Abstract
Background:
Allan-Herndon-Dudley syndrome (AHDS) is an X-linked type of mental retardation resulting from hindered thyroid hormone access to neurons. Clustered nonrecurrent deletions of SLC16A2 exon 1 have been described in three patients with AHDS. We report a fourth patient with such a deletion and discuss possible mechanisms leading to these rearrangements.
Case Presentation:
A three-and-a-half-year-old male with clinical and biochemical AHDS phenotype and a history of normal neonatal screening for hypothyroidism underwent SLC16A2 molecular analysis. Unexpectedly, he showed skeletal signs of hypothyroidism.
Methods and Results:
The exons of the SLC16A2 (MCT8) gene and the sequences surrounding exon 1 were amplified using PCR. The patient had a 36-kb deletion affecting exon 1 of SLC16A2. The deletion junction was subjected to bioinformatic analyses, along with two other reported exon 1 deletion junctions, identifying possible sequence features and mechanisms responsible for such genomic rearrangements.
Discussion/Conclusion:
This patient had a classic AHDS phenotype with an unexpectedly large anterior fontanel and delayed bone age and dentition. Bioinformatic analyses suggested that exon 1 deletions in patients with AHDS are caused by microhomology-mediated replicative-based and nonhomologous end-joining mechanisms. Rearrangement susceptibility may be due to the size of intron 1 and the percentage of repeat sequences.
Introduction
A
To date, over 60 heterogeneous mutations have been described in the SLC16A2 gene, including point mutations (53%), small insertions or deletions (30%), and deletions of entire exons (17%) (2). Even if a clear genotype–phenotype correlation has not been established, patients with mutations that retain significant residual transport activity show a milder phenotype (3,4), while deletions of entire exons predict complete MCT8 inactivation, resulting in a severe phenotype (2).
Although most of the mutations identified to date are private, small mutations have been observed in two families each (2,5 –8). Large deletions affecting whole exons have also been identified in more than one family. Deletions of exon 1 have been described in three families, even though the exact breakpoints have been identified in only two of them, suggesting that deletions cluster in this region (4,6,8). It has been described that nonrecurrent rearrangements with scattered breaking points can originate from several different mechanisms including nonhomologous end joining (NHEJ), microhomology mediated end joining, nonallelic homologous recombination, and replicative-based mechanisms (9) (see the Glossary located at the end of this article). The mechanism that leads to a specific rearrangement is thought to be influenced by the local genomic architecture; therefore, analyzing the characteristics of breakpoints and their surrounding sequences permits the delineation of possible mechanisms influencing that rearrangement.
We describe a fourth patient with an AHDS phenotype and a deletion of exon 1 of the SLC16A2 gene, and perform bioinformatic analyses of the breakpoints, surrounding regions and junction fragments of the available SLC16A2 deletion sequences in order to discuss possible mechanisms leading to these rearrangements.
Case Report
The patient was a 3.5-year-old male, the second son of a young and healthy nonconsanguineous couple. Pregnancy was uneventful until polyhydramnios during the seventh month prompted a Cesarean section during week 39 of gestation. At birth, the patient weighed 2,700 g, was 49 cm in length, both within normal percentiles (pc), and had Apgar scores of 8/9. Neonatal screening for congenital hypothyroidism (CH) was reported normal.
At age 3 months, the patient showed somnolence, difficulty feeding, and a wide anterior fontanel, prompting medical consultation. The initial evaluation in our institute at age 9 months showed a psychomotor development delay, severe hypotonia with altered swallowing mechanics that led to a gastrostomy at age 13 months, and an atypical thyroid profile, consisting of a total triiodothyronine (T3) of 344 ng/dL (normal range, 72–170 ng/dL), a total thyroxine (T4) of 5.4 ng/dL (normal range, 4.5–12.5 ng/dL), and a thyrotropin (TSH) of 3.09 μU/mL (normal range, 0.4–4.0 μU/mL).
Physical examination at an age of 2 years, 3 months showed a weight of 8,900 g (pc <3); a height of 78 cm (pc <3); a head circumference of 45.6 cm (pc <3); a wide anterior fontanel; dysmorphism (ptosis, long lashes, depressed nasal bridge, telecanthus, bulbous nose, anteverted nares, and prominent upper lip); central hypotonia; and hyperreflexia. One year later, his height and weight remained under the third percentile, the anterior fontanel was no longer palpable, dentition was delayed (presence of only 7 teeth), and hypotonia prevailed, with a myopathic face, ptosis, and spasticity of the pelvic limbs. Peripheral muscle hypotrophy was evident and hyperreflexia persisted. A Gesell developmental schedule at age 2 years, 6 months showed a global retardation coefficient of 85 percent. Assessment of bone age at a chronological age of 3 years, 6 months showed a severe disharmonic delay ranging from 1 year at the metacarpals to 3 years at the phalanxes.
Thyroid tests have repeatedly shown a particular profile, consisting of high total and free T3, normal-low total and free T4 and normal TSH levels. Despite thyroid hormone replacement starting at age 10 months with a levothyroxine dose ranging from 5 to 8 mcg/kg/day, there was only slight improvement in total and free T4 concentrations and clinically, only marginal motor improvement was noted. These findings suggested a diagnosis of AHDS and prompted molecular analysis of the SLC16A2 gene.
Thyroid hormone tests of the patient's mother showed a TSH of 1.0 μU/mL (normal range, 0.4–4.0 μU/mL); total T3 of 170 ng/dL (normal range, 72–170 ng/dL); and total T4 of 6.5 ng/dL (normal range, 4.5–12.5 ng/dL).
Methods and Results
The genomic DNA of the patient, his mother, healthy brother, and maternal grandmother were isolated from peripheral blood according to standard procedures. A 3-fold PCR amplification assay (5) in the patient's sample showed the integrity of exons 2 to 6 and a deletion of exon 1. In addition, a PCR-based mapping strategy, utilizing sets of primers distributed in the vicinity of exon 1, led to the amplification and sequencing of a 2.8-kb fragment that allowed the characterization of a 36,108-bp deletion (Fig. 1A). The deletion junction sequence was deposited in the Genebank database (accession number KJ608500).

Overview of three reported deletions of exon 1 of SLC16A2 (University of California, Santa Cruz [UCSC], Human Genome Browser, hg18) showing the extension of each deletion
Analysis of the DNA sample of the proband's brother showed integrity of the six SLC16A2 exons. The PCR product containing the breakpoint (2.8 kb) was identified in the mother of the proband, establishing her as a carrier, but it was not detected in his maternal grandmother.
The genomic architecture around the known SLC16A2 exon 1 deletions was analyzed using bioinformatic programs that allow the identification of features frequently found around genomic rearrangements: microhomology, presence of repetitive elements, non-B DNA structures and DNA sequence motifs (see Glossary). Microhomology of one or more base pairs between the proximal and distal reference sequences surrounding the breakpoints was analyzed by performing multiple sequence alignments with the Clustal W software (10). The presence of repetitive elements was evaluated using the RepeatMasker track of the University of California, Santa Cruz (UCSC) genome browser in order to screen DNA sequences for interspersed repeats and low complexity DNA sequences. Screening of DNA sequences leading to non-B DNA conformations was performed using the Repeat Around program to identify direct, inverted and mirror repeats, while oligo(G)n tracts were found using the QGRS Mapper program. Finally, 40 known sequence motifs found to predispose to DNA breakage were analyzed with the Fuzznuc program (11) (see Glossary).
The deletions of exon 1 with characterized breaking points, which include the patient described by Friesema et al. (1), the one reported by Visser et al. (4), and the one described here, were plotted using the UCSC genome browser BLAT tool (12) (Fig. 1A); their breakpoints, surrounding regions, and junction fragments were bioinformatically analyzed as described previously (11). Microhomology at the junction point, and a repeat sequence at the 5′ rupture point, were observed in our patient (4 bp of microhomology and a composite retrotransposon SVA composed of
Discrepancies between the number of bases involved in long interspersed elements (LINEs) and long terminal repeats (LTRs) in intron 1 vs. the total SLC16A2 gene is expected due to the adjustments made by the RepeatMasker program when a repeat is fragmented through deletions, insertions, or an inversion to create a more legible annotation closer to the biological reality. Further explanation can be found in (13) (see Glossary).
GC, guanine-cytosine; SINEs, short interspersed elements.
Discussion
We report a patient with AHDS due to a previously unreported inherited deletion in the SLC16A2 gene. The classic clinical manifestations of patients with MCT8 deficiency resulting from hindered T3 transport into neurons (14) include hypotonia, severe developmental retardation, and a distinctive thyroid profile with elevated T3, normal low T4, and normal TSH. Other clinical manifestations include poor head control, dysarthric or absent speech, muscle hypotrophy, difficulty feeding, spastic paraparesia, and paroxysmal dyskinesia. The patient described here presented with all these manifestations, except for paroxysmal dyskinesia.
Although the somatic symptoms associated with CH, such as macroglossia, umbilical hernia, myxedematous skin changes, and prolonged jaundice, have not been observed in patients with AHDS (5), our patient presented skeletal signs of CH, including a palpable anterior fontanel at age 27 months and delayed bone age and dentition. These manifestations could be explained, in part, by the patient's poor nutritional state. However, little is known about bone phenotype in these patients. Normal bone age has been reported in patients with the c.671C>T (15) and c.1535T>C (16) mutations, while advanced bone age was described in a patient with the c.1003C>T mutation (7). Albeit, evidence from mouse models has shown that a deficiency or excess of thyroid hormone can manifest in a tissue-specific manner (17), both effects in a single tissue have not been reported in AHDS. The discrepancy in bone phenotype among patients with AHDS may be due to the underlying mutation. However, descriptions of skeletal phenotype in other patients with this condition and experimental evidence are necessary to support this hypothesis.
