DUOX2 variants are a frequent cause of congenital primary hypothyroidism in Thai patients.

OBJECTIVE: To identify the genetic etiologies of congenital primary hypothyroidism (CH) in Thai patients. DESIGN AND METHODS: CH patients were enrolled. Clinical characteristics including age, signs and symptoms of CH, pedigree, family history, screened thyroid-stimulating hormone results, thyroid function tests, thyroid imaging, clinical course and treatment of CH were collected. Clinical exome sequencing by next-generation sequencing was performed. In-house gene list which covered 62 potential candidate genes related to CH and thyroid disorders was developed for targeted sequencing. Sanger sequencing was performed to validate the candidate variants. Thyroid function tests were determined in the heterozygous parents who carried the same DUOX2 or DUOXA2 variants as their offsprings.
RESULTS: There were 118 patients (63 males) included. Mean (SD) age at enrollment was 12.4 (7.9) years. Forty-five of 118 patients (38%) had disease-causing variants. Of 45 variants, 7 genes were involved (DUOX2, DUOXA2, TG, TPO, SLC5A5, PAX8 and TSHR). DUOX2, a gene causing thyroid dyshormonogenesis, was the most common defective gene (25/45, 56%). The most common DUOX2 variant found in this study was c.1588A>T. TG and TPO variants were less common. Fourteen novel variants were found. Thyroid function tests of most parents with heterozygous state of DUOX2 and DUOXA2 variants were normal.
CONCLUSIONS: DUOX2 variants were most common among Thai CH patients, while TG and TPO variants were less common. The c.1588A>T in DUOX2 gene was highly frequent in this population.

Introduction

Congenital primary hypothyroidism (CH) is classified into thyroid dysgenesis (TD) and thyroid dyshormonogenesis (TDH) (1). TDH has increasingly been reported while the incidence of TD has remained stable (2, 3). Genetic studies have provided more information on the causes of CH (2, 4, 5, 6). To date, more than 20 disease-causing genes have been reported to be linked with the pathogenesis of CH (1, 7, 8, 9, 10). TD is defined as abnormal thyroid gland development including ectopic gland, hypoplasia and athyreosis. Genetic etiologies of TD include TSHR, NKX2-1, FOXE1, PAX8, NKX2-5, GLIS3, JAG1, TBX1, NTN1 and CDCA8 variants (1, 11). TDH is characterized by thyroid hormone biosynthetic defect. Genetic defects involved in the steps of thyroid hormone synthesis pathway include SLC5A5, SLC26A4, DUOX1, DUOX2, DUOXA1, DUOXA2, TPO, TG, IYD and GNAS genes (1, 11).

Identifying genetic causes of CH has several advantages for patients. Genetic diagnosis provides a risk estimation of thyroidal and extrathyroidal defects in affected patients and families and helps in predicting long-term prognosis in affected individuals (1). Owing to the fact that CH is a genetically heterogeneous disorder which is caused by variants of various genes, traditional sequencing of candidate genes of CH demonstrated pathogenic variants in only approximately 10% of the reported cases (12). Currently, next-generation sequencing (NGS) analysis has been reported to provide an efficient, cost-effective and multigenic screening tool to establish the genetic causes of CH with the diagnostic yield of 46–59% (4, 5, 6).

The incidence of CH has been increasing worldwide. Previous studies reported varied incidences of CH depending on race and ethnicity (13, 14, 15). The CH incidence was reported at 1:1200–2380 in Asians and 1:3533–11,000 in Caucasians (13, 14). TDH was found to be more frequent than TD in patients from China, Iran and United Arab Emirates (2, 16, 17). Genetic analysis revealed that TDH was more frequently associated with DUOX2 variants in patients of Asian origin, including Japan, Korea and China, and with TG and TPO variants in patients from United Kingdom and Finland (2, 5, 6, 18, 19). This study aimed to investigate the clinical and molecular characteristics of Thai patients with CH.

Materials and methods

Patients

All enrolled CH patients were regularly treated at the Departments of Pediatrics and Medicine, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand. CH was diagnosed based on the findings of elevated serum thyroid-stimulating hormone (TSH) and low free thyroxine (FT4) concentrations on either confirmatory test for positive newborn screening (NBS) or thyroid function tests for other signs and symptoms suggesting CH. Patients with transient CH secondary to maternal conditions, sick euthyroid syndrome and obvious syndromic features were excluded.

Provisional clinical diagnoses of TDH and TD were made in patients who had clinical features, and possibly thyroid scintigraphic or ultrasonographic findings suggestive of the particular diagnoses. Patients with goiter, or normal eutopic or enlarged thyroid gland on the thyroid imaging were classified as having TDH while patients who had absent or small or ectopic thyroid gland on thyroid imaging were considered to have TD. Patients who were not compatible with the two groups were classified as having undetermined cause. Patients with persistently high TSH after levothyroxine (LT4) discontinuation after 3 years of age were diagnosed as having permanent CH. ‘Transient’ CH was diagnosed based on having normal thyroid function test results following discontinuation of LT4 therapy after 3 years of age and thereafter.

The study was approved by the Ethics Committee on Human Research of the Faculty of Medicine Ramathibodi Hospital, Mahidol University (MURA 2018/844, dated 6 December 2018). The study conformed with the Declaration of Helsinki. Written informed consent was obtained from the patients or their legal guardians.

Clinical data collection

Clinical characteristics including age, signs and symptoms of CH, pedigree, family history of CH, TSH screening results, thyroid function tests, thyroid imagings, clinical course and treatment of CH were collected.

