The Thyroid and its Diseases
	Chapter 16 Section 16 A--TSH Receptor Diseases,  Section 16 B--Defects in Thyroid Hormone Supply, Section 16 C--Abnormal Thyroid Hormone Transport Section 16 D--Thyroid Hormone Resistance Syndromes TO OBTAIN A DOWNLOAD OF THIS CHAPTER IN PDF OR WORD FORMAT,  CLICK HERE

Chapter 16D
Resistance to Thyroid Hormone
Samuel Refetoff, M.D., September 28, 2004

Resistance to thyroid hormone (RTH) is a syndrome of reduced responsiveness of target tissues to thyroid hormone (TH). More than 1,000 cases have been described that appear to fit this definition. In practice, such patients are identified by their persistent elevation of circulating free thyroxine (FT₄) and free triiodothyronine (FT₃) levels in association with nonsuppressed serum thyrotropin (TSH), and in the absence of intercurrent illness, drugs, or alterations of TH transport serum proteins. More importantly, higher doses of exogenous TH are required to produce the expected suppressive effect on the secretion of pituitary TSH and on the metabolic responses in peripheral tissues.

Although the apparent resistance to TH may vary in severity, it is always partial. The variability in clinical manifestations may be due to the severity of the hormonal resistance, the effectiveness of compensatory mechanisms, the presence of modulating genetic factors, and the effects of prior therapy. The magnitude of the hormonal resistance is, in turn, dependent on the nature of the underlying genetic defect. The latter is usually but not always, a mutation in the TH receptor (TR) ß gene (1, 2)

Despite a variable clinical presentation, the common features characteristic of the RTH syndrome are: 1) elevated serum levels of free T₄ and T₃, (2) normal or slightly increased TSH level that responds to TRH, (3) absence of the usual symptoms and metabolic consequences of TH excess, and (4) goiter.

CLINICAL CLASSIFICATION

The diagnosis is based on the clinical findings and standard laboratory tests and, when possible, confirmed by genetic studies. Before TR gene defects were recognized, the proposed subclassification of RTH was based on symptoms, signs and laboratory parameters of tissue responses to TH (3). Not withstanding the assessment of TSH feedback regulation by TH, the measurements of most other responses to the hormone are insensitive and relatively nonspecific. For this reason, all tissues other than the pituitary have been grouped together under the term peripheral tissues, on which the impact of TH was roughly assessed by a combination of clinical observation and laboratory tests.

The majority of patients appeared to be eumetabolic and maintained a near normal serum TSH concentration. They were classified as having generalized resistance to TH (GRTH). In such individuals, the defect seemed to be compensated by the high levels of TH. In contrast, patients with equally high levels of TH and nonsuppressed TSH that appeared to be hypermetabolic, because they were restless or had sinus tachycardia, were classified as having selective pituitary resistance to TH (PRTH). Finally, the occurrence of isolated peripheral tissue resistance to TH (PTRTH) was reported in a single patient studied in detail (4). In this individual with partial thyroid gland ablation, although serum TSH was suppressed with physiological doses of liothyronine (L-T₃), supraphysiological doses of this hormone failed to produce symptoms and signs of thyrotoxicosis or to increase oxygen consumption and pulse rate. No mutation in the TRß gene of this patient was found (5) and no similar cases have been reported. More common in clinical practice is the apparent tolerance of some individuals to the ingestion of supraphysiological doses of TH.

The earliest suggestion that PRTH may not constitute an entity distinct from GRTH can be found in a study by Beck-Peccoz et al (6). In a group of 15 patients with PRTH, these authors found that the serum level of the peripheral tissue marker of TH action, sex hormone-binding globulin (SHBG), was not increased as it should have been if the hormonal resistance were confined to the pituitary and hypothalamus. In a recent, comprehensive study involving 312 patients with GRTH and 72 patients with PRTH, it was conclusively shown that the response of SHBG and other peripheral tissue markers of TH action, were equally attenuated in GRTH and PRTH (7). The frequency of symptoms and signs were also similar in patients under both classifications. More importantly, identical mutations were found in individuals classified as having GRTH and PRTH, many of whom belonged to the same family (8). It was, therefore, concluded that these two forms of RTH are the product of the subjective nature of symptoms and poor specificity of signs (see section on the Molecular Basis of the Defect). Thus, it is uncertain whether PRTH exist as a true TH resistance entity, except in association with TSH-producing pituitary adenomas (9, 10).

INCIDENCE AND INHERITANCE

The precise incidence of RTH is unknown. Because routine neonatal screening programs are based on the determination of TSH, RTH is rarely identified by this means (11). A limited neonatal survey by measuring blood T₄ concentration, suggested the occurrence of one case per 40,000 live births (12). Currently published cases surpass 1,000 of which 349 have been previously reviewed in detail (1).

Although most thyroid diseases occur more commonly in women, RTH has been found with equal frequency in both genders. The condition appears to have wide geographic distribution and has been reported in Caucasians, Africans, and Asians. The prevalence may vary among different ethnic groups.

Familial occurrence of RTH has been documented in approximately 75% of cases. Taking into account only those families in whom both parents of the affected subjects have been studied, the true incidence of sporadic cases, is 21.3%. This is in agreement with current estimate of the frequency of de novo mutations of 22.5% (Table 1). The reports of acquired RTH are seriously questioned.

Inheritance is autosomal dominant. Transmission was clearly recessive in only one family (27, 28). Consanguinity in a family with dominant inheritance of RTH has produced a homozygous child with very severe resistance to the hormone (29) who died at the age of 7 years. There is only partial information on another patient with very severe RTH, homozygote for a TRß gene mutation (30).

ETIOLOGY AND GENETICS

Using the technique of restriction fragment length polymorphism (RFLP), Usala et al (31) were first to demonstrate linkage between a TRß locus on chromosome 3 and the RTH phenotype, 21 years after description of the syndrome (27). Subsequent studies at the University of Chicago and at the National Institutes of Health have identified distinct point mutations in the TRß gene of two unrelated families with RTH (32, 33). In both families only one of the two TRß alleles was involved, compatible with the apparent dominant mode of inheritance. These mutations resulted in the substitutions of single amino acids in the T₃-binding domain of the TRß.

Mutations in the TRß gene have now been identified in subjects with RTH belonging to 300 families (Table 1; Fig.1). They comprise 122 different mutations. With the exception of the index family, found to have complete deletion of the TRß gene (28), all others harbor minor alterations at the DNA level. The majority (285 families) have single nucleotide substitutions resulting in single amino acid replacements in 281 instances and stop codons in 4 others, producing truncated molecules of 433-448 amino acids in the TRß1 protein. In two families dinucleotide substitutions produce a single amino acid substitution and a premature stop, respectively. Six families have trinucleotide deletions producing the loss of single amino acid, while insertion of a trinucleotide produced the addition of a single amino acid in another family. A single nucleotide deletion in 1 family and an insertion in 4 others, resulted in a frameshift, the latter producing a protein containing two extra amino acids.

