Role of Thyroid Hormone Receptor Subtypes � and �

on Gene Expression in the Cerebral Cortex andStriatum of Postnatal Mice

Pilar Gil-Ibañez, Beatriz Morte, and Juan Bernal

Instituto de Investigaciones Biomédicas Alberto Sols, Consejo Superior de Investigaciones Científicas, andUniversidad Autónoma de Madrid, and Center for Biomedical Research on Rare Diseases, 28029 Madrid,Spain

The effects of thyroid hormones (THs) on brain development and function are largely mediated bythe control of gene expression. This is achieved by the binding of the genomically active T3 totranscriptionally active nuclear TH receptors (TRs). T3 and the TRs can either induce or represstranscription. In hypothyroidism, the reduction of T3 lowers the expression of a set of genes, thepositively regulated genes, and increases the expression of negatively regulated genes. Two mech-anisms may account for the effect of hypothyroidism on genes regulated directly by T3: first, theloss of T3 signaling and TR transactivation, and second, an intrinsic activity of the unliganded TRsdirectly responsible for repression of positive genes and enhancement of negative genes. To an-alyze the contribution of the TR subtypes � and �, we have measured by RT-PCR the expression ofa set of positive and negative genes in the cerebral cortex and the striatum of TR-knockout maleand female mice. The results indicate that TR�1 exerts a predominant but not exclusive role in theregulation of positive and negative genes. However, a fraction of the genes analyzed are not oronly mildly affected by the total absence of TRs. Furthermore, hypothyroidism has a mild effect onthese genes in the absence of TR�1, in agreement with a role of unliganded TR�1 in the effects ofhypothyroidism. (Endocrinology 154: 0000–0000, 2013)

The effects of thyroid hormones (THs) on brain devel-opment and function are mediated primarily by

genomic actions involving the control of the expression ofmany genes (1–3). This is achieved by binding to nuclearTH receptors (TRs), which are ligand-regulated transcrip-tion factors (4). In mammals TRs are encoded by twogenes, THRA and THRB. The THRA gene produces sev-eral proteins of which only TR�1 binds the active hor-mone T3 and therefore functions as a hormone receptor.The THRB gene produces two T3-binding isoforms, TR�1and TR�2. The TRs regulate transcription by binding tospecific sequences present in regulatory regions of the tar-get genes (5, 6). The TRs can also regulate transcription bydirect interactions with other transcription factors (7, 8).TH can exert positive or negative regulation (1). The pos-itively regulated genes are repressed in the absence of T3

and stimulated in its presence. The negatively regulated

genes are stimulated in the absence of T3 and inhibited inits presence. In the present paper, we use the term positivegenes to mean genes with decreased expression in hypo-thyroid animals and negative genes those with increasedexpression in hypothyroid animals, independently of themechanism of regulation, which is known for some of thegenes that we have studied (9–11).

Despite the role of TRs in mediating the physiologicaleffects of TH, it is known that the absence of TRs is notequivalent to TH deprivation, ie, hypothyroidism (12).For example, the morphological changes induced by hy-pothyroidism in the cerebellum of wild-type (WT) mice,are not observed in mice devoid of TR�1 (13), indicatingthat the effects of hypothyroidism are partly due to unli-ganded receptor activity. In agreement with this, a hypo-thyroid phenotype is produced by expression of mutant

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in U.S.A.Copyright © 2013 by The Endocrine SocietyReceived December 6, 2012. Accepted February 28, 2013.

Abbreviations: P21, postnatal day 21; TH, thyroid hormone; TR, TH receptor; WT, wild-type.

T H Y R O I D - T R H - T S H

doi: 10.1210/en.2012-2189 Endocrinology endo.endojournals.org 1

Endocrinology. First published ahead of print March 14, 2013 as doi:10.1210/en.2012-2189

Copyright (C) 2013 by The Endocrine Society

TR�1 with dominant-negative activity due to intact DNAbinding but deficient in hormone binding (14).

Differences in physiological roles between the TR sub-types � and �, or the TR� isoforms, TR�1 and TR�2,depend mainly on their patterns of expression in differenttissues. Hepatic gene expression profiles in vivo were verysimilar in TR� or TR� knockout mice (15), indicating thatin liver, TR� and TR� regulate a similar set of genes. Thisagrees with recent findings demonstrating that when in-dividually expressed in the same cell types, the TR sub-types regulate largely overlapping sets of genes, with dif-ferences mostly in the kinetics of regulation (16, 17).However, the TRs also display some selectivity on generegulation in some experimental settings (18 –20). Forexample, Klf9 is induced by T3 in N2a neuroblastomacells expressing TR�1 but not in N2a cells expressingTR�1 (21).

