Background Resistance to thyroid hormone is characterised by a lack of response of peripheral tissues to the active form of thyroid hormone (triiodothyronine, T3). In about 85% of cases, a mutation in THRB, the gene coding for thyroid receptor β (TRβ), is the cause of this disorder. Recently, individual reports described the first patients with thyroid hormone receptor α gene (THRA) defects.
Methods We used longitudinal clinical assessments over a period of 18 years at one hospital setting combined with biochemical and molecular studies to characterise a novel thyroid hormone resistance syndrome in a cohort of six patients from five families.
Findings Using whole exome sequencing and subsequent Sanger sequencing, we identified truncating and missense mutations in the THRA gene in five of six individuals and describe a distinct and consistent phenotype of mild hypothyroidism (growth retardation, relatively high birth length and weight, mild-to-moderate mental retardation, mild skeletal dysplasia and constipation), specific facial features (round, somewhat coarse and flat face) and macrocephaly. Laboratory investigations revealed anaemia and slightly elevated cholesterol, while the thyroid profile showed low free thyroxine (fT4) levels coupled with high free T3 (fT3), leading to an altered T4 : T3 ratio, along with normal thyroid-stimulating hormone levels. We observed a genotype–phenotype correlation, with milder outcomes for missense mutations and more severe phenotypical effects for truncating mutations.
Interpretation THRA mutations may be more common than expected. In patients with clinical symptoms of mild hypothyreosis without confirmation in endocrine studies, a molecular study of THRA defects is strongly recommended.
- thyroid hormone resistance syndrome
- thyroid receptor
- thyroid hormone receptor alpha gene
Statistics from Altmetric.com
Thyroid hormones (THs, comprising triiodothyronine—T3 and thyroxine—T4) have diverse actions, which include the regulation of skeletal growth, maturation of the central nervous system, cardiac and gastrointestinal function and energy homeostasis.1 THs exert their effects through α (TRα) and β (TRβ) receptors,1 which belong to the nuclear receptors superfamily including receptors for glucocorticoids, mineralocorticoids, oestrogens, progesterone and vitamin D. TRα and TRβ are encoded by two genes (THRA and THRB), each of which undergoes alternate splicing to generate receptor subtypes (TRα1, TRα2, TRβ1, TRβ2 and TRβ3) with distinct tissue distribution.1 TRα1 and TRα2 are located mainly in bones, digestive tract, cardiac and skeletal muscle and central nervous system.2 Three subtypes of TRβ are mainly expressed in the liver, kidney, hypothalamus and pituitary, where they regulate the release of thyroliberin or thyrotropin, subject to the current concentration of THs in blood serum and in the thyroid gland.3 ,4
Resistance to TH (RTH) is characterised by a lack of response of peripheral tissues to the active form of TH (T3). Two forms of inheritable RTH have been described, leading to clinically distinct phenotypes: RTHβ and RTHα. In >85% of the cases, RTHβ is caused by a mutation in THRB, the gene coding for TRβ. This disorder affects 1 per 50 000 live births and can appear both sporadically and in an inherited autosomal recessive or dominant manner.5 Affected individuals clinically present with goitre, short stature, decreased body mass, enlarged heart and mild psychomotor retardation. Furthermore, the disruption of TRβ function in the hypothalamus, pituitary and thyroid gland results in a loss of the negative feedback in this axis. This is reflected in the thyroid function tests in affected individuals, which show increased levels of T4 and T3 coupled to inappropriately normal levels of thyroid-stimulating hormone (TSH).6 Approximately 10% of patients with RTH do not harbour a mutation in the THRB gene, which has led to the suspicion that this disorder may be caused by alterations in other TH receptor-related genes. These candidate genes include coregulators, coactivators and corepressors regulating TRβ function. However, efforts to screen individuals with RTH for mutations in these genes have failed so far.7
Particularly, the absence of mutations in THRA in patients with RTH has been met with puzzlement, leading to speculation that mutations in this gene could be extremely rare or undiagnosed. Recently, several case reports described the first patients with heterozygous truncating mutations in THRA presenting with clinical features suggestive of TH resistance, however different from RTHβ.1 ,5 ,8 ,9
Here we present the clinical and laboratory features of five paediatric patients (one male and four female patients, aged 4–18 years) as well as one adult patient (aged 39 years) with RTHα, a novel TH resistance syndrome due to truncating and missense THRA mutations. To our knowledge, this is the first study with a larger number of patients and a description of a distinct and consistent phenotype of mild hypothyroidism, associated with macrocephaly, specific facial features, skeletal abnormalities and altered T4 : T3 ratio, along with normal TSH levels.
