Novel Sodium/Iodide Symporter Compound Heterozygous Pathogenic Variants Causing Dyshormonogenic Congenital Hypothyroidism

Introduction

Iodide transport defect (ITD) (Online Mendelian Inheritance in Man no. 274400) is an autosomal recessive disorder caused by impaired active iodide accumulation in the thyroid follicular cell, thus leading to dyshormonogenic congenital hypothyroidism (1). Iodide accumulation, the first step in thyroid hormonogenesis, is mediated by the sodium/iodide symporter (NIS) (2). To date, 16 different loss-of-function variants in the NIS-coding SLC5A5 gene have been identified in patients with ITD (3). In this study, we report novel compound heterozygous NIS pathogenic variants in a pediatric patient with dyshormonogenic congenital hypothyroidism due to a defect in iodide accumulation.

Patient

The proposita was a full-term female patient, born in 1992 from nonconsanguineous healthy parents. On day 20, an abnormally high thyrotropin (TSH) level was detected by newborn screening (>200 μU/mL, cutoff <15 μU/mL). The diagnosis of congenital hypothyroidism was then confirmed: serum TSH 150 μU/mL (1.3–10 μU/mL), total thyroxine (T4) 1.1 μg/dL (6.0–14.0 μg/dL), and total triiodothyronine (T3) 30 ng/dL (80–240 ng/dL). Thyroid autoantibodies were negative. Thyroid function analyses were performed using a DELFIA system (PerkinElmer, United Kingdom). X-ray revealed absent ossification of knee epiphyseal nuclei. Thyroid hormone supplementation was started at day 28 with a daily dose of 12.5 μg/kg levothyroxine; however, treatment adherence was suboptimal during the first years of life. At 2.5 years of age, after levothyroxine withdrawal for a month, thyroid function evaluation indicated persistent congenital hypothyroidism (TSH >65 μU/mL, total T4 < 1 μg/dL, and T3 37 ng/mL). Ultrasonography showed a normal size and properly located thyroid gland. Radionuclide scintigraphy revealed nondetec table ¹³¹I uptake by the thyroid gland, as well as the salivary glands, suggesting a defect in iodide accumulation. At 13 years of age, evaluation using the WISC-III test demonstrated a total IQ of 82 (verbal IQ 84; performance IQ 82) with poor processing speed. The parents were clinically and biochemically euthyroid.

Results

The genetic analysis was approved by institutional ethics committee and performed under informed written consent of the proposita's mother. Genomic DNA was extracted from peripheral blood. SLC5A5 coding exons were analyzed using Sanger sequencing (4). The proposita was found to be compound heterozygous for two novel missense variants in the SLC5A5 gene: c.991G>A (exon 8), p.D331N, and c.1.641C>A (exon 13), p.S547R (RefSeq: NM_000453) (Fig. 1A). These variants were not reported in the Exome Aggregation Consortium, 1000 Genomes, and Single Nucleotide Polymorphism databases. Consistent with the recessive nature of the disease, the euthyroid mother was heterozygous carrier of the SLC5A5 variant p.S547R (Fig. 1B). The father was unavailable for analysis.

OPEN IN VIEWER

FIG. 1.

