The human THRB gene spans a region of 376609 bp and is divided into 20 different exons inlcuding 16 separated by large introns which may give rise to 19 various TRβ1 transcripts mostly differing in 5 untranslated region (5UTR), 1 truncated variant TRβ4 (TRβ1 isoform) and 1 TRβ2 transcript with differential promoter usage (see diagram 2 and 3). Note that several sense and antisense non-TRβ transcripts were shown to be expressed in the THRB locus, in which may affect the expression of the thyroid hormone receptors (see diagram 2). The THRB gene is one of two thyroid hormone receptor genes in the human genome located on chromosome 3. The other gene, THRA located on chromosome 17, encodes the related thyroid hormone receptors TRα1, TRα2 and TRΔα1 (p28 and p43). These genes were initially identified by their homology to the avian retroviral oncogene v-erbA encoding a mutated variant of chicken TRα1.

According to the NCBI Gene, the THRB maps to NC_000003.11 (24158644..24536772, complement) and spans a region over 376 kilo bases. Within the sequence, some other transcriptionally active genes have been identified (see diagram 1 and 2). Two precursor mRNA (pre-mRNA) are transcribed using two THRB promoters: P1 (GeneBank: S37458.1), updated according to the sequence of chromosome 3 genomic contig (GenBank: NT_022517.18) allowing for expression of TRβ1 and P2 promoter of TRβ2 isoform (see diagram 3), located between exon 4 and 5 (GeneBank: NG_009159.1; 330058..334363), however the sequence range of P2 has not been exactly established. The THRB gene consists of 20 different exons including 16 separated by large introns. The exons are described in details above (see diagram 3).
The alternative TRβ1 splicing results in expression of multiple mRNAs encoding the same protein, however its translation is differentially regulated by the 5UTR variants. These transcripts consist of 10 - 14 different exons forming a 5UTR region (exons 1, 2 and the beginning of exon 3), protein coding region (the end of exon 3, 4-9, first 242nt of exon 10) and long 3UTR (last part of exon 10) (Frankton et al., 2004). The exons 5-10 are common to the most of transcripts including TRβ2 mRNA that contains one more, first exon named "a" (see diagram 2 and 3). The variant TRβ4 contains additional 137-bp insertion (exon) between exons 7^th and 8^th shown in diagram 3 (intron 5^th according to Tagami et al., 2010) that results in synthesis of a truncated protein lacking the ligand binding domain. This isoform may modulate T3 action as an endogenous antagonist (according to Tagami et al., 2010). The multiple 5UTR splice variants of the TRβ1 can differently regulate translation of the TRβ1 coding sequence. The 5UTRs vary in length, GC-content, secondary structures, number of upstream open reading frames (uORFs) and internal ribosome entry site (IRES) predicted in the variant A (GeneBank: AY286465.1). These regulatory sequences may be organized in secondary and tertiary RNA structures that are recognized by trans-acting factors such as protein translation factors and naturally occurring small RNAs. Moreover, the most of TRβ mRNA variants contain long 3UTR (see diagram 3) with multiple microRNA binding sites that may affect the expression of these receptors. Disturbances in the expression of the TRβ mRNA variants have been reported in various cancers. Some of the disturbances seem to be a cancer specific. For example, the loss of transcript variants F1 (GQ456950) and IVS4B (GQ919288.1) has been observed in clear cell renal cell cancer (ccRCC). In the same tissues, over 70% reduction in TRβ1 mRNA (coding sequence) has been reported. Simultaneously, expression of 5UTR variants A and F (AY286470.1) has been reduced in ccRCC by 75% and 62%, respectively, compared to control samples (Master et al., 2010).
Note also that several sense and antisense non-TRβ transcripts have been shown to be expressed in the THRB locus, in which may affect the expression of the thyroid hormone receptors (see diagram 2 and 3).

Two precursor mRNAs (transcribed from two different promoters of the THRB) undergo extensive co- and post-transcriptional modification in the nucleus that includes tissue specific, alternative splicing of the pre-mRNAs.

No pseudogene has been reported for the THRB gene. Nevertheless, the THRB may be regulated by pseudogenes identified within the THRB gene sequence or in the same locus (see diagram 1 and 2). The pseudogenes may produce long naturally occurring antisense transcripts (cis-NATs) forming sense-antisense pairs (see Homology). These sense-antisense pairs may activate numerous mechanisms, similar to those observed in a pseudogene-mediated regulation of a target gene via pseudogene-derived small interfering RNAs or on the level of RNA-directed DNA methylation, pre-mRNA transcription, alternative splicing as well as RNA editing, transport and localized protein translation. However, this regulation remains to be established for the THRB gene.

The human THRB encodes three protein isoforms, the TRβ1 and TRβ2 that are T3-dependent receptors mediating genomic and nongenomic actions of the thyroid hormone, and TRβ4 isoform, which is a carboxyl-terminal splicing variant of TRβ1 that lacks the ligand binding domain and thus, may modulate T3 action as an endogenous antagonist in the tissue or cellular context (Tagami et al., 2011). The TRβ proteins are implicated in regulation of transcription of target genes and control key cellular processes including differentiation, proliferation, apoptosis and metabolism.

There are three human TRβ isoforms TRβ1, TRβ2 and TRβ4, which are differentially expressed in various tissues. The TRβ1 and TRβ2 receptors have the typical domain structure of a nuclear receptor with an N-terminal domain (A/B), central DNA binding domain (C) consisting of a double C4-type zinc fingers, hinge region (D), C-terminal ligand binding domain (E), and AF2 domain (F). These receptors have a variable N-terminal domain (A/B) that differs between TRβ1 and TRβ1 isoforms (see diagram 4). The N-terminus is responsible for cofactor and regulatory protein binding, T3-independent transactivation, receptor dimerization and DNA recognition. Local dimerization sites have been found in domain C, E, corepressor interaction sites in domain D, E and coactivator interaction sites in domain A/B, E and F. The hinge region (D) is also found to be required for TRβ dependent suppression of ras-mediated responses.
The TRβ1 can bind to target DNA sites regardless of T3 binding status as a homodimer or heterodimer with retinoid X receptors (RXR) or as a monomer (more weakly). TRβ can bind to DNA in the absence of ligand and therefore is thought to have the potential to mediate both T3-dependent and T3-independent regulation of target gene transcription. The protein phosphorylation has been shown to enhance its cytoplasmic-nuclear import (Maruvada et al., 2003). Phosphorylated T3 receptors can exhibit increased TRE binding as a homodimer, but not as heterodimer or monomer. Moreover, integrin-mediated non-genomic action of T4 (Davis et al., 2009) may result in phosphorylation of TRβ1 Ser142 leading to dissociation of the corepressors and transactivation of target genes (Davis et al., 2000). In case of typical positively regulated genes such as DIO1 encoding iodothyronine deiodinase type I and GH1 of human growth hormone 1, T3 binding stimulates a conformational changes in the TRβ allowing for dissociation of corepressors followed by recruitment of coactivators to form an activating complex stimulating the transcription. There is also identified a group of negatively regulated genes that includes hypothalamic thyrotropin-releasing hormone (TRH) and pituitary thyroid stimulating hormone (TSH) encoded by TSHB and controlling the hypothalamic-pituitary thyroid (HPT) axis. In this regulation liganded nuclear receptors down-regulate target gene transcription, with the cooperative binding of various transcription factors to multiple regulatory elements on DNA (for more see genomic actions of T3 mediated by TRβ receptors).
The TRβ4 is a C-terminal splicing variant of TRβ1 which contains a stop codon in an 137-bp insertion (exon) insertion that results in synthesis of a truncated protein lacking the ligand binding domain (see diagram 4). The TRβ4 contains full A/B, C, D domain of TRβ1 and a fragment of domain E (is identical for the first 246 amino acids) but the carboxyl-terminal 215 amino acids of TRβ1 are replaced by an entirely distinct sequence of 13 residues, which results from an insertion (see RNA). Therefore, this truncated protein is unable to bind thyroid hormone, thus may modulate T3 action as an endogenous antagonist in the tissue or cellular context. The TRβ4 cannot mediate T3-dependent gene regulation but may inhibit the negative regulation of TSH mediated by TRβ1 or TRβ2, that was shown in TSA-201 cells, a clone of human embryonic kidney 293 cells (according to Tagami et al., 2010 and Tagami et al., 2011). These findings are consistent with current model for T3-dependent negative regulation of TSHB gene (see genomic actions below).
The levels of TRβ proteins depend on the protein stability and transcription/translation efficiency that is tissue specific and differentially regulated in various mRNA variants. The frequently reported lack of correlation between the mRNA and protein suggests that apart from transcriptional control the expression of TRβ receptors is accurately controlled at the level of translation. In fact, multiple 5UTR variants of TRβ1 have been shown in vitro to differentially regulate the protein translation. The major renal TRβ1 transcript contains a 5UTR relevant to variant A (GeneBank: AY286465.1). Analysis of the effects of 5UTR variants on protein expression in JEG-3 choriocarcinoma cells (Francton et al., 2004) and Caki-2 renal cancer cells (Master et al., 2010) indicated that the weakly folded variant A permitted also the highest level of protein expression. In contrast, the strongly folded 5UTR variants: F (AY286470.1) or F1 (GQ456950) was identified to be transcribed and translated at the lowest levels. Although, structured 5UTR such as variant F treated in vitro with a trans-acting factor (an antisense oligonucleotide) significantly up-regulated the translation efficiency of a downstream sequence up to the level of the variant A. This may estimate the potential of the 5UTR-mediated translational control during expression of TRβ receptors. The translation of the TRβ can be also modulated by multiple regulatory ORFs that exist upstream of the primary ORF. On the other side, the protein synthesis may be controlled by long TRβ 3UTR through binding of various microRNAs including miR-21 and miR-146a. The microRNA can trigger the RNA interference (RNAi) phenomenon that may lead to translational repression or even degradation of TRβ mRNAs (Jazdzewski et al., 2011). Since UTR-mediated translation initiation is a key rate-limiting phase affecting efficiency of the protein synthesis, the translational control mediated by the multiple UTR variants is emerging to be a major regulator of the final protein levels in cells. The TRβ receptors have been also reported to be affected by aberrant promoter methylation, alternative splicing and impaired cell signaling.