Based on the concentrations of TR protein in brain,predominantly in the form of TR�1 (22), it is thought thatthe effects of TH on the brain are mostly mediated throughTR�1, but a detailed study on the relative contributions ofTR� and TR� on brain gene expression has not beenmade. In the present work, we have studied the role of theTRs in the mouse cerebral cortex and striatum. To thisend, we analyzed the expression of a set of positively andnegatively regulated genes (2, 3) in these regions from WTmice and from mice devoid of TR�1, TR�, or both. Theresults indicate a primary, but not exclusive, role of TR�1in positive and in negative regulation. We also studiedwhether the combined absence of TR�1 and TR� hadsimilar effects as the absence of the hormone. The resultsindicated that hypothyroidism induced a more profoundeffect on brain gene expression than complete absence ofTRs, in agreement with a role of unliganded TR in hypo-thyroidism. The role of unliganded TR�1 was demon-strated on some genes by showing that hypothyroidismhad no effect, or the effect was milder, on mice devoid ofTR�1 than on WT mice.

Materials and Methods

Handling of animalsProtocols for animal handling were approved by the local

institutional Animal Care Committee, following the rules of theEuropean Union. Animals were housed in temperature-con-trolled (22°C � 2°C) and light-controlled (12-hour light, 12-hour dark cycle; lights on at 7:00 AM) conditions and had freeaccess to food and water. Mice of a hybrid genetic backgroundof 129/OLa�129/Sv� BALB/c�C57BL/6 (23–26) were used. Westarted by crossing TR�1�/�TR��/� male mice with female ofthe same genotype (F0) to generate all possible genotypic com-binations (F1). The mice used in the experiments were obtained

by appropriate crossings of F1 to generate WT, TR�1�/�TR��/�,TR�1�/�TR��/�, and TR�1�/�TR��/� mice (F2) and were in-distinctly male or female of age postnatal day 21 (P21). TheTR�1 deletion was detected by using the following combinationof primers: forward 5�-caagatcgagaagagtcagga-3�, reverse 5�-gtatgggagctgcatctatccaag-3� and the TR�1�/�-specific reverseprimer 5�-cactgcattctagttgtggt-3�. The TR� deletion was de-tected by using the following combination of primers: forward5�-gcacaggcaggaagtaggctgttct-3�, reverse 5�-ccctggaggccaaaggt-catcaatg-3� and the TR��/�-specific reverse primer 5�-gtgc-cagcggggctgctaaag-3�. Hypothyroidism was induced in pregnantand lactating dams by administering a drinking solution con-taining 0.02% 1-methyl-2-mercapto-imidazole (MMI; SigmaChemical Co, St Louis, Missouri) plus 1% KClO4 ad libitum.These antithyroid compounds were given from gestational day 9and throughout the lactating period until the end of the exper-iment on P21. These compounds cross the placenta and are pres-ent in the milk, so that the pups derived from the treated damswere also hypothyroid.

RNA preparation and quantification by real-timeRT-PCR

The pups were killed by decapitation on P21. The cerebralcortex and the striatum were rapidly dissected out, frozen on dryice, and kept at �80°C until RNA isolation. RNA was isolatedusing the Trizol procedure (Invitrogen, Carlsbad, California)with an additional step of chloroform extraction for RNA iso-lation. The quality of RNA was analyzed using a BioAnalyzer(Agilent, Santa Clara, California). cDNA was prepared from 250ng of RNA using the High Capacity cDNA Reverse Transcrip-tion kit (Applied Biosystems, Foster City, California). Quanti-tative PCR assays were performed on microfluidic cards or sin-gle-tube PCR. For the microfluidic cards, we used TaqMan low-density arrays (Applied Biosystems), format 48a (P/N 4342253).cDNA aliquots corresponding to 10 ng of starting RNA fromindividual mice were used, with TaqMan Universal PCR MasterMix, No Amp Erase UNG (Applied Biosystems) on a 7900HTFast Real-Time PCR System (Applied Biosystems). The PCR pro-gram consisted in a hot start of 95°C for 10 minutes, followed by40 cycles of 15 seconds at 95°C and 1 minute at 60°C. Foranalysis, we used the 2-Ct method. As internal control we in-cluded 18S RNA and Ppia (peptidylprolyl cis/trans isomerase Aor cyclophilin A). The use of either reference RNA gave similarresults, so that the data were all normalized to the 18S RNA.Data are expressed relative to the values obtained for the controlWT, which was given a value of 1.0 after correction for 18SRNA. For single-tube PCR, a cDNA aliquot corresponding to 5ng of the starting RNA was used, with TaqMan Assay-on-De-mand primers and the TaqMan Universal PCR Master Mix, NoAmp Erase UNG (Applied Biosystems) on a 7900HT Fast Real-Time PCR System (Applied Biosystems). The PCR program con-sisted in a hot start of 95°C for 10 minutes, followed by 40 cyclesof 15 seconds at 95°C and 1 minute at 60°C. PCRs were per-formed in triplicates, using the 18S gene as internal standard andthe 2-Ct method for analysis.

