The Thyroid Hormone Receptor β-Specific Agonist GC-1 Selectively Affects the Bone Development of Hypothyroid Rats


  • Fatima RS Freitas,

    1. Department of Anatomy, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Luciane P Capelo,

    1. Department of Histology, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Patrick J O'Shea,

    1. Molecular Endocrinology Group, Division of Medicine and Medical Research Council Clinical Sciences Centre, Imperial College London, Hammersmith Hospital, London, United Kingdom
    Search for more papers by this author
  • Vanda Jorgetti,

    1. School of Medicine, University of Sao Paulo, Sao Paulo, SP, Brazil
    Search for more papers by this author
  • Anselmo S Moriscot,

    1. Department of Histology, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil
    Search for more papers by this author
  • Thomas S Scanlan,

    1. Departments of Pharmaceutical Chemistry and Cellular and Molecular Pharmacology, University of California, San Francisco, California, USA
    Search for more papers by this author
  • Graham R Williams,

    1. Molecular Endocrinology Group, Division of Medicine and Medical Research Council Clinical Sciences Centre, Imperial College London, Hammersmith Hospital, London, United Kingdom
    Search for more papers by this author
  • Telma MT Zorn,

    1. Department of Histology, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil
    Search for more papers by this author
  • Cecilia HA Gouveia PhD

    Corresponding author
    1. Department of Anatomy, Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil
    • Department of Anatomy, Institute of Biomedical Sciences, University of Sao Paulo, Av. Prof. Lineu Prestes, 2415, Sao Paulo 05508-900, Brazil
    Search for more papers by this author

  • The authors have no conflict of interest.


We investigated the effects of GC-1, a TRβ-selective thyromimetic, on bone development of hypothyroid rats. Whereas T3 reverted the IGF-I deficiency and the skeletal defects caused by hypothyroidism, GC-1 had no effect on serum IGF-I or on IGF-I protein expression in the epiphyseal growth plate of the femur, but induced selective effects on bone development. Our findings indicate that T3 exerts some essential effects on bone development that are mediated by TRβ1.

Introduction: We investigated the role of the thyroid hormone receptor β1 (TRβ1) on skeletal development of rats using the TRβ-selective agonist GC-1.

Materials and Methods: Twenty-one-day-old female rats (n = 6/group) were rendered hypothyroid (Hypo) and treated for 5 weeks with 0.3 ug/100 g BW/day of T3 (1xT3), 5xT3, or equimolar doses of GC-1 (1xGC-1 and 5xGC-1). Serum triiodothyronine (T3), thyroxine (T4), thyroid-stimulating hormone (TSH), and insulin-like growth factor (IGF)-I concentrations were determined by radioimmunoassay (RIA). BMD and longitudinal bone growth were determined by DXA. Trabecular bone histomorphometry and epiphyseal growth plate (EGP) morphometry were performed in the distal femur. Expressions of IGF-I protein and of collagen II and X mRNA were evaluated by immunohistochemistry and in situ hybridization, respectively. To determine hormonal effects on ossification, skeletal preparations of hypothyroid-, 5xGC-1-, and 5xT3-treated neonatal rats were compared.

Results: Hypothyroidism impaired longitudinal body growth and BMD gain, delayed ossification, reduced the number of hypertrophic chondrocytes (HCs; 72% versus Euthyroid ‘Eut’ rats; p < 0.001), and resulted in disorganized columns of EGP chondrocytes. Serum IGF-I was 67% reduced versus Eut rats (p < 0.001), and the expression of IGF-I protein and collagen II and X mRNA were undetectable in the EGP of Hypo rats. T3 completely or partially normalized all these parameters. In contrast, GC-1 did not influence serum concentrations or EGP expression of IGF-I, failed to reverse the disorganization of proliferating chondrocyte columns, and barely affected longitudinal growth. Nevertheless, GC-1 induced ossification, HC differentiation, and collagen II and X mRNA expression and increased EGP thickness to Eut values. GC-1-treated rats had higher BMD gain in the total tibia, total femur, and in the femoral diaphysis than Hypo animals (p < 0.05). These changes were associated with increased trabecular volume (48%, p < 0.01), mineralization apposition rate (2.3-fold, p < 0.05), mineralizing surface (4.3-fold, p < 0.01), and bone formation rate (10-fold, p < 0.01).

Conclusions: Treatment of hypothyroid rats with the TRβ-specific agonist GC-1 partially reverts the skeletal development and maturation defects resultant of hypothyroidism. This finding suggests that TRβ1 has an important role in bone development.


THYROID HORMONE IS essential for normal bone development and metabolism. In hyperthyroidism, the activity of osteoblasts and osteoclasts are increased; however, the latter predominates favoring resorption, negative balance of calcium, and bone loss.(1) During development, hypothyroidism causes a generalized delay in endochondral and intramembranous ossification in addition to important alterations in the epiphyseal growth plates (EGPs), such as reduced thickness, disorganized columns of chondrocytes, and impaired differentiation of hypertrophic chondrocytes,(2, 3) resulting in reduced growth and skeletal abnormalities.(4) On the other hand, thyroid hormone excess results in accelerated skeletal maturation, with premature closure of the EGPs and subsequent decreases in longitudinal bone growth.(4–6)