Materials and methods
All patients (n=6, age range 2–39 years, mean age 16 years, median age 11.5 years) were admitted and followed at The Children's Memorial Health Institute (Warsaw, Poland). All five paediatric patients were selected based on strikingly similar dysmorphic features along with developmental and metabolic alterations, suggestive for a common underlying genetic defect.
Molecular genetic studies
Blood and/or saliva samples from the patients and their parents were obtained, from which DNA was extracted and purified. Exome libraries were prepared using a SureSelect human exome v.2 kit (Agilent, Santa Clara, USA) and sequencing was performed on patient DNA on a SOLiD4 platform (Life Technologies, Foster City, USA). Reads were mapped to the hg19 reference genome and variant calling was done with Lifescope software v.2.1 (Life Technologies, Foster City, USA). Variant annotation, filtering and prioritisation were performed as described previously.10 Sanger sequencing was performed to validate candidate variants and prove de novo occurrence or show segregation.
The protocol was approved by the human subjects institutional review board at the Children's Memorial Health Institute. The study was designed and conducted in compliance with the principles of the International Conference on Harmonisation of Technical Requirements for registration of Pharmaceuticals for Human use Guidelines for Good Clinical Practice.
Written informed consents for the genetic investigations and picture publication were provided by the parents or legal guardians.
The clinical syndrome was similar in all patients and occurred due to de novo mutations or was transmitted as a dominant trait in one family (see online supplementary figure S1). Online supplementary table S1 summarises the clinical data for the six patients studied.
All patients presented with a similar clinical outcome and phenotype, which comprise growth retardation (with relatively short limbs, hands and feet and long thorax), mild-to-moderate mental retardation, mild skeletal dysplasia, constipation, deep voice and specific facial features (round/puffy, somewhat coarse, flat face, flat nasal bridge, upturned nose and hypertelorism), with relative macrocephaly (figures 1 and 2, see online supplementary figure S2). Other changes included puffy hands and feet, club feet, tortuosity of arteries of the dorsal area of the feet and hands and skin with rough and doughy texture (figure 3). In two patients with more pronounced clinical features mild cardiomyopathy was also noted.
The developmental phenotype was characterised by delayed delivery >40 weeks, relatively high birth weight and length, large head circumference relation to the chest circumference and floppiness (see online supplementary table S1). Newborn screening allowed exclusion of hypothyreosis. From the age of 2 years, the height gain became slower, leading to the marked short stature at the later age (see online supplementary figure S3). Growth retardation was disproportionate, affecting the lower segment more than the upper segment. The Polish reference charts for sitting height are available for children aged >4 years. The following indexes were calculated:
Trunk-lower extremities index (according to Giuffrida-Ruggeri) using formula: (sitting height/body height)×100.
Skelic Index (according to Manouvrier) using formula: (body height−sitting height/sitting height)×100.
Trunk-lower extremities index in all patients except one (patient 1 after puberty had skeletal deformity) revealed long trunk. Skelic index revealed all patients to be hyperbrachyskelic.
From infancy, all patients presented mild-to-moderate motor and mental retardation, deep voice and constipation.
Laboratory investigations revealed anaemia (low red blood cells and haemoglobin level), slightly elevated creatine kinase and cholesterol. The thyroid profile showed low free thyroxine (fT4) levels coupled with high free T3 (fT3), leading to an altered T4 : T3 ratio, along with normal TSH levels (see online supplementary tables S1 and S2).
X-ray examination revealed ovoid immature form of the vertebral bodies, anterior superior ossification defect in the lower thoracic and upper lumbar bodies and hypoplasia of the acetabular and supra-acetabular portions of the ilia and coxa vara. X-rays of hands showed minimal changes in the tubular bones of the hands, which became abnormally short, wide and deformed. One of the consulting radiologists suggested some similarities of X-ray examination to hypothyreosis (see online supplementary figure S4).
Attempts to treat affected patients with levothyroxine at standard doses did not improve any of the clinical symptoms.
