A resistance to thyroid hormone syndrome mutant operates through the target gene repertoire of the wild-type thyroid hormone receptor

A resistance to thyroid hormone syndrome mutant operates through the target gene repertoire of the wild-type thyroid hormone receptor

Molecular and Cellular Endocrinology 447 (2017) 87e97 Contents lists available at ScienceDirect Molecular and Cellular Endocrinology journal homepag...

2MB Sizes 0 Downloads 18 Views

Molecular and Cellular Endocrinology 447 (2017) 87e97

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology journal homepage: www.elsevier.com/locate/mce

A resistance to thyroid hormone syndrome mutant operates through the target gene repertoire of the wild-type thyroid hormone receptor Robyn Jimenez, Martin L. Privalsky* Department of Microbiology and Molecular Genetics, College of Biological Sciences, University of California at Davis, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 August 2016 Received in revised form 26 February 2017 Accepted 27 February 2017 Available online 28 February 2017

Thyroid hormone receptors (TRs) play crucial roles in vertebrates. Wild-type (WT) TRs function primarily as hormone-regulated transcription factors. A human endocrine disease, Resistance to Thyroid Hormone (RTH)-Syndrome, is caused by inheritance of mutant TRs impaired in the proper regulation of target gene expression. To better understand the molecular basis of RTH we compared the target genes regulated by an RTH-TRb1 mutant (R429Q) to those regulated by WT-TRb1. With only a few potential exceptions, the vast majority of genes we were able to identify as regulated by the WT-TRb1, positively or negatively, were also regulated by the RTH-TRb1 mutant. We conclude that the actions of R429Q-TRb1 in RTHSyndrome most likely reflect the reduced hormone affinity observed for this mutant rather than an alteration in target gene repertoire. Our results highlight the importance of target gene specificity in defining the disease phenotype and improve our understanding of how clinical treatments impact RTHSyndrome. © 2017 Elsevier B.V. All rights reserved.

Keywords: Thyroid hormone receptor Target genes RTH RNA-seq Endocrine disease

1. Introduction Thyroid hormones (chiefly T3 tri-iodothyronine and T4 tetraiodothyronine) play multiple crucial roles in vertebrate differentiation and homeostasis, including control of overall metabolic rate, body temperature, neural development, heart rate, color vision, and hearing (Forrest and Vennstrom, 2000; Larsen et al., 2003; Mullur et al., 2014; Yen, 2015). Thyroid hormones manifest most of these actions by binding to thyroid hormone receptors (TRs): members of the nuclear receptor family of ligand-regulated transcription factors (Brent, 2012; Cheng et al., 2010; Flamant et al., 2006). Three major TR isoforms, denoted TRa1, TRb1, and TRb2, are encoded from distinct genetic loci and by alternative splicing; they are expressed at different abundances in different tissues and play distinct biological roles (Chan and Privalsky, 2009a,b; Cheng, 2005a,b; Forrest and Vennstrom, 2000; Lazar, 1993; Lin et al., 2013). TRs are thought to function primarily through their ability to bind to specific target genes and to regulate gene expression either up or down in response to binding of thyroid hormone (Cheng et al., 2010; Flamant et al., 2006; Viguerie and Langin, 2003; Zhang and

* Corresponding author. Department of Microbiology and Molecular Genetics, One Shields Avenue, University of California at Davis, Davis, CA 95616 USA. E-mail address: [email protected] (M.L. Privalsky). http://dx.doi.org/10.1016/j.mce.2017.02.044 0303-7207/© 2017 Elsevier B.V. All rights reserved.

Lazar, 2000). The best understood of these target genes display “positive regulation” by thyroid hormone. TRs bound to these genes recruit auxiliary proteins, denoted corepressors, in the absence of hormone and repress transcription (Astapova and Hollenberg, 2013; Moehren et al., 2004; Privalsky, 2004; Zhang and Lazar, 2000). Conversely in the presence of thyroid hormone the TRs release corepressors, bind a distinct set of coactivators, and induce gene expression (Brent, 2012; Feige and Auwerx, 2007; Stashi et al., 2014). However, additional, less well-elucidated modes of transcriptional regulation by TRs have also been observed, including the “negative regulation” of target genes that are induced in the absence of thyroid hormone yet repressed in its presence (e.g. (Abel et al., 2001; Chan and Privalsky, 2009a,b; Feng et al., 2000; Furumoto et al., 2005; Gloss et al., 2005; Kim et al., 2005; Lin et al., 2013; Miller et al., 2004; Moehren et al., 2004; Nakano et al., 2004; Ortiga-Carvalho et al., 2005; Viguerie and Langin, 2003; Wang et al., 2009; Weitzel et al., 2003; Yen et al., 2003)). Not only the directionality, but also the magnitude and hormonesensitivity of the T3-mediated gene response differs at different target genes (e.g. (Chan and Privalsky, 2009a,b; Chatonnet et al., 2014; Cheng et al., 2010; Lin et al., 2013)). These gene-specific regulatory effects by TRs are likely due to recruitment of distinct sets of coregulatory proteins when the receptor is bound to different genes and/or interactions of the TRs with additional transacting factors co-assembled on each given target gene (e.g. (Ayers

88

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

et al., 2014; Wang et al., 2010)). Human Resistance to Thyroid Hormone (RTH)-Syndrome is a genetic disorder most often resulting from inherited mutations in TRb1 (Brent, 2012; Chatterjee, 2008; Cheng, 2005a,b; Dumitrescu and Refetoff, 2013; Jameson, 1994; Weiss et al., 1997). RTH-TRb1 mutants typically have sustained base substitutions within their hormone binding and/or transcriptional regulatory domains that render them defective in the ability to bind and/or respond to hormone. As a result RTH-TRs typically bind corepressors in the absence of T3, yet fail to release these corepressors and/or acquire coactivators properly on addition of T3 (Privalsky, 2008; Safer et al., 1998; Yen, 2003; Yoh et al., 1997). When coexpressed with WT-TR, RTH-TRs can act as dominant negatives and attenuate WT-TR function in response to T3 (Collingwood et al., 1994; Hayashi et al., 1995; Jameson, 1994; Matsushita et al., 2000; Miller et al., 2004; Wondisford, 2003). Two broad categories of clinical RTH have been proposed: generalized (GRTH) and pituitary-specific (PRTH) (Chatterjee, 2008; Cheng, 2005a,b; Jameson, 1994; Olateju and Vanderpump, 2006; Refetoff and Dumitrescu, 2007; Wondisford, 2003; Yen, 2003). GRTH manifests as a broad failure to respond to circulating T3, resulting in a loss of both negative feedback on T3 synthesis in the hypothalamus/pituitary/thyroid axis (HPT) and of T3 sensing in the peripheral tissue; GRTH mimics several of the symptoms of hypothyroidism despite the presence of elevated T3, TSH, and TRH levels. In contrast, PRTH reflects a selective loss of negative feedback sensing of T3 in the HPT axis (leading to large increases in circulating T3) but with significant retention of T3 responsiveness in the periphery; PRTH therefore mimics certain symptoms of hyperthyroidism. The divergent PRTH versus GRTH phenotypes likely reflect differences in the ability of a given RTH-TRb mutant to interact with coactivators and corepressors when expressed as the TRb2 isoform (primarily in the hypothalamus and pituitary) versus as the TRb1 isoform (primarily in peripheral tissues) (Lee et al., 2011; Machado et al., 2009; Wan et al., 2005). Significantly, several outwardly similar dominant-negative TRb1 (and TRa1) mutants are associated not with endocrine disorders but with two forms of human neoplasia: hepatocellular carcinoma (HCC) and renal clear cell carcinoma (RCCC) (Kamiya et al., 2002; Lin et al., 2001). Most of these neoplasia-associated TRa1 and TRb1 mutants have lesions not only in their hormone-binding/ transcriptional regulatory regions, but also in regions able to alter their DNA recognition specificity (Chan and Privalsky, 2006; Rosen et al., 2011). Apparently as a result, the HCC and RCCC TR mutants display target gene repertoires distinct from those of the corresponding WT-TR isoforms (Chan and Privalsky, 2006, 2009, Chan and Privalsky, 2009a,b; Chan and Privalsky, 2006; Rosen et al., 2011). We proposed that the HCCe and RCCC-TR mutations alter both the transcriptional properties and target gene recognition of the encoded TRs, resulting in novel gene repertoires that lead to an aberrant regulation of oncogenes and/or tumor suppressors not normally regulated by the WT-TRs and thereby contributing to the neoplastic outcomes associated with these particular mutants. In contrast we suggested that the RTH-TR mutations change the transcriptional properties/hormone binding of the encoded receptor, but not its DNA recognition or gene repertoire, leading to the endocrine disruptions characteristic of RTH-Syndrome. Importantly the target gene repertoires of the RTH-TRs compared to the WT-TRs were not previously investigated comprehensively, and our prior, very limited qRT-PCR analysis raised the possibility that even the RTH-mutants might differ in certain of their target gene preferences from those of the WTreceptors (Rosen et al., 2011). Given that such gene-specific aberrations would have profound consequences for understanding and treating the RTH-Syndrome phenotype, we returned to this

