Sensitive Sequencing Analysis Suggests Thyrotropin Receptor and Guanine Nucleotide-Binding Protein G Subunit Alpha as Sole Driver Mutations in Hot Thyroid Nodules

Abstract

Background:

Constitutively activating mutations in the thyrotropin receptor (TSHR) and the guanine nucleotide-binding protein G subunit alpha (GNAS) are the primary cause of hot thyroid nodules (HTNs). The reported prevalence of TSHR and GNAS mutations in HTNs varies. Previous studies show TSHR mutations in 8–82% of HTNs and GNAS mutations in 8–75% of HTNs. With sensitive and comprehensive targeted next-generation sequencing (tNGS), we re-evaluated the prevalence of TSHR and GNAS mutations in HTNs.

Methods:

Samples from three previous studies found to be TSHR and GNAS mutation negative were selected and re-evaluated using high-resolution melting (HRM) PCR. Remaining mutation negative samples were further reanalyzed by tNGS with a sequencing depth between 3000 × and 10,000 × . Our tNGS panel covered the entire TSHR coding sequence along with mutation hot spots in GNAS. Sequencing reads were aligned to reference and variants were called using Torrent Suite software v5.8.

Results:

In total, 154 of 182 previously mutation negative HTNs were positive for TSHR or GNAS mutations, resulting in an 85% prevalence of TSHR and GNAS mutations in HTNs, 79% and 6%, respectively. In a subset of 25 HTNs with multiple samples per nodule, and analyzed by tNGS at high sequencing depth, TSHR mutations were detected in 23 (92%) HTNs and 1 GNAS mutation was detected in 1 (4%) HTN, 96% mutation positive HTNs in this subset.

Conclusions:

Owing to the higher sensitivity of tNGS as compared with denaturing gradient gel electrophoresis and HRM-PCR, TSHR or GNAS mutations could be detected in 85% of HTNs. The detection of TSHR and GNAS mutations occurred in 96% of HTNs in a sample set with multiple samples per nodule analyzed by tNGS. Taken together with the fact that no other driver mutations could be identified by whole exome sequencing, our study strongly supports the hypothesis that TSHR and GNAS mutations are the main somatic mutations leading to HTNs.

Introduction

Constitutively activating mutations in the thyrotropin receptor (TSHR) are the primary cause of hot thyroid nodules (HTNs) (1,2). Somatic point mutations that constitutively activate the TSHR were first identified by Parma et al. in HTN (3). These mutations explain the clinical phenotype of HTNs through their activation of the G_s or the G_s and G_q/11 pathways, resulting in a stimulation of proliferation, differentiation, and thyroid hormone synthesis (4). Sanger sequencing was the original method of detection used (3). Owing to the low sensitivity and the variable extent of sequencing (e.g., only parts of exon 10 vs. exons 9 and 10 sequenced), variable frequencies for TSHR mutations in HTNs have been previously described. The prevalence of TSHR mutations in HTNs has been reported to vary from 8% to 82% (3,5 –16). To increase sensitivity of detection, ease of application, and cost effectiveness, denaturing gradient gel electrophoresis (DGGE) was employed (17). Despite the increased sensitivity of DGGE (17), a comprehensive study revealed a frequency of only 57% TSHR mutations in 75 consecutive HTNs (18).

Further increased sensitivity and higher throughput were achieved by the use of high-resolution melting (HRM) PCR. In addition to higher sensitivity, the advent of HRM-PCR offered further improvements in detection of somatic mutations through simplification of analysis. The use of HRM-PCR showed higher frequency of TSHR mutations of 66% in 33 HTNs (19).

In addition to TSHR mutations, mutations in the guanine nucleotide-binding protein G subunit alpha (GNAS) can also result in an HTN phenotype (9). The prevalence of GNAS mutations in HTNs has been reported to vary from 8% to 75% (3,5 –16). These mutations are specific to two hotspot loci, c.201 and c.227. Sanger sequencing or pyrosequencing has been the primary methods of detection of GNAS mutations in HTNs.

Much of the reported variability in the prevalence of TSHR and GNAS mutations in HTNs can likely be attributed to the sensitivity of the detection method used. To determine whether another gene may be responsible for the remaining TSHR and GNAS mutation negative cases of HTNs, whole exome sequencing (WES) was performed for 13 TSHR and GNAS mutation negative samples (20). This study found no further causative mutation for HTNs. Therefore, by using the more sensitive and comprehensive targeted next-generation sequencing (tNGS), our aim was to determine an accurate prevalence of TSHR and GNAS mutations based on increased sensitivity and coverage. This would allow us to further evaluate the impact of these mutations on the molecular etiology of HTNs in a large cohort of HTNs derived from two different regions in Europe (Turkey and Germany) (18,21,22).

