Thyroid Hormones Regulate Hypertrophic Chondrocyte Differentiation and Expression of Parathyroid Hormone-Related Peptide and Its Receptor During Endochondral Bone Formation†
David A. Stevens,
Imperial College School of Medicine Molecular Endocrinology Group, Division of Medicine and Medical Research Council Clinical Sciences Center, Imperial College School of Medicine, Hammersmith Hospital, London, United Kingdom
Imperial College School of Medicine Molecular Endocrinology Group, Division of Medicine and Medical Research Council Clinical Sciences Center, Imperial College School of Medicine, Hammersmith Hospital, London, United Kingdom
Address reprint requests to: Dr. Graham R. Williams, ICSM Molecular Endocrinology Group, MRC Clinical Sciences Center, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London W12 0NN, U.K.
This paper was presented in part at the 72nd Annual Meeting of the American Thyroid Association, Palm Beach, Florida, U.S.A., 1999
Hypothyroidism in children causes developmental abnormalities in bone and growth arrest, while thyrotoxicosis accelerates growth rate and advances bone age. To determine the effects of thyroid hormones on endochondral bone formation, we examined epiphyseal growth plates in control, hypothyroid, thyrotoxic, and hypothyroid-thyroxine (hypo-T4)-treated rats. Hypothyroid growth plates were grossly disorganized, contained an abnormal matrix rich in heparan sulfate, and hypertrophic chondrocyte differentiation failed to progress. These effects correlated with the absence of collagen X expression and increased parathyroid hormone-related protein (PTHrP) messenger RNA (mRNA) expression. In thyrotoxic growth plates, histology essentially was normal but PTHrP receptor (PTHrP-R) mRNA was undetectable. PTHrP is a potent inhibitor of hypertrophic chondrocyte differentiation that acts in a negative feedback loop with the secreted factor Indian hedgehog (Ihh) to regulate endochondral bone formation. Thyroid hormone receptor α1(TRα1), TRα2, and TRβ1 proteins were localized to reserve zone progenitor cells and proliferating chondrocytes in euthyroid rat cartilage; regions in which PTHrP and PTHrP-R expression were affected by thyroid status. Thus, dysregulated Ihh/PTHrP feedback loop activity may be a key mechanism that underlies growth disorders in childhood thyroid disease.
UNTREATED JUVENILE hypothyroidism causes growth arrest, delayed bone age, and epiphyseal dysgenesis.(1) Furthermore, a proportion of patients with resistance to thyroid hormone (RTH), caused by dominant negative mutant thyroid hormone receptor β (TRβ) proteins, suffer from developmental abnormalities of bone and growth retardation.(2) Studies of the long-term effects of childhood hypothyroidism on linear growth indicate that catch-up growth after thyroxine (T4) replacement is incomplete because bone age advances at a faster rate than height.(3) Thus, skeletal maturation and epiphyseal fusion occur before final height is achieved, leading to speculation that skeletal responses to T4 may be exaggerated in hypothyroidism.(3) Accordingly, childhood thyrotoxicosis causes accelerated growth and advanced bone age, which may lead to craniosynostosis, premature growth plate closure, and eventual short stature.(1, 3, 4) These clinical observations suggest that bone formation and growth are exquisitely sensitive to thyroid hormones. Studies of TR knockout mice support this view because mice that lack either TRα,(5) or both TRα and TRβ genes,(6) exhibit characteristic features of growth retardation and epiphyseal dysgenesis. These abnormalities result from severe hypothyroidism caused by impaired production of thyroid hormones. The phenotype can be reversed by T4 replacement in TRα-deficient mice, indicating that TRβ can compensate for loss of TRα in the growth plate.(5) These data add to conclusions from cell culture systems, which suggest that thyroid hormones are essential to support differentiation and columnar organization of chondrocytes in vitro(7,8) but do not provide mechanistic information regarding how thyroid hormones regulate endochondral ossification or an understanding of the pathogenesis of skeletal abnormalities in thyroid disease.
