Two novel mutations in Gli-similar 3 in patients with congenital hypothyroidism and thyroid dysgenesis

Jie Lan The A liated Hospital of Qingdao University Chunhui Sun The Third People's Hospital of Qingdao Xinping Liang Qingdao Women and Children's Hospital Ruixin Ma The A liated Hospital of Qingdao University Yuhua Ji Yantaishan Hospital Miaomiao Li The A liated Hospital of Qingdao University Shiguo Liu The A liated Hospital of Qingdao University Fang Wang (  18700840614@126.com ) The A liated Hospital of Qingdao University


Introduction
Congenital hypothyroidism (CH) is the most common endocrine disease in infants, with a recent reported incidence of 1 in 1,400-2,800 and a sex ratio of approximately 1:2 (male:female) [1; 2]. Although its clinical features are not obvious at birth, if not diagnosed and treated in a timely manner, severe CH can lead to growth failure and permanent intellectual disability. CH is divided into permanent and temporary forms, which are in turn divided into disorders with primary, secondary, and peripheral causes [3]. Two pathophysiological mechanisms are responsible for primary, permanent CH: thyroid dysgenesis (TD) and thyroid hormone (TH) synthesis disorders.TD accounts for 80-85% of CH cases, including thyroid dysplasia, ectopic thyroid, or the absence of thyroid tissue [4],and approximately 10-15% of CH is caused by TH dyshormonogenesis [5]. Although TD usually occurs sporadically, the rate of asymptomatic thyroid developmental anomal among rst-degree relatives of people with TD was signi cantly higher than that of normal people [6]. Furthermore , genetic factors have been described in about 5% of cases. Previous studies have shown that mutations in one of several genes involved in thyroid formation may be related to TD, including thyroid stimulating hormone receptor (TSHR) or transcription factors PAX8, NKX2-1 or FOXE1 [7][8][9]. More recently, some additional genes have been associated with TD, including NKX2-5, JAG1, CDCA8, TUBB1, NTN1 ,GLIS3 [10].
The GLIS3 gene is located on chromosome 9p24.2, containing 11 exons, encoding for a 90 KD -size protein. It is highly expressed in the kidney, thyroid gland, endocrine pancreas, thymus, testis, and uterus. Lower levels of expression were also described in brain, lung, ovary, and liver [11]. GLIS3 is a member of the GLI-similar 1-3 (GLIS1-3) subfamily of Krüppel-like zinc nger protein transcription factors which play a key regulatory role in embryo-genesis and many biological processes ,such as thyroid hormone biosynthesis, pancreatic β cell generation and insulin expression ,and spermatogenesis [12]. Since Taha [15]. Thyroid hemiagenesis is a rare type of TD , but it does not associated with reduced thyroid function [16]. Additionally, rare heterozygous GLIS3 missense variants were identi ed in a recent study of a large cohort (18/177) of Caucasian patients with isolated CH.
Investigation of the thyroid phenotype of these affected cases shows that TD and in-situ thyroid gland are accounted for halves [17]. However, to date, the relationship between the GLIS3 mutations and its associated CH phenotype remains unclear.Therefore, we screened GLIS3 exons in the peripheral blood genomic DNA of 50 Chinese patients with CH and TD to detect mutations and tried to elucidate the functions of the identi ed mutations.

Patients
We selected 50 children diagnosed with CH and TD (23 with athyreosis, 21 with ectopy, and 6 with hypogenesis; male: 22, female: 28; average age: 2.5 ± 0.5 years) from Shandong province, including Jinan, Qingdao, and Weifang, between 2007 and 2016, via a neonatal screening program. There were three criteria for inclusion. First, the child was diagnosed with CH. Heel blood samples were collected from the newborn and analyzed to determine serum thyroid-stimulating hormone (TSH) levels. Children with TSH levels ≥10 μIU/mL were recalled, and their serum TSH, free thyroxine (FT4), and free triiodothyronine (FT3) levels were further evaluated. CH was diagnosed based on a high level of TSH (>4.2 μIU/mL) and low level of FT4 (<12 pmol/L). Second, the children were diagnosed with TD after thyroid ultrasound or thyroid nucleus scanning. Third, other congenital diseases, such as blood and immune system diseases, malignant tumors, and mental disorders, were excluded. This study was approved by the ethics committee of the A liated Hospital of Qingdao University, and blood samples were collected from 50 subjects and 100 healthy controls (male: female ratio ,1:1.2, age range ,1-40 years) after obtaining their written informed consent.
Genetic analysis DNA samples were extracted from the peripheral blood leukocytes of 50 patients with TD using a QIAamp Blood DNA Mini Kit (QIAGEN, Hilden, Germany). Primers for the 13 exons covering the coding sequence, anking intronic sequence, 5′untranslated region, and 3′ untranslated region of GLIS3 were designed using Primer 5.0. Polymerase chain reaction (PCR) was performed in a total volume of 25 µL, which contained 250 nM dNTPs, 100 ng template DNA, 0.5 μM each of the forward and reverse primers, and The PCR products were identi ed by agarose gel electrophoresis and analyzed using the Bio-Rad Gel Doc™ XR+ imaging system (Hercules, CA, USA). Products appearing as distinct, single bands were subjected to Sanger sequencing and compared with the GLIS3 reference sequence (NC-000009.12) to identify mutations.