The thyroid hormone receptor isoforms are products of both the THRB and THRA genes. During development, TRβ and TRα isoforms are differentially expressed in a temporospatial and tissue-specific patterns and in adult tissues are present in distinct ratios (Williams, 2000; Francton et al., 2004).
TRβ1 is widely expressed in all human tissues, but is prominent in brain, thyroid, kidney and liver. TRβ1 controls liver and kidney metabolism and mediates cholesterol lowering effects of thyroid hormones. Both the TRβ1 and TRβ2 are essential for regulation of thyroid hormone levels through the hypothalamic-pituitary-thyroid (HPT) axis, a negative feedback loop, which includes auto-control of hypothalamic Thyrotropin-Releasing Hormone (TRH) and pituitary Thyroid-Stimulating Hormone (TSH) by the thyroid hormones (TH). In this regulation, the TRβ receptors mediate TRH- and TSH- lowering effects of TH.
TRβ2 is restricted to the hypothalamus, anterior pituitary, developing brain, cochlea (inner ear) and retina, wherein TRβ2 alone is crucial for development of mid-wavelength (MW) cones photoreceptors, which play a significant role in circadian clock light entrainment and in phase shifting of the circadian oscillator (Dkhissi-Benyahya et al., 2007). This isoform is therefore important for visual and auditory function. Down regulation of hypothalamic TRH is TRβ2 specific. TRβ2 mRNA has been also identified in situ in human chondrocytes and osteoclasts (Abu et al., 2000).
TRβ4 isoform is expressed in various human tissues but is highly abundant in testis and skeletal muscle. This isoform lacks T3 -binding domain and may act as a dominant-negative protein. It has been identified in a TSH-secreting pituitary adenoma (TSHoma) as well (Tagami et al., 2011).
In contrast, TRα mediates T3 actions during development of heart, bone, intestinal and is responsible for body temperature and basal heart rate in adults. TRα1 and TRα2 isoforms are highly expressed in the brain, with lower abundance in the kidneys, skeletal muscle, lungs, heart, and testes. The TRα1, TRβ1, and TRβ2 isoforms can bind DNA and T3 acting as functional thyroid hormone receptors, whereas TRα2 and TRα3 do not bind T3 due to the presence of the longer AF2 (F) domain, and act as antagonists. There are some truncated isomers of TRα with specific mitochondrial functions (p28, p43) or may act as dominant-negative receptors (TRΔα1, TRΔα2). There are also known two additional TRβ receptors expressing only in rats: functional T3 receptor - TRβ3 and its dominant negative isoform -TRΔβ3 (Williams et al., 2000).
Both, TRβ and TRα isoforms are involved in circadian cycle that was demonstrated in an in vivo mouse model, in different metabolic tissues including white adipose tissue (WAT), brown adipose tissue (BAT), liver, and skeletal muscle (Yang et al., 2007). While TH levels are generally constant, the TRβ and TRα along with their key target genes dramatically cycle in a coordinated manner that is in agreement with known cyclic behaviour of lipid and glucose metabolism. TRβ has been shown to cycle in WAT, whereas TRα in WAT, BAT and liver. Analysis of TRβ mRNA expression revealed a unique rhythmic pattern in which their transcripts spike at ZT4 (Zeitgeber time, 4h after lights were turned on) followed by a precipitous decline in the next 4h, and remain at low levels through the rest of the day. Interestingly, mRNA levels of heme-liganded nuclear receptor Rev-erbβ encoded by NR1D2 gene, which may interact with THRB expression (see diagram 1 and 2), dramatically cycle in the all examined metabolic tissues. Although precisely how the circadian clock acts to control metabolic rhythms is not clear, it is known that the ratio of TRα1/TRα2 isoforms is closely determined by co-expression of Rev-erbα (NR1D1) - a key component of the circadian core oscillator complex. The Rev-erbα represses the expression of BMAL changing the expression of other genes of the cycle including CLOCK, PER1, PER2, PER3 and cryptochrome genes CRY1 and CRY2 responsible for blue light photoreceptor activity (see specific functions of TRβ2). The cell -autonomous feedback loop allows for cyclic expression of these oscillator genes at various phases with the same period length of approximately 24r (Zhang et al., 2009). All this findings suggest a role for TRβ, TRα and heme receptors (Rev-erbβ, α) in coupling the peripheral circadian rhythm to divergent metabolic outputs. Since the expression of TRβ is prominent in brain and metabolic tissues, it is expected that these receptors may serve peripheral clock input pathways that can integrate signals from the light-sensing central clock in the suprachiasmatic nuclei (SCN) and other cues including xenobiotic metabolism pathways. Interplay between the circadian clock and TH receptors may be a part of a large -scale signaling network that links biological timing to metabolic physiology (Yang et al., 2006).

Subcellular localization and changes in expression of TRβ receptors can vary depending on the cell cycle phase, cell density, cellular stress, signaling events, tissue types, metabolic rate or even circadian cycle (Maruvada et al., 2003).
Typically, TRβ1 and TRβ2 isoforms are predominantly localized to the nucleus and retained in the nucleus regardless of the ligand binding status (in the absence and presence of T3). However, T3 can induce a nuclear reorganization of TRβ receptors. The nuclear localization is essential for TRβ-mediated genomic actions of T3. In standard conditions, a minor fraction of TRβ proteins resides in the cytoplasm, wherein the receptors are thought to be mediators of nongenomic actions of thyroid hormone (TH). The cytoplasmic-nuclear shuttling is facilitated by the presence of a nuclear localization signal (NLS) in the TRβ hinge region (D). The TRβ nuclear import is ATP-dependent and can be regulated by nongenomic actions of TH through 1) T4-dependent activation of plasma membrane integrin αvβ3 (Davis et al., 2009) followed by activation of downstream pathways leading to phosphorylation of ERK1/ERK2 and TRβ1 proteins and/or 2) via T3-dependent formation of cytoplasmic TRβ1 complexes with p85 subunit of PI3K that may activate downstream pathways. The TRβ complexes with p85, ERK1/2 or nuclear receptor coactivators may facilitate nuclear import as well. TRβ rapidly shuttles between the nuclear and the cytoplasmic compartments. Energy-dependent blockade (ATP depletion) enhances TRβ nuclear export to cytoplasm. Nevertheless, coexpression of nuclear corepressors (NCoRs) and/or retinoid X receptors (RXRs) can markedly decrease the shuttling by maintaining unliganded TRβ within the nucleus. A TRβ mutant defective in DNA binding has a slightly altered nuclear-cytoplasmic distribution when compared with wild-type TRβ. TRβ mutants that abrogate its interaction with the NCoRs accumulates within the cytoplasm due to an increase in the rate of nuclear export when compared with nuclear import. Nuclear-cytoplasmic shuttling has been proposed as a mechanism for modulating TRβ-mediated regulation of transcription (Maruvada et al., 2003).
Subcellular localization of TRβ4 is unknown; however the lack of T3 binding domain might suggest an altered nuclear/cytoplasmic distribution.

TRβ proteins are high affinity receptors for thyroid hormone (TH) functioning as ligand-dependent (T3) and sequence-specific DNA binding (TRE) transcription factors (see genomic actions below) that regulate expression of target genes affecting cell growth, development, proliferation, differentiation, apoptosis, organ morphogenesis, heart rate, body fat distribution, bone density. These receptors are required for the development of the auditory system and of the cone photoreceptors that mediate colour visual function. TRβ1 isoform is expressed in most tissues, whereas TRβ2 is restricted to the hypothalamus, pituitary, cochlea, and retina that may indicate the functional specificity of the isoforms. The TRβ1 control the major responses of the liver and kidney to T3 and play a critical role in mediating changes in metabolism and thermogenesis. The T3 receptors are able to increase metabolic rate by accelerating fuel oxidation in most of tissues wherein they may activate lipolysis, glucose metabolism and protein synthesis. TH receptors are responsible for T3-dependent homeostasis and maintenance of a steady body temperature that is realized via an auto-regulatory negative feedback loop controlling the hypothalamic-pituitary thyroid (HPT) axis. Its regulation is mainly determined by local T3 concentration in the anterior pituitary as well as paraventricular nucleus (PVN) of the hypothalamus wherein low levels of T3 abolish TRβ-mediated transcriptional repression of pituitary thyroid-stimulating hormone (TSH) promoter and hypothalamic thyrotropin-releasing hormone (TRH) promoter. However, the local tissue T3-levels are dependent on the availibility of circulating thyroid prohormone - thyroxine (T4), its intracellular transport and cytoplasmic T3 production, catalyzed by iodothyronine deiodinases. Since TRs can form heteroduplexes with retinoid X receptors, TH is found to modulate the skin response to retinoids. Genomic effects of TRβ1 have been recognized as tumor suppressive and disturbances of the TRβ1 expression have been found in different cancers (Martínez-Iglesias et al., 2009b; Kim at al., 2013).