Statistical analysisDifferences between means were obtained by 1- or 2-way

ANOVA, depending on the experiment, and the Tukey or Bon-ferroni’s post hoc tests, respectively. Calculations were done us-

2 Gil-Ibañez et al TRs and Brain Gene Expression Endocrinology

ing the GraphPad Prism software (http://www.graphpad.com/prism/). The experimental groups were formed with about thesame number of male and female pups, and sex of the animalswas not considered a factor in statistical analysis.

Results

The goal of this work was to analyze the relative roles ofTR�1 and TR� on the expression of TH-dependent genesin the cerebral cortex and the striatum of P21 mice. Weanalyzed 27 genes regulated positively and 14 genes reg-ulated negatively by TH (Supplemental Data 1, publishedon The Endocrine Society’s Journals Online web site athttp://endo.endojournals.org). These genes were identi-fied in previous microarray analysis of the cerebral cortex(3) and the striatum (manuscript in preparation). The cri-terion to define these genes as positively or negatively reg-ulated by TH was that hypothyroidism induced a de-creased or increased expression, respectively, in at least 1brain region. The expression of each gene was measuredby quantitative PCR in RNA samples of individual micefrom four genotypes: WT, TR�1�/�, TR��/�, andTR�1�/�TR��/�. As a reference for the TH dependenceof each gene, we also included a group of WT mice ren-dered hypothyroid from prenatal stages. The effects ofhypothyroidism on the cortex expression of most geneswere as previously described (2), with the exception ofCol6a2. This gene was included in the study as a negativegene based on previous findings. However, in the presentstudy, the higher mean expression value in hypothyroid-ism was not statistically significant (Supplemental Data 1).

Figure 1 shows the expression patterns in cortex andstriatum of 4 selected genes, Kcnj10, Cbr2, Cirbp, and

Angptl4. Individual results for all 41genes are shown in SupplementalFigure 1. The 4 genes selected in Fig-ure 1 are examples illustrating theobserved patterns of regulation (seealso Figure 7). When hypothyroid-ism was performed in the WT mice,expression of Kcnj10 and Cbr2 de-creased, whereas expression ofCirbp and Angptl4 increased, classi-fying these genes among the positiveand negative genes, respectively. Theeffect of hypothyroidism may becompared with the effect of the com-bined absence of TR�1 and TR� (ab-breviated as ��). In the cortex andthe striatum, the absence of �� in-creases the expression of the nega-tively regulated gene Cirbp to the

same level as in hypothyroidism. For Kcnj10, the effect of�� deficiency was in the same direction but not as strongas the effect of hypothyroidism. In contrast to the strongeffects of hypothyroidism, �� deficiency was without ef-fect on Cbr2 in the cortex and the striatum and on Angptl4in the striatum. The effect of �� deficiency on cortex Ang-ptl4 was less clear. Single inactivation of TR�1 or TR� hadvariable effects, with a decreased expression of Kcnj10andCirbp in the striatumand increasedexpressionof Ang-ptl4 in the cortex of TR�1-deficient mice.

Figures 2 through 5 show the relative changes of geneexpression in reference to the expression in the WT, givena value of 1.0 and represented as a dotted line in the fig-ures. The genes were ordered in the figures in relation tothe value obtained for the hypothyroid WT mice. For thepositive genes (Figures 2 and 3), the strongest effect ofhypothyroidism was on Agxt2l1 in the cortex and Cd72 inthe striatum with more than 90% reduction. There was noquantitative correlation between the effects of hypothy-roidism in the cortex and striatum (r � 0.13, P � .52, notshown). As extreme examples, Cd72 and Vdr were amongthe most strongly affected genes in the striatum, whereasin the cortex, they were little or not affected. Neither thelack of TR�1 nor the lack of TR� induced consistentchanges, although the mean expression was below the WTvalue for the TR�1�/� and above the WT for the TR��/�.The absence of �� had in general the strongest effect andin the same direction as hypothyroidism. The effect of ��

deficiency on some genes approached that of hypothy-roidism. This was the case for Hr, Pvalb, Kcnj10, Nefm,Nefh, or Sema7a in the cortex and for Cd72, Vdr, Pvalb,Aldh1a1, Nefm, Hr, or Nefh in the striatum. Other genesremained at or near WT levels (Flywch2, Ier5, Itih3, Nrtn,

Figure 1. Expression of selected genes in the P21 mouse cerebral cortex and striatum: effects ofhypothyroidism or TR deficiency. For WT mice, n � 7; WT hypothyroid mice (H), n � 6; TR�1�/�

mice (�), n � 6; TR�1�/� mice (�), n � 6; and TR�1�/� TR�1�/� mice (��), n � 5. Statisticalcomparisons were made by 1-way ANOVA followed by the Tukey test, between WT, H, and ��and between WT, �, and �. Abbreviations: Ctx, cortex; ns, not significant; Str, striatum. *P �.05; **P � .01; ***P � .001.