DNA extraction and targeted sequencing of candidate genes

Genomic DNA was extracted from peripheral blood using the QuickGene DNA Whole Blood Kit L (Kurabo, Japan). DNA of the patients was submitted for clinical exome sequencing (CES). CES by NGS was performed by Illumina MiSeq® system (Illumina, USA) using the TruSight One Sequencing Panel®. The TruSight One Sequencing Panel® focused on 4811 known disease-causing genes that have been reported to be associated with human diseases. Sequences were aligned with the human reference genome version hg19. Thyroid disorder gene list including genes related to CH, secondary hypothyroidism, thyroid hormone resistance, thyroid hormone metabolism defects and thyroid test abnormalities without thyroid pathology (such as ALB and SERPINA7) was developed in-house. It covered 62 potential candidate genes (Table 1) which are known to be related to thyroid disorders according to the previous reports (7, 8, 9, 10). Some genes related to syndromic CH were included to detect genetic variants in patients who might not have recognizable features. Of the 62 genes, there were 16 genes that are related to TD and TDH.

Table 1

Sixty-two genes that are related to thyroid disorders and covered by the panel used in this study.

Classification	Genes	OMIM number	Phenotypes	Inheritance
Thyroid dysgenesis	NKX2-1	600635	Choreoathetosis, hypothyroidism and neonatal respiratory distress	AD
	FOXE1	602617	Bamforth-Lazarus syndrome	AR
	PAX8	167415	Thyroid dysgenesis or hypoplasia	AD
	NKX2-5	600584	Congenital nongoitrous hypothyroidism	AD
	GLIS3	610192	Neonatal diabetes mellitus with congenital hypothyroidism	AR
	TSHR	603372	Congenital nongoitrous hypothyroidism	AR, AD
	JAG1	601920	Alagille syndrome	AD
	TBX1	602054	DiGeorge syndrome	AD
Thyroid dyshormonogenesis	SLC5A5	601843	Thyroid dyshormonogenesis	AR
	TPO	606765	Thyroid dyshormonogenesis	AR
	SLC26A4	605646	Pendred syndrome	AR
	TG	188450	Thyroid dyshormonogenesis	AR
	IYD	612025	Thyroid dyshormonogenesis	AR
	DUOX2	606759	Thyroid dyshormonogenesis	AR, AD
	DUOXA2	612772	Thyroid dyshormonogenesis	AR, AD
	GNAS	139320	Pseudohypoparathyroidism	AD
Central hypothyroidism	TSHB	188540	Congenital nongoitrous hypothyroidism	AR
	TRHR	188545	Congenital nongoitrous hypothyroidism	AR
	TBL1X	300196	Congenital nongoitrous hypothyroidism	XLR
	HESX1	601802	Combined pituitary hormone deficiencies	AD, AR
	LHX3	600577	Combined pituitary hormone deficiencies	AR
	LHX4	602146	Combined pituitary hormone deficiencies	AD
	SOX3	313430	Panhypopituitarism	XLR
	OTX2	600037	Combined pituitary hormone deficiencies	AD
	POU1F1	173110	Combined pituitary hormone deficiencies	AD, AR
	PROP1	601538	Combined pituitary hormone deficiencies	AR
	IRS4	300904	Congenital nongoitrous hypothyroidism	XLR
Thyroid hormone resistance and abnormal thyroid hormone metabolism	THRBTHRASLC16A2SECISBP2	190160190120300095607693	Thyroid hormone resistanceCongenital nongoitrous hypothyroidismAllan-Herndon-Dudley syndromeAbnormal thyroid hormone metabolism	AD, ARADXLRAR
Syndromes or transcription factors which may be associated with congenital hypothyroidism	SALL1UBR1DYRK1AELNKDM6AKMT2D	602218605981600855130160300128602113	Townes-Brocks syndromeJohanson-Blizzard syndromeMental retardationSupravalvular aortic stenosisKabuki syndrome Kabuki syndrome	ADARADADXLDAD
	KAT6B	605880	Genitopatellar syndrome and Say-Barber-Biesecker-Young-Simpson syndrome	AD
	ALB	103600	Dysalbuminemic hyperthyroxinemia	AD
	ALMS1	606844	Alstrom syndrome	AR
	DIO1	147892	Asymptomatic hyperthyroxinemia	AD
	DIO2	601413	Asymptomatic hyperthyroxinemia	ND
	FGF8	600483	Hypogonadotropic hypogonadism with or without anosmia	AD
	HHEX	604420	Thyroid dysgenesis	ND
	NKX2-3	606727	Thyroid dysgenesis	ND
	NKX2-6	611770	Conotruncal heart malformations and persistent truncus arteriosus	AR
	PTH1R	168468	Pseudohypoparathyroidism	ND
	PTRH2	608625	Infantile-onset multisystem neurologic, endocrine, and pancreatic disease	AR
	RYR2	180902	Hyperemesis gravidarum	ND
	SERPINA7	314200	Thyroxine binding globulin deficiency	XLR
	SLC30A10	611146	Hypermanganesemia with dystonia	AR
	SLCO1C1(OATP1C1)	613389	Thyroid hormone transporter deficiency	AR
	TTR	176300	Dystransthyretinemic hyperthyroxinemia	AD
	MC2R	607397	Glucocorticoid deficiency due to ACTH unresponsiveness	AR
	MRAP	609196	Glucocorticoid deficiency	AR
	PDE4D	600129	Acrodysostosis with or without hormone resistance	AD
	PRKAR1A	188830	Acrodysostosis with or without hormone resistance	AD
	TBC1D24	613577	Deafness, onychodystrophy, osteodystrophy, mental retardation and seizures (DOORS) syndrome	AR
	TRAPPC9	611966	Mental retardation	AR
	TXNRD2	606448	Glucocorticoid deficiency	AR
	FOXI1	601093	Enlarged vestibular aqueduct	AR
	KCNJ10	602208	Enlarged vestibular aqueduct	AR

ACTH, adrenocorticotropic hormone; AD, autosomal dominant; AR, autosomal recessive; OMIM, online Mendelian inheritance in men; ND, no data; XLD, X-linked dominant; XLR, X-linked recessive.