question by using RNA-seq to more broadly survey target gene regulation by an RTH-TRb1 mutant versus by WT-TRb1 (Fig. 1A). We focused our initial analysis on R429Q-TRb1, a mutant associated with PRTH and possessing several altered molecular properties that potentially might impact on its target gene repertoire (Adams et al., 1994; Clifton-Bligh et al., 1998; Collingwood et al., 1994; Kong et al., 2005; Machado et al., 2009; Safer et al., 1997). We report here that the WT-TRb1 displayed a variety of gene-specific regulatory properties, both repressive and activating, by this analysis and that within the statistical limitations of our study the R429Q-TRb1 mutant displayed virtually the same regulatory properties and target gene repertoire as did the WT-TRb1 on the genes surveyed here. This was observed both in the absence of added T3 and in the presence of saturating hormone. We propose that the R429Q-TRb1 mutant retains the inherent potential to regulate most of the same genes as does WT-TRb1 at very high T3 levels, but contributes to RTH-Syndrome in vivo primarily due to its impaired ability to bind and respond to the lower T3 concentrations found physiologically. These results are consistent with the proposal that TRb1 mutants displaying a primarily WT-TR target gene repertoire produce primarily endocrine disorders, whereas TRb1 mutants displaying altered target gene repertoires, such as those associated with hepatocellular and renal clear cell carcinomas, contribute to the neoplastic phenotypes of these diseases by regulating genes not associated with normal thyroid endocrinology. 2. Materials and methods 2.1. Cell culture HepG2 cells and HEK293T cells were maintained in Dulbecco's Modified Eagles Medium (DMEM) containing glutamine and 4.5 g/L glucose (Gibco ThermoFisher, Waltham MA), supplemented with 10% fetal bovine serum (FBS, Gibco/ThermoFisher low-endotoxin “Performance-Grade”) at 37  C in a humidified 5% CO2 atmosphere. 2.2. Creation of the adenoviral vectors Adenovirus vectors lacking an insert (AdEmpty), containing the

Fig. 1. Schematic of TRb1 protein and its expression in vector-infected cells. A. The TRb1 protein is presented schematically with the DNA and hormone binding domains indicated. Domains that mediate interaction with or exchange of coregulators (“CoReg”) are depicted, as is the position of the R429Q mutation. B. The expression of TRb1 mRNA is shown for HepG2 cells infected by the empty, WT, or R429Q adenoviral vectors in the absence or presence of added T3. The mean and standard deviation is shown (n ¼ 3).

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

WT-TRb1 open reading frame (AdWT-TRb1), or containing the R429Q-TRb1 open reading frame (AdR429Q-TRb1) were prepared using Ad-Easy constructs (Addgene, Cambridge MA) (Luo et al., 2007). In brief, standard PCR, restriction cleavage, and ligation methods were employed on each TR cDNA to generate pAdTrackCMV bacterial shuttle vectors containing the relevant inserts (or the no-insert pAdTrack-Empty control). These shuttle vectors were then linearized with EcoRI and electroporated into the BJ5183 AdEasier Escherichia coli strain, permitting in vivo recombination with the BJ5168 pAdEasy-1 bacmid and generating full-length recombinant adenoviral vector DNAs. Bacteria that had undergone this recombination were selected by growth in kanamycin. The resulting vector DNA constructs, which contained the CMV promoter-only (the Empty control), the WT-TRb1 allele, or the R429Q-TRb1 allele, also expressed Green Fluorescent Protein (GFP), allowing the extent of infection of host cells to be monitored by fluorescence. The recombinant viral vector DNAs were purified from the E. coli and used in a Lipofectamine (ThermoFisher)mediated transfection of HEK293T cells, which package the recombinant vector DNA into infectious adenoviral particles. After 21 days (or after GFP fluorescence was widely detected in the transfected cells) the infectious adenoviruses were harvested from the HEK293T cells by freeze/thaw. These stocks were further amplified by successive rounds of HEK293T cell infection, and were harvested. Viral titers were determined by endpoint dilution on HEK293T cells and visible cytopathicity. Viral stocks were stored at 80  C. 2.3. Infection and T3 treatment of HepG2 cells HepG2 cells were plated at 3  105 cells per well in 6 well tissue culture plates in 2 mL DMEM-10% hormone-stripped FBS (Sharif and Privalsky, 1992) and were incubated overnight at 37  C in a 5% CO2 atmosphere. The next day 1 mL of fresh DMEM-10% hormone-stripped FBS medium containing the desired adenoviral stock (AdEmpty, AdWT-TRb1, or AdR429Q-TRb1) was added per well at a multiplicity of infection of 10. Although adenovirus vectors efficiently deliver ectopic genes into HepG2 cells, these cells lack the helper functions that permit the secondary rounds of viral replication required for our cytological end-point assay; we therefore confirmed by GFP-fluorescence that a majority of the HepG2 cells were infected and expressed the adenovector-encoded gene products (consistent with the qRT-PCR and RNA-seq experiments, below). After an additional 24 h incubation, 100 nM T3 (or the equivalent volume of ethanol carrier alone) was added to the appropriate cultures and the cells were incubated for a further 6 h. The cells were then harvested by washing and centrifugation, and were stored in RNAProtect (Qiagen, Valencia, CA) at 80  C prior to RNA isolation (below). Three biological replicates, performed on separate days, were analyzed for each adenoviral vector derivative and hormone concentration. 2.4. RNA isolation and cDNA synthesis RNA was isolated using QiaShredder cell-homogenization kit (Qiagen) and an RNeasy RNA-isolation kit (Qiagen) employing the manufacturer's recommended protocols. All RNA samples were quality checked using a 2100 Bioanalyzer and RNA 6000 Nano LabChips (Agilent, San Jose CA). RNA Integrity Number (RIN) scores were 8.0 or higher. High-throughput cDNA sequencing libraries were prepared from 1 mg of each RNA sample using a NEBNext Ultra RNA Library Prep kit (New England Biolabs, Ipswich MA). An Illumina bar code was used for each sample during the PCR library enrichment step by employing NEB Index Primers. Quality determinations were performed on each cDNA library using a DNA

