46.7.1

Targeting thyroid hormone transporters

Genetically modified mice are a powerful tool to investigate the individual components of thyroid hormone transport, metabolism and action, and their contributions to skeletal physiology ( Table 46.1 ). Thyroid hormone transporters are necessary for the active transport of thyroid hormones into cells. Mice lacking the thyroid hormone transporter MCT8 have elevated T3 but decreased T4 levels . Mct8 ^y/− mice have only a minor delay in postnatal growth suggesting that other transporters may compensate for deficiency of MCT8 . Nevertheless, analysis of Mct8 knockout mice revealed delayed endochondral ossification due to impaired T3 signaling in chondrocytes, and Mct8 ^y/− mice also exhibited decreased bone mineralization and strength in adulthood . Although MCT8 is important for endochondral ossification and adult bone maintenance, other studies provide further evidence to suggest functional redundancy between thyroid hormone transporters in the skeleton. Thus although Mct10 mutant mice and Oatp1c1 ^−/− knockout mice have normal weight gain and linear growth, Mct8 and Oatp1c1 double knockout mice display significant postnatal growth retardation .

46.7.2

Targeting thyroid hormone metabolism

The skeletal consequences of mutation or deletion of the deiodinase enzymes have also been investigated. Dio1 is not expressed in the skeleton and, accordingly, DIO1-deficient mice display normal weight gain and linear growth, although the skeleton has not been studied in detail .

Dio2 is expressed mainly in osteoblasts in the skeleton , and Dio2 ^−/− mice have a threefold increase in TSH, a small increase in T4 and normal T3 levels due to impaired deiodination of T4 in the pituitary . DIO2 deficiency and absent T4 to T3 conversion in osteoblasts in Dio2 ^−/− mice resulted in decreased bone turnover, increased bone mineralization and brittle bones in adults, whereas skeletal development and postnatal growth were normal . Nevertheless, mutant mice harboring a Thr92Ala polymorphism in DIO2, which has been associated with increased bone turnover markers and decreased BMD in humans, were recently shown to have normal linear growth and normal bone mineral content and strength . Taken together, these data demonstrate that DIO2 in osteoblasts is indispensable and essential for the maintenance of normal adult bone mass and strength, whereas a Thr92Ala polymorphism in DIO2 that results in minor biochemical changes in enzyme activity has no effect on bone mineralization and strength.

Finally, the role of DIO3 in bone has not been investigated specifically. Although Dio3 ^−/− mice have severe growth retardation, they also have a multisystem disorder and high perinatal mortality that confounds interpretation of the skeletal consequences following Dio3 deletion . Thus cell-specific gene targeting will be required to determine the role of DIO3 in bone.

46.7.3

Targeting thyroid hormone receptors

Studies in mutant mice have also revealed the roles of individual TR isoforms in the skeleton. Mutation or deletion of TRα causes severe skeletal abnormalities, including delayed endochondral ossification, impaired chondrocyte differentiation, defective growth plate development, and decreased bone mineralization. TRα ^0/0 mice have complete deletion of all isoforms arising from the Thra locus and are systemically euthyroid with moderate skeletal abnormalities . Thus juveniles have transient growth delay with impaired endochondral and intramembranous ossification, whereas adults display increased bone volume and mineralization with osteosclerosis due to decreased osteoclastic bone resorption.

The key role of TRα1 in bone was further demonstrated in TRα1 ^PV mice that express a potent dominant-negative mutant receptor. The homozygous mutation is lethal, but TRα1 ^PV/− heterozygotes have mildly raised TSH and T3, but normal T4 levels. Despite their near normal thyroid status, TRα1 ^PV/+ mice have severely impaired skeletal development and grossly abnormal adult bone maintenance, displaying a much more severe phenotype than TRα ^0/0 knockout mice and thus demonstrating an important deleterious role for the dominant-negative receptor in bone . TRα1 ^PV/+ mice that were also homozygous for a mutant nuclear corepressor NCoR1 that cannot interact with TR had a milder skeletal phenotype, indicating that interaction with NcoR1 is required for dominant-negative activity of the TRα1 ^PV mutant receptor . A similar developmental phenotype was seen in mice expressing a dominant-negative TRα1 ^L400R mutant in chondrocytes, indicating the important role of TRα1 in chondrocytes during skeletal development .

