyrosine hydroxylase enzyme (TyrH) using iGT as described above. We found that in controls transfected with shCt (figure 2B left columns), T 3 significantly increased TyrH-luc transcription (p,0.01). The same transcriptional profile is obtained when shTRa2 is co-transfected, with T 3 significantly increasing TyrH-luc transcription (p,0.01) (figure 2B, right columns). Thus, no effect was seen on T 3independent and dependent TyrH-luc transcriptional levels when shTRa2 is co-transfected as compared to controls ( Figure 2B). Only a significant increased effect on T 3 -independent TyrH-luc transcription was obtained (p,0.001) when TRa2 was overexpressed as compared to Ct ( Figure 2C). We conclude that TRa2 overexpression abrogates T 3 -dependent transcription on both positively regulated T 3 target genes tested, ME as previously mentioned ( Figure 1A) and TyrH ( Figure 2C), whereas TRa2 transient knockdown maintains T 3 -dependent TyrH-luc transcription ( Figure 2B, far right histograms).
The effects of TRa2 knockdown on the negatively regulated Trh gene were examined, using the same iGT paradigm with the mix of shTRa2 constructs. In control mice, expression from the Trh-luc construct (co-transfected with shCt) was repressed significantly by 60.5% (p,0.001) in animals injected with T 3 as compared with animals receiving saline (left columns in figure 2D). When the shTRa2 was co-transfected, it abolished the physiological T 3 regulation of Trh (right columns in figure 2D). Transcriptional activity was equivalent as in the group injected with shCt in absence of T 3 , and was unchanged whereas T 3 was present or not. Thus, loss of TRa2 function seems to allow TRa1 effect on Trh promoter being unmasked, resulting in about the same Trh promoter activity than when TRa1 is over-expressed. We conclude that, both gain or loss of TRa2 function seems to block ME-tk-luc transcription. ME-tk-luc transcription was measured in hypothyroid (PTU) 2 days old mice treated with T 3 (2.5 mg/g b.w.) (PTU+T3) or saline (PTU), 18 h after hypothalamic injection of 1 mg reporter construct and 100 ng expression vector (empty pSG5 (Ct) or pSG5-TRa1 (TRa1) or pSG5-TRa2 (TRa2)). Transcription from ME-tk-luc is significantly increased in the presence of T 3 when TRa1 is overexpressed (as compared to Ct) (p,0.001). In contrast TRa2 overexpression significantly increases basal, T 3 -independent ME transcription as compared to Ct and TRa1 (p,0.001), but addition of T 3 does not modify transcription further. B: TRa2 exerts dominant negative activity on negatively Trh-luc transcription. Trh-luc transcription was measured in hypothyroid (PTU) 2 days old mice as described above (100 ng expression vector and 1 mg reporter gene, Trh-luc per pup). Transcription from a Trh-luc construct is significantly decreased both in absence (PTU) and presence of T 3 (PTU+T3) when TRa1 is overexpressed (as compared with Ct). In contrast, overexpression of TRa2 has no effect on T 3 -independent Trh transcription, but blocks its T 3 -dependent repression. SEMs are given, n$10 per point. In each case, the whole experiment was repeated twice giving similar results. *, p,0.05; **, p,0.01; ***, p,0.001. doi:10.1371/journal.pone.0095064.g001 Trh transcription at an intermediate level between activated and repressed control levels. Indeed, an average TRH activity remains, whereas fine physiological T 3 regulation is lost.
To test this hypothesis, we next investigated the consequences of gain or loss of TRa2 or TRa1 function on endogenous TRH production.

