The Coexistence of a Novel Inactivating Mutant Thyrotropin Receptor Allele with Two Thyroid Peroxidase Mutations: A Genotype-Phenotype Correlation

The presence of a TPO gene mutation in one allele does not aggravate the mild hyperthyrotropinemia of TSHR gene mutation present in the other.

Abstract

Context:

TSH receptor (TSHR) and thyroid peroxidase (TPO) gene mutations occur independently. This is the first report of their coexistence in the same individuals.

Objectives:

The objective of the study was to evaluate the genotype-phenotype correlations when mutations in both genes are present alone or together in the same individual.

Patients and Methods:

Thirty subjects from an extended Arab kindred underwent clinical investigation and molecular studies of the mutant TSHRs.

Results:

A novel mutant TSHR was identified, involving four nucleotides at three sites on the same allele, c.267G>T (L89L), c.269/270AG>CT (Q90P), and c.790C>T (P264S). In addition, two known TPO gene mutations, G493S and R540X, were identified. Thirteen heterozygotes for the mutant TSHR allele had mild hyperthyrotropinemia. In nine of theses, the coexistence of a TPO mutation in one allele did not magnify the hyperthyrotropinemia. Homozygotes for the mutant TSHR and a compound heterozygote for the TPO mutations presented frank hypothyroidism. In vitro studies showed increasing loss of function for Q90P less than P264S less than Q90P/P264S TSHR mutants, the latter being that expressed in the subjects under investigation. The two interchangeably used WT TSHR vectors, L87 and V87, although functionally identical, differed in structure and function in the presence of the Q90P mutation.

Conclusions:

TSHR and TPO gene mutations were identified alone and together in individuals of a consanguineous kindred. Homozygotes for the TSHR and a compound heterozygote for the TPO mutations were hypothyroid. The mild hyperthyrotropinemia of heterozygotes for the mutant TSHR allele was not aggravated by the coexistence of a TPO defect in one allele.

Resistance to TSH (RTSH) is a syndrome of reduced sensitivity to TSH. The defect is characterized by elevated serum TSH concentration, normal or hypoplastic thyroid gland, and normal to very low levels of thyroid hormones (1). Several gene defects can cause RTSH including loss-of-function mutations in TSHR, PAX8, and GNAS genes. An autosomal dominant form of RTSH has been linked to a locus on chromosome 15q25.3–26.1 (2, 3).

Inactivating mutations in the TSHR gene are recessively inherited (1), although some heterozygotes manifest mild TSH elevations (4). The spectrum of phenotypes, ranging from euthyroid hyperthyrotropinemia to severe congenital hypothyroidism (CH) is dependent on the magnitude of the functional impairment of the mutant TSH receptor (TSHR) (5).

Inheritance of thyroid peroxidase (TPO) gene defects is recessive with clinical manifestations of CH and goiter. Rapid radioiodine uptake with a significant discharge after perchlorate administration is consistent with iodide organification defect (6).

This study is the first to report the coexistence of TPO and TSHR mutations in various combinations in the same individuals of a consanguineous kindred. It allowed assessing the genotype-phenotype correlation when mutations in both genes are present in isolation and combined in the same individual. In vitro functional studies of the various components of the mutant TSHR and their correlation to structure and phenotype are also provided.

Subjects and Methods

Thirty family members of an Arab kindred living in northern Israel were studied and were approved by the Institutional Review Boards of Ha'Emek Medical Center and the University of Chicago. Index cases belonging to three nuclear families had CH described in Supplemental Data, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org.

Thyroid function test and gene sequencing

Details can be found in the Supplemental Data.

In vitro analyses

Functional analysis of the mutant TSHR was performed using expression vectors of each mutant component as well as the complete allele and compared with the wild-type (WT) TSHR. Construction and origin of all vectors are described in Supplemental Data. Human embryonic kidney-293 cells at 50–80% confluence were cotransfected with 0.5 μg reporter plasmid, 10 ng pRL-Tk internal control vector, and 150 ng effector plasmid/empty vector using FuGene6 (Roche, Indianapolis, IN). After 5 h, fresh medium containing none or various amounts of bovine TSH (Sigma, St. Louis, MO) was added, and cells were harvested 48 h later for sequential analysis of firefly and Renilla luciferase activities (dual luciferase reporter assay system; Promega, Madison, WI). The ratios between the measured firefly and Renilla luciferase activities or relative light units (RLU)(CRE-Luc) were calculated. All experiments were performed in triplicate.

Statistical analyses

Data expressed as mean ± sem were analyzed by two-tailed Student's t test for unpaired observations and by ANOVA for comparison among three or more groups.

TSHR modeling

The molecular model of the TSHR-TSH complex was assembled as described in Supplemental Data.

Results

Genetic results

A novel mutant TSHR allele was identified, involving four nucleotide substitutions, c.267G>T, c.269/270AG>CT, and c.790C>T. They produced, respectively, a silent mutation (L89L), replacement of the normal glutamine 90 with a proline (Q90P), and replacement of the normal proline 246 with a serine (P264S). In addition, we identified two TPO gene mutations (G493S and R540X), previously found in the same geographic region (7).