89

High Sensitivity LabChip and a 2100 Agilent Bioanalyzer. Acceptable cDNA fragment distributions of 500e700 bp were achieved for each sample library and any remaining enrichment primers were removed before pooling. 2.5. Library pooling and high-throughput sequencing Samples were quantified employing Qubit and a 2100 Bioanalyzer. The bar-coded cDNA libraries were pooled to generate equimolar representations of each cDNA sample and the pools were subjected to high-throughput DNA sequence analysis on a HiSeq 3000 (Illumina, San Diego CA) by the UC Davis DNA Technology Core facility. A majority of the single-end reads yielded 49 to 51 base pair lengths, as desired; a total of 16 million reads (representing some 900,000 reads per individual bar-coded sample, or 2.7 million reads per vector/hormone combination as analyzed in triplicate) were obtained. These numbers allow a survey, but not a complete in-depth cataloging, of gene expression in these cells (discussed in section 4.3). 2.6. RNA-sequence bioinformatics Raw data from the HighSeq 3000 was analyzed using R on the U.C. Genome Center Galaxy Instance at the UC Davis Bioinformatics Core Facility Further analysis and filtering were achieved by sequential application of Fast QC, Scythe, Sickle, and TopHat 2.0 (Anders et al., 2015; Blankenberg et al., 2010; Goecks et al., 2010; Raghavachari et al., 2012; Trapnell et al., 2012). Count reads and pairwise comparisons of differential expression were determined using HTseq-count and EdgeR (Anders et al., 2015; Nikolayeva and Robinson, 2014). Quality scores confirmed an acceptable average Phred score of 40 or above. 2.7. Quantitative reverse transcriptase-polymerase chain reaction qRT-PCR analysis was performed using a LightCycler 480 II (Roche Life Sciences, Indianapolis IN), employing the manufacturer's reagents and protocols, and using gene-specific oligonucleotide primers (Chan and Privalsky, 2009; Rosen et al., 2011). 3. Results 3.1. Adenovirus vector infection led to efficient expression of WT or mutant TRs in HepG2 cells Thyroid hormone, operating primarily through the TRb1 receptor, plays a key role in the regulation of normal hepatic physiology and metabolism (Brent, 2000; Mullur et al., 2014; Vatner et al., 2013; Weitzel et al., 2003; Yen et al., 2003). The established HepG2 cell line expresses many normal liver markers and is a widely studied model of hepatocyte function (e.g. (Meex et al., 2011)). Unmanipulated HepG2 cells express only low levels of endogenous TRs (Chan and Privalsky, 2009a,b), allowing us to introduce a TR allele of choice by infection with an adenoviral vector expressing WT-TRb1 (AdWT-TRb1), with an adenovirus vector expressing R429Q-TRb1 (AdR42Q-TRb1), or with an empty adenovirus vector control (AdEmpty). The infected cells were permitted to express the receptor for 24 h and then treated with ethanol carrier alone or with 100 nM T3 for an additional 6 h. The 6 h time course was chosen to focus primarily on the target genes directly responsive to T3, rather than on any secondary effects mediated indirectly by T3-induction/repression of downstream transcription factors. Consistent with prior observations that HepG2 cells exhibit little or no endogenous TR function, RNA-seq analysis of HepG2 infected

90

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

by the AdEmpty control displayed very low levels of endogenous TR mRNA (Fig. 1B) and only a small number of statistically-significant changes in the expression of specific host genes in response to addition of 100 nM T3 (Table 1). In contrast, HepG2 cells infected by either the AdWT-TRb1 or the AdR429Q-TRb1 viral vector expressed high levels of the corresponding TRb1 mRNA (Fig. 1B), indicating that these vectors efficiently expressed the ectopic TR alleles. Expression of either ectopic TRb1 mRNA was decreased modestly by addition of T3 (likely due to T3-inhibition of the pCMV promoter used in the Adeno-vectors; unpublished results) but remained robust and at comparable levels for the WT and R429Q receptors (Fig. 1B). Attempts to detect either of the TR proteins by immunoblotting were not successful (most likely due to the limitations of the antisera available to us; unpublished observations); however the protein levels of the mutant and WT receptors are likely to be comparable to one another based on the close identity of their engineered transcripts and on in vitro translation experiments ((Lee et al., 2011) and MLP unpublished results). Expression of either TR allele resulted in the robust regulation of characteristic panels of target genes whose expression patterns were detectably altered from that seen in HepG2 cells infected by the AdEmpty vector. Many genes exhibited differences in their level of expression in the cells infected by the Ad-WT-TRb1 vector compared to cells infected by the Ad-empty vector even in the absence of added T3 (i.e. hormone-stripped media) (Fig. 2). Under these conditions, the bulk of these candidate TR-target genes displayed lower expression in the presence of WT-TRb1 than in the absence of the receptor (shades of red in Fig. 2); a smaller additional panel of genes displayed the opposite pattern by being expressed more in the presence of the WT-TRb1 than it its absence (shades of green in Fig. 2). The alteration in expression of many of these genes in response to the presence of WT-TRb1, up or down, was consistent with a simple p value of 0.05; however only seven met the statistically more stringent and meaningful requirement of an adjusted p value of 0.05 (Fig. 2 and Supplemental Table 1). Of these seven, the expression of two genes displayed the “classical” pattern of WT-TRb1-associated repression in the absence of T3, whereas five other genes exhibited the converse pattern of higher expression in the presence of the AdWT-TRb1 than in its absence (Fig. 2, lane 2 and Supplemental Tables 1 and 2). We performed the same type of analysis comparing gene expression minus or plus WT-TRb1 but now both in the presence of 100 nM T3 (this high T3 level in the media was chosen in part to help establish receptor-saturating T3 concentrations within the HepG2 cells as rapidly as possible during the short 6 h time course of the experiments). In this latter experiment expression of 121 genes displayed statistically significant differences (adjusted p value ¼ 0.05) when cells expressing WT-TRb1 were compared to those not (Fig. 2, lane 3, and Supplemental Tables 1 and 2); yetadditional genes displayed similar trends but did not reach statistical significance by our adjusted p value criterion (Fig. 2 and

Table 1 Changes in gene expression in HepG2 in the absence of ectopic TR. HepG2 cells infected by the AdEmpty control vector were analyzed by RNA-seq minus or plus 100 nM T3. Only 6 genes displayed a statistically significant response under these conditions (i.e. adjusted p values ¼ 0.05). Fold-change (log2) and adjusted p values shown. Gene