Description
TRβ is involved in many processes that compose thyroid hormone (TH) actions and gene expression. This T3-dependent receptor act as a transcription factor regulating expression of genes involved in the cell cycle progression, differentiation, apoptosis and cellular metabolic rate. TH exerts a pleiotropic effect on development and homeostasis. This effect results from genomic and nongenomic TH actions mediated by both, the TRβ and TRα receptors (TRs) that regulate hundreds of genes responding to T3-liganded or unliganded receptors. A large number of genes have been identified to respond to T3 stimulation (~10% of all expressed genes) and the divergence between T3-treated and untreated cells can grow rapidly over time. The initial studies of TRs actions revealed near complete overlaps in their effects (TRα and TRβ can regulate similar gene sets). However, gene-specific differential TRα and TRβ actions are reported as well. These differences appear to result from the 1) differences in tissue-specific expression patterns, 2) diurnal rhythm (see expression and specific functions), 3) various time-courses of actions with different kinetics 4) target gene-specific variations in pattern of response to T3 concentration. For instance, TRβ has been shown to exhibit gene-specific requirements for higher T3 levels (compared to TRα) for regulation of HR, MYH6, ALPI and FURIN genes, whereas HIF1A is identified to have lower T3 requirements during TRβ-mediated regulation. ANGPTL4 encoding a PPARG angiopoietin related protein is a verified in parental HepG2 cells direct TRβ target. Prolonged T3 treatment selectively augments TRβ action in the context of the TRβ-dependent genes. Moreover, several T4- and T3- analogues (see ligands) have been also reported to induce TRβ-specific response.
Genomic and nongenomic actions
Genomic actions of TRβ are initiated with nuclear translocation of the newly synthesized receptors (Maruvada et al., 2003) that is facilitated by nuclear localization signal (NLS) found in the hinge region (D) of the receptors (see diagram 2 and 3). The TRβ proteins are classified as type II nuclear receptors, which are retained in the nucleus regardless of the ligand binding status (free or occupied by T3) and bind to TREs (thyroid hormone response elements) as heterodimers with retinoid X receptors (RXR), rarely as homodimers or monomers. The TREs are specific DNA sequences of the consensus core recognition motif AGGTCA, AGGACA or (A/G)GGT(C/A/G)A hexamers (half-site →), in which two or more motifs are positioned as direct repeats separated by 4, 0 or 6-nucleotide: (TRE-DR4, →n4→), palindromes (TRE-P0, →←) or inverted palindromes (TRE-IP6, →n6←). The analysis of the response elements formed by direct repeats of the half-sites demonstrated that a spacer of 4-nucleotides can provide maximal transactivation by TRβ in TRβ-RXR heterodimers but the transactivation efficiency may depends on sequence context of the TRE, tissue specific trans-acting factors and ligand concentration. It has been also shown that for an efficient genomic action of the heterodimers, the presence of TRβ and RXR ligands: triiodothyronine (T3) and 9-cis retinoic acid (9-cis RA) may be necessary. However synergistic effects of the RXR ligand and T3 on the heterodimers (RXR/TR)-mediated transcription have been reported for specific promoters, the other studies suggest that RXR ligands may inhibit T3-dependent transactivation, possibly by promoting the formation of RXR homodimers. The biological effects of TRE binding by the unoccupied versus the T3-occupied receptor are quite different. In many cases, binding of the unliganded or antagonist-liganded receptor alone to DNA may lead to repression of transcription, whereas binding of the agonist (T3)- liganded receptor complex activates transcription. However, the receptor activity is mainly regulated by ligand-dependent interactions with corepressor (CoR) and coactivator (CoA) proteins. In the absence of ligand (T3), the TRβ receptors are often complexed with corepressor proteins (CoRs, NCoR, SMRT) and histone deacetylases (HDACs) allowing the histones to wrap the DNA more tightly. T3 binding to these nuclear receptors causes conformational changes leading to dissociation of corepressors and recruitment of coactivator proteins (CoAs, SRC3, CBP, p300, pCAF), as well as mediators (TRAP-DRIP multi-subunit complex) that are thought to target the entire complex to a liganded receptor through a single subunit, TRAP220. The thyroid hormone (TH) receptors, similarly as vitamin D3 receptors (VDRs), and peroxisome proliferator-activated receptors (PPAR), exhibit a strong activation function 2 (AF2)-dependent preference for receptor binding domain 2 (RBD-2) of the TRAP220 protein. It has been also demonstrated that RXR receptor (in TRβ/RXR heterodimer) displays a weak yet specific AF2-dependent preference for another TRAP220 RBD-1 domain. Addition of ligand for the RXR receptor (9-cis RA), in addition to T3 for TRβ partner, might further strengthen the RXR-RBD-1 interaction and presumably stabilize the overall association of TRAP220 with the heterodimer. After formation of the coactivator multi-subunit complex, general transcription factors (TAFs, TFIIA, TBP) and RNA polymerase II are recruited to the complex that initiates transcription DNA into pre-RNA. The pre-mRNA undergoes further modifications including alternative splicing, mRNA editiing and translation into protein that may result in a change in cell function. The nuclear receptors including TRβ proteins can regulate gene expression by binding directly or indirectly (via other proteins) to specific sequences in the promoters of target genes (Puzianowska-Kuznicka, 2013). Apart from the current model of HRE/TRE - dependent transcriptional control, it has been proposed that one transcription factor may repress the activity of a second transcription factor through a protein-protein interaction, without the requirement of two different DNA binding sites. These proteins, notably AP1 and NF-kB, can act by interfering with transcriptional complex formation in a DNA-independent manner and may lead to transrepression of target gene transcription, what has been also shown in TRβ-mediated negative regulation of the hypothalamic-pituitary thyroid (HPT) axis. TH responsive genes can be both positively and negatively regulated by T3-liganded TRβ receptors. Iodothyronine deiodinase type I (DIO1) or GH1 of human growth hormone 1 are up-regulated in the presence of T3 (see positive regulation model - diagram 5). In contrast, pituitary TSH encoded by TSHB gene is down-regulated by T3 and rises during T3 deprivation. In this regulation, liganded nuclear receptors down-regulate target gene transcription, with the cooperative binding of various transcription factors to multiple regulatory elements on DNA. It has been proposed that the negative regulation of the TSHB gene may require at least two response elements on promoter DNA (nTRE-"negative" TRE and GATA-RE) as well as dimerization of liganded TRβ with a GATA transcription factor (e.g. GATA2), that can repress the activity of the GATA. In fact, the nTRE in the promoter of TSHB gene contains a single half site-like sequence (GGGTCA). The GATA2 alone can activate TSHB promoter, and this activation was repressed by Zn-finger region of liganded TRβ. Upon T3 binding, the T3/TRβ complex is thought to be released from nTRE, favoring its translocation and interaction with GATA-2 zinc finger region (GATA2-Zf) on GATA-RE located next to the nTRE. The T3/TRβ/GATA2-Zf/DNA complex formation is the principal complex responsible for the TSHB gene down regulation. This interaction occurs via the highly conserved zinc-finger in DNA binding domains that may resemble the trans-repression-like mechanism in which direct binding of the receptor to another transcription factor occurs in a DNA-independent manner. In the absence of T3 TRβ prefers nTRE element, allowing for the gene transcription. This negative regulation shows that T3 weakens TRβ binding to the regulatory element (nTRE) in the TSHB promoter. The displacement of the liganded TRβ from nTRE to another site (GATA-RE/GATA2-Zf) seems to be critical for the mechanism used by the cell for silencing the TSHB gene transcription (Figueira et al., 2010). Down-regulation of hypothalamic thyrotropin-releasing hormone TRH gene by T3/TRβ complexes involves acetylation and methylation of specific residues of histone tails in the TRH promoter region and relies on changing amount of the TRβ receptors on TRH nTRE after T3 binding (Umezawa et al., 2009). Prolonged administration of T3 causes demethylation of specific histones and subsequently the release of TRβ receptors from the gene to suppress it. Both negative regulations may require T3/TRβ to be removed from its nTRE to down-regulate the genes, although the mechanism used by a ligand-bound TRβ leading to repression of transcription is still a subject of contention. The classical mechanism of TRβ activity suggests that receptors typically complexed with RXR bind more strongly to DR4 in a clearly cooperative binding between the transcription factors and DNA (see diagram 5). Nevertheless, coexpression of RXR and TRβ2 has been also shown to slightly reduce both the transactivation and transrepression by liganded TRβ2 that may act more strongly as homodimer or monomer, depending on the architecture of the TRE. Thus, the mechanism responsible for TRβ-mediated genomic action of T3 is still unclear. Several recent studies indicate that at the genomic level TRβ may act as a tumor suppressor. For instance, the gain-of-function approach by stably expressing TRβ in a human breast cancer cell line MCF-7 (MCF-7-TRβ), which normally lacks the TR expression, has been reported to inhibit the growth of the MCF-7 cell tumors in xenograft models (Park et al., 2013). These estrogen (E2) dependent cells show elevated JAK2-STAT3-cyclin D signal that is repressed by the TRβ expression at the level of transcription. Interestingly, other studies indicate that TH has anti-apoptotic and pro-proliferative effects in ERα-positive human breast cancer cells (Perri et al., 2013). However, these opposite effects are recognized as a result of the αvβ3 integrin -dependent nongenomic actions of TH (see below).
