| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
RAPID COMMUNICATION |
Dipartimento di Endocrinologia ed Oncologia Molecolare e Clinica (M.D., A.M.F., D.S., G.F., P.E.M.) and Dipartimento di Biologia e Patologia Cellulare e Molecolare L. Califano (L.S., R.D.L.), Università degli Studi di Napoli Federico II, 80131 Naples, Italy; Dipartimento di Biologia Cellulare e Neuroscienze, Istituto Superiore di Sanità (V.C., C.F., A.O., M.S., M.T.), 00161 Rome, Italy; Stazione Zoologica A. Dohrn, CEINGE (A.R., M.D.F., R.D.L.), 80145 Naples, Italy; Fondazione S. Maugieri, IRCCS (L.C.), 27100 Pavia, Italy; Dipartimento di Endocrinologia e Metabolismo, Ortopedia e Traumatologia, Medicina del Lavoro, Università di Pisa (A.P.), 56124 Pisa, Italy; Istituto di Puericoltura, Università La Sapienza (L.M., P.C.), 00185 Rome, Italy; Centro Screening Neonatale, Ospedale Santa Croce (V.S.), 61132 Fano, Italy; and Dipartimento di Medicina Sperimentale, Università degli Studi della Magna Grecia (M.B.), Catanzaro, Italy
Address all correspondence and requests for reprints to: Dr. Paolo E. Macchia, Dipartimento di Endocrinologia ed Oncologia Molecolare e Clinica, Università di Napoli Federico II, Via S. Pansin, 5, 80131 Naples, Italy. E-mail: pmacchia{at}unina.it.
Abstract
Context: Congenital hypothyroidism (CH) is a common endocrine disorder with an incidence of 1:3000–4000 at birth. In 80–85% of cases, CH is caused by defects in thyroid organogenesis, resulting in absent, ectopically located, and/or severely reduced gland [thyroid dysgenesis (TD)]. Mutations in genes controlling thyroid development have demonstrated that in a few cases, TD is a Mendelian trait. However, accumulating evidence supports the view that the genetics of TD are complex, possibly with a polygenic/multifactorial basis. A higher prevalence of congenital heart disease has been documented in children with CH than in the general population. Such an association suggests a possible pathogenic role of genes involved in both heart and thyroid development. NKX2–5 encodes a homeodomain-containing transcription factor with a major role in heart development, and mutations affecting this gene have been reported in individuals with congenital heart disease.
Objective: In the present work we investigated the possible involvement of NKX2–5 mutations in TD.
Results: Our results indicate that Nkx2–5–/– embryos exhibit thyroid bud hypoplasia, providing evidence that NKX2–5 plays a role in thyroid organogenesis and that NKX2–5 mutations contribute to TD. NKX2–5 mutational screening in 241 patients with TD allowed the identification of three heterozygous missense changes (R25C, A119S, and R161P) in four patients with TD. Functional characterization of the three mutations demonstrated reduced DNA binding and/or transactivation properties, with a dominant-negative effect on wild-type NKX2–5.
Conclusion: Our results suggest a previously unknown role of NKX2–5 in the pathogenesis of TD.
PRIMARY CONGENITAL HYPOTHYROIDISM (CH) is the most frequent endocrine-metabolic disease in infancy, with an incidence of about 1/3000–4000 newborns. In most cases (80–85%), primary permanent CH is caused by alterations in thyroid gland morphogenesis, with a gland that can be absent (athyreosis), ectopically located, and/or severely reduced in size (hypoplasia) (1). All these entities are grouped under the term thyroid dysgenesis (TD).
Accumulating evidence indicates that genetic factors are involved in the pathogenesis of TD. Mutations in TITF1 [NKX2–1 or thyroid transcription factor 1 (TTF-1)], Forkhead box E1 (FOXE1; TTF-2), PAX8, and TSHR have been demonstrated as causes of TD in animal models and in a small percentage of patients with TD (1).
