*Corresponding author. E-mail: paolo.beckpeccoz@unimi.it

Introduction
TSH resistance (OMIM #275200) is a condition of thyroid refractoriness to TSH stimulation. The first description dates back to 1968 (1) when a child with congenital hypothyroidism, in situ thyroid gland of normal volume, and no increase of iodine uptake after exogenous TSH administration was reported. Since then, a number of other cases have been described. Beginning from the early reports, alterations of TSH receptor (TSHR), Gsα or other downstream signaling elements were considered as possible causes of this particular disorder. However, it was only after the cloning of TSHR gene (2) that loss-of-function TSHR mutations were recognized as the molecular defect in most familial cases of TSH resistance (3-6). This review will focus on the clinical presentation, differential diagnosis and molecular genetics of this rare disorder.

Clinical and biochemical features of TSH resistance
High levels of serum TSH, normal/reduced levels of serum thyroid hormone and normal/hypoplastic thyroid gland characterize TSH resistance. Depending on the degree of TSH insensitivity, the presentation may be extremely variable, ranging from severe congenital hypothyroidism to only mild elevations of TSH in the absence of signs and symptoms of hypothyroidism.

Table I. Clinical and biochemical phenotypes associated with inactivating mutations of TSHR depend on the degree of impairment in receptor function.



As TSH is the major physiological stimulus of both thyrocyte function and proliferation, a profound reduction of thyroid sensitivity to TSH, as observed in patients with complete TSH resistance, leads to severe hypothyroidism with a hypoplastic thyroid gland. Because of reduced thyroid hormone feedback, TSH is markedly elevated. Such cases are typically diagnosed at the neonatal screening for congenital hypothyroidism. In the first description by Abramowicz et al. (4), thyroid hypoplasia was so profound and radioiodine uptake so impaired that thyroid agenesia was diagnosed at scintigraphy. However detectable serum Tg levels disclosed the presence of thyroid tissue.
When thyroid refractoriness to TSH is incomplete, a condition known as partial TSH resistance, TSH elevation can somewhat compensate for the reduced sensitivity of the thyroid and milder forms of hypothyroidism are seen. Patients typically have a thyroid gland of normal/reduced size, high TSH levels, but concentrations of free thyroid hormones in the normal range (3,5,6).
More recently, we described three families with only mild elevations of TSH levels and a normal thyroid gland, in which the defect was caused by heterozygous inactivating TSHR mutations (see below) (7). Similar alterations can be found in some heterozygous relatives of patients with homozygous TSHR mutations (8). Although such cases can be positive at neonatal TSH screening (7), they are generally diagnosed later on in life, when they can be easily confounded with the more prevalent condition of subclinical hypothyroidism due to autoimmune thyroid disease (AIT).

Differential Diagnosis
The occurrence of elevated TSH serum levels in the presence of low/normal free T4 concentration and thyroid volume is rather frequent in the general population. However, only a minority of these patients is affected with TSH resistance. So, the diagnostic workup should exclude other potential causes such as AIT, defects in TSH molecule (particularly those due to TSH beta gene mutation) and TSH biological activity or other forms of congenital primary hypothyroidism, including abnormalities in thyroid transcription factors (6).
Differential diagnosis with AIT, by far the most frequent cause of TSH elevations in the adult population, is based on clinical history, biochemical evaluation of anti-thyroid antibodies (anti-TPO and anti-thyroglobulin) and thyroid ultrasound (7). The positivity of anti-thyroid antibodies and/or the presence of the typical heterogeneous hypoechoic pattern at thyroid ultrasound are strongly suggestive of AIT. Other elements in favor of AIT are diagnosis reached in adult/advanced age as well as progressive evolution from subclinical toward overt disease. The only exception is the very rare association of TSH resistance with AIT (9).
Some patients with central hypothyroidism may present with elevated serum TSH levels. These alterations are generally found in patients with hypothalamic (tertiary) hypothyroidism (10,11), but can also occur in the presence of TSH? gene mutations (12). In the former situation, despite high serum concentrations of the immunoreactive hormone, TSH biological activity is reduced, which explains the condition of hypothyroidism. In these cases, the presence of hypothalamic-pituitary disorders supports the diagnosis of central hypothyroidism and prompts to test the biological activity of circulating TSH molecules in vitro.
Complete TSH resistance due to loss-of-function TSHR mutations can lead to congenital hypothyroidism with thyroid gland in situ. Other known causes are defects in thyroid transcription factors, such as NKX2.1 (also named TITF1) or PAX8 (13). Due to the expression of these transcription factors in tissues other then the thyroid, complex phenotypes have been reported in these cases. Patients with NKX2.1 mutations may have neurological (choreoathetosis) and pulmonary alterations, while PAX8 mutations can be suspected in the presence of kidney abnormalities. NKX2.1 and PAX8 mutations can be excluded by genetic analysis.
TSH resistance can also occur in the context of the multiple hormone resistance syndrome (Albright’s Hereditary Osteodystrophy) caused by inactivating mutations of the gene encoding Gsα protein. The presence of high PTH levels associated with hypocalcemia and hyperphosphatemia, as well as typical clinical features are suggestive of this disorder.

