Thyroid Hormone Receptor (cid:2) Mutation Causes a Severe and Thyroxine-Resistant Skeletal Dysplasia in Female Mice

A new genetic disorder has been identified that results from mutation of THRA , encoding thyroid hormone receptor (cid:2) 1 (TR (cid:2) 1). Affected children have a high serum T 3 :T 4 ratio and variable degrees of intellectual deficit and constipation but exhibit a consistently severe skeletal dysplasia. In an attempt to improve developmental delay and alleviate symptoms of hypothyroidism, patients are receivingvaryingdosesanddurationsofT 4 treatment,butresponseshavebeeninconsistentsofar. Thra1 PV/ (cid:2) mice express a similar potent dominant-negative mutant TR (cid:2) 1 to affected individuals, andthusrepresentanexcellentdiseasemodel.Wehypothesizedthat Thra1 PV/ (cid:2) micecouldbeused to predict the skeletal outcome of human THRA mutations and determine whether prolonged treatment with a supraphysiological dose of T 4 ameliorates the skeletal abnormalities. Adult female Thra1 PV/ (cid:2) mice had short stature, grossly abnormal bone morphology but normal bone strengthdespitehighbonemass.AlthoughT 4 treatmentsuppressedTSHsecretion,ithadnoeffect onskeletalmaturation,lineargrowth,orbonemineralization,thusdemonstratingprofoundtissue resistance to thyroid hormone. Despite this, prolonged T 4 treatment abnormally increased bone stiffnessandstrength,suggestingthepotentialfordetrimentalconsequencesinthelongterm.Our studies establish that TR (cid:2) 1 has an essential role in the developing and adult skeleton and predict that patients with different THRA mutations will display variable responses to T 4 treatment, which depend on the severity of the causative mutation. ( Endocrinology 155: 3699–3712, 2014) between calcein labels at 20 locations per specimen beginning 0.2 mm below the growth plate. BS and mineralizing surface were measured using ImageJ, and the bone formation rate was calculated by multiplying mineralizing surface and mineral apposition rate.

