Molecular description of non-autoimmune hyperthyroidism at a neonate caused by a new thyrotropin receptor germline mutation

The patient was born after an uneventful pregnancy in the 36 week of gestation without signs of prior fetal tachycardia. In utero ultrasound data of the thyroid gland were not available. At the age of one month tachycardia was detected in a routine examination without further clinical signs of hyperthyroidism. At one month of age the weight was 3085 g and length 52 cm length. The thyroid status revealed severely elevated T3 (5.5 µg/L; normal range: 0.8 – 2.0 µg/L) and fT4 (≥ 7.8 ng/dL; normal range: 0.9 – 1.9 ng/d) levels, and supposed TSH below 0.01 mU/L (normal range: 0.27 – 4.2 mU/L). Thyroid autoantibodies were negative. Ultrasound of the thyroid gland showed no increase in volume (1.9 mL; normal range: 0.2 – 2.0 mL).

Due to severe tachycardia with 220/min the child was treated with Propanolol 3.1 mg/kg body weight/d. Antithyroid drug treatment with Thiamazol was initiated with a dose of 0.5 mg/kg body weight per day. In light of severe hyperthyroidism in terms of very high thyroid hormone levels the clinical course beside severe tachycardia was surprisingly asymptomatic including the relatively small volume of the thyroid gland. Actually at the age of two months the patient was treated with 2.5 mg Propanolol four times a day, and with Thiamazol 1 mg/d and D-Fluorette 500 IE/d.

All clinical investigations described were conducted in consultation with the investigated patient and her family and genetic analyses were in accordance with the guide-lines proposed in the Declaration of Helsinki.

Genomic DNA was prepared from peripheral white blood cells from the patient using a commercial kit (Blood amp kit, Qiagen, Hilden, Germany). Exon 9 and 10 of the TSHR gene were PCR-amplified and sequenced using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer, Weiterstadt, Germany) and an automatic sequencer (ABI 3710xl, Applied Biosystems, Foster City, CA, USA).

The PCR product of the mutation I486N was cloned into the pCR2.1-TOPO vector and then transformed into competent E.coli DH5α. DNA was extracted from selected colonies of bacteria using QIAprep Miniprep kit (QIAGEN GmbH, Germany), digested with Eco RI (New England Biolabs, MA, USA) and additionally sequenced for the detection of mutation I486N. The mutant was inserted into the eucaryotic expression vector encoding the wild-type (WT) TSHR (N-terminal hemagglutinine (HA)-tagged between amino acid 25 and 26) via Bsu36I and BstEII.

COS-7 cells were grown in Dulbecco´s modified Eagle medium (DMEM) and HEK-293 cells in Minimum Essential Medium (MEM) Earle´s, both supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 µg/ml streptomycin and 2 mM L-glutamine at 37°C in a 5% CO₂ incubator. For functional characterization WT TSHR and mutant receptor were transiently transfected into COS-7 or HEK 293 cells and investigated regarding cell surface expression and TSH-induced accumulation of intracellular cAMP and IP3 accumulation.

For all functional tests 2.5x10⁴ COS-7 cells (for cAMP accumulation) or 5 X 10⁴ HEK 293 cells (for IP accumulation) were seeded in 48-well culture plates. Transient transfection was performed by using metafectene (Biontex, Munich, Germany) according to the manufactures protocol. Cell surface expression was determined by immunological detection of the N-terminal HA-tag as described previously [10]. Intracellular cAMP was determined via alpha-screen technology [11] after stimulation with bovine TSH (bTSH). For investigation of Gq/11 activation, we co-transfected HEK-293 cells with TSHR cDNAs and a reporter construct containing the firefly luciferase gene under the control of NF-AT. Subsequently, the cells were stimulated with bTSH and lysed. Gq/11 activation was reported by luciferase activity expression in the luciferase reporter gene assay according to the manufacturer’s instructions (Promega, Madison, USA).

Using a previously described modeling procedure [12] we designed a new TSHR homology model of the serpentine-domain assembled with the extreme N- and C-terminal parts of the hinge region. As suggested by recently published experimental data [13] a cysteine-bridge between C284 at the N- and C408 at the C- terminus of the hinge region was added. To improve the model and to increase the quality (refined side-chain orientations) we performed in contrast to the previous procedure an extended molecular dynamic simulation of 6ns for the side-chains by fixed backbone atoms. Structure images were produced using the PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC.

For functional characterization the mutated TSHR (I486N) and wild type TSHR were transiently transfected into either COS-7 cells to detect cell surface localization and cAMP signal transduction or HEK-293 cells to investigate IP3 signaling properties.

