Introduction
Thyroid hormone action is mediated by nuclear thyroid hormone receptors, encoded by 2 genes, thyroid hormone receptor α (THRA) and thyroid hormone receptor β (THRB), with distinct tissue distributions and phenotypes [1-3]. Resistance to thyroid hormones (RTH) is defined by reduced tissue sensitivity, presenting with elevated free thyroxine and triiodothyronine levels and inappropriately normal or mildly elevated thyroid-stimulating hormone (TSH). It is most often caused by mutations in THRB (OMIM *190160), and more rarely in THRA (OMIM *190120).
Resistance to thyroid hormone β (RTHβ; OMIM #188570) affects approximately 1 in 18 750 to 1 in 40 000 births [4, 5], and was first described by Refetoff et al in a family with deafness, stippled epiphyses and goiter [6]. RTHβ is usually inherited in an autosomal dominant manner or occurs de novo [7]. Clinical presentation varies widely, ranging from asymptomatic individuals to severe manifestations [8].
Goiter is the most common finding [8, 9]. Symptoms depend on the tissue-specific balance between thyroid hormone receptor β (TRβ) and thyroid hormone receptor α (TRα). Patients with RTHβ develop symptoms of hypothyroidism in tissues expressing predominantly TRβ, such as pituitary (leading to impaired negative feedback on TSH and goiter), brain (altered neuronal maturation), cochlea (deafness) and liver (neonatal jaundice).
Conversely, in tissues expressing mainly TRα (eg, heart, bone, brain), RTHβ results in thyrotoxic manifestations, including tachycardia, abnormal ossification, developmental delay, and attention deficit disorder [2, 8, 10].
Common causes of neonatal goiter are maternal Graves disease, maternal iodine excess, and dyshormonogenesis disorders [11]. We report the case of a neonate with RTHβ presenting with compressive goiter.
Case presentation
A male newborn was the second child of unrelated parents. His mother's first-trimester TSH level was normal. His father and brother had no medical history.
The infant was born spontaneously at 35 weeks of gestation in the context of preeclampsia. Antenatal ultrasounds were normal. His birth weight was 2470 g (53rd percentile).
Apgar scores were 4 and 8 at 1 and 5 minutes, respectively. He presented with immediate and persistent respiratory distress, requiring noninvasive ventilation. Immediate intubation attempts failed due to the presence of an anterior cervical mass.
Diagnostic assessment
Medical examination on day 1 revealed a large, bilobed anterior cervical mass (Fig. 1). Respiratory distress did not improve, and orotracheal intubation was performed.
Thyroid ultrasound showed a large goiter with an isthmus measuring 7 mm, compressing the trachea, which had a reduced diameter of 6 mm. The total thyroid volume was 9.7 mL (normal range, 0.45-1.53 mL in full-term neonates [12]). The first thyroid hormone assay showed elevated TSH at 73.8 mIU/L (normal range, 1.5-15.6); free thyroxine was markedly elevated at 77.1 pmol/L (59.9 pg/mL; normal range, 11.6-36 pmol/L [9.0-28.0 pg/mL]); free triiodothyronine exceeded 30 pmol/L (>19.5 pg/mL; normal range, 3.5-7.8 pmol/L [2.3-5.1 pg/mL]).
Thyroid autoantibodies, including TSH receptor, thyroperoxidase, and thyroglobulin, were negative. Cervical and brain magnetic resonance imaging revealed a large goiter extending posteriorly to the pharynx.
The diagnosis was genetically confirmed: the patient carries a de novo heterozygous nonsense mutation in exon 10 of the THRB gene, classified as pathogenic according to American College of Medical Genetics and Genomics criteria: c.1352_1353delinsAA; p.(Phe451Stop). This novel variant introduces a premature stop codon within the T3-binding domain. Given that the phenotype could not be fully explained by the THRB mutation alone, a next-generation sequencing panel targeting 25 genes involved in thyroid disorders was performed.