T he THRA and THRB genes encode the nuclear receptors (thyroid hormone receptor [TR␣] and TR␤), which mediate thyroid hormone action in target tissues (1). Autosomal-dominant resistance to thyroid hormone (RTH) was recognized in 1967 (2), and the first causative mutations affecting THRB were identified 22 years later (3). More thanr 1000 RTH families have since been de-scribed, and affected individuals have increased thyroid hormone levels with an inappropriately normal or elevated TSH concentration due to disruption of the hypothalamus-pituitary-thyroid axis (4).
After the identification of THRB mutations in individuals with RTH it was a further 23 years before the first THRA mutations were reported in 2012 and 2013 (5)(6)(7).
A six year-old girl with skeletal dysplasia and growth retardation was found to have a heterozygous THRA nonsense mutation resulting in expression of a truncated TR␣1 E403X protein. She had normal serum TSH with low/ normal T 4 and high/normal T 3 concentrations. Further investigations revealed macrocephaly with patent and abnormal skull sutures, delayed tooth eruption and bone age, disproportionate short stature, and epiphyseal dysgenesis with delayed mineralization of secondary ossification centers. Treatment with T 4 for 9 months resulted in suppression of TSH and an increased basal metabolic rate but did not improve linear growth or skeletal development (5). A second girl with similar thyroid function and skeletal dysplasia was found to have a heterozygous frameshift mutation in THRA resulting in expression of a truncated TR␣1 F397fs406X protein (6). She presented at the age of 3 years with macrocephaly, delayed tooth eruption, absent secondary ossification centers, and congenital hip dislocation. Reducing growth velocity became evident between 3 and 6 years of age. T 4 treatment between 6 and 11 years of age only resulted in a small increase in growth velocity for a 2-month period but ultimately had no effect on her height, which continued along the 20th centile and was accompanied by persistently delayed bone age. The girl's 47-year-old father had the same THRA mutation and displayed short stature with a height 3.77 SDs below normal and acquired hearing loss due to otosclerosis (6,8). Recently, a 45-year-old female with similar thyroid function, macrocephaly, and disproportionate short stature was identified and found to have a heterozygous frameshift mutation in THRA, resulting in expression of a truncated TR␣1 P382fs388X protein. She presented in infancy with developmental delay and was treated intermittently with T 4 , which resulted in some improvement in growth velocity although her final adult height remained 2.34 SDs below normal (7).
These recent reports define a new genetic disorder characterized by a severe developmental phenotype with profound skeletal abnormalities that are thought to result from impaired T 3 action in bone and cartilage (5)(6)(7)(8). In an attempt to ameliorate the phenotype, three children have already received intermittent T 4 at different doses and for varying durations. However, responses to date have been limited, and it is unknown whether long-term T 4 treatment will be beneficial or detrimental. Thus, it is now essential to define the adult skeletal consequences of THRA mutations and determine the longterm effects of T 4 supplementation, because life-long therapy is likely to be required. Importantly, Van Mullem et al (8) showed that dominant-negative inhibition of TR␤ by TR␣1 F397fs406X in vitro could be overcome partially by a high concentration of thyroid hormone. Furthermore, sev-eral studies have demonstrated that TR␤ may mediate T 3 actions in bone and cartilage (9 -12), even though the principal physiological effects are mediated by TR␣1. These observations suggest, therefore, that treatment of patients with supraphysiological doses of T 4 may improve their skeletal abnormalities via TR␤-mediated actions. To address this timely question we investigated the effects of prolonged T 4 treatment in a mouse model of this novel disease.
Mice with dominant-negative mutations affecting Thra (Thra1 PV ) and Thrb (Thrb PV ) were generated to investigate the tissue-specific roles of TR␣ and TR␤ and aid the identification of patients with THRA mutations (13,14). The PV mutation, first recognized in a patient with RTH, is a C-insertion in exon 10 of THRB that results in a frameshift affecting the C-terminal 16 amino acids (15). The equivalent Thra1 PV mutation comprises a homologous Cinsertion followed by the PV sequence described in THRB PV (14,16). The Thra1 PV mutation disrupts helix 12 of TR␣1, which is essential for T 3 binding and coactivator recruitment (17) and lies within a 21-amino acid region containing the described human THRA mutations (5-7). Accordingly, TR␣1 PV cannot bind T 3 or activate target gene transcription but acts as a potent dominant-negative inhibitor of wild-type (WT) TR␣1 or TR␤ (14,18,19). Thus, the functional characteristics of TR␣1 PV closely resemble those reported for TR␣1 E403X , TR␣1 F397fs406X , and TR␣1 P382fs388X (5-7). Importantly, PV and none of the described human mutations affect the sequence of the TR␣2 isoform that is also expressed from the THRA locus but which cannot bind T 3 and has no known physiological function. Consistent with this, juvenile Thra1 PV/ϩ mice display the same characteristics as children with heterozygous THRA mutations. They have a reduced T 4 :T 3 ratio (14), delayed closure of the skull sutures with enlarged fontanelles, and severe postnatal growth retardation with delayed bone age. These abnormalities result from impaired TR␣1-mediated T 3 action in bone and cartilage (20 -22), indicating that Thra1 PV mice represent an excellent disease model in which to investigate the consequences of prolonged T 4 treatment.
We hypothesized that the adult phenotype of Thra1 PV/ϩ mice would predict the skeletal outcome of human THRA mutations and determine whether affected individuals may be susceptible to fracture or osteoarthritis, both of which are associated with altered thyroid hormone action in bone (23)(24)(25)(26). We also hypothesized that prolonged treatment of Thra1 PV/ϩ mice with a supraphysiological dose of T 4 would ameliorate the developmental skeletal phenotype and improve bone structure and strength in adulthood.
The current studies demonstrate that adult Thra1 PV/ϩ mice have short stature but normal bone strength despite high bone mass, suggesting that patients with THRA mutations are unlikely to have an increased risk of fracture. By contrast, gross morphologic abnormalities of the bones and joints predict that individuals with THRA mutations may be predisposed to osteoarthritis (27,28). Although treatment with a supraphysiological dose of T 4 completely suppressed TSH secretion, it had no effect on skeletal maturation, linear growth, or bone mineralization, thus demonstrating profound tissue resistance to thyroid hormone in Thra1 PV/ϩ mice. However, prolonged T 4 treatment increased bone stiffness and strength abnormally due to progressive enlargement of cortical bone diameter and thickness. Overall, the findings suggest that T 4 treatment of individuals with dominant-negative THRA mutations is unlikely to improve their skeletal abnormalities substantially and may even be detrimental in the long term. Nevertheless, Thra1 PV/ϩ mice represent an important disease model in which to identify and evaluate new therapeutic approaches.