Because congenital malformations, mostly cardiac, have higher frequency in children with CH than in the general population (2, 3), it is possible to hypothesize that the genes involved in heart organogenesis might also participate in thyroid development. Based on these findings, genes expressed during embryogenesis in precursors of thyroid and heart are promising candidate genes to be investigated for their involvement in the pathogenesis of TD.
One of these is the gene encoding for NKX2–5 (CSX or NKX2.5), a homeodomain-containing transcription factor that plays a crucial role in heart morphogenesis (4, 5, 6). In the mouse, Nkx2–5 expression precedes the onset of heart myogenic differentiation and continues in the cardiomyocytes of embryos, fetuses, and adults (4, 7, 8). In addition, on embryonic d 8.5–9.5 (E8.5–9.5), the Nkx2–5 transcript has been demonstrated in the precursors of thyroidal cells in the pharyngeal floor (9). At later stages, Nkx2–5 pharyngeal expression is limited to the area corresponding to the thyroid primordium (9).
Several mutations in the NKX2–5 gene have been identified in patients with congenital heart disease (CHD) (10, 11, 12, 13); however, no data on the thyroidal status of these patients are currently available.
Given the foregoing, we considered NKX2–5 as an excellent candidate gene involved in TD. In the present work, NKX2–5 mutational screening was carried out in a group of 241 patients with TD, allowing identification of a heterozygous missense mutation in four cases. Functional characterization of the identified mutations showed a reduction in DNA binding and/or transactivation ability. These data suggest a possible pathogenic role for NKX2–5 mutations in TD.
Patients and Methods
Embryonic expression of Nkx2–5 and analysis of thyroid phenotype in Nkx2–5–/– embryos
In situ hybridization on paraffin sections was performed as previously described (14) on staged Nkx2–5–/– embryos (gift from Dr. Seigo Izumo, Cardiovascular Division, Department of Medicine, Israel Deaconess Medical Center, Harvard Medical School, Boston, MA). The Nkx2–5 riboprobe was previously described (15).
Immunohistochemistry assays to detect the expression of Titf1, Pax8, and Foxe1 (TTF2) were carried out in Nkx2–5–/– and age-matched wild-type (WT) embryos using standard procedures (16).
Patient recruitment
Two hundred forty-one individuals with persistent CH (53 with athyreosis, 98 with thyroid ectopy, 15 with hypoplasia, and 75 with diagnosis of CH without goiter) were included in the study; 561 Italian volunteers with no thyroid-related disorders were also enrolled as controls. The study was approved by the ethical committees of the participating institutes. All individuals or parents of minors gave informed consent.
Mutation analysis
The entire NKX2–5 coding region was screened for mutations on gnomic DNA extracted from peripheral blood lymphocytes. Oligonucleotide primers (sequences available upon request) were designed based on reported human NKX2–5 cDNA and genomic sequences. PCRs were carried out in a 25-µl reaction volume with the following cycling parameters: 94 C for 8 min, 94 C for 45 sec, 64 C for 30 sec, and 72 C for 45 sec, for 33 cycles, then 72 C for 15 min.
PCR products were analyzed by denaturing HPLC, using the Wave 2100 DNA Fragment Analysis System (Transgenomics, Omaha, NE). Amplimers with abnormal denaturing profiles were sequenced bidirectionally.
Functional characterization of the mutant NKX2–5
Plasmids encoding murine NKX2–5 (17) and human type II deiodinase (DIO2)-chloramphenicol acetyltransferase (CAT) (18), thyroglobulin (TG)-CAT (19), and thyroperoxidase (TPO)-luciferase (Luc) (20) have been previously described.
Mutations were introduced by PCR-based mutagenesis with the QuikChange site-direct mutagenesis kit (Stratagene, La Jolla, CA) and were verified by sequencing.