doi: 10.1210/en.2012-2189 endo.endojournals.org 3

and Paqr6 in the cortex and Agxt2l1, Cbr2, Flywch2, Ier5,Itih3, Klf9, and Pdp1 in the striatum) (Figures 2 and 3 andSupplemental Figure 1). The results partially agree with apredominant role of TR�1 in TH-mediated brain geneexpression. The absence of TR� was associated with nor-mal or increased gene expression, probably due to the

increased TH levels in these mice (24). In the absence ofTR�1, TR� was able to sustain gene expression to near WTeuthyroid levels.

For the negative genes (Figure 4 for the cortex and Fig-ure 5 for the striatum), the effects of hypothyroidismshowed a positive correlation between the cortex and the

Figure 3. Expression of positive genes in the striatum of hypothyroid(Hypo), TR�1�/�, TR��/�, and TR�1�/�TR��/� mice using microfluidiccards (TaqMan arrays). The figure shows the comparisons with the WTmice, which were given a value of 1.0, represented by the dotted line.For individual data and statistical comparisons, see SupplementalFigure 1. The number of animals was as indicated in the legend forFigure 1.

Figure 4. Expression of negative genes in the cerebral cortex ofhypothyroid (Hypo), TR�1�/�, TR��/�, and TR�1�/�TR��/� mice usingmicrofluidic cards (TaqMan arrays). The figure shows the comparisonswith the WT mice, which were given a value of 1.0, represented by thedotted line. For individual data and statistical comparisons, seeSupplemental Figure 1. The number of animals was as indicated in thelegend for Figure 1.

Figure 2. Expression of positive genes in the cerebral cortex ofhypothyroid (Hypo), TR�1�/�, TR��/�, and TR�1�/�TR��/� mice usingmicrofluidic cards (TaqMan arrays). The figure shows the comparisonswith the WT mice, which were given a value of 1.0, represented by thedotted line. For individual data and statistical comparisons, seeSupplemental Figure 1. The number of animals was as indicated in thelegend for Figure 1.

Figure 5. Expression of negative genes in the striatum of hypothyroid(Hypo), TR�1�/�, TR��/�, and TR�1�/�TR��/� mice using microfluidiccards (TaqMan arrays). The figure shows the comparisons with the WTmice which were given a value of 1.0, represented by the dotted line.For individual data and statistical comparisons, see SupplementalFigure 1. The number of animals was as indicated in the legend forFigure 1.

4 Gil-Ibañez et al TRs and Brain Gene Expression Endocrinology

striatum (r � 0.76, P � .0014, not shown). Comparedwith the positive genes, a strongest effect of the absence ofTR�1 was observed as a whole. Hypothyroidism in-creased the expression of 13 genes in the cortex from 1.5-to4-foldand14genes in the striatumfrom1.3- to5.5-fold.The absence of TR�1 increased the expression of severalgenes in the cortex and in the striatum at least 1.5-fold. Incontrast, the absence of TR� induced minimal changes.The absence of �� was similar to the absence of TR�1,with 9 genes increasing at least 2-fold in the cortex and 6genes increasing at least 2-fold in the striatum. We mayconclude that also for the negative genes, TR�1 was morerelevant for gene expression than TR�. As for the positivegenes, the effects of hypothyroidism were in general stron-ger than the absence of TRs.

The effects of hypothyroidism on gene expression couldbe due to 2 factors. First, the reduction of T3 signalingdirectly related to the reduction of TR occupancy andtransactivation. In this case, the absence of receptorsshould be similar to the effects of TH deprivation. Second,the unliganded TRs might have intrinsic activity and di-rectly inhibit or stimulate the expression of positive ornegative genes, respectively. The correlations between theeffects of the lack of TRs and hypothyroidism for the pos-

itive and negative genes are shown in Figure 6. In all cases,the correlations were significant, but the slopes of the re-gression lines were lower than 1, indicating that the effectof hypothyroidism was stronger than the effect of TR de-ficiency. From the values of the y-intercepts it may be cal-culated that the effect of TR deprivation, ie, the loss of T3

signaling accounted for about 70%–80% of the effect ofhypothyroidism on the positive genes and 60% for thenegative genes on average.