Thra1 PV mice
WT and heterozygous Thra1 PV/ϩ mice have a mixed C57BL/6J and NIH Black Swiss genetic background and were bred and genotyped as described elsewhere (14,21). Detailed characterization of the adult skeleton in Thra1 PV/ϩ mice was performed in 14-week-old female mice after cessation of growth, and in fully mature 20-week-old female mice that had been treated with vehicle or T 4 from weaning at 4 weeks of age until death. All mice were given ip injections of calcein (10 mg/kg in 100 L PBS) 14 and 7 days before tissue collection (29).

Ethics
Animal studies were performed according to the National Institutes of Health Guide for Care and Use of Laboratory Animals, and the National Cancer Institute Animal Care and Use Committee granted ethical approval for all experiments.

Manipulation and measurement of thyroid status
TSH, T 4 , and T 3 levels were determined in serum from mice (n ϭ 5-13 per group) treated with vehicle or T 4 (1.2 g/mL in the drinking water) between 4 -20 weeks of age. T 4 -supplemented water was changed every 3 days, with the T 4 concentration adjusted to intake in 2-week cycles to ensure all animals received the same amount of T 4 and did not become markedly thyrotoxic (14, 30 -32).

Histology
Tibias were fixed in 10% neutral buffered formalin and decalcified in 10% EDTA, embedded in paraffin wax. Sections (5 m) were stained with alcian blue and van Gieson (29,33). Measurements from at least 4 separate positions across the growth plate were obtained to calculate the mean height using a Leica DM LB2 microscope and DFC320 digital camera (Leica Microsystems). Results from 2 levels of sectioning were compared.

Faxitron digital x-ray microradiography
Femurs were imaged at 10 m resolution using a Faxitron MX20 (Qados). Bone mineral content was determined relative to steel, aluminum, and polyester standards. Images were calibrated with a digital micrometer, and bone length, cortical bone diameter, and thickness were determined (33,34).

Micro-computed tomography (CT)
Femurs were analyzed by micro-CT (Skyscan 1172a) at 50 kV and 200 A with a detection pixel size of 4.3 m 2 , and images were reconstructed using Skyscan NRecon software. A 1-mm 3 region of interest was selected 0.2 mm from the growth plate, and trabecular bone volume as proportion of tissue volume (BV/TV), trabecular number, and trabecular thickness were determined (29,33). Representative femurs from each treatment group were rescanned using a SCANCO CT 40 (SCANCO Medical AG) operating at 55 kVp peak energy detection, 6 m resolution to obtain approximately 2500 cross-sections per specimen in 766 ϫ 763 pixel 16 bit DICOM files. Raw data were imported using 32-bit Drishti v2.0.221 (Australian National University Supercomputer Facility, http://anusf.anu.edu.au/Vizlab/drishti/) and rendered using 64-bit Drishti v2.0.000 to generate high-resolution images.

Back scattered electron-scanning electron microscopy (EM) (BSE-SEM)
Femurs were fixed in 70% ethanol and opened longitudinally (33). Carbon-coated samples were imaged using backscattered electrons with a Zeiss DSM962 digital scanning electron microscope (EM) at 20-kV beam potential (KE Electronics). Highresolution images were quantified using ImageJ to determine the fraction of trabecular and endosteal bone surfaces displaying osteoclastic resorption (33).