Binding experiments, cotransfection, and Western blot experiments were performed as previously described (17). All experiments were performed in duplicate, and at least three separate transfections were carried out. Data (Luc/CAT or CAT/Luc ratios) are shown as the mean ± SD. Statistical analysis was performed using Student’s t test.
Results
Thyroid phenotype in Nkx2–5–/– embryos
We examined the expression of Nkx2–5 during thyroid development. Consistent with previous data (9), in situ hybridization demonstrated Nkx2–5 expression in the ventral region of the pharynx and in the thyroid bud on E8.5. Nkx2–5 was expressed in the thyroid primordium up to E11.5; thereafter, Nkx2–5 transcript was no longer detected in the thyroid bud, whereas it was present in the heart region (data not shown).
Because Nkx2–5 mRNA was present in the thyroid primordium at an early stage of development, we investigated whether its expression was required for thyroid morphogenesis. To such a goal, we analyzed the phenotype of the thyroid primordium in Nkx2–5–/– mouse embryos by immunohistochemistry on E9.5, using the anti-Pax8, anti-Titf-1, and anti-Foxe1 antibodies (Fig. 1
). At this stage, the thyroid primordium appeared as a midline epithelial thickening in the ventral wall of the primitive pharynx. The endodermal cells fated to become thyroid follicular cells were univocally identified by the expression of Nkx2–1, Pax8, and Foxe1. The identification in the foregut Nkx2–5–/– embryos of a thyroid primordium as a small outgrowing bud of endodermal cells indicated that thyroid morphogenesis occurred, and thyroid transcription factors were normally expressed. However, serial sections of Nkx2–5–/– and WT embryos revealed that the number of cells stained with Nkx2–1 and Pax8 were strongly reduced (<50%). Foxe1-positive cells also appeared reduced; however, the presence of a positive signal in the entire pharynx makes a precise quantification of such a reduction difficult. Our findings indicate that the thyroid bud appears smaller in the Nkx2–5–/– embryo compared with the WT embryo, suggesting that Nkx2–5 is required during thyroid development.
|
Given the role of Nkx2–5 in thyroid development, we screened for mutations in the coding region of the NKX2–5 gene a cohort of 241 individuals affected by TD. Three distinct heterozygous missense mutations were identified in four subjects (Fig. 2): two of them (355G
T and 482GC) had not previously been reported, whereas the third (73CT) was previously described in patients with CHD (12). The 355GT and 482GC transversions, predicting, respectively, the A119S change, occurred a few residues upstream from the beginning of the homeodomain, and the R161P substitution occurred within the homeodomain. The third mutation (73CT) changed an arginine residue to cysteine at position 25 (R25C).
|
The R161P change was found in a 6-yr-old girl (patient 2) with thyroid ectopy and hypothyroidism (TSH, 47 mU/liter; FT3, 2.8 pg/ml; FT4, 1.2 pg/ml). Of note, this patient exhibited patent foramen ovale at birth that resolved spontaneously and minor mitral valve insufficiency.
Finally, the R25C substitution was observed in one case with thyroid ectopy (patient 3) and in one case with athyreosis of the gland (patient 4). The former was a 24-yr-old woman with no evidence of cardiac malformation (TSH, 300 mU/liter; FT3, 2.74 pg/ml; FT4, 5.4 pg/ml). Patient 4 was a 15-yr-old boy with athyreosis (TSH, 419 mU/liter; FT3, 2,20 pg/ml; FT4, 1.6 pg/ml) born at the 35th week of gestation. Since birth, the baby presented bilateral cortex atrophy and attention deficit hyperactivity disorder, but his intelligence quotient was normal, and no history of cardiac abnormalities was reported.
The 355GT and 482GC changes were not observed among 561 controls, whereas the 73CT substitution was identified in a single, apparently healthy individual. Statistical analysis indicates a significant association between these defects and disease (P = 0.0305, by Fisher’s exact test).