Figure 1. Location of natural mutations in the TRß molecule associated with RTH. TOP PORTION: Schematic representation of the TRß and its functional domains for interaction with TREs (DNA-binding), with hormone (T3-binding), with activating (13), repressing (14-16) cofactors and with nuclear receptor partners (dimerization) (17-20). Their relationship to the three clusters of natural mutations is also indicated. BOTTOM PORTION: The T3-binding domain and distal end of the hinge region, which contain the three mutation clusters, are expanded and show the positions of CpG dinucleotide mutational "hot spots" in the corresponding RTß gene. The location of the 121 different mutations detected in 299 unrelated families (published and our unpublished data) are each indicated by a symbol. Identical mutations in members of unrelated families are represented by the same shading pattern of vertically placed symbols. "Cold regions" are areas devoid of mutations associated with RTH. Amino acids are numbered consecutively starting at the amino terminus of the TRß1 molecule according to the consensus statement of the First International Workshop on RTH (21). TRß2 has 15 additional residues at the aminoterminus. AF2, Hormone-dependent activation function (12th amphipatic helix) (22, 23) RBE, corepressor-binding enhancer; RBI, corepressor-binding inhibitor (23) SSD, silencing subdomain (16) NucL, nuclear localization (24) SigM, signature motif (25) [Modified from Refetoff et al (26)]

Given that there are 178 more families than the number of different mutations, 45 of the mutations are shared by more than one family. Haplotyping of intragenic polymorphic markers showed that, in most instances, identical mutations have developed independently in different families (34). Eleven of these occur in more than 6 apparently unrelated families and, with the exception of two, all others occur in mutagenic CpG dinucleotide hot spots; the mutation R338W (CGG -> TGG) has been identified in 25 different families. In agreement with this finding is the relatively high prevalence (16.7%) of de novo mutations. In addition different mutations producing more than one amino acid substitution at the same codon have been found at 37 different sites; mutations at codon 345 and 453 each produced 5 different amino acid replacements, G345R,S,A,V,D and P453T,S,A,Y,H.

All TRß gene mutations are localized in the functionally relevant domain of T₃-binding and its adjacent hinge region. Three mutational clusters have been identified with intervening cold regions (Fig. 1). With the exception of the family with TRß gene deletion, in all others inheritance is autosomal dominant. No mutations have so far been detected in the TRa gene. Recent findings in mice with targeted mutations in the TRa gene (knock in mice) indicate that such mutations do not produce the phenotype of RTH (35-37) (see animal models of RTH, below). The significance of somatic TRa gene mutations identified in non functioning pituitary adenomas (38) is unclear.

Somatic mutations in the TRß gene have been identified in some TSH-secreting pituitary tumors (9, 10). These mutations can be identical to those occurring in the germline. However, because their expression is limited to the thyrotrophs, the phenotype, as in other TSHomas, is that of TSH induced thyrotoxicosis. It is postulated that defective TR interfering with the negative regulation of TSH by TH is responsible for the development of the pituitary tumor.

In a growing number of individuals, RTH occurs in the absence of mutations in the TRa or TRß genes (nonTR-RTH) (39). They represent 15% of families and 7% of cases with RTH. Such individuals may have a defect in one of the cofactors involved in the mediation of TH action (see animal models of RTH, below).

MOLECULAR BASIS OF THE DEFECT

Thyroid Hormone Action at the Gene Level

Genes on chromosome 17 and 3, each generate TRa and TRß molecules, respectively, that have substantial structural and sequence similarities. Both genesw produce two isoforms; a1 and a2 by alternative splicing and ß1 and ß2 by different transcription start points. TRa2 binds to TH response elements (TREs) but due to a sequence difference at the ligand-binding domain (LBD) site, it does not bind TH and thus does not function as a proper TR (40) and appears to have a weak antagonistic effect (41). Additional TR isoforms, including a TRß with shorter amino terminus (TRß3) and truncated TRß3, TRa1 and TRa2 [lacking the DNA-binding domain (DBD)], have been identified in different tissues of rodents (42, 43). The latter three exhibit dominant antagonistic effect but their significance in humans remains unknown (44). Finally, a p43 protein, translated from a downstream AUG of TRa1, is believed to mediate the TH effect in mitochondria (45).

The relative expression of the two TR genes and the distribution of their products vary among tissues and during different stages of development (46-48). Recent studies have shown several splice variants involving the 5'-untranslated region of the human TRß1 (49, 50). Their relative abundance is developmentally and tissue regulated, thus potentially controlling the expression of the receptor protein. To a certain degree, TRß and TRa are interchangeable (51, 52). However, the resulting compensatory effects, observed in the absence of one of the receptors, do not represent functional equivalence during normal physiological conditions. Some TH effects are absolutely TR isoform specific (see animal models of RTH, below). It is of interest that nuclear receptors, which arose from two waves of gene duplication during the evolution of metazoans, were initially orphan receptors that subsequently acquired ligand-binding ability (53).

TREs, located in TH regulated genes, consist of half-sites having the consensus sequence of AGGTCA and vary in number, spacing and orientation (54, 55). Each half-site usually binds a single TR molecule (monomer) and two half-sites bind two TRs (dimer) or one TR and a heterologous partner (heterodimer), the most prominent being the retinoid X receptor ß(RXR). Dimer formation is facilitated by the presence of an intact "leucine zipper" motif located in the middle of the LBD of TRs. Occupation of TREs by unliganded (without hormone) TRs, also known as aporeceptors, inhibits the constitutive expression of genes that are positively regulated by TH (56) through association with corepressors such as the nuclear corepressor (NCoR) or the silencing mediator of retinoic acid and TH receptors (SMRT) (57). Transcriptional repression is mediated through the recruitment of the mammalian homologue of the Saccaromyces transcriptional corepressor (mSin3A) and histone deacetylases (HDAC) (58). This latter activity compacts nucleosomes into a tight and inaccessible structure, effectively shutting down gene expression. This effect is relieved by the addition of TH, which releases the corepressor, reduces the binding of TR dimers to TRE, enhances the occupation of TREs by TR/RXR heterodimers (59) and recruits coactivators (CoA) such as p/CAF (CREB binding protein-associated factor) and NCoA [steroid receptor coactivator-1 (SRC-1)] with HAT (histone acetylation) activity (57, 60). This results in the loosening of the nucleosome structure making the DNA more accessible to transcription factors. Actually, the ligand-dependent association with TR associated proteins (TRAP), in conjunction with the general coactivators PC2 and PC4, act to mediate transcription by RNA polymerase II and general initiation factors (61). TR dimerization is not required for hormone-binding and the latter does not induce dimerization. Furthermore, it is believed that T₃ exerts its effect by inducing conformational changes of the TR molecule and that TRAP stabilizes the association of TR with TRE.