Effects of TRa2 or TRa1 gain or loss of function on endogenous TRH production in euthyroid mice
First, TRa1 or TRa2 or a control vector were transfected into the hypothalamus of newborn euthyroid mice. mRNA were extracted and endogenous Trh levels were followed by qPCR. As seen in figure 3A, the results show that mRNA Trh levels were not significantly modified in either 3 or 5-days old mice when TRa2 or TRa1 was overexpressed, as compared to controls (taken at the same ages). Similarly, no significant effects were seen on endogenous Trh levels when shTRa2 or shTRa1 was transfected as compared to controls ( Figure 3B). Thus, these results arise the question of determining if at later time points (as compared to shorter times; 1day post transfection corresponding to 3 days old mice), we could see a differential effect of TRa2 or TRa1 on thyroid hormone circulating levels. ShTRa2 has no effect on TyrH-luc transcriptional activity either in absence or presence of T 3 . shCt or shTRa2 (400 ng as above) were co-transfected with 1 mg of TyrH-luc construct/hypothalamus of hypothyroid 2-day old mice treated (PTU+T3) or not (PTU) by T 3 (2.5 mg/g b.w.). C: TRa2 overexpression abrogates T 3 -independent repression of the positively regulated TyrH promoter. TRa2 overexpression significantly increases T 3 -independent TyrH-luc transcription as compared to Ct (p,0.001), but addition of T 3 does not increase transcription further. Empty pSG5 vector (Ct) or pSG5-TRa2 (TRa2) was used at 100 ng and co-transfected with 1 mg of TyrH-luc construct/hypothalamus of hypothyroid 2-day old mice. D: TRa2 transient knockdown abolishes T 3 -dependent repression of the negatively regulated Trh promoter. ShTRa2 has no effect on T 3 -independent Trh promoter activity (p = 0.07) and when T 3 is added, Trh-luc transcription is not repressed anymore (p,0.05) as compared to shCT. The same experimental conditions as in B were used (400 ng expression vector and 1 mg reporter gene, Trh-luc per pup). SEMs are given, n$10 per point. In each case, the whole experiment was repeated twice giving similar results. *, p,0.05; **, p,0.01; ***, p,0.001. doi:10.1371/journal.pone.0095064.g002 The effects of TRa2 or TRa1 overexpression on Trh-luc transcription are correlated with modifications of thyroidal status Given the differential effects of TRa2 versus TRa1 on the Trh promoter activity obtained by iGT, we next examined the effects of their overexpression on circulating T 4 levels. As seen in figure 4, TRa1 overexpression resulted in a significantly decreased circulating T 4 level at P7 as compared to controls (p,0.01) (Figure 4, far right columns). However no effect in circulating T 4 was observed at the same age when TRa2 was overexpressed. The results of TRa2 or TRa1 overexpression on Trh-luc transcription are correlated with modifications of thyroidal status.

Discussion
It is intriguing to note that of all the four main products (TRa1, TRa2, TRb1 and TRb2) of the two TR genes (NR1A1 and NR1A2), the mRNA of the non-hormone-binding variant TRa2 is by far the most highly expressed in the brain [6,26]. Indeed in the rat brain, temporal expression of TRa2 mRNA follows the same spatial pattern of expression of TRa1, but its levels are markedly higher [26], suggesting that TRa2 might be a critical non T 3 dependent regulator of thyroid hormone action by modulating T 3binding TR effects on the expression of brain-specific genes [7]. One line of investigation to address the role of TRa2 in general thyroid hormone dependent signalling has been to generate mice lacking TRa2. These mice show an overexpression of TRa1 in all tissues examined, including brain. The mice have significantly lower circulating free T 3 and free T 4 and their thyroid glands show features of dysfunction, suggesting decreased activity of the Hypothalamic Pituitary Thyroid axis (H-P-T) [20]. This phenotype (insufficient stimulation of the thyroid and of the production of TH) raises the question of the physiological function of TRa2 in brain and notably in the hypothalamus at the level of Trh transcription.