In all cases the mutation was inherited from one of the parents (Fig. 2
). Electrocardiogram, thyroid function test, and cardiac and thyroidal ultrasounds were performed in all family members of the patients carrying the mutation. The results documented the presence of minor mitral valve insufficiency in the father transmitting the R161P change (family 2), whereas in the family segregating the A119S change, the mother transmitting the mutation exhibited autoimmune hypothyroidism and was on life-long treatment with L-T4.
Functional characterization of the TD-associated NKX2–5 mutants
The transcriptional properties of WT and R25C, A119S, and R161P NKX2–5 mutants were investigated by cotransfection assays in the HeLa cells as previously described (15, 17). All NKX2–5 mutants were able to activate the reporter genes in a dose-dependent manner, although their activity was reduced compared with WT in all the assays performed (DIO2, TG, and TPO promoters) at all tested concentrations (10, 25, 50, and 100 ng vector; Fig. 3A). This effect was not related to a reduction in the mutant protein concentration, as confirmed by Western blot analysis (data not shown).
|
. Coexpression of WT with the same amount of the three mutated proteins resulted in a reduction of luciferase activity, suggesting a dominant-negative effect of the mutants.
DNA binding properties of the three NKX2–5 mutants were also analyzed by EMSAs using a short DNA stretch corresponding to the NKX2–5-binding site within the DIO2 (oligo D) (17). The results are shown in Fig. 4. A significant reduction in DNA binding was observed for both A119S and R161P mutants, whereas the R25C mutation apparently did not interfere with the binding, as previously reported (21). WT, R25C, A119S, and R161P NKX2–5 protein levels were similar, as demonstrated by Western blots with the same extracts used for the EMSA assays.
|
In this study we provided evidence for the contribution of NKX2–5 gene mutations to the pathogenesis of TD. Specifically, we showed that although thyroid budding occurs normally in Nkx2–5–/– embryos, the number of thyroid precursor cells is clearly reduced compared with that in WT cells. Consistently, the expression of Titf1, Foxe1, and Pax8 in the thyroid primordium does not seem to be impaired in Nkx2–5–/– embryos.
Mutational screening of the entire NKX2–5 coding sequence in a cohort of 241 patients allowed the identification of three mutations, R25C, A119S, and R161P, in four subjects with TD (three patients with thyroid ectopy and one with athyreosis). Functional studies demonstrated that these mutants exhibited a significant functional impairment, with reduction of transactivation properties and dominant-negative effect, which was associated with reduced DNA binding in the A119S and R161P mutants. The R25C amino acid change was previously characterized, showing a defect only in dimer formation, without any significant effect on DNA binding as monomer or transactivation of the atrial natriuretic factor (ANF) promoter (21). Our results indicate that although the R25C NKX2–5 mutant normally also binds the DIO2 promoter, its activity on the DIO2, TG, and TPO promoters is significantly impaired. In addition the R25C NKX2–5 exerts a dominant-negative action on the activity of WT NKX2–5 measured on DIO2 and TG promoters. The observed discrepancies from previously published data could be the consequence of the different experimental models used, and they support the hypothesis that NKX2–5 regulates the expression of target genes through complex mechanisms. These mechanisms can be different in various tissues and at each developmental stage and can be mediated via the interaction with other proteins. In this scenario, it has been recently demonstrated that the expression of a dominant-negative Nkx2–5 (N188K) isoform can interfere with Titf1 action (22).
NKX2–5 is essential for normal heart morphogenesis, myogenesis, and function (6), and several loss of function mutations in NKX2–5 have been described in patients with CHD (10, 11, 12, 13). The coinheritance of NKX2–5 mutations and different CHD suggests that this transcription factor is involved in diverse developmental processes. In our patients a very mild (patient 2) or a completely absent (patients 1, 3, and 4) cardiac phenotype was observed. This observation raises the question of whether the effect of the mutations could be specifically critical for thyroid development rather than that of the heart. Analyses of animal models with tissue-specific gene targeting will probably address this point. In fact, other genes of the NK2 family, including Titf1 or Nkx2–6 and Nkx2–3, are expressed in addition to Nkx2–5 in the endodermal layer of the developing pharynx and in the thyroid anlage, with possible redundant functions (6, 23, 24).
In all cases, mutations were not de novo, but were transmitted by one of the parents. Of note, only the father in family 2 presented a minor mitral valve insufficiency. In addition, of 561 controls, one normal subject presented the same R25C mutation observed in two of our patients. The identification of one control carrying the mutation does not change the significant statistical association between TD and a mutation in NKX2–5; over all, these data suggest that NKX2–5 mutations have a variable penetrance and clinical significance, as previously demonstrated in both humans (10, 11, 12) and mice (5, 6), where a strong influence of the genetic background has been demonstrated (5). The incomplete penetrance and variable expressivity of mutations at both thyroidal and cardiac levels strongly suggest that other genes are expected to contribute to the phenotype and support the view that TD is a polygenic disorder (25) or that stochastic epigenetic mechanisms can play a relevant role (26).
In conclusion, we identified NKX2–5 as a novel gene involved in the pathogenesis of congenital hypothyroidism with ectopy or athyreosis of the thyroid gland. The variable phenotypes in patients with loss of function NKX2–5 mutations suggest that other factors/genes may play a role modulating penetrance and expressivity at both the cardiac and thyroidal levels.
Acknowledgments
We thank Dr. S. Izumo for providing staged embryos of NKX2–5–/– mice. We also thank Prof. S. Refetoff for critical comments on the manuscript.
Footnotes
This work was supported by Ricerca Finalizzata 1% FSN 2000 Basi Molecolari dell’Ipotiroidismo Congenito: Predizione, Prevenzione e Intervento (to R.D.L. and M.T.) and Ricerca Finalizzata 1% FSN 2003 Studio Molecolare di Nuovi Geni Candidati Nella Disgenesia Tiroidea Isolata o Associata a Malformazioni del Sistema Nervoso Centrale e/o a Disturbi Neurologici o Psichici (to M.T.).
Current address for L.S.: Dipartimento Clinico Sperimentale di Medicina e Farmacologia, Sezione di Endocrinologia, Via Consolare Valeria, 98125 Messina, Italy.
The authors have no conflict of interest.
First Published Online January 17, 2006
1 M.D. and V.C. contributed equally to the study. 
Abbreviations: CAT, Chloramphenicol acetyltransferase; CH, congenital hypothyroidism; CHD, congenital heart disease; DIO2, type II deiodinase; E, embryonic day; Foxe1, Forkhead box E1 (thyroid transcription factor 2); FT3, free T3; Luc, luciferase; TD, thyroid dysgenesis; TG, thyroglobulin; TPO, thyroperoxidase; TSHR, TSH receptor; TTF2, thyroid transcription factor 2; WT, wild type.
Received June 16, 2005.
Accepted January 5, 2006.
References
This article has been cited by other articles:
 |
|
 |
|
  E. Al Taji, H. Biebermann, Z. Limanova, O. Hnikova, J. Zikmund, C. Dame, A. Gruters, J. Lebl, and H. Krude Screening for mutations in transcription factors in a Czech cohort of 170 patients with congenital and early-onset hypothyroidism: identification of a novel PAX8 mutation in dominantly inherited early-onset non-autoimmune hypothyroidism Eur. J. Endocrinol., May 1, 2007; 156(5): 521 - 529. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Coding region. HD, Homeobox region. Identified amino acid substitutions and positions of affected residues are shown.
) or 50 ng WT NKX2–5 plus 50 ng of each construct (
). Results are expressed as the percentage of activation on both DIO2 and TG promoters compared with 50 ng WT NKX2–5. Coexpression of WT protein and each mutant resulted in a significant reduction in the transactivation activity as a consequence of a dominant-negative effect. Statistical significance was calculated with Student’s t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001).