Thyroid Hypoplasia in Congenital Hypothyroidism Associated with Thyroid Peroxidase Mutations

Introduction

Congenital hypothyroidism (CH) is the most common neonatal endocrine disorder, affecting about 1/3000 newborns. Sixty-five percent result from thyroid dysgenesis (TD) (1,2), including a large spectrum of developmental anomalies. Only in <10% of patients can a genetic cause be identified. Many of the remaining cases of primary CH with a (sometimes goitrous) eutopic gland (gland in situ [GIS]) are due to thyroid dyshormonogenesis related to mutations in TPO, TG, SLC5A5/NIS, SLC26A4/PDS, IYD/DEHAL1, DUOX2, and DUOXA2 genes, usually with autosomal recessive inheritance. Over the past years, the incidence of GIS has significantly increased. However, in a number of patients with GIS and the aforementioned majority of patients with TD, no molecular diagnosis is made (3). To identify novel candidate genes in patients with CH, whole exome sequencing (WES) was applied. Herein, the findings in two families with non-goitrous CH are reported.

Patients

Family 1

A male (patient #1) born to consanguineous parents (first cousins) from Sfax, Tunisia, was diagnosed with CH at day 10 of life, with a thyrotropin (TSH) level of 52 mIU/L (normal range [N] 0.7–10.1 mIU/L) and a free thyroxine (fT4) level of 14.2 pmol/L (N 9.3–23.5 pmol/L). Thyroid ultrasound (US) showed a non-goitrous eutopic thyroid, confirmed by scintigraphy. Levothyroxine (LT4) treatment was initiated immediately. The patient had a normal growth and puberty pattern, normal cognitive development, and mild bilateral sensorineural hearing loss. At 25 years of age, the patient was treated with LT4 at a dose of 200 μg/day (2.2 μg/kg/day), and he had a TSH of 3.16 mIU/L. The thyroid US showed hypoplasia, with a volume of 1.3 mL (N 12 ± 5 mL). At 27 years of age, LT4 was withdrawn for 49 days, and a thyroid US performed after withdrawal (TSH >100 mIU/L (N 0.27–5 mIU/L) and fT4 < 0.3 pmol/L (N 12–22 pmol/L) confirmed thyroid hypoplasia (2.9 mL; N 12 ± 5 mL). Scintigraphy showed normal ^99mTc uptake by the small normally located thyroid. Thyroid US and thyroid function tests were normal in the two parents and two other siblings (Fig. 1A).

OPEN IN VIEWER

FIG. 1.

Pedigrees of the two families harboring TPO mutations and clinical characteristics of described patients. (A) Family F1: Whole exome sequencing (WES) identified the homozygous missense mutation c.875 C>T, p.S292F in this consanguineous family. Both parents and one sibling are heterozygous for the mutation (wt/c.875C>T), whereas the index case (patient #1, II2) is homozygous. The mutation and familial segregation were confirmed by Sanger sequencing, represented in chromatograms; black dots show the nucleotide variant. On the right side, thyroid ultrasound showing thyroid gland hypoplasia. (B) Family F2: WES identified the compound heterozygous mutation c.387delC/c.2578G>A, p.N129Kfs*80/ p.G860R. The index cases (patients #2 and #3, II1 and II3, respectively) have CH with thyroid hypoplasia and thyroid in situ, respectively. Each parent is a carrier for one of the mutations. Representative chromatograms are shown for each family member; black dots show the nucleotide variant. For both families, no pathogenic variants were identified by sequencing of coding and exon–intron boundary regions of other genes involved in thyroid dysgenesis, including NKX2-1, PAX8, FOXE1, TSHR, GLIS3, BOREALIN, and NKX2-5. (C) Clinical characteristics of the patients and results of TPO gene analysis. hom, homozygous; het, heterozygous; RS number, reference single nucleotide variants number; NA, not available; n/a: not applicable; MAF, minor allele frequency. SIFT scores <0.05 predict a deleterious effect of the variant. PolyPhen-2 (Polymorphism Phenotyping v2 HumVar) values of 0–0.2 predict a benign variant, 0.2–0.85 a possibly damaging variant, and 0.85–1 a probably damaging variant. Color images available online at www.liebertpub.com/thy