The variant annotation was performed with VarSeq® Software version 2.1.1 (Golden Helix, USA). Candidate variants were filtered based on in-house developed thyroid disorder gene list and minor allele frequency (MAF) of less than 0.05 across the online databases (e.g. gnomAD, 1000 Genomes, ExAC, dbSNP and ClinVar) and in-house Thai database (455 persons). Using the American College of Medical Genetics and Genomics (ACMG) 2015 variant classification guidelines together with Varsome® software (Saphetor, Switzerland), the clinical interpretation of selected variants was determined (20, 21). Computational and prediction data using in silico tools were done as one of the ACMG criteria. Variants that were classified as pathogenic or likely pathogenic were considered to be definite causes of CH in the patients. Variants that did not meet the criteria of pathogenic, likely pathogenic, benign or likely benign, would be classified as variant of uncertain significance (VUS). Sanger sequencing was performed to validate the candidate variants in all patients and their parents. In index cases who had siblings with CH, their CH siblings were analyzed for the same variants by Sanger sequencing. Thyroid function tests including FT4, TSH and thyroglobulin (Tg) concentrations were determined in the heterozygous parents who carried the same DUOX2 or DUOXA2 variants as their offsprings. Genotype and phenotype correlation of CH was analyzed.

Statistical analysis

Data were analyzed using SPSS version 22.0 (IBM Corp). Normally and non-normally distributed data were expressed as mean and s.d., and median and interquartile range (IQR), respectively. Mann–Whitney U test was used for comparison between two groups of non-normally distributed data. A P-value of less than 0.05 was considered statistically significant.

Results

A total of 120 Thai patients with CH were enrolled. Two patients with syndromic features were excluded. Therefore, 118 patients from 109 families were included in the analysis. Eighteen patients were siblings in 9 families. There was no history of consanguinity. There were 55 females and 63 males. Mean (s.d.) age at enrollment was 12.4 (7.9) years. Of the 118 patients, 41 (35%), 22 (19%) and 55 (46%) patients were clinically classified as having TDH, TD and undetermined cause, respectively. Ninety-one patients (77%) were identified through positive NBS. The remaining 27 patients presented with hypothyroid-related symptoms (21 patients), ectopic thyroid gland (5 patients) and non-autoimmune thyroid goiter (1 patient). There were 92 and 11 patients with permanent and transient CH, respectively. The remaining 15 patients were less than 3 years of age at the time of enrollment, therefore their permanence awaited to be determined.

CES analysis revealed seven CH-causing genes in 39 out of 109 families (45 out of 118 patients, 38%). Thirty-six out of 45 patients (80%) had variants in the genes related to TDH, including DUOX2 (n = 25), DUOXA2 (n = 6), TG (n = 2), TPO (n = 2) and SLC5A5 (n = 1); and the remaining 9 patients (20%) had variants in the genes related to TD, including TSHR (n = 5) and PAX8 (n = 4). There were 14 novel pathogenic variants, including 4 DUOX2 variants, 2 DUOXA2 variants, 2 TG variants, 1 SLC5A5 variant, 3 PAX8 variants and 2 TSHR variants (Table 2). There were no pathogenic or likely pathogenic variants in SLC26A4, IYD, GNAS, NKX2-1, FOXE1, NKX2-5, GLIS3, TBX1 and JAG1 genes. VUS were demonstrated in 8 additional patients among the 118 patients (7%). Among these 8 patients, there were 2 patients who had heterozygous VUS; one had DUOX2 variant (c.2830G>A) and the other had DUOXA2 variant (c.122T>C) which might be responsible for their CH phenotype. VUS were not included in the reported positive variants.

Table 2

Details of pathogenic and likely pathogenic variants of seven genes identified in the study.