We chose to examine the effects of TRa2 on Trh transcription using an in vivo reporter gene approach. Three reasons, besides the  phenotype of the TRa2 -/mouse, made Trh promoter of a particular interest in terms of function of this enigmatic TRa2 isoform. First, Trh gene regulation allows one to investigate TR isoforms specificity as TRb and TRa have distinct roles in the negative transcriptional regulation by T 3 [1,27]. Second, Trh is a critical component of the H-P-T axis and is thus a critical regulatory gene. Third, Trh is a T 3 negatively regulated target gene and is of particular mechanistic interest from the transcriptional point of view. It is important to discuss here the fact that it is often considered that the HPT axis is immature in the postnatal mouse. This concept is largely based on the observations of the low levels of circulating T 3 and T 4 levels that increase steadily during the first two weeks of postnatal life peaking at p15 and then declining slightly to reach adult levels [28]. However, the feedback system is active as decreasing T 3 and T 4 levels by administrating PTU increase Trh expression. Thus even if the axis is not fully mature, the components of negative feedback are present (TRs, NCoR, SMRT, etc… [29]). In fact, just because circulating levels of T 3 and T 4 climb during this post-natal phase does not actually imply that the axis is not functional until adult levels are attained. First, the low levels of circulating hormone indicate more that is could be due to low feed forward drive at any of the levels, TRH, TSH or even T 3 /T 4 production. Second, these low levels do not rule out the possibility that the feedback system can respond to high levels of T 3 . Thus more knowledge is required on the manner at which hypothalamic setpoints are established, and modulated, during this critical post-natal period.
We started our study by iGT experiments conducted on hypothyroid newborn mice, to reduce high variability in endogenous thyroid hormone levels, which could compromise transcriptional regulation study. The results on Trh transcription were compared to those obtained on Malic Enzyme (ME). In both cases we compared the effects of TRa2 to those obtained with TRa1, because TRa1 action on both genes of interest has already been well characterised in vivo. We observed that in absence of T 3 , TRa1 fails to repress ME expression, suggesting that level of endogenous TRa1 was already sufficient to repress basal ME expression. In contrast, TRa2 was able to increase basal ME transcription, suggesting that when TRa2 is overexpressed it acts as a dominant-negative receptor, competing with endogenous TRa1 as to lead to an increase in ME transcription.
Regarding the negatively T 3 -regulated Trh gene, TRa1 prevents the T 3 -indepedent Trh activation, and increases the T 3 -dependent repression observed in the control group. Thus, TRa2 acts as a dominant-negative receptor on both positively and negatively regulated T 3 target genes. Our in vivo result on ME-tk-luc transcription was in accordance with data conducted on transfected cells where TRa2 exerted a negative effect on T 3 -positive response element-mediated transcription [30]. The molecular mechanisms underlying the dominant negative activity of TRa2 are not yet completely elucidated, even in vitro. Two different mechanisms have been proposed: the first, described by Katz et al. [15] involves a passive repression, in which TRa2 blocks TRs action by competing for binding to TREs; the second mechanism has been proposed by Liu et al. [17], who demonstrated that TRa2 inhibitory effect does not require binding to TRE and suggested that interactions with components of the general transcription machinery might instead play a crucial role.
For the T 3 negatively regulated gene Trh, the gain of function of TRa2, when compared to the gain of function of TRa1, results in strong activation of transcription that is unmodified by presence or absence of T 3 . In contrast, TRa1 overexpression down-regulates Trh transcription and this regulation is equally T 3 insensitive. Each of these regulations, contrast with the physiological T 3 -dependent repression of Trh in the presence of TRb isoforms [2,21]. This T 3independent activation of Trh transcription by TRa2 suggests a possible role of TRa2 in vivo, acting as a dominant-negative receptor on negative T 3 -regulated genes. In vitro experiments have not been able to reveal such a role. When transfected into JEG-3 cells, TRa2 isoform was inactive on positively and negatively regulated T 3 response genes whereas TRa1 and TRb stimulated transcription from TRE-tk-CAT (pTRE), and repressed TSHa-CAT (nTRE) reporter genes in T 3 -dependent manners. When coexpressed with TRa1 or TRb at relatively high doses, TRa2 inhibited regulation of positive TREs but did not affect negative regulation [14]. The difference between these findings and our data are probably due to different cellular contexts and different target genes studied.