Molecular genetics
The congenital/childhood occurrence and the frequent family setting of TSH resistance suggest the genetic origin of the disorder. Therefore, after its cloning in 1989 (2), TSHR became the more obvious candidate gene. Shortly after, inactivating TSHR mutations were identified in a family with recessive transmission of elevated TSH levels and normal thyroid hormone secretion (3). Since then, 23 different TSHR loss-of-function mutations have been documented to be responsible for TSH resistance (5,6). An updated database of TSHR mutations can be found on the web (14). However, some cases of TSH resistance not associated with TSHR or GNAS (Gsα) mutations have been reported, suggesting the probable involvement of other not identified genes (7,15-17).
In the earlier studies, in which only probands with large TSH elevations were screened for mutations, the disease was linked to homozygous or compound heterozygous mutations and was described to follow a recessive pattern of inheritance. More recently, we described patients with a mild form of partial TSH resistance due to heterozygous TSHR mutations. In these cases, the defect had a dominant pattern of inheritance (7). Such genetic heterogeneity is well reflected by diversity of clinical presentation (Table I).
Natural mutations leading to resistance to TSH action are distributed all along the receptor backbone and affect either the extracellular or transmembrane domains. Almost all types of alterations (missense or nonsense mutations, deletions, insertions or alterations in intron-exon boundaries) have been reported. However, missense P162A and C41S mutations appear to be relatively more frequent. All of these mutations are associated with a defective cAMP response to TSH stimulation. This was shown by in vitro functional assays using different end points, such as determination of cAMP concentrations or using reporter genes (e.g., luciferase) under cAMP regulation.
The molecular mechanisms responsible for the loss of receptor function are probably multiple (5,6). Lack of receptor expression on the plasma membrane, ligand binding or coupling with G-proteins may be expected in case of mutations causing the synthesis of a truncated receptor. Similarly, the skipping of exon 6 as reported in two patients, which causes the deletion of one leucine-rich repeat in the aminoterminal hormone binding domain, is expected to impair TSH receptor binding. Initially, all missense mutations located in the extracellular domain were thought to alter ligand binding. However, several exceptions to this rule have now been reported (7,18). For instance, mutations C41S and I167N lead to a profound alteration of TSHR native conformation with hampers receptor targeting to the cell membrane (7,18). Another interesting exception is represented by mutation R310C which is instead associated with a unexpected increase in ligand-independent activity, possibly contributing to the euthyroidism observed in the affected patients. In contrast, mutations located in the transmembrane domain may result either in defective transmission of the stimulatory signal or in deranged routing of mutant TSHR to the cell plasma membrane. The mechanism of TSH resistance in patients with heterozygous TSHR mutations is less obvious. We recently demonstrated the presence of an in vitro dominant negative effect exerted by some of these mutants (C41S, L467P, C600R) at the level of wild-type receptor maturation and routing to the cell membrane. In fact, in cells co-transfected with wild-type and mutant TSHRs, we observed a reduction of both basal and TSH-stimulated cAMP production as compared to that of cell transfected with wild-type receptor alone.



Figure. Dominant negative effect of TSH receptor L467P mutant. Plasmids encoding wild-type TSHR or L467P mutant were either single- or co-transfected in Cos-7 cells. A mutant to wild-type DNA ratio of 3:1 was used to enhance the effect of L467P. Cells were then stimulated with increasing concentrations of bovine TSH (bTSH). No cAMP acumulation was seen in cells transfected with L467P mutant. A partial impairment of cAMP production was observed in cells co-expressing wild-type and mutant receptor as compared to cells expressing wild-type receptor. Similar results were obtained with C41S and C600R mutants.

In addition, the co-expression of mutant TSHRs caused intracellular entrapment, mainly in the endoplasmic reticulum, of wild-type receptor. Finally, we documented, by fluorescence resonance energy transfer and co-immunoprecipitation, the existence of a physical interaction between wild-type and mutant receptors. Based on these data, the occurrence of TSH resistance in these cases appears to be due to a reduction of the amount of wild-type TSHR present at the cell membrane caused by oligomerization with intracellularly retained TSHR mutants (8).

Concluding Remarks
Several issues about TSH resistance remain to be addressed. First, some well-documented cases are not associated with mutations in TSHR coding regions. However, introns and 5’ or 3’ flanking regions of TSHR gene have not been thoroughly analyzed so far. In addition, other genes could be involved, as recently shown by a linkage analysis study (17).
Another debated point is whether to treat or not patients with apparently well-compensated (i.e. partial) TSH resistance. Indeed, clinical evidence suggests that treatment may not be necessary. In fact, several of these patients were diagnosed in childhood or adult age and had quite a normal somatic and neurological development despite untreated since neonatal age. Moreover, one patient reported by our group was diagnosed at neonatal TSH screening, but was not treated with L-thyroxine because of thyroid hormone levels close to the upper limit of the normal range. He is also doing quite well without L-thyroxine supplementation. However, no general rule can be derived from these limited observations and a decision about starting or not therapy should be made in every single patient. In case no therapy is established, we advise a close clinical and biochemical follow-up, including pituitary MRI. In fact, conditions in which thyroid hormone needs are increased, such as puberty or pregnancy, as well as any insult to the thyroid, e.g. AIT, may easily turn euthyroid hyperthyrotropinemia into mild or even overt hypothyroidism.