Genes	Nucleotide position	Amino acid position	Mutation types	SIFT	Polyphen-2	Allele frequency gnomAD	Thai allele frequency (n= 455)	Number of alleles	Status (accession number)	RS number
DUOX2 (NM_014080.4)	c.1588A>T	p.Lys530Ter	Nonsense	NA	NA	0.000675966	0.00989	10	Reported	rs180671269
	c.2654G>A	p.Arg885Gln	Missense	0.006Deleterious	0.999Damaging	0.000115324	0.0010989	4	Reported	rs181461079
	c.2048G>T	p.Arg683Leu	Missense	0.002Deleterious	1.000Damaging	0.000342275	–	3	Reported	rs8028305
	c.2104_2106delGGA	p.Gly702del	In-frame deletion	NA	NA	0.000075580	0.0010989	3	Reported	rs779340990
	c.2654G>T	p.Arg885Leu	Missense	0.003Deleterious	0.999Damaging	0.000405621	–	3	Reported	rs181461079
	c.4027C>T	p.Leu1343Phe	Missense	0.054Tolerated	0.831Damaging	0.000592662	0.0054945	3	Reported	rs147945181
	c.1304A>G	p.Asp435Gly	Missense	0.000Deleterious	1.000Damaging	0.000031812	–	2	Reported	rs772040742
	c.1310G>C	p.Gly437Ala	Missense	0.000Deleterious	1.000Damaging	0.000123275	0.0010989	2	Novel (SCV001250672)	rs769796932
	c.2101C>T	p.Arg701Ter	Nonsense	NA	NA	0.000031826	0.0032967	2	Reported	rs201109959
	c.3693+1G>T	–	Splice site	NA	NA	0.000103437	0.0032967	2	Reported	rs200717240
	c.989T>G	p.Val330Gly	Missense	0.226Tolerateda	0.016Benigna	–	–	1	Novel (SCV001250732)	–
	c.1232G>A	p.Arg411Lys	Missense	0.033Deleterious	0.372Benign	0.000059650	–	1	Reported	rs764353021
	c.1295G>A	p.Arg432His	Missense	0.038Deleterious	0.933Damaging	0.000067603	–	1	Reported	rs530736554
	c.2635G>A	p.Glu879Lys	Missense	0.000Deleterious	1.000Damaging	0.000075555	0.0010989	1	Reported	rs774556391
	c.2895_2898delGTTC	p.Phe966Serfs*29	Frameshift	NA	NA	0.00293655	–	1	Reported	rs530719719
	c.3115C>T	p.Arg1039Trp	Missense	0.000Deleterious	1.000Damaging	0.000011936	–	1	Novel (SCV001245530)	rs752176935
	c.3329G>A	p.Arg1110Gln	Missense	0.003Deleterious	0.994Damaging	0.000194847	0.0010989	1	Reported	rs368488511
	c.3340delC	p.Leu1114Serfs*56	Frameshift	NA	NA	0.000015905	0.0010989	1	Reported	rs748194265
	c.3478_3480delCTG	p.Leu1160del	In-frame deletion	NA	NA	0.000027845	–	1	Reported	rs758318135
	c.3631C>T	p.Arg1211Cys	Missense	0.000Deleterious	1.000Damaging	0.000043753	–	1	Reported	rs374410986
	c.4080G>T	p.Lys1360Asn	Missense	0.068Toleratedb	0.379Benignb	0.000003992	–	1	Novel (SCV001250737)	rs374891282
	c.4408C>T	p.Arg1470Trp	Missense	0.000Deleterious	1.000Damaging	0.000159584	0.0010989	1	Reported	rs200785525
DUOXA2 (NM_207581.3)	c.738C>G	p.Tyr246Ter	Nonsense	NA	NA	0.000143084	0.0021978	5	Reported	rs4774518
	c.604G>A	p.Ala202Thr	Missense	0.016Deleterious	0.643Possibly damaging	0.000016892	–	2	Novel (SCV001250673)	rs770148072
	c.232G>A	p.Val78Met	Missense	0.076Tolerated	1.000Damaging	0.000044113	–	1	Reported	rs746132852
	c.501C>A	p.Cys167Ter	Nonsense	NA	NA	0.000004014	–	1	Novel (SCV001250908)	rs781126484
TG (NM_003235.4)	c.48G>A	p.Trp16Ter	Nonsense	NA	NA	0.000004367	–	1	Novel (SCV001250735)	rs780846892
	c.274+2T>G	–	Splice site	NA	NA	0.000003991	0.0010989	1	Reported	rs1398373161
	c.1348delT	p.Ser450Profs*29	Frameshift	NA	NA	0.000055760	–	1	Reported	rs776553164
	c.6791G>A	p.Cys2264Tyr	Missense	0.001Deleterious	1.000Damaging	0.000011931	–	1	Novel (SCV001250736)	rs1229345000
TPO (NM_000547.5)	c.670_672delGAC	p.Asp224del	In-frame deletion	NA	NA	0.000059679	–	2	Reported	rs772164623
	c.2422delT	p.Cys808Alafs*24	Frameshift	NA	NA	0.000083532	–	2	Reported	rs763662774
SLC5A5 (NM_000453.2)	c.794A>G	p.Gln265Arg	Missense	0.008Deleterious	0.999Damaging	–	–	2	Novel (SCV001245529)	–
PAX8 (NM_003466.3)	c.92G>A	p.Arg31His	Missense	0.000Deleterious	1.000Damaging	–	–	1	Reported	rs104893657
	c.203C>T	p.Thr68Ile	Missense	0.000Deleterious	1.000Damaging	–	–	1	Novel (SCV001245528)	–
	c.236C>T	p.Ser79Phe	Missense	0.000Deleterious	1.000Damaging	–	–	1	Novel (SCV001250734)	–
	c.457_458delCT	p.Leu153Glufs*47	Frameshift	NA	NA	–	–	1	Novel (SCV001250738)	–
TSHR (NM_000369.2)	c.1960A>T	p.Ile654Phe	Missense	0.000Deleterious	1.000Damaging	0.000011929	0.0021978	4	Novel (SCV001250733)	rs767239688
	c.545+5G>T	–	Splice site	NA	NA	–	–	2	Novel (SCV001250739)	–
	c.1825C>T	p.Arg609Ter	Nonsense	NA	NA	0.000003978	–	2	Reported	rs763679435
	c.1349G>A	p.Arg450His	Missense	0.000Deleterious	1.000Damaging	0.000234637	0.0021978	1	Reported	rs189261858

–Absent in database; ac.989T>G variant was predicted to be probably deleterious (0.960) by Mutation Taster; bc.4080G>T variant was predicted to be probably deleterious (0.999) by Mutation Taster.

gnomAD, Genome Aggregation Database (version 2.1.1); NA, not available; Polyphen-2, Polymorphism Phenotypic version 2 (used to predict the effects of missense mutations); RS number, reference single nucleotide variants number; SIFT, Sorting Intolerant from Tolerant.

Clinical characteristics and details of patients with genetic variants are summarized in Table 3. All pathogenic and likely pathogenic variants are shown in Table 2.

Table 3

Clinical characteristics and details of patients with genetic variants (n = 45).

Family	Age at enrollment (years)	Sex	At diagnosis			Thyroid scintigraphy/USG	Transient or permanent	Genetic variant information
			Age (years)	FT4 (ng/dL)	TSH (mU/L)			Genes	Variants		Zygosity
1	17.8	F	NBS	1.9	21.5	Eutopic	Permanent	DUOX2	c.2048G>T (p.Arg683Leu), c.4027C>T (p.Leu1343Phe)	c.2635G>A (p.Glu879Lys)	ComHet
2	3.1	F	NBS	0.3	>100.0	Eutopic	Permanent	DUOX2	c.2654G>T (p.Arg885Leu)	c.2895_2898delGTTC (p.Phe966Serfs*29)	ComHet
3	11.7	M	NBS	1.0	>50.0	NA	Permanent	DUOX2	c.1310G>C (p.Gly437Ala)	c.3115C>T (p.Arg1039Trp)	ComHet
4	17.0	M	NBS	0.7	>100.0	Eutopic	Permanent	DUOX2	c.1588A>T (p.Lys530Ter)	c.3631C>T (p.Arg1211Cys)	ComHet
5	16.8	M	NBS	T4 4.5 µg/dL (N, 6-15)	>50.0	Eutopic	Permanent	DUOX2	c.1304A>G (p.Asp435Gly)	c.1588A>T (p.Lys530Ter)	ComHet
6	17.4	M	NBS	T4 2.8 µg/dL (N, 6-15)	48.2	Eutopic	Permanent	DUOX2	c.1588A>T (p.Lys530Ter)	c.2101C>T (p.Arg701Ter)	ComHet
7	12.7	F	NBS	0.3	>100.0	NA	Permanent	DUOX2	c.2654G>T (p.Arg885Leu)	c.3329G>A (p.Arg1110Gln)	ComHet
8	18.6	F	NBS	0.5	>100.0	NA	Permanent	DUOX2	c.1310G>C (p.Gly437Ala)	c.3478_3480delCTG (p.Leu1160del)	ComHet
9	17.5 (1st twin)	M	NBS	0.9	31.0	NA	Permanent	DUOX2	c.1588A>T (p.Lys530Ter)	c.2654G>A (p.Arg885Gln)	ComHet
	17.5 (2nd twin)	M	NBS	0.5	41.0	NA	Permanent	DUOX2	c.1588A>T (p.Lys530Ter)	c.2654G>A (p.Arg885Gln)	ComHet
10	11.1	F	NBS	NA	NA	NA	Permanent	DUOX2	c.1588A>T (p.Lys530Ter)	c.1588A>T (p.Lys530Ter)	Hom
11	6.2	F	NBS	0.4	>100.0	Eutopic	Permanent	DUOX2	c.3693+1G>T	WT	Het
12	3.1	M	NBS	0.9	60.8	Eutopic	Permanent	DUOX2	c.3340delC (p.Leu1114Serfs*56)	WT	Het
13	11.7	M	NBS	1.2	10.3	NA	Permanent	DUOX2	c.1295G>A (p.Arg432His)	WT	Het
14	20.8	M	NBS	0.9	>50.0	Eutopic	Permanent	DUOX2	c.2048G>T (p.Arg683Leu), c.4027C>T (p.Leu1343Phe)	WT	Het
15	17.2	M	NBS	1.5	17.7	NA	Permanent	DUOX2	c.4408C>T (p.Arg1470Trp)	WT	Het
16	5.0	M	NBS	0.4	>100.0	NA	Transient	DUOX2	c.2048G>T (p.Arg683Leu), c.4027C>T (p.Leu1343Phe)	c.2654G>A (p.Arg885Gln)	ComHet
17	13.4	M	NBS	T4 1.5 µg/dL (N, 6-15)	>50.0	Eutopic	Transient	DUOX2	c.2654G>A (p.Arg885Gln)	c.3693+1G>T	ComHet
18	6.3	F	NBS	1.4	7.1	Eutopic	Transient	DUOX2	c.1304A>G (p.Asp435Gly)	c.4080G>T (p.Lys1360Asn)	ComHet
19	4.2	M	NBS	0.5	>100.0	Eutopic	Transient	DUOX2	c.1232G>A (p.Arg411Lys)	WT	Het
20	8.4	F	NBS	T4 4.1 µg/dL (N, 6-15)	45.1	Eutopic	Transient	DUOX2	c.2101C>T (p.Arg701Ter)	WT	Het
21	1.1 (1st twin)	M	NBS	0.9	38.9	Eutopic	Unknowne	DUOX2	c.1588A>T (p.Lys530Ter)	c.2104_2106delGGA (p.Gly702del)	ComHet
	1.1 (2nd twin)	M	NBS	0.8	92.7	Eutopic	Unknowne	DUOX2	c.1588A>T (p.Lys530Ter)	c.2104_2106delGGA (p.Gly702del)	ComHet
22	0.1	F	NBS	0.5	>100.0	NA	Unknowne	DUOX2	c.1588A>T (p.Lys530Ter)	c.2654G>T (p.Arg885Leu)	ComHet
23	0.4	F	0.2a	0.6	>100.0	NA	Unknowne	DUOX2	c.989T>G (p.Val330Gly)	c.2104_2106delGGA (p.Gly702del)	ComHet
24	8.4	F	NBS	T4 2.7 µg/dL (N, 6-15)	>50.0	NA	Permanent	DUOXA2	c.738C>G (p.Tyr246Ter)	c.738C>G (p.Tyr246Ter)	Hom
25	7.3	M	NBS	0.4	>100.0	NA	Permanent	DUOXA2	c.232G>A (p.Val78Met)	WT	Het
26	28.1	F	5b	NA	NA	Ectopic	Permanent	DUOXA2	c.738C>G (p.Tyr246Ter)	WT	Het
27	5.5	F	NBS	0.5	>100.0	Eutopic	Transient	DUOXA2	c.604G>A (p.Ala202Thr)	c.738C>G (p.Tyr246Ter)	ComHet
	2.2	M	0.1a	0.6	>100.0	NA	Unknowne	DUOXA2	c.604G>A (p.Ala202Thr)	c.738C>G (p.Tyr246Ter)	ComHet
28	5.6	M	0.1a	0.1	>100.0	Eutopic	Transient	DUOXA2	c.501C>A (p.Cys167Ter)	WT	Het
29	6.2	M	NBS	0.7	>100.0	NA	Permanent	TG	c.274+2T>G	c.1348delT (p.Ser450Profs*29)	ComHet
30	14.6	M	NBS	0.7	>100.0	NA	Permanent	TG	c.48G>A (p.Trp16Ter)	c.6791G>A (p.Cys2264Tyr)	ComHet
31	25.5	M	8.6c	NA	NA	Eutopic	Permanent	TPO	c.670_672delGAC (p.Asp224del)	c.2422delT (p.Cys808Alafs*24)	ComHet
	24.0	F	6.7c	NA	NA	Eutopic	Permanent	TPO	c.670_672delGAC (p.Asp224del)	c.2422delT (p.Cys808Alafs*24)	ComHet
32	11.2	M	NBS	0.2	>100.0	NA	Permanent	SLC5A5	c.794A>G (p.Gln265Arg)	c.794A>G (p.Gln265Arg)	Hom
33	31.2	M	NBS	T4 2 µg/dL (N, 6-15)	>100.0	NA	Permanent	TSHR	c.545+5G>T	c.1825C>T (p.Arg609Ter)	ComHet
	24.4	M	NBS	NA	NA	NA	Permanent	TSHR	c.545+5G>T	c.1825C>T (p.Arg609Ter)	ComHet
34	6.4 (1st twin)	F	NBS	0.4	>100.0	Athyreosis	Permanent	TSHR	c.1960A>T (p.Ile654Phe)	c.1960A>T (p.Ile654Phe)	Hom
	6.4 (2nd twin)	F	NBS	0.6	>100.0	Athyreosis	Permanent	TSHR	c.1960A>T (p.Ile654Phe)	c.1960A>T (p.Ile654Phe)	Hom
35	9.4	M	NBS	1.1	6.6	NA	Permanent	TSHR	c.1349G>A (p.Arg450His)	WT	Het
36	10.8	F	NBS	1.7	97.5	Athyreosis	Permanent	PAX8	c.203C>T (p.Thr68Ile)	WT	Het
37	22.5	M	19.1d	0.6	>100.0	NA	Permanent	PAX8	c.92G>A (p.Arg31His)	WT	Het
38	22.8	F	5.4d	0.3	>100.0	Hypoplasia	Permanent	PAX8	c.236C>T (p.Ser79Phe)	WT	Het
39	9.5	M	NBS	T4 8.1 µg/dL (N, 6-15)	7.7	Absent uptake, but present thyroid on USG	Permanent	PAX8	c.457_458delCT (p.Leu153Glufs*47)	WT	Het