In contrast to TRα mutant mice in which thyroid status is near normal, TRβ ^−/− knockout mice have elevated circulating T4, T3, and TSH due to an impaired negative feedback response to thyroid hormones in the hypothalamus and pituitary . The elevated circulating thyroid hormones activate TRα signaling in bone, resulting in a phenotype of accelerated growth and endochondral ossification in juveniles and progressive bone loss and osteoporosis in adults that is similar to the effects of thyrotoxicosis . Furthermore, TRβ ^PV/PV mice that express dominant-negative TRβ mutant receptors have a more severe phenotype than TRβ ^−/− mice that consist of advanced endochondral and intramembranous ossification, craniosynostosis, short stature during development, and severe osteoporosis in adulthood . Together, these observations in TRβ mutant mice further demonstrate major roles for TRα1 in the developing and adult skeleton. Thus T3 stimulates anabolic actions mediated by TRα1 that promotes bone growth and mineralization during development but exerts catabolic actions of accelerated bone loss in adults that lead to osteoporosis.

46.7.4

Targeting the thyroid-stimulating hormone receptor

Tshr ^−/− knockout mice have congenital hypothyroidism and do not survive to adulthood unless supplemented with thyroid extract . Nevertheless, analysis of Tshr ^−/− mice revealed a 20% reduction in BMD, suggesting the hypothesis that TSH inhibits bone remodeling and turnover . Importantly, Tshr ^−/− mice received thyroid extract only after weaning and thus remained hypothyroid during the time of physiological peak thyroid hormone levels when growth velocity is at a maximum . Nevertheless, in hyperthyroid Tshr ^−/− mice treated with supraphysiological doses of thyroxine, bone loss was greater than observed in similarly treated wild-type control animals, suggesting further that TSH inhibits bone resorption in vivo and supporting the proposal that bone loss in hyperthyroidism may result from both elevated thyroid hormone concentrations and suppressed levels of TSH.

One study compared two strains of mutant mice, which both have grossly elevated circulating TSH levels and differ only by expressing a normal or nonfunctional TSHR. Thus Tshr ^hyt/hyt mice with congenital hypothyroidism harbor a loss-of-function mutation in Tshr , whereas Pax8 ^−/− mice lack the thyroid-specific transcription factor PAX8 essential for thyroid follicular cell development but express a normal fully functional TSHR. Both Tshr ^hyt/hyt and Pax8 ^−/− mice displayed similar skeletal abnormalities that are independent of TSHR signaling and consist of delayed endochondral ossification, growth retardation, reduced cortical bone volume, defective trabecular bone remodeling, and decreased bone mineralization. Furthermore, TRβ ^−/− and TRβ ^PV/PV mice each have disruption of the HPT axis with elevated circulating concentrations of both TSH and thyroid hormones. Both mutants display accelerated skeletal development and osteoporosis in adults that is consistent with elevated T3 signaling in bone but not with their high circulating TSH levels.

Overall, experimental studies in a series of mutant mice targeting all levels of the HPT axis and thyroid hormone signaling pathway are consistent with major physiological roles for T3, acting via TRα1, during skeletal development, and in adult bone. TSH also inhibits bone resorption via the TSHR in osteoclasts and suppression of TSH in thyrotoxicosis may thus compound the detrimental effects of thyroid hormone excess on the skeleton. Nevertheless, it is less clear whether TSH has important physiological effects during normal bone formation and maintenance.

46.8.1

Mutations affecting thyroid hormone transporters

Inactivating mutations in MCT8 cause Allan–Herndon–Dudley X-linked psychomotor retardation syndrome (OMIM 300523 ) with low T4 and elevated T3 levels. Affected individuals have severe developmental delay with ataxia, dystonia, and hypotonia, but the skeletal consequences have not been described in detail and are likely to be confounded by the neurological abnormalities of the condition .

46.8.2

Mutations affecting thyroid hormone metabolism

Selenocysteine (Sec) insertion sequence binding protein 2 (SBP2) mediates incorporation of the rare amino acid selenocysteine into the active site of the thyroid hormone deiodinases and is thus essential for conversion of T4 to T3. Affected individuals with mutations in SBP2 (OMIM 609698 ) have an elevated T4:T3 ratio and display transient delayed bone maturation, prepubertal growth retardation, and craniofacial abnormalities , demonstrating the importance of thyroid hormone metabolism and action for skeletal development in humans.