The mean age of the 30 subjects studied was 25.35 ± 3.27 yr (range 0.4–57), 17 of whom were female. As shown in Fig. 1A, 15 had at least one mutant TSHR allele. Because the CH index case of family 3 (subject III-8) had no mutant TSHR allele, other gene defects were sought. He was found to be compound heterozygous for the TPO mutations G493S and R540X, whereas 18 others had one of these TPO mutations.

Fig. 1.

A, Pedigree of the three nuclear families contained in a larger kindred with TSHR and/or TPO gene mutations. Square symbols indicate males, circles females, and lozenges any gender. The genotype of each subject is indicated within the symbol, as identified in the legend below the pedigree. Results of TSH and FT₄I are aligned with each symbol and were obtained on no hormonal replacement. Roman numerals to the left of the pedigree indicate the generation, and numerals on the right of each symbol indicate individual family member. Double lines across spouses indicate consanguinity. Asterisk indicates the presence of circulating TG and/or TPO antibodies. B, Serum TSH concentrations in individual subjects grouped according to the genotype. TPO mut1 is G493S and TPO mut2 is R540X. NS, not significant, represents P ≥ 0.05. C, Serum TSH and FT₄I concentrations in WT, heterozygotes, and homozygotes for the TSHR gene mutation irrespective of TPO gene mutations. Subject III-8, compound heterozygote for the TPO gene mutations, was not included in the figure. Data were expressed as mean ± sem.

Of the 13 individuals heterozygous for the mutant TSHR allele, nine were also heterozygous for one of the two TPO mutations. Of the 17 individuals heterozygous for one of the two TPO mutations, eight had normal TSHR alleles. Of the four individuals on L-T₄, two were homozygous for the mutant TSHR (II-3 and III-6), one was compound heterozygous for the TPO mutations (III-8), and one was heterozygous for both the TSHR and TPO mutations (I-2) (see Fig. 1A).

Thyroid function tests

Individual serum TSH concentrations are plotted in Fig. 1B. Genotype-phenotype correlation based on mean TSH concentrations revealed no significant difference between heterozygotes for the two TPO mutations (G493S, n = 4; TSH 3.60 ± 0.61 mU/liter; and R540X, n = 4; TSH 2.75 ± 0.23 mU/liter) or between all heterozygotes for the TPO mutations (n = 8; TSH 3.24 ± 0.40 mU/liter) and unaffected (WT) family members (n = 6; TSH 2.22 ± 0.35 mU/liter). Also, TSH concentrations of heterozygotes for the mutant TSHR without TPO mutation (n = 4; TSH 5.73 ± 1.59 mU/liter) and with TPO mutation in one allele (n = 9; TSH 8.13 ± 1.42 mU/liter) were not significantly different. However, the heterozygotes for the mutant TSHR had higher (P < 0.01) mean TSH concentration than all groups without mutant TSHR allele, except the compound heterozygote for the two TPO defects. The statistical significance persisted with exclusion of the four subjects with positive thyroglobulin (TG) and TPO antibodies (I-1, I-6, I-8, and II-14). Accordingly, data were reanalyzed by pooling WT and TPO heterozygotes (n = 14) as well as those with mutant TSHR alleles irrespective of whether they were also heterozygous for a TPO mutation (n = 13) (Fig. 1C).

All subject heterozygotes for the mutant TSHR had significantly (P < 0.0001) higher mean serum TSH concentration (n = 13; 7.39 ± 1.11 mU/liter) than that of WT (n = 14; 2.80 ± 0.30 mU/liter). With one exception (subject II-10), heterozygotes for the mutant TSHR had free T₄ index (FT₄I) values in the normal range but a mean FT₄I of 8.23 ± 0.35 that was significantly (P < 0.01) lower than the WT of 9.54 ± 0.26 (Fig. 1C). All other tests results, including total T₃, free T₃ index, total rT₃, and TG were not significantly different among the groups (results not shown).

The two homozygotes for the mutant TSHR and one compound heterozygote for the TPO mutations had, when untreated, TSH level of 39.7, 58, and greater than 75 mU/liter, respectively, and FT₄I of 5.1, 2.8, and less than 2, all below normal.

Functional characterization of mutant TSHR

Functional activity of the TSHR was determined in terms of cAMP production by measuring the RLU generation by the reporter CRE-Luc at baseline and after TSH stimulation. To compare data from separate experiments, RLU was expressed relative to the maximum TSH response of the WT TSHR in each experiment (Fig. 2A).

Fig. 2.

In vitro functional analysis of TSHR. A, Analysis of the Q90P, P264S, and Q90P/P264S TSHR compared with WT and P162A TSHR. All constructs were in the WT background of V87, including the silent mutation, L89L, to simulate the precise nucleotide sequences of family members TSHR. The reporter gene pCRE-Luc is driven by cAMP. Human embryonic kidney-293 cells transfected with the empty pSVL vector served to determine the background luciferase activity. Results are expressed relative to those obtained with WT TSHR. B, TSH induced cAMP response of P264S in comparison with Q90P/P264S TSHR. Results are from a single experiment using six replicate transfection incubated with 10 and 100 mU/ml TSH. C, Functional activity of transfections simulating the different genotypes observed in humans. Error bars represent sem. All transfections contained the same amount of total TSHR vector DNA so that in simulating the heterozygous state, there was half of the amount of the WT and half of the Q90P/P264S vectors, a total same as homozygous of each vector.