Linear growth occurs by endochondral bone formation, a process in which epiphyseal growth plate chondrocytes undergo an organized program of maturation, proliferation, and differentiation resulting in the formation of hypertrophic chondrocytes. Terminally differentiated hypertrophic chondrocytes secrete an extracellular matrix rich in collagen X and eventually undergo apoptosis to leave a cartilage scaffold that is mineralized and ossified by osteoblasts invading from the bone marrow during neovascularization.(9) These fundamental processes are coupled to ensure that endochondral bone formation is organized and growth progresses normally (9–13) but it is not known how they are coordinated by circulating hormones. A series of experiments have established that the pace of chondrocyte differentiation in embryonic chicken and mouse long bones is regulated by a complex feedback loop involving the paracrine factors Indian hedgehog (Ihh),(14,15) parathyroid hormone-related protein (PTHrP),(16,17) and the bone morphogenetic protein receptor 1A (BMPR-1A).(18,19) Details of this feedback loop have not been investigated during postnatal growth, but studies of patients with Jansen's metaphyseal chondrodysplasia(20) and Blomstrand chondrodysplasia,(21) conditions that result from constitutively active or nonfunctional PTHrP receptors (PTHrP-R), respectively, confirm a fundamental role for PTHrP in human endochondral bone formation.
Thus, physiological postnatal growth and skeletal maturation rates are dependent on thyroid status and these parameters are governed directly by the Ihh/PTHrP regulated pace of chondrocyte differentiation during endochondral bone formation. We propose that the effects of thyroid hormones on linear growth result from changes in activity of the Ihh/PTHrP negative feedback loop. To investigate this hypothesis, we examined the expression of components of this loop and localized TR proteins in epiphyseal growth plates from euthyroid, hypothyroid, thyrotoxic, and hypothyroid-T4 (hypo-T4)-treated rats.
MATERIALS AND METHODS
Fifteen normal and 15 thyroidectomized male, 6-week-old Sprague-Dawley rats were studied for 6 weeks. Normal rats were treated with saline or T4 (50 μg/kg per day) and thyroidectomized rats were treated with saline or T4 (10 μg/kg per day) to form euthyroid control (n = 7), thyrotoxic (n = 8), hypothyroid (n = 7), and hypo-T4 (n = 8) groups. All rats were fed a normal diet and thyroidectomized animals received calcium lactate supplementation to drinking water. Rats were weighed and tibias were measured at the start of the study, after 3 weeks, and at death. Animal studies were performed under license in compliance with the Animals (Scientific Procedures) Act 1986 and were approved by the local Imperial College School of Medicine Biological Services Unit ethical review process. Plasma T4 concentrations were determined by immunoradiometric assay (Euro/DPC, Ltd., Caernarfon, Gwynedd, Wales, U.K.). Thyroid-stimulating hormone (TSH) was measured using reagents provided by the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK) and the National Hormone and Pituitary Program (Dr. A. Parlow, Harbor University of CA, Los Angeles Medical Center, CA, U.S.A.) as described.(22)
Histology and immunohistochemistry
Tibias were fixed at 4°C for 24 h in 4% paraformaldehyde and decalcified in 12.5% EDTA at 4°C for 2 weeks or fixed for 24 h in 10% neutral buffered formalin and subsequently decalcified for 1 week in 10% formic acid + 10% neutral buffered formalin at 20°C. Three-micrometer sections were cut onto 3-aminopropyltriethoxysilane (APES; Sigma, Dorset, U.K.)-coated slides, deparaffinized in xylene, and rehydrated. Sections were stained with hematoxylin and eosin, or Alcian blue 8GX and van Gieson. Critical electrolyte concentration staining with Alcian blue was performed in buffers containing graded concentrations of MgCl2.(23)
Ihh expression was determined using a polyclonal goat anti-human Ihh antibody (Santa Cruz Laboratories, Santa Cruz, CA, U.S.A.). Endogenous peroxidase was quenched with 0.3% H2O2 in methanol and nonspecific binding blocked with 10% normal rabbit serum (NRS; Vector, Peterborough, U.K.) in phosphate-buffered saline (PBS) + 1% bovine serum albumin (BSA) for 1 h at room temperature. Sections were incubated for 1 h at room temperature with Ihh antibody diluted 1:40 in PBS + 1% NRS + 0.1% BSA and washed in PBS. Controls were treated with goat immunoglobulin G (IgG; Sigma). Bound antibody was detected by a biotinylated rabbit anti-goat secondary antibody (DAKO, Cambridge, U.K.) diluted 1:400 in PBS for 45 minutes at room temperature and exposed to preformed ABC complex (Vector) for 40 minutes. Peroxidase was visualized with diaminobenzidine tetrahydrochloride (DAB) and 0.2% H2O2.