46.8.3

Mutations affecting thyroid hormone receptors

Resistance to thyroid hormone (RTH) β (OMIM 188570 ) is an autosomal-dominant condition caused by dominant-negative mutations of THRB , while RTHα results from dominant-negative mutations affecting THRA .

In RTHβ, mutant TRβ disrupts the negative feedback loop in the HPT axis, causing increased thyroid hormone levels with inappropriately normal or elevated TSH levels . Individuals have a variable and complex mixed phenotype of hyperthyroidism and hypothyroidism in different tissues, depending on (1) the degree of disruption to the HPT axis and consequent increase in circulating thyroid hormones that depends on the potency of the specific mutant receptor and (2) whether a particular tissue predominantly expresses TRα or TRβ. In the skeleton, affected individuals may develop craniofacial abnormalities, craniosynostosis, delayed or advanced bone age, short stature, and increased bone turnover causing osteoporosis and fracture, but a broad range of abnormalities have been observed even within affected kindreds .

Individuals affected by RTHα (OMIM 614450 ) have classical features of congenital hypothyroidism, but unlike RTHβ, there is no disruption to the normal feedback mechanisms of the HPT axis. TSH levels are normal, with a reduced T4:T3 ratio with normal/low T4 and normal/high T3 levels. Skeletal features result from severe disruption of thyroid hormone action in bone and include disproportionate short stature due to short lower limbs, delayed tooth eruption, delayed closure of fontanels, delayed bone age, epiphyseal dysgenesis, and a thickened calvarium. To date, 14 heterozygous frameshift, missense and nonsense mutations in THRA have been described that cause RTHα . The severity of the RTHα phenotype varies according to the type and location of the THRA mutations, with frameshift and nonsense mutations causing a more severe clinical syndrome than that seen in patients with milder missense mutations who may benefit from levothyroxine treatment .

46.8.4

Mutations affecting thyroid-stimulating hormone

Rare mutations have been described in TSHB that cause inactive TSH and congenital nongoitrous hypothyroidism (OMIM 275100 ). Normal BMD has been observed in two individuals aged 10 and 7 years with this condition who received thyroid hormone replacement from birth . Thus the lifelong absence of TSH did not affect bone mineralization during childhood.

Loss-of-function mutations in TSHR cause congenital nongoitrous hypothyroidism of varying severity (OMIM 275200 ) . Individuals with severe thyroid hypoplasia had delayed bone age and intramembranous ossification at birth but normal growth and skeletal development occurred following T4 replacement .

Nonautoimmune autosomal-dominant hyperthyroidism (OMIM 609152 ) is caused by rare gain-of-function mutations in TSHR and treated by surgical and radioiodine thyroid ablation followed by T4 replacement . At presentation, individuals display classical features of juvenile hyperthyroidism, including advanced bone age and craniosynostosis, but amelioration of the phenotype has been observed after treatment that normalizes systemic thyroid status . These conditions that affect TSHB and TSHR demonstrate that correction of thyroid hormone levels improves or prevents skeletal abnormalities, regardless of the levels and activity of TSH or the TSH receptor.

46.9.1

Population genetic studies

Thyroid status is a highly heritable trait with complex inheritance. Normal individuals have a unique HPT axis set point that varies by less than 50% of the population reference range . Twin studies have supported these findings with heritability estimates of 23%–64% for circulating free T3 concentrations, 39%–65% for free T4 levels and 64%–65% for TSH . An initial meta-analysis of genome-wide association studies (GWASs) identified 20 loci independently associated with TSH levels and 6 loci independently associated with T4 levels . A recent meta-analysis that included over 72,000 individuals identified 109 variants associated with thyroid dysfunction and also led to the identification of a novel thyroid hormone transporter SLC17A4 and metabolizing enzyme AADAT . Genetic variation that affects tissue-specific thyroid hormone sensitivity is likely to affect variability in bone mineral density (BMD). BMD is also a highly heritable trait, and GWASs have identified over 500 loci to be independently associated with variability in BMD . Though the THRB locus was identified in a recent GWAS of BMD estimated by heel quantitative ultrasound in 426,824 UK Biobank participants , no other loci containing genes that regulate thyroid status have yet been associated with BMD or fracture susceptibility.