Materials and Methods

DNA samples

In this study, DNA from fresh frozen tissue samples of 148 HTNs from 3 studies performed by Trulzsch et al., Gozu et al., and Sancak et al. was reanalyzed (18,21,22). At the time of the first analyses (18,21,22), fresh frozen tissue samples derived from the HTNs were grinded, homogenized, and DNA was extracted from a portion of the homogenized tissue. If DNA from these initial studies was no longer available for the present investigations, DNA was re-extracted from the original homogenized tissue samples.

The samples analyzed were 56 samples from Germany (Trulzsch et al.) and 92 samples from Turkey (60 samples from Gozu et al. and 32 samples from Sancak et al.). Samples excluded were 12 TSHR and GNAS mutation negative HTNs from these studies for which no adequate DNA or frozen tissue was available, and 7 mutation positive HTNs from the previous publications that could not be matched to the tissues. An additional 34 frozen samples were collected after publication of these studies by the investigators. This included 14 samples from Germany and 20 samples from Turkey (11 from Sancak et al. and 9 from Gozu et al.).

Genomic DNA was extracted from the additional frozen tissue samples after grinding and homogenizing the tissue samples using the QiaAmp DNA kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. Tissue samples collected postpublication had undergone only HRM-PCR analysis before tNGS. The additional DNA samples were also included for a total of 182 HTNs analyzed to calculate prevalence of TSHR and GNAS mutations in HTNs. All samples from the study by Trulzsch et al. (18) were from solitary autonomous nodules. Fifteen samples from Sancak et al. (22) were obtained from solitary autonomous nodules, and the additional 17 samples were from toxic multinodular goiter (TMG). Fifteen samples from Gozu et al. (21) were obtained from solitary autonomous nodules and the additional 45 samples were obtained from TMG.

The study was approved by the local ethics committees. Written informed consent was obtained from all patients before study participation.

Denaturing gradient gel electrophoresis

PCR, DGGE, and sequencing analyses were performed as described previously by Trulzsch et al. (18). PCR products showing mutations in exons 9 and 10 of the TSHR gene by DGGE were subsequently sequenced using Big Dye-terminator chemistry (Applied Biosystems, Darmstadt, Germany) according to the manufacturer's instructions and analyzed on an automatic sequencer ABI 3130xl (Applied Biosystems).

High-resolution melting PCR

Previously DGGE mutation negative samples were re-evaluated using the more sensitive HRM-PCR. TSHR point mutations were detected by real-time PCR and HRM, using primers reported in the study by Eszlinger et al. (19), encompassing exons 9 and 10 (those exons in which all the previously reported constitutively activating TSHR mutations were detected) using the LightCycler 480 High Resolution Melting Master chemistry (Roche, Mannheim, Germany) on a LightCycler 480 (Roche). PCRs to detect TSHR point mutations were processed through an initial denaturation at 95°C for 10 minutes followed by 55 cycles of a 3-step PCR, including 3 seconds of denaturation at 95°C, a 10 seconds annealing phase at 56°C, and a 10 seconds elongation phase at 72°C. Thereafter, an HRM curve was assessed from 75°C to 95°C with an increase of 0.02°C/s and 25 acquisitions per degree.

Sanger sequencing

DNA from samples that tested positive by HRM-PCR for TSHR mutations was subsequently sequenced using Big Dye-terminator chemistry (Applied Biosystems) according to the manufacturer's instructions and analyzed on an automatic sequencer ABI 3730xl (Applied Biosystems). The same methodology was applied for the known hotspots c.201 and c.227 of GNAS.

Pyrosequencing

GNAS point mutations in position c.201 and c.227 were detected by pyrosequencing using self-designed primers and the following PCR conditions: PCRs were processed through an initial denaturation at 95°C for 15 minutes followed by 42 cycles of a 3-step PCR, including 20 seconds of denaturation at 95°C, a 30 seconds annealing phase at 60°C, and a 30 seconds elongation phase at 72°C, followed by a final 5 minutes extension phase at 72°C. PCR products were immobilized to streptavidin sepharose beads and single-stranded DNA was prepared allowing subsequent annealing of the sequencing primer to the template DNA. Then, the primed single-stranded DNA was transferred to a PyroMark Q24 (QIAGEN) for pyrosequencing (19).