Polyclonal rabbit anti-TRα1 and TRα2 or mouse monoclonal anti-TRβ1 antibodies (Affinity Bioreagents, Inc., Golden, CO, U.S.A.) were used to determine expression of TRs.(24,25) Endogenous peroxidase was quenched and nonspecific binding was blocked with 10% normal goat serum (NGS) for TRα1 and TRα2 or 10% normal horse serum (NHS) for TRβ1, diluted in PBS + 1% BSA + 1% Triton-X 100 + 0.5% casein. Sections were incubated for 1 h at room temperature with primary antibody diluted 1:400 for TRα1, 1:75 for TRα2, and 1:100 for TRβ1 in PBS + 1% NGS or NHS + 0.1% BSA and washed in PBS. Controls were treated with rabbit or mouse IgG (DAKO). Bound antibody was detected by a goat anti-rabbit secondary antibody (Vector) diluted 1:500 in PBS for TRα1 and 1:200 for TRα2 or a horse anti-mouse secondary antibody (Vector) diluted 1:200 for TRβ1.
Cloning of rat PTHrP, Ihh, and BMPR-1A cDNAs
PTHrP and Ihh partial complementary DNAs (cDNAs) were amplified by polymerase chain reaction (PCR) from genomic DNA (200 ng) in a 50-μl reaction containing 0.2 mM deoxynucleoside triphosphate (dNTPs), 50 pmol/μl primers (Sigma-Genosys Biotechnologies, Cambridge, U.K.), Taq DNA polymerase (1 U), 10× PCR buffer (5 μl), and 1.5 mM MgCl2. After denaturation at 94°C for 1 minute, 35 cycles of 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C followed by a final extension of 2 minutes were performed. A BMPR-1A partial cDNA was generated from UMR106 cell RNA by reverse-transcription (RT)-PCR. RNA (7 μg) was reverse transcribed for 30 minutes using a BMPR-1A reverse primer (500 pmol/μl) and random hexamers (0.016 A260 units; Pharmacia LKB, Uppsala, Sweden) in the presence of RNAse inhibitor (20 U), dNTPs (0.8 mM), Avian myeloblastosis virus reverse transcriptase (AMV-RT; 10 U; NBL Gene Sciences, Cramlington, Northumberland, U.K.), and 10× AMV-RT buffer (2 μl) in a 20-μl reaction. PCR amplification for 40 cycles was performed using 10 μl of cDNA template and an annealing step of 62.5°C for 30 s. cDNA products were sequenced (Thermo-Sequenase I dye terminators; Amersham Pharmacia, Little Chalfont, Bucks., U.K.) using a semiautomated detection system (ABI 373A; Applied Biosystems, Foster City, CA, U.S.A.). The following primers were used: rat PTHrP (GenBank M31603), forward primer nucleotides 191-213 and reverse 593-571; rat BMPR-1A (GenBank D38082), forward 261-280 and reverse 1034-1012; rat Ihh was cloned using primers from the mouse cDNA (GenBank X76291), forward 422-444 and reverse 800-777. The rat Ihh partial cDNA sequence was deposited in GenBank (AF175209) and contained 93% and 85% identity to the mouse and human cDNAs, respectively, and 77% and 76% amino acid identity. The reduced amino acid homology is caused by a single nucleotide deletion in the rat cDNA, followed by a nucleotide insertion that results in a frame shift involving 26 amino acids.
cRNA probes and in situ hybridization
cDNAs were subcloned into pGEM-T and a PTHrP-R cDNA (clone R15B(26)) was a gift from Dr. G.V. Segre (Boston, MA, U.S.A.). PTHrP and Ihh were linearized with Nco I; BMPR-1A was linearized with Spe I; and PTHrP-R was linearized with HindIII. Digoxigenin-labeled probes of 402, 378, 773, and 831 base pairs (bp), respectively, were synthesized with SP6 RNA polymerase (Boehringer Mannheim, Lewes, Sussex, U.K.).