The two patients with the most severe phenotype were selected for exome sequencing (patients 1 and 2; see online supplementary table S1). Both affected individuals were found to have heterozygous stop mutations in the last exon of the coding sequence for THRA isoform 1 (RefSeq NM_199334; c.1176C>A, [p.C392X] in patient 1 and c.1207G>T, [p.E403X] in patient 2). These mutations were absent in inhouse exomes (2094 exomes), as well as 6500 public exomes from Exome Variant Server (http://evs.gs.washington.edu/EVS/). Truncating mutations in the last exon of THRA isoform 1 were previously reported to cause a similar phenotype in three individuals from two families.1 We validated these mutations by Sanger sequencing and confirmed that they had occurred de novo in the patients. Using Sanger sequencing, we screened the rest of our cohort for mutations in the coding regions of THRA. We identified a heterozygous missense mutation in the last exon of THRA isoform 1 in two additional unrelated individuals (c.1207G>A, [p.E403K] and c.1193C>G, [p.P398R] in patients 3 and 4, respectively). The mutation in patient 3 was inherited from the father, who was also affected, while the mutation in patient 4 occurred de novo. All of these mutations cluster to the C-terminus of the TRα1 protein, leaving TRα2 unaltered (see online supplementary figure S5), which is consistent with the previous case reports.1 ,8
Despite sharing the same clinical features, no mutations in THRA were found in patient 5 after screening by Sanger sequencing. We performed additional genetic studies to identify the disease-causing genetic lesion in this individual, including exome sequencing and a high-resolution microarray copy number variation analysis. The absence of mutations in the THRA gene was confirmed by exome sequencing and no potentially pathogenic variants were identified in any candidate genes (including genes coding for TR cofactors and other proteins involved in TH metabolism and function). Furthermore, a high-resolution microarray failed to show potentially pathogenic copy number variations. Therefore, despite genome-wide screening by different methods, we failed to identify the disease-causing genetic hit in patient 5.
We describe a novel clinical and biochemical phenotype associated with truncating mutations in the THRA gene, a disorder that is of particular interest for several reasons. While resistance to thyroid hormone (RTH) due to mutations in THRB gene is a quite well-known disorder, patients with RTH caused by mutations in THRA had not, until recently, been identified. As no patients with classic RTH have been found to have mutations in THRA, it has been often suggested that mutations in THRA may be extremely rare or clinically unrecognised. However, our study suggests that milder clinical features and thyroid profile in patients with this novel form of RTHα may be the major cause. This discrepancy is most likely due to the differences in tissue expression of THRA and THRB; THRA is expressed in the cardiac and skeletal muscle, digestive tract, bones and brain, while THRB is found in the liver, kidneys, hypothalamus, hypophysis and thyroid gland. In both forms of RTH, organs expressing a dysfunctional TH receptor will respond in a tissue-specific way to the lack of TH stimulus. Mutations in THRB lead to selective pituitary RTH, releasing the negative feedback in the hypothalamus–pituitary–thyroid axis, which finally results in an increase in production of TH by the thyroid gland. The clinical picture of individuals with RTHβ often includes a mixture of features of hyperthyroidism, such as tachycardia or weight loss, and of classic hypothyroidism (for instance, delayed growth or alterations in bone ossification). Interestingly, although the uptake of THs is altered in several target organs, the signalling in the hypothalamus–pituitary–thyroid axis is preserved in RTHα because of the regular function of TRβ2 in the hypothalamus and pituitary gland. As a result, despite the deprivation of TH stimulus on peripheral tissues, the concentration of TSH in these individuals is found within the normal range. Despite normal/high T3 blood serum concentration, the patients present symptoms of hypothyreosis as the result of an impaired TRα function in T3 target tissues. All patients described so far with RTHα presented with clinical features resembling those of untreated mild congenital hypothyroidism such as mild-to-moderate intellectual disability, short stature, alterations in bone ossification, macroglossia and chronic constipation.1 ,8 ,11 Our study revealed additional hallmark features of RTHα such as macrocephaly, specific facial features, skeletal abnormalities (planovalgus foot, sandal gap deformities, club foot or spine deformities), mild features of skeletal dysplasia characteristic of hypothyreosis, anaemia, elevated creatine kinase and moderate hypercholesterolaemia. Additionally, our findings suggest a certain degree of correlation between the genotype and phenotype, as patients with nonsense mutations have intellectual disability, while patients with missense mutations show low IQ levels that are still within normal ranges. In line with this is the observation that the only inherited disorder was due to a THRA missense mutation, while all other mutations leading to more severe outcome were due to de novo mutations.
Although the clinical picture of RTHα reminds of hypothyroidism, the laboratory data in these patients are inconsistent. The results of laboratory analyses in the study group are not typical for hypothyroidism. Despite a low fT4 level, the concentration of fT3, physiologically the strongest of the THs, is high. Such correlation between the free THs results from a TRα mutation, which prevents complete utilisation of THs at the cellular level. Free THs fall within a broad, normal range. This results in a thyroid laboratory profile, which without detailed analysis is unremarkable for thyroid disease and thus hampers the correct diagnosis in these patients. The above may lead to a discontinuation of attempts to find causes of hypothyroidism and a misdiagnosis. However, if results of TH investigations show normal levels of TSH with low-normal T4 and high-normal T3 resulting in a low T4/T3 ratio, the possibility of deficit of TRα should be taken into consideration. Newborn screening for hypothyreosis, which is run in many countries, does not reveal TRα deficit. We strongly recommend that patients with some clinical signs/symptoms of hypothyroidism, without confirmation in routine endocrine studies should have a molecular study of THRA defects.