Targeted next-generation sequencing

A custom AmpliSeq panel (ThermoFisher Scientific, Inc., Waltham, MA) was designed using AmpliSeq Designer software to detect variants in the entire TSHR coding sequence and exons 8 and 9 of GNAS. Panel size was 4.62 Kb with 24 amplicons divided among two pools. For each pool, sample libraries were generated from 10 ng of DNA using Ion AmpliSeq Library Kit 2.0 (ThermoFisher Scientific) and barcoded with Ion Xpress barcode adapters (ThermoFisher Scientific). Sample pools were combined and quantified using the Agilent High Sensitivity DNA kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). Sample libraries were then combined and diluted to 30 pM and loaded on the Ion Chef instrument for automated template preparation and chip loading onto an Ion 520 chip.

Loaded chips were sequenced using 200 bp sequencing chemistry with 500 flows on an Ion S5 XL sequencer (ThermoFisher Scientific). Sequencing reads were generated with Torrent Suite software v5.8 (ThermoFisher Scientific) and aligned to human reference genome 19 (hg19). Variant calling was performed with the Torrent Variant Caller plugin and annotated using Ion Reporter v5.10. In brief, variants were called using somatic stringency parameters optimized for low frequency variant detection >2.5%. All variants were manually reviewed using Integrated Genomics Viewer (23). Mean depth of sequence was ∼8800 × .

Relevance of novel TSHR mutations

Novel TSHR mutations were evaluated based on SIFT and PolyPhen2 scores to predict whether amino acid substitutions were deleterious (24,25).

Statistical analysis

Pearson's chi-squared test and Fisher's exact test were used as appropriate to determine difference in frequencies using standard R packages.

Results

Prevalence of TSHR and GNAS mutations in HTNs

The aim of this study was to accurately determine the true prevalence of TSHR and GNAS mutations in HTNs. To do so, we analyzed 182 HTNs by combining the previously published DGGE and Sanger sequencing results (18,21,22) with results obtained by DGGE in 34 samples collected following our publication, HRM-PCR and pyrosequencing of 23 DGGE mutation negative samples collected following our publication, and of 67 published DGGE mutation negative samples, and tNGS for 38 HRM-PCR and pyrosequencing mutation negative samples, 4 of which were collected after publication (18,21,22). Only samples that were mutation negative were included for mutation testing in the next methodology. Samples with positive results were not reanalyzed. Ten samples with known polymorphisms were included in the mutation negative category and reanalyzed.

Of the 67 published samples with wild type DGGE results, reanalysis showed 20 had TSHR mutations detected by HRM-PCR, 2 had GNAS mutations detected by pyrosequencing (Supplementary Table S1), 6 and 3 samples had TSHR or GNAS mutations, and 1 sample had both a TSHR and GNAS mutation detected by tNGS. Of 34 additional samples collected following our prior publication, 11 samples had TSHR mutations detected by DGGE (Supplementary Table S2), 4 had TSHR mutations detected by HRM-PCR, 1 had a GNAS mutation detected by pyrosequencing (Supplementary Table S3), and no additional TSHR mutation was detected by tNGS.

A total of 38 HTNs were analyzed by tNGS; 34 of these samples were from prior publications and tested as mutation negative by HRM-PCR, while 4 were collected after our publication (18,21,22). The 34 previously published HTNs were earlier found to be TSHR and GNAS mutation negative as tested by Sanger sequencing, DGGE, HRM-PCR, and pyrosequencing. Of these 34 HTNs, TSHR or GNAS mutations could be identified in an additional 6 and 3 HTNs, respectively, with an additional sample found to have co-occurring TSHR and GNAS mutations by tNGS (Table 1).

Table 1.

Somatic Mutations Identified Using tNGS, Organized by Ascending Variant Allele Frequency of GNAS and TSHR Mutations

Sample ID	DGGE result	HRM result	tNGS result	Variant allele frequency as detected by tNGS, %	SIFT/PolyPhen2 scores for newly identified mutations	Amino acid change	Described in TSHR mutation database
S15	WT	WT	GNAS c.681G>T	37		GNAS Q277H	n/a
T131	WT	WT	GNAS c.681G>T	25		GNAS Q227H	n/a
T136	WT	WT	GNAS c.601C>T	14		GNAS R201C	n/a
S25	WT	WT	TSHR c.1802A>T	35	0.0/1.0	TSHR Y601P	Newly identified
			TSHR c.1458C>G	26		TSHR I486M	Yes
			TSHR c.1358T>C	13		TSHR M453T	Yes
			TSHR c.1535T>A	8		TSHR L512Q	Yes
			TSHR c.1714_1715delATinsGC	6	0.0/0.998	TSHR M572A	Newly identified
			TSHR c.2009A>G	5		TSHR N670S	Yes
S7	WT	WT	TSHR c.1919T>C	5	0.0/1.0	TSHR I640T	Newly identified
G8.2	WT	WT	TSHR c.1658C>T	4	0.0/1.0	TSHR A553V	Newly identified
G8.2	WT	WT	TSHR c.1714A>G	4	0.0/0.993	TSHR M572V	Newly identified
S24	WT	WT	TSHR c.1878G>C	4		TSHR M626I	Yes
T73	WT	WT	TSHR c.1387A>G	3		TSHR M463V	Yes
S10	WT	WT	TSHR c.1919T>C	3	0.0/1.0	TSHR I640T	Newly identified
G48.2	WT	WT	GNAS c.602G>A	34		GNAS R201H	n/a
G48.2	WT	WT	TSHR c.1919T>C	4	0.0/1.0	TSHR I640T	Newly identified