Nongenomically initiated, TRβ -mediated thyroid hormone (TH) actions occur at the plasma membrane or in cytoplasm and may also culminate in complex, nucleus-mediated cellular events leading to the transcription of specific genes such as HIF1A, MCL1 or GLUT1 (Moeller et al., 2006; Davis et al., 2009). The nongenomic actions of TH can interfere with genomic effects of T3, hence, a clear distinction between nongenomic and genomic effects may no longer be practical. However, initiation events leading to these effects are quite different and may have different kinetics or timing. For instance, enhancement of the antiviral activity of interferon-γ by TH is achieved nongenomically and genomically by T4-dependent activation of the mitogen activated protein kinase /extracellular signal-regulated kinase 1/2 (MAPK/ ERK1/2) which can phosphorylate serine (Ser142) located on DNA binding domain of TRβ1 - a nuclear receptor of T3. This phosphorylation leads to dissociation of TRβ1 from corepressor proteins, the NCoR (nuclear receptor co-repressor) and SMRT (silencing mediator of retinoid and TH receptors) allowing for transactivation of T3-specific target genes. TETRAC and TRIAC (see ligands), which can inhibit nongenomic actions of TH, can also block the T4 potentiation of the antiviral and immunomodulatory actions of the interferon-γ, even though these analogues have no direct effect on the interferon-γ action. There is increasing evidence that nongenomic and genomic actions of TH may overlap with genomic and nongenomic effects of estrogens and testosterone in tumor cells. The thyroid hormone, estrogene and dihydrotestosterone have similar ERK1/2-dependent proliferative actions on the estrogen receptor α (ERα)-positive human breast cancer cells. TH and steroids also have interacting nongenomic and genomic actions in heart and brain cells. The binding affinity of T4 to its plasma-membrane receptor modulates intracellular protein trafficking of the ERα and TRβ1 receptors from the cytoplasm to nucleus. T4 -transduced activation of the ERK1/2 promotes its nuclear uptake and ERK1/2-dependent phosphorylation of the TRβ1, ERα and signal transducers and activators of transcription 3 and 1α (STAT1α, STAT3). In the nucleus of T4-treated cells, the TRβ1, ERα, STAT1α and STAT3 transcription factors were found to be co-immunoprecipitated with the activated ERK1/2. The complexing of TRβ1 and ERK1/2 was relatively rapid and detected with 1.4- and 7.8-fold increases in TRβ1 in 30 and 40 min of T4 treatment, respectively. This effect of T4 was observed as early as in 10 min and persisted for up to 90 min. The increase in nuclear TRβ1 was assumed to be originated from the pool of cytosolic fraction of the TRβ1. Interestingly, the nongenomic action of T4 has been shown to block the p53-mediated proapoptotic activity of resveratrol, a polyphenolic compound found in grapes and wine (SIRT1 activator), by disrupting an ERK1/2-nucleoprotein complex. Although, inhibition of T4 binding at the cell surface receptor can restore the apoptotic action of the resveratrol (Lin et al., 2002).
The action of T4 on cellular signal transduction is initiated at a cell-surface by activation of an integrin receptor - αvβ3 (identified in 2005). The TH plasma-membrane receptor was previously reported as a putative G protein-coupled receptor (GPCR) that preferentially binds T4 and 3,5,3,5-tetraiodothyroacetic acid (TETRAC), an antagonist of the nongenomic actions of TH (Lin et al., 1998; Lin et al.,1999). Extracellular binding of T4 to its transmembrane receptor was demonstrated to activate the signal transduction cascade which included G-proteins, PLC, PKC, Ras, Raf-1, MAPK kinase (MEK), the MAPK (ERK1/2) and downstream pathways (Davis et al., 2000).
The current model of nongenomic actions of TH is based on transduction of the hormone signal through the membrane by the integrin αvβ3, which can preferentially bind T3 to S1 and T4 to S2 sites of the integrin (Davis et al., 2009). The signal transduction may result in pro-proliferative and pro-angiogenic effects achieved in a ligand -dependent manner via the αvβ3/ERK1/2 cascade or by the hormone activated phosphatidylinositol 3-kinase (PI3K), respectively. The αvβ3 integrin is concentrated largely in plasma membranes of endothelial cells, vascular smooth muscle cells, various cancer cells, osteoclasts and platelets. Hormone-binding domain of the integrin receptors includes two recognition sites that are capable of binding T3 (S1) or T4 and T3 (S2). S2 site can bind both T4 and T3, though the affinity for T3 binding to the S2 site is lower than that for T4. Binding T4 to integrin αvβ3 (without cell entry) may mediate nongenomic actions of TH by activation of extracellular-regulated kinases 1/2 (ERK1/2), which transduces the hormone signal into complex cellular and nuclear events including angiogenesis and tumor cell proliferation. This T4-induced pathway can stimulate shuttling of TRβ1 receptor from the cytoplasm to the nucleus and increase the TRβ-mediated transactivation of specific genes. Moreover, phosphorylation of serine 142 located on DNA binding domain (DBD) of the cytoplasmic/nuclear TRβ1 is thought to facilitate transcription derepression by dissociation of corepressors (NCoR and SMRT) and recruiting coactivators and mediators (SRC-1, CBP, p300, pCAF, TRAP-DRIP). The nongenomic action via integrin receptors may be specifically inhibited by TETRAC, a deaminated T4 analogue that can displace T4 and T3 from both sites (S1, S2) but does not mimic the agonist functions of the hormones through the integrin receptors. This T4-analogue blocks thyroid hormone effects on angiogenesis and cancer cell proliferation and might have some benefits in cancer treatment. In the chick chorioallantoic membrane (CAM) model, the cells treated with TETRAC - an agonist for TRs (in genomic action), significantly enhance the expression of thrombospondin, an antiangiogenic gene. This effect has been shown to complement the anti-VEGF and anti-bFGF actions of TETRAC. In contrast, protein ligands present in extracellular matrix (ECM) and containing an Arg-Gly-Asp (RGD) motif fully inhibits T3 actions initiated at S1 site (PI3K pathway) and does not affect T3 actions initiated at S2 site and ERK1/2-dependent cell proliferation. The S1 site can bind T3 exclusively and rapidly transduces the hormone signal via PI3K leading to cytoplasm-nucleus shuttling of TRα1 and expression of HIF1A gene encoding the hypoxia-inducible factor-1α, alpha subunit (HIF-1α) (Davis et al., 2009).
The similar activation of the HIF-1α transcription may be also initiated intracellularly by interaction of T3 with cytoplasmic fraction of TRβ1 (see diagram 5).The liganded TRβ1 mediates action of T3 on expression of specific genes, including pro-angiogenic genes through binding to the regulatory subunit p85α of PI3K (see diagram 5) followed by activation of downstream cascade of protein kinase Akt/PKB, mammalian target of rapamycin (mTOR) and its substrate - p70S6K that is involved in cellular protein synthesis (Kenessey and Ojamaa, 2006; Moeller et al., 2006). Activation of mTOR is rapid, with detectable phosphorylation within the minutes after T3 treatment. This pathway may lead to up-regulation of specific genes including: HIF1A and HIF-1α target genes; SLC2A1 (GLUT1) - solute carrier family 2 (facilitated glucose transporter), member 1; PFKP - phosphofructokinase, platelet; SLC16A4 (MCT4) - solute carrier familymonocarboxylate transporter 4; RCAN2 regulator of calcineurin 2 (ZAKI-4). Rapamycin - an inhibitor of mTOR, abrogates the thyroid hormone-dependent induction of the ZAKI-4α, suggesting the role of sequential activation of the kinases in the PI3K pathway. Treatment of endothelial cells with T3 leads to activation of the Akt/PKB and endothelial nitric oxide synthase (eNOS) that is abolished by PI3-kinase inhibitors, but not by the inhibitors of transcription. The PI3K-mediated effects are important to angiogenesis and other recently appreciated cell functions but not to tumor cell division. The T3 liganded TRβ1 in cytoplasm may also nongenomically initiate a process that result in transcription of the ATP1A1 (ATPase, Na+/K+-ATPase) and insertion of the gene product into the plasma membrane (Davis et al., 2009).
Regardless of TRβ1, also TRα1 can interact with the p85α subunit of PI3K in a T3 -dependent manner, leading to phosphorylation of Akt and activation of downstream signaling pathways. The other nongenomic actions of the TH have been also shown to modulate cellular ion fluxes, sodium current (I(Na)), inward rectifying potassium current (IKir), sodium pump (Na, K-ATPase) and ERK1/2-regulated Na/H exchanger (NHE) encoded by SLC9C1 - solute carrier family 9 subfamily C (Na+-transporting carboxylic acid decarboxylase), member 1. Both, the genomic and nongenomic actions of TH have been shown to proceed the transcription and activity of the sarcoplasmic reticulum Ca(2+)-ATPase (calcium pump). T4 and rT3 but not T3 may act through a truncated form of TRα1 (TRΔα1) located in cytoplasm, wherein the liganded TRΔα1 may take part in conversion of soluble actin to fibrous (F) actin that is important to cell motility. Certain of these actions appear to interfere with genomic function of the TRβ receptors. The nongenomic effects of TRβ ligands occur rapidly and are unaffected by inhibitors of transcription or translation processes.