Cell culture and quantitative reverse transcription PCR
For in vitro studies, we used 293T cells (passage 4) cultured in Dulbecco's modi ed Eagle's medium (Biological Industries, Beit HaEmek, Israel) containing 10% fetal bovine serum, 100 IU/mL penicillin, and 100 µg/mL streptomycin. The cells were incubated at 37°C in 5% CO 2 in a humidi ed atmosphere. Total RNA was isolated with TRIzol™ reagent (Ambion, Austin, TX, USA) from 293T cells transfected for 24 h. Reverse transcription (RT) was performed in a reaction mixture comprised of 1 μg total RNA, 1 μL random primer, 10 μL 2× TS Reaction Mix, 1 μL TransScript® RT/RI Enzyme Mix, 1 μL gDNA Remover, and RNasefree water to a nal volume of 20 μL (all reagents from TransGen Biotech). The reaction mixtures were incubated at 25°C for 10 min and then at 42°C for 30 min. Next, the samples were heated to 85°C for 5 s to inactivate TransScript RT/RI and gDNA Remover. Real-time PCR for each cDNA was performed in triplicate in a 20-μL reaction mixture containing 5 ng cDNA, 0.4 μL each forward and reverse primer (both 10 µM), and 10 μL 2× TransStart® Green qPCR SuperMix (TransGen Biotech, Beijing, China); the nal volumes were achieved by adding an appropriate volume of water. The samples were incubated in 96well plates on an Applied Biosystems 7900HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) at 95°C for 15 min, followed by 40 cycles at 95°C for 10 s and 56°C for 32 s. The 2 −ΔΔCt method was used to determine the relative quantitative levels of individual cDNA expression. Transcript levels were normalized to the level of β-actin, and values were expressed as relative differences compared to those in their corresponding controls. The GLIS3 primers used were forward, TTCAACGCCCGCTATAAACTG, and reverse, ATACGGCTTCTCGCCTGTGT.

Western blot analysis
After 48h of transfection, the cells were lysed and collected with radioimmunoprecipitation assay buffer and phenylmethylsulfonyl uoride at 100:1 ratio (Beyotime, Shanghai, China). Protein concentrations were determined using a BCA kit (Thermo Scienti c, Waltham, MA, USA). We separated 40 μg of each sample by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blotting was performed according to a routine process as previously described [18] with a rabbit anti-GLIS3 primary antibody (1:200; Proteintech, Rosemont, IL, USA) and secondary goat anti-rabbit antibody conjugated to horseradish peroxidase (1:5,000). The results were analyzed on a FluorChem Q (ProteinSimple, San Jose, CA, USA) with a chemiluminescent horseradish peroxidase substrate (Immobilon® Western, Millipore, Billerica, MA, USA).
Dual-luciferase reporting system The luciferase reporter system uses luciferin as a substrate to detect re y luciferase activity. We transfected 293T cells with 250 ng wild-type (WT) or mutant GLIS3 expression vector, with the Renilla luciferase expression vector as an internal reference and pcDNA3.1 as a control vector. We detected luciferase activity 48 h after transfection with the Dual-Luciferase® Reporter Assay System kit (Promega, Madison, WI, USA). Fire y luciferase activity was recorded and calculated as a ratio to Renilla luciferase activity.

Statistical analysis
Statistical analysis was performed using the paired Student's t-test. A p-value less than 0.05 was considered as statistically signi cant; ** represents p < 0.01 and * represents p < 0.05. Graphs were prepared using GraphPad Prism 7.0 software (GraphPad, Inc., La Jolla, CA, USA).