Properties of Mutant TRß Receptors and Dominant Negative Effect

TRß gene mutations produce two forms of RTH. The less common, described in only one family (27), is caused by deletion of all coding sequences of the TRß gene and is inherited as an autosomal recessive trait (28). In addition to the hormonal and metabolic abnormalities typical of the syndrome of RTH, these individuals have severe deafness resulting in mutism (27), as well as monochromatic vision (62). This is caused by the complete lack of TRß, which is required for the cochlear maturation and the development of cone photoreceptors that mediate color vision (63) (see animal models of RTH, below). Heterozygous individuals that express a single TRß gene have no clinical or laboratory abnormalities. This is not due to compensatory overexpression of the single normal allele of the TRß gene nor that of the TRa gene (64). However, because some responses to TH could be demonstrated in subjects homozygous for TRß gene deletion, it is logical to conclude that TRa1 is capable of partially substituting for the function of TRß (see animal models of RTH, below).

The more common form of RTH is inherited in a dominant fashion and is characterized by minor defects in one allele of the TRß gene, principally missense mutations. This is in contrast to individuals that lack one allele of the TRß that do not exhibit the RTH phenotype. These findings indicated that RTH is not simply the consequence of a reduced amount of a functional TR (haploinsufficiency) but is caused by the interference of the mutant TR (mTR) with the function of the wild-type (WT)-TR (dominant negative effect; Fig. 2). This has been clearly demonstrated in experiments in which mTRs are coexpressed with WT-TRs (65, 66). The most severe resistance to the action of TH observed in one child homozygous for a TRß mutation (67) underscores the role of dominant negative effect exerted by the expression of two mTR alleles in the pathogenesis of RTH.

Figure 2. Mechanism of the dominant expression of RTH: In the absence of T3, occupancy of TRE by TR heterodimers (TR-TRAP) or dimers (TR-TR) suppresses transactivation through association with a corepressor (CoR). (A) T3-activated transcription mediated by TR-TRAP heterodimers involves the release of the CoR and association with coactivators (CoA) as well as (B) the removal of TR dimers from TRE releases their silencing effect and liberates TREs for the binding of active TR-TRAP heterodimers. The dominant negative effect of a mutant TR (mTR), that does not bind T3, can be explained by the inhibitory effect of mTR-containing-dimers and heterodimers that occupy TRE. Thus, T3 is unable to activate the mTR-TRAP heterodimer (A') or release TREs from the inactive mTR homodimers (B'). [Modified from Refetoff et al (1)].

Studies carried out during the last decade have established two basic requirements for mTRs to exert a dominant negative effect. These include, (1) preservation of binding to TREs on DNA and (2) the ability to dimerize with a homologous (68-70) or heterologous (20, 71) partner. These criteria apply to mTRs with predominantly impaired T₃-binding activity (Fig. 2). In addition, a dominant negative effect can be exerted through impaired association with a cofactor even in the absence of important impairment of T₃-binding. Increased affinity of a mTR to a corepressor (CoR) (72, 73), or reduced association with a coactivator (CoA) (74-76), have been found to play a role in the dominant expression of RTH. These conclusions are based on direct experimental evidence as well as observations that correlate the location of mutations on the receptor molecule and the clinical consequences. In the first instance, the introduction in a mTR of an additional artificial mutation that abolishes either DNA binding, dimerization or the association with a CoR results in the abrogation of its dominant negative effect (71, 77, 78).

Examination of the distribution of TRß mutations associated with RTH reveals conspicuous absence of mutations in regions of the molecule that are important for dimerization, for the binding to DNA and for the interaction with CoR (Fig. 1). These "cold regions" are not devoid of CpG hot spots, suggesting that these regions of the molecule may not be devoid of natural mutations. However, they would escape detection owing to their failure to produce clinically significant RTH in heterozygotes. This has been indirectly deduced from in vitro studies with mTRßs harboring artificial mutations placed in the CpGs of the cold region 1 (79). Structural studied of the DBD and LBD have provided further understanding regarding the clustered distribution of mTRßs associated RTH and defects in the association with cofactors (80-84) .

Based on the early finding that RTH is associated with mutations confined to the LBD of the TRß it was anticipated that the clinical severity of RTH will correlate with the degree of T₃-binding impairment. In fact some studies appeared to be in agreement with this hypothesis (85). Other investigators found a better correlation with the potency of the dominant negative effect of the mTR assessed in vitro (70, 86). Examination of 18 different mTRßs suggested that both opinions are correct (87). The serum free T₄ concentration, used as indicator of thyrotroph hyposensitivity to TH, correlated with the degree of T₃-binding impairment of the corresponding mutant receptor in 12 of these mTRßs, designated as group I. This correlation was not found in 6 of the mTRßs studied; a discrepancy explained in some of them by the demonstration of reduced dominant negative potency due to diminished ability to form homodimers (for example R316H and E338W) (70, 87). Weakened association with DNA of CoR can produce the same effect.

While reduced dominant negative effect can explain a mild impairment of function in the heterozygote, despite severe impairment of T₃-binding, the reason for the occurrence of the reverse situation was less readily apparent. Indeed, more severe RTH and interference with the function of the WT-TRß, despite very mild impairment of T₃-binding or no binding defect at all, has been also observed (88). Two such mTRßs, R243Q and R243W, located in the hinge region of the receptor, have no significant impairment of T₃-binding when tested in solution yet, both clinically and in vitro manifested relatively severe RTH and impairment of transactivation function, respectively. The demonstration of normal nuclear translocation but reduced ability of T₃ to dissociate homodimers formed on TRE suggested that these mutant TRßs have reduced affinity for T₃ only after they bound to DNA (88). This has been confirmed by measurement of T₃-binding after complexing to TRE (89). Another mTRß, L454V located in the AF2 domain (Fig. 1) and with near normal T₃-binding, exhibited altered transcriptional function and RTH because of attenuated interaction with the CoA (74). Finally, some mTRßs, such as R383H, exhibit a delay in CoR release despite minimal reduction of T₃-binding (90).

In general the relative degree of impaired function among various mTRßs is similar whether tested using TREs controlled reporter genes that are negatively or positively regulated by T₃. Exceptions to this rule are the mTRßs, R383H and R429Q that show greater impairment of transactivation on negatively than positively regulated promoters (87, 90, 91). The reason for this discrepancy is a matter of conjecture. Given the fact that R429Q binds T₃ normally, it is possible that T3-binding to these mTRßs is allosterically modulated by the different TREs and cofactors (92, 93).

MOLECULAR BASIS OF THE VARIABLE PHENOTYPE OF RTH

The very extremes of the RTH phenotype have a clear molecular basis. Subjects heterozygous for a TRß gene deletion are normal because the expression of a single TRß allele is sufficient for normal function. RTH manifests in homozygotes completely lacking the TRß gene and in heterozygotes that express a mTRß with dominant negative effect. The most severe form of RTH, with extremely high TH levels and signs of both hypothyroidism and thyrotoxicosis, occur in a homozygous individual expressing only a mutant TRs (29, 67). The severe hypothyroidism manifesting in bone and brain of this subject can be explained by the silencing effect of a double dose mTR and its strong interference with the function of TRa (94); a situation which does not occur in homozygous subjects with TRß deletion. In contrast, the manifestation of thyrotoxicosis in other tissues, such as the heart, may be explained by the effect high TH levels have on tissues that normally express predominantly TRa1 (95, 96) (see animal models of RTH, below). It is for the same reason that tachycardia is a relatively common finding in RTH (97).