THs exert important physiological actions by both genomic and nongenomic effects in mitochondria. T3 and T2 - a thyroid hormone metabolite, regulate mitochondrial genome transcription and nongenomically initiated mitochondrial processes such as cellular respiration and thermogenesis. T2 metabolite binds and activates the mitochondrial cytochrome-c-oxidase Va, whereas T3 binds to two truncated TRα1 isoform (p28 and p43). Whereas the role of p28 remains unknown, p43 protein is a T3-dependent transcription factor of the mitochondrial genome, acting via dimeric complexes involving two other truncated forms of nuclear receptors: mtRXR and mtPPAR. All these mitochondrial actions as well as expression of other nuclear TH receptors (e.g. TRα1, TRα2) may have an impact on thyroid hormone availability and the TRβ function in the cell.
Concluding this section, the nongenomic actions of TH that are mediated by TRβ is fast (10-40min), frequently reported to result in pro-proliferative, pro-angiogenic, anti-apoptotic effects and may be simplified into two main signaling cascades: 1) extracellular-T4/ αvβ3-integrin/ PLC/ PKCα/ ERK1/2/ TRβ1-Ser142 phosphorylation that among others can result in specific gene transactivation or transrepression (see diagram 5N4); 2) cytoplasmic-T3/ TRβ1/ CSH2-p85α-p110(PI3K)/ Akt/ mTOR phosphorylation leading to transcription of PI3K-specific genes such as HIF1A and GLUT1 (diagram 5N1, 5N1b, 5N1c). Specific inhibitors of these pathways are shown on diagram 5.
Interaction
TRβ has been shown to interact with: BRD8, CCND1, NCOA1, NCOA6, NCOR2, NR2F6, PPARGC1A and RXRA.
Crosstalk signaling with proteins of nuclear hormone receptor superfamily.
Apart from TRβ proteins, the other T3 regulated nuclear receptors may affect the thyroid hormone levels and the TRβ function in cells. T3 and T4 exert a pleiotropic effect on cellular homeostasis and are mediated by protein products of both the THRB and THRA genes. The THRA encodes TRα1, TRα2, TRα3 and some truncated variants TRΔα1, TRΔα2, p28 and p43. The TRα1 can display both a nuclear and cytoplasmic location, and is the only thyroid hormone receptor that is imported into the mitochondrial matrix as p28 and p43 truncated variants. The thyroid hormone receptor beta gene may produce TRβ1, TRβ2, TRβ4 (in humans) and two additional isoforms expressed in rats: TRβ3 and TRΔβ3. The TRα1, TRβ1, TRβ2 as well as TRβ3 can bind T3 and mediate T3-dependent actions, thus, the proteins are bona fide receptors, whereas the TRα2, TRα3 or TRΔα1 TRΔα2, TRΔβ3 and TRβ4 do not bind the hormone and their function remains to be elucidated. The TRα2 and TRα3 have longer carboxy-terminal domains (AF2) that does not bind T3 and weakly binds DNA, thus, the variants act as dominant negative antagonists of T3 signalling. The truncated variant TRΔβ3 lacks the DNA-binding domain but retains T3 binding activity and acts as a dominant-negative antagonist. The human variant TRβ4 that is a C-terminal spliced variant of TRΒ1 lacks T3 -binding ability and acts as an endogenous dominant-negative isoform. The TRβ4 weakly but significantly inhibits transcription mediated by functional T3 receptors.
The function of the TRβ proteins may be influenced by the other members of the receptor superfamily that may lead to either synergistic or antagonistic effects. The nuclear receptor superfamily includes the estrogen receptor-like subfamily liganded by estrogene (ER) or 3-ketosteroids: glucocorticoid (GR), progesterone (PR), androgen (AR), mineralocortycoid (MR) and the thyroid hormone receptor-like subfamily that consists of the nuclear receptors for the thyroid hormone (TRα, TRβ), liver X receptor-like proteins (LXR, FXR), vitamin D receptor-like proteins (VDR, PXR, CAR), retinoic acid receptors (RARs) and peroxisome proliferator-activated receptors (PPARs). This subfamily also include the heme receptors: Rev-ErbAα encoded by NR1D1 gene that regulates various cellular function including circadian cycle and the ratio of TRα1/TRα2 isoforms as well as Rev-erbβ (NR1D2) identified in the THRB locus (3p24.2, see diagram 2). These heme binding receptors were identified previously as "orphan" (unknown ligand and/or DNA target) receptors. The members of the thyroid hormone receptor-like subfamily are classified as type II nuclear receptors, which are retained in the nucleus and can bind to DNA regardless of the ligand binding status and usually form heterodimers with Retinoid X Receptor-like subfamily transcription factors, which includes the retinoid X receptor (RXR), hepatocyte nuclear factor-4 (HNF4), testicular receptor (TR2, TR4) and photoreceptor cell-specific nuclear receptor (PNR). The class II receptors include the members of the estrogen receptor-like subfamily as well.
Interactions in the nuclear superfamily could be illustrated on the basis of crosstalk signaling between TRs and PPARs. This interaction appears to be important in TRs -mediated adipogenesis and carcinogenesis. Majority of the effects of PPARs and TRs were found to be opposing during the crosstalk, however the cooperative effects were reported as well (Lu and Cheng, 2009). PPARs liganded by prostaglandins, prostacyclins or a conjugated linoleic acids (CLAs) and T3 liganded TRs can reciprocally affect the target gene expression. In humans, the DBDs of TRs and PPARs are highly homologous and can bind to the same half-site sequence that is present in both, TRE and PPRE elements. The canonical DR1 PPRE consists of two direct repeats of AGGTCA with a 1 bp spacer, whereas the DR4 TRE is built by the same direct repeats, separated with a 4 bp spacer. TRs can competitively bind to PPREs of PPAR target genes. Moreover, either TRs or PPARs compete with the other receptor for binding to RXR that may result in decreased availability of RXR and reduced transcriptional activity of the PPAR target genes. Consequently, the PPARγ agonist rosiglitazone is able to reverse the effects of dexamethasone and to increase serum T3 and T4 levels. Interestingly, in rats with a high-fat diet and in a hyperthyroid state, administration of the PPARα agonist Wy14,643 restores glucose tolerance by enhancing glucose-stimulated insulin secretion and relieves the effect of hyperthyroidism. These data suggest that PPARα activity may restore the pancreatic islet function affected by abnormal T3/TR signaling. Furthermore, the genomic crosstalk of these two receptors may occur also via nongenomic actions of the receptors (Lu and Cheng 2009).
Ligands and metabolites
L-thyroxine (T4), a major secretory product of thyroid gland, and 3, 5, 3-triiodo-L-thyronine (T3), the most active form of thyroid hormone are naturally occurring ligands for TRβ receptors. T3 is a tyrosine-based derivative of T4 that is produced by the thyroid gland in response to thyroid-stimulating hormone (TSH) from the anterior pituitary. The TSH is released by thyrotropin-releasing hormone (TRH) from the paraventricular nucleus (PVN) of the hypothalamus and T4 synthesis is controlled by negative feedback loop: hypothalamic-pituitary thyroid (HPT) axis. Free thyroid hormones in the circulation act negatively on the pituitary and hypothalamus, thus reducing the release of TRH, TSH and finally T4 and T3 concentration in plasma. In the nucleus, the thyroid hormones bind TRβ with high affinity and specificity with Kd values in the nM range. Most T4 and T3 in the circulation are bound to proteins including Thyroxine-binding globulin (TBG), transthyretin and albumin. This proteins are responsible for carrying the thyroid hormones in the bloodstream.
Under normal conditions only a small fraction of T3 is generated by the thyroid gland, the remainder of T3, which is available for binding sites in the plasma and body cells, is synthesized by mono-deiodination of T4, that is inactive and needs to be converted to T3 what occurs in peripheral tissues. The reaction is catalyzed by type 1 (DIO1, EC1.97.1.10) or type 2 (DIO2, EC 1.97.1.11) iodothyronine deiodinases (selenoproteins), the first is abundant in kidney, liver and thyroid whereas the last one is mainly present in brown adipose tissue, pituitary and central nervous system. DIO1 is sensitive to inhibition by the anti-thyroid drug propylthiouracil (PTU). The enzyme activity of the kidney and liver is responsive to the nutritional status of an organism and is found to be more active during states of accelerated glucose metabolism. Most T3 molecules are produced by enzymatic outer ring deiodination (ORD) of T4. Alternative, inner ring deiodination (IRD) of T4 yields the metabolite rT3. ORD is regarded as an activating pathway and IRD as an inactivating pathway. DIO1 shows the ORD and IRD activity, DIO2 only ORD activity and the third iodothyronine deiodinase - DIO3 (expressed above all in brain tissue) mediates only the degradation of thyroid hormone since it has only IRD activity. T3 and rT3 undergo further deiodination to the common metabolite 3,3-diiodothyronine (3,3T2), which is generated by IRD of T3 and by ORD of rT3. Recent evidence for binding of T2 by a subunit of mitochondrial cytochrome c oxidase and its stimulation appears to be of a receptor/effector nature, showing as well that the T2 metabolite may have an important biological role that influences cellular respiration.
Apart from T4, T3, rT3 and 3,3T2, the other thyroid hormone derivatives have been shown to take part in iodothyronine-like endogenous signaling, which may includes T2 (3,5-diiodo-L-thyronine), TAMs (thyronamines), and sulfate or glucuronic acid derivatives of thyroid hormones.