Candidate gene studies of thyroid signaling genes have yielded variable results when assessing the relationships between polymorphisms and changes in BMD. A large study of polymorphisms at the TSHR locus found no association with changes in femoral neck or lumbar spine BMD, and Mendelian randomization analysis found no evidence of a causal relationship between variation in TSH concentrations and BMD . However, the D727E polymorphism in TSHR had separately been associated with increased femoral neck BMD , but in apparent contradiction, the same polymorphism was found to have higher frequency in those with osteoporosis . Consistent with the UK Biobank GWAS , polymorphisms in THRB have been associated with trabecular but not cortical BMD at the hip , but variants in THRA have not been associated with BMD or fracture risk despite the autosomal-dominant skeletal dysplasia recently identified in patients with dominant-negative mutations of THRA . A polymorphism in DIO2 has been linked to reduced femoral neck BMD and increased levels of bone turnover markers, but the findings were not repeated in a larger study .

46.9.2

Epidemiological studies

Many observational studies have reported the consequences of altered thyroid function on BMD and fracture risk in population cohorts. Recently published studies have had larger cohort sizes, and some have included prospective follow-up. Larger cohorts reduce the potential for bias as many biological and environmental factors have large effects on BMD and fracture risk. Nonetheless, outcomes also depend on the composition of cohorts and may be confounded by inclusion of subjects with prior thyroid disease or varying combinations of pre- and postmenopausal female or male participants.

There have only been a few studies of the effects of thyroid status on bone turnover markers in euthyroid populations. In a study of 60 healthy postmenopausal women, high TSH levels were associated with low urinary deoxypyridinoline concentrations but not with serum procollagen type I C propeptide (PICP) levels . Two larger studies of postmenopausal women and men over 70 years of age , however, showed no association between TSH and serum osteocalcin, procollagen type 1 N-terminal propeptide (P1NP), beta-C-terminal telopeptide (CTX), and urinary N-terminal telopeptide (NTX).

The effect of variation of thyroid status within the normal reference range on BMD has recently been studied in a number of larger cohorts. Five cross-sectional studies of postmenopausal or elderly women with between 581 and 15,316 subjects found that low–normal TSH was associated with reduced hip or spine BMD . Three cross-sectional studies in men or both sexes (between 1017 and 1961 subjects) showed a consistent association between TSH and BMD in the forearm , lumbar spine , or hip . Only one recent study with 598 men and 719 women over 65 years of age showed no association between TSH levels and BMD . Other prospective and cross-sectional population studies (between 593 and 2957 subjects) have shown that high T4 is associated with reduced BMD or trabecular bone score , and a study of 4204 children showed that high–normal T4 is associated with lower BMD at 6 and 10 years of age . Although these studies have typically reported the association of either TSH or T4 with BMD, since both parameters are in a physiological inverse relationship, the results should be considered as an association with thyroid status rather than the effect of either TSH or T4 alone.

Low BMD is a major risk factor for fragility fractures, and so a number of studies have investigated associations between thyroid status and fracture risk directly. Recent prospective and cross-sectional studies of large female cohorts have shown that low TSH is associated with an increased risk of hip or vertebral fracture. Raised levels of T4 or T3 were also found to be associated with an increased risk of nonvertebral fracture by 20% and 33%, respectively . In one study of males aged over 65 years, low–normal TSH was associated with an increased risk of hip fracture . However, other population studies in postmenopausal or elderly women and/or men showed no significant association between TSH and fracture risk , including one large study of 16,610 women and 8595 men over 40 years of age . Despite these inconsistent data, a meta-analysis of individual participant data from 13 prospective studies of fracture risk in euthyroid individuals showed that low–normal TSH and high–normal T4 are associated with an increased risk of hip fracture . Importantly, this meta-analysis of individual participant data limited confounding due to study heterogeneity and led to a clear and robust conclusion.

In summary, these studies show that thyroid status in the upper–normal range with higher T4 and/or lower TSH is associated with decreased BMD and an increased risk of fracture.