Results

Three variants (with allele frequencies <1%) in the TPO gene (RefSeq transcript NM_000547.5) were found by WES. Patient #1 was homozygous for the missense mutation c.875C>T, p.S292F in exon 8 (5). Both parents and one sibling were carriers. The in silico assessment (SIFT, http://sift.jcvi.org; Polyphen-2, http://genetics.bwh.harvard.edu/pph2/) predicted this variant to be deleterious. No pathogenic variant was identified in SLC26A4/PDS, a candidate gene in this patient with CH and hearing defect. Patients #2 and #3 were compound heterozygous for the frameshift mutation c.387delC, p.N129Kfs*80 in exon 5 (inherited from the father) and the missense mutation c.2578G>A, p.G860R in exon 15 (inherited from the mother). The frameshift mutation leads to an early termination signal in exon 7 (6), while the missense mutation in exon 15 (7) is predicted in silico to be deleterious (SIFT) and possibly damaging (Polyphen-2; Fig. 1C). Notably, no variants were identified in the genes involved in TD, explaining thyroid hypoplasia, or dyshormonogenesis and goitrogenesis in these patients, after careful analysis of WES data (Supplementary Data and Supplementary Fig. S1).

Discussion

Biallelic TPO mutations are known to cause CH with thyroid in situ and most frequently goiter, which may develop during the fetal/neonatal period or at any time in childhood (8). Here, patients harboring TPO mutations with unexpected thyroid hypoplasia at birth are reported.

TPO encodes for thyroid peroxidase, located on the apical thyrocyte membrane and participating in thyroid hormonogenesis. TPO mutations impair thyroid function by impeding iodine organification, thereby leading to thyroid gland enlargement through chronic TSH stimulation (9). All three variants reported here have been published in patients with dyshormonogenesis (5 –7) and are in highly conserved and functional TPO domains. The biallelic mutation in exon 8, p.S292F, has been associated with sensorineural hearing loss (as a consequence of the CH), a feature in patient #1, and with multinodular goiter and follicular carcinoma (5). The p.N129Kfs*80 variant may be responsible for the loss of a crucial region for TPO function (6). Finally, exon 15 encodes the membrane-spanning part of the enzyme, and the nucleotide variation in patients #2 and #3 (c.2578G>A, p.G860R) is likely to impair insertion of the TPO protein into the apical membrane (7). The combination of these two TPO variants is likely to result in altered TPO function and CH.

To the authors' knowledge, the patients reported here are the first to have TPO mutations and small thyroid glands. These unexpected findings are consistent with recent data on SLC26A4/PDS (10) and DUOX2 (11) mutations in patients with CH phenotypes presenting with hypoplasia and non-goitrous hypothyroidism. Furthermore, in a previously described rat model, WIC-rdw rats have dwarfism and non-goitrous hypothyroidism due to a missense mutation in the Tg gene (12). The thyroid is hypoplastic in rdw/rdw rats, in contrast to previously reported Tg mutations. A suggested hypothesis to explain the absence of goitrogenesis may be the cytotoxic effect of the misfolded protein (13). This could be a hypothesis for hypoplasia in the patients in this study. However, it does not explain why previously described patients harboring the same mutation presented with a goiter. A second additional genetic or epigenetic hit in the thyroid tissue could explain the small and not enlarged thyroid in these patients at birth (14). Moreover, in these patients, treatment was initiated immediately after diagnosis, and it is unclear whether their thyroid glands would have responded and have grown under postnatal TSH stimulation. TPO is expressed at the late stages of thyroid development, making a primary structural defect unlikely. It is also noteworthy that WES did not identify any variants in genes involved in defects associated with TD such as TSHR, NKX2-1, PAX8, FOXE1, GLIS3, BOREALIN, or NKX2-5. A genetic cause is known only for about 10% of TD cases. In a large cohort of CH patients screened by a next-generation sequencing approach, including 11 TD and dyshormonogenesis genes, one or more disruptive variants were found for only 32.9% of patients, confirming that the genetic background of TD with a gland in situ still remains poorly understood and may have a complex etiology (15). Conceivably, additional epigenetic or genetic factors (modifier genes) and environmental influences (endocrine disruptors, iodine status) may also influence the functional and structural phenotype.

In conclusion, patients with CH harboring TPO mutations with unexpected thyroid hypoplasia at birth are described. However, it remains elusive with how these biallelic TPO mutations explain the hypoplastic phenotype and additional (epi)genetic or non-genetic mechanisms, not yet unraveled, may also be required.