Overall, the female carriers of these SLC16A2 mutations have a normal appearance, growth and intellectual function (5), still there have been some incidental reports of neurologically impaired carriers (18), which may be due to skewed X inactivation (19). In addition, around 25% of female carriers have shown an abnormal thyroid profile, with elevated T3 levels (18). The carrier mother of our patient had an IQ of 90 on the Wechsler Adult Intelligence Scale, a value considered normal, while her thyroid profile, although within reference ranges, was reminiscent of that of affected individuals.
To date, mutations in SLC16A2 have been reported in 69 families. Of the 61 different mutations in this gene, 80% are point mutations and small insertions or deletions, while the remainder consists of large rearrangements affecting one or more exons. Most of the described mutations are private, but some have been reported more than once (Supplementary Table S1; Supplementary Data are available online at
Of the 69 families with mutations in SLC16A2 and an AHDS phenotype, four (5.8%) have deletions in exon 1, making this the most frequently found mutation in these patients. Four other exon deletions and a translocation involving a rupture in intron 1 (18) have also been identified in AHDS patients (Supplementary Table S1), this results in an intron 1 rupture in 13% (9/69) of the mutations in SLC16A2. Moreover, two different deletions of exon 1 have been reported in which an AHDS phenotype is not clearly defined and cannot be deducted from the given information (Supplementary Table S1). These observations suggest that the region surrounding exon 1 is a hotspot for genomic rearrangements.
Mechanisms leading to clustered microdeletions have been studied in different pathologies. Non-homologous end-joining was initially regarded as the main mechanism that causes these clustered nonrecurrent deletions (see Glossary), but over half of the analyzed junctions were found to have sequence features that suggest the occurrence of replicative-based repair mechanisms (11).
Intron 1 of SLC16A2 possesses two characteristics that may account for its frequent participation in genomic rearrangements: its large size of almost 99 kb and the fact that it consists of almost 55% repeated sequences. Analysis of partial deletions of the DMD gene responsible for dystrophinopathies showed that these features are involved in the genesis of genomic rearrangements. For example, large introns at the 5′ end of the DMD gene, such as intron 7 (over 109 kb), have a higher incidence of deletions (21). Moreover, the sequence of intron 7 of DMD was shown to be rich in long insertions (Table 1), which may have contributed to an increased distance between matrix attachment region sequences, favoring looping out of the chromatin in that region and an increase in the number of rearrangements (22).
Bioinformatic analyses of the deletions in exon 1 of the SLC16A2 gene (Fig. 1B) revealed that the 337-kb deletion described by Visser et al. (4) and the 36-kb deletion herein described, display microhomologies of 1–4 bp at the deletion junction, with only one of their breakpoints occurring in a repetitive element. Moreover, the presence of non-B DNA structures in only one of the breakpoint surrounding regions of each of these deletions could favor the collapse of replication forks or the occurrence of double strand breaks (11). These findings support the occurrence of a replicative-based repair mechanism, even if NHEJ cannot be rejected. Regarding the 24 kb deletion reported by Friesema et al. (1), the absence of microhomology and the occurrence of each breakpoint in a repetitive element from a different family, leaving no visible signature, suggest that this deletion resulted from NHEJ. Taken together, these observations support the hypothesis that microhomology-mediated mechanisms can also be responsible for the unique, clustered microdeletions of this region (11).
Conclusion
To our knowledge, this is the first Mexican AHDS patient and the fourth in which an exon 1 deletion has been identified. The phenotype resulting from this mutation is very severe. Although AHDS is characterized by the absence of somatic CH manifestations, this patient presented a wide anterior fontanel with delayed closure as well as retardation of bone age and dentition. A detailed description of skeletal manifestations in other AHDS patients should result in further characterization of the bone phenotype. Finally, a review of previously reported SLC16A2 mutations showed that intron 1 is susceptible to genomic rearrangements, which may be due to the size and percentage of repeat sequences in this intron.
Footnotes
Acknowledgments
We would like to thank Dr. T.J. Visser, who kindly provided the primer sequences used at the beginning of the PCR-based fine mapping strategy. This work was supported by Research Funding from the National Institute of Pediatrics, México D.F. (2014).
MAAO and BGT designed the study. BGT performed the experiments. BGT and MERF performed the bioinformatic analyses. BGT, AGA, MLRR, RCL, and BPE analyzed the clinical data. BGT, AGA, MERF, and MAAO prepared the draft manuscript. All authors contributed to discussion of the results and approved the manuscript.
Patient consent was obtained from all participants. No approval from an ethics committee or institutional review board was obtained, since the patients' parents gave informed consent for genetic testing to determine the cause of his disease.
Author Disclosure Statement
No competing financial interests exist.