Change ± T3

Adjusted p Value

TP53INP2 COQ10A ARSB KLF9 KCNJ10 ICOSLG

1.13 0.55 0.62 0.46 0.52 0.31

3.89E-05 0.009 0.023 0.023 0.023 0.024

Supplemental Table 1). As in the absence of added T3, both genes repressed by and genes activated by the WT-TRb1 (relative to the empty vector control) were detected in the presence of 100 nM T3. Similar patterns of gene expression were observed when representative individual examples of these target genes were analyzed by qRT-PCR (Supplemental Figure S1). We note that yet-additional WT-TRb1 target genes, minus or plus T3, may exist but which would not have been detected given the read depth obtained in our RNA-seq. The above comparison focused on the changes in target gene expression in response to the introduction of WT-TRb1 versus its absence, with each comparison performed separately either minus or plus added T3. We next performed a related, but conceptually distinct comparison focusing on the changes in target gene expression specifically in response to adding 100 nM T3 versus not, with both hormone conditions assayed in the presence of the WTTRb1 (Fig. 3A and Supplemental Tables 1 and 2). We discerned at least three general patterns of hormone-regulated genes under these conditions (i.e. all in the presence of the WT-TRb1): (a) genes whose expression was altered by the WT-TRb1 compared to no receptor in Fig. 2, but was not significantly further altered by T3 (black in Fig. 3A) and (b) genes whose expression was increased in response to T3 (green in Fig. 3A), (c) genes whose expression was reduced in response to T3 (red in Fig. 3A). Within each T3-regulated class the magnitude of the change in response to hormone and the absolute levels of expression varied from target gene to target gene (Fig. 3A and quantified/displayed as bar graphs for a panel of representative genes in Fig. 4A). In contrast, and as noted earlier, very few genes responded to T3 in the absence of the ectopic WTTRb1 (AdEmpty control; Table 1). We conclude that the effects of WT-TRb1 and of T3 on gene expression are specific for each individual gene, presumably reflecting the specific coregulatory proteins recruited by the WT-TRb1 and the contributions of adjacent trans-acting factors at each specific target gene. 3.2. Within the statistical limitations of our analysis, the R429QTRb1 mutant displayed a target gene repertoire that closely paralleled that of WT-TRb1 The mutant TRs isolated from human RCCCs and HCCs display altered target gene repertoires compared to WT-TRs, and we have proposed that these alterations in target gene specificity contribute to the neoplastic nature of the diseases associated with these mutants. We further speculated that RTH-TRb1 mutants would, in contrast, display a WT-TRb1 repertoire, thus accounting for the endocrine rather than oncogenic nature of RTH-Syndrome. However this had not been rigorously examined, and a preliminary analysis, examining only a limited number of genes, raised the possibility that even RTH-TRb1 mutants might diverge from WTTRb1 in their target gene specificities. To help resolve this question and test our hypothesis, we employed RNA-seq to more broadly evaluate the target gene repertoire of an RTH-mutant receptor, R429Q-TRb1. As with the WT receptor, we performed these experiments on the R429Q-TRb1 mutant both in the absence and in the presence of 100 nM added T3 (the latter is saturating for T3 binding for both the WT and mutant receptors despite the 2e5 fold impairment in T3 binding reported for the mutant; notably both receptors display otherwise similar sigmoidal T3 binding and EC50 curves (Adams et al., 1994; Clifton-Bligh et al., 1998; Lee et al., 2011; Wan et al., 2005)). This strategy therefore allowed us to compare the target gene repertoires of the two receptor forms at two extremes: when unliganded or when fully engaged by hormone. Notably the overall pattern of target gene regulation observed for WT-TRb1 was also observed for the R429Q-TRb1 mutant whether in the absence of added T3 (Fig. 2, lanes 1 versus 2) or in

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

91

Fig. 2. Heat map representation of changes in cell gene expression in response to WT-TRb1 or R429Q-TRb1 versus an empty vector. HepG2 cells were infected by the empty viral vector control, the WT-TRb1 viral vector, or by the R429Q-TRb1 viral vector and the cells were analyzed by for gene expression by RNA-seq. The change in gene expression in response to the WT-TRb1 (lanes 2 and 3) or to R429Q-TRb1 (lanes 1 and 4), each relative to the no receptor control, was determined (relative expression ¼ expression in TRb1-cells divided by expression in empty vector cells). The log2 of this relative expression value, down or up, is displayed as a color (see key). The comparisons were performed separately in the absence (lanes 1 and 2) or the presence (lanes 3 and 4) of 100 nM T3. Blue dots indicate responses with an adjusted p ¼ 0.05 (n ¼ 3 for each condition). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the presence of saturating 100 nM T3 (Fig. 2, lanes 3 versus 4). Please note that most differences in the shading between R429Q versus WT in Fig. 2 did not rise to the level of statistically significant differences in expression (as detailed in Supplemental Tables 1 and 2). Similarly, the three generic patterns of T3 response observed in cells infected by the WT-TRb1 vector (no detectable response, T3driven repression, or T3-driven activation) were also observed in the cells infected by the R429Q-TRb1 vector (Fig. 3B).

The general congruence between R429Q-TRb1 and WT-TRb1mediated regulation was also observed in analysis of individual examples of these target genes by RNA-seq (compare Fig. 4B versus Fig. 4A) or by qRT-PCR (Supplemental Figure S1). We conclude that the R429Q mutant possesses “inherent” transcriptional properties, both negative and positive, that for the target genes identified here largely overlap those of the WT-TRb1 receptor and that are unmasked at the saturating 100 nM T3 levels employed in these

92

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

Fig. 3. Effect of T3 on target genes in the presence of WT-TRb1 or R429Q-TRb1. The change in expression in response to T3 (expression in 100 nM T3 versus expression in the absence of added T3) was calculated for the genes shown and is displayed as a color (see key). Blue dots indicate responses with an adjusted p ¼ 0.05 (n ¼ 3 for each condition). A. HepG2 cells infected by the WT-TRb1 vector. B. HepG2 cells infected by the R429Q-TRb1 vector. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

experiments. In contrast, we propose that at the lower physiological hormone levels found in patients the R429Q-TRb1 mutant is impaired in its overall transcriptional response to T3 compared to WT-TRb1 and we speculate that it is this divergence in hormone affinity, rather than target gene specificity, that is an important mechanistic basis behind the RTH-Syndrome phenotype conferred by this mutant in vivo. Contrary to this general congruity in target gene repertoire between the R429Q mutant and WT-TRb1, approximately 1.2% of the total target genes analyzed exhibited differences in their responses to the mutant versus WT-receptors when evaluated by non-

Fig. 4. Quantitation of expression of representative TR target genes. The expression levels (determined as frequency of reads in the RNA-seq analyses) are presented as bar graphs for a panel of representative TRb1 target genes. A. WT-TRb1. B. R429Q-TRb1. Means and standard errors (n ¼ 3) are shown.

adjusted p value, although none when evaluated by the more stringent (and more appropriate) adjusted p value criterion (e.g. Wnt9A in Fig. 5). We therefore do not rule out the possibility that certain target genes may indeed be divergently regulated by WTTRb1versus R429Q-TRb1; however our existing data does indicates that if these differentially regulated target genes do exist they are relatively rare within the portion of the TR-transcriptome characterized here. 3.3. Gene ontology (GO) analysis indicated that WT-TRb1 and R429Q-TRb1 regulate genes that are involved in a variety of cellular pathways including metabolic, cell signaling, and developmental processes A PANTHER GO analysis (Mi et al., 2013) revealed that the gene regulated by WT-TRb1 in the HepG2 cells participate in a broad

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

Fig. 5. Regulation of WNT9A by WT-TRb1 appears to differ from that by R429QTRb1. Analysis was performed as in Fig. 4.

range of biological pathways (Fig. 6). Many of these target genes are associated with metabolic processes, including several implicated in ATP synthesis, glycolysis, and the TCA cycle; this observation is consistent with known effects of TRb1 in liver. Additional WT-TRb1 responsive genes identified in this study are associated with cell signaling and developmental processes, also consistent with known biological roles of TRs. Due to the highly overlapping nature of the WT-TRb1 and R429Q-TRb1 gene repertories, the comparable analysis for R429Q-TRb1 revealed essentially the same gene ontologies as for WT-TRb1 (not shown). 4. Discussion 4.1. WT-TRb1 regulates, both up and down, a large panel of target genes in the liver-derived HepG2 cell line Our strategy was to not restrict our analysis to identification of genes whose expression changed in response to T3 hormone, but to more broadly identify genes whose expression was altered by the presence of WT-TRb1 compared to that of a no-receptor control and to examine this question separately in the absence and in the presence of added T3. This approach allowed us to identify several genes whose expression is altered by the introduction of the WT-

93

TRb1 in the absence of added T3 yet not further altered by a subsequent addition of 100 nM T3. TRs have been shown to possess both constitutive and hormone-dependent regulatory properties in other experimental contexts (e.g. (Chan and Privalsky, 2009a,b; Helmer et al., 1996)). TRs were initially proposed to function by repressing the transcription of target genes in the absence of T3 and by activating these same genes in the presence of T3. Examination of our RNAseq data revealed a set of HepG2 genes that, as predicted by this model, was repressed compared to a no receptor control in the absence of added T3. This overt repression in the absence of added T3 displayed statistical significance for only a few of the genes analyzed, although many more genes appeared to parallel this trend (if insufficiently to achieve significance by adjusted p value). In contrast over one hundred genes exhibited statistically significant activation by the WT-TRb1 in the presence of 100 nM T3 when compared to the no-receptor controls. Greater numbers of WTTRb1 target genes were also observed in the presence versus in the absence of T3 in a prior study of HepG2 cells using stable transformation and microarray analysis (Chan and Privalsky, 2006). The larger number of target genes displaying statistically significant differences in response to WT-TRb1 in the presence of added T3 (121 genes) versus its absence (7 genes) may reflect authentic differences in the numbers of target genes regulated under each hormone condition. Alternatively this difference may reflect a technical limitation of the assay (e.g. if comparable numbers of target genes are regulated minus or plus T3 but the fold-change is less in the former, which would make identifying the minus-T3 target genes more difficult). Although our RNA-seq. analysis represents a survey of many, rather than all possible, TR target genes in these cells (see section 4.3), we were nonetheless able to identify several broad categories of hormone response for the different target genes that were identified in our study. As noted above, one class of target genes responded to WT-TRb1 in an apparently T3-independent fashion. A second class displayed the “classical” repression by WT-TRb1 in the absence versus activation in the presence of T3. A third class displayed the opposite phenotype: activation by the WT-TRb1 in the absence of T3 versus repression in the presence of T3. This