Thyroid hormone directly affects bone development but also acts indirectly through the growth hormone (GH) and insulin-like growth factor-I (IGF-I) axis.(7, 8) GH gene transcription is profoundly regulated by triiodothyronine (T3) through its interaction with the thyroid hormone receptor-thyroid hormone responsive element complex (TR-TRE) in the GH promoter.(9, 10) Thus, the adverse effects of hypothyroidism on skeletal development are accompanied by inhibition of the GH/IGF-I axis.(11) Nevertheless, an important and unresolved issue is whether the skeletal effects of T3 result primarily from direct T3 actions in bone cells or through interactions with the GH/IGF-I axis. Recent evidence suggests that direct effects are likely to be critical to bone development. It was shown, for instance, that GH, without T3, is unable to stimulate the maturation(12) and organization(2) of growth plate chondrocytes and that GH replacement does not rescue the ossification defect seen in TRα and TRβ-null mice.(8) Furthermore, in primary cultures of growth plate chondrocytes, T3 inhibits cell proliferation while concurrently inducing hypertrophic chondrocyte differentiation.(13)

T3 actions are mediated by nuclear receptors (TRs), which are T3-inducible transcriptional factors(14) encoded by two genes,(15)TRα and TRβ. The TRα gene encodes TRα1 and TRα2 and two truncated forms, Δα1 and Δα2.(16) TRα2 and the truncated isoforms act as antagonists of TRα1 and TRβ receptors in vitro, although their physiological roles are unknown, and the antagonistic effects of TRα2 are weak.(14, 16, 17) The TRβ gene encodes TRβ1, TRβ2, TRβ3, and a dominant negative truncated isoform TRΔβ3.(14, 18) TRβ2 is restricted mainly to pituitary, hypothalamus, ear, and retina and controls the pituitary-thyroid axis and development of the inner ear and retina.(19) The other TRβ isoforms are more widely distributed, although isoform-specific functions have not yet been identified. TRα1, TRα2, and TRβ1 are expressed in osteoblasts,(20) osteoclasts,(21) and chondrocytes.(22) Nevertheless, the functional roles of each TR isoform in the skeleton are incompletely understood.

Animal models in which TR genes were targeted for disruption have brought important insights into the knowledge of the role of TRs in mediating T3 effects of skeletal development.(19) It was shown that the TRβ(23) and the TRα1(24) knockout (KO) mice do not present defects of bone development. However, the absence of both receptors, seen in the double KO mice TRα1−/−TRβ−/−,(25) results in a number of defects, including delayed ossification and dysgenesis of EGPs. The findings of these studies suggest a substantial overlap and, therefore, functional redundancy of the TRα1 and TRβ isoforms on bone development. TRα−/− KO mice, which do not express TRα1 and TRα2 but continue to express the Δα1 and Δα2 isoforms, also present defects of bone development; however, these animals are hypothyroid.(26) On the other hand, TRα0/0 mice, which lack all TRα isoforms, including Δα1 and Δα2, are biochemically euthyroid and show a less severe phenotype of delayed bone development than TRα−/− mice, suggesting a physiological role for non-T3-binding TRα variants in bone development.(27) Indeed, Pax-8−/− mice, which express all TR isoforms but lack the follicular cells producing thyroxine (T4) and T3 in the thyroid gland, present even more pronounced defects in bone development than mice devoid of all TRs (TRα0/0TRβ−/−).(28) This finding has been interpreted to result from repressor effects of the unliganded TRs (aporeceptors) on thyroid hormone responsive genes during development of key T3-target organs such as bone. In support of this, Pax-8−/−TRα0/0, but not Pax-8−/−TRβ−/−, compound mutants present a partial rescue of the bone phenotype, reinforcing the physiological importance of TRα1 aporeceptor activity during bone development.(19, 28)

Recently, GC-1, a synthetic analog of thyroid hormone that is selective for both binding and activation functions of TRβ1 over TRα1, was developed.(29) GC-1 contains several structural changes with respect to the natural hormone T3, which include replacement of the three iodines with methyl and isopropyl groups, replacement of the biaryl ether linkage with a methylene linkage, and replacement of the amino-acid side-chain with an oxyacetic-acid side-chain.(29) GC-1 binds TRβ1 with the same affinity as T3 but binds TRα1 with an affinity that is 10 times lower compared with T3. Accordingly, GC-1 presents selective actions in vitro(29) and in vivo.(30, 31) The TRβ selectivity of GC-1 is particularly interesting and makes GC-1 a useful tool to probe the functional roles of TR isoforms, especially in physiological situations where all endogenous TRs are intact. The differential effects of GC-1 in the central nervous system,(32–34) brown adipose tissue,(31) and tadpole metamorphosis(35) correspond with the specific gene recognition by individual TR isoforms within a tissue. On the other hand, GC-1 effects are also related to body distribution of TR isoforms. GC-1 has no effect on heart function, which expresses mainly TRα1, but decreases serum levels of cholesterol and triglycerides, which agrees with a higher expression of TRβ1 in the liver.(30) In a recent study, we showed that high doses of T3 cause generalized bone loss in rats, whereas equivalent treatment with GC-1 spares the bone mass, which suggests that TRα1 is the main isoform to mediate the osteopenic effects of T3.(36) This is also consistent with data showing that TRα1 mRNA is 10–12 times higher than TRβ1 mRNA in the femur and tibia of mice.(5) Nevertheless, the co-expression of both TRα1 and TRβ1 in bone and cartilage and evidence of a compensatory role for TRβ1 in the skeleton of TRα1-deficient mice is all suggestive of a physiological role for TRβ1 as well as TRα1 in bone.