To confirm that the effects of hypothyroidism on somegenes were due at least in part by the activity of unligandedreceptors, especially TR�1, we analyzed the effect of hy-pothyroidism on gene expression in TR�1-deficient mice.Figure 7 shows the response of the genes described in Fig-ure 1 (Kcnj10, Cbr2, Cirbp, and Angptl4). In WT hypo-thyroid mice, the expression of Kcnj10 and Cbr2 de-creased and the expression of Cirbp and Angptl4increased. On the other hand, whereas hypothyroidismhad a similar effect on the expression of Kcnj10 and Cirbpin the presence or absence of TR�1, it was without effecton Cbr2 and Angptl4 in the absence of TR�1, indicatingthat the effects of hypothyroidism on these 2 genes was dueto the repressing (Cbr2) or inducing (Angptl4) activity ofunliganded TR�1. The results obtained using 12 positive

Figure 6. Correlations of gene expression between hypothyroid WT mice and TR�1�/�TR��/� mice. The data correspond to expression ofpositive and negative genes in the cerebral cortex and the striatum. Shown are the best fit for linear regression and the 95% confidence limits. Thedotted line represents the equation y � x.

doi: 10.1210/en.2012-2189 endo.endojournals.org 5

genes and 7 negative genes in the cerebral cortex and/or thestriatum are shown in Supplemental Figure 2. Comparedwith the effect of hypothyroidism on the WT mice, in theabsence of TR�1, hypothyroidism had a much lower effecton Agxt2l1, Gls2, Cbr2, Flywch2, Itih3, Angptl4, Gpc3,Mmdc2, and Ly75 in at least 1 region. In addition, the testfor interaction in the 2-way ANOVA comparing the re-sponse of WT and TR�1-deficient mice to hypothyroidismwas highly significant (P � .001) for most genes, indicat-ing an influence of genotype in the response.

Discussion

In the present work, we have analyzed the role of TR� andTR� in the control of brain gene expression. To this end,we have measured the expression of genes that we havepreviously identified as TH-dependent after microarrayanalysis of the cerebral cortex and striatum from normaland hypothyroid P21 mice. The genes analyzed cover awide range of physiological and biochemical processes,reflecting the complexity of TH action in the brain and thepleiotropic effects of hypothyroidism (Supplemental Data1). Although the goal of this paper was not to discuss thebrain processes regulated by THs, it is pertinent to brieflysummarize the main functions of the genes analyzed. Theyinclude genes involved in different aspects of metabolism,such as Hmgcs2 (ketogenesis), Aldh1a1 (retinol metabo-lism), Cbr2 (nicotinamide adenine dinucleotide phosphate-dependent oxidoreductase), Gls2 (glutamine metabo-lism), Pla2g5 (membrane phospholipids), and Angptl4

(glucose metabolism); genes in-volved in cell proliferation (Ly75),differentiation (Cd72), and survival(Cirbp); neurofilaments, such asNefh and Nefm, or related to tubulinprocessing (Agbl3), axon growth(Sema7a) and neuron survival(Nrtn), extracellular matrix and ad-hesion proteins (Itih3, Col6a1,Mamdc2, and Cntn2), development(Shh), transcription (Dbc1, Hr,Klf9), Ca2�/calmodulin signaling inthe dendritic spines (Nrgn), K�

channel (Kcnj10), and Ca2�-bindingproteins (Pvalb and Calb). We havepreviously studied the expression ofmany of these genes in other con-texts, such as inactivation of themonocarboxylate transporter 8, andtypes 2 and 3 deiodinases (2, 3).

By comparing their expression inTR�1�/� and in TR� �/� mice rela-

tive to the WT, we conclude that both receptor subtypesare involved in the regulation of gene expression in brain.TR�1 appears to have a primary role, but the lack of thisreceptor affects only a subset of the genes. Absence of bothreceptor types increases the number of genes affected, andin many cases, the effect approaches quantitatively theeffect attained by hypothyroidism, indicating that in theabsence of TR�1, TR� maintains gene expression nearnormal levels. On the other hand, the absence of TR�

results in little change, with increased expression of a fewpositive genes and decreased expression of a few negativegenes.

In these effects of receptor deficiency, we have to takeinto account possible changes of TH concentrations thatmight have contributed to the observed changes. The ef-fects of TR�1 deficiency are not probably due to lower T3

concentration, because cerebral cortex concentrations ofT4 and T3 are not modified in the TR�1�/� mice (13).However, the increased or decreased expression of somegenes observed in the absence of TR� is most likely due tothe known enhancement of TH production in the TR��/�

mice (24). This would result in increased T3 actionthrough the remaining TR�1, and increased or decreasedexpression of positive and negative genes, respectively.More difficult is to analyze the contribution of the highlyincreased TH concentrations in the double �� deficiency(25). We cannot discard that overactivation of non-genomic pathways (1) might play a role in these mice.However, we think that this possibility is unlikely giventhat �� deficiency either has no effect on gene expression

Figure 7. Effects of hypothyroidism in the presence and absence of TR�1. This figure shows thedata from 4 different genes, 2 of them positively regulated (Kcnj10 and Cbr2) and 2 of themnegatively regulated (Cirbp and Angptl4) in the cerebral cortex and the striatum of WT andTR�1�/� mice in the basal state and after induction of hypothyroidism. Statistical comparisonswere by 2-way ANOVA, with 8 mice for each experimental condition. Significance symbols abovethe bars represent the comparison between hypothyroidism (filled bars) with the respectiveuntreated WT or TR�1�/� mice (open bars). Abbreviations: Ctx, cortex; ns, not significant; Str,striatum. *P � .05; **P � .01; ***P � .001.