Messenger RNA (mRNA) in growth plate sections was exposed by digestion with 20 μg/ml of proteinase K in TE (10 mM Tris-Cl, 1 mM ethylenediaminetetracetic acid [EDTA] pH 7-8) for 12 minutes. Sections were acetylated in 0.1 M triethanolamine and 0.3 M acetic anhydride for 10 minutes before washing, dehydrating, and drying. Hybridization solution (1× Denhardt's solution, 50% deionized formamide, 20% dextran sulfate, and 2-μg probe) was heated to 80°C for 1 minute, added to the section, and incubated in humidified chambers at 55°C overnight. Slides were washed in 50% formamide in 0.15 M NaCl, 5 mM NaH2PO4, 5 mM Tris/HCl, and 2.5 mM EDTA, pH 6.8 and twice for 30 minutes at 55°C followed by 5 washes in 0.5 M NaCl, 0.1 M Tris/HCl, 0.2 M EDTA, pH 7.4 for 10 minutes at room temperature. Sections were blocked in 2.5% BSA in 100 mM Tris/HCl and 150 mM NaCl, pH 7.5, for 1 h before adding alkaline phosphatase (AP) or horseradish peroxidase (HRP)-conjugated antidigoxigenin Fab fragments (Boehringer Mannheim) and then diluted 1:500 in blocking solution for 1 h at room temperature. Slides were washed in PBS and developed in either nitroblue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate-4-toluidine (NBT/BCIP; Boehringer Mannheim) to detect AP or DAB to detect HRP.
In situ hybridization controls
A bacterial neomycin resistance gene complementary RNA (cRNA) probe (Boehringer Mannheim) was used as a negative control for all hybridizations (data not shown). Specificity of RNA:RNA hybridization was tested by preincubating sections with RNases A and T before the acetylation step and processing as mentioned previously. To ensure that detection of AP was not caused by endogenous enzyme activity in bone, controls were performed in which HRP-conjugated Fab fragments were used in place of AP-conjugated fragments and slides were developed using NBT/BCIP. Reciprocal controls were performed to ensure no endogenous peroxidase activity was present. In all controls no signal was detected (data not shown). All in situ hybridization studies were performed using both AP and HRP detection systems and identical results were obtained (data not shown). Thermolability studies were performed to confirm the specificity of hybridization. No signal was detected with any probe when annealing was performed at 65°C compared with the normal 55°C (data not shown).
Thyroidectomy resulted in severe hypothyroidism, linear growth retardation, and failure to gain weight (Table 1). Treatment of thyroidectomized rats with subphysiological doses of T4 (10 μg/kg per day) incompletely reversed the hypothyroidism but resulted in weight gain equivalent to controls and restored growth. Normal rats treated with T4 (50 μg/kg per day) became thyrotoxic, with elevated T4 concentrations and suppressed TSH levels. In these animals, there was a small increase in weight, not significantly different from control or hypo-T4 groups, and linear growth was equivalent to euthyroid controls.
Table Table 1.. Thyroid Status and Growth Characteristics of Animal Treatment Groups
Thyroid hormones are essential for hypertrophic chondrocyte differentiation and endochondral ossification
Growth plates from hypothyroid rats were grossly disorganized compared with euthyroid control, thyrotoxic, and hypo-T4 animals (Fig. 1; data not shown). Proliferating chondrocytes failed to form discrete columns and the hypertrophic zone was markedly diminished and morphologically indistinct. The growth plate was separated from the primary spongiosum by a mineralized layer in which neovascularization and osteoblast invasion were markedly reduced or absent. Similar appearances were reported in studies that revealed a major role for thyroid hormone in growth plate cartilage.(27,28) In addition, bone trabeculae in the primary spongiosum of euthyroid rats were organized parallel to the columns of proliferating chondrocytes, reflecting the functional continuity between maturing chondrocytes and mineralizing osteoblasts required for normal endochondral ossification. (9–13) In hypothyroid rats this relationship was disrupted in that bone trabeculae were fewer, thinner, and disorganized relative to controls (Fig. 1).
