Biochemical and molecular issues
  The first patients described with RTH in 1967 had, as mentioned above, a very particular phenotype. The inheritance in these cases was autosomal recessive. After a tight linkage was reported between the TRb locus (Usala, 1988) and RTH about 600 other cases have been described. The first patients had a homozygous deletion of the TRb1 allele, which as it turned out was the exception to the rule since all other cases had point mutations or small deletions in the TRb1 gene. As a result of these changes in the TRb1 gene, amino acids change or the synthesis of the receptor protein is stopped prematurely. Interestingly all mutations found to date cluster in three areas of the receptor with some amino acid positions very prone to mutation (Fig. 5)(Collingwood, 1998 and references therein). The mutations do not occur in areas of the receptor that confer corepressor binding and dimerisation, indicating that these regions are important for receptor function. The inheritance of the point mutations is autosomal dominant and patients are heterozygous for the mutation. In 15% of cases RTH is sporadic and a mutation has arisen "de novo".

Figure 5 Position of the mutations in the ligand binding domain of the TRb



The mutations that have been found in the TRb cluster in three areas of the ligand binding domain straddling the sites for corepressor binding and dimerisation. The amino acid positions between which the mutations are found are indicated above the boxes.


How do the mutant receptors give rise to the resistance phenomenon? From the first case described it is clear that losing the complete gene is only a problem when both alleles are lost. The dominant nature of the inheritance of all other mutants described indicates that the mutated receptors do interfere with the action of their normal counterparts. From in vitro experiments it has become clear that this is indeed the case and that the mutated receptors act in a dominant negative manner, i.e. they decrease the effect of the hormone even though the normal receptor is present (Sakurai, 1990; Collingwood, 1994). For the mutant receptor to act in a dominant negative manner, DNA binding and heterodimer formation are essential (Nagaya, 1993). When mutations which abolish either of the two are tested they will not work as dominant negatives. On the basis of this the interference with the normal way in which the TR works can be envisaged to take place at several different levels (Fig. 6).

As was shown in figure 5 all mutations cluster in three particular areas. When these areas are plotted onto the 3D structure of the TRb it becomes apparent that they are surrounding the binding site for thyroid hormone (Fig. 7). Therefore the first possibility is loss of hormone binding. In this case the hormone cannot bind to the receptor which will then not be able to release the corepressor (see Fig. 6) and the result of this is that the gene is not activated. Other receptor mutants have been found that do bind T3 but which release the co-repressor slower than normal (Safer, 1998; Clifton-Bligh, 1998). This will also lead to a decrease in gene activation. Another possibility is that the mutant receptor molecules form heterodimers and go back to the TRE but then fail to attract the coactivator in which case the gene remains silent.

Fig. 6: Effects of the mutations on gene activation by the TR



The mutated TR can still bind to the TRE as a homodimer with a normal TR. This complex can still attract corepressors and because T3 cannot bind it will form a stable inactive complex on the TRE of the gene. When there is some residual T3 binding the homodimer will be released but it is possible that the mutated TR binds the RXR or other cofactor preferentially. In the case where the mutated receptor has been able to heterodimerise and then again binds to the TRE it will block access for the normal receptor heterodimer with coactivator again resulting in silencing of the gene.


Figure 7 Three dimensional structure of the TRb ligand binding domain



The 3D structure of the TRb ligand binding domain is depicted as a ribbon following the peptide backbone (derived from NCBI-MMDB database, structure number 1BSX). The top of the structure is the side of the DNA binding domain. The red ribbons indicate the three areas where mutations preferentially occur. It can be clearly seen that these areas concentrate themselves around the T3 binding site.


Disruption of the TRb gene in mouse models shows a phenotype reminescent of the first RTH patients identified to harbour a homozygous deletion (Forrest, 1996). These animals also have serious hearing defects. When the deletion is heterozygous, normal thyroid test results are found.

The expression of mutant TRb in mice results in an animal model of RTH with lower body weight, hyperactivity and learning problems, similar to the problems found in humans (Zhu, 1999; Wong, 1997; McDonald, 1998; Abel, 1999).

Another interesting point is that of the variation in peripheral symptoms encountered, which brings us back to the GRTH/PRTH distinction. The same mutation has led to the diagnosis PRTH in one family whereas it led to GRTH in another. There are even reports that PRTH and GRTH can exist within one family harboring the same TRb mutation. It has therefore been argued that the distinction between the two is artifactual and based on the poor definition of the symptoms. PRTH and GRTH can therefore be viewed as two sides of a spectrum of a single gene disease. A possible reason for the variability in the symptoms could be that not all individuals express the same levels of TR (both mutant and normal) in their tissues (Mixson, 1993). When the ratio between mutant and normal TR changes so will the final effect of the mutant receptor (see Fig. 6). Furthermore not all mutations have the same effect on T3 binding (Hayashi, 1995). Recently it was found that there is a relation between the T3 binding impairment and the outcome of thyroid function tests (Ercan-Fang, 2000). Another factor could be the different tissue distribution of the TR (see Fig. 2). For instance the heart is a predominantly TRa tissue. Since the TRa is normal in RTH patients but their FT3 levels are high, it can be expected that they will react to the extra amount of T3 in a hyperthyroid manner as far as the heart is concerned. Furthermore not every individual will express the same amount of TR or corepressors/coactivators in a particular tissue leading to differences between individuals. An interesting observation in this context is that some mutations are more deleterious when present in the TRb2 then in the TRb1. Since the TRb2 expression is restricted to the pituitary it may be expected to give rise to a "PRTH" phenotype.

Fig. 2: Schematic representation of the four thyroid hormone receptor (TR) isoforms



The receptor isoforms are shown schematically with the tissues where their expression is most prominent. The black box represents the DNA binding domain and the red and dark green boxes the T3-binding domain. TRb1 and b2 differ in their N-terminal domain whereas TRa1 and a2 differ in their C-terminal domain. In both cases an mRNA is transcribed from one gene which upon alternative splicing yields the a or b isoform.


A number of cases have recently been reported where no mutation has been found in the TRb gene even though the biochemical evidence was there (Pohlenz, 1999). It has been argued that the origin of the resistance in these cases is a faulty cofactor (Weiss 1993; Weiss 1996). This is supported by the recent finding that retinoid-X receptor (RXR)g knock-out mice display PRTH like symptoms (Brown, 2000), and that steroid receptor coactivator (SRC)-1 knock-out mice manifest RTH symptoms (Weiss, 1999b).

With the rapid increase of our knowledge of receptor structure and the way agonists and antagonists interact with the receptors it can be envisaged that it would be possible to design receptor agonists which will correct the receptor defect. It has been shown for instance that one of the RTH mutations causes a shift of 0.3 Å of one of the helices in the ligand-binding domain which results in decreased T3 binding. If an agonist were found which can "live" with this small shift and thus activate the receptor, patients harboring this particular mutation could be treated.