Thus, in this study, our goal was to investigate the role of TRβ1 in the bone development of rats using the TRβ-selective agonist GC-1 as a pharmacological tool.


Animals and drugs

All experimental procedures were performed in accordance with the guidelines of the Standing Committee on Animal Research of the University of Sao Paulo. In a first set of experiments, 36 female Wistar rats were obtained from our breeding colony and maintained under controlled conditions of light and temperature (12-h/12-h dark/light cycle at 25°C). All animals were kept in plastic cages, six per cage, and had free access to food (rat chow containing 1.4% Pi, 0.7% Ca, and 4.5 IU/g of vitamin D) and water. At the age of 21 days and weighing 56 ± 4 g, 36 rats were randomly divided into six groups (n = 6 per group): (1) euthyroid (Eut); (2) hypothyroid (Hypo), hypothyroidism was induced by maintaining the animals on metimazole (MM; 0.1%) and sodium perclorate (P; 1%) in the drinking water; (3) 1xT3, hypothyroid animals receiving 0.3 μg/100 g BW of T3 per day (intraperitoneally), which is equivalent to a physiological dose of T3(37); (4) 5xT3, hypothyroid animals receiving 1.5 μg/100 g BW; (5) 1xGC-1; and (6) 5xGC-1. The latter two groups include hypothyroid animals receiving GC-1 in 1xT3 and 5xT3 equimolar doses (0.15 and 0.75 μg/100 g BW/day, respectively). The equimolar doses of GC-1 were calculated from the molecular weight of T3 (molecular weight = 651) and GC-1 (molecular weight = 328.4). T3 and GC-1 treatments started after 1 week of hypothyroidism induction and lasted 5 weeks. Body weights were measured three times a week, and the amount of hormone administered was adjusted in accordance with the changes in body weight, to maintain the proper dosage. Because the effects of 1xT3 versus 5xT3 and 1xGC-1 versus 5xGC-1 were very similar on all parameters of bone development, only the results of 5xT3 and 5xGC-1 treatments will be presented in this study. In a second set of experiments, neonatal Wistar rats were divided in four groups (n = 3/group): (1) Eut; (2) Hypo, hypothyroidism was induced by administration of MM+P in the drinking water of the dams of the neonatal rats for two weeks; (3) 5xT3; and (4) 5xGC-1 dams given MM+P and neonates left untreated for 1 week and given T3 or GC-1 (subcutaneous injections) for 1 week. At 2 weeks of age, all animals were killed, and the skeletons were prepared for the analysis of ossification.

Serum parameters

At the end of the experimental period, the animals were killed by decapitation, and the blood of the trunk was collected. The serum was separated by centrifugation and immediately frozen. Thyroid-stimulating hormone (TSH) and IGF-I were measured using radioimmunoassay (RIA) kits specifically designed for rat TSH (Biotrak; Amersham Pharmacia Biotech, Piscataway, NJ, USA) and rat IGF-1 (Diagnostic Systems Laboratories, Webster, TX, USA). Total T4 and T3 serum levels were measured by commercial RIA kits (RIA-gnost T4 and RIA-gnost T3; CIS bio international). For the T4 and T3 assays, standard curves were generated using charcoal-stripped rat serum. To test if GC-1 could cross-react in the T4 and T3 RIAs, serum levels of T4 and T3 were measured in hypothyroid rats treated with different doses of GC-1. Serum T4 was undetectable in all animals, and there were no differences in serum levels of T3 between the GC-1-treated animals and the hypothyroid-untreated animals, which indicate that T4 and T3 RIAs cannot detect GC-1.

Bone densitometry

BMD was measured by DXA using the pDEXA Sabre Bone Densitometer and the pDEXA Sabre Software version 3.9.4 (Norland Medical Systems, Fort Atkinson, WI, USA), both especially designed for small animals. The research mode scan option was used for the measurements. Pixel spacing for the scan was set to 0.5 × 0.5 mm and the scan speed to 4 mm/s. To limit the scan area, which allows the scans to be performed in a higher resolution mode, the scans were performed from the first lumbar vertebra to the hind limbs. The regions of interest (ROIs) for analysis were (1) hind body (HB), which includes L2-L6, pelvic bones, hind limbs, and the first four caudal vertebrae; (2) lumbar vertebrae (L2-L5); (3) both femurs; and (4) both tibias. Considering the different proportions of cortical bone and trabecular bone in different regions of femur and tibia, the BMD of these bones was also analyzed as three segments: proximal and distal regions (each one determined by one-third of the total bone length) and diaphysis, the region between the proximal and distal regions. Taking into account the possibility of bone mass differences between the left and right limbs as a result of functional bilateral asymmetry,(38) the BMD of femur and tibia were expressed as the mean of the left and right limbs for each animal. For the scans, the animals were anesthetized with a ketamine-xylazine cocktail (10 and 30 mg/kg BW) and scanned in the prone position. The animals were scanned 1–2 days before hypothyroidism induction and after 6 weeks (after 5 weeks of treatment).