Samples were separated into three groups: samples with GNAS mutations, samples with TSHR mutations, and a single sample with both a TSHR and a GNAS mutation. The first letter in sample ID indicates which publication the sample originated from: S [Sancak et al. (22)], T [Trulzsch et al. (18)], and G [Gozu et al. (21)]. TSHR Mutation database found at tsh-receptor-mutation-database.org, updated 2019. Mutations described in the mutation database have been characterized as constitutively activating. Mutations detected for the first time in this study have SIFT/PolyPhen2 score indicating their relevance (23,24).

DGGE, denaturing gradient gel electrophoresis; GNAS, guanine nucleotide-binding protein G subunit alpha; HRM, high-resolution melting; tNGS, targeted next-generation sequencing; TSHR, thyrotropin receptor; WT, wild type.

Variant allele frequencies (VAFs) of detected mutations varied from 3% to 37%. Twelve TSHR and GNAS mutation negative HTNs analyzed in the three previous studies did not have remaining DNA for further analysis by tNGS. Therefore, in total 154 HTNs were positive for TSHR or GNAS mutations resulting in an 85% prevalence of TSHR and GNAS mutations in HTNs. In detail, 144 HTNs (79% of HTNs, 93.5% of mutation positive HTNs) were found to have TSHR mutations, and 10 HTNs (6% of total HTNs or 6.5% of mutation positive HTNs) were found to have GNAS mutations. The number of mutation positive samples was not significantly different between solitary nodules and nodules from TMG (Fischer exact test, p = 1).

Analysis of multiple samples per nodule and intra-nodule mutation variability

For 51 HTNs from the study by Gozu et al. (21), a single tissue sample per nodule was available for mutation screening. A TSHR mutation was detected in 42 of the 51 HTNs (82%) and 1 GNAS mutation was detected in 1 HTN (2%), resulting in a total of 84% mutation positive HTNs.

For 25 HTNs from the study by Gozu et al. (21), multiple samples from different regions of the same HTN were available for molecular testing. While 18 HTNs of this subgroup yielded consistent results for all samples analyzed per nodule, in 7 samples discrepant results were detected (Table 2). Of the seven HTNs with discrepant results for samples derived from different regions of a nodule, all had at least one sample with a wild type result (Table 2). While in four cases (G31.2, G15.3, G4.2 and G24.3), the wild type mutation status was associated with cystic, necrotic, or degenerative features reported by the pathologist during the gross examination of the surgical specimens, the cystic degenerative areas reported for nodule G12 cannot precisely be linked to the two wild type samples G12.2 and G12.3 (Table 2).

Table 2.

Results of Analysis by tNGS and Pathology Findings for Hot Thyroid Nodules from Gozu et al. Samples

Sample ID	Nodule size	Region of the nodule where the tissue sample was taken from	Pathology notes	Mutations detected
G7.2	2.5 × 1.6 cm	Lateral part	No notes for cystic changes	WT
G7.3	2.5 × 1.6 cm	Medial part	No notes for cystic changes	TSHR N372T
G48.2	3 × 3 cm	Lateral part	No notes for cystic changes	TSHR I640TGNAS R201H
G48.3	3 × 3 cm	Central part	No notes for cystic changes	WT
G31.1	3.5 × 3 cm	Superior part		TSHR T632A
G31.2	3.5 × 3 cm	Central part	Cystic degeneration	WT
G12.1	4 × 4 cm	Superior part	Cystic degeneration	TSHR M453T
G12.2		Lateral part		WT
G12.3		Medial part		WT
G15.1	2 × 2 cm	Superior part		TSHR S505N
G15.2		Medial part		TSHR V656F
G15.3		Lateral part	Cystic degeneration	WT
G4.1	5 × 3.5 cm	Central part		TSHR F631I
G4.2		Superior part	Necrosis	WT
G4.3		Inferior part		TSHR F631I
G24.1	6 × 5.5 cm	Superior part		TSHR T632I
G24.2		Lateral part		TSHR T632I
G24.3		Medial part	Cystic degeneration	WT
G24.4		Inferior part		TSHR T632I

Gozu et al., (21).

