| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Division of Endocrinology and Diabetes, Department of Medicine, University of Minnesota, Minneapolis, Minnesota 55455
Address correspondence and requests for reprints to: Jack H. Oppenheimer, M.D., University of Minnesota, Department of Medicine, Division of Endocrinology and Diabetes, 6-124-PWB, 516 Delaware Street SE, Box 101, Minneapolis, Minnesota 55455
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The clinical features of the THR syndrome (3) and its molecular basis have been extensively reviewed (4, 5). The mutated receptor has been implicated in delayed bone and CNS maturation in some patients. In such patients, RTH in these peripheral tissues is not adequately compensated by an appropriate increase in thyroid hormone secretion, even though there is reduced suppression of TSH secretion at the pituitary and hypothalamic level. Other patients exhibit symptoms of peripheral hyperthyroidism because THR in the central pituitary-hypothalamic area exceeds that observed at the peripheral level. Theoretically, if resistance at the central level were perfectly matched to peripheral resistance, one would not expect any clinical manifestations of the disorder, with the exception of the thyroid enlargement accompanying overproduction of TSH. These considerations may well explain the broad diversity of responses associated with thyroid hormone receptor mutations.
From a practical point of view, suspicion of RTH rests on the identification of patients with goiter or a history of goiter who exhibit an inappropriately high TSH value for a given value of the free T4 index (FTI). We here present data suggesting that the degree of THR at the central level can be assessed, at least in some patients, from the slope of the natural logarithm of serum TSH (lnTSH)/FTI ratio, as generated by incremental changes in the daily dose of L-T4 administered in the course of clinical management designed to define the optimal treatment dose of L-T4. In the three mutations studied, the degree of THR assessed in vivo correlates with the T3 affinity of the in vitro translated receptor.
The present studies were initiated independently of earlier reports by Hayashi et al. (6) and Yagi et al. (7). Hayashi calculated in vitro resistance on the basis of the relationship between the increment in free T4 induced by the injection TRH and the association constant determined in vitro. In these studies, the authors noted that, in 12 mutations studied, there was a clear correlation between the association constant determined in vitro and the in vivo measurements; but no such correlation could be established in 6 other mutations.
In the present study, we have also explored the feasibility of assessing the in vivo association constant from the slope of the lnTSH as a function of the FTI.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patients were under the care of members of the Division of Endocrinology and Diabetes. Written permission for the performance of these studies and the genotyping of the suspected TRß mutation was provided by all subjects, in line with the policies adopted by the Human Research Committee of the University of Minnesota.
The T4 index used in this study is the product of the total serum T4 multiplied by a serum-binding correction factor, as determined by standard techniques (8). The product of the free T3 fraction multiplied by the total serum T4 concentration generated a value that we have shown to be proportional to the level free T4 (Sandhoffer, C. R., J. H. Oppenheimer, C. N. Mariash, and D. Brown, unpublished observations). Twenty-four-hour Holter monitoring and evaluation of cardiac parameters were obtained with the assistance of the Division of Cardiology. Achilles-tendon relaxation time was measured using a kinemometer (10).
Patient histories
Patient A1 (index patient). Patient A1, a 24-yr-old graduate
student, was referred to us in 1994 for evaluation of her thyroid
status. She indicated that in 1987 she developed tachycardia and
thyroid enlargement and exhibited elevated levels of thyroid hormone,
features which led her physicians, at the time, to the diagnosis of
hyperthyroidism, presumably as a result of Graves’ disease. Treatment
with radioiodine led to alleviation of her symptoms but resulted in the
subsequent development of hypothyroidism. The specific issue that
prompted her referral to us in 1994 was the difficulty in establishing
a L-T4 replacement regime that would
result in the simultaneous normalization of serum thyroid hormones and
TSH. These findings raised the possibility of RTH. A normal magnetic
resonance imaging of the pituitary gland and a normal level of the
-subunit of pituitary glycoprotein excluded the possibility of a
pituitary TSH-producing tumor. Initially, her heart rate was noted to
be mildly elevated; but after discontinuation of nortriptyline, which
had been prescribed for mood disorder, her heart rate normalized, as
confirmed by Holter monitoring. Achilles-tendon relaxation time, serum
lipid values, sex hormone-binding globulin, urinary excretion of
pyridinoline, and deoxypyridinoline were normal. These negative
findings, together with persistent elevations of TSH levels for given
levels of serum thyroid hormones, led to the diagnosis of THR.
Patient A2. This patient is the father of Patient A1. He was 58 yr old at the time of these studies. At age 24, Patient A2 developed symptoms and signs that were attributed to an attack of so-called thyroid storm. This was successfully treated with radioiodine. Medical records, documenting presenting symptoms and signs, are no longer available. The patient is currently maintained on 200 µg L-T4 per day. Although blood samples were provided, the patient declined further evaluation.
Patient B1. This patient was first seen at the University of Minnesota in 1986, at age 46, for an evaluation of her thyroid status. In 1971, she was treated with T3 tablets, presumably to shrink her goiter. T3 was stopped in 1981 and L-T4 substituted. In 1985, she was treated with 6 mCi 131I, to shrink her goiter. Subsequently, her TSH was noted to be more than 64 mIU/L and her serum T4 to be 9.0 µg/dL, suggesting the likelihood of THR. Her last set of thyroid function tests in 1986 revealed a serum TSH of 5.8 mIU/L and a T4 index of 21.8.
Patient B2. The daughter of Patient B1 was referred to the University of Minnesota by her mother in 1986 at age 24. She had also been treated with radioiodine for symptoms and signs presumed to be attributable to primary hyperthyroidism. Subsequently, she developed signs, symptoms, and laboratory values of hypothyroidism and was started on thyroid replacement therapy. In April 1986, on an unknown dose of L-T4, she was found to have a high FTI of 17.5 and an elevated TSH of 30 mIU/L.
Patient C1. This patient has been previously reported by Bantle et al. (11) as Case 1.
Patients presumed to exhibit normal sensitivity
Our initial objective in these studies was to compare the relationship between the TSH level and the FTI of patients diagnosed as THR, to that of patients presumed to exhibit normal thyroid hormone sensitivity. We have previously analyzed the relationship between serum TSH and the serum free T4 indices in 157 patients with the following disorders: 1) hyperthyroidism in response to treatment; 2) hypothyroidism treated with progressive increases in L-T4 dose; and 3) thyroid cancer under treatment with L-T4. Because we did not identify, in these groups, any indications of THR, we used the slope measurements in this control population as a reference for quantitating thyroid hormone sensitivity in patients at the pituitary-hypothalamic level.
Molecular biology techniques
Genomic sequencing. Genomic DNA was isolated from peripheral leukocytes using the Wizard (Promega Corp., Madison, WI) peripheral genomic DNA isolation system. Using specific synthesized oligomeric primers (12), TRß exons 9 and 10 were PCR-amplified (Expand PCR System, Roche Molecular Biochemicals, Indianapolis, IN) using 1 µg of genomic DNA from each patient (denaturation, 94.0 C for 10 sec; annealing, 55.0 C for 30 sec; elongation, 68.0 C for 2 min for 30 cycles). Controls in which polymerase or genomic DNA was omitted were included. PCR products were individually gel-isolated [4% NuSeive Agarose (FMC, Rockland, ME) in 0.5 x Tris-Acetate-EDTA], extracted using QIAquick Gel Extraction Kit (QIAGEN, Chatsworth, CA) and eluted in 50 µl Tris-EDTA. One microliter of the eluted product was reamplified using the same primers and was gel-isolated as above. The exons were initially sequenced using the Sequenase 2.0 (USB, Cleveland, OH) system with separation on a 5% polyacrylamide denaturing gel and later sequenced using the ABI/Prism Dye-Terminator labeling system with separation on an 310 Genetic Analyzer (Perkin-Elmer Corp., Foster