Fig. 6. PANTHER Gene Ontogeny (GO) analysis of the patterns of gene regulation altered by WT-TRb1. Each pathway is listed below; the height of each bar represents the number of genes in that pathway altered in expression in response to WT-TRb1 in the presence of 100 nM T3 (pathways with only one gene altered by TRb1 are not displayed).

94

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

“negative regulation” of a subset of target genes, although initially unanticipated, has now been revealed in a variety of other studies. (e.g. (Abel et al., 2001; Chan and Privalsky, 2009a,b; Feng et al., 2000; Furumoto et al., 2005; Gloss et al., 2005; Kim et al., 2005; Lin et al., 2013; Miller et al., 2004; Moehren et al., 2004; Nakano et al., 2004; Ortiga-Carvalho et al., 2005; Viguerie and Langin, 2003; Wang et al., 2009; Weitzel et al., 2003; Yen et al., 2003)). The percentage of class 3 target genes in our own study is more modest than that reported in several of these prior studies; this may be a consequence of the specific cell type and receptor-expression methodology we employed here, and/or some other technical aspect of our analysis. In this regard we emphasize that our experiments were restricted to examining gene expression at 6 h post T3 treatment, and therefore would not reveal possible gene-specific differences in the kinetics of gene induction or repression in between these “end-point” values (e.g. (Yen et al., 2003)). There were additional target genes whose expression decreased or increased in response to T3, but did not display a complete inversion of transcriptional regulation. Several genes were not detectably affected by the TR in the absence of added T3, yet were either activated or repressed by the TR in the presence of T3. Yet other genes were mildly repressed or activated by the TR in the absence of T3, and were still more strongly repressed or activated by the TR in the presence of T3. Presumably the specific coregulatory proteins and trans-acting factors present on each TRoccupied enhancer/promoter fine-tune the transcriptional response of each target gene so as best fit its particular biological role (e.g. (Abel et al., 2001; Chan and Privalsky, 2009a,b; Feng et al., 2000; Furumoto et al., 2005; Gloss et al., 2005; Kim et al., 2005; Lin et al., 2013; Miller et al., 2004; Moehren et al., 2004; Nakano et al., 2004; Ortiga-Carvalho et al., 2005; Viguerie and Langin, 2003; Wang et al., 2009; Weitzel et al., 2003; Yen et al., 2003)). 4.2. The R429Q-TRb1 mutant exhibited a target gene repertoire and regulatory properties generally parallel to those of WT-TRb1 when its hormone-affinity deficient was rescued We wished to determine whether the TRb1 mutants associated with RTH Syndrome are simple anti-morphs of wild-type receptor, or if they also display alterations in their target gene repertoire. The latter was observed for the TR mutants associated with renal and hepatic carcinomas, and several published studies suggested that RTH-TR mutants might similarly be able to mediate target-gene specific effects distinct from those of wild-type receptor (e.g. (Miller et al., 2004; Rosen et al., 2011)). To better answer this question we tested the R429Q-TRb1 mutant either with no added hormone, or with a concentration of T3 sufficient to overcome the hormone-affinity impairment of this mutant. Under either of these conditions the panel of target genes for the R429Q-TRb1 mutant was, within statistical error, strongly overlapping with those for the WT-TRb1. Notably this shared regulatory properties of both the WT- and R429Q-TRs included the ability to not only repress certain genes in the absence of T3 but also to activate others under the same conditions. Therefore whatever the molecular mechanism behind the ability of the WTTRb1 to activate certain genes even in the absence of T3, the RTH mutant was unaffected in this regard. This was not previously established and reveals a potentially important mechanism by which this RTH-mutant can, under specific circumstances, faithfully mimic WT-TRb1 function. Nonetheless we did identify a relatively small subset of TR target genes that displayed differences in their regulation by the R429QTRb1 mutant versus WT-TRb1 by non-adjusted P value criteria but not by adjusted P value criteria. We therefore do not fully rule out the possibility that a small panel of these “non-congruent” WT and

RTH-TR target genes may exist and, as such, might influence the RTH-phenotype in a distinct fashion from that of simple hypothyroidism. 4.3. Is the HepG2 cell/adenovirus-TR system an adequate approximation of hepatic physiology in vivo? Hepatocytes serve as key sites of T3 action in both normal and aberrant T3 endocrinology and HepG2 cells retain many of the biochemical markers found in these normal hepatocytes (Meex et al., 2011; Mullur et al., 2014; Viguerie and Langin, 2003; Webb, 2010). HepG2 cells do diverge from normal hepatocytes by expressing only low levels of endogenous TRs (Chan and Privalsky, 2009a,b; Lin et al., 2013). This provided an important experimental advantage in that these cells represent a clean background for the study of specific TR alleles via their ectopic expression, as was exploited for the experiments described here. This goal is difficult or impossible to achieve in intact animals or in primary hepatocytes. Of note: although our viral system expresses the ectopic TRs at higher levels than those found in vivo in normal liver, our prior studies indicate that the levels of TR expression in HepG2 cells are unlikely to significantly alter their target gene repertoire (e.g. (Chan and Privalsky, 2009a,b)). The vast majority of the HepG2 cells used in each of our analyses were infected by the corresponding adenoviral vector based on multiplicity of infection and on GFP-staining criteria; however it is plausible that a small percentage of each cell population escaped infection and therefore was not expressing the ectopic TR. Given that comparable levels of infection were employed for each TR allele and that HepG2 cells not expressing an ectopic TR exhibit little or no T3 response, the existence of these uninfected cells would, at most, modestly dilute the overall magnitude of the TR response observed for the cell population as a whole, but it should not alter the identities of the target genes that were identified or our overall conclusions. We also emphasize that our RNA-seq analysis, representing approximately 900,000 reads per sample, generated a survey but not a complete catalog of gene expression in the HepG2 cells. Although sufficient to allow a general comparison of the target gene regulatory properties of the WT and R429Q, this read depth falls short of a fully detailed statistical sampling, particularly of lower abundance mRNAs. Although we have no evidence to suggest so, it remains possible that a more highly powered analysis could reveal additional target genes whose regulatory properties differ from those analyzed here. In summary: as with any ex vivo study, it is unlikely that the HepG2 cell/ectopic TR system captures all aspects of endogenous TRs in mouse liver in vivo. Nonetheless this approach was sufficient to achieve our primary goal: a general comparison of the target gene repertoires of WT-TRb1 versus R429Q-TRb1 in a model cell type possessing many of the features found in a physiologically relevant T3-target tissue. At least several of the TR-regulated genes identified here were found previously to be regulated by T3 in livers in intact animals (e.g. (Viguerie and Langin, 2003; Yen et al., 2003)). 4.4. Is the R429Q-TRb1 mutant an appropriate model of RTHSyndrome? The R429Q mutation has been identified in multiple human patients, as well as engineered into a mouse model, and presents with a cluster of symptoms characteristic of PRTH: highly elevated T3 and T4 with unsuppressed TSH and TRH together with tachycardia and other symptoms of peripheral thyrotoxicity (Adams et al., 1994; Collingwood et al., 1994; Kong et al., 2005; Machado et al., 2009; Safer et al., 1997). This mutation has been reported to exert a wide range of effects on the molecular biology of the

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

encoded TRb1/TRb2 receptors, including alterations in receptor conformation, in T3 binding affinity, in homodimerization, in coregulator interactions, and in regulation of negative response genes versus positive response genes (e.g. (Adams et al., 1994; Chatonnet et al., 2011; Lee et al., 2011; Machado et al., 2009; Safer et al., 1997; Togashi et al., 2005; Wan et al., 2005)). Many of these effects of the R429Q mutation differ whether it is expressed as TRb1 or as TRb2, including the ability of the mutant receptors to recruit corepressors and/or coactivators, and these isoform-specific differences in coregulator recruitment have in particular been implicated as contributing to the tissue-specificity that is characteristic of the PRTH phenotype (although it remains unresolved which coregulators are of the greatest relevance in this regard) (Lee et al., 2011; Machado et al., 2009; Privalsky et al., 2009; Wan et al., 2005). Given its many known pleiotropic effects on receptor function, we viewed (and selected for analysis) the R429Q genetic lesion as a particularly good candidate for an RTH mutation that might also affect target gene repertoire. It was therefore unexpected that, when tested here, the R429Q-TRb1 mutant largely recapitulated the WT-TRb1 pattern of gene specificity, leading us to now believe that the R429Q-TRb1 mutant mediates many aspects of its disease phenotype as a consequence of its defect in hormone-sensing rather than due to substantial changes in target gene specificity. Nonetheless it remains possible that the target gene specificities of the R429Q-TR mutant differ from WT if expressed in other cells types or as the TRb2 isoform (e.g. (Adams et al., 1994; Chatonnet et al., 2011; Lee et al., 2011; Machado et al., 2009; Safer et al., 1997; Togashi et al., 2005; Wan et al., 2005)). Furthermore, other RTH-mutations may alter target gene specificity even though the R429Q mutation does not; notably a very strong dominantnegative TR mutant (“PV”) has been reported to confer a variety of gene-specific regulatory effects in mouse adipose, heart, or brain that differ from those in WT animals (liver was not examined) (Miller et al., 2004). Finally we emphasize that our use of a single, highly saturating concentration of T3 in these experiments (a concentration likely to produce thyrotoxicity if maintained in vivo) tests the target gene specificities of the WT and R429Q TRb1 receptors only at an extreme point of their hormone response and does not fully recapitulate the lower levels and more nuanced changes in T3 availability that occur in different tissues and at different times in normal physiology. We ultimately hope to elucidate the target-specific gene response for several different mutant TRs over a range of cell types, time courses, T3 concentrations, and when expressed as TRb1 or as TRb2; however such studies will require experimental resources not currently available to us. It has been reported that introduction of the R429Q mutation into the TRb locus (THRB) in mice specifically disrupts receptormediated repression of a number of target genes in a variety of tissues, including liver (Machado et al., 2009). In contrast, our data presented here indicates that the R429Q-TRb mutant retains the ability to repress target gene expression in at least several contexts. Our studies were performed in cultured human HepG2 cells rather than in an in vivo mouse model and our survey of target genes was broad but not exhaustive; it remains possible that the R429Q-TRb1 fails to repress a specific subset of hepatic genes that were not among those achieving statistical significance in our current study. Of interest: little or no impact of the R429Q mutation on TRmediated gene activation was noted in either our or the Machado et al. study. 4.5. Conclusions Among the genes surveyed by our experimental system we observed an extensive overlap between the target gene repertoire

95

of the R429Q RTH-TRb1 mutant and that of WT-TRb1. This overlap was seen both in the absence and in the presence of added saturating T3 and extended to both negative and positive gene regulation. We suggest therefore that the defects of this mutant in a clinical setting may be, at least in part, a consequence of its overall impaired ability to bind and respond to T3 at physiological hormone concentrations rather than target gene-specific defects, at least in hepatocytes. Further, although recognizing the experimental limitations inherent in our utilization of an ex vivo cultured cell line and single T3 concentration employing one time point, our results are consistent with a model in which the nature of the corresponding target gene repertoire helps account for why the R429Q-RTH mutation is associated with endocrine disease whereas otherwise other dominant-negative TR mutations are associated with HCC and RCCC neoplasia. Acknowledgements This work was funded in part by NIDDK52528. RJ was also supported by a PHS predoctoral training award from the National Institutes of Health, T32-GM007377. We sincerely thank Liming Liu and Elsie L. Campbell for superb technical assistance. We are also grateful to the skills and advice of the University of California Davis Expression Analysis Core and Bioinformatics Core. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.mce.2017.02.044. References Abel, E.D., Ahima, R.S., Boers, M.E., Elmquist, J.K., Wondisford, F.E., 2001. Critical role for thyroid hormone receptor beta2 in the regulation of paraventricular thyrotropin-releasing hormone neurons. J. Clin. Invest. 107, 1017e1023. Adams, M., Matthews, C., Collingwood, T.N., Tone, Y., Beck-Peccoz, P., Chatterjee, K.K., 1994. Genetic analysis of 29 kindreds with generalized and pituitary resistance to thyroid hormone. Identification of thirteen novel mutations in the thyroid hormone receptor beta gene. J. Clin. Invest. 94, 506e515. Anders, S., Pyl, P.T., Huber, W., 2015. HTSeqea Python framework to work with highthroughput sequencing data. Bioinformatics 31, 166e169. Astapova, I., Hollenberg, A.N., 2013. The in vivo role of nuclear receptor corepressors in thyroid hormone action. Biochim. Biophys. Acta 1830, 3876e3881. Ayers, S., Switnicki, M.P., Angajala, A., Lammel, J., Arumanayagam, A.S., Webb, P., 2014. Genome-wide binding patterns of thyroid hormone receptor beta. PLoS One 9, e81186. Blankenberg, D., Von Kuster, G., Coraor, N., Ananda, G., Lazarus, R., Mangan, M., Nekrutenko, A., Taylor, J., 2010. Galaxy: a web-based genome analysis tool for experimentalists. Curr. Protoc. Mol. Biol. 10, 1e21 (Chapter 19), Unit 19. Brent, G.A., 2012. Mechanisms of thyroid hormone action. J. Clin. Invest. 122, 3035e3043. Brent, G.A., 2000. Tissue-specific actions of thyroid hormone: insights from animal models. Rev. Endocr. Metab. Disord. 1, 27e33. Chan, I.H., Privalsky, M.L., 2009a. Isoform-specific transcriptional activity of overlapping target genes that respond to thyroid hormone receptors alpha1 and beta1. Mol. Endocrinol. 23, 1758e1775. Chan, I.H., Privalsky, M.L., 2009b. Thyroid hormone receptor mutants implicated in human hepatocellular carcinoma display an altered target gene repertoire. Oncogene 28, 4162e4174. Chan, I.H., Privalsky, M.L., 2006. Thyroid hormone receptors mutated in liver cancer function as distorted antimorphs. Oncogene 25, 3576e3588. Chatonnet, F., Flamant, F., Morte, B., 2014. A temporary compendium of thyroid hormone target genes in brain. Biochim. Biophys. Acta 1849, 122e129. Chatonnet, F., Picou, F., Fauquier, T., Flamant, F., 2011. Thyroid hormone action in cerebellum and cerebral cortex development. J. Thyroid. Res. 2011, 145762. Chatterjee, V.K., 2008. Nuclear receptors and human disease: resistance to thyroid hormone and lipodystrophic insulin resistance. Ann. Endocrinol. Paris. 69, 103e106. Cheng, S.Y., 2005a. Isoform-dependent actions of thyroid hormone nuclear receptors: lessons from knockin mutant mice. Steroids 70, 450e454. Cheng, S.Y., 2005b. Thyroid hormone receptor mutations and disease: beyond thyroid hormone resistance. Trends Endocrinol. Metab. 16, 176e182. Cheng, S.Y., Leonard, J.L., Davis, P.J., 2010. Molecular aspects of thyroid hormone actions. Endocr. Rev. 31, 139e170. Clifton-Bligh, R.J., de Zegher, F., Wagner, R.L., Collingwood, T.N., Francois, I., Van

96

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97

Helvoirt, M., Fletterick, R.J., Chatterjee, V.K., 1998. A novel TR beta mutation (R383H) in resistance to thyroid hormone syndrome predominantly impairs corepressor release and negative transcriptional regulation. Mol. Endocrinol. 12, 609e621. Collingwood, T.N., Adams, M., Tone, Y., Chatterjee, V.K., 1994. Spectrum of transcriptional, dimerization, and dominant negative properties of twenty different mutant thyroid hormone beta-receptors in thyroid hormone resistance syndrome. Mol. Endocrinol. 8, 1262e1277. Dumitrescu, A.M., Refetoff, S., 2013. The syndromes of reduced sensitivity to thyroid hormone. Biochim. Biophys. Acta 1830, 3987e4003. Feige, J.N., Auwerx, J., 2007. Transcriptional coregulators in the control of energy homeostasis. Trends Cell Biol. 17, 292e301. Feng, X., Jiang, Y., Meltzer, P., Yen, P.M., 2000. Thyroid hormone regulation of hepatic genes in vivo detected by complementary DNA microarray. Mol. Endocrinol. 14, 947e955. Flamant, F., Baxter, J.D., Forrest, D., Refetoff, S., Samuels, H., Scanlan, T.S., Vennstrom, B., Samarut, J., 2006. International Union of Pharmacology. LIX. The pharmacology and classification of the nuclear receptor superfamily: thyroid hormone receptors. Pharmacol. Rev. 58, 705e711. Forrest, D., Vennstrom, B., 2000. Functions of thyroid hormone receptors in mice. Thyroid 10, 41e52. Furumoto, H., Ying, H., Chandramouli, G.V., Zhao, L., Walker, R.L., Meltzer, P.S., Willingham, M.C., Cheng, S.Y., 2005. An unliganded thyroid hormone beta receptor activates the cyclin D1/cyclin-dependent kinase/retinoblastoma/E2F pathway and induces pituitary tumorigenesis. Mol. Cell Biol. 25, 124e135. Gloss, B., Giannocco, G., Swanson, E.A., Moriscot, A.S., Chiellini, G., Scanlan, T., Baxter, J.D., Dillmann, W.H., 2005. Different configurations of specific thyroid hormone response elements mediate opposite effects of thyroid hormone and GC-1 on gene expression. Endocrinology 146, 4926e4933. Goecks, J., Nekrutenko, A., Taylor, J., Galaxy, T., 2010. Galaxy: a comprehensive approach for supporting accessible, reproducible, and transparent computational research in the life sciences. Genome Biol. 11, R86. Hayashi, Y., Weiss, R.E., Sarne, D.H., Yen, P.M., Sunthornthepvarakul, T., Marcocci, C., Chin, W.W., Refetoff, S., 1995. Do clinical manifestations of resistance to thyroid hormone correlate with the functional alteration of the corresponding mutant thyroid hormone-beta receptors? J. Clin. Endocrinol. Metab. 80, 3246e3256. Helmer, E.B., Raaka, B.M., Samuels, H.H., 1996. Hormone-dependent and -independent transcriptional activation by thyroid hormone receptors are mediated by different mechanisms. Endocrinology 137, 390e399. Jameson, J.L., 1994. Mechanisms by which thyroid hormone receptor mutations cause clinical syndromes of resistance to thyroid hormone. Thyroid 4, 485e492. Kamiya, Y., Puzianowska-Kuznicka, M., McPhie, P., Nauman, J., Cheng, S.Y., Nauman, A., 2002. Expression of mutant thyroid hormone nuclear receptors is associated with human renal clear cell carcinoma. Carcinogenesis 23, 25e33. Kim, S.W., Ho, S.C., Hong, S.J., Kim, K.M., So, E.C., Christoffolete, M., Harney, J.W., 2005. A novel mechanism of thyroid hormone-dependent negative regulation by thyroid hormone receptor, nuclear receptor corepressor (NCoR), and GAGAbinding factor on the rat cD44 promoter. J. Biol. Chem. 280, 14545e14555. Kong, A.P.S., Lam, C.W., A.O.K., C, Yiu, S.F., Tiu, S.C., 2005. Resistance to thyroid hormone in a Chinese family with R429Q mutation in the thyrid hormone receptor beta gene. Hong Kong Med. J. 11, 125e129. Larsen, P.R., Davies, T.F., Schlumberger, M.-J., Hay, I.D., 2003. Thyroid physiology and diagnostic evaluation of patients with thyroid disorders. In: Larsen, P.R., et al. (Eds.), Williams Textbook of Endocrinology. Saunders, Philadelphia, pp. 331e373. Lazar, M.A., 1993. Thyroid hormone receptors: multiple forms, multiple possibilities. Endocr. Rev. 14, 184e193. Lee, S.H., Young, B.M., Wan, W., Chan, I.H., Privalsky, M.L., 2011. A mechanism for pituitary-resistance to thyroid hormone (PRTH) syndrome: a loss in cooperative coactivator contacts by thyroid hormone teceptor (TR){beta}2. Mol. Endocrinol. 25, 1111e1125. Lin, J.Z., Sieglaff, D.H., Yuan, C., Su, J., Arumanayagam, A.S., Firouzbakht, S., Cantu Pompa, J.J., Reynolds, F.D., Zhou, X., Cvoro, A., Webb, P., 2013. Gene specific actions of thyroid hormone receptor subtypes. PLoS One 8, e52407. Lin, K.H., Wu, Y.H., Chen, S.L., 2001. Impaired interaction of mutant thyroid hormone receptors associated with human hepatocellular carcinoma with transcriptional coregulators. Endocrinology 142, 653e662. Luo, J., Deng, Z.L., Luo, X., Tang, N., Song, W.X., Chen, J., Sharff, K.A., Luu, H.H., Haydon, R.C., Kinzler, K.W., Vogelstein, B., He, T.C., 2007. A protocol for rapid generation of recombinant adenoviruses using the AdEasy system. Nat. Protoc. 2, 1236e1247. Machado, D.S., Sabet, A., Santiago, L.A., Sidhaye, A.R., Chiamolera, M.I., OrtigaCarvalho, T.M., Wondisford, F.E., 2009. A thyroid hormone receptor mutation that dissociates thyroid hormone regulation of gene expression in vivo. Proc. Natl. Acad. Sci. U. S. A. 106, 9441e9446. Matsushita, A., Misawa, H., Andoh, S., Natsume, H., Nishiyama, K., Sasaki, S., Nakamura, H., 2000. Very strong correlation between dominant negative activities of mutant thyroid hormone receptors and their binding avidity for corepressor SMRT. J. Endocrinol. 167, 493e503. Meex, S.J., Andreo, U., Sparks, J.D., Fisher, E.A., 2011. Huh-7 or HepG2 cells: which is the better model for studying human apolipoprotein-B100 assembly and secretion? J. Lipid Res. 52, 152e158. Mi, H., Muruganujan, A., Casagrande, J.T., Thomas, P.D., 2013. Large-scale gene function analysis with the PANTHER classification system. Nat. Protoc. 8, 1551e1566.