It identified an additional heterozygous missense variant in exon 10 of the TSHR gene: c.1703T>A; p.(Ile568Asn). This variant was inherited from the patient's father and grandmother, both of whom have normal thyroid function and no goiter (Fig. 2).
This familial segregation pattern suggests that the variant is of uncertain significance (class 3 according to American College of Medical Genetics and Genomics criteria). The patient's mother, who has a euthyroid multinodular goiter, carries the wild-type (WT) allele for both variants.
To investigate the underlying pathophysiological mechanisms and understand the patient's phenotype, we conducted functional analyses. In vitro assays were performed using HepG2 cells for THRB and Cos7 cells for TSHR transfections (Fig. 3).
For transactivation assays, HepG2 cells were transfected with 0.5 mcg of WT THRB (kindly provided by S. Refetoff, University of Chicago) or the Phe451* mutant, together with 1 mcg of a luciferase reporter plasmid containing thyroid hormone response elements from the chicken lysozyme gene (inverted repeat, TRE-F2) and stimulated with T3 (5 nM). Transient transfection of COS-7 cells with WT TSHR (hTSHR-pCMV6 plasmid; Origene) or the Ile568Asn mutant was performed using 0.5 mcg of WT or mutant plasmid together with 0.2 mcg of the luciferase reporter vector containing a cAMP-response element (pGL4.29; Promega), followed by stimulation with TSH (100 mIU/mL) for 30 minutes.
Luciferase activity was normalized to β-galactosidase activity. In the absence of T3, the TRβ mutant exhibited a significant 3-fold increase in transcriptional activity compared with the WT receptor (constitutive activity, P < .01). T3 stimulation failed to induce any change in luciferase activity for the mutant compared with the WT receptor, confirming that the mutation results in resistance to T3.
Furthermore, when equimolar WT and mutant receptors were cotransfected, the mutated receptor exerted a strong dominant-negative effect on the WT receptor (Fig. 3A). The familial TSHR variant did not affect the inositol 1,4,5-trisphosphate signaling pathway (not shown).
However, it led to a constitutive activation of the cAMP signaling pathway, showing a significant 3.8-fold increase in the absence of TSH (P < .01). The TSH-induced cAMP production was conserved in the mutated TSHR, as compared with the WT receptor (Fig. 3B).
Cotransfection of the TSHR mutant with the mutated THRB did not alter its transcriptional activity in comparison with cells expressing the THRB mutant alone (not shown).
Treatment
A single dose of L-thyroxine (50 mcg) was given before thyroid hormone test results were available, in the hypothesis of a severe dyshormonogenesis disorder. Following the availability of hormonal data, therapy with tiratricol, a T3 analog, was initiated on day 5 at an initial dose of 35 mcg/8 hours (40 mcg/kg/day), progressively increased to 300 mcg/day in 4 doses (100 mcg/kg/day). The patient exhibited symptoms of thyrotoxicosis, including episodes of supraventricular tachycardia on days 3 and 19, which were initially managed with adenosine, and then with propranolol and digoxin.
Tiratricol therapy led to a reduction of serum TSH levels to the upper limit of the normal range, with a decrease in goiter volume (Fig. 4) enabling well-tolerated extubation at 17 days of life (12 days after the start of tiratricol).
Outcome and follow-up
The patient exhibits borderline normal psychomotor development: he was able to sit unaided at 9 months, walked at 21 months, and began speaking single words at 13 months and word combinations by 22 months. He experienced premature dental eruption and has moderate hearing loss. At the most recent evaluation, at 4 years of age, he presents with language delay and attention-deficit/hyperactive disorder; his stature is within normal limits, at −1 SD score.
The goiter is now invisible, with a volume of 8 mL (normal range 2.8 ± 0.6 mL [13]), and serum TSH remains within the normal range. He is still treated with tiratricol at the dose of 75 mcg 3 times a day (15 mcg/kg/day). Cardiac evaluation is currently normal, and the patient maintains a sinus rhythm with well-tolerated tachycardia under β-blocker therapy.
