| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
COMMENT |
T Mutation Is Caused by a Founder Effect
Department of Pediatric Endocrinology (H.Br., H.Bi., D.S., D.D., A.G.), Charité, Humboldt University Berlin, D-13353 Berlin; Institute of Human Genetics (A.P.), Technical University Munich, D-81675 Munich; and Institute of Human Genetics (A.P.), GSF National Research Center, D-85764 Neuherberg, Germany
Address all correspondence and requests for reprints to: Arne Pfeufer, M.D., M.Sc., Institute of Human Genetics, Technical University Munich, Klinikum Rechts der Isar, and GSF Federal Research Center, Neuherberg, Trogerstr. 32, 81675 München, Germany. E-mail: arne.pfeufer{at}web.de.
Abstract
Neonatal TSH screening has been a major achievement for the early detection and treatment of primary congenital hypothyroidism. It nevertheless fails to reveal cases of central hypothyroidism caused by TSH levels in the low normal range. In the last 10 yr, homozygous mutations in the TSHß-subunit gene have been recognized as a cause of central hypothyroidism with isolated TSH deficiency. The most frequent TSHß mutation 313
T (C105V) has been described in six apparently unrelated families. We investigated the frequency and possible monophyletic origin of the different TSHß 313T alleles of the three affected families. Haplotype analysis of five polymorphic single-nucleotide polymorphism loci in the TSHß region revealed the presence of seven different haplotypes in the general population. In all six parental lines, the mutation occurred on the same haplotype. Extending the haplotype by two flanking microsatellite markers led to a mutational age estimate of about 150 generations. In 500 unrelated individuals from the general population, we did not detect any TSHß 313T allele, suggesting a population heterozygote carrier frequency less than 1:170 with more than 95% probability. Accordingly, the disease risk in the general population because of homozygosity is low. Our data suggest a monophyletic origin of the TSHß 313T mutation from a common ancestor and no significant population prevalence. Therefore, identification and genetic counseling of heterozygous carriers in affected families seems to be more advisable than population-wide neonatal T4 screening programs for an early detection of this rare condition.
MEASURING OF NEONATAL TSH levels in dried blood spots has led to the identification of about 1 in 3500 newborns with primary congenital hypothyroidism. It proved enormously effective and cost efficient because a normal development is achieved in the vast majority of patients. However, patients suffering from congenital central hypothyroidism are missed in this screening procedure because of TSH levels in the low to normal range (1). In very rare cases, central hypothyroidism is caused by an isolated TSH deficiency (2). Within the last 10 yr, a number of patients have been reported to suffer from autosomal recessive mutations in the gene of the ß-subunit of TSH (TSHß) as the molecular cause of the disease (3).
Three different TSHß mutations leading to severe central hypothyroidism have been described and appear to be geographically focused: 1) a nonsense mutation in codon 12 (E12X) was exclusively detected in two Greek families (4); 2) in five Japanese families, a G29R mutation in exon 2 was found (5); and 3) recently a nonsense mutation at codon 49 (Q49X) has been described in a single geographic region (6). In homozygous carriers of any of the three variants peripheral TSH is not detectable by standard assays.
In contrast, thus far, the most frequent TSHß mutation, a frame-shifting 1-bp deletion in codon 105 (313T: C105V) in exon 3 of the TSHß gene, is less geographically focused. Having initially been described in two Brazilian families with consanguineous parents (7), affected children show low TSH levels and no increase in TSH after TRH stimulation. The same mutation has since been described in two German families (8, 9) and one Belgian family (10). We subsequently have identified this mutation in two additional families from Germany. All the above families were unrelated to our knowledge.
The TSHß gene is located in the chromosomal position 1p13 in tight physical linkage with the genes for the nerve growth factor ß-subunit (NGFß) and the n-ras oncogene (NRAS) (11). Physical positions of all three genes have been established by the ongoing sequencing project of the human genome (12). All three genes have been mapped between the anchor markers D1S2881 and D1S467, which are both included in the Généthon (13, 14) and Marshfield (15) genetic linkage maps.
The aim of this study was to investigate whether the apparently independent appearance of TSHß 313T mutation carriers was due to a mutational hot spot at this position or to a founder effect. A hot spot with repeated independent de novo mutational events would result in a more even distribution of mutation carriers in the general population. In addition, mutant alleles might be more frequent than previously recognized and have a potentially significant public health impact. In this case, additional newborn screening strategies (e.g. neonatal T4 screening) could become advisable to detect affected patients early.