Normal range for FT4 (ng/dL): neonates age 0–2 weeks 0.9–5.0; infants 0.8–2.1; children and adults 0.7–1.4. Normal range for TSH (mU/L): neonates age 4–7 days 1.3–16.0; infants 0.9–7.1; children and adults 0.6–4.5. To convert FT4 in ng/dL to pmol/L, multiply by 12.9; T4 in µg/dL to nmol/L, multiply by 12.9 and TSH in mU/L to µIU/mL multiply by 1.0.

aPresented with prolonged jaundice; bPresented with ectopic thyroid; cPresented with short stature and goiter; dPresented with short stature; eLess than 3 years of age, permanence awaited to be determined.

ComHet, compound heterozygous; F, female; FT4, free thyroxine; Het, heterozygous; Hom, homozygous; M, male; N, normal range; NA, not available; NBS, newborn screening; T4, thyroxine; TSH, thyroid-stimulating hormone; USG, ultrasonography; WT, wild type.

Variants of genes related to TDH

DUOX2 variants were the most frequent cause of TDH. Twenty-two different DUOX2 variants were identified in 25 patients (23 families). Eighteen out of 25 patients (72%) carried either compound heterozygous or homozygous variants; and the remaining 7 patients (28%) had heterozygous variants. The most common pathogenic DUOX2 variant was c.1588A>T, in 10 alleles in 9 patients. While this variant is rare in overall population with MAF of 0.0007 from gnomAD database, it is relatively common in Thai population with MAF of approximately 0.01 in 455 ethnic-matched normal control subjects from our in-house Thai database. Four different DUOXA2 variants were identified in 6 patients (5 families), of which three of them had either compound heterozygous or homozygous variants; and the other three had heterozygous variants. The most common DUOXA2 variant was c.738C>G, in 5 alleles in 4 patients. Five patients with DUOX2 variants and 2 patients with DUOXA2 variants had transient CH and 16 patients with DUOX2 variants and 3 patients with DUOXA2 variants had permanent CH. The remaining 4 patients with DUOX2 variants and 1 patient with DUOXA2 variant were less than 3 years of age at the time of enrollment, so their permanence awaited to be determined. Hypothyroidism in 27 out of 31 patients (87%) with DUOX2 and DUOXA2 variants was detected by NBS while 3 patients had negative NBS results and prolonged jaundice was the presentation of hypothyroidism. The remaining 1 patient who had DUOXA2 variant presented with enlargement of an ectopic thyroid gland at 5 years of age.

SLC5A5 variant was identified in 1 patient. At 12 years of age following LT4 therapy discontinuation, his thyroid scintigraphy showed no radiotracer uptake but ultrasonography showed normal thyroid gland. TPO variants were detected in 2 patients from the same family. The older brother presented with short stature and diffuse goiter at 8.6 years of age and his sister presented with short stature and multinodular goiter at 6.7 years of age. Additionally, TG variants were found in 2 patients.

Variants of genes related to TD

The majority of variants of the genes related to TD were found in TSHR gene. Four TSHR variants in 5 patients were detected. Of these 5 patients, 4 had either homozygous or compound heterozygous variants and one patient with subclinical hypothyroidism had heterozygous variant. Four patients with PAX8 variants had varied thyroid phenotypes, including athyreosis, hypoplasia and gland in situ, but absent uptake on thyroid scintigraphy. Two patients presented with short stature during childhood and adolescence.