Discussion
We report RTH in a newborn presenting with a compressive goiter. Genetic analysis revealed a THRB variant, a finding that was not sufficient to fully explain the clinical phenotype. Further genetic analysis revealed an additional TSHR variant constitutively activating the cAMP signaling pathway, and we assume that this exceptional combination of genetic anomalies leads to the excessive in utero growth of thyroid tissue.
Goiter is a common feature of RTHβ but develops progressively. There is no phenotype-genotype correlation with regard to goiter. In a retrospective study, 56% of children with RTHβ exhibited goiter, with a mean thyroid volume of +2.87 SD score [14].
Our understanding is that chronically mildly elevated TSH levels activate the TSHR and lead to chronic overproduction of thyroid hormones and goiter. The youngest report of compressive goiter was in a 4-year-old girl [15]. She had, like our patient, a truncating variant in exon 10 of THRB; truncating variants have been described to be more severe than missense variants [16].
A few case reports describe neonatal goiters [17, 18] or compressive symptoms [15, 19, 20], but cohort data are missing. We cannot exclude the possibility of somatic mosaicism, leading to localized homozygosity. Ferrara et al described 3 new cases of homozygous RTHβ in 2 consanguineous families [21], in which patients exhibited more severe symptoms, including profound developmental delay.
While childhood goiter was observed, compressive goiter was not reported.
Functional analysis demonstrated that the THRB mutation exerts a significant dominant-negative effect in vitro, abolishing transcriptional activation in response to T3. This mechanism has been previously described and can affect both WT TRβ and TRα [2]. Despite the presence of 1 normal allele, heterodimerization between WT and mutant receptors fails to achieve a transcriptional response.
TSHR activating mutations usually result in nonautoimmune hyperthyroidism with a wide degree of variation [22]. Centrally, this indirectly leads to suppressed TSH levels through TSH-independent thyroid hormone production. Gozu et al reported no clear genotype-phenotype correlation, although diffuse goiter was a common finding [23].
Thus, a diffuse goiter in the neonatal period is a diagnostic criterion for persistent sporadic nonautoimmune hyperthyroidism [22]. The degree of receptor activation observed in vitro was not predictive of clinical severity. Other variants affecting codon 568 have been reported to exert a stimulating effect on TSHR [24, 25], which aligns with our functional findings.
Although intrafamilial variability is frequent [22, 23], it is unusual that the patient's relatives, who carry the same TSHR variant, remain completely asymptomatic in adulthood.
It is therefore exceptional to have at the same time an activating TSHR mutation and elevated TSH levels, and we assume that the co-occurrence of these 2 events lead to the massive in utero thyroid growth in our patient. The TSHR promoter contains thyroid response elements leading to decreased TSHR expression when thyroid hormone levels increase [26]. In asymptomatic carriers, receptor hyperactivity could potentially be compensated by reduced membrane expression.
In our patient, defective thyroid hormone signaling could result in high TSH levels, therefore increasing the functional impact of the TSHR mutation and contributing to goiter development.
Tiratricol is approved for the management of peripheral thyrotoxicosis in patients with monocarboxylate transporter 8 deficiency and has been used for a long time in RTH [27, 28]. Structural differences between tiratricol and T3 allow the metabolite to bind and activate certain mutant receptors [28, 29]. Tiratricol binds to TRα with similar affinity to T3 and shows even higher affinity for TRβ.
Consequently, the drug alleviates hypothyroidism symptoms without exacerbating thyrotoxicosis signs, and several pediatric cases also reported successful use of tiratricol [30]. Therapeutic management of RTH aims to modulate the TRα/TRβ balance, thereby adjusting the clinical expression of hyper- and hypothyroid features [10]. Tiratricol cross-reacts with T3 assays, making T3 levels unreliable for dose adjustment, except for assessing patient adherence [30].
Therapeutic monitoring is thus based primarily on clinical signs.
Learning points