Subjects and Methods
We examined children from three affected families. Family B (Fig. 1
) has been published previously (9). All individuals included in the study gave informed consent. The investigation was approved by the local ethics committee.
|
The patient is the first child of unrelated parents. Neonatal TSH screening was inconspicuous (TSH, <15 mU/liter). At the age of 4 months, the girl was transferred to hospital in a critical condition. She presented the full clinical picture of severe congenital hypothyroidism. Free T3 (fT3) was low (0.83 pmol/liter; age-related normal range, 2.2–8.8 pmol/liter), and free T4 (fT4) was undetectable (age-related normal range, 15–30 ng/liter). TSH was 0.15 mU/liter. Ultrasound examination revealed a normal thyroid. Under T4 substitution fT3 and fT4 normalized rapidly.
Patient III-2 of family A
The girl was born 16 months later than her sister (patient A:III-1). In the neonatal screening, TSH was normal (TSH, <15 mU/liter). Because of her sister’s medical history, thyroid hormones were investigated immediately after birth. Laboratory data confirmed the diagnosis of central hypothyroidism also in this girl: total T3 and T4 were diminished at 0.56 µg/liter (age-related normal range, 1.32–3.71 µg/liter) and 24 µg/liter (age-related normal range, 89–236 µg/liter), respectively. TSH was low (0.06 mU/liter). Under therapy with T4 (50 µg/d), which started at the second day of life, her physical and mental development was normal.
Patient III-2 of family C
The patient is the second child of unrelated parents. As a neonate he had icterus, feeding problems, and elevated liver enzymes. After 10 wk the patient was transferred to the hospital to rule out liver disease. The boy presented the typical symptoms of congenital hypothyroidism. Total T3 and T4 were very low at 0.29 µg/liter (age-related normal range, 1.05–2.45 µg/liter) and 5 µg/liter (age-related normal range, 59–163 µg/liter), respectively. TSH was 0.3 mU/liter. Under T4 substitution there was a fast normalization of T3 with 1.93 µg/liter, T4 with 117 µg/liter, fT4 with 19 ng/liter (age-related normal range, 11–27 ng/liter), but TSH was still very low with 0.07 mU/liter. At a chronological age of 10 wk, bone age was neonatal. Thyroid ultrasound and stimulated GH levels in the clonidine-arginin-test were normal. The boy’s development was delayed, first teeth with 11 months, walk with 16 months, and speech with 24 months of life. But in the following years, under consequent T4 substitution of 50 and 75 µg/d, the development ameliorated; logopedic therapy was done between yr 4 and 6. At an age of 9 yr, physical examination was normal. School performances were very good, and hearing and intelligence tests were normal (IQ 97).
Patient III-3 of family C
In the sister, who was 7 yr younger, neonatal screening TSH was normal. Because of the history of the brother, thyroid hormones were investigated and central hypothyroidism was also diagnosed. fT3 was 1.7 pmol/liter (age-related normal range, 2.2–8.8 pmol/liter), fT4 was 2 ng/liter (age-related normal range, 15–30 ng/liter), and TSH was 0.02 mU/liter. Under therapy with 50 µg/d T4 since the second week of life, physical and mental development was normal, height and weight between the 75th and 90th percentiles, first teeth with 9 months, walk with 13 months, and speech with 15 months. Development tests at the age of 2 yr were also normal.
Normal population controls
To determine the allele frequency for the TSHß 313T mutation, we selected 500 individuals from the random population. This control group was made up of students and members of the laboratory staff. None of the individuals in the control group showed any signs of thyroid dysfunction.
Methods
Genomic analysis. Genomic DNA was prepared from peripheral white blood cells using a DNA extraction kit (QIAmp blood kit; QIAGEN, Hilden, Germany). Patients and relatives were analyzed for mutations in the TSHß gene by amplification of all TSHß exons by PCR with primers and conditions published previously (9). Sequencing was performed using an ABI Prism 377 DNA sequencer and BigDye terminator kit (Applied Biosystems, Weiterstadt, Germany) according to the conditions supplied by the manufacturer.