Miller, L.D., McPhie, P., Suzuki, H., Kato, Y., Liu, E.T., Cheng, S.Y., 2004. Multi-tissue gene-expression analysis in a mouse model of thyroid hormone resistance. Genome Biol. 5, R31. Moehren, U., Eckey, M., Baniahmad, A., 2004. Gene repression by nuclear hormone receptors. Essays Biochem. 40, 89e104. Mullur, R., Liu, Y.Y., Brent, G.A., 2014. Thyroid hormone regulation of metabolism. Physiol. Rev. 94, 355e382. Nakano, K., Matsushita, A., Sasaki, S., Misawa, H., Nishiyama, K., Kashiwabara, Y., Nakamura, H., 2004. Thyroid-hormone-dependent negative regulation of thyrotropin beta gene by thyroid hormone receptors: study with a new experimental system using CV1 cells. Biochem. J. 378, 549e557. Nikolayeva, O., Robinson, M.D., 2014. edgeR for differential RNA-seq and ChIP-seq analysis: an application to stem cell biology. Methods Mol. Biol. 1150, 45e79. Olateju, T.O., Vanderpump, M.P., 2006. Thyroid hormone resistance. Ann. Clin. Biochem. 43, 431e440. Ortiga-Carvalho, T.M., Shibusawa, N., Nikrodhanond, A., Oliveira, K.J., Machado, D.S., Liao, X.H., Cohen, R.N., Refetoff, S., Wondisford, F.E., 2005. Negative regulation by thyroid hormone receptor requires an intact coactivator-binding surface. J. Clin. Invest. 115, 2517e2523. Privalsky, M.L., 2004. The role of corepressors in transcriptional regulation by nuclear hormone receptors. Annu. Rev. Physiol. 66, 315e360. Privalsky, M.L., 2008. Thyroid hormone receptors, coregulators, and disease. In: Kumar, R., O”Malley, B.W. (Eds.), NR Coregulators and Human Diseases. World Scientific Publishing, Singapore, pp. 243e280. LTD. Privalsky, M.L., Lee, S., Hahm, J.B., Young, B.M., Fong, R.N., Chan, I.H., 2009. The p160 coactivator PAS-B motif stabilizes nuclear receptor binding and contributes to isoform-specific regulation by thyroid hormone receptors. J. Biol. Chem. 284, 19554e19563. Raghavachari, N., Barb, J., Yang, Y., Liu, P., Woodhouse, K., Levy, D., O'Donnell, C.J., Munson, P.J., Kato, G.J., 2012. A systematic comparison and evaluation of high density exon arrays and RNA-seq technology used to unravel the peripheral blood transcriptome of sickle cell disease. BMC Med. Genomics 5, 28. Refetoff, S., Dumitrescu, A.M., 2007. Syndromes of reduced sensitivity to thyroid hormone: genetic defects in hormone receptors, cell transporters and deiodination. Best. Pract. Res. Clin. Endocrinol. Metab. 21, 277e305. Rosen, M.D., Chan, I.H., Privalsky, M.L., 2011. Mutant thyroid hormone receptors (TRs) isolated from distinct cancer types display distinct target gene specificities: a unique regulatory repertoire associated with two renal clear cell carcinomas. Mol. Endocrinol. 25, 1311e1325. Safer, J.D., Cohen, R.N., Hollenberg, A.N., Wondisford, F.E., 1998. Defective release of corepressor by hinge mutants of the thyroid hormone receptor found in patients with resistance to thyroid hormone. J. Biol. Chem. 273, 30175e30182. Safer, J.D., Langlois, M.F., Cohen, R., Monden, T., John-Hope, D., Madura, J., Hollenberg, A.N., Wondisford, F.E., 1997. Isoform variable action among thyroid hormone receptor mutants provides insight into pituitary resistance to thyroid hormone. Mol. Endocrinol. 11, 16e26. Sharif, M., Privalsky, M.L., 1992. V-erb A and c-erb A proteins enhance transcriptional activation by c-jun. Oncogene 7, 953e960. Stashi, E., York, B., O'Malley, B.W., 2014. Steroid receptor coactivators: servants and masters for control of systems metabolism. Trends Endocrinol. Metab. 25, 337e347. Togashi, M., Nguyen, P., Fletterick, R., Baxter, J.D., Webb, P., 2005. Rearrangements in thyroid hormone receptor charge clusters that stabilize bound 3,5',5-triiodo-Lthyronine and inhibit homodimer formation. J. Biol. Chem. 280, 25665e25673. Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg, S.L., Rinn, J.L., Pachter, L., 2012. Differential gene and transcript expression analysis of RNA-seq experiments with tophat and cufflinks. Nat. Protoc. 7, 562e578. Vatner, D.F., Weismann, D., Beddow, S.A., Kumashiro, N., Erion, D.M., Liao, X.H., Grover, G.J., Webb, P., Phillips, K.J., Weiss, R.E., Bogan, J.S., Baxter, J., Shulman, G.I., Samuel, V.T., 2013. Thyroid hormone receptor-beta agonists prevent hepatic steatosis in fat-fed rats but impair insulin sensitivity via discrete pathways. Am. J. Physiol. Endocrinol. Metab. 305, E89eE100. Viguerie, N., Langin, D., 2003. Effect of thyroid hormone on gene expression. Curr. Opin. Clin. Nutr. Metab. Care 6, 377e381. Wan, W., Farboud, B., Privalsky, M.L., 2005. Pituitary resistance to thyroid hormone syndrome is associated with T3 receptor mutants that selectively impair beta2 isoform function. Mol. Endocrinol. 19, 1529e1542. Wang, D., Xia, X., Liu, Y., Oetting, A., Walker, R.L., Zhu, Y., Meltzer, P., Cole, P.A., Shi, Y.B., Yen, P.M., 2009. Negative regulation of TSHalpha target gene by thyroid hormone involves histone acetylation and corepressor complex dissociation. Mol. Endocrinol. 23, 600e609. Wang, D., Xia, X., Weiss, R.E., Refetoff, S., Yen, P.M., 2010. Distinct and histonespecific modifications mediate positive versus negative transcriptional regulation of TSHalpha promoter. PLoS One 5, e9853. Webb, P., 2010. Thyroid hormone receptor and lipid regulation. Curr. Opin. Investig. Drugs 11, 1135e1142. Weiss, R.E., Tunca, H., Knapple, W.L., Faas, F.H., Refetoff, S., 1997. Phenotype differences of resistance to thyroid hormone in two unrelated families with an identical mutation in the thyroid hormone receptor beta gene (R320C). Thyroid 7, 35e38. Weitzel, J.M., Hamann, S., Jauk, M., Lacey, M., Filbry, A., Radtke, C., Iwen, K.A., Kutz, S., Harneit, A., Lizardi, P.M., Seitz, H.J., 2003. Hepatic gene expression patterns in thyroid hormone-treated hypothyroid rats. J. Mol. Endocrinol. 31, 291e303.

R. Jimenez, M.L. Privalsky / Molecular and Cellular Endocrinology 447 (2017) 87e97 Wondisford, F.E., 2003. Thyroid hormone action: insight from transgenic mouse models. J. Investig. Med. 51, 215e220. Yen, P.M., 2015. Classical nuclear hormone receptor activity as a mediator of complex biological responses: a look at health and disease. Best. Pract. Res. Clin. Endocrinol. Metab. 29, 517e528. Yen, P.M., 2003. Molecular basis of resistance to thyroid hormone. Trends Endocrinol. Metab. 14, 327e333. Yen, P.M., Feng, X., Flamant, F., Chen, Y., Walker, R.L., Weiss, R.E., Chassande, O.,

97

Samarut, J., Refetoff, S., Meltzer, P.S., 2003. Effects of ligand and thyroid hormone receptor isoforms on hepatic gene expression profiles of thyroid hormone receptor knockout mice. EMBO Rep. 4, 581e587. Yoh, S.M., Chatterjee, V.K., Privalsky, M.L., 1997. Thyroid hormone resistance syndrome manifests as an aberrant interaction between mutant T3 receptors and transcriptional corepressors. Mol. Endocrinol. 11, 470e480. Zhang, J., Lazar, M.A., 2000. The mechanism of action of thyroid hormones,. Annu. Rev. Physiology 62, 439e466.