Body longitudinal growth

The lengths of femur, tibia, and lumbar spine were measured indirectly by DXA using the ruler tool provided by the pDEXA Sabre Software. The length of the femur and tibia was measured from the proximal to the distal epiphysis. The length of the lumbar spine was measured from the proximal end of L2 to the distal end of L6. The body length was measured directly with a ruler (Norland Medical Systems) from the tip of the snout to the base of the tail. The longitudinal growth (LG) of each segment was calculated by the difference of the basal to the final measurements (LG = final length − basal length).

Bone histology and morphometry

The fourth lumbar vertebra (L4) and the right femur were dissected and fixed in 4% paraformaldehyde for 1 week at room temperature. The bones were demineralized in 4% EDTA, pH 7.2, at room temperature, under agitation, for 4–5 weeks. Samples were dehydrated and embedded in Paraplast (Oxford, St Louis, MO, USA). Five-micrometer-thick sections were stained with H&E. The histological sections were photographed under light microscopy and analyzed. Morphometric measurements were performed with the aid of the image analyzer Image-pro Plus (Media Cybernetics, Silver Spring, MD, USA). The thickness of the distal EGP of the femur and of the reserve zone (RZ), proliferative zone (PZ), and hypertrophic zone (HZ) were determined at five equidistant points over the width of the GP. The number of proliferative chondrocytes (PC) was determined per column of chondrocytes and the number of hypertrophic chondrocytes (HC) was determined in a total of five fields and expressed per millimeter squared of GP.

Immunohistochemistry and in situ hybridization

IGF-I protein expression was determined using a mouse monoclonal antibody against human IGF-I (Upstate Biotechnology, Lake Placid, NY, USA). Sections were treated with 3% (v/v) H2O2 (Merck, Darmstadt, Germany) in PBS for 1 h to block endogenous peroxidase activity. Nonspecific reaction was blocked by incubating the sections for 30 minutes with normal horse serum diluted 1:1 in PBS-1% bovine serum albumin followed by incubating with SuperBlock buffer solution (Pierce, Rockford, IL, USA) for 1 h. Sections were incubated with IGF-1 antibody diluted in 1:100 PBS-0.3% (v/v) Tween 20 overnight at 4°C. The sections were washed thoroughly with PBS followed by incubation with biotinylated horse antimouse IgG (Vector Laboratories, Burlingame, CA, USA) diluted 1:1000 in PBS for 1 h at room temperature. After extensive rinsing in PBS, the sections were treated with streptavidin/peroxidase using the Vectastain ABC kit (Vector Laboratories) for 1 h at room temperature. The reaction was visualized using 0.03% (w/v) 3,3′-diaminobenzidine (Sigma) plus 0.03% (v/v) H2O2 in PBS. After immunostaining, sections were lightly stained with Mayer's hematoxylin. The control sections were performed by omitting the primary antibody. Trachea sections were used as positive controls. mRNA expression in EGP sections was analyzed by in situ hybridization as previously described(3, 5) using collagen II (coll II) and collagen X (coll X) cRNA probes. A bacterial neomycin resistance gene cRNA probe (Roche Molecular Biochemicals, Lewes, Sussex, UK) was used as a negative control for all hybridizations.(3) Rat coll II (nucleotides 2982–3689; GenBank accession no. L48440) and coll X (nucleotides 418–858; GenBank accession no. AJ131848) partial cDNA was isolated by RT-PCR as described(3) using RNA from chondrogenic FTC5:3 cells(39) and osteoblastic ROS 17/2.8 cells.(20) PCR products were subcloned into the pGEM-T vector (Promega) and sequenced. Coll II and coll X constructs were linearized with NcoI and SpeI, and digoxigenin-labeled antisense cRNA probes were synthesized using SP6 and T7 RNA polymerases (Roche Molecular Biochemicals).

Bone histomorphometry

At the end of the experimental period, the left distal femur was carefully dissected out and processed as previously described.(40) Five-micrometer sections were stained with 0.1% toluidine blue, pH 6.4, and at least two nonconsecutive sections were examined for each sample. Static and structural parameters of bone formation and resorption were measured at the distal metaphyses (magnification, ×250), 195 μm far from the EGP, in a total of 30 fields, using a semiautomatic method (Osteometrics, Atlanta, GA, USA). For dynamic histomorphometric analysis of bone formation, all animals were double-labeled with tetracycline (Terramycine; Pfizer, Sao Paulo, Brazil) administered intraperitoneally in a dose of 20 mg/kg BW 19 and 18 days before death, for the first labeling, and 2 and 1 days before death, for the second labeling. Kinetic bone parameters were obtained from unstained 10-μm sections examined by fluorescent light microscopy (Nikon, Tokyo, Japan). The mineralization apposition rate (MAR) was expressed in micrometers per day. The mineralizing surface was expressed as a percentage of total bone surface (MS/BS; %). The bone formation rate was expressed per unit of bone surface (BFR/BS; μm3/μm2/year). The static histomorphometric indices evaluated were trabecular volume (BV/TV; %), trabecular thickness (Tb.Th; μm), trabecular separation (Tb.Sp; μm), trabecular number (Tb.N; mm−1), eroded surface (ES/BS; %), number of osteoclasts per bone surface (Oc.S/BS; %), number of osteoblasts per bone surface (Ob.S/BS; %), osteoid volume (OV/BV; %), osteoid thickness (O.Th; μm), and the osteoid surface per bone surface (OS/BS; %). All histomorphometric indices were reported according to the standardized nomenclature recommended by the American Society of Bone and Mineral Research.(41) All animal data were obtained by blind measurements.