46.10.1

Hypothyroidism in children

Congenital hypothyroidism is the most common congenital endocrine disorder with an incidence of approximately 1:2000 to 1:4000 newborns and causes severe effects on growth with epiphyseal dysgenesis, delayed bone age, growth arrest, and short stature . This complex skeletal dysplasia includes hip subluxation, immature vertebral bodies, epiphyseal dysgenesis, absent secondary ossification centers in the femoral heads, delayed closure of the fontanelles, and a flattened nasal bridge . These features are remarkably similar to the phenotype observed in patients with RTHα, clearly demonstrating the role of TRα in the human skeleton. If congenital hypothyroidism is treated early, thyroxine replacement induces rapid catch-up growth with potential to achieve predicted adult height and BMD , though in one report adult BMD remained low despite treatment from the neonatal period .

Children with juvenile acquired hypothyroidism similarly display growth arrest and delayed bone maturation, but thyroxine treatment induces catch-up growth . Nonetheless, these individuals may fail to achieve the final predicted height, depending on the duration of thyroid hormone deficiency prior to initiation of treatment. There are no clinical data on the long-term consequences to the adult skeleton in individuals treated during childhood and adolescence for juvenile hypothyroidism.

46.10.2

Hypothyroidism in adults

In newly diagnosed patients with severe hypothyroidism, the bone resorption markers urinary pyridinoline and deoxypyridinoline are decreased . Histomorphometric analysis of bone biopsies of patients presenting with hypothyroidism has shown that bone turnover is also decreased with a net increase in bone formation . However, in these analyses, there was no significant increase in trabecular bone volume because structural consequences of low bone turnover require a long period of time to become evident. Consequently, BMD has also been found to be normal in newly diagnosed patients with hypothyroidism

Patients treated for hypothyroidism have been shown to have an increased risk of fracture . Nevertheless, fracture risk was proportional to the cumulative time and degree of thyroxine overreplacement, suggesting that bone loss and fracture susceptibility in hypothyroidism does not result from the disease but rather from the consequences of its treatment .

46.10.3

Subclinical hypothyroidism

Subclinical hypothyroidism is defined as a raised TSH in the presence of circulating thyroid hormone levels within the normal reference range, and it most commonly results from autoimmune thyroiditis. It has a high population prevalence, affecting 3% of men and 8% of women (10% of women over 55 years of age) . Most patients have no symptoms, though some questionnaire-based studies have shown an increased prevalence of tiredness and other nonspecific neurocognitive symptoms . Recent guidelines support treatment of selected individuals with subclinical hypothyroidism with thyroxine , although elderly populations in whom the age-specific reference range for TSH is increased are the least likely to benefit from treatment, and they are also the most likely to suffer unwanted side effects affecting the heart and cardiovascular system .

In large individual participant data meta-analyses, subclinical hypothyroidism was neither associated with a change in BMD nor with an increased risk of fracture . Overall, subclinical hypothyroidism has no significant effect on BMD or fracture risk, and there is no evidence of benefit to the skeleton from intervention. Importantly, thyroid hormone replacement stimulates bone turnover and may result in a detrimental reduction in BMD .

46.11.1

Hyperthyroidism in children

Thyrotoxicosis is rare in children, and Graves’ disease is the most common cause. In juvenile thyrotoxicosis, skeletal development is accelerated resulting in early cessation of growth and short stature due to premature fusion of the growth plates. In very young children, premature closure of the cranial sutures can cause craniosynostosis, resulting in cognitive impairment . Even in the developing skeleton, increased thyroid hormones can cause reduced BMD and decreased cortical bone density , though BMD returns to its predicted values after 1–2 years of treatment with antithyroid medication .

46.11.2

Hyperthyroidism in adults

Histomorphometry data from bone biopsies in patients with active thyrotoxicosis show increased bone turnover with a net increase in bone resorption . In thyrotoxicosis, bone turnover markers are increased, including both resorption markers such as urinary pyridinoline and deoxypyridinoline, and formation markers such as alkaline phosphatase and osteocalcin . Severe osteoporosis due to uncontrolled thyrotoxicosis is now rare because of prompt treatment with antithyroid medications or definitive treatment with thyroidectomy or radioactive iodine in refractory cases. Nonetheless, studies have revealed that BMD is reduced at the time of diagnosis and that fracture risk is increased . Moreover, the increase in fracture risk is proportionate to the duration of hyperthyroidism prior to treatment and persists for up to 5 years after treatment .