6 Gil-Ibañez et al TRs and Brain Gene Expression Endocrinology

or the effect goes in the same direction as inhypothyroidism.

Another important concern is how the cellular hetero-geneity of the brain regions might have influenced the geneexpression changes. Indeed, genetic studies have revealeda cellular complexity that goes well beyond the classicalcell type subdivisions of the brain based on morphologyand neurotransmitter production (27, 28). Different cellgroups might respond differently to THs in the regulationof expression of individual genes. Also, the responsive cellsmight be a minor component of the total cellular repertoireof the region under study. It is well known that some in-dividual genes may be sensitive to THs in some cell pop-ulations and not in others despite expressing TRs in ade-quate amounts (29). As an example, Nrgn, a generegulated directly by T3 at the transcriptional level (30), isvery sensitive to TH in the striatum, dentate gyrus, andlayer 6 of cerebral cortex, whereas other layers of the cor-tex and hippocampus are not sensitive (31). This is thereason for the lower effect of hypothyroidism on Nrgnexpression and on other genes such as Vdr and Cd72 in thewhole cortex compared with the striatum. Only quanti-tative in situ hybridization techniques, with a detailed ac-count of the gene expression responses by individual cellgroups, should be able to provide a complete picture.

A full interpretation of the data would also require aprevious understanding of the mechanisms of regulationby T3 on each individual gene, specifically whether theyare direct or indirect responses. Several lines of evidenceindicate that many of the genes studied in this work aremost probably direct targets of TH. Some of them havebeen specifically studied in this regard, with confirmationof direct regulation at the cellular level and identificationof thyroid-responsive elements. These include Nrgn (9,30), Hr (10), Klf9 (11, 21), Gbp3 (32), and Angptl4 (33).Others respond to the administration of a single T3 dose toadult rats (Aldh1a1, Bcar3, Hr, Itih3, and Klf9) (34) or toprimary mouse cerebrocortical neurons (Aldh1a1, Bcar3,Calb, Flywch2, Gbp3, Gpc3, Hr, Kcnj10, Klf9, Nefh,Nefm, Paqr6, Sema7a, and Shh) (35). Studies are in prog-ress to analyze gene responses in primary cells derivedfrom TR-deficient mice.

With the above limitations in mind, we found little ev-idence for receptor subtype specificity among the positivegenes. Some of the genes analyzed in this work have alsobeen examined in other cellular contexts. In HepG2 cells,Kcnj10 was positively regulated by T3 through TR�1 andGpc3 was negatively regulated, in general agreement withour findings (16). In contrast, Angptl4, which behaves asa negative gene in cortex and striatum, in HepG2 cells isa TR�-specific positive gene (33). Concerning the negativegenes, the absence of TR�1 was in general more effective

than for the positive genes, indicating that TR�1 was moreinvolved in negative regulation of brain genes than TR�.These results contrast with the predominant effect of TR�

on negative regulation of the Tshb gene (20).Compared with the effects of hypothyroidism, the ab-

sence of TR�1 and TR�, and therefore complete lack of T3

signaling through the TR, led either to no changes in geneexpression or to changes that were in the same direction asin hypothyroidism but generally of less severity. Takingtogether all genes, the effects of hypothyroidism and ofTR�1 plus TR� inactivation were correlated, but the re-gression line indicated a stronger effect of hypothyroid-ism. The results are compatible with the idea that the ef-fects of hypothyroidism on the expression of some genesis due to the activity of the unliganded TR, consisting ofrepression of positive genes and activation of negativegenes. It is likely that the effects of unliganded TR in brainhypothyroidism are primarily due to TR�1. This was dem-onstrated by showing that on some genes, hypothyroidismhad no effect on the TR�1�/� mice. A clear example wasCbr2 in cortex and striatum. On others, there was a sig-nificant effect of hypothyroidism on the TR�1�/� micebut milder than the effect on the WT, for example, Kcnj10also in both regions. This may indicate that unligandedTR� might also play a role in hypothyroidism in agree-ment with the effects of mutant TR�1 (36, 37).

A final conclusion of these studies is that TR�1 exertsa predominant but not exclusive role in the regulation ofgene expression in the cerebral cortex and the striatum.This may be a direct consequence of the higher abundanceof TR�1 relative to TR� in brain or may reflect differencesin the action kinetics between the two receptor subtypes,as pointed out recently for nonneural cell lines in culture(16, 17).