Various mechanism can be postulated to explain the tissue differences in TH resistance within the same subject and among individuals. The distribution of receptor isoforms varies from tissue to tissue (46, 98, 99). This likely accounts for greater hormonal resistance of the liver as compared to the heart. Differences in the degree of resistance among individuals harboring the same mTRß could be explained by the relative level of mutant and WT-RT expression. Such differences have been found in one study (100) but not in another (64).

Although in a subset of mTRßs a correlation was found between their functional impairment and the degree of thyrotroph hyposensitivity to TH, it is surprising that this correlation was not maintained with regards to the hormonal resistance of peripheral tissues (87). Subjects with the same mutations, even belonging to the same family, showed different degrees of RTH. A most striking example is that of family G.H. in which the mTRß R316H did not cosegregate with the RTH phenotype in all family members (101). This variability of clinical and laboratory findings was not observed in affected members of two other families with the same mutation (7, 102). A study in a large family with the mTRß R320H, suggests that genetic variability of factors other than TR may modulate the phenotype of RTH (103).

RTH WITHOUT TR GENE MUTATIONS (nonTR-RTH)

The molecular basis of nonTR-RTH remains unknown. Since the first demonstration of nonTR-linked RTH (2), 29 subjects belonging to 23 different families have been identified (39, 104, 105). The phenotype is indistinguishable from that in subjects harboring TRß gene mutations (see differential diagnosis, below). Distinct features are an increased female to male ratio of 2.5:1 and the high prevalence of sporadic cases. As a matter of fact of the 21 families in whom both parents, all sibling and progeny were examined, the occurrence of RTH in another family member was documented in only 4. In those instances, and as in the case of TRß-linked RTH, the inheritance pattern is dominant. While it has been postulated that nonTR-RTH is likely caused by a defect in one of the cofactors involved in the mediation of TH action, proof supporting this contention is lacking. Indeed a search for abnormalities in several corepressors, coactivators, coregulators and a TH membrane transporter, yielded negative results (106, 107).

ANIMAL MODELS OF RTH

TRß Gene Manipulations

Understanding of TH action in vivo, and in particular the mechanisms underlying abnormalities observed in patients with RTH, have been bolstered by observations made in genetically manipulated mice. Three types of genetic manipulations have been applied: (a) transgenic mice that overexpress a receptor; (b) deletion of the receptor (knockout or KO mice); and (c) introduction of mutations in the receptor (knockin or KI mice). The latter two types of gene manipulation, species differences not withstanding, have yielded true models of the recessively and dominantly inherited forms of RTH, respectively (108).

The features of RTH found in patients homozygous for TRß deletion also manifest in the TRß deficient mouse (109-111) (Table 2). Special features, such as sensorineural deafness and monochromatic vision are characteristic and shared by mouse (112) (113) and man (27, 62). This indicates that the lack of TRß per se, rather than deletion of other genetic material, is responsible for the deaf mutism and color blindness in affected members of the family (28). The mouse model allowed for investigations in greater depth into the mechanisms responsible for the development of these abnormalities. Thus, TRß1 deficiency retards the expression of fast-activating potassium conductance in inner hair cells of the cochlea that transforms the immature cells from spiking pacemaker to high-frequency signal transmitters (114). TRß2 interacts with transcription factors involved in photoreceptor development to provide timed and spatial order for cone differentiation resulting in the selective loss off M-opsin (113). The down regulation of hypothalamic TRH is also TRß2 specific (115). Mice deficient in TRß have increased heart rate that can be decreased to the level of the WT mouse by reduction on the TH level (111). This finding, together with the lower heart rate in mice selectively deficient in TRa1 (116), indicates that TH dependent changes in heart rate are mediated through TRa, and explains the tachycardia observed in some patients with RTH. TRß is also required for the T3-mediated regulation of hepatic cholesterol metabolism, which cannot be compensated by TRa1 (117).

The combined deletion of TRa1 and a2, produces no important alterations in TH or TSH concentrations in serum (51). However, expression of the carboxyl terminal fragment of the TRa, due to the presence of a natural promoter and a transcription start site in intron 5, produces in the context of deficient intact TRa, severe neonatal hypothyroidism that is lethal but can be rescued by short term treatment with TH (118). The complete lack of TRs, both a and ß, is compatible with life (51, 52). This contrasts with the complete lack of TH which, in the athyreotic Pax8 deficient mouse, results in death prior to weaning, unless rescued by TH treatment (119). The survival of the combined TR deficient mice is not due to expression of a yet unidentified TR but to the absence of the noxious effect of aporeceptors. Indeed, removal of the TRa gene rescues the Pax8 KO mice from death (120). The combined TRß and TRa deficient mice have serum TSH levels that are 500-fold higher than those of the WT mice and T₄ concentrations 12-fold above the average normal mean (51). Thus, in this instance, the presence of an aporeceptor does not seem to be required for the upregulation of TSH.

The first, partial model of the dominantly inherited RTH utilized somatic gene transfer of a mutant TRß1, G345R, by means of a recombinant adenovirus (121). The liver of these mice was resistant to TH, as demonstrated by the reduced responsiveness of TH controlled genes to the administration of L-T₃. Overexpression of the WT TRß increased the severity of hypothyroidism in the TH deprived mouse, confirming that the unliganded TR has a constitutive effect in vivo as in vitro. Similarly, the response to TH is enhanced in animals that overexpress the WT TR. Transgenic mice have also been developed that express mTRß1 genes (122). True mice models of dominantly inherited RTH were recently generated by targeted mutations in the Trß gene (123, 124). Mutations were modeled on those formerly identified in humans with RTH (frame-shift resulting in 16 carboxylterminal nonsense amino acids (PV mouse) and T337D). As in humans, the phenotype manifested in the heterozygous KI animals and many of the abnormalities observed in man were reproduced in these mice (Table 2).

Recent work has shown that NcoA (SRC-1) deficient mice have RTH manifesting with elevated T₄, T₃ and TSH concentrations (125). A more mild form of RTH was identified in mice deficient in RXRg (126). Animals show reduced sensitivity to T3 in terms of TSH downregulation but not metabolic rate These data indicate that abnormalities in cofactors can produce RTH.