Histochemical staining revealed that the disrupted architecture of hypothyroid growth plates was associated with deposition of an abnormal cartilage matrix. The van Gieson stain shows collagen-containing bone and osteoid in red and Alcian blue 8GX stains cartilage matrix mucopolysaccharides blue.(29) The extent of Alcian blue staining in hypothyroid cartilage was reduced, patchy, and grossly disorganized. Cartilage islands containing chondrocytes were present in hypothyroid epiphyseal bone, which lacked the lamellar structure evident in trabeculae from euthyroid animals (Fig. 1).
Expression of collagen X, a specific marker of hypertrophic chondrocyte differentiation that is expressed before the onset of endochondral ossification,(30,31) was restricted to a discrete layer of mature hypertrophic chondrocytes in euthyroid, thyrotoxic, and hypo-T4 rats but was undetectable in growth plates of hypothyroid rats (Fig. 1; data not shown). Thus, hypertrophic chondrocyte differentiation was impaired severely in hypothyroidism.
Cartilage matrix in hypothyroid growth plates is abnormal
Alcian blue critical electrolyte concentration histochemistry(23) was used to characterize differences between acidic mucopolysaccharide groups present in the cartilage matrix of growth plates from hypothyroid rats and other groups of animals (Fig. 2; data not shown). Mg2+ salts in the staining solution reversibly compete with dye molecules for binding sites in the polyionic substrate, thereby increasing Alcian blue specificity when graded Mg2+ concentrations are used.(23,29) Hypothyroid growth plate cartilage stained at a critical electrolyte concentration of 0.7-0.8 M MgCl2, as indicated by the blue color of the matrix surrounding proliferating chondrocytes (Fig. 2G). Alcian blue staining at this critical electrolyte concentration of 0.7-0.8 M MgCl2 indicates the presence of strongly sulfated matrix mucopolysaccharide residues consisting mainly of heparan sulfate.(23, 29, 32) These blue-stained sulfated matrix components were located exclusively around proliferating chondrocytes in hypothyroid growth plates. Alcian blue staining was absent from euthyroid growth plates at 0.7-0.8 M MgCl2 concentration (Fig. 2C), which appeared red as a result of counterstaining. Growth plate cartilage from other groups of animals only stained blue at a critical electrolyte concentration of 0.2-0.3 M MgCl2 (Fig. 2B; data not shown) indicating a preponderance of mucopolysaccharides containing carboxylate and phosphate residues composed mainly of chondroitin sulfate and hyaluronic acid.(23, 29, 32) The negative staining at 1.0 M MgCl2, indicated by neutral red counterstaining, is caused by the absence of keratan sulfate mucopolysaccharides from growth plate cartilage matrix in the rat.
PTHrP mRNA expression is enhanced in hypothyroidism and PTHrP-R mRNA is absent in thyrotoxicosis
Next, we determined whether the effects of hypothyroidism on endochondral ossification and cartilage matrix deposition correlated with changes in expression of components of the Ihh/PTHrP feedback loop.(14,15) Ihh mRNA was expressed mainly throughout the proliferative zone and prehypertrophic region, but protein was detected predominantly in hypertrophic chondrocytes in keeping with its effects on embryonic bone development.(14,15) The distribution of Ihh mRNA and protein was similar in euthyroid, thyrotoxic, and hypo-T4 groups but was localized more predominantly in the upper regions of the proliferative zone and in the reserve zone in hypothyroid animals (Fig. 3). Expression of BMPR-1A mRNA was undetectable by in situ hybridization in all animal groups (data not shown) suggesting that mRNA levels were below the sensitivity of the method because BMPR-1A protein has been detected recently in proliferating and maturing chondrocytes in 12-week-old rats.(33)
In contrast, expression of both PTHrP and PTHrP-R was clearly regulated by alterations of thyroid status. PTHrP mRNA expression was restricted to a discrete layer of prehypertrophic and hypertrophic chondrocytes in euthyroid, thyrotoxic, and hypo-T4 animals but was increased and included all chondrocytes extending throughout the proliferative and reserve zones in hypothyroid growth plates (Fig. 4).
PTHrP-R mRNA, in contrast, was expressed throughout all zones of the growth plate in euthyroid and hypothyroid animals but was restricted to proliferative and prehypertrophic chondrocytes in hypo-T4 growth plates. Strikingly, PTHrP-R expression was undetectable in growth plates from thyrotoxic animals (Fig. 4). This data strongly implicates PTHrP signaling components of the Ihh/PTHrP feedback loop as important targets for thyroid hormone in growth plate cartilage. The more restricted expression of PTHrP-R mRNA in hypo-T4 rats, compared with euthyroid and hypothyroid groups, supports the view that the hypothyroid skeleton may be more sensitive to T4(3) and further implicates this pathway as a direct target of thyroid hormones.
TRs localize to regions of the growth plate in which PTHrP and PTHrP-R are expressed
Previously, we have used TRα1-, TRα2-, and TRβ1-specific antibodies in rat skeletal cells and shown that functional TR proteins are expressed in osteoblastic cells.(24,25) These antibodies were used to determine the expression of TR isoforms in growth plate chondrocytes. TRα1 and TRβ1 were localized to the nuclei of reserve zone progenitor cells, proliferating chondrocytes, and invading osteoblasts but were absent from hypertrophic chondrocytes. TRα1 expression was widespread in all cells throughout these regions whereas TRβ1 staining was weak and present in only a subset of cells. TRα2 was present in the nuclei of reserve and proliferative zone chondrocytes but was not detected in osteoblasts. In hypothyroid growth plates, TRα1 and TRα2 were undetectable in all areas but weak TRβ1 staining was maintained in prehypertrophic chondrocytes (Fig. 5). No changes in TR expression were seen in thyrotoxic or hypo-T4 animals (data not shown).
In these studies endochondral bone formation was impaired severely in hypothyroid rats but appeared normal in all other groups including hypothyroid animals treated with subphysiological doses of T4. Furthermore, growth retardation and failure to gain weight was only evident in hypothyroid rats but not in other groups, supporting the hypothesis that growth plate chondrocytes are exquisitely sensitive to thyroid hormones in vivo.
Identification of TRα1, TRα2, and TRβ1 proteins in progenitor cells and proliferating chondrocytes and TRα1 and TRβ1 in invading osteoblasts is consistent with the recent identification of TRs in whole rat growth plate extracts by RT-PCR and Western blotting,(34) although TRα2 was undetectable by immunoblotting techniques in that study. TRα1 and TRβ1 bind T3 and act as hormone-inducible transcription factors, whereas α2 does not bind T3 and antagonizes the actions of other TR isoforms in vitro.(35) The absence of TRα1 and TRα2 expression and persistence of TRβ1 in hypothyroid animals is reminiscent of TRα knockout mice. These animals express only TRβ1 in the skeleton and are grossly hypothyroid, yet the growth plate is highly sensitive to thyroid hormones because T4 replacement for 1 week rescues the phenotype.(5) Furthermore, double knockout of TRα and TRβ genes does not worsen growth retardation and epiphyseal dysgenesis compared with TRα null mice,(5,6) although the double knockout phenotype is not rescued by T4 replacement. Thus, altered TR expression in hypothyroid rats in these studies mimics the hypothyroid state induced in TRα null mice. Taken together, this data indicates that chondrocyte activity is highly sensitive to changes in local thyroid hormone concentrations and strongly suggests that the epiphyseal growth plate is a primary T3 target tissue. Colocalization of TR and PTHrP/PTHrP-R expression in growth plate chondrocytes further supports this view.
In the developing limb, Ihh is secreted by committed prehypertrophic chondrocytes and stimulates production of PTHrP from the periarticular region of the epiphysis via a relay system involving BMPR-1A. (14–19) PTHrP acts on PTHrP-R expressing prehypertrophic chondrocytes to maintain cells in a proliferative state. This results in reduced Ihh production and completes a regulatory feedback loop in which PTHrP exerts a potent negative signal that inhibits hypertrophic chondrocyte differentiation. It is not known whether this feedback loop is maintained during postnatal growth but the phenotypes of Jansen's and Blomstrand chondrodysplasias(20,21) confirm a major role for PTHrP in human endochondral ossification, and the loop also is conserved among the vertebrates.
In these studies, the region of the epiphyseal growth plate expressing PTHrP was expanded in hypothyroid animals, whereas expression of its receptor was undetectable in thyrotoxicosis. These findings suggest a mechanism to account for the effects of thyroid hormones on linear growth and skeletal maturation. Increased PTHrP expression in hypothyroid growth plates is predicted to transmit an enhanced negative signal that potently inhibits hypertrophic chondrocyte differentiation to result in arrest of linear growth. The absence of collagen X mRNA expression in hypothyroid growth plates, but persistence of Ihh protein expression in prehypertrophic chondrocytes, indicates that terminal hypertrophic chondrocyte differentiation, rather than the commitment of proliferating cells toward hypertrophy, is blocked in hypothyroidism. Conversely, absence of PTHrP-R expression in growth plates of thyrotoxic animals is predicted to prevent negative PTHrP signaling and facilitate progression of hypertrophic chondrocyte differentiation to accelerate linear growth. In the proposed model, thyroid hormones regulate the set point of the Ihh/PTHrP feedback loop to modulate the pace of chondrocyte differentiation and endochondral bone formation during postnatal growth. This hypothesis is supported by clinical observations of childhood thyroid disease and RTH syndromes in man, (1–4) by the skeletal phenotypes of TRα(5) and TRα/β(6) null mice, by studies that show that T4 replacement in hypothyroid children produces rapid catch-up growth and reversal of epiphyseal stippling,(3) and by the rapid rescue of the skeletal phenotype in TRα null mice by T4 replacement.(5)
It cannot be determined from our studies whether the effects on PTHrP and PTHrP-R expression represent direct actions of T3 or whether they are secondary effects. Thus, other paracrine factors acting in the growth plate also should be considered in the pathogenesis of hypothyroid growth disorders. It is significant that epiphyseal dysgenesis in hypothyroidism was associated with deposition of an abnormal cartilage matrix, separation of the growth plate from the underlying primary spongiosum, and diminished growth plate neovascularization. Coupling of hypertrophic chondrocyte differentiation and angiogenesis is a prerequisite for endochondral ossification and involves numerous growth factors (10–13) that influence chondrocyte differentiation but may require components of the extracellular matrix to be active in cartilage.(36) For example, fibroblast growth factor (FGF) binds to cell surface receptors and matrix binding sites containing heparan sulfate, and the inhibitory effects of FGF on hypertrophic chondrocyte differentiation require heparan sulfate.(37,38) Furthermore, it has been proposed that heparan sulfate serves as a depot for basic FGF in cartilage during endochondral bone formation.(38) Thus, the finding of a heparan sulfate-rich matrix in hypothyroid growth plate cartilage suggests that T3 may regulate growth factor signaling by altering the structure of matrix secreted by chondrocytes. This raises the novel hypothesis that the cartilage matrix provides a link between paracrine pathways in the growth plate and the actions of thyroid hormone during endochondral bone formation.
These studies will facilitate new areas of investigation to elucidate how chondrocyte differentiation, matrix production, angiogenesis, and osteoblast activity are coordinated by circulating hormones during endochondral ossification. Ultimately, an understanding of these events will provide approaches to manipulate bone formation for the generation of novel strategies to manage growth disorders and to enhance skeletal remodeling during fracture repair, a process that is highly sensitive to thyroid status. (39–41)
G.R.W. was supported by a Medical Research Council Career Establishment Grant (G9803002) and Wellcome Project Grant (050570) and T.S. was supported (in part) by a European Society for Pediatric Endocrinology Research Fellowship, sponsored by Novo Nordisk A/S.