Acknowledgments

We acknowledge the technical assistance of Eulalia Moreno,Inmaculada G. Martín, Irene Cuevas, and Daniel Alvarez.

Address all correspondence and requests for reprints to: JuanBernal or Beatriz Morte, Instituto de Investigaciones Biomédi-cas, Arturo Duperier 4, 28029 Madrid, Spain. E-mail:jbernal@iib.uam.es or bmorte@iib.uam.es.

This work was supported by Grants SAF2011-25608 fromthe Plan Nacional de I�D�i, the Ramón Areces Foundation, andthe Center for Research on Rare Diseases (Ciberer), Instituto deSalud Carlos III, Madrid, Spain. P.G.-I. is recipient of a Junta deAmpliación de Estudios predoctoral fellowship from theConsejo Superior de Investigaciones Científicas.

Disclosure Summary: The authors have nothing to disclose.

doi: 10.1210/en.2012-2189 endo.endojournals.org 7

References

1. Cheng SY, Leonard JL, Davis PJ. Molecular aspects of thyroid hor-mone actions. Endocr Rev. 2010;31:139–170.

2. Hernandez A, Morte B, Belinchon MM, Ceballos A, Bernal J. Crit-ical role of types 2 and 3 deiodinases in the negative regulation ofgene expression by T3 in the mouse cerebral cortex. Endocrinology.2012;153:2919–2928.

3. Morte B, Ceballos A, Diez D, et al. Thyroid hormone-regulated mousecerebral cortex genes are differentially dependent on the source of thehormone: a study in monocarboxylate transporter-8- and deiodinase-2-deficient mice. Endocrinology. 2009;151:2381–2387.

4. Flamant F, Baxter JD, Forrest D, et al. International Union of Phar-macology. LIX. The pharmacology and classification of the nuclearreceptor superfamily: thyroid hormone receptors. Pharmacol Rev.2006;58:705–711.

5. Umesono K, Murakami KK, Thompson CC, Evans RM. Direct re-peats as selective response elements for the thyroid hormone, reti-noic acid, and vitamin D3 receptors. Cell. 1991;65:1255–1266.

6. Williams GR, Zavacki AM, Harney JW, Brent GA. Thyroid hor-mone receptor binds with unique properties to response elementsthat contain hexamer domains in an inverted palindrome arrange-ment. Endocrinology. 1994;134:1888–1896.

7. Fondell JD, Brunel F, Hisatake K, Roeder RG. Unliganded thyroidhormone receptor � can target TATA-binding protein for transcrip-tional repression. Mol Cell Biol. 1996;16:281–287.

8. Saatcioglu F, Claret FX, Karin M. Negative transcriptional regula-tion by nuclear receptors. Semin Cancer Biol. 1994;5:347–3459.

9. Martinez de Arrieta C, Morte B, Coloma A, Bernal J. The humanRC3 gene homolog, NRGN contains a thyroid hormone-responsiveelement located in the first intron. Endocrinology. 1999;140:335–343.

10. Thompson CC. Thyroid hormone-responsive genes in developingcerebellum include a novel synaptotagmin and a hairless homolog.J Neurosci. 1996;16:7832–7840.

11. Denver RJ, Williamson KE. Identification of a thyroid hormoneresponse element in the mouse Kruppel-like factor 9 gene to explainits postnatal expression in the brain. Endocrinology. 2009;150:3935–3943.

12. Bernal J, Morte B. Thyroid hormone receptor activity in the absenceof ligand: Physiological and developmental implications [publishedonline April 24, 2012]. Biochim Biophys Acta. doi:10.1016/j.bbagen.2012.04.014.

13. Morte B, Manzano J, Scanlan T, Vennstrom B, Bernal J. Deletion ofthe thyroid hormone receptor �1 prevents the structural alterationsof the cerebellum induced by hypothyroidism. Proc Natl Acad Sci US A. 2002;99:3985–3989.

14. Wallis K, Sjogren M, van Hogerlinden M, et al. Locomotor defi-ciencies and aberrant development of subtype-specific GABAergicinterneurons caused by an unliganded thyroid hormone receptor �1.J Neurosci. 2008;28:1904–1915.

15. Yen PM, Feng X, Flamant F, et al. Effects of ligand and thyroidhormone receptor isoforms on hepatic gene expression profiles ofthyroid hormone receptor knockout mice. EMBO Rep. 2003;4:581–587.

16. Chan IH, Privalsky ML. Isoform-specific transcriptional activity ofoverlapping target genes that respond to thyroid hormone receptors�1 and �1. Mol Endocrinol. 2009;23:1758–1775.

17. Lin JZ, Sieglaff DH, Yuan C, et al. Gene specific actions of thyroidhormone receptor subtypes. PLoS One. 2013;8:e52407.

18. Zhu XG, Kim DW, Goodson ML, Privalsky ML, Cheng SY. NCoR1regulates thyroid hormone receptor isoform-dependent adipogene-sis. J Mol Endocrinol. 2011;46:233–244.