The Phenotype of TR

The question of why mutations in the TRa gene have not been identified in man was partially answered by the study of mice with targeted gene manipulations. As stated in the preceding section, TRa gene deletions, total or only a1, failed to produce a RTH phenotype. Similarly, mice with targeted TRa gene mutations failed to manifest the phenotype of RTH. Human mutations known to occur in the TRß gene were targeted in homologous regions of the TRa gene of the mouse. These were, the PV frame-shift mutation in the carboxylterminus of TRa1 (35), TRa1 R348C [equivalent to TRß R438C (36)] and P398H [equivalent to TRß P453H (37)]. The resulting phenotypes were variable but did not exhibit RTH. In the heterozygous state, the former two show sever retardation in post-natal development and growth while the latter has an increase in body fat and insulin resistance. Decreased heart rate and cold-induced thermogenesis, as well as reduced fertility, were also observed. The TRaPV KI mice exhibited severe reduction in brain glucose utilization and synaptic density (127, 128). All three TRaKIs were lethal in the homozygous state, recapitulating the noxious effect of unliganded TRa1.

PATHOGENESIS

The reduced sensitivity to TH in subjects with RTH is shared to a variable extent by all tissues. The hyposensitivity of the pituitary thyrotrophs results in nonsuppressed serum TSH, which in turn, increases the synthesis and secretion of TH. The persistence of TSH secretion in the face of high levels of free TH contrasts with the low TSH levels in the more common forms of TH hypersecretion that are TSH-independent. This apparent paradoxical dissociation between TH and TSH is responsible for the wide use of the term "inappropriate secretion of TSH" to designate the syndrome. However, TSH hypersecretion is not at all inappropriate, given the fact that the response to TH is reduced. It is compensatory and appropriate for the level of TH action mediated through a defective TR. As a consequence most patients are eumetabolic, though the compensation is variable among affected individuals and tissues in the same individual. However, the level of tissue responses do not correlate with the level of TH, probably owing to a discordance between the hormonal effect on the pituitary and other body tissues. Thyroid gland enlargement occurs with chronic, though minimal, TSH hypersecretion due to increased biological potency of this glycoprotein through increased sialylation (129). Administration of supraphysiologic doses of TH are required to suppress TSH secretion without induction of thyrotoxic changes in peripheral tissues.

When sought, TSH-binding antibodies could not be detected (130, 131). Thyroid-stimulating antibodies, which are responsible for the thyroid gland hyperactivity in Graves' disease, have been conspicuously absent in patients with RTH. Another potential thyroid stimulator, human chorionic gonadotropin, is also not involved (132, 133).

The selectivity of the resistance to TH has been convincingly demonstrated. When tested at the pituitary level, both thyrotrophs and lactotrophs were less sensitive only to TH. Thyrotrophs responded normally to the suppressive effects of the dopaminergic drugs L-dopa and bromocriptine (131, 134) as well as to glucocorticoids.(131, 135, 136). Studies carried out in cultured fibroblasts confirm the in vivo findings of selective resistance to TH. The responsiveness to dexamethasone, measured in terms of glycosaminoglycan (137) and fibronectin synthesis (138), was preserved in the presence of T3 insensitivity (Fig. 3).

Figure 3. The inhibitory effect of T3 and dexamethasone on glycosaminoglycan synthesis in fibroblasts from subjects with TH and glucocorticoid resistance compared to that in fibroblasts from a normal (nonresistant) subject. Note that fibroblasts from the patient with RTH had an attenuated response to T3 but not to dexamethasone, while those from the patient with glucocorticoid resistance showed an attenuated response to dexamethasone only. [From Refetoff et al (1)]

Several of the clinical features encountered in some patients with RTH may be the manifestation of selective tissue deprivation of TH during early stages of development. These clinical features include retarded bone age, stunted growth, mental retardation or learning disability, emotional disturbances, attention deficit/hyperactivity disorder (ADHD), hearing defects, and nystagmus (1). A variety of associated somatic abnormalities appear to be unrelated pathogenically and may be the result of involvement of other genes such as in major deletions of DNA sequences (28). However, no gross chromosomal abnormalities have been detected on karyotyping (27, 139).

PATHOLOGY

Little can be said about the pathologic findings in tissues other than the thyroid gland because of unavailability of autopsy data from patients with RTH. Electron microscopic examination of striated muscle obtained by biopsy from one patient revealed mitochondrial swelling, also known to be encountered in thyrotoxicosis (27). This is compatible with the predominant expression of TRa in muscle (140), resulting in a low level of expression of the mutatnt TRß and unobstructed action of the excessive amount of circulating TH available to this tissue. Light microscopy of skin fibroblasts stained with toluidine blue showed moderate to intense metachromasia (141) as described in myxedema and is probably a manifestation of a decreased TH action in this tissue. However, in contrast to patients with myxedema due to TH deficiency, treatment with the hormone failed to induce a disappearance of the metachromasia in fibroblasts from patients with RTH.

Thyroid tissue, obtained by biopsy or at surgery, revealed various degrees of hyperplasia of the follicular epithelium (130, 131, 142, 143). The follicles may vary in size, from small to large. Some specimens have been described as "adenomatous goiters", others as "colloid goiters, and still others as normal thyroid tissue. Occasional lymphocytic infiltration, a finding not related to the syndrome, is due to the fortuitous coexistence of thyroiditis (144).

CLINICAL FEATURES

Characteristic of the RTH syndrome is the paucity of specific clinical manifestations. When present, manifestations are variable from one patient to another. Investigation leading to the diagnosis has been undertaken because of the presence of goiter, hyperactive behavior or learning disabilities, developmental delay and sinus tachycardia (Fig. 4). The finding of elevated serum TH levels in association with nonsuppressed TSH is usually responsible for the pursuit of further studies leading to the diagnosis.

Figure 4. The reasons prompting further investigation of the key member of each family with RTH.

The majority of untreated subjects maintain a normal metabolic state at the expense of high levels of TH. The degree of this compensation of tissues hyposensitivity to the hormone is, however, variable among individuals as well as in different tissues. As a consequence, clinical and laboratory evidence of TH deficiency and excess often coexist. For example, RTH can present with a mild to moderate growth retardation, delayed bone maturation and learning disabilities suggestive of hypothyroidism, along with hyperactivity and tachycardia, compatible with thyrotoxicosis. The more common clinical features and their frequency are given in Table 3. Frank symptoms of hypothyroidism are more common in those individuals who, because of erroneous diagnosis, have received treatment to normalize their circulating TH levels. In such patients, symptoms of fatigue, somnolence, depression, weight gain and bradycardia were noted. In children inappropriate treatment has aggravated the delay in growth and development (1).

On physical examination, goiter is by far the most common abnormality. It has been reported in 66-95% of cases. In some patients, without clinically obvious thyroid gland enlargement, goiter could be detected by ultrasonography, was absent due to prior surgery, or was present in other affected members of the family. Gland enlargement is usually diffuse; nodular changes and gross asymmetry are found in recurrent goiters after surgery.

Sinus tachycardia is also very common, with some studies reporting frequency as high as 80% (7). Palpitations often bring the patient to the physician and the finding of tachycardia is the most common reason for the erroneous diagnosis of autoimmune thyrotoxicosis or the suspicion of PRTH.