The mutant TSHR expression at the cell surface was reduced to 51% ± 7% of WT TSHR (Fig.1 A).

We determined a 4-fold increase in ligand independent activation of the Gαs signaling pathway compared with the WT (I486N: 404% ± 67% of WT; Fig.1B). The maximal level of activation of the mutant was slightly reduced compared with WT (I486N: 74% ± 8% of WT; Fig. 1B), whereby the EC₅₀ values were similar (EC50 WT: 0.18 ± 0.02 mU/ml; I486N: 0.35 ± 0.17 mU/ml).

In contrast, determination of basal and TSH- induced activation of the Gαq/11 signaling pathway revealed a 13-fold impairment of signaling capability of the I486N mutation (I486N: 8% ± 2% of WT). The basal activity was like wild type (Fig.1C).

It can be assumed that the extreme N- and C-terminus of the extracellular hinge region are localized and tightly embedded between the ECLs (figure 2). Three facts do support this three-dimensional package:

Constitutively activating mutations in the TSHR are a frequent molecular cause of non-autoimmune hyperthyroidism (reviewed in [16‐19]). We here describe a new germline TSHR mutation I486N located in the ECL1. In the affected neonate hyperthyroidism was diagnosed due to tachycardia without further signs of hyperthyroidism and without goiter. Thyroid hormone values were severely elevated revealed non-autoimmune hyperthyroidism.

Three somatic mutations at position 486 were found previously in toxic thyroid adenomas: I486F [20], I486M [20], I486N [21]. Functional characterization for mutations I486F and I486M demonstrated constitutive activation of both Gs and Gq mediated signaling. The mutation I486N we identified in this study as a germline mutation was already found as a somatic mutation [21]. In that study mutation I486N lead to constitutive activation of the Gs pathway, however, activation of Gq activation was not investigated. We here show that I486N did not result in constitutive activation of Gq/11 (figure 1) in contrast to the I486F and I486M mutations. Moreover, the capability for bTSH-induced Gq-mediated signaling was impaired in the I486N mutation.

Interestingly, at the neighbouring position 485 a germline mutation (A485V) was described previously [22]. This mutation also reveals constitutive activation of the Gs mediated pathway, whereas the basal level of IP3 formation remained unchanged. Stimulation of that mutant TSHR with high concentrations of bTSH failed to activate the IP3-Gq pathway. The patient presented with elevated thyroid hormone levels and a suppressed TSH level in the absence of thyroid hormone antibodies, leading to the diagnosis of non-autoimmune hyperthyroidism. No measurable goiter was found in this patient and the phenotype was described as a mild form of hyperthyroidism, although the functional characterization revealed a high constitutive basal activity. This phenotype and functional characterization of the mutated TSHR is similar to our findings regarding the patient-phenotype and molecular characteristics of the TSHR variant I486N. We recently described a patient`s phenotype due to a TSHR mutation with comparable functional characteristics as non-autoimmune and non goitrous hyperthyroidism [23].

To understand the functional and structural role of TSHR-I486N consideration of naturally occurring TSHR mutations in the extracellular loops provides valuable information. Naturally occurring activating mutations in the ECLs are reported several times in the TSHR: ECL1- I486M,F,N [20, 21] and A485V [22]; ECL2 - I568T,V [20, 24], ECL3 - N650Y [25] and V656F [26]. These activating mutations are indicators that the extracellular loops acting as important determinants for receptor activation. In support to this assumption also few naturally occurring inactivating mutations were published previously: ECL1 - W488R [27], Q489H [28]; ECL3 - L653V [29].

The potential interplaying counterparts of the ECLs (figure 2) are likely the extreme N- and C-terminus of the extracellular hinge region where also activating as well as inactivating mutations were reported (reviewed in [1]). Mutations at I486 likely modify the conformation of loop 1 and by this lead to re-justification of the extracellular receptor parts to each other. This spatial shift induces interruption of side-chain interactions or induces new interactions between amino acids which finally effects the arrangement of connected transmembrane helices (intramolecular signal transduction). The specific importance of the TSHR ECL1 in this scenario was evidenced and specified by site-directed mutagenesis studies [15, 30]. In these studies the ECL1 was identified as a key-transducer and amplifier of extracellularly provided signals, which is supported by findings at other GPCRs [31, 32].

Of note are the diverse functional characteristics of different I486 mutations tested in cell-systems. Two observations which might be exemplarily also for other pathogenic mutations we would like to highlight: