A Single Copy of the Recently Identified Dual Oxidase Maturation Factor (DUOXA) 1 Gene Produces Only Mild Transient Hypothyroidism in a Patient with a Novel Biallelic DUOXA2 Mutation and Monoallelic DUOXA1 Deletion

Transient congenital hypothyroidism in a patient with a single functional DUOXA2 allele indicates that the system can operate with only one functional maturation factor allele.

Abstract

Context:

Dual oxidases (DUOX1 and DUOX2) play a crucial role in the generation of hydrogen peroxide required in the oxidation of iodide and the synthesis of thyroid hormone. Heterodimerization with specific maturation factors (DUOXA1 and DUOXA2) is essential for the maturation and function of the DUOX enzyme complexes. Biallelic loss-of-function mutations of DUOX2 result in congenital hypothyroidism (CH), whereas a single reported case of homozygous DUOXA2 mutation (Y246X) has been associated with mild CH.

Objective:

We now report an infant with transient CH due to a complex genetic alteration of the DUOX/DUOXA system.

Results:

Our patient was born to euthyroid nonconsanguineous parents and presented with an elevated TSH and enlarged thyroid gland at neonatal screening. Genetic analysis revealed a missense mutation (C189R) on the maternal DUOXA2 allele. The mutant DUOXA2 protein showed complete loss-of-function in reconstituting DUOX2 in vitro. The apparent C189R homozygosity of the proband in the absence of the same mutation in the father led to detailed gene mapping, revealing an approximately 43-kb pair deletion encompassing DUOX2, DUOXA1, and DUOXA2. Thus, in addition to being deficient in DUOXA2, the proband lacks one allele of DUOX2 and DUOXA1 but has two functioning DUOX1 alleles.

Conclusion:

The transient CH in the presence of only one functional maturation factor allele indicates a high level of functional redundancy in the DUOX/DUOXA system.

Hydrogen peroxide (H₂O₂) is the essential cosubstrate for the oxidation of iodide and iodination of thyroglobulin (TG) by thyroid peroxidase (TPO) in thyroid hormone synthesis (1). It is produced at the apical membrane of follicular thyroid cells by dual oxidase (DUOX)/DUOX maturation factor (DUOXA) reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complexes. The dual oxidases (DUOX1 and DUOX2) constitute the catalytic core of the thyroidal H₂O₂ generator (2, 3). Heterodimerization of a DUOXA (DUOXA1 or DUOXA2) subunit with a DUOX subunit is essential for the maturation, membrane translocation, and function of the respective DUOX isoenzymes (4). In vitro studies suggest that only DUOX2/DUOXA2 and DUOX1/DUOXA1 complexes are fully functional, whereas the DUOX2/DUOXA1 and DUOX1/DUOXA2 complexes tend to dissociate at the cell surface, resulting in reduced or absent production of H₂O₂, respectively (5, 6).

The DUOX and DUOXA genes are contiguous on the long arm of chromosome 15 with the DUOXA genes oriented head-to-head to the DUOX genes in the DUOX1/DUOX2 intergenic region (Fig. 1A). Each of the two DUOXA genes thus shares the core promoter region with one of the DUOX genes, providing a mechanism for coordinated expression (4). DUOX2 loss-of-function mutations have been identified in patients with congenital hypothyroidism (CH), supporting a role of DUOX2 in thyroid hormonogenesis (7, 8). CH caused by biallelic loss-of-function mutations of DUOX2 is typically permanent (7, 9–11), whereas it is transient in monoallelic defects (7). In some patients with complete loss of DUOX2 activity, full compensation of the defect is possible after the neonatal period (transient CH with biallelic loss of DUOX2) (12, 13). Conceivably, the DUOX1 isoenzyme can support sufficient hormone synthesis in these patients. No mutations have been reported in DUOX1, and the phenotype of DUOX1 defects, if any, is unknown. Because DUOXAs are essential for reconstitution of functional DUOX-based oxidases in vitro, mutations in DUOXA were proposed to also cause hormone synthesis defects. A single reported case of a homozygous DUOXA2 nonsense mutation has been associated with mild permanent CH (5).

Fig. 1.

Molecular genetic analysis of the propositus and his parents. A, Schematic presentation of the DUOX/DUOXA locus showing the orientation of the genes on the long arm of chromosome 15. B, Summary of the genetic analysis of the proband's DUOX/DUOXA system illustrating the presence of only one functional DUOXA allele (maternal DUOXA1). C, Sequence electropherogram of a section of the proband's DUOXA2 gene showing a homozygous missense mutation, c.T565C (p.C189R). D, DUOXA2 topology model illustrating the location of the C189R mutation in the third transmembrane helix and the first documented DUOXA2 mutation (Y246X) in the second extracellular loop. E, Haplotyping of the interval on chromosome 15 spanning the region of apparent homozygosity in the proband. Results for the indicated markers are aligned with the symbols in the pedigree. The markers in red indicate a deletion on the paternal chromosome as evidenced by loss of heterozygosity in the proband. F, Results of CGH of proband DNA with pooled normal DNA as reference. The x-axis displays the probe position on chromosome 15, and the y-axis is the log₂ ratio of the paired proband/normal DNA hybridization signals. Each red dot in the lower two panels indicates an individual probe. The cyan curve represents data smoothed with a “fused lasso” regression method (15); the blue curve represents CGHseq breakpoint estimates (16). The analysis reveals a deletion that encompasses DUOX2, DUOXA2, and DUOXA1, and a duplication upstream between 42.7 and 42.9 Mbp. SORD, Sorbitol dehydrogenase.

We report a novel loss-of-function DUOXA2 missense mutation that is found in a compound heterozygous state together with a large deletion comprising DUOX2, DUOXA2, and DUOXA1 (Fig. 1B) in a patient with surprisingly transient CH.

Subjects and Methods

Case report

The proband, born at term by uncomplicated vaginal delivery, was the first child of nonconsanguineous euthyroid parents of German ethnic origin. The Apgar score was 9/10/10. Weight was 3.44 kg, length was 53 cm, and head circumference was 39 cm. He had a positive result in the newborn screen for CH (TSH, 58 mU/liter). At 14 d of age, high serum TSH level persisted at 151 mU/liter, with total T₄ being slightly low at 5.2 μg/dl. Treatment with 50 μg l-T₄ was started, reducing the serum TSH to 7.0 mU/liter within 1 wk. Serum TSH was subsequently maintained in the normal range with a reduced l-T₄ dose of 25 μg/d; serum TG level remained elevated (71 ng/ml). These results and an enlarged thyroid gland (6.2 ml) were compatible with dyshormonogenesis. The child grew normally on l-T₄ replacement, above the 50th percentile for weight and length. Two years later, after 1 month of l-T₄ withdrawal, the serum TSH rose to 5.3 mU/liter (reference range, 0.3–4.2 mU/liter), although the serum free T₄ concentration of 1.4 ng/dl remained within the reference range (0.9–1.7 ng/dl). Since then, the boy was seen at 3, 6, and 12 months by a pediatric endocrinologist and remained clinically euthyroid. TSH levels were 4.2, 3.2, and 3.4 mU/liter, respectively. Corresponding free T₄ concentrations of 1.4, 1.5, and 1.4 ng/dl, respectively, remained unchanged.

Both parents had no goiters, and the results of their thyroid function tests were within the reference range. Mother and father, respectively, had free T₄ of 1.2 and 1.6 ng/dl, free T₃ of 3.4 and 3.9 ng/liter (reference range, 1.9–5.1 ng/liter), and TSH of 1.3 and 0.9 mU/liter.

Screening of candidate genes

Written informed consent was obtained from both parents for participation in the clinical and genetic studies. Genomic DNA from blood samples was isolated with a QIAamp DNA Blood Kit (QIAGEN, Hilden, Germany). All coding sequences of the candidate genes pendrin (PDS), TPO, DUOX2, DUOXA2, sodium iodide symporter (NIS), and TG were amplified by PCR (detailed protocol and primer sequences available upon request from the authors), purified by isopropanol precipitation, and subjected to automated sequencing (A3100 Avant Genetic Analyzer; Applied Biosystems, Weiterstadt, Germany).

Expression vectors, cell culture, and transfection

The c.T565C (p.C189R) mutation was introduced into an expression vector encoding N-terminal myc-epitope tagged human DUOXA2 by site-directed mutagenesis as previously described (5). The expression vector for DUOX2 was prepared as described (4). All constructs were verified by sequencing. HeLa cells were cultured and transfected as described previously (14).

Determination of NADPH oxidase activity

DUOX2-generated H₂O₂ release was assessed by endpoint fluorescence assay using cell-impermeable 10-acetyl-3,7-dihydroxyphenoxazine (14).

Western blot analysis

Whole cell lysates were prepared, and Western blot analysis of DUOXA2 protein was performed as previously described using anti-c-myc clone 9E10 (Roche Applied Science, Indianapolis, IN) at a concentration of 0.1 μg/ml (14).

Array comparative genomic hybridization (CGH)

Proband DNA was cohybridized with pooled normal DNA as a reference on a chromosome 15-specific CGH array with a mean probe spacing of 175 nucleotides (Roche NimbleGen, Madison, WI). Log-transformed data were smoothed with a “fused lasso” regression function (15), and probability of breakpoints was estimated by the CGHseq method (16).

Agilent array CGH 2.1M

Genomic DNA from blood leukocytes of the patient and his father was isolated with the Purgene DNA Isolation Kit (QIAGEN) according to the manufacturer's recommendations. The array CGH has been carried out using the Agilent array-CGH 2.1M according to the manufacturer's instructions. The data were analyzed using the Genomic Workbench Standard Edition v.5.0 (Agilent Technologies, Inc., Santa Clara, CA) and the Human Mar 2006 (NCBI36/hg18) Assembly.

Results

Direct sequencing of the proband's genomic DNA revealed no abnormalities in the coding sequences and adjacent introns of the PDS, TPO, DUOX2, NIS, or TG genes. However, we found a thymine-to-cytosine transition in codon 189 of the DUOXA2 gene resulting in the replacement of a highly conserved cysteine with an arginine (C189R) (Fig. 1C). This mutation, not found in 100 control alleles, introduces a charged residue into the third transmembrane helix of the DUOXA2 protein (Fig. 1D).