19. Lee S, Young BM, Wan W, Chan IH, Privalsky ML. A mechanismfor pituitary-resistance to thyroid hormone (PRTH) syndrome: aloss in cooperative coactivator contacts by thyroid hormone recep-tor (TR)�2. Mol Endocrinol. 2011;25:1111–1125.

20. Chiamolera MI, Sidhaye AR, Matsumoto S, et al. Fundamentallydistinct roles of thyroid hormone receptor isoforms in a thyrotrophcell line are due to differential DNA binding. Mol Endocrinol. 2012;26:926–939.

21. Denver RJ, Ouellet L, Furling D, Kobayashi A, Fujii-Kuriyama Y,Puymirat J. Basic transcription element-binding protein (BTEB) is athyroid hormone-regulated gene in the developing central nervoussystem. Evidence for a role in neurite outgrowth. J Biol Chem. 1999;274:23128–23134.

22. Schwartz HL, Strait KA, Ling NC, Oppenheimer JH. Quantitationof rat tissue thyroid hormone binding receptor isoforms by immu-noprecipitation of nuclear triiodothyronine binding capacity. J BiolChem. 1992;267:11794–11799.

23. Wikstrom L, Johansson C, Salto C, et al. Abnormal heart rate andbody temperature in mice lacking thyroid hormone receptor � 1.EMBO J. 1998;17:455–461.

24. Forrest D, Hanebuth E, Smeyne RJ, et al. Recessive resistance tothyroid hormone in mice lacking thyroid hormone receptor �: evi-dence for tissue-specific modulation of receptor function. EMBO J.1996;15:3006–3015.

25. Gothe S, Wang Z, Ng L, et al. Mice devoid of all known thyroidhormone receptors are viable but exhibit disorders of the pituitary-thyroid axis, growth, and bone maturation. Genes Dev. 1999;13:1329–1341.

26. Contreras-Jurado C, Garcia-Serrano L, Gomez-Ferreria M, CostaC, Paramio JM, Aranda A. The thyroid hormone receptors as mod-ulators of skin proliferation and inflammation. J Biol Chem. 2011;286:24079–24088.

27. Sugino K, Hempel CM, Miller MN, et al. Molecular taxonomy ofmajor neuronal classes in the adult mouse forebrain. Nat Neurosci.2006;9:99–107.

28. Diaz E. A functional genomics guide to the galaxy of neuronal celltypes. Nat Neurosci. 2006;9:10–12.

29. Guadano-Ferraz A, Escamez MJ, Morte B, Vargiu P, Bernal J. Tran-scriptional induction of RC3/neurogranin by thyroid hormone: dif-ferential neuronal sensitivity is not correlated with thyroid hormonereceptor distribution in the brain. Brain Res Mol Brain Res. 1997;49:37–44.

30. Morte B, Iniguez MA, Lorenzo PI, Bernal J. Thyroid hormone-reg-ulated expression of RC3/neurogranin in the immortalized hypo-thalamic cell line GT1–7. J Neurochem. 1997;69:902–909.

31. Iniguez MA, De Lecea L, Guadano-Ferraz A, et al. Cell-specificeffects of thyroid hormone on RC3/neurogranin expression in ratbrain. Endocrinology. 1996;137:1032–1041.

32. Chatonnet F, Guyot R, Picou F, Bondesson M, Flamant F. Genome-wide search reveals the existence of a limited number of thyroidhormone receptor � target genes in cerebellar neurons. PLoS One.2012;7:e30703.

33. Yuan C, Lin JZ, Sieglaff DH, et al. Identical gene regulation patternsof T3 and selective thyroid hormone receptor modulator GC-1. En-docrinology. 2012;153:501–511.

34. Diez D, Grijota-Martinez C, Agretti P, et al. Thyroid hormone ac-tion in the adult brain: gene expression profiling of the effects ofsingle and multiple doses of triiodo-L-thyronine in the rat striatum.Endocrinology. 2008;149:3989–4000.

35. Gil-Ibañez P, Bernal J, Morte B. 2012 Thyroid hormone action oncerebrocortical neurons in primary culture. In: Program of the 94thAnnual Meeting of The Endocrine Society, Houston, TX, 2012.Abstract SAT-397.

36. Hashimoto K, Curty FH, Borges PP, et al. An unliganded thyroidhormone receptor causes severe neurological dysfunction. Proc NatlAcad Sci U S A. 2001;98:3998–4003.

37. Portella AC, Carvalho F, Faustino L, Wondisford FE, Ortiga-Car-valho TM, Gomes FC. Thyroid hormone receptor � mutation causessevere impairment of cerebellar development. Mol Cell Neurosci.2010;44:68–77.

8 Gil-Ibañez et al TRs and Brain Gene Expression Endocrinology