Screening for the TSHß 313T allele via restriction analysis.
The TSHß 313T 1-bp deletion introduces a new Sna BI restriction site in exon 3 of the TSHß gene. Genotyping for the TSHß 313T allele was performed by PCR and restriction enzyme analysis. A 332-bp fragment was amplified from exon 3 by PCR (9), restriction digested with Sna BI (New England Biolabs, Inc., Beverly, MA), and separated on a 2.5% agarose gel. To control for correct assay performance, a heterozygous-positive control sample was analyzed in all experiments.
Microsatellite analysis.
Two flanking microsatellite markers were selected within a 1-cM interval around the TSHß gene. We selected two anchor markers from the Gene Map 1999, D1S2881 and D1S467, that had been included both in the Généthon (13, 16) and Marshfield (17) genetic maps (Table 1). Analysis of markers was performed using fluorescently labeled primers (TIB Molbiol, Berlin, Germany) by standard semiautomated methods using the ABI 377 DNA sequencer (Applied Biosystems). Primer sequences and amplification conditions were obtained from the Généthon map. After addition of 2 vol 95% formamide, the reaction products were electrophoresed in 4% denatured polyacrylamide gel using standard procedures.
|
. The SNPs were located in three physical clusters of less than 500-bp size each, which allowed SNP genotyping by amplification and direct sequencing of three PCR fragments: PCR fragment A containing SNPs rs6327, rs6328, rs6329, and rs6331 was amplified with the primers 5'-GAAGAGGTGCCTGGACTAAG-3' and 5'-GAGTTCTCTGAAGCTTAAGG-3' and had a length of 333 bp, PCR fragment B containing SNPs rs6325 and rs6330 was amplified with the 5'-CAGGTGCATAGCGTAATGTC-3' and 5'-GCTCTGCGAAGGGCAGTGTC-3' and had a length of 167 bp, PCR fragment C containing SNPs rs12254 and rs8453 was amplified with the primers 5'-CAGGGGAATGGATGGGTCTG-3' and 5'-GCGCTAGGTGCTAGTGTCTC-3' and had a length of 291 bp.
Amplification of PCR fragments A, B, and C was performed in a 50-µl volume using a 9700 thermal cycler (Perkin-Elmer, Weiterstadt, Germany). Initial denaturation for 3 min at 95 C was followed by 30 cycles of 1 min at 95 C, 1 min at 58 C (for PCR fragment 1) 60 C (for PCR fragment B and C), and 1 min at 72 C. Amplification was terminated by a 8-min extension at 72 C. Sequencing of the three fragments for SNP genotype determination was performed as described above.
Haplotype assembly and estimation of the time of origin of the mutation.
Haplotypes were assembled from the SNP genotyping data of the three affected and eight unaffected families by an inferential method (20, 10). To estimate the evolutionary age of the haplotypes, we used a previously described method (22). Briefly the linkage disequilibrium measure 
between the TSHß 313T mutation and each of the microsatellite markers D1S2881 and D1S467 were calculated as = (Pd - Pn)/(1 - Pn), with Pd being the frequency of the ancestral microsatellite allele among the carriers of the mutation and Pn being the frequency of the same microsatellite allele on unmutated haplotypes. Given the genetic distance between the two markers in the Généthon and Marshfield genetic maps, the age of the mutation in generations (G) was calculated according to G = log /log (1 - 
).
Results
Families and genotyping results of affected children and their relatives
The three affected families from Germany included in our studies lived in different geographic regions of Germany. Family history gave no indication of a common ancestor. Genotyping revealed homozygosity for the 313T allele in all affected children from the families. All parents were heterozygous carriers of the allele. The result of the genotyping analysis for family C is shown exemplarily in Fig. 1.
Screening for the 313T allele in a 500-person control group
The screening of 500 individuals from the general population by the same method as described above did not reveal the presence of any 313T allele. The probability not to find a randomly distributed allele with an allele frequency of 1/x when screening n alleles from the general population is P = [(x - 1)/x]n. Resolving this formula for x and accepting P to be 0.05 the absence of TSHß 313T in 1000 alleles suggests that its allele frequency in the general population is less than 1:333 and the heterozygote carrier frequency is less than 1:167 with 95% (1-P) probability.
Given these data, the probability of a random mating of two heterozygotes in the population is less than 1:25,000 and the probability of an homozygously affected newborn is less than 1:100,000 corresponding to less than 10 cases per year in Germany given a total of 700,000 annual live births. Therefore, among the investigated German population, the TSHß 313T mutation does not seem to have significant population prevalence.
Haplotype analysis of SNPs in the TSHß locus
We genotyped 30 individuals from the three affected families and another 19 individuals from six unaffected control families with identical ethnic background for eight selected SNPs from the TSHß region. Detailed characteristics of gene and marker positions in the region are given in Table 1. In this sample of 49 individuals, five SNPs showed significant heterozygosity (rs12254: het = 0.27; rs6330: het = 0.33; rs6329: het = 0.11; rs6328: het = 0.46; rs6327: het = 0.33), and three SNPs did not show any variation and were, therefore, not included in further analysis.
Haplotype assembly revealed the presence of seven different haplotypes (1, 2, 3, 4, 5, 6, 7) in nine investigated families (Table 2). Family structure and haplotypes for the three affected families are shown in Fig. 1. Overall, the most frequent haplotype observed was haplotype 1, which had a prevalence of 44% (31 of 71) among unmutated alleles from all nine families (Tables 2 and 3). In all six parental lines of the three affected families, the TSHß 313T mutation occurred exclusively on this haplotype. The mutant SNP haplotype was referred to as haplotype 1*.
|
|
Extending the SNP haplotypes by two flanking microsatellite markers with 0.8-cM interval spacing according to the Généthon map revealed some allelic heterogeneity of the mutant 1* haplotypes from the six investigated family lines (Table 3). In three of the pedigrees, both lines of family C and the paternal line of family B, the mutation occurred on the microsatellite haplotype 1*a in a total of 14 alleles. The corresponding nonmutant haplotype 1a was observed in only 1 of 31 alleles. In the three other family trees, the mutation occurred on three different microsatellite haplotypes (1*b, 1*c, 1*d) in a total of 13 alleles with at least one of the two microsatellite alleles altered, compared with haplotype 1*a. Only one corresponding nonmutant haplotype (1d) was observed once in 31 alleles. Within each family line, microsatellite haplotype heterogeneity was not observed.
Using the combined microsatellite-SNP haplotype less than 3% (2 of 71) of the investigated nonmutant alleles share a haplotype identical with the mutant haplotypes in all markers except the TSHß 313T mutation.
Estimation of the age of the mutation
The ancestral microsatellite alleles occurring together with the TSHß 313T mutation were 167 for the D1S467 polymorphism and 213 for the D1S2881 polymorphism. D1S467 allele 167 occurred on 19 of 27 mutant alleles (Pd = 0.70), but it was observed in only 31 of the 71 normal alleles (Pn = 0.44). D1S2881 allele 213 also occurred on 19 of 27 mutant alleles (Pd = 0.70), and it was observed in 14 of the 71 normal alleles (Pn = 0.20). The linkage disequilibrium measure between the TSHß 313T mutation and each of the microsatellite markers D1S2881 was calculated as = 0.625 (D1S2881) and = 0.460 (D1S467). The recombination fraction was determined from the genetic positions and the distances of the two markers, which differed slightly between the Généthon and Marshfield genetic maps. For all calculations the genetic distance between the TSHß 313T mutation and the marker on either side was assumed to be half the genetic distance between both markers. The estimate of the TSHß 313T mutational age was between 117 and 194 generations using the Généthon map data and between 177 and 293 generations using the Marshfield map data. Because the Généthon genetic map is based on a larger number of recombinants, its estimates are supposedly more accurate. According to the genomic map, the physical distance between the marker D1S467 and the TSHß gene is smaller (367 kb) than between the TSHß gene and the marker D1S2881 (1226 kb). The assumption of genetic equidistance between the gene and both markers may not be justified, but more detailed genetic distance data are not available. The approximate genetic age of the TSHß 313T mutation given the Généthon genetic map can be estimated to be about 150 generations or approximately 3700 yr. Because mutation rates of the microsatellite markers were not taken into account, these results tend to overestimate genetic age.