Skeletal preparations

Two-week-old rats were eviscerated and had the skin and muscles removed. After a 3-h fixation in methacarn (60% methanol, 30% chloroform, 10% acetic acid), they were transferred to 100% ethanol overnight and 95% ethanol overnight. Specimens were stained in Alcian blue solution (150 mg alcian blue, 80 ml ethanol, 20 ml acetic acid) for 48 h. After 72 h in 95% ethanol, they were transferred to 1% KOH for 24 h. Skeletons were stained in alizarin red solution (50 mg/liter alizarin red in 2% KOH) for several hours and cleared in 1% KOH/20% glycerol solution. Stained skeletons were stored in 50% ethanol/50% glycerol.

Statistical analysis

In a previous study, we found that rat bone mass varies according to a normal Gaussian distribution,(40) which allowed us to use parametric statistical tests for the analysis of BMD. One-way ANOVA was used to compare more than two groups and was always followed by the Student-Newman-Keuls test to detect differences between groups. To analyze the histomorphometric results expressed as percentages, we used the Kruskal-Wallis nonparametric ANOVA test, followed by the Dunn's test. For all tests, p < 0.05 was considered statistically significant. All results are expressed as the mean ± SE. For statistical analysis, we used the GraphPad Instat Software (GraphPad Software, San Diego, CA, USA).


T3- and GC-1-induced effects on serum parameters

In all MM+P-treated animals, serum levels of T4 were undetectable and serum levels of TSH were elevated, except in the T3-treated rats (Table 1). In Hypo rats, serum TSH was 7-fold higher than in Eut animals (p < 0.001). 5xT3 treatment resulted in serum levels of T3 that was 3.7-fold higher than in Eut (p < 0.01) and suppressed TSH in 42.4% (p < 0.001 versus Eut; Table 1). The doses of GC-1 used in this study had no effect on serum levels of T3 and were unable to suppress TSH; however, they slightly reduced serum levels of TSH by 3.5-4.6% compared with Hypo animals (p < 0.05). As expected, serum levels of IGF-I (Table 1) were significantly lower in Hypo animals (3.1-fold versus Eut, p < 0.001). Replacement therapy with 5xT3 significantly increased serum IGF-I compared with Hypo animals (∼2-fold, p < 0.001), but was unable to completely reverse the IGF-I deficiency (∼1.6-fold lower than Eut, p < 0.001). GC-1 treatment had no effect on serum levels of IGF-I.

Table Table 1. Serum Parameters
original image

GC-1 barely affects longitudinal growth

As expected, the longitudinal growth of the body, tibia, femur, and lumbar spine were decreased, in a range of 66–82%, in the Hypo group, compared with Eut animals (Table 2). The replacement therapy with T3 resulted in significant increases in the rates of longitudinal growth (final length − basal length); however, these rates were still significantly lower than those of Eut animals, which resulted in lower final lengths of all skeletal segments. The total body growth, however, was completely restored by treatment with 5xT3. GC-1 treatment did not have any significant effect on bone longitudinal growth, except in the femur and lumbar spine where 5xGC-1 slightly induced growth (1.6- and 1.9-fold versus Hypo animals, respectively).

Table Table 2. Longitudinal Growth of the Body and Skeletal Segments
original image

GC-1 had selective effects on the epiphyseal growth plate

The thickness of the EGP of the distal femur of Hypo rats decreased 43% compared with that of the Eut rats (Table 3; Figs. 1A and 1B). This resulted from reductions of 45% and 62% in the thickness of the PZ and HZ, respectively. The columns of chondrocytes were disorganized (Figs. 1A and 1B), and the numbers of PC and HC were 34% and 72% reduced compared with Eut rats, respectively (Table 3). The T3 replacement therapy completely reversed all of these alterations (Table 3; Fig. 1C). In contrast, GC-1 treatment partially restored the hypothyroidism-induced defects in the EGP. 5xGC-1 therapy resulted in recovery of the thickness of the EGP, PZ, and HZ. This was followed by a significant but incomplete increase in the number of PC (28%) and HC (163%) compared with Hypo animals (Table 3; Fig. 1D). GC-1 treatment, however, failed to induce organization of the columns of PC in the EGP (Fig. 1D). To better characterize the effect of T3 and GC-1 on PC and HC, the expression Coll II and Coll X mRNAs, which are specific markers of these cells, respectively, were detected by in situ hybridization. In the EGPs of Hypo animals, Coll II and Coll X mRNAs were barely expressed. However, in Eut and 5xT3-treated animals, Coll II mRNA was expressed mainly throughout the PZ. In GC-1-treated animals, Coll II mRNA was restricted to prehypertrophic proliferating chondrocytes. It is noteworthy that the expression of Coll X mRNA, a specific marker of HC,(42) was detected predominantly in HZ of Eut-, T3-, and GC-1-treated animals, suggesting that GC-1 was able not only to induced morphological differentiation but also functional differentiation of PC to HC. The immunohistochemical localization of IGF-I in the EGP of the distal femur showed that IGF-I was expressed in the PZ and HZ of the Eut animals and undetectable in the EGPs of Hypo rats and GC-1-treated rats (Figs. 1E, 1F and 1H), which suggests that the GC-1 effect on HC differentiation was not IGF-I mediated. In the T3-treated animals, IGF-I was expressed in the PZ but not in the HZ (Fig. 1G). These findings are in keeping with the localization of TRs in the PZ but not HZ of the EGP(13) and suggest that T3, but not GC-1, exerts direct actions on proliferating chondrocytes to induce IGF-I expression.