The most commonly studied TRs agonists are: TRIAC (3,5,3-triiodothyroacetic acid showing thyromimetic activity, TRβ selective), TETRAC (tetraiodothyroacetic acid, an inhibitor of T4-transduced, αvβ3-integrin-mediated pathway), DITPA (3,5-diiodothyropropionic acid), DIMIT (3,5-dimethyl-3-isopropylthyronine), GC-1 (3,5-dimethyl-4-(4-hydroxy-3-isopropylbenzyl)-phenoxy acetic acid (sobetirome), a TRβ-selective), KB-141 (3,5-dichloro-4-[(4-hydroxy-3-isopropylphenoxy) phenyl] acetic acid), SKF 94901(3,5-dibromo-3-pyridazinone-L-thyronine). TETRAC expresses nucleus-mediated thyromimetic activity, though current studies demonstrated its role in suppression of T4-induced αvβ3-integrin/ERK1/2 pathway (Davis et al., 2009). Covalently bonded to a nanoparticle, TETRAC acts exclusively at the cell surface TH αvβ3 integrin receptor without the thyromimetic activity and TR-mediated effects in nucleus.
TRs agonist relative potencies: GC-1~T3~TRIAC~T4>>rT3 have been determined for both coactivator recruitment and corepressor dissociation.
The TRβ and liver selective thyromimetics (STRMs) such as GC-1 (sobetirome), KB2115 (eproterome), KB141 and MB07811 have been developed to selectively lower serum total and LDL cholesterol without affecting HDL cholesterol levels and induce weight loss without deleterious effects on heart and combat other aspects of metabolic disease. GC-1 is the first TRβ agonist. KB141 is a selective TRβ agonist that can bind with an affinity 14-fold higher than that for TRα and is a useful candidate for attenuating the features of metabolic syndrome. This agonist lowers cholesterol, causes significant weight reduction in primates and has a 10-fold window in which therapeutic increases in metabolic rate are seen without cardiac hypertrophy or tachycardia therefore is a promising candidate for treating obesity, hyperlipidemia and diabetes. KB2115 has been described as a TRβ-selective agonist that is preferentially taken up by the liver and is also effective in patients on statin therapy. MB07811 is a liver-selective pro-drug that after hepatic activation through enzymatic cleavage is converted to the active form MB07344 - a strong TRβ ligand binding the other TRs with significantly lower affinity. Selective targeting with TRβ agonists may present an innovative strategy for searching for TRβ-specific effects.
The most commonly studied antagonists are: DIBRT (3,5-dibromo-4-(35-diisopropyl-4-hydroxyphenoxy)benzoic acid) and NH-3 (4-(4-hydroxy-3-isopropyl-5-(4-nitrophenylethynyl) benzyl)-3,5-dimethylphenoxy) acetic acid (NH-3, a low affinity TR antagonist). The TRβ-selective (but not TRα-selective) agonists and antagonists are expected to alter gene expression in a TRβ biased manner that can differ from T3, which binds the two TRs with similar affinity. Amiodarone (Amio) and dronedarone (Dron) are drugs used to discriminate between TRα1 or TRβ1 regulated genes in central and peripheral TH metabolism. A major metabolite of Amio, desethylamiodarone, acts as a TRα1 and TRβ1 antagonist, whereas the major metabolite of Dron -debutyldronedarone acts as a selective TRα1 antagonist that allows the TRβ1 effects to become apparent.

Human THRB gene is reported to be conserved in Euteleostomi and has orthologs identified in: Pan troglodytes (Chr3, NCBI protein reference sequence: XP_001163850.1), Macaca mulatta (Chr2, XP_001090554.1), Canis lupus (Chr23, XP_862690.2), Bos taurus (Chr27, XP_002698801.1), Mus musculus (Chr14 7.08 cM, NP_001106888.1), Rattus norvegicus (Chr15 p16, NP_036804.2), Danio rerio (Chr19, NP_571415.1) and Gallus gallus (Chr2, erythroblastic leukemia viral (v-erb-a) oncogene homolog 2, avian, NP_001239150.1); according to the NCBI HomoloGene. Since the THRB shows multiple alignment, pairwise similarity scores and evolutionary distances, additional putative orthologs are likely in a variety of different species and can be viewed using alignment tools (BLAST, NCBI Resources).
This gene is highly homologous to ERBB2 (v-erb-b2 erythroblastic leukemia viral oncogene homolog 2, neuro/glioblastoma derived oncogene homolog (avian)). A connection between TRβ1 and cancer became evident when the chicken TRα1 ortholog was characterized as the c-erbA proto-oncogene, the cellular counterpart of the retroviral v-erbA oncogene. However, a growing number of evidences suggest that TRβ could serve as a tumor suppressor.
THRB is a member of a large superfamily of nuclear hormone receptor genes (see interactions) and displays extensive structural and functional similarity with the paralogous gene - THRA, which evolved from a common ancestor gene, duplicated 500 Mya. The gene is duplicated at least in vertebrates, birds and amphibians. Divergent evolution of the genes increased the variety of cellular responses to TH in vertebrates. The paralogous THRB and THRA show 77% of mRNA coding sequence (CDS) identity and 84% of protein identity. The similarities were determined based on the NCBI BLAST protein seq.: TRβ1 (NP_000452.2) ver. TRα1 (NP_955366.1); results: identities: 310/367(84%); positives: 342/367(93%); gaps: 0/367(0%); method: Compositional matrix adjust. NCBI BLAST mRNA CDS seq.: TRβ1 (NM_000461.4) ver. TRα1 (NM_199334.3); results: identities: 854/1113(77%); score 814 bits(902); gaps: 16/1113(1%). Comparative DNA analysis of the THRB and THRA revealed 10 conserved fragments of at least 261bp in length (73-90% of DNA identity between the genes), located in THRB introns: downstream of exon 1 (4x); 2a (2x), 2b, 3 (2x) and 4. These Alu-like elements were found in 5-end of THRA sequence as well as in NR1D1 and NR1D2 genes. Moreover, two full-length of ADAR-binding sequence motif flanking A-to-I RNA editing sites (Ramaswami et al., 2012) are present in introns before exon 2a and 2c of the THRB. This consensus sequence is also present in THRA and NR1D1 genes.
The other reported paralogs for THRB gene are: VDR, NR4A3, NR4A1, NR1I2, RARG, NR1I3, RARA, NR1H2, NR4A2, NR1H4, RARB, NR1H3.
Interestingly, the THRB has close structural and functional relatives in NR1D2 (in the same locus), which encodes Rev-erbβ as well as THRA encoding TRα1 and TRα2 isoforms and NR1D1 gene expressing Rev-erAα. The THRB and NR1D2 genes are also linked to the RARB encoding RARβ. Since it has been shown that the THRA/NR1D1 locus is also linked to the RARA gene, these data suggest that the two receptor gene clusters RARA/THRA/NR1D1 and RARB/THRB/NR1D2 were generated by a single large-scale duplication. Moreover, the THRA gene shares a partial overlap with the NR1D1 gene that influences the TRα1/TRα2 ratio. Similar regulation may occur in case of the TRβ transcripts and the products of the genes sharing the same DNA sequence with the THRB (see diagram 1 and 2). The genes located in the same locus may produce long naturally occurring antisense transcripts (cis-NATs) forming sense-antisense pairs with a single stranded DNA or transcripts of the THRB. These sense-antisense pairs may activate a pseudogene-mediated regulation (see Pseudogene).
Reference conserved domains on TRβ1 protein (Homo sapiens, NP_001121649.1): 1) cd06961: NR_DBD_TR superfamily, DNA-binding domain of thyroid hormone receptors, Heterodimer interface of human Thyroid Hormone; 2) cd06935: NR_LBD_TR superfamily, The ligand binding domain of thyroid hormone receptor, coactivator recognition site (polypeptide binding site), dimer interface (polypeptide binding site).
Reference conserved domains on TRβ2 (Homo sapiens): 1) cd06929: NR_LBD_F1 (aa : 281-454), Ligand-binding domain of nuclear receptor family 1; 2) cd06916 : NR_DBD_like (aa : 122- 194) DNA-binding domain of nuclear receptors is composed of two C4-type zinc fingers.
Animal models
Animal models imply close associations between aberrant expression of TRβ or TRβ mutants and pathogenesis of some diseases, such as dominant or reccessice Generalized Resistance to Thyroid Hormone (GRTH) and Follicular Thyroid Carcinoma (FTC). Creation of a mouse model that harbors a knockin mutation of TRβ has facilitated the study of the molecular actions of TRβ mutants in vivo. Knock-out studies in mice suggest that the different TH receptors may mediate different functions, though several studies of TRs actions revealed near complete overlaps in their effects. Tissue-specific expression of TRβ isoforms is thought to be a major factor responsible for the observed differences in phenotypes, including those that are affected by TRβ mutations. In mice knock-outs, the TRβ abnormalities may affect the following systems: endocrine/exocrine glands (increased thyrotroph cell number, enlarged thyroid gland, abnormal pituitary gland physiology); hearing/vestibular/ear (abnormal cochlea morphology, increased or absent threshold for auditory brainstem response, sensorineural hearing loss); homeostasis/metabolism (increased circulating thyroid-stimulating hormone level); nervous system (increased thyrotroph cell number, abnormal pituitary gland physiology, decreased cochlear outer hair cell number, abnormal retinal cone cell morphology); skeleton (spiral ligament degeneration), vision/eye (abnormal retinal cone cell morphology).