Discrepant results were found when more than one sample was analyzed per hot thyroid nodule.

Interestingly, one 2 × 2 cm HTN with 3 available samples from different regions of the nodule was found to have 2 different TSHR mutations as well as 1 wild type result (sample G15; Table 2). Overall, a TSHR mutation could be detected in 23 of the 25 HTNs (92%) and a GNAS mutation was detected in 1 HTN (4%), resulting in a total of 96% mutation positive HTNs (Fig. 1).

FIG. 1.

Prevalence of TSHR and GNAS mutations in HTNs from the Gozu et al. (21) sample set. (A) HTNs with one sample (n = 51) vs. (B) HTNs with more than one sample per nodule (n = 25). GNAS, guanine nucleotide-binding protein G subunit alpha; HTNs, hot thyroid nodules; TSHR, thyrotropin receptor.

Multiple TSHR mutations in a single nodule detected by tNGS

In addition to HTN G15 with 3 available samples (G15.1, G15.2, and G15.3) from different regions of the nodule (Table 2), 2 HTNs (G8.2 and S25) were found to have more than 1 TSHR mutation in a single sample (Table 1). A single sample from the 2 × 2 cm HTN G8.2 was found to have two TSHR mutations by tNGS (Table 1). In addition to these two somatic TSHR mutations, tNGS also found a TSHR c.154C>A (p.P52T) mutation, which is a known polymorphism (26).

Only one tissue sample was retrieved from HTN S25, a 4.3 × 2.1 cm nodule. tNGS analysis of this sample found six TSHR mutations of various VAFs ranging from 5% to 35% (Table 1).

Two samples were taken from the 2.8 × 1.4 cm HTN G48. While sample G48.2 taken from the lateral part of the HTN was found to have both a TSHR and GNAS mutation (Table 1), sample G48.3 taken from the central part of the HTN was wild type.

Discussion

With the aim of determining the accurate prevalence of TSHR and GNAS mutations and to further evaluate the impact of these mutations on the molecular etiology of HTNs, we reanalyzed 38 TSHR and GNAS mutation negative HTNs from 182 HTNs previously analyzed by DGGE and HRM using a comprehensive tNGS panel covering the entire coding sequence of the TSHR gene and hot spots for GNAS constitutively activating mutations. This included 34 previously published TSHR and GNAS mutation negative HTNs and an additional 4 mutation negative HTNs collected after the 3 original publications (18,21,22).

Overall, our findings show an 85% prevalence of TSHR and GNAS mutations in HTNs. This is significantly higher than previous reports: Trulzsch et al. (18), Sancak et al. (22), and Gozu et al. (21) who reported 60%, 43%, and 72% of samples to harbor TSHR or GNAS mutations, respectively (p = 1.187 × 10⁻⁷, Pearson's chi-squared test). In all subsets, the predominant reason for this increased prevalence is likely the higher sensitivity in mutation detection methodologies from Sanger sequencing to tNGS.

Based on a mean depth of coverage of ∼8800 × in our tNGS panel, it is highly likely that we can detect mutations at frequencies down to 3%. Detection of lower mutation frequencies might have been hampered by PCR error (which is ∼1%) due to the fact that we did not use molecular identifiers to control for this source of error. In comparison with the previously used methods, this increased tNGS sensitivity allows for more accurate detection of mutations in HTNs than previously described. Therefore, although we identified low VAFs between 3% and 8% for 11 out of 20 mutations detected by tNGS (Table 1), we do interpret these results as accurate based on our high sensitivity and depth of coverage.

Moreover, all of these mutations have been validated through visual inspection of sequencing reads in the Integrated Genomics Viewer (27). The low VAF findings and also wild type findings could be explained by sampling in an area of the nodule with a low number of mutated cells due to bleeding, necrosis, cystic changes, calcification, degeneration, or other unknown factors contributing to an allele frequency that is not representative of the whole nodule. Furthermore, DNA was extracted from homogenized grinded tissue, therefore, low VAFs may be explained by dilution if the mutation was only present in a subclone within the nodule. Available pathology reports of tNGS wild type samples frequently described at least one of the mentioned characteristics (Table 2). While less likely, contamination by normal thyroid tissue could also lead to decreased VAFs.