Quantitative BSE-SEM
Bone mineralization was determined by quantitative BSE-SEM at 1-m 3 resolution. Specimens were embedded in methacrylate and block faces polished to an optical finish for scanning electron microscopy (EM) analysis at 20 kV, 0.5nA with a working distance of 11 mm (33). Gradations of micromineralization density were represented in 8 equal intervals by a pseudocolor scheme (33,35).

Osteoclasts
Sections from decalcified tibias were stained for tartrate-resistant acid phosphatase, counterstained with aniline blue, and imaged using a Leica DM LB2 microscope and DFC320 digital camera (29,33). A montage of 9 overlapping fields covering an area of 1 mm 2 located 0.2 mm below the growth plate was constructed for each bone. BV/TV was measured, and osteoclast numbers and surface were determined in trabecular bone normalized to total bone surface (BS) (29,33).

Osteoblasts
Methacrylate-embedded specimens were imaged with a Leica SP2 reflection confocal microscope at 488-nm excitation to determine the fraction of BS undergoing active bone formation (33,36). Mineral apposition rate was calculated by determining the separation between calcein labels at 20 locations per specimen beginning 0.2 mm below the growth plate. BS and mineralizing surface were measured using ImageJ, and the bone formation rate was calculated by multiplying mineralizing surface and mineral apposition rate.

Bone strength
Three-point bend tests were performed on tibias, with a constant rate of displacement of 0.03 mm/s until fracture, using an Instron 5543 load frame and 100N load cell (Instron Limited). Biomechanical variables reflecting cortical bone strength were derived from load displacement curves (33,37).

Statistics
Data were analyzed by unpaired two-tailed Student's t test; P Ͻ .05 was considered significant. Frequency distributions of mineralization densities obtained by Faxitron and quantitative BSE were compared using the Kolmogorov-Smirnov test (29,33,34).

Thyroid status and response to T 4 administration in Thra1 PV/؉ mice
The thyroid status of adult WT and Thra1 PV/ϩ mice was determined following treatment with vehicle or a supraphysiological dose of T 4 from weaning until 14 weeks of age ( Figure 1). The basal T 4 concentration did not differ between WT and Thra1 PV/ϩ mice, whereas T 3 and TSH levels were increased in Thra1 PV/ϩ mice by 1.5-fold (P Ͻ .01) and 6-fold (P Ͻ .001), respectively. Thus, the characteristically reduced T 4 :T 3 ratio identified in individuals with THRA mutations (5-7) was also present in Thra1 PV/ϩ mice (T 4 :T 3 ratio: Thra1 PV/ϩ 23 vs WT 39). Supraphysiological T 4 treatment completely suppressed TSH in both WT and Thra1 PV/ϩ mice. Despite profound and similar suppression of TSH, the increases in circulating T 4 and T 3 concentrations were attenuated in Thra1 PV/ϩ mice (T 4 , 3.5-fold increase; T 3 , 1.5-fold) compared with WT (T 4 , 6-fold increase, P Ͻ .001; T 3 , 4-fold, P Ͻ .01) indicating that they are resistant to T 4 administration.

Delayed ossification and impaired bone modeling in Thra1 PV/؉ mice
Delayed bone development in juvenile Thra1 PV/ϩ mice (21) led to severe skeletal abnormalities in adults. Growth plates in 14-and 20-week-old Thra1 PV/ϩ mice were 39% and 70% wider than in WT mice ( Figure 2, A and B), demonstrating persistent delay of endochondral ossification. An increased degree of retention of mineralized cartilage within trabeculae revealed that bone modeling was also impaired ( Figure 2C). T 4 administration did not affect either of these abnormalities in mutant mice (Figure 2 and data not shown).