THRB gene, like mice Thrb ortholog, encodes TH receptor isoforms TRβ1 TRβ2 and TRβ4. Moreover, function of these receptors may be influenced by two additional isoforms: TRβ3 and TRΔβ3, expressed in rat models (Williams et al., 2000). TRβ1, TRβ2, and TRβ3 are bona fide T3 receptors that bind DNA, T3 and regulate expression of T3-responsive target genes. Studies of Trβ and Trβ2 knockout mice indicated that Trβ1 is essential for development of auditory function, whereas Trβ2 is not required, but that Trβ2 alone is essential for development of mid-wavelength (MW) cones photoreceptors. In contrast, both Trβ1 and Trβ2 are required for regulation of hypothalamic-pituitary-thyroid axis. The Trβ2 deletion in mice induces a complete and selective loss of MW-cone opsin without significant changes in total cone numbers. TRβ3 and TRΔβ3 variants are transcribed using third promoter (P3) positioned upstream of human exon 5 and downstream of second promoter (P2) of TRβ2 (see diagram 3). TRΔβ3 mRNA lacks rat exon B (315 nt) encoding a fragment of DNA-binding domain, present in TRβ3, which contains both exon A (342 nt) and B of rat TRβ. Start codons of TRβ3 has been identified in frame in various animal Thrb sequences including mouse, dog, chicken but not in human, chimpanzee and macaque. Nevertheless, none of these ATG codons are positioned within a favorable Kozak translation initiation sequence context and the lack of murine TRβ3 or TRΔβ3 expressed sequence tags (NCBI EST) suggests that in rats, expression from P3 promoter is differentially regulated. TRβ3 is a functional T3 receptor and the most potent isoform, but dependent on the sequence context of TRE elements, whereas TRΔβ3 retains T3-binding activity but lacks a functional DNA-binding domain and does not activate target gene transcription. Therefore, this isoform acts as a modulator (potent antagonist) of TRβ3 when coexpressed at low concentrations. At higher concentrations, TRΔβ3 is a TRE-selective and cell-specific antagonist of TRα1, Trβ1, and TRβ3. Identified in humans TRβ4 is a C-terminal variant that lacks the ligand binding domain (truncated variant of TRβ1), thus may function as a potent endogenous antagonist, however its expression and function in animals is unknown.
Mice with targeted deletions in TR genes have provided understanding of the possible roles of the different TRβ isoforms. Knockout mice that are unable to produce theTRα1 receptor show subnormal body temperature and mild abnormalities in cardiac function, whereas mice which lack expression of both TRα isoforms were severely hypothyroid and die within the first few weeks of life. Mice with disruptions of the entire beta gene (TRβ1 and TRβ2) exhibit elevated TSH levels and deafness suggesting its role in auditory system, while mice with mutations disturbing only TRβ2 had elevated TSH, but normal hearing. These mutants allow determination of which functions of the different receptor isoforms are redundant and which are not. TRs play important roles in the pathogenesis of thyroid cancers and hepatocellular carcinoma (HCC). For instance, v-erbA, a mutant form of TR lacking ligand-binding ability, triggers HCC development in transgenic mice. Similarly, TRβ^PV (Kaneshige et al., 2000) mutation harmoring mice develope thyroid cancers (see exemplary mutants below).
There are various mouse knock-outs for THRB:
- TRβ^PV/PV; mutant thyroid hormone receptor kindred PV (Kaneshige et al., 2000); Synonyms: TRβPV; Allelic composition: homozygous TRβ^PV/PV and heterozygous TRβ^PV/+; Mutation details: PV has an unusual mutation in exon 10, a C-insertion at codon 448, which produces a frameshift of the carboxyl-terminal 14 amino acids of TRβ1. PV was derived from a patient (called PV) with severe RTH characterized by elevated thyroid hormone levels accompanied by normal TSH, short stature, goiter, and tachycardia. This naturally occurring mutation shows lost T3-binding, transactivation activities, and displays dominant negative activity. Moreover, PV strongly interferes with the transactivation activity of wild-type TRs in vitro and unlike the missense mutations or single amino acid deletion of TRβ found in other patients, this unique frame-shifted mutated sequence is immunogenic, for which high-affinity specific antibodies have been developed. TRβ PV mutant has been obtained by using homologous recombination and the CreyloxP System. Affected systems: endocrine/exocrine glands, homeostasis/metabolism; Human disease model: 1) Thyroid Hormone Resistance (NCBI OMIM: 274300, 188570); TRβ PV/+ and TRβ PV/PV mice faithfully reproduce human RTH - the TRβPV mutation was initially identified in a patient who has the syndrome of resistance to thyroid hormone (RTH). Mice expressing a single PV allele showed the typical abnormalities of thyroid function found in heterozygous humans (mutation of one copy of human THRB) with RTH that is characterized by a reduced sensitivity of tissues to the action of thyroid hormones. 2) Thyroid Carcinoma (OMIM: 188470), pituitary tumors; TRβ^PV/PV mice but not TRβ^PV/+, spontaneously develop follicular thyroid carcinoma and pituitary tumors with tumor progression similar to human cancer. The homozygous PV mice exhibit severe dysfunction of the pituitary-thyroid axis, abnormal bone development and impaired weight gains. This phenotype is distinct from that seen in mice with a null mutation in the TRβ gene (TRβ^-/-) demonstrating the interference of the mutant TRβ with the functions of the wild-type TRβ. TRβ^PV/PV mutants may serve also as molecular model for nongenomic actions of T3 condributing to thyroid carcinogenesis. The mutant mice allows the elucidation of oncogenic activity of the TRβ(PV) through cytoplasmic effects of T3 (see diagram 5 N1). This nongenomic action is mediated by interaction of PV with p85α regulatory subunit of PI3K to activate the downstream AKT/mTOR, p70^S6K and PI3K-integrin-linked kinase-matrix metalloproteinase-2 signaling pathways. The PV-mediated PI3K activation leads to increased cell proliferation, motility, migration, and metastasis. In this regulation, a nuclear receptors corepressor NCoR competes with PV for binding to the p85α that result in reduction of the AKT-mTOR-p70^S6K signaling. The NCoR protein levels are significantly lower in thyroid tumor cells than in wild type thyrocytes, allowing more effective binding of PV to p85α to activate the PI3K pathway, thereby contributing to tumor progression in the TRβ^PV/PV mutant mice (Guigon and Cheng, 2009).
- Thrb^tm1Df; thyroid hormone receptor beta; targeted mutation 1, Douglas Forrest ; Synonyms: Thrb-, TRbeta- Allelic composition: homozygous, Thrb^tm1Df/Thrb^tm1Df involves: 129 S1/Sv * C57BL/6J; Mutation details: Insertion of a neomycin cassette into exon 3, disrupts both the beta1 and beta2 isoforms of this gene. This transcript revealed a deletion of exon 3 sequences, and fused beta1 exon 2 to exon 4 resulting in an aberrant open reading frame, which terminates early into exon 4. No functional protein is predicted from this transcript, as the essential DNA binding and T3 binding domains not present; Affected systems: endocrine/exocrine glands, hearing/vestibular/ear, homeostasis/metabolism, nervous system but not behavior/neurological phenotype; Human disease model : Thyroid Hormone Resistance, Generalized, Autosomal Recessive - GRTH (OMIM: 274300).
- Thrb^tm3Few; thyroid hormone receptor beta; targeted mutation 3, Frederic E Wondisford; Synonyms: GS125 KI, TR-BetaGS; Allelic composition: homozygous, Thrb^tm3Few/Thrb^tm3Few, involves: 129* C57BL/6; Mutation details: Missense mutations were introduced at codons 125 and 126 (exon 3), resulting in Glu to Gly and Gly to Ser substitutions. The substitutions were within the P-box of the first zinc finger and were shown, in vitro, to abolish DNA-binding while retaining the ability to interact with T3 and cofactors. Western blot analysis showed endogenous levels of protein in homozygous mutant mice; Affected systems: endocrine/exocrine glands, hearing/vestibular/ear, homeostasis/metabolism, nervous system; vision/eye; Human disease model: Thyroid Hormone Resistance, Generalized, Autosomal Recessive, GRTH (OMIM: 274300).
- Thrb^tm2Few; thyroid hormone receptor beta; targeted mutation 2, Frederic E Wondisford; Synonyms: TRbeta^delta337T; Allelic composition: homozygous, Thrb^tm2Few/Thrb^tm2Few, involves: 129X1/SvJ * C57BL/6; Mutation details: The deletion of 3 base pairs in exon 6, corresponding to a deletion that results in thyroid horomone resistance in humans, was introduced via site-directed mutagenesis along with a neomycin selection cassette inserted into intron 5. The mutation in exon 6 affects the ligand-binding domain which is common to both isoforms produced from this locus; Affected systems: behavior/neurological, homeostasis/metabolism, nervous system; Human disease model: Thyroid Hormone Resistance, Generalized, Autosomal Dominant, GRTH (OMIM: 188570).
- Thrb^tm2Few/+; thyroid hormone receptor beta; targeted mutation 2, Frederic E Wondisford; Synonyms: TRbeta^delta337T; Allelic composition: heterozygous, Thrb^tm2Few/Thrb⁺, involves: 129X1/SvJ * C57BL/6; Mutation details: (see Thrb^tm2Few/Thrb^tm2Few); Affected systems: behavior/neurological, homeostasis/metabolism, nervous system; Human disease model: Thyroid Hormone Resistance, Generalized, Autosomal Dominant, GRTH (OMIM: 188570).
