Introduction
Thyroid hormone is important for brain development which is dramatically manifested by the severe neurological deficits caused by untreated fetal and neonatal hypothyroidism (1). In the brain, neurons are the major target cells for T3 during brain development, and this T3 is largely derived from outer ring deiodination of the prohormone T4 by the type 2 deiodinase (D2) located in neighbouring astrocytes (Fig. 1) (2-4). As in all T3-sensitive cells, the effect of T3 on neurons is largely initiated by its binding to nuclear receptors (TRs) associated with T3 response elements in the promoter region of T3-responsive genes, resulting in an altered transcription of these genes (Fig. 1) (5). In neurons, important T3-responsive genes are involved in the control of the migration, differentiation and arborization of these cells (1, 2). For the termination of T3 action, many neurons also express the type 3 deiodinase (D3) which catalyzes the inactivation of T3 via inner ring deiodination (Fig. 1) (2, 3). Normal brain development requires the coordinated time-dependent and tissue-specific expression of the deiodinases in the different cell types as demonstrated recently also in the developing human brain (6).

Fig. 1 Local regulation of thyroid hormone bioactivity in the brain in functional units of astrocytes, which express D2, and the major target cells for thyroid action during brain development, the neurons, which express MCT8 and D3. MCT8 is essential for neuronal T3 uptake and thus for T3 action and metabolism in these cells.

Not only the nuclear T3 receptors but also the deiodinase active centers are located intracellularly. Therefore, both action and metabolism of thyroid hormone require the transport of iodothyronines across the plasma membrane (7, 8). This transport does not occur by simple diffusion but is facilitated by transporters. In recent years several iodothyronine transporters have been identified, including organic anion transporters (Na/taurocholate cotransporting polypeptide [NTCP] and organic anion-transporting polypeptides [OATPs]), the L-type amino acid transporter (LAT), and fatty acid translocase (FAT) (7, 8). However, most of these transporters show low activity and specificity towards iodothyronines. A notable exception is OATP1C1 that shows strong preference for T4 as the ligand and is specifically expressed in brain (9-11). This transporter is particularly abundant in brain capillaries, suggesting that it plays an important role in the transport of T4 across the blood-brain barrier (4, 9-11).

Mutations in the MCT8 thyroid hormone transporter
Recently, we have identified rat and subsequently human monocarboxylate transporter 8 (MCT8) as an active and specific iodothyronine transporter (12, 13). The human MCT8 gene was already cloned in 1994 but its role has remained elusive until recently (14). It is located on the X chromosome (Xq13.2), consists of 6 exons and, depending on which of the two translation start sites (TLSs) is used, codes for a protein of 613 or 539 amino acids. The protein contains 12 putative transmembrane domains (TMDs), and both N- and C-terminal domains are located inside the cell. In particular the N-terminal domain is large (170 or 96 amino acids) and contains a PEST domain (rich in P [Pro], E [Glu], S [Ser] and T [Thr] residues) that is probably important for MCT8 turnover (14).
We have recently examined 9 unrelated young boys suffering from severe psychomotor retardation (Fig. 2), who also show strongly elevated serum T3 levels. Serum T4 and FT4 levels are low-normal to clearly decreased, serum rT3 is low, and serum TSH is normal or increased. In all patients this novel syndrome was found to be associated with different mutations in the MCT8 gene, i.e. 4 deletions of ~24 kb, 2.4 kb, 14 bp, and 3 bp (delPhe230 in TMD2); 3 missense mutations (Ala224Val in TMD2; Leu471Pro in TMD9; and Arg271His in the 2nd extracellular loop); a nonsense mutation (Arg245stop); and a splice site mutation resulting in the loss of 94 amino acids, including TMD4-6 (15, 16). Mutations in MCT8 have recently also been reported by Dumitrescu et al. in 2 boys with a very similar phenotype, and by Schwartz et al. and Maranduba et al. in 8 different families with the Allan-Herndon-Dudley syndrome (AHDS), named after the authors of the paper describing one of these families in 1944 (17-20). In most families with AHDS, the affected males show again a very similar phenotype as in our patients (Fig 2), whereas in some families the phenotype is somewhat milder with some of the patients reaching old age and being capable of walking and speaking albeit with great difficulties. Only recently it has become clear that AHDS is also associated with elevated serum T3 levels (18, 19). None of the patients with MCT8 mutations has been recognized at neonatal screening for congenital hypothyroidism.

Fig. 2 Most common clinical characteristics in patients with MCT8 mutations


Five of the 9 mutations identified in our patients are obviously deleterious for functional MCT8 expression. This is less obvious for the 3 missense mutations and the 3-bp deletion, resulting in the substitution or deletion of single amino acids. These mutations were therefore introduced in the cDNA coding for human MCT8 and their effects were studied by measurement of T3 uptake in cells transfected with wild-type or mutant MCT8. In addition, T3 metabolism was analyzed in cells transfected with wild-type or mutant MCT8 in combination with cDNA coding for human D3. In both systems, the single amino acid substitutions or deletion were found to result in an almost complete inactivation of MCT8 (16). Another interesting mutation was identified in one of the above-mentioned families with AHDS, namely a single nucleotide deletion in the 2nd last codon (18). This results in the by-passing of the natural stop codon until an alternative stop codon is encountered 196 nucleotides further downstream, resulting in the elongation of the protein with 64 amino acids which may include an additional TMD. Also this mutation was introduced in MCT8 cDNA and found to result in a major decrease in T3 uptake and metabolism in transfected cells, although the mutant showed significant residual activity (18). The genotype-phenotype relationship of the MCT8 mutations identified in the various families remains to be fully explored.
The neurological defect caused by mutations in MCT8 is at least as dramatic as that associated with iodine deficiency or untreated congenital hypothyroidism (1). It is hypothesized that inactivation of MCT8 blocks the neuronal entry of T3 and thus its access to its intracellular (nuclear) receptor and degrading enzyme D3 (Fig 1). As a consequence, T3 cannot exert its crucial action in the developing brain, resulting in an impaired neurological development. The reduced breakdown of T3 by D3 supposedly leads to an initial rise in serum T3 that subsequently stimulates the expression of D1 in liver and kidney with a further increase in T3 production. The secondary increase in hepatic D1 expression may also explain the low serum T4 and rT3 levels in patients with MCT8 mutations. This explanation is based on the assumption that the elevated serum T3 results in an increased T3 effect in the liver and kidneys despite the fact that MCT8 is also expressed in these tissues. However, liver and kidneys also express other thyroid hormone transporters that may be responsible for a relatively unhindered T3 uptake even if MCT8 is inactive. This is supported by the findings that serum SHBG levels are strongly elevated in patients with hemizygous MCT8 mutations, suggesting that hepatic SHBG production is increased because of an increased intrahepatic T3 concentration (16).
The effects of MCT8 mutations on the thyroid state of different tissues thus depends on the extent to which T3 uptake in these tissues requires a functional MCT8 transporter. MCT8 appears essential for T3 uptake in central neurons, and its inactivation will thus result in a hypothyroid state in different brain regions despite the high extracellular T3 levels. If MCT8 is only one out of many transporters facilitating T3 uptake in liver and kidneys, these tissues may well be thyrotoxic because of the high circulating T3 concentrations. In other tissues where MCT8 is an important but not the only T3 transporter, inactivation of MCT8 may result in a partial block of T3 uptake which in combination with the high serum T3 concentration could lead to relatively normal intracellular T3 concentrations. The heart which is know to express MCT8 may be an example of such a tissue which function appears to be normal in patients with MCT8 mutations.

Conclusion
Much remains to be learned about exactly how hemizygous mutations in MCT8 cause the severe X-linked psychomotor retardation in affected males. Obviously, mutations in thyroid hormone transporters represent a novel mechanism of thyroid hormone resistance with dramatic consequences.