Genotype-phenotype analysis of patients with variants

Among 18 patients with biallelic DUOX2 variants, 11 (61%) had permanent CH, 3 (17%) had transient CH and the remaining 4 (22%) were under 3 years of age, whose permanence awaited to be determined. Out of 7 patients with monoallelic DUOX2 variants, 5 had permanent CH and 2 had transient CH. Median (IQR) serum TSH and FT4 concentrations at diagnosis of patients with monoallelic and biallelic variants were not statistically different [TSH: 50.0 (17.7, 100.0) and 50.0 (39.5, 100.0) mU/L, p=0.604; FT4: 0.9 (0.4, 1.3) and 0.6 (0.4, 0.9) ng/dL, p= 0.482, respectively]. There was no evidence of genotype-phenotype correlation.

Segregation analysis of patients with and variants

Serum FT4, TSH and Tg concentrations were determined in 29 heterozygous parents from 17 families of patients who carried variants of the DUOX2 and DUOXA2 genes (Fig. 1 and Table 4). Regarding patients with compound heterozygous and homozygous variants in the DUOX2 and DUOXA2 genes from 12 families which were inherited as an autosomal recessive manner, 22 heterozygous parents had normal FT4, TSH and Tg concentrations, while 2 parents (Families 6 and 16) had mildly elevated Tg concentrations, but normal FT4 and TSH (Fig. 1 and Table 4).

Figure 1

Pedigree of patients with DUOX2 and DUOXA2 variants CH, congenital primary hypothyroidism; WT, wild type; *, no DNA available.

Table 4

Thyroid function tests of the parents of the patients with DUOX2 and DUOXA2 variants.

Family	Member	FT4 (ng/dL)	TSH (mU/L)	Tg (ng/mL)
1	Father	1.0	0.8	5.6
	Mother	1.0	2.0	18.3
2	Father	ND	ND	ND
	Mother	ND	ND	ND
3	Father	0.8	1.8	9.2
	Mother	0.8	3.7	3.7
4	Father	0.9	1.6	16.7
	Mother	1.0	1.9	14.6
5	Father	0.9	1.3	14.3
	Mother	1.1	1.0	4.0
6	Father	1.1	1.7	81.2
	Mother	0.9	0.9	9.3
7	Father	ND	ND	ND
	Mother	ND	ND	ND
8	Father	0.9	0.5	5.8
	Mother	0.9	1.0	5.1
9	Father	ND	ND	ND
	Mother	ND	ND	ND
10	Father	ND	ND	ND
	Mother	ND	ND	ND
11	Father	ND	ND	ND
	Mother	ND	ND	ND
12	Father	ND	ND	ND
	Mother	ND	ND	ND
13	Father	1.0	0.5	11.7
	Mother	ND	ND	ND
14	Father	ND	ND	ND
	Mother	0.8	10.5	3.7
15	Father	ND	ND	ND
	Mother	ND	ND	ND
16	Father	1.2	1.3	8.5
	Mother	1.0	1.1	83.1
17	Father	0.9	2.7	4.6
	Mother	0.9	0.6	3.6
18	Father	ND	ND	ND
	Mother	ND	ND	ND
19	Father	ND	ND	ND
	Mother	1.0	0.6	7.3
20	Father	ND	ND	ND
	Mother	ND	ND	ND
21	Father	1.1	1.5	13.9
	Mother	0.9	2.0	7.2
22	Father	1.0	1.4	4.2
	Mother	0.8	2.5	12.0
23	Father	1.0	0.9	12.1
	Mother	1.0	1.3	38.0
24	Father	ND	ND	ND
	Mother	ND	ND	ND
25	Father	ND	ND	ND
	Mother	ND	ND	ND
26	Father	1.3	0.7	34.7
	Mother	ND	ND	ND
27	Father	0.8	0.7	7.3
	Mother	0.9	2.8	26.8
28	Father	ND	ND	ND
	Mother	T4 7 µg/dL (N, 4-13)	1.7	ND

FT4, free thyroxine; T4, thyroxine; TSH, thyroid-stimulating hormone; Tg, thyroglobulin; ND, not done.

Adult normal ranges for FT4 0–7-1.4 ng/dL, TSH 0.6–4.5 mU/L, Tg 3.5–77.0 ng/mL. To convert FT4 in ng/dL to pmol/L, multiply by 12.9; TSH in mU/L to µIU/mL multiply by 1.0 and Tg in ng/mL to µg/L multiply by 1.0.

Some heterozygous variants of the DUOX2 and DUOXA2 genes have been described as an autosomal dominant inheritance. Four out of five parents who were tested and carried the same heterozygous variants as their offsprings had normal FT4, TSH and Tg concentrations. Only the mother of a patient with DUOX2 defect who carried two variants in the same allele (c.2048G>T and c.4027C>T) had subclinical hypothyroidism which was subsequently found to be related to autoimmune thyroiditis (Fig. 1, family 14 and Table 4).

Discussion

This study demonstrated that the frequency of genetic defects in the genes causing TDH was more common than that of the genes causing TD (36/118 (30%) vs 9/118 (8%)) which was in agreement with the previous studies (4, 5, 22, 23, 24). The most frequently affected gene in this study was DUOX2 (25 out of 45, 56%). This finding is consistent with the frequency reported in other Asian countries (Korea, Japan and China) at 53–74% (4, 18, 19). In contrast, TG and TPO variants were demonstrated in 4 out of 45 patients (9%) which was much less than that of DUOX2 variants. TG and TPO variants have been reported as the most frequent cause of TDH in Western populations (5, 6). The high rate of DUOX2 variants in Asians could be explained by the founder effect which contributed to more frequent occurrence of the particular variants compared with other populations. MAF of normal control Thai database of 11 out of 22 DUOX2 variants identified in this study was greater than that of the general population from the gnomAD (0.001–0.01 vs 0.00002–0.0007) (Table 2).