Careful evaluation of subjects with RTH has shown that about one-half have some degree of learning disability with or without ADHD (1, 146). One-quarter have intellectual quotients (IQ) lesser than 85% but frank mental retardation (IQ <60) has been found only in 3% of cases. Impaired mental function was found to be associated with impaired or delayed growth (<5th percentile) in 20% of subjects though growth retardation alone is rare (4%) (1). Despite the high prevalence of ADHD in patients with RTH, the occurrence of RTH in children with ADHD must be very rare, none having been detected in 330 such children studied (147, 148). Current data do not support a genetic linkage of RTH with ADHD. Rather the higher prevalence of low IQ scores may confer a higher likelihood for subjects with RTH to exhibit ADHD symptoms (102).

A variety of physical defects that cannot be explained on the basis of TH deprivation or excess have been recorded. These include major or minor somatic defects, such as winged scapulae, vertebral anomalies, pigeon breast, prominent pectoralis, birdlike facies, scaphocephaly, craniosynostosis, short 4th metacarpals, as well as Besnier's prurigo, congenital ichthyosis, and bull's eye type macular atrophy (1). No particular defect appears to be prevalent in RTH. A distinct body habitus and deaf-mutism occurred in all three affected members of a single family with TRß gene deletion (27).

COURSE OF THE DISEASE

The course of the disease is as variable as is its presentation. Some subjects have normal growth and development, and lead a normal life at the expense of high TH levels and a small goiter. Others present variable degrees of mental and growth retardation. Symptoms of hyperactivity tend to improve with age but this is not characteristic of RTH since it has been also observed in subjects with ADHD only.

Goiter has recurred in every patient who underwent thyroid surgery. As a consequence, some subjects have been submitted to several consecutive thyroidectomies or treatments with radioiodide (143, 149-151).

Only in one patient is RTH believed to have contributed to his demise. This child, homozygous for a dominantly inherited TRß mutation and a resting heart rate of 190 beats/minute (29), died from cardiogenic shock complicating staphiloccocal pneumonia (B. B. Bercu, personal communication).

LABORATORY FINDINGS

Thyroid Hormone and its Metabolites in Serum

In the untreated patient, elevation in the concentration of serum free T₄ is a sine qua non requirement for the diagnosis of RTH. It is accompanied by high serum levels of T₃. Serum TBG and TTR concentrations are normal. The resin T₃ uptake is usually high as is the case in patients with thyrotoxicosis.

Serum T₄ and T₃ values vary from just above to several fold the upper limit of normal. Although the levels may vary in the course of time in the same patient (9), the degree of T₄ and T₃ elevation is usually congruent, resulting in a normal T₃:T₄ ratio (1). This is in contrast to the disproportionate increase in serum T₃ concentration relative to that of T₄, characteristic of autoimmune thyrotoxicosis (152). This observation has led some investigators to consider the possibility of reduced extrathyroidal conversion of T₄ to T₃ in patients with RTH (143).

Reverse T₃ concentration is also high in patients with RTH as are those of another product of T₄ degradation, 3,3'-T2 (130). Serum thyroglobulin concentration tends also to be high and the degree of its elevation reflects the level of TSH induced thyroid gland hyperactivity.

Thyrotropin and Other Thyroid Stimulators

A characteristic, if not pathognomonic, feature of the syndrome is the presence of TSH in serum and preservation of its response to TRH despite elevated TH level (153). In most cases, the basal serum TSH concentration is normal and the circadian rhythm is unaltered (154, 155). TSH values above 10 mU/L occur in subjects that have received treatment aimed at reducing their high level of TH. The TSH response to TRH is either normal or exaggerated.

Thyrotropin has increased biological activity (129, 156) and the concentration of its free a subunit (SU) is not disproportionately high. Except for the rare occurrence of coincidental autoimmune thyroiditis (157), serum is free of thyroid stimulating immunoglobulins as well as antibodies against thyroglobulin and thyroid peroxidases.

Thyroid Gland Activity and Integrity of Hormone Synthesis

The fractional uptake of radioiodide by the thyroid gland is high as is the absolute amount of accumulated iodide. The latter is normally organified since no discharge of trapped iodide has been observed following the administration of perchlorate (27, 149, 158). No abnormal iodide containing compounds have been detected in the circulation (27, 141).

Turnover of Thyroid Hormone

In vivo turnover kinetics of T₄ showed a normal or slightly increased volume of distribution and fractional disappearance rate of the hormone. However, because of the high concentration of T₄ in serum, the extrathyroidal pool and absolute daily production of T₄ are increased, up to twofold the upper limit of normal and that of T₃ is increased by about two- to four-fold (133, 141, 149, 159, 160). The extrathyroidal conversion of T₄ to T₃ is however normal (160).

In-vivo Effects of Thyroid Hormone

The impact of TH on peripheral tissues has been assessed in vivo by a variety of tests. Results have been, by and large normal and, given the high serum levels of TH, suggest a reduced biologic response to the hormone (1, 7). The metabolic status has been evaluated by measurements of the basal metabolic rate (BMR), serum cholesterol, carotene, triglycerides, creatine kinase, alkaline phosphatase, angiotensin-converting enzyme, sex hormone-binding globulin (SHBG), ferritin and osteocalcin, all of which usually have been within the normal range. Urinary excretion of magnesium, hydroxyproline, creatine, creatinine, carnitine, and cyclic adenosine monophosphate (cAMP), all found to be elevated in thyrotoxicosis, have been normal or low, suggesting normal or slightly reduced TH effect. With the exception of increased resting pulse rate in about one half of the patients with RTH, the cardiac function is only minimally altered. Two-dimensional and Doppler echocardiography showed mild hyperthyroid effect on cardiac systolic and diastolic function of the myocardium whereas other parameters, such as ejection and shortening fractions of the left ventricle, systolic diameter, and left ventricle wall thickness, were normal (97). The Achilles tendon reflex relaxation time has also

been normal or slightly prolonged. The PRL response to TRH was not blunted as it is in patients with thyrotoxicosis. In fact, the PRL hyperresponsiveness in some patients with RTH may be due to the functional TH deprivation at the level of the lactotrophs (153).

Other Endocrine Tests

Evaluation of endocrine function by a variety of tests has failed to reveal significant defects other than those related to the thyroid. The following laboratory analyses have been carried out in patients with RTH and were found to be within the normal range: serum levels of cortisol and its diurnal rhythm; testosterone, estrogens, and progesterone; gonadotropins and their response to gonadotropin-releasing hormone; adrenocorticotropic hormone; insulin; prolactin and its response to TRH, L-dopa and glucocorticoids; growth hormone and its response to insulin hypoglycemia, arginine, and pyrogen; as well as the urinary excretion of 17-hydroxycorticoids, 17-ketosteroids, vanillylmandelic acid, adrenaline, and noradrenaline. Radiologic and magnetic resonance examinations of the pituitary gland and sella turcica have shown no anatomic abnormalities.

Bone age

Radiologic evidence of delayed bone maturation has been observed in one-half of patients with RTH diagnosed during infancy or childhood (1). However, the majority achieve normal adult stature. It is unclear whether the presence, in some cases, of stippled epiphyses is also the consequence of reduced TH action.