Of the Gozu sample set (21), 7 of 25 HTNs with multiple tissue samples per nodule had discrepant results where at least 1 sample was wild type (Table 2). Had these wild type samples been the only samples analyzed, the mutations reported by DGGE, HRM, and our tNGS analysis would have been lower in number. The comparison of the subgroups of HTNs with one sample (n = 51) versus HTNs with more than one sample per nodule (n = 25) within the Gozu sample set reveals a substantially, but not significant, higher mutation prevalence of 96% versus 84% in the HTNs with multiple samples per nodule (Fig. 1; p = 0.26, Fisher's exact test for Count Data). While the analysis of multiple samples per nodule in this sample set explains the greater number of mutations detected, it also substantiates our hypothesis that issues of sampling may lead to undetectable mutation levels in one sample/area of an HTN, and a detectable mutation in another sample/area of the same nodule.

It is important to note that we detected new mutations that have not yet been characterized in cell culture and included those in our prevalence calculations. Based on SIFT/PolyPhen2 scores, we can assume that these mutations have an effect on the protein function, however, without cell culture experiments we cannot make a definitive claim that these mutations are responsible for the hot phenotype. While we are aware of this limitation and its effect on our ability to draw strong conclusions, functional characterization was not the focus of this study and is outside the scope of our aims.

Recently, WES of 13 TSHR and GNAS wild type HTNs detected no further causative driver mutations, which could explain the clinical phenotype of the HTNs (20). Although, based on our current findings, the analysis of single samples per nodule seems to be a limitation of that study, the fact that no further driver mutation could be detected by WES supports the hypothesis that TSHR and GNAS mutations are the only driver mutations in HTNs. This hypothesis gains further support by our current finding of a 96% TSHR/GNAS mutation prevalence in HTNs detected in an optimal sample set with multiple samples per nodule and analyzed by a comprehensive tNGS panel with high sequencing depth.

Of particular interest are the 3 HTNs with more than 1 mutation. To the best of our knowledge, HTNs with multiple TSHR mutations in the same nodule have not previously been described. Both mutations (A553V and M572V) detected in HTN G8.2 are newly identified with very low VAFs of 4%. Histology of this HTN described signs of bleeding and cystic areas. Both A553V and M572V have not yet been characterized as constitutively activating and these amino acids have previously not been identified with constitutively activating mutations but are found in exon 10. Based on their SIFT and PolyPhen2 scores, both mutations are described with a high probability as deleterious mutations.

Interestingly, in amino acid position 553, an inactivating mutation has been previously described 4 times (28 –31). There is one other instance (amino acid 593) in the TSHR mutation database (tsh-mutation-database.org) (32) where different mutations at a particular locus are both activating and inactivating. Based on the low VAF of 4%, it is unclear what the biological relevance of these mutations is. More importantly, the VAF may be heterogeneous throughout HTNs and this low VAF finding is not necessarily representative of the entire HTN.

HTN G48.2 was found to have both a previously undescribed TSHR I640T (c.1919T>C, VAF 4%) and a GNAS R201H (VAF 34%) mutation. While the TSHR mutation, I640T, is not found in the TSHR mutation database (tsh-mutation-database.org) (32), another somatic mutation, I640K, at this position has been found to be constitutively activating (21). The GNAS mutation is known to be constitutively activating. The biologic and signaling effect and the relevance of these co-occurring mutations are not known.

tNGS analysis of HTN S25, a 4.3 × 2.1 cm HTN, found 4 known constitutively activating TSHR mutations with varying VAFs (5–26%), and 2 additional newly identified mutations TSHR c.1802A>T (Y601P) and TSHR c.1714_1715delATinsGC (M572A) with VAFs of 35% and 6%, respectively (Table 1). For the amino acid position of one of these new mutations (Y601P), a constitutively activating mutation (Y601N) has been described (33). Based on their SIFT (25) and PolyPhen2 (24) scores, both mutations are described with a high probability as deleterious mutations.

None of these mutations were detected in other samples, making contamination error an unlikely reason for this finding. However, the biological relevance, the signaling effect, and the etiologic mechanism of more than one TSHR mutation in the same HTN are unknown. The phenotype of this HTN is not different from other nodules in size or hormone levels before treatment onset. The sequencing results alone do not indicate whether the TSHR mutations accumulated over time in a cell population derived from a single progenitor or whether the TSHR mutations occur in isolated distinct populations.

In conclusion, owing to the higher sensitivity of the tNGS panel compared with DGGE and HRM-PCR, TSHR or GNAS mutations could be detected in 85% of HTNs. In an optimal sample set with multiple samples per nodule analyzed by a comprehensive tNGS panel with high sequencing depth, the prevalence of TSHR and GNAS mutations rose to 96%. These results, in addition to the fact that no further causative driver mutations explaining the phenotype of HTNs could be identified by previous WES, strongly support the hypothesis that TSHR and GNAS mutations are the only somatic mutations leading to HTNs.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by a Laura Linders Grant to Alexandra Stephenson.