The clustering of all the mutations identified to the last exon of THRA supports that the mechanism through which this mutation leads to pathogenicity is by acting as a dominant-negative TRα1 protein constitutively binding a corepressor.1 Remarkably, one of the missense mutations identified in our cohort, TRα1 P398R, corresponds to a mutation of the homologous residue in TRβ1 (P452R), which has been previously identified to lead to RTHβ.12 The presence of equivalent pathogenic mutations in both TH receptors points to a common molecular mechanism underlying RTH in RTHα and RTHβ. Regarding the stop mutations in THRA, their localisation to the last exon allows them to avoid mRNA nonsense-mediated decay, permitting transcription of the protein. In our cohort, the two patients with nonsense mutations in THRA had the most severe clinical phenotypes accompanied by the largest deviations from the norm in the results of thyroid-related laboratory tests. At the same time, these two patients are the oldest in the observed group and their clinical symptoms seem to be intensifying with age. However, one can speculate that, despite that in dominant inheritance there is a wide spectrum of clinical expression, in our cohort there is genotype/phenotype correlation.
Despite presenting with the same clinical phenotype, the absence of mutations in THRA in one patient in our cohort suggests genetic heterogeneity in this syndrome. Additional genome-wide studies failed to identify the genetic lesion responsible for the phenotype in this patient. This is reminiscent of RTHβ in which 85% of affected individuals have mutations in THRB, while mutations in this gene have been excluded in the remaining 15% without thereby identifying the disease-causing genetic lesion.7 It has been hypothesised that individuals with so-called non-TR-RTH may have a mutation in a gene involved in TR function or TH metabolism. We speculate that this may also be the case for RTHα; a mutation or small deletion in a gene involved in TR function or TH metabolism may be the cause of the disease in patient 5. Our results suggest that this novel syndrome of TH resistance is genetically heterogeneous and there may be additional genes involved in the pathogenesis of this disorder.
The results from studies performed on mutant Thra mouse models suggest that the recruitment or action of coregulators, mainly corepressors, on TRα1 could be a target for treatment in individuals with THRA mutations.13 A better understanding of the role that TRα1 coregulator dysfunction plays in this genetic syndrome is necessary. However, pharmaceutical compounds targeting such nuclear receptor coregulators could be a promising treatment, for rare genetic syndromes such as RTH and for hormone-dependent cancers.14
Our study also has a practical aspect, as it is the first description of numerous cases of this defect, which turns out to be quite frequent in the practice of dysmorphologists, endocrinologists and paediatricians.
Truncating mutations in THRA gene lead to RTHα, a distinct and consistent phenotype of mild hypothyroidism associated with macrocephaly, specific facial features and skeletal abnormalities.
THRA mutations may be more common than expected and may escape standard laboratory/hormone level tests.
In patients with clinical signs/symptoms of mild hypothyreosis, without confirmation in endocrine studies, a molecular study of THRA defects is an important investigation.
The authors express their gratitude to Professor Kazimierz Kozłowski for his suggestions and X-ray descriptions, Ms Halina Witkowska for her invaluable help with patients’ identification and Dr Dorota Birkholz-Walerzak for her suggestions concerning patient 2. We are pleased to thank all patients and their families for their co-operation during the study. The authors are grateful to the Radboudumc Genomics Technology Center and all members of the Genomic Disorders Group and Developmental Genomics Group for technical assistance.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
AT-S and RA-H contributed equally.
Twitter Follow Rocio Acuna-Hidalgo at @burstingcell
Contributors AT-S: Conception and design, conduct of the work (patients’ clinical assessments), drafting and revising the article. RA-H: Conception and design, conduct of the work (performance of Sanger sequencing, analysis of exome sequencing data), drafting and revising the article. MK-W: Conception and design, conduct of the work (patients’ clinical assessments), drafting and revising the article. AL-A: Conception and design, conduct of the work, drafting and revising the article. MS: Conception and design, conduct of the work (preparation of libraries for exome sequencing), drafting and revising the article. CG: Conception and design, conduct of the work (development of the bioinformatics annotation pipeline for analysing exome sequencing data), drafting and revising the article. HGB: Conception and design, conduct of the work, drafting and revising the article. AJ: Conception and design, conduct of the work, drafting and revising the article. AR-Ś: Conception and design, conduct of the work (anthropometric measurements), drafting and revising the article. AH: Conception and design, conduct of the work, drafting and revising the article. KHC: Conception and design, conduct of the work and drafting and revising the article.
Funding AH was supported by the Netherlands Organization for Health Research and Development (ZonMW 916-12-095). RA-H was supported by a Radboudumc PhD grant.
Competing interests None.
Patient consent Obtained.
Ethics approval The Children's Memorial Health Institute.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.