Table Table 3. Morphometry of the Distal Femur EGP
original image
Figure Fig. 1..

Sections of the distal femur from (A and E) Eut, (B and F) Hypo, (C and G) 5xT3-treated, and (D and H) 5xGC-1-treated rats showing the growth plate zones. RZ, reserve zone; PZ, proliferative zone; HZ, hypertrophic zone; PS, primary spongiosum. A-D show H&E staining. E-H show immunolocalization of IGF-I. Magnification, ×200.

GC-1 moderately induced ossification

The histological analysis of the proximal epiphysis of L4 showed that Hypo animals presented a delayed ossification in this skeletal region (Fig. 2). Whereas epiphyseal trabecular bone (ETB) is present in Eut animals, it is absent in Hypo animals (Figs. 2A and 2B). T3 treatment completely reverted this defect but not GC-1 (Figs. 2C and 2D). Similar to Hypo animals, GC-1-treated rats did not show ETB in this region. However, a more careful observation revealed the presence of a greater number of HC in the proximal epiphysis of L4 of GC-1-treated animals than in that of Hypo animals (Figs. 2B and 2D), which suggested that the ossification process was more advanced in the GC-1-treated animals. To further investigate the effect of GC-1 on ossification, the skeletons of 2-week old hypothyroid rats treated with 5xGC-1 or 5xT3 every day for 1 week before death were stained with alizarin red and alcian blue. As expected, Hypo animals presented a generalized delay in endochondral ossification. As shown in Figs. 2F and 2J, there was a clear ossification delay in the tarsal bones and in the epiphysis of tibia, femur, and metatarsals bones of Hypo rats (Figs. 2G and 2K). 5xT3 treatment almost completely reverted the delays in ossification of all skeletal regions. GC-1 clearly induced endochondral ossification, however, the T3 effect on ossification was more prominent in all skeletal sites. The GC-1 effect on endochondral ossification is very clear in the distal epiphysis of the femur and in the proximal epiphysis of the tibia, where larger ossification centers can be seen in GC-1-treated animals compared with those of the Hypo animals (Figs. 2F and 2H). Note that ossification centers are present in the distal epiphysis of the metatarsal bones (II-IV), in the navicular and in the intermediate cuneiform of GC-1 rat, but absent in the Hypo rat (Figs. 2J and 2L). In addition, the ossification of the cuboid and the medial and the lateral cuneiforms are advanced in GC-1 animal compared with that in Hypo rat. Advanced ossification in GC-1 and T3 animals compared with Hypo animas was also clearly observed in the carpal and long bones of the forelimbs (data not shown). However, the effect of GC-1 treatment on ossification was intermediate between hypothyroidism and T3 treatment in all bone segments evaluated.

Figure Fig. 2..

Analysis of endochondral ossification in the skeleton of (A, E, and I) Eut, (B, F, and J) Hypo, (C, G, and K) 5xT3-treated, and (D, H, and L) 5xGC-1-treated rats. A-D show the proximal end of L4 (magnification, ×50); (B) Hypo and (D) GC-1 vertebrae do not present epiphyseal trabecular bone (ETB); details (magnification, ×100) show hypertrophic chondrocytes in the region between the growth plate (GP) and the articular cartilage (AC); H&E staining. E-H show the distal and proximal epiphyses of the femur and tibia, respectively. I-L show hind paws stained with alcian blue and alizarin red (cartilage stains blue and bone stains red); arrows indicate ossification centers. F, femur; T, tibia; M, metatarsus (I-V); mC, iC, and lC, medial, intermediate, and lateral cuneiform; Cu, cuboid; N, navicular.

GC-1 slightly induced bone mass gain

Hypothyroidism significantly impaired the gain of bone mass (ΔBMD) in all regions of the skeleton that were analyzed. ΔBMD was on average 3.8-fold lower in Hypo animals than in Eut animals (Table 4). The region that was most affected by thyroid hormone deficiency was the femoral diaphysis (5.3-fold lower than Eut, p < 0.001) and the least affected region was the distal tibia (2.6-fold lower than Eut, p < 0.001). Excluding L2-L5, T3 treatment significantly increased the gain of bone mass in every skeletal site compared with Hypo animals. ΔBMD was on average 2-fold greater in T3-treated rats than in Hypo rats; however, ΔBMD was on average 1.8-fold lower than in Eut animals. The femoral diaphysis was also the most responsive skeletal site to treatment with T3 (2.8-fold higher than Hypo, p < 0.001) and the lumbar spine was the least responsive region (1.7-fold higher than Hypo, p > 0.05). GC-1 had a small positive effect on bone mass. On average, the GC-1-treated animals presented ΔBMDs that were 1.4-fold greater than those of the Hypo animals. GC-1 only significantly increased the ΔBMD in the total femur (1.6-fold versus Hypo, p < 0.05), femoral diaphysis (1.9-fold versus Hypo, p < 0.01), total tibia (1.5-fold versus Hypo, p < 0.05), and tibial diaphysis (1.6-fold versus Hypo, p < 0.05).