Identification and functional characterization of novel pathogenic NIS variants causing dyshormonogenic congenital hypothyroidism. (A) Partial sequence chromatograms covering the region of the variants in exons 8 (c.991G>A, p.D331N) and 13 (c.1.641C>A, p.S547R) of the SLC5A5 gene. The nucleotide and corresponding amino acid sequences—using the one-letter code—are indicated. The proposita is compound heterozygous for the mutant alleles p.D331N and p.S547R NIS. (B) Pedigree analysis of the members of the proposita's family. Arrowhead denotes the proposita. The mutated alleles for p.D331N and p.S547R NIS are shown in gray and black, respectively. (C) NIS secondary structure model. Numbered cylinders represent the 13 transmembrane segments and triangles indicate the 3 N-glycosylation sites. The amino-terminus faces the extracellular space and the carboxy-terminus the cytosol. The pathogenic variants p.D331N and p.S547R are outlined in red. (D) Steady-state iodide uptake in HEK-293T cells transiently expressing empty vector (Mock) or WT, D331N, or S547R NIS. The expression vector encoding WT human NIS cloned into pcDNA3.1 has been previously reported (8). Cells were incubated with 10 μM I⁻ in the absence (black bars) or presence (gray bars) of 40 μM perchlorate. Data are expressed in pmol of I⁻/μg of DNA ± SEM (n = 3) and standardized by transfection efficiency assessing total NIS expression by flow cytometry. *p < 0.05 (ANOVA, Newman–Keuls test). (E) Representative Western blot analysis of whole-cell lysates from transiently transfected HEK-293T cells probed with antihuman NIS and α-tubulin (T9026; Sigma-Aldrich) antibodies. Labels on the right side of the blot indicate the relative electrophoretic mobilities of the corresponding NIS polypeptides depending on their glycosylation status. (F) Representative merged immunofluorescence analysis of transiently transfected HEK-293T cells probed with antihuman NIS (green) and anti-SEKDEL epitope (sc-58774; Santa Cruz Biotechnology; red) antibodies under permeabilized conditions. Nuclei were stained with DAPI (blue). Scale bars: 10 μm. (G, H) PSI/TM-Coffee-generated multiple amino acid sequence alignment of NIS orthologs from different metazoan species (G) and the human SLC5A family members (H) around the mutated amino acids D331 and S547 of human NIS. Arrowheads denote the mutated amino acids. ANOVA, analysis of variance; NIS, sodium/iodide symporter; SEM, standard error of the mean; WT, wild-type. Color images are available online.

According to the current secondary structure model for NIS (2), D331 is located in the fourth extracellular loop connecting transmembrane segments 8 and 9, and S547 in the intracellularly facing carboxy-terminus, which is essential for NIS plasma membrane trafficking (5) (Fig. 1C). In silico analysis suggested that the variants D331N and S547R are pathogenic: PolyPhen-2, 0.6 and 1.0 (0.0: benign; 1.0: damaging); MutationTaster2, p = 1.0 and p = 0.98 (p-values close to 1 indicate a high security of the prediction); SIFT 0.01 and 0.00 (0.0: damaging; 1.0: tolerated), respectively. Functional analysis revealed that HEK-293T cells, which do not express NIS endogenously, transiently transfected to express D331N or S547R NIS displayed minimal to undetectable perchlorate-sensitive iodide accumulation compared with wild-type (WT) NIS-expressing cells (Fig. 1D). Moreover, on Western blots, the plasma membrane-located fully glycosylated NIS polypeptide was barely detected in D331N NIS-expressing cells, while it was absent in those expressing S547R NIS, indicating that these mutant proteins exhibit only partial maturation (Fig. 1E). In contrast, the electrophoretic pattern of WT NIS showed nonglycosylated (∼55 kDa, band A), immaturely glycosylated (∼60 kDa, band B), and fully glycosylated (∼90–100 kDa, band C) polypeptides (6) (Fig. 1E). In agreement, immunofluorescence analysis revealed that D331N and S547R NIS mostly colocalize with the endoplasmic reticulum marker SEKDEL, indicating that these mutant proteins are retained in the endoplasmic reticulum and are thus not fully targeted to the plasma membrane (Fig. 1F). Interestingly, multiple sequence alignment of NIS orthologs from different metazoan species revealed that missense variants D331N and S547R affect highly conserved residues (Fig. 1G), supporting their importance in NIS processing. Moreover, charged and serine residues at the positions corresponding to D331 and S547 in NIS, respectively, are highly conserved in SLC5A family members (Fig. 1H).

Discussion

The molecular characterization of loss-of-function NIS variants revealed that most of the mutant proteins are intracellularly retained (4). In this study, we identified two novel pathogenic NIS variants causing deficient iodide accumulation, leading to dyshormonogenic congenital hypothyroidism. The variants p.D331N and p.S547R severely decrease targeting of NIS to the plasma membrane, and consequently reduce NIS-mediated iodide accumulation. Significantly, normalizing the levels of iodide accumulation in terms of the amount of WT or D331N NIS expressed at the plasma membrane revealed that the decrease in iodide transport by D331N NIS results from impaired targeting to the plasma membrane, not from reduced intrinsic activity.

Importantly, Mizokami et al. (7) demonstrated deficient NIS-mediated iodide accumulation in the breast milk of a lactating mother carrying the homozygous loss-of-function NIS variant T354P. This observation is particularly relevant to the case presented here as the proposita is now a woman of reproductive age. Therefore, prophylactic iodine supplementation should be considered during a future pregnancy if exclusive breastfeeding is provided to the infant to prevent iodide deficiency in the nursing newborn.