WT TSHR showed the highest activity compared with all other TSHR variants. The mutant Q90P TSHR showed lower activity compared with the WT TSHR but higher than the mutant P264S and Q90P/P264S. Similar to the WT TSHR, the response of the Q90P TSHR reached a plateau at 10 mU/ml TSH, albeit 55% lower than the WT. The response of the P264S TSHR alone or in combination with the Q90P TSHR was approximately 10% the baseline. In contrast, the mutant P162A TSHR, described previously (8), although requiring more TSH for a given cAMP response, reached a maximal activity similar to that of the WT TSHR, suggesting that the mutant Q90P and P264S have relatively lower cell surface expression.

To determine whether the presence of Q90P in the combined Q90P/P264S mutant affected the magnitude of loss of function of P264S TSHR, we compared their relative response to TSH in the same experiment. The activity of the P264S TSHR mutant was significantly higher than that of the Q90P/P264S TSHR at TSH concentrations of 10 and 100 mU/ml (Fig. 2B). Also, Q90P/P264S TSHR showed a greater response at TSH concentration of 10 than 100 mU/ml, whereas that of P264S TSHR was maximal at 10 mU/ml TSH.

To simulate the heterozygous state of the subjects, we cotransfected the WT TSHR with the Q90P/P264S TSHR mutant in equal amount. Response to TSH was intermediate to that of the WT TSHR and Q90P/P264 TSHR alone, the latter simulating the homozygous state (Fig. 2C).

Because two different WT TSHR clones differing in amino acid 87, leucine (9), or valine (10) have been used by many laboratories, we studied their effect without and with the mutations. Results are detailed in Supplemental Data.

Discussion

We present a large consanguineous Arab kindred containing three nuclear families each with CH but no goiter. The three novel nucleotide substitutions, located in the same TSHR allele, resulted in a silent mutation (L89L) and two missense mutations, in exon 3 (Q90P) and in exon 9 (P264S). It is likely that these substitutions represent successive rather than concomitant events. We also identified two previously described TPO gene mutations, G493S and R540X (7).

This is the first report on the coexistence of TSHR and TPO gene mutations in the same family and individuals. Various combinations of different mutations gave rise to eight genotypes but only three TSH phenotypes. The TSH concentrations of the WT and heterozygotes for the mutant TSHR were not affected by the presence of a TPO defect in one allele. Because the TPO promoter is cAMP and TSH responsive (11) and the mutant TSHR lowered cAMP production, we hypothesized that heterozygotes for both TSHR and TPO mutations should have higher TSH than heterozygotes for the TSHR mutation alone. This did not occur, presumably because the amount of TPO synthesized from a single functional allele is still optimal for thyroid hormone synthesis. Unfortunately, we were unable to perform an in vivo evaluation of the TPO enzymatic activity using the perchlorate discharge test.

When tested in transfection studies, TSHR constructs containing each of the two amino acid substitutions, located in the extracellular domain of the molecule, showed impaired cAMP production in response to TSH. Cells expressing the mutant Q90P TSHR had lower cAMP production than the WT TSHR but higher than those expressing the mutant P264S and Q90P/P264S. All reached the plateau response at 10 mU/ml TSH, lower than that of the WT TSHR. In contrast, the previously described P162A TSHR (8) achieved a maximal cAMP as did the WT TSHR, albeit at higher TSH concentration. Heterozygotes for P162A TSHR had milder phenotype than heterozygotes for Q90P/P264S. This observation is compatible with a reduced cell surface expression of the mutant Q90P and P264S TSHR.

Transfections simulating the heterozygous state of the subjects harboring the mutant TSHR showed intermediate response between the WT TSHR and the mutant alone. Results were in agreement with the phenotype of subclinical hypothyroidism in heterozygotes. Their partial RTSH may be due to reduced surface expression of the mutant TSHR as well as to dominant effect through dimerization. Both phenomena have been observed in vitro with mutant TSHR (12). Finally, as shown in the Supplemental Data, caution should be exerted in the interchangeable use of the two variant clones of the WT TSHR, containing leucine or valine 87 because another amino acid change in the vicinity or interacting with codon 87 may have a significant functional effect, affecting the interpretation of data.

In conclusion, we identified mutations of TSHR and TPO genes in the same individuals of a consanguineous kindred. This is an interesting in vivo demonstration of oligogenic defects in the same functional pathway. Yet the mild hyperthyrotropinemia of heterozygotes for the mutant TSHR was not aggravated by the coexistence of a TPO defect in only one allele. It is possible that more stringent environmental exposures for this functional pathway, such as mild iodine deficiency, can stretch the limited resources of the system and manifest as overt hypothyroidism.