Supplementary Material

Supplementary Tables S1

Supplementary Tables S2

Supplementary Tables S3

References

Paschke

, Ludgate

. 1997. The thyrotropin receptor in thyroid diseases. N Engl J Med, 337:1675–1681.

Krohn

, Maier

, Paschke

. 2007. Mechanisms of disease: hydrogen peroxide, DNA damage and mutagenesis in the development of thyroid tumors. Nat Rev Endocrinol, 3:713.

Parma

, Duprez

, Van Sande

, Cochaux

, Gervy

, Mockel

, Dumont

, Vassart

. 1993. Somatic mutations in the thyrotropin receptor gene cause hyperfunctioning thyroid adenomas. Nature, 365:649.

Kero

, Ahmed

, Wettschureck

, Tunaru

, Wintermantel

, Greiner

, Schütz

, Offermanns

. 2007. Thyrocyte-specific G q/G 11 deficiency impairs thyroid function and prevents goiter development. J Clin Investig, 117:2399–2407.

Führer

, Holzapfel

H-P

, Wonerow

, Scherbaum

, Paschke

. 1997. Somatic mutations in the thyrotropin receptor gene and not in the Gsα protein gene in 31 toxic thyroid nodules. J Clin Endocrinol Metab, 82:3885–3891.

Führer

, Kubisch

, Scheibler

, Lamesch

, Krohn

, Paschke

. 1998. The extracellular thyrotropin receptor domain is not a major candidate for mutations in toxic thyroid nodules. Thyroid, 8:997–1001.

Lyons

, Landis

, Harsh

, Vallar

, Grunewald

, Feichtinger

, Duh

Q-Y

, Clark

, Kawasaki

, Bourne

. 1990. Two G protein oncogenes in human endocrine tumors. Science, 249:655–659.

O'Sullivan

, Barton

, Staddon

, Brown

, Lemoine

. 1991. Activating point mutations of the gsp oncogene in human thyroid adenomas. Mol Carcinog, 4:345–349.

Parma

, Duprez

, Van Sande

, Hermans

, Rocmans

, Van Vliet

, Costagliola

, Rodien

, Dumont

, Vassart

. 1997. Diversity and prevalence of somatic mutations in the thyrotropin receptor and Gsα genes as a cause of toxic thyroid adenomas. J Clin Endocrinol Metab, 82:2695–2701.

10.

Paschke

, Tonacchera

, Van Sande

, Parma

, Vassart

. 1994. Identification and functional characterization of two new somatic mutations causing constitutive activation of the thyrotropin receptor in hyperfunctioning autonomous adenomas of the thyroid. J Clin Endocrinol Metab, 79:1785–1789.

11.

Porcellini

, Ciullo

, Laviola

, Amabile

, Fenzi

, Avvedimento

. 1994. Novel mutations of thyrotropin receptor gene in thyroid hyperfunctioning adenomas. Rapid identification by fine needle aspiration biopsy. J Clin Endocrinol Metab, 79:657–661.

12.

Russo

, Arturi

, Suarez

, Schlumberger

, Du Villard

J-A

, Crocetti

, Filetti

. 1996. Thyrotropin receptor gene alterations in thyroid hyperfunctioning adenomas. J Clin Endocrinol Metab, 81:1548–1551.

13.

Russo

, Arturi

, Wicker

, Chazenbalk

, Schlumberger

, DuVillard

, Caillou

, Monier

, Rapoport

, Filetti

. 1995. Genetic alterations in thyroid hyperfunctioning adenomas. J Clin Endocrinol Metab, 80:1347–1351.

14.

Vanvooren

, Uchino

, Duprez

, Costa

, Vandekerckhove

, Parma

, Vassart

, Dumont

, Van Sande

, Noguchi

. 2002. Oncogenic mutations in the thyrotropin receptor of autonomously functioning thyroid nodules in the Japanese population. Eur J Endocrinol, 147:287–291.

15.

Georgopoulos

, Sykiotis

, Sgourou

, Papachatzopoulou

, Markou

, Kyriazopoulou

, Papavassiliou

, Vagenakis

. 2003. Autonomously functioning thyroid nodules in a former iodine-deficient area commonly harbor gain-of-function mutations in the thyrotropin signaling pathway. Eur J Endocrinol, 149:287–292.

16.

Garcia-Delgado

, Gonzalez-Navarro

, Napal

, Baldonado

, Vizmanos

, Gullon

. 1998. Higher sensitivity of denaturing gradient gel electrophoresis than sequencing in the detection of mutations in DNA from tumor samples. Biotechniques, 24:72, 74, 76.

17.

Trulzsch

, Krohn

, Wonerow

, Paschke

. 1999. DGGE is more sensitive for the detection of somatic point mutations than direct sequencing. Biotechniques, 27:266–268.