Discussion
This study investigated the prevalence and genetic origin of the TSHß 313T mutation leading to autosomal recessive congenital central hypothyroidism. The relatively frequent diagnosis of the condition due to this single mutation in the last 3 yr is noteworthy because the recognized incidence of isolated TSH deficiency had been very low in previous times. This fact may most likely be due to a missed diagnosis of the condition until recently. Given the now apparent importance of the 313T mutation in congenital hypothyroidism, questions arose as to whether the observed 313T mutations were due to independent de novo mutational events presuming a potential hot spot at this site or to a common founder. The answer to this question is of high importance because in the case of a hot spot, the introduction of an additional neonatal T4 screening might be advisable.
The history of three German families affected with the TSHß 313T mutation has not revealed any consanguinity among family members. Analysis of SNP haplotypes around the TSHß locus revealed the presence of at least seven haplotypes in the general population. The haplotypes could be fitted into a cladistic model assuming only one mutational step between each haplotype. Haplotypes could be divided into two clades (haplotypes 1–5 and haplotypes 6–7). The differences between both clades were two mutational steps at SNP loci rs6327 and rs6330 with none of the intermediate haplotypes observed. For each of the haplotypes 1, 3, and 6, a corresponding haplotype with the opposite stated allele at the SNP locus rs12257 in the NRAS gene, 382 kb distant from the NGFß locus was present. This observation is more likely explained by recombination events between the NGFß and NRAS loci than by independent mutations. The haplotypes 1* and 5 do not have corresponding haplotypes differing in the rs12257 locus, suggesting their relatively younger evolutionary age without sufficient time for recombination.
All mutant TSHß 313T alleles occurred on haplotype 1. Although this haplotype was the most frequent in the general population with a frequency of 0.4, the independent occurrence of the mutation in all six parental lines on this haplotype by chance alone would be unlikely. The haplotype analysis data, therefore, suggest a monophyletic origin of the TSHß 313T mutation and a common founder.
While this study was performed, three more families with TSHß 313T homozygous children have been identified with one living in Germany, one in Brazil, and one in Belgium. This observation of local accumulation also in these cases supports the possibility of a common ancestor of all families. The estimated dating of the last common ancestor between the members of our six family lines about 150 generations ago fits with the migrational hypothesis of mutation carriers found in different geographic regions because of migration. Haplotype investigation of members from the other families will be necessary to further substantiate this result.
The low population frequency of the TSHß 313T resulting in the statistical probability of less than 10 affected newborns per 700,000 births per year is well in concordance with the observations from our own center. We observe 30,000 newborns every year, resulting in a total of 600,000 newborns over the last 20 yr. Among these children, only one case with congenital central hypothyroidism because of a TSHß mutation was diagnosed.
Considering the monophyletic origin of the TSHß 313T mutation and its low population frequency with no evidence of a higher than normal de novo mutation rate, the identification and genetic counseling of heterozygous carriers from affected families seems more advisable than population-wide neonatal T4 screening.
Acknowledgments
We sincerely thank Jasmin Denhardt from TIB Molbiol Berlin for providing labeled DNA standards and for her appreciated assistance during microsatellite analysis.
Footnotes
Abbreviations: fT3, Free T3; fT4, free T4; NGFß, nerve growth factor ß-subunit; NRAS, n-ras oncogene; SNP, single-nucleotide polymorphism.
H.Br. and A.P. contributed equally to the work described in the manuscript.
This work was supported by a scholarship from the Deutsche Gesellschaft für Endokrinologie (to H.Br.) and a Bundesministerium für Bildung und Forschung grant (FK 01GS0109) within the NGFN network (to A.P.).
Received February 25, 2002.
Accepted June 26, 2002.
References
-subunit, and variable circulating thyrotropin (TSH) levels as hallmark of central hypothyroidism due to mutations of the TSH ß gene. J Clin Endocrinol Metab 86:1600–1604This article has been cited by other articles:
 |
|
 |
|
  G. Borck, A. K. Topaloglu, E. Korsch, U. Martine, G. Wildhardt, N. Onenli-Mungan, B. Yuksel, U. Aumann, G. Koch, G. Ozer, et al. Four New Cases of Congenital Secondary Hypothyroidism due to a Splice Site Mutation in the Thyrotropin-{beta} Gene: Phenotypic Variability and Founder Effect J. Clin. Endocrinol. Metab., August 1, 2004; 89(8): 4136 - 4141. [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 |