DUOX2 requires DUOX1 and their maturation factors (DUOXA1 and DUOXA2) to maintain normal hydrogen peroxide (H2O2) production (1, 25). Twenty-two different DUOX2 variants (Table 2) were identified in this cohort. The c.1588A>T in DUOX2 gene was highly recurrent in 9 out of 25 patients (36%) with DUOX2 variants in our cohort. The c.1588A>T variant had population-specificity and was mainly reported from Asian countries (26, 27, 28). Interestingly, among these 9 patients who carried c.1588A>T in both compound heterozygous and homozygous patterns, 6 of them had permanent CH and the remaining 3 were less than 3 years of age whose permanence awaited to be determined. Therefore, most patients with c.1588A>T variant in this study had permanent CH. However, previous studies demonstrated that the clinical phenotype of patients carrying c.1588A>T in each different genotype (biallelic and monoallelic variants) had both transient and permanent CH (27, 29). The difference in the phenotype of patients who had the same variants among studies could be explained by the difference in thyroid hormone requirement with various ages, iodine status, variable variants in the other allele and variable H2O2 supply by DUOX1/DUOXA1 system (27). This study found double variants in the same allele (c.2048G>T and c.4027C>T) in 3 patients (Table 3, families 1, 14 and 16). Although, there was a study which demonstrated increased severity in patients who had greater number of variants (29), this study demonstrated that 2 patients with compound heterozygous variants 3 variants) had both transient and permanent CH, but the patient who had heterozygous variant (2 variants) experienced permanent CH. These heterozygous variants have never been reported as a cause of CH, so functional studies of these variants are required. Additionally, c.2895_2898delGTTC variant which was commonly reported in Western population (30), was found in only one patient in this study. Therefore, the variant frequency seemed to be ethnic specific.

Four different variants in DUOXA2 gene were identified in this study. The nonsense variant c.738C>G was the most frequent DUOXA2 variant. Its functional studies have already been performed (31, 32). In normal control Thai database, this variant had low MAF of 0.002. Interestingly, this variant in DUOXA2 gene which is usually related to TDH, was found in a heterozygous pattern in the patient who had an ectopic thyroid gland (Table 3, family 26). A previous study reported an association of ectopic thyroid gland with DUOX2 variants (33). We postulate that DUOXA2 variants might also be related to thyroid gland development. However, the functional impact of the heterozygous c.738C>G variant in DUOXA2 gene was not assessed, and the finding could not exclude DUOX2 or other gene deletions.

Both parents of the patient with homozygous variants of SLC5A5 had a heterozygous state of the variant confirmed by Sanger sequencing. This variant was not identified in our in-house Thai database (455 persons). The parents absolutely denied a history of consanguinity. The homozygous state in the patient could be caused by unrecognized consanguineous history of the family because the parents’ hometown was in the northeastern region of Thailand.

Two patients with compound heterozygous TG variants were identified in this study. The c.274+2T>G variant found in 1 patient was a common variant reported in Chinese patients (34). Although TG variants have been reported as the most prevalent cause of TDH in Europeans, they were infrequent in our cohort.

In this study, the compound heterozygous, in-frame deletion (c.670_672delGAC) and frameshift mutations (c.2422delT) in TPO gene were identified in two siblings. Both variants have previously been reported (35, 36). Both patients developed goiter during childhood as a CH presentation which was in accordance with that reported in a Japanese patient who carried the same c.670_672delGAC variant and developed large goiter at 8 years of age (37). Retaining about 50% of residual peroxidase activity might explain the mild phenotype (35, 37). The development of multinodular goiter was possibly caused by delay in diagnosis and treatment (37, 38).

TSHR variants cause variable CH phenotypes. Hypothyroidism in our patients with either compound heterozygous or homozygous TSHR variants was more severe than those carrying heterozygous variant which was similar to previous reports (39, 40, 41).

PAX8 variants were inherited via autosomal dominant pattern with variable expressivity (42). Interestingly, our patient with novel c.457_458delCT variant had an absent thyroidal uptake on thyroid scintigraphy, but normal appearance of thyroid gland on ultrasonography which is a characteristic finding of iodide transport defect. Therefore, PAX8 variants might affect sodium iodide symporter expression (43).

This study did not find the variants in the genes related to syndromic defects such as NKX2-1, FOXE1, JAG1 and TBX1 because the patients with obvious syndromic features and typical phenotypes were excluded from the CES analysis.

The strengths of this study include being the first relatively large study of genetic diagnosis of CH in Thai patients, having comprehensive clinical courses to be analyzed with genetic diagnosis and having thyroid function tests of heterozygous parents of the patients with DUOX2 and DUOXA2 variants. However, there were some limitations. First, DUOX1 and DUOXA1 genes which are required for full-function of DUOX2 and DUOXA2 genes were not included in TruSight One Sequencing Panel®. Second, patients with heterozygous variants of DUOX2 and DUOXA2 genes might carry undetected variants in the other allele, because NGS cannot detect a large gene deletion or variants in non-coding regions. Third, some recently identified genetic defects causing CH which were not included in the panel used in this study such as SLC26A7 could have been missed. In conclusion, DUOX2 variants were the most common cause of CH among Thai patients, while TG and TPO variants were less common. The c.1588A>T in DUOX2 gene was a common variant in this population.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This work was supported by a research grant from the Faculty of Medicine Ramathibodi Hospital, , Bangkok, Thailand (CF62002, 2019). T T is a recipient of the Research Career Development Award from the Faculty of Medicine Ramathibodi Hospital, , Bangkok, Thailand.

Author contribution statement

K S, T T, P M and P P designed the work, collected, analyzed and interpreted data for the work, and drafted the article. W C, N I, I S, B P, P J and S N undertook the laboratory work, analyzed and interpreted data for the work. P K, S P, C S, M K and C S collected the data. All authors read and approved the final article.