Structural consequences of defective ossification, modeling, and remodeling in adult Thra1 PV/؉ mice
Bones from 14-and 20-week-old Thra1 PV/ϩ mice were grossly dysmorphic. They were 17% and 15% shorter than WT and had splayed metaphyses, an abnormal crosssection throughout the diaphysis, and misshapen joint surfaces ( Figure 3A). Micro-CT analysis indicated that trabecular bone volume, number, and thickness were increased in 20-week-old Thra1 PV/ϩ mice (BV/TV, 2.1fold; trabecular number, 1.9-fold; trabecular thickness, 1.1-fold greater) (Supplemental Figure 1), and these findings were confirmed by back-scattered electron-scanning EM (BSE-SEM) ( Figure 3B). Similarly, cortical bone thickness (48% wider at 14 weeks, 43% at 20 weeks) and periosteal diameter (13% larger at 14 weeks, 20% at 20 weeks) were markedly increased in Thra1 PV/ϩ mice (Supplemental Figure 1). T 4 administration had no effect on these morphologic abnormalities ( Figure 3A) but resulted in a gradual increase in cortical bone thickness and diameter in Thra1 PV/ϩ mice (Supplemental Figure 1). Importantly, the endosteal diameter did not change in Thra1 PV/ϩ mice following T 4 treatment, whereas in WT mice it increased by 16% (P Ͻ .01). Thus, the increase in cortical bone thickness in Thra1 PV/ϩ mice resulted from a failure of endosteal bone resorption combined with a likely increase in periosteal bone deposition.
Increased bone mineral content but reduced mineralization in Thra1 PV/؉ mice X-ray microradiography revealed that 14-week-old Thra1 PV/ϩ mice had lower bone mineral content than WT mice, consistent with reduced mineral accrual during postnatal growth (21). Thus, in Figure 4A, the pseudocolored images in 14-week-old mice show more yellow and fewer red pixels in Thra1 PV/ϩ mice compared with WT, indicating reduced bone mineral content. These differences are shown graphically in Figure 4B, in which the frequency distribution for Thra1 PV/ϩ mice is shifted to the left. By contrast, in 20-week-old mice there was a small shift to the right in the pixel frequency distribution for Thra1 PV/ϩ , mice indicating higher, rather than lower, bone mineral content in older animals (Figure 4, A and B). Remarkably, supraphysiological T 4 treatment further increased bone mineral content in Thra1 PV/ϩ mice even though, as expected, it was reduced in WT mice following treatment (Figure 4, A and B). Thus, Thra1 PV/ϩ mice were resistant to T 4 -induced bone loss and had a paradoxical increase in bone mineral content following treatment. Despite this, BSE-SEM revealed that cortical and trabecular bone mineralization density was reduced in 20 week-old Thra1 PV/ϩ mice, the difference being greater in cortical bone, and that T 4 treatment did not affect mineralization ( Figure 5, A-D). Thus, Thra1 PV/ϩ mice have an increase in bone mineral content (Figure 4) despite the reduction in tissue mineralization density ( Figure 5) because their trabecular and cortical bone volume is substantially increased (Figure 2 and Supplemental Figure 1). Overall, therefore, Thra1 PV/ϩ mice have increased cortical and trabecular bone volume compared with WT, but their bone is less mineralized.

Reduced osteoclastic bone resorption in Thra1 PV/؉ mice
Consistent with micro-CT and BSE-SEM analysis, histomorphometry studies demonstrated increased bone volume and surface in Thra1 PV/ϩ mice. Furthermore, osteoclast surfaces were reduced and fewer osteoclasts were present in Thra1 PV/ϩ mice compared with WT ( Figure 6, A-C). Thus, Thra1 PV/ϩ mice had a smaller proportion of their increased BS covered by osteoclasts (see also Supplemental Figure 2). The differences in BS, BV/TV, osteoclast surface/BS, and osteoclast number/BS between WT and Thra1 PV/ϩ mice were accentuated following T 4 treatment ( Figure 6, A-C). Consistent with these findings, bone resorption was generally lower in Thra1 PV/ϩ mice (Supplemental Figure  2) but bone formation parameters were similar (Supplemental Figure 3). However, it is important to note that small differences in dynamic bone formation may not have been detected in these studies because only 3 mice were analyzed per group.

Abnormal bone stiffness and strength after prolonged T 4 treatment of Thra1 PV/؉ mice
Biomechanical testing revealed no difference in bone strength between untreated WT and Thra1 PV/ϩ mice (Figure 7, A and B). Nevertheless, T 4 treatment resulted in gradual increases in yield load, maximum load, fracture load, and stiffness of bones from Thra1 PV/ϩ mice (Figure 7, A and B). Thus, prolonged T 4 administration abnormally and progressively increased bone stiffness and strength in Thra1 PV/ϩ mice.

Skeletal phenotype resulting from mutation of Thra
During development Thra1 PV/ϩ mice have delayed closure of the skull sutures, severe growth retardation, delayed bone age, and impaired bone mineral accrual (22). The delayed ossification persists into adulthood and is accompanied by impaired bone modeling and remodeling, resulting in short stature, increased bone mass, and gross morphologic abnormalities of the bones and joints, but normal bone strength. These findings suggest that, despite severe skeletal abnormalities, adults with THRA mutations are unlikely to have an increased risk of fracture. However morphologic abnormalities affecting the bones and joints predict that they may be at increased risk of osteoarthritis (27,28).

Cellular and molecular mechanisms
The abnormalities in Thra1 PV/ϩ mice are consistent with effects of prolonged hypothyroidism on the growing and adult skeleton (38 -42). Hypothyroidism disrupts growth plate chondrocyte differentiation leading to delayed endochondral ossification and linear growth, impairs bone modeling, and uncouples the processes of osteoclastic bone resorption and osteoblastic bone formation (43). In adults, even though it is well established that thyroid hormones increase bone resorption and promote bone loss, it is not known whether T 3 acts directly in osteoclasts or whether effects on osteoclasts are secondary to the direct actions of T 3 in osteoblasts (43). In Thra1 PV/ϩ mice, prolonged impairment of chondrocyte differentiation is manifest by growth retardation and short stature in adulthood. Similarly, defective osteoclastic bone resorption is evidenced by reduced metaphyseal in-wasting, ab- normal diaphyseal cross-section, and increased trabecular bone volume with retention of mineralized cartilage. Moreover, the grossly delayed formation of secondary ossification centers and reduced bone mineral accrual in Thra1 PV/ϩ mice persisted throughout growth when mice were active and gaining weight. Thus, unmineralized epiphyses were exposed to abnormal and greater mechanical loads, resulting in compensatory enlargement of the epiphyses and metaphyses and culminating in adult joint deformity. Surprisingly, the strength of adult Thra1 PV/ϩ bones was normal despite these structural abnormalities and is accounted for by the increased cortical bone thickness and diameter (33,44). A series of studies in genetically modified mice have shown that TR␣1 is the principal mediator of T 3 action in bone and cartilage (12,21,(45)(46)(47)(48). The finding of an identical skeletal phenotype in patients with THRA mutations (5-7) now demonstrates that TR␣1 has a similar essential role in human bone development. Analysis of the mechanisms underlying the skeletal phenotypes in Thra mutant mice revealed decreased expression of T 3 target genes including GH receptor (Ghr), insulin like growth factor-1 (Igf1), Igf1 receptor (Igf1r), fibroblast growth factor receptor-1 (Fgfr1) and Fgfr3, and reduced downstream signaling responses mediated by the MAPK, signal transducer and activator of transcription 5, and AKT signaling pathways in chondrocytes and osteoblasts (12,20,21,45,49,50). These data demonstrate impaired T 3 action in cartilage and bone in Thra mutant mice despite a normal systemic T 3 concentration and thus indicate the skeletal phenotype in individuals with THRA mutations is a consequence of local resistance to thyroid hormone.
The phenotypes in Thra1 PV/ϩ mice and patients with THRA mutations result from the actions of potent dominant-negative mutant receptors. However, we have pre-viously reported that mice harboring a less severe Thra1 R384C mutation have a milder phenotype with only transiently delayed ossification and growth retardation, although modeling and remodeling defects resulting in increased bone mass, cortical thickness, and diameter were present in adults (45,47). Importantly, and in contrast to Thra1 PV/ϩ mice, treatment of Thra1 R384C mice with a dose of T 3 that overcomes the reduced ligand binding affinity and dominant-negative activity of the mutant receptor did ameliorate their skeletal abnormalities (45).