GC-1 selectively induced changes in bone histomorphometric parameters of trabecular bone

In Hypo animals (Table 5), BV/TV was 50% lower than in Eut animals (p < 0.001). It was associated with reduced Tb.Th (−18% versus Eut, p < 0.01) and Tb.N (−38% versus Eut, p < 0.01) and with a 2-fold increase in Tb.Sp (p < 0.05). The treatment of hypothyroid animals with both T3 and GC-1 partially or totally restored BV/TV and Tb.Th and significantly reduced Tb.Sp to near Eut values, but did not restored Tb.N (Table 5). All bone formation parameters were reduced in Hypo animals in a range of 13–84% (Table 5); however, only osteoblast number, accessed by Ob.S/BS, was significantly decreased compared with Eut animals (84%, p < 0.01). Treatment with T3 had a positive effect on osteoblast number. In the 5xT3 group, Ob.S/BS was significantly higher than in Hypo (17-fold, p < 0.001) and Eut animals (1.9-fold, p < 0.01). OV/BV and OS/BS were also significantly increased over the Eut values (in a range of 2.7- to 3.6-fold, p < 0.01) in the T3-treated animals. On the other hand, GC-1 failed to increase all these parameters (Ob.S/BS, OV/BV, and OS/BS) above the Hypo values. The dynamic parameters of bone formation (MAR, MS/BS, and BFR/BS) were significantly increased by T3 treatment above the Eut values in a range of 2.6-5.2 times. GC-1 treatment also had positive and significant effects on these parameters. MAR was significantly higher in 5xGC-1−treated rats than in Hypo rats (2.1-fold, p < 0.05). Likewise T3, 5xGC-1 increased MS/BS above the Eut values (2.9-fold, p < 0.001). 5xGC-1 animals also presented significantly higher BFR/BS compared with Hypo (10-fold, p < 0.001) and Eut animals (3.5-fold, p < 0.05). The bone resorption markers (Table 5), Oc.S/BS and ES/BS, were significantly lower in Hypo animals compared with Eut animals (84% and 74%, respectively). 5xT3 animals presented Oc.S/BS and ES/BS 7.4- (p < 0.01) and 5.6-fold (p < 0.001), respectively, greater than Hypo animals. Compared with Hypo animals, these two bone resorption parameters were also significantly increased by 5xGC-1 treatment (4.8- and 3.4-fold, respectively).

Table Table 5. Effect of GC-1 and T3 on Histomorphometric Variables of Trabecular Bone of the Distal Femur
original image


We used the TRβ-selective agonist GC-1 as a pharmacological tool to investigated the role of TR isoforms in mediating T3 effects on bone development of rats. Hypothyroid rats were treated with T3 or with equimolar doses of GC-1 and the skeletal development was analyzed. As expected, hypothyroidism resulted in low serum levels of IGF-I, which is in keeping with the known suppression of the GH/IGF-I axis that accompanies hypothyroidism.(11, 43) This was followed by a severe skeletal phenotype of growth retardation, decreased bone mass, and a generalized defective endochondral ossification. Whereas T3 treatment almost completely reverted all the skeletal defects induced by thyroid hormone deficiency, GC-1 treatment had only partial effects (Fig. 3).

Figure Fig. 3..

Diagram comparing the main effects of GC-1 and T3 on bone development. EGP, epiphyseal growth plate; PZ, proliferative zone; HC, hypertrophic chondrocytes. −, no effect; +, effect; −/+, effect and no effect depending on skeletal site or bone parameter.

An important effect of GC-1 was the induction of endochondral ossification. Considering the selectivity of GC-1, our findings indicate that TRβ1 isoform mediates T3 effects on ossification. However, because the effect of GC-1 was less accentuated than either euthyroidism or the effect of T3, it is likely that the normal induction of endochondral ossification by T3 requires actions mediated by both TRα1 and TRβ1. The fact that, differently of T3, GC-1 was not able to increase serum levels of IGF-I or the protein expression of IGF-I in the growth plates suggests a direct effect of thyroid hormone on ossification. The skeletal phenotype of TRα1−/−β−/− mice, which present growth retardation and defective ossification of the epiphyses associated with inhibition of the GH/IGF-I axis, reinforce the view that T3 directly affects ossification. In these animals, it was shown that GH substitution reverses the growth phenotype but not the defective ossification, suggesting that TRs are essential for the direct effects of T3 on ossification.(8) On the other hand, a synergist or additive action between GH/IGF-I and thyroid hormone in endochondral ossification can not be excluded, because mice with inactivated GH or IGF-I receptors present delayed ossification.(44, 45) Further studies are necessary to clarify the participation of TRα1 and GH/IGF-I in this process.