18.

Trulzsch

, Krohn

, Wonerow

, Chey

, Holzapfel

H-P

, Ackermann

, Fuhrer

, Paschke

. 2000. Detection of thyroid-stimulating hormone receptor and Gs mutations: in 75 toxic thyroid nodules by denaturing gradient gel electrophoresis. J Mol Med, 12:684–691.

19.

Eszlinger

, Niedziela

, Typlt

, Jaeschke

, Huth

, Schaarschmidt

, Aigner

, Trejster

, Krohn

, Bösenberg

, Paschke

. 2014. Somatic mutations in 33 benign and malignant hot thyroid nodules in children and adolescents. Mol Cell Endocrinol, 393:39–45.

20.

Calebiro

, Grassi

, Eszlinger

, Ronchi

, Godbole

, Bathon

, Guizzardi

, De Filippis

, Krohn

, Jaeschke

, Schwarzmayr

, Bircan

, Gozu

, Sancak

, Niedziela

, Strom

, Fassnacht

, Persani

, Paschke

. 2016. Recurrent EZH1 mutations are a second hit in autonomous thyroid adenomas. J Clin Investig, 126:3383–3388.

21.

Gozu

, Bircan

, Krohn

, Müller

, Vural

, Gezen

, Sargin

, Yavuzer

, Sargin

, Cirakoglu

, Paschke

. 2006. Similar prevalence of somatic TSH receptor and Gsα mutations in toxic thyroid nodules in geographical regions with different iodine supply in Turkey. Eur J Endocrinol, 155:535–545.

22.

Sancak

, Jaeschke

, Eren

, Tarcin

, Guellueoglu

, Sen

, Sever

, Gozu

, Bircan

, Akalin

, Paschke

, Eszlinger

. 2011. High prevalence of TSHR/Gsα mutation-negative clonal hot thyroid nodules (HNs) in a Turkish cohort. Horm Metab Res, 43:562–568.

23.

Robinson

, Thorvaldsdottir

, Winckler

, Guttman

, Lander

, Getz

, Mesirov

. 2011. Integrative genomics viewer. Nat Biotechnol, 29:24–26.

24.

Adzhubei

, Schmidt

, Peshkin

, Ramensky

, Gerasimova

, Bork

, Kondrashov

, Sunyaev

. 2010. A method and server for predicting damaging missense mutations. Nat Methods, 7:248–249.

25.

, Henikoff

. 2001. Predicting deleterious amino acid substitutions. Genome Res, 11:863–874.

26.

Tonacchera

, Pinchera

. 2000. Thyrotropin receptor polymorphisms and thyroid diseases. J Clin Endocrinol Metab, 85:2637–2639.

27.

Robinson

, Thorvaldsdóttir

, Winckler

, Guttman

, Lander

, Getz

, Mesirov

. 2011. Integrative genomics viewer. Nat Biotechnol, 29:24.

28.

Abramowicz

, Duprez

, Parma

, Vassart

, Heinrichs

. 1997. Familial congenital hypothyroidism due to inactivating mutation of the thyrotropin receptor causing profound hypoplasia of the thyroid gland. J Clin Investig, 99:3018–3024.

29.

Khoo

, Parma

, Rajasoorya

, Ho

, Vassart

. 1999. A germline mutation of the thyrotropin receptor gene associated with thyrotoxicosis and mitral valve prolapse in a Chinese family. J Clin Endocrinol Metab, 84:1459–1462.

30.

Nicoletti

, Bal

, De Marco

, Baldazzi

, Agretti

, Menabo

, Ballarini

, Cicognani

, Tonacchera

, Cassio

. 2009. Thyrotropin-stimulating hormone receptor gene analysis in pediatric patients with non-autoimmune subclinical hypothyroidism. J Clin Endocrinol Metab, 94:4187–4194.

31.

Cangul

, Morgan

, Forman

, Saglam

, Aycan

, Yakut

, Gulten

, Tarim

, Bober

, Cesur

. 2010. Novel TSHR mutations in consanguineous families with congenital nongoitrous hypothyroidism. Clin Endocrinol, 73:671–677.

32.

Lüblinghoff

, Nebel

, Huth

, Jäschke

, Schaarschmidt

, Eszlinger

, Paschke

. 2012. The leipzig thyrotropin receptor mutation database: update 2012. Eur Thyroid J, 1:209–210.

33.

Arseven

, Wilkes

, Jameson

, Kopp

. 2000. Substitutions of tyrosine 601 in the human thyrotropin receptor result in increase or loss of basal activation of the cyclic adenosine monophosphate pathway and disrupt coupling to Gq/11. Thyroid, 10:3–10.