Therapeutic approaches in individuals with THRA mutations
The response to thyroid hormone treatment in Thra1 R384C mice suggests that individuals with THRA mutations may benefit from similar treatment. Unfortunately, however, doses of T 4 sufficient to normalize circulating hormone concentrations have been largely ineffective in the patients treated so far (5)(6)(7)(8), presumably because the currently identified individuals have mutations that result in expression of mutant receptors with little or no T 3 binding affinity. Despite this, Van Mullem et al (8) showed that dominant-negative inhibition of TR␤ by TR␣1 F397fs406X in vitro could be overcome partially by increasing concentrations of thyroid hormones.  In this context, several studies have suggested that TR␤ can mediate T 3 action in bone and cartilage (9 -12), even though the principal physiological effects are mediated via TR␣1. Thus, we hypothesized that treatment of Thra1 PV/ϩ mice with a supraphysiological dose of T 4 might improve bone structure and strength. However, such treatment of Thra1 PV/ϩ mice had no beneficial effect on growth or skeletal deformity but did, nevertheless, increase cortical bone thickness and diameter. These responses were likely mediated by TR␤ and resulted in abnormal increases in bone stiffness and strength that may adversely affect the optimal compromise between strength and flexibility that is essential to minimize fracture risk (51). Thus, prolonged treatment of individuals harboring THRA mutations with high doses of T 4 may also have adverse consequences in other tissues where T 3 action is predominantly mediated via TR␤.
GH therapy represents an alternative approach to improve linear growth and skeletal maturation in children with THRA mutations, but treatment in one individual so far was ineffective (6). The reduced expression of Ghr, Igf1, and Igf1r, together with impaired signal transducer and activator of transcription 5 and AKT signaling in growth plate chondrocytes in Thra mutant mice (12,21,50), suggests a mechanism to account for this lack of clinical response to GH.