Accordingly to previous findings,(2, 3) we showed that the PZ was disorganized and the HC differentiation was blocked in the distal femur of Hypo rats, leading to EGPs nearly absent of chondrocytes with hypertrophic features. It is remarkable that, whereas T3 reverted these growth plate defects, GC-1 induced morphological (increased number of HC and HZ thickness) and molecular (expression of Coll X mRNA) changes indicative of HC differentiation but did not induce the organization of the columns of PC (Fig. 1; Table 3). It is noteworthy that GC-1 did not induce the expression of IGF-I protein in the EGP, which suggests that the GC-1-induced effects in the growth plate were not IGF-I mediated. Considering that the HC differentiation and the PZ organization are essentially regulated by thyroid hormone,(2) our findings suggest that TRβ1 mediates the T3 induction of HC differentiation, whereas TRα1 mediates T3 effects on the columnar organization of the PC. Despite the inability of GC-1 to organize the PZ, GC-1 induced the expression of Coll II in prehypertrophic chondrocytes, and similarly to T3, rescued the thickness of the PZ and increased the number of cells per column of PC (28% above the Hypo values), which suggests a positive effect on the proliferation of growth plate chondrocytes in vivo. Similar effects of thyroid hormone were observed in the EGP of the proximal tibia of hypothyroid rats.(2) This effect, however, contrasts with the clear inhibitory effect of T3 on chondrocyte proliferation in vitro(13, 46) and probably is the result of complex systemic and local interactions that certainly occur in vivo. Another point that deserves attention is the fact that GC-1 did not induce longitudinal growth of the body and tibia and barely promoted femoral and vertebral longitudinal growth regardless of its clear effects in the EGP. This is probably and partially explained by the deficiency of serum IGF-I in GC-1-treated rats. It was previously shown that circulating levels of IGF-I seem to play a major role on bone linear growth.(47) However, classical studies show that T4 alone stimulates longitudinal bone growth in hypophysectomized rats,(48, 49) besides the well known T4 function of potentiating the growth promoting effect of GH/IGF-I.(50) Therefore, the absence or minimal growth observed in GC-1-treated animals may also be the result of the lack of TRα1-mediated actions of T3 or be caused by repression by TRα1 aporeceptor activity. Further studies, using GH and/or IGF-I in combination with GC-1 will be important to clarify this hypothesis. Nevertheless, all these findings suggest that both TRα1 and TRβ1 are important for the EGP architecture and functionality and that some TRα1-mediated actions of T3 can not be compensated by TRβ1 mediation. This is consistent with the expression of both isoforms in the EGP of rats.(13)

An important finding of this study was the striking effect of hypothyroidism of impairing the gain of bone mass during the skeletal development of rats. The femoral diaphysis, which is composed basically of cortical bone, was the most affected region by thyroid hormone deficiency (Table 4). This may reflect serum IGF-I deficiency. It was shown that circulating IGF-I is critical for bone mass acquisition, particularly cortical bone.(47) However, trabecular bone was also affected by hypothyroidism. The histomorphometric analysis of the distal femur showed that Hypo rats presented a lower trabecular bone volume (BV/TV) that was associated with a lower trabecular thickness and number. This was followed by a generalized decrease in bone resorption parameters and by a remarkable decrease (9.2-fold versus Hypo) in the number of osteoblasts (Ob.S/BS). The T3 replacement therapy increased practically all the bone formation and bone resorption parameters above the Eut values, which indicates an increase in bone turnover and probably partially explains why T3 treatment did not restore the BMD gain to the Eut level. The treatment of hypothyroid rats with GC-1 had a small but significant and positive effect on ΔBMD of the total femur, femoral diaphysis, total tibia, and tibial diaphysis. It suggests that thyroid hormone has anabolic effects on bone mass that are TRβ1-mediated. In fact, GC-1 was able to increase the trabecular bone volume (BV/TV) of the distal femur beyond that of the Hypo rats. This was accompanied by a positive effect of GC-1 on the dynamic parameters of bone formation (MAR, MS/BS, and BFR), which indicates that the activity of the existing osteoblasts was increased by GC-1. These results are consistent with direct stimulatory effects of T3 on osteoblasts in vitro(51) and suggest that TRβ1 mediates some of the T3 effects on osteoblastic activity. On the other hand, contrary to T3, GC-1 was unable to increase Ob.S/BS, which suggests that TRα1 but not TRβ1 activity is important to sustain the proper number of osteoblasts during development.

Table Table 4. Effect of GC-1 and T3 Treatment on BMD
original image

An issue to be considered is the possibility that GC-1, at the doses used (1xGC-1 and 5xGC-1), could also bind TRα1 isoform. However, in previous studies, similar or higher doses of GC-1 were used, and its TRβ-selectivity was maintained.(30–33, 36) In addition, the fact that some effects of GC-1 on bone physiological and structural endpoints occurred at the same magnitude than those induced by T3, whereas some T3 effects on bone could not be rescued by GC-1 treatment suggest that the TRβ selectivity was maintained. Another point to be considered is the fact that treatment of hypothyroid rats with GC-1 activates TRβ1 but leaves TRα1 unoccupied. Thus, the final skeletal responses of GC-1 should be considered to be the result of TRβ1 activation plus the aporeceptor activity of TRα1. It may explain why some GC-1 effects on bone are partial and highlight the importance of TRα1 activity to bone development.

In conclusion, chronic treatment of hypothyroid rats with GC-1 partially reverts the skeletal development and maturation defects resultant of the thyroid hormone deficiency. Considering that GC-1 is selective for TRβ1 over TRα1, these findings suggest that TRβ1 has an important role in bone development. On the other hand, our findings show that TRβ1-mediated actions cannot completely compensate the lack of TRα1-mediated actions on this process, which suggests the importance of both receptors to skeletal development.


This work was supported by a grant from Fundação de Amparo à Pesquisa do Estado de São Paulo, Brazil (CHAG) and an MRC Career Establishment Grant (G9803002) (GRW).