46.11.3

Subclinical hyperthyroidism

Subclinical hyperthyroidism, defined as normal T4 and T3 levels in the presence of suppressed or undetectable circulating TSH, is less common than subclinical hypothyroidism and has a prevalence of 1% and 1.5% in men and women over 60 years of age respectively . Though patients with subclinical hyperthyroidism are typically asymptomatic, there are significant long-term health consequences. Individuals with a TSH concentration below 0.1 mIU/L have increased all-cause mortality, coronary heart disease-related deaths, and risk of atrial fibrillation .

In patients with subclinical hyperthyroidism, the resorption markers urinary deoxypyridinoline and pyridinoline, and formation markers osteocalcin, alkaline phosphatase, and procollagen I C-terminal extension propeptide may be elevated or normal . Large meta-analyses have demonstrated that BMD is decreased and that fracture risk is increased with the largest effects in postmenopausal women . One meta-analysis of five large cohort studies found an increased risk of fracture in elderly subjects only . In a sub-analysis of the Fracture Intervention Trial cohort of 15,316 postmenopausal women, subjects with TSH <0.5 mIU/L had an increased risk of vertebral fracture, although individuals with overt thyrotoxicosis were not formally excluded . Only one recent meta-analysis failed to demonstrate a significant association between subclinical hyperthyroidism and fracture risk, but in this study, a broad definition of subclinical hyperthyroidism was adopted that included individuals with TSH only partially suppressed to <0.45 mIU/L . One prospective cohort study showed an association with hip fracture in men (increased hazard ratio of 3.27) but no clear association in women; however, only 47 out of 3646 study subjects were in the higher risk category of individuals with complete suppression of TSH .

Overall, the evidence from large individual participant data meta-analyses demonstrates that subclinical hyperthyroidism is associated with detrimental consequences for BMD and fracture risk . The association with osteoporosis in these studies is consistent with the population effects of moderate increases in thyroid function within the normal range. In postmenopausal women the risk of subclinical hyperthyroidism causing osteoporosis is higher than in premenopausal women or men and is increased in proportion to the severity of TSH suppression.

46.11.4

Patients treated for hyperthyroidism

Hyperthyroidism results in a state of high bone turnover with a net increase in bone resorption that leads to osteoporosis. Various studies have examined the effects of treatment for thyrotoxicosis on the skeleton. Following treatment with antithyroid drugs, the elevated bone resorption and formation markers return to normal within 2 months . The low BMD at the time of diagnosis similarly returns to normal within 1–4 years , though if thyrotoxicosis is treated by thyroidectomy, improvement in BMD may occur as rapidly as 6 months following surgery . Despite the rapid effects of antithyroid medication on bone turnover and BMD, in a large case–control fracture risk study comprising half a million subjects, the increased relative risk of fracture persisted to 5 years after initiation of treatment and may result from inadequate treatment or noncompliance .

46.11.5

Thyroid-stimulating hormone suppression for treatment of differentiated thyroid cancer

Differentiated thyroid cancer comprises both papillary and follicular thyroid cancer, and treatment after total thyroidectomy and radioiodine ablation may require long-term suppression of serum TSH with thyroxine . Similar to other causes of subclinical hyperthyroidism, this exogenous cause carries the risk of osteoporosis and fragility fracture, in addition to an increased risk of atrial fibrillation and cardiovascular mortality .

Small cohorts of patients receiving suppressive doses of T4 have been studied in detail and were found to have increased levels of bone resorption and formation markers . Most studies of BMD have shown TSH suppression therapy to be associated with reduced BMD at the lumbar spine, femur or radius in postmenopausal women proportionate to the duration of TSH suppression . Despite this, TSH suppression therapy was not associated with change in BMD in male patients .

As differentiated thyroid cancer frequently has a favorable prognosis, international guidelines now recommend dynamic risk stratification of patients with the aim to limit the duration and severity of required TSH suppression . Patients with low or intermediate risk of recurrent disease should have moderate TSH suppression to <2.0 or 0.1–0.5 mIU/L, respectively . The highest risk patients should have target TSH levels of <0.1 mIU/L rather than complete suppression of TSH to undetectable levels. This is because complete suppression of TSH is associated with the highest risk of osteoporotic fractures, with no additional reduction in risk of tumor recurrence . Since suppressive doses of thyroxine have a greater effect on the skeleton in postmenopausal women , monitoring of BMD has been recommended in this group of patients .