The patient appeared to be hemizygous for C189R because sequencing of his genomic DNA revealed only the maternal mutant allele, although his father was not a carrier of the mutation. These results suggested the presence of a deletion comprising DUOXA2 on the paternal chromosome. To test this hypothesis, we resequenced regions containing multiple database single nucleotide polymorphisms (dbSNPs) in all three family members to potentially identify informative polymorphisms. Lack of allele sharing between the proband and his father for two dbSNPs including rs2554452 (centromeric of DUOX2) was suggestive of a larger deletion on the paternal chromosome comprising at least DUOXA2 and DUOX2 (Fig. 1E).

Because no additional informative dbSNPs were found to more precisely define the deleted interval and assess the involvement of DUOXA1, the proband's genomic DNA was analyzed by high-density array CGH against pooled normal reference DNA (Fig. 1F). The results revealed a monoallelic deletion encompassing the coding regions of DUOX2, DUOXA2, and DUOXA1, and a duplication upstream of the deletion involving genes SPG11, PATL2, B2M, and TRIM69, none of which are known to be involved in thyroid function. The same result was obtained when we analyzed samples from the patient and his father using the Agilent CGH array. It furthermore proved that the father is a heterozygous carrier of the deletion and duplication in chromosome 15.

To assess the functional significance of the C189R mutation in reconstitution of DUOX2-based NADPH oxidase activity, we performed transient cotransfection experiments in a heterologous cell system. Wild-type (WT) DUOXA2, but not C189R DUOXA2, was able to reconstitute a H₂O₂-generating NADPH oxidase when coexpressed with DUOX2 in HeLa cells (Fig. 2A). Furthermore, Western blot analysis of the epitope-tagged WT and C189R DUOXA2 proteins showed virtually no steady-state expression of the mutant DUOXA2, similarly to the previously reported Y246X DUOXA2 (Fig. 2B) (5). Overall, these in vitro studies indicate that the C189R mutation results in an unstable protein that is rapidly degraded and does not allow the reconstitution of detectable NADPH oxidase activity.

Fig. 2.

In vitro functional analysis of the C189R DUOXA2 mutation. A, DUOX2-mediated H₂O₂ generation of HeLa cells cotransfected with the indicated expression vectors. Error bars represent sd (n = 3). B, Western blot analysis showing relative expression levels of WT (DUOXA2-myc) and mutant (C189R-myc) proteins in transiently transfected HeLa cells.

Discussion

We report here the second documented case of mutations in the DUOXA2 gene in an infant with mild CH. Our in vitro studies indicate that the C189R mutation results in a complete loss of functional DUOXA2 protein, similar to the previously described Y246X nonsense mutation (5). We cannot exclude the remote possibility that C189R is hypomorphic and that the finding of complete lack of function and expression in a heterologous system may not reproduce its properties in a follicular cell of the thyroid. Unfortunately, this hypothesis cannot be tested without thyroid tissue from the patient because there is no thyroid cell line that is devoid of endogenous DUOXA that could serve for transfection.

Our genetic analysis revealed that the proband was hemizygous for the C189R mutation, having inherited the mutant allele from his mother and a copy of chromosome 15 with a deletion comprising DUOX2, DUOXA2, and DUOXA1 from his father. The failure to amplify across the mapped deletion breakpoints using different primer sets and the presence of a centromeric duplication could indicate that the genetic rearrangement of the paternal chromosome is complex, involving an inversion event (17).

Despite having only a single functional DUOXA allele, specifically the maternal DUOXA1 copy, the child displayed transient CH compared with patients with biallelic DUOX2 loss-of-function mutations. Consistent with in vitro studies, in which DUOX2 can at least be partially activated by coexpression of DUOXA1 (4, 5), this suggests that a single functional allele of DUOXA1 provides sufficient residual capacity to synthesize thyroid hormone beyond the activity of DUOX1/DUOXA1 by partial rescue of DUOX2. This conclusion is further strengthened by the observation that the CH phenotype of the proband is similar in intensity to that of the patient with biallelic Y246X DUOXA2 mutations (see Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). In fact, after the age of 3 yr, he remained euthyroid without hormone replacement at the expense of TSH values in the 75th to 90th percentiles (for age adjusted values, see Robert Koch Institute, http://www.kiggs.de/experten/downloads/dokumente/KiGGS_Laborparameter%5B1%5D.pdf) and a modest thyroid gland enlargement. This shows that DUOX2 and DUOXA1 haploinsufficiency temporarily aggravates the CH phenotype caused by biallelic DUOXA2 defects. Also, in agreement with the initial description of DUOXA2 deficiency, his mother, being a heterozygous carrier of the mutation, did not display a thyroid phenotype. His father, having the heterozygous deletion comprising DUOX2/DUOXA2/DUOXA1, also had normal thyroid function tests when tested at an adult age. Maintenance of euthyroidism in members of this family illustrates the high degree of redundancy within the DUOX/DUOXA system for thyroid function.