Results
Screening of GLIS3 mutations in a cohort of patients with CH and TD Among the 50 unrelated patients with CH and TD, we identi ed two variants of GLIS3. The variant c.2710G>A is located on exon 11, resulting in a glycine-to-arginine substitution at codon 904 of the protein (p.G904R) ( Figure 1A). The variant c.2507C>A is located on exon 10 and causes a proline-toglutamine substitution at codon 836 (p.P836Q) ( Figure 1B). The mutations were not detected in 100 healthy individuals.
Clinical characteristics of patients with GLIS3 mutations Patient 1 (P1), a 2-year-old boy with the G904R mutation, was born by vaginal delivery after full-term gestation with a birth weight of 3,000 g. He was rst diagnosed with CH at 20 days of age with a serum TSH level >110 μIU/mL, FT3 level of 5.8 pmol/L, and FT4 level of 2.67 pmol/L. The patient had no family history of thyroid disease. B-ultrasound inspection indicated athyreosis, suggesting a diagnosis of CH. He was initially prescribed 25 μg/d levothyroxine (L-T4), and his current dose is 50 μg/d at 5 years old. P2, a 3-year-old girl harboring the P836Q mutation, was a full-term infant from unrelated parents with no family history of thyroid disease. Fourteen days after birth, she was diagnosed with CH based on a serum TSH level >100 μIU/mL, FT3 level of 4.1 pmol/L, and FT4 level of 5.31 pmol/L. Her B-ultrasound results, similar to those of P1, showed athyreosis. After diagnosis, L-T4 replacement therapy was initially administered at a dose of 25 μg, and her current dose is 83 μg/d at 7 years of age. Both patients exhibit normal growth and intelligence.
Effects of G904R and P836Q mutations on GLIS3 mRNA and protein levels We successfully expressed WT and mutant GLIS3 in 293T cells. As shown in Figure 3A, quantitative RT-PCR revealed that the mRNA expression levels of G904R and P836Q were 59.95% and 31.23% lower than that of WT GLIS3, respectively. We also examined the expression of the proteins generated by 293T cells after transfection with the control and mutant vectors. As shown in Figure 3B, the protein expression level of the G904R mutant was signi cantly lower than that of the WT protein, and the protein expression level of the P836Q mutant did not signi cantly differ from that of the WT protein.

Effect of mutations on GLIS3-mediated transcriptional activation
After co-transfection of GlisBS and G904R, luciferase activity was signi cantly lower than that in cells transfected with WT GLIS3 (p < 0.05) and it was partially recovered by additional co-transfection with WT GLIS3 (Figure 4). However, the ability of P836Q to promote reporter gene expression was intact.