In vitro Tests of Thyroid Hormone Action

The normal stimulatory effect of T₃ on the degradation rate of low-density lipoproteins was reduced in cultured skin fibroblasts obtained from three affected members of one family (151). Similarly, T₃ and T₄, but not dexamethasone, failed to produce the normal inhibitory effect on the synthesis of glycosaminoglycans in fibroblasts from 4 of 6 patients with RTH (137) (Fig. 3). In contrast, T₃ normally stimulated glucose consumption by cultured fibroblasts from a patient with RTH (136, 161). In vitro demonstration of TH resistance has been most consistent by measurement of the normal inhibitory effect of T₃ on the synthesis of fibronectin and its mRNA in skin fibroblasts maintained in culture. Of 12 patients with RTH who were studied, 11 showed either an attenuated or paradoxical response (138).

RESPONSES TO THE ADMINISTRATION OF THYROID HORMONE

Because reduced responsiveness to TH is central in the pathogenesis of the syndrome, patients have been given TH in order to observe their responses and thereby establish the presence of hyposensitivity to the hormone. Unfortunately, data generated have been discrepant, not only because of differences in the relative degree of resistance to TH among patients but, also because of lack of uniformity in the manner the hormone trials have been carried out. These include differences in hormonal preparations, dosages, duration of treatment and the type of observations and measurements carried out, not to mention the conspicuous lack of adequate control studies.

Administration of TH ultimately suppresses TSH secretion resulting in a decrease and eventually the abolition of the TSH response to TRH. The amount of TH necessary to produce such an effect has been variable, as has been the relative effectiveness of L-T₃ as compared to L-T₄. Such observations have led to earlier speculations that some patients may have an abnormality in conversion of T₄ to T₃ (142, 143). The decreased TSH secretion during the administration of supraphysiologic doses of TH is accompanied by a reduction in the thyroidal radioiodide uptake and, when exogenous T₃ is given, a reduction in the pretreatment level of serum T₄ (132, 133, 143, 149, 151).

Various responses of peripheral tissues to the administration of TH have been quantitated. Most notable are measurements of the BMR, pulse rate, reflex relaxation time, serum cholesterol, lipids, enzymes, osteocalcin and SHBG, and urinary excretion of hydroxyproline, creatine, and carnitine. Either no significant changes were observed, or they were much reduced relative to the amount of TH given (1).

Of great importance are observations on the catabolic effect of exogenous TH. In some subjects with RTH, L-T₄ given in doses of up to 1000 µg/day, and L-T₃ up to 400 µg/day, failed to produce weight loss without a change in calorie intake, nor did they induce a negative nitrogen balance (131, 132, 141). In contrast, administration of these large doses of TH over a prolonged period of time was apparently anabolic as evidenced by a dramatic increase in growth rate and accelerated bone maturation (11, 131).

EFFECTS OF OTHER DRUGS

As expected, administration of the TH analogue, 3,5,3'-triiodo-L-thyroacetic acid (TRIAC) to patients with RTH produced attenuated responses (141, 155, 162).

Administration of glucocorticoids promptly reduced the TSH response to TRH and the serum T₄ concentration (131, 132, 135, 142, 159).

Administration of L-dopa and bromocriptine produced a prompt suppression of TSH secretion, as well as a diminution of the thyroidal radioiodide uptake and serum T₃ level (131, 134, 142). Domperidone, a dopamine antagonist, caused a rise in the serum TSH level when given to patients with RTH (155). These observations indicate that, in this syndrome, the normal inhibitory effect of dopamine on TSH is intact.

The response to antithyroid drugs has shown some variability. Methimazole and propylthiouracil, in doses usually effective in reducing the high serum TH level of hyperthyroidism of autoimmune etiology, had no effect in two patients (141). However, in other cases of RTH, antithyroid drugs induced some decrease in the circulating level of TH, producing a reciprocal change in the TSH concentration (32, 139, 158, 163). Administration of 100 mg of iodine daily had a similar effect in one patient (133), but 4 mg potassium iodide per day produced no changes in another (141).

Observations on the effect of other drugs such as diazepam, and chlorpromazine are limited. With one exception (141), propranolol and atenolol caused a significant reduction in heart rate.

DIFFERENTIAL DIAGNOSIS

Because the clinical presentation of RTH is variable, detection requires a high degree of suspicion. The differential diagnosis includes all possible causes of hyperthyroxinemia. The sequence of diagnostic procedures listed in Table 4 is suggested.

The presence of elevated serum T₄ concentration with nonsuppressed TSH needs to be confirmed by repeated testing. The possibility of an inherited or acquired increase in serum TBG must be excluded by direct measurement and by estimation of the circulating free T₄ level. The presence of a high serum T₃ level must also be documented since reduced conversion of T₄ to T₃ by peripheral tissues may occasionally give rise to the elevation of total and free T₄ but not T₃ levels. This may occur transiently in a variety of nonthyroidal illnesses or during the administration of some drugs (see Chapters 6E amd 5). A familial form of hyperthyroxinemia, presumably due to a defective T₄ monodeiodination, has been also described (164). The inherited abnormality of T₄ binding to an albumin presents with high serum T₄ but normal T₃ concentration (see Chapter 16C). A rare cause of elevated serum T₄ and T₃ level is the endogenous production of antibodies directed against these hormones, which can be excluded by direct testing

Most useful is the measurement of the serum TSH. Under most circumstances, patients with high concentrations of circulating free TH have virtually undetectable serum TSH levels, which characteristically fail to increase in response to TRH. This is true even when the magnitude of TH excess is minimal and therefore subclinical both on physical examination or by other laboratory tests (see Chapters 6A). The combination of elevated serum levels of TH and non suppressed TSH, narrows the differential diagnosis to RTH and autonomous hypersecretion of TSH associated with pituitary tumors. The latter should be suspected when other members of the family, and particularly the parents of the patient, fail to exhibit thyroid test abnormalities. Rarely, endogenous antibodies to TSH or some of the test components, can give rise to false increase in serum TSH values.

In addition to symptoms and signs of thyrotoxicosis, some patients with TSH-producing (thyrotroph) pituitary adenomas may present with acromegaly due to the concomitant hypersecretion of growth hormone by the tumor. Galactorrhea and amenorrhea in association with hyperprolactinemia have also been reported (see Chapter 13). The tumors may be demonstrated by computerized tomography or by magnetic resonance imaging of the pituitary. A typical finding in patients with TSH-producing pituitary adenomas is a disproportionate abundance in serum of free a-SU of TSH relative to whole TSH (165). Moreover, with rare exceptions (166), serum TSH fails to increase above the basal level in response to TRH or to decrease during the administration of TH. Ectopic production of TSH has not been unequivocally demonstrated (see Chapter 13). It is uncertain that endogenous TRH hypersecretion could maintain high TSH levels and induce hyperthyroidism.