To explore further the function of TRa2 in Trh regulation, we used transient knockdown of TRa2 using an shRNA approach. Effects were also followed on positively T 3 -regulated target genes (Tyrosine hydroxylase (TyrH) and ME enzymes). TRa2 overexpression abrogated T 3 -dependent transcription on both of these positively T 3 regulated target genes, whereas TRa2 transient knockdown maintains T 3 -dependent TyrH-luc transcription. This result confirms the dominant negative action of TRa2 on positively regulated T 3 target genes since its knockdown unmasks functional TRs transcriptional effects. Intriguingly, transient knockdown of TRa2 has the same effect as its overexpression on Trh gene transcription (Trh-luc transcriptional levels are similar to those obtained in controls in absence or presence of T 3 ). Indeed, when TRa2 is overexpressed it leads to an imbalance in the transcriptional machinery, thus impairing the well-defined effect of TRb on T 3 -dependent Trh transcriptional repression [23,31]. Conversely, when shTRa2 is transfected, a shift in the balance of the transcriptional machinery towards TRa1 results in equivalent Trh-luc transcriptional levels as when TRa1 is overexpressed.
In order to study the effects of TRa1 and TRa2 on endogenous Trh mRNA levels, qPCR analyses were conducted on euthyroid mice so as to examine the dynamics of feedback in physiologically normal animals. We obtained no differential effects of the two overexpressed isoforms in the shorter frame at one day posttransfection (3 days old mice) nor at three days post-transfection. Similarly, no detectable effect was seen on endogenous Trh mRNA levels in the shorter frame (at 36 h post-transfection) when shTRa2 or shTRa1 was transfected. These data showing no detectable variations in Trh mRNA levels in the shorter term fit with those published by a number of authors [32,33] who showed Trh mRNA varied within longer time frames.
To propose a model of TRa1 and TRa2 interaction, it is easiest to start from the TRa2 loss of function studies. The mutant mice show loss of TRa2 to increase TRa1 expression. In effect, we observe that hypothalamic TRa2 loss of function has the same effect on Trh-luc transcriptional activity as TRa1 overexpression. We therefore suggest that in physiological conditions there is a balance between the effects of TRa2 and TRa1 allowing the overriding effects of the TRb isoforms that provide physiological T 3 -dependent Trh regulation.
Indeed, the results from the mutant mice studies [20] suggest that the loss of TRa2 or the changed balance of TRa2/TRa1 perturbs a range of functions (metabolism and growth) notably at the central level, suggesting a role for TRa2 in regulating central T 3 -dependent transcription genes. Similarly, a changed balance of TRa2/TRa1 in a context where TRa2 is overexpressed would have consequent transcriptional effects on brain gene expression. A biological activity that can be attributed to TRa2 would be an adjustment of the T 3 binding TRa1 protein activity to physiologically appropriate levels, implying thus an important, widespread regulatory role in mammalian physiology of the ratio of TRa1/ TRa2 expression. Our data showing that TRa2 does indeed act as a dominant-negative receptor on both negatively and positively T 3 regulated target genes in the brain strongly bolster this hypothesis.
In this report we define an in vivo function for the nonbinding TRa2 isoform in regulating brain genes, supported by findings on athyroid Pax8 -/-TRa1 -/mice who die around weaning unless they are substituted with thyroid hormones due to the negative effects of the TRa1 aporeceptor, but, rather, including a more complex mechanism involving TRa2 and unliganded TR isoform TRDa2 [34].
Taken together, our results emphasize the as yet neglected physiological importance of TRa2, naturally acting as dominant-negative receptor on hypothalamic Trh transcription, and consequently, on the regulation of HPT axis.