Thyroid hormone metabolism and response to T 4 administration in Thra1 PV/؉ mice
Thyroid hormone metabolism is mediated by 3 iodothyronine deiodinases. The type 1 enzyme (D1) catalyzes removal of an inner or outer ring iodine from T 4 to generate T 3 or rT 3 , or an outer ring iodine from rT 3 to generate 3,3Ј-diiodothyronine. D1 is expressed in liver, kidney, and thyroid and contributes to the circulating concentration of T 3 (52). The type 2 enzyme (D2) converts the prohormone T 4 to the active hormone T 3 : it is expressed in the hypothalamus and pituitary and peripheral target tissues, where it generates a local A, Normal response in WT mice. High concentrations of T 4 are metabolized in the liver. D1 converts T 4 to rT 3 or T 3 , and rT3 is metabolized to 3,3Ј-diiodothyronine (T 2 ). Acting via TR␤1, T 3 increases D1 expression to complete a feed-forward loop. However, T 3 also acts via TR␣1 to increase D3 expression and thus limit feed-forward activation of D1. Thus, T 4 excess results in a parallel increase in both D1 and D3 so that levels of T 3 , rT 3 , and T 2 in the circulation rise to reflect increased T 4 metabolism. The high levels of circulating thyroid hormones suppress TRH and TSH expression and inhibit endogenous T 4 and T 3 production. At steady state, most circulating T 3 is derived from increased D1-mediated metabolism of T 4 . The TR␣1-mediated actions of T 3 in bone are increased. B, Abnormal response in Thra1 PV/ϩ mice. High concentrations of T 4 are metabolized in the liver. D1 converts T 4 to rT 3 or T 3 , and rT 3 is metabolized to T 2 . Acting via TR␤1, T 3 increases D1 expression to complete a feed-forward loop. However, in Thra1 PV/ϩ mice the mutant TR␣1 PV prevents T 3 stimulation of D3 expression, thus maintaining feed-forward activation of D1. Administration of T 4 fuels this feed-forward activation and would result in enhanced metabolism of T 4 , and ultimately increased accumulation of T 2 . Thus, although circulating T 3 and T 4 levels rise to a lesser degree than in WT animals, they are still sufficient to suppress the hypothalamus-pituitary-thyroid axis. At steady state, the grossly increased D1 activity thus accounts for resistance of Thra1 PV/ϩ mice to T 4 administration. Despite exogenous thyroid hormone administration, T 3 action in bone remains inhibited by dominant-negative TR␣1 PV (21). doi: 10.1210/en.2013-2156 endo.endojournals.org supply of T 3 and is subject to substrate-mediated inactivation (53). By contrast, the type 3 enzyme (D3) catalyzes removal of an inner ring iodine from T 4 or T 3 to generate the inactive metabolites rT 3 or 3,3Ј-diiodothyronine. D3 expression is induced by thyroid hormone, thus limiting the supply of T 3 in conditions of thyroid hormone excess (52).
Remarkably, and despite complete suppression of TSH, Thra1 PV/ϩ mice had a blunted increase in circulating thyroid hormones following a supraphysiological dose of T 4 . This discrepancy indicates that the hypothalamus-pituitary-thyroid axis is intact in Thra1 PV/ϩ mice, but metabolism of thyroid hormones must be increased. Indeed, we previously showed that untreated Thra1 PV/ϩ mice have a 9-fold increase in hepatic D1 mRNA expression (14) resulting in a 4.8-fold increase in enzyme activity (54). It is well established that T 3 acts via TR␤1 to stimulate D1 expression in the liver (55,56) and, accordingly, hepatic D1 activity is increased further in Thra1 PV/ϩ mice following treatment with T 3 (54,57). By contrast, T 3 acts via TR␣1 to stimulate expression of D3 (58) and we previously demonstrated that T 3 treatment of Thra1 PV/ϩ mice fails to induce the normal increase in D3 activity observed in WT animals (54,57).
We propose, therefore, that the resistance to T 4 administration observed in Thra1 PV/ϩ mice results from the markedly increased D1 activity combined with this absent D3 response (Figure 8). Consistent with this model, TSH in individuals with THRA mutations was suppressed readily following T 4 treatment despite only small increases in T 4 and T 3 concentrations (7,8). Detailed future metabolic studies will be required to confirm the precise underlying mechanisms responsible for these findings. For example, because defects in TR␣1 action may result in intestinal problems, it is possible that absorption of orally administered T 4 could be impaired in Thra1 PV/ϩ mice. However, it should also be noted that, following oral treatment with T 4 , the TSH concentration was suppressed completely in both WT and Thra1 PV/ϩ mice, indicating that intestinal absorption of T 4 was unlikely to be markedly impaired in Thra1 PV/ϩ mice. Nevertheless, it would be instructive to investigate whether differences in serum T 4 and T 3 levels persist between WT and Thra1 PV/ϩ mice following parenteral administration of T 4 .

Conclusions
The overall resistance of the skeleton to T 4 treatment in Thra1 PV/ϩ mice and the patients studied so far is likely to be a consequence of the potent dominant-negative activities of their mutant TR␣1 proteins (5-7, 14, 18). It is inevitable, however, that individuals with less severe THRA mutations will be identified in the future and, in such cases, T 4 treatment is likely to be beneficial. Thus, treatment of Thra1 R384C mice with doses of T 4 that overcome the reduced binding affinity of TR␣1 R384C rescued their skeletal phenotype by preventing delayed ossification and growth retardation, ultimately ameliorating adult bone structure and mineralization (45). Taken together, these studies predict that individuals with THRA mutations will display variable degrees of skeletal deformity and different responses to T 4 treatment that correlate with the functional consequences of the particular diseasecausative mutation. Therefore, in patients with THRA mutations, it will be important to characterize the functional properties of their mutant TR␣1 because this may predict their response to T 4 treatment and the optimal systemic T 4 concentration required.