Since the etiology of PRTH is not distinct from that of RTH, failure to demonstrate an anatomic defect in the pituitary gland by imaging does not exclude the presence of a small adenoma or thyrotroph hyperplasia. Furthermore, absence of conspicuous signs and symptoms of hypermetabolism are not sufficient to rule out thyrotoxicosis. The occurrence of an apparent selective hyposensitivity to T₄ but not T₃, has been reported in one family (167). Initially attributed to a putative defect in type II 5'-deiodinase, it was recently shown that affected family members had the TRß R320L mutation (D. Gross, P.R. Larsen and W.W. Chin, personal communication).

Proving the existence of peripheral tissue resistance to TH is not simple. Lack of clinical symptoms and signs of hypermetabolism are not sufficient to establish the diagnosis of RTH and no single test objectively proves the existence of eumetabolism. Because resistance to the hormone is variable in different tissues, no single test measuring a particular response to TH is diagnostic. Furthermore, results of most tests that measure the effect of TH on peripheral tissues show considerable overlap among thyrotoxic, euthyroid and hypothyroid subjects. The value of these tests is enhanced if measurements are obtained before and following the administration of supraphysiologic doses of TH.

While the demonstration of TRß gene mutation is sufficient to establish the diagnosis of RTH, in vivo demonstration of tissue resistance to TH is required when RTH is not associated with a TR gene mutation (2). A standardized diagnostic protocol, using short-term administration of incremental doses of L-T₃ and outlined in Fig. 5, is recommended. It is designed to assess several parameters of central and peripheral tissue effects of TH in the basal state and in comparison to those elicited following the administration of L-T₃. The three doses given in sequence are a replacement dose of 50 µg/day and two supraphysiologic doses of 100 and 200 µg/day. The hormone is administered in a split dose every 12 hours and each incremental dose is given for the period of 3 days. Doses are adjusted in children and in adults of unusual size to achieve the same level of serum T₃. L-T₃, rather than L-T₄, is used because of its direct effect on tissues, bypassing potential defects of T₄ transport and metabolism, which may also produce attenuated responses. In addition, the more rapid onset and shorter duration of T₃ action reduces the period required to complete the evaluation and shortens the duration of symptoms that may arise in individuals with normal responses to the hormone. Responses to each incremental dose of L-T₃ are expressed as increments and decrements or as a percent of the value measured at baseline. The results of such a study are shown in Figs.-6 and 7.

Figure 5. Schematic representation of a protocol for the assessment of the sensitivity to TH using incremental doses of L-T3. For details see text. [Adapted from Refetoff et al (1)]

Figure 6. Thyrotroph responses to TRH stimulation at baseline and after the administration of graded doses of L-T3. The hormone was given in three incremental doses, each for 3 days as depicted in Fig. 81-5. Results are shown for patients with RTH in the presence (left) or absence (right) of a TRß gene mutation, together with the unaffected mother of the patient with nonTR-RTH (center).

Figure 7. Responses of peripheral tissues to the administration of L-T3 in the presence or absence of mutations in the TRß gene. The hormone was given as described in Figure 6. Note the stimulation of ferritin and sex hormone binding globulin (SHBG) and the suppression of cholesterol and creatine kinase (CK) in the normal subject. Responses in affected subjects, with or without a TRß gene mutation, were blunted or paradoxical.

Failure to differentiate RTH from ordinary thyrotoxicosis has resulted in the inappropriate treatment of nearly one-third of the patients. The diagnosis requires awareness of the pessible presence of RTH, usually suspected when high levels of circulating TH are not accompanied by a suppressed TSH.

TREATMENT

Although no specific treatment is available to fully and specifically correct the defect, the ability to identify specific mutations in the TRs provides a means for prenatal diagnosis and appropriate family counseling. This is particularly important in families whose affected members show evidence of growth or mental retardation. Fortunately, in most cases of RTH, the partial tissue resistance to TH appears to be adequately compensated for by an increase in the endogenous supply of TH. Thus, treatment need not be given to such patients. This is not the case in patients with limited thyroidal reserve due to prior ablative therapy. In these patients, the serum TSH level can be used as a guideline for hormone dosage.

Not infrequently, some peripheral tissues in patients with RTH appear to be relatively more resistant than the pituitary. Thus, compensation for the defect at the level of peripheral tissues is incomplete. In such instances, judicious administration of supraphysiologic doses of the hormone is indicated. Since the dose varies greatly among cases, it should be individually determined by assessing tissue responses. In childhood, particular attention must be paid to growth, bone maturation and mental development. It is suggested that TH be given in incremental doses and that the BMR, nitrogen balance and serum SHBG be monitored at each dose, and bone age and growth on a longer term. Development of a catabolic state is an indication of overtreatment.

The exact criteria for treatment of RTH in infancy have not been established. This is often an issue when the diagnosis is made at birth or in early infancy. In infants with elevated serum TSH levels, subclinical hypothyroidism may be more harmful than treatment with TH. Indications for treatment may include a TSH level above the upper limit of normal, retarded bone development and failure to thrive. The outcome of affected older members of the family who did not receive treatment may serve as a guideline. Longer follow-up and psychological testing of infants who have been given treatment will determine the efficacy of early intervention.

A recent survey has shown an increased miscarriage rate and low birth weight of normal infants born to mothers with RTH (168). It unclear at this time whether intervention during early gestation is appropriate. However, mothers with RTH should be followed carefully during pregnancy and not allowed to have low serum TSH values. The wisdom of utero treatment is questionable (169, 170).

Patients with more severe thyrotroph resistance and symptoms of thyrotoxicosis may require therapy. Usually symptomatic treatment with an adrenergic ß blocking agent, preferably atenolol, would suffice. Treatment with antithyroid drugs or thyroid gland ablation increase TSH secretion and may result in thyrotroph hyperplasia. Development of true pituitary tumors, even after long periods of thyrotroph overactivity, is extremely rare (171).

It has been recently shown that treatment with supraphysiological doses of L-T₃, given as a single dose every other day, is successful in reducing goiter size with remarkable cosmetic benefits and without causing side effects (172). This appears to be the treatment of choice considering that postoperative recurrence of goiter is the rule. The L-T₃ dose must be adjusted in increments until TSH and TG are suppressed and reduction of goiter size is observed.

Patients with presumed isolated peripheral tissue resistance to TH present a most difficult therapeutic dilemma. The problem is, in reality, diagnostic rather than therapeutic. Many, if not most, patients falling into this category are habitual TH users. Gradual reduction of the TH dose and psychotherapy is recommended.

Acknowledgments

Supported in part by US Public Health Grants DK15070 and RR00055. Reproduced from Refetoff, S: Resistance to thyroid hormone. In Werner and Ingbar's The Thyroid: A Fundamental and Clinical Text. Braverman, L.E., and Utiger, R.E, (eds.), LIppincott, Williams and Wilkins, Philadelphia, PA., 2005.

Continue to Section 16C- Abnormal Thyroid Hormone Transport