Discussion
The GLIS3 protein has a relatively large N-terminus, the zinc nger domain (ZFD) which including ZF1-ZF5, and C-terminal transactivation domain (TAD).Targeted deletion of the N-terminus increases GLIS3 transcriptional activity suggesting the N-terminal domain may act as a repressor [19].The ZFD and TAD play important roles in GLIS3-mediated transcriptional activation, and ZF2-5 are required for GLIS3 DNAbinding site sequence recognition [20]. In-vitro studies have determined the optimal sequence for GLIS3 binding known as the Glis binding site (GlisBS), 5'-(G/C)TGGGGGG(A/C) [21]. the GlisBS are located within the regulatory regions of target genes. Once bound to the DNA, GLIS3 can repress or enhance the expression of target genes [22].
In the current studies, patients with loss-of-GLIS3-function mutations most develop a syndrome as neonatal diabetes and congenital hypothyroidism (NDH). The patients showed decreased levels of T3 and T4 along with elevated TSH and thyroglobulin (TG). Furthermore, depending on the nature of the mutation, NDH patients may develop a wider range of abnormalities, such as polycystic kidney, hepatic brosis, glaucoma, osteopenia and mild mental retardation [14; 23]. Similar phenotypes have been observed in Glis3-de cient mice, however, histological examination of the thyroid gland suggested that Glis3 does not signi cantly affect thyroid gland development [24].Three types of GLIS3 alterations associated with NDH have been described. Homozygous frameshift mutations, p.Arg780Profs*79, p.Gly311Alafs*15, and p.Pro772Leufs*35 resulting in an early termination codon and the loss of the GLIS3 transactivation domain. Deletion mutation encompassing exons 5-9, 3-4,9-11, 10-11 and larger deletions covering regions >100 kb that include exons 1-2, 1-4, or 5-9 and several homozygous missense mutations, such as p.Arg589Trp, p.Cys536Trp and p.His561Tyr [25]. Notably, the individuals with NDH have widely ranging thyroid structures, including athyreosis, hypoplasia, perifollicular brosis, interstitial brosis, and normal thyroid anatomy [22], which may be attributed to the tissue-speci c expression of variable-length transcripts derived from exons 11 of the GLIS3 gene [26]. Besides, in several patients with apparently normal thyroid morphology, elevated TSH and TG levels seem to be resistant to levothyroxine treatment [21].
As the inconsistent clinical features of CH caused by GLIS3 mutation, which has made it di cult to ascertain its causative mechanism. In the study of Kang et al, the development of hypothyroidism in Glis3KO mice seems to be related to dyshormonogenesis, its mechanism of action is that GLIS3 acts downstream of TSH/TSHR and is essential for the induction of TSH/TSHR-mediated TH biosynthesis and the proliferation of thyroid follicular cells [27]. In addition, despite a relevant role in thyroid cell proliferation, no signi cant thyroid developmental defects were observed in Glis3 knock-out mice [27].These ndings are different from those observed in patients with NDH, in which the phenotype of the thyroid gland varies from aplasia/ dysplasia to dyshormonogenesis. Therefore, it is possible that in a different genetic background, GLIS3 de ciency might also affect embryonic thyroid development [26].
More recent evidence demonstrates that the down-regulation of GLIS3 in zebra sh embryos leads to thyroid developmental defects. GLIS3 morphants showed a decreased expression of the early transcription factors nkx2.4 and pax2a in the thyroid primordium [28].
Nevertheless, in the study of Glis3KO mice, the expression of genes essential for thyroid development (including Pax8, Tttf1 (Nkx2.1) and TttF2 (Foxe1)) was slightly increased or unchanged in the thyroid glands [27]. A reasonable explanation about the difference in thyroid phenotype between the mouse and zebra sh models may be due to the time difference in Glis3 expression during embryonic development [29]. As a result, the current studies suggested that Glis3 might regulate multiple levels of thyroid function.
In this study, we identi ed two novel heterozygous missense mutations of GLIS3 in two unrelated patients among 50 Chinese patients with CH and TD. They are both athyreosis, only patient carried G904R mutation presenting with abdominal distention and lethargy when diagnosed with CH. This is different from Fu C et al 's study, the patient with GLIS3 heterozygous missense variants (c.2159G >A /p.R720Q) has an increased size thyroid gland, and also carried a compound heterozygous DUOX2 variant inherited from parents [30]. However, due to the rejection of the children's parents, we did not conduct a family study. As both mutations we detected are located in the key C-terminal transcriptional activation region of GLIS3 and conserved across various species, we explored the effects of these mutations on the function of GLIS3 protein and its ability to activate transcription. We found that both mutant transcripts were expressed at a lower level than the WT mRNA, whereas only the G904R mutation caused lower expression of GLIS3 protein. In addition, G904R had impaired transactivation, indicating that this mutant failed to activate transcription in the GlisBS region. This may provide an explanation for why the patient carrying G904R mutation featured TD. We also found the mutation had no dominantnegative effect on the WT protein, which suggests that a nonfunctional dimer is not formed between the mutant protein and the WT protein, so that the function of the WT protein is not inhibited. Another mutation (P836Q) promotes the reporter gene expression unaffected, it may cause the phenotype of TD through other means. Our research also showed that different mutations of GLIS3 can cause CH with TD through different pathogenic mechanisms.

Conclusion
In summary, our ndings broaden the GLIS3 mutation spectrum and provide a possible explanation for how GLIS3 defects cause TD. These two mutations (p.G904R; p.P836Q) were reported for the rst time in the Chinese population. Since only 50 CH children were selected in the case group in this study, the mutation rate of GLIS3 gene is not representative. We are currently using high-throughput sequencing technology to expand sample collection across the country, further screening GLIS3 gene mutation points to provide a more reliable GLIS3 gene mutation rate.
Declarations Figure 1 Partial GLIS3 sequences from patients with missense mutations and individuals with the WT sequence.  Comparison of the GLIS3 protein amino acid sequences across different species. Red rectangular frames indicate the locations of G904R and P836Q.

Figure 3
Expression of WT and mutant GLIS3 in transfected 293T cells. A. At 24 h after transfection, we observed the mRNA expression of WT and mutant GLIS3. B. G904R mutation, but not the P836Q mutation, reduced the expression of GLIS3 compared to that with WT GLIS3. C. Histograms represent the relative intensity of GLIS3. GAPDH was used as a quantitative protein control.