- Thrb^{tm1Df/tm1.1Syc}; thyroid hormone receptor beta; targeted mutation 1, Douglas Forrest; Synonyms: Thrb-, TRbeta-; Allelic composition: heterozygous, Thrb^tm1Df/Thrb^tm1.1Syc, involves: 129S1/Sv * 129S6/SvEvTac * C57BL/6 * C57BL/6J; Mutation details: see above; Affected systems: endocrine/exocrine glands, mortality/aging, tumorigenesis, respiratory system; Human disease model: Thyroid Carcinoma, Follicular; FTC (OMIM: 188470).
- Thrb^tm1Few; thyroid hormone receptor beta; targeted mutation 1, Fredric E Wondisford; Synonyms: TRbeta2 null; Allelic composition: homozygous mutant mice Thrb^tm1Few/Thrb^tm1Few, involves: 129S4/SvJae; Mutation details: A PGK-neo cassette replaced the transcription site, the entire Thrb2-specific exon, and the splice donor acceptor site. RT-PCR analysis of pituitary RNA confirmed the preservation of the Thrb1 isoform, as well as the absence of the Thrb2 isoform. Affected systems: endocrine/exocrine glands, homeostasis/metabolism but not hearing/vestibular/ear.
- Thrb^tm4Few; thyroid hormone receptor beta; targeted mutation 4, Frederic E Wondisford; Synonyms: TR-Beta-, TrbetaKO ; Allelic composition: homozygous, Thrb^tm4Few/Thrb^tm4Few involves: involves: 129 * C57BL/6; Mutation details: Exon 3 was replaced with a self-excising PGK-neo cassette. The deletion of exon 3 putatively results in an aberrant open reading frame caused by the fusion of exons 2 and 4. Using a C-terminal mAb, protein was undetected by Western blot analysis of homozygous mutant mice; Affected systems: all mentioned at the begining of this section.
More information on the Mouse Genome Informatics website (Mouse Genome Database (MGD)).

9555 human THRB variants (NCBI dbSNP) includind 8007 single nucleotide polymorphisms (SNPs) and 326 human variants (20 studies, NCBI dbVar) have been recorded in the NCBI databases. Moreover, 42 pathogenic variants of clinical significance including 31 germline SNP, 4 copy gain, 3 deletions and 3 insertions of the gene have been catalogued in the NCBI ClinVar database (see Table 1). These allelic variants have been identified to impair hormone binding, DNA binding or ligand-dependent conformational changes. Some of them inhibits homodimer formation or stabilizes homodimer ligand-dependent conformational changes. The majority of the mutated TRβ receptors lost their trans-activation function and exhibited dominant-negative activity.

There are 31 annotated germline THRB allelic variants in the NCBI dbSNP and ClinVar but only 15 reported in dbVar of the NCBI. Most of them are SNPs and insertions or deletions are less frequently found. The most relevant SNPs are those located on hormone binding domain (see diagram 6 and table 1). Some of them inhibits homodimer formation or stabilizes homodimer ligand-dependent conformational changes. The majority of the mutated TRβ receptors lost their function of transactivation (e.g. DIO1, GH1) or transrepression (e.g. TSHB, TRH) in T3-dependent manner. These variants usually act as dominant-negative mutants. Most of the germline, clinically associated mutations of the TRβ receptor have been identified in patients with autosomal recessive or dominant, generalized thyroid hormone resistance (RTH, OMIM: 188570, 274300 respectively) as well as selective pituitary thyroid hormone resistance (OMIM: 145650). Patients with RTH have got impairment of the mechanism negatively regulating the feedback of T4/T3 to the hypothalamic TRH and pituitary TSH genes by the mutated TRβ receptors.
There are also studies showing the mutations in TRβ-DNA binding domain, in 5 and 3 mRNA untranslated regions (UTRs) and intronic variants with potential disease association. These allelic variants have been identified to impair transcription, alternative splicing, translation or TRβ protein function such as hormone binding, DNA binding, ligand-dependent conformational changes or corepressors/coactivators dissociation/association function (see references).

There are no reference variants reported as somatic in the NCBI dbSNP and ClinVar databases, however 3 allele variants (ID: nsv429603, nsv429566, nsv429555) are present in the NCBI dbVar. Morover, several studies tested the hypothesis that the functions of TRβ could be impaired in various cancer tissues by somatic mutations (see references). For instance, Puzianowska-Kuznicka et al. (2002) tested this hypothesis in selected human thyroid papillary cancer. Based on cancer-derived cDNAs, they found that the mean expression levels of TRβ1 mRNA and TRα1 mRNA were significantly lower, whereas the protein levels of both were higher in cancer tissues compared to healthy thyroid samples. Sequencing of TRβ1 and TRα1 cDNAs, cloned from 16 papillary cancers, revealed that mutations affected receptor amino acid sequences in 93.75% and 62.5% of cases, respectively. In contrast, no mutations were identified in healthy thyroid controls, and only 11.11% and 22.22% of thyroid adenomas had such TRβ1 or TRα1 mutations, respectively. The authors summarized that the findings suggest a possible role for mutated thyroid hormone receptors in the tumorigenesis of human papillary thyroid carcinoma (NCBI OMIM: 188550). The similar conclusion can be found in studies of Kamiya et al. (Kamiya et al., 2002), who have cloned and sequenced 22 cDNAs obtained from the human renal clear cell carcinoma (OMIM: 144700). Somatic mutations were found in 7 TRβ1 and 3 TRα1 samples. These findings are consistent with the results obtained on hepatocellular carcinoma HCC (Lin et al., 1999). However, some data from direct genomic DNA sequencing provided an evidence that somatic THRB gene mutations may not be as common in differentiated thyroid cancers, in which hypermethylation of the gene was shown to be a major mechanism responsible for down-regulation of the gene expression. Indeed, THRB has been proposed to serve as a novel epigenetic marker for early detection and prognosis of high grade serous ovarian cancer (Kashuba et al., 2013) and identified to be frequently methylated in prostate cancer, breast cancer, non-small cell lung cancer and acute lymphoblastic leukemia (Dmitriev et al., 2009; Dmitriev et al., 2012; Ling et al., 2010; Vasiljevic et al., 2011).
Searching the cBio Cancer Genomics Portal (Cerami et al., 2012), currently providing access to data from more than 5000 tumor samples from 20 cancer studies revealed that somatic mutations were the most frequently found in Skin Cutaneous Melanoma (up to 5.7% cases altered in the database). The mutations were also reported in stomach adenocarcinoma (4.1%), kidney renal clear cell carcinoma (3.6%), colorectal cancer (2.8%), pancreatic adenocarcinoma (2.4%), colon and rectum adenocarcinoma (1.9%), head and neck squamous cell carcinoma (1.8%), lung adenocarcinoma (1.6%), ovarian serous cystadenocarcinoma (1.6%), uterine corpus endometrioid carcinoma (1.3%), sarcoma (1.2%), lung squamous cell carcinoma (1.1%), breast invasive carcinoma (0,9%) and acute myeloid leukemia (0,5%). Most of the cancer mutations were identified as substitutions, however deletions were the most frequently found in renal clear cell. There were also reported amplifications in pancreatic adenocarcinoma, kidney chromophobe renal cell carcinoma and sarcoma (Gao et al., 2013; see cBioPortal in external links). It is noticeable that the reported alteration frequency can vary in different studies depending on the target populations, number of tested samples, analytical methods and histopathological classification of the tumors.
Somatic mutations/sequence variants have been shown to be created post-transcriptionally in various cancer-derived transcripts (Klimek-Tomczak et al., 2006; Chen et al., 2013). Adenosine (A) to inosine (I) RNA editing of AZIN1 was demonstrated to be increased in the hepatocellular carcinoma and suggested as a potential driver in the pathogenesis of human cancers, particularly HCC. ADAR1-mediated A-to-I RNA editing was shown to change the RNA nucleotide sequence relative to that of the encoding DNA that was reported to result in cancer development and progression (Huang et al., 2013). ADAR belongs to the family of RNA specific adenosine deaminase, which acts on double-stranded RNA (dsRNA) substrats including those created by long naturally occurring antisence transcripts (cis- and trans-NATs) such as Rev-erbα/TRα2 or intra-THRB transcripts (see diagram 1). Then, inosine, which in RNA can pairs with cytosine (C) uracil (U) or adenine (A, generating wobble pairs), is recognized as guanosine (G) by ribosomes (translation) and reverse transcriptase during cDNA synthesis. In fact, the comparison of whole genome sequences and deeply sequenced transcriptomes of human lymphoblastoid cell line GM12878 and a Han Chinese individual (YH) allowed for identification of two mismatches corresponding to A/I(G) ediging sites within THRB sequence (DNA ref. sec. position: 24167619, 24321108) as well as several editing sites in NR1D2 gene (see diagram 1, genomic context, Ramaswami et al., 2012). The next-generation sequencing used in the study permitted comparing genomic DNA and RNA sequencing data from the same individuals. These preliminary findings, however obtained from non-cancerous samples, may indicate that the ADAR-mediated RNA editing may change the sequence of THRB pre-mRNA post-transcriptionally (Ramaswami et al., 2012) and that this process may be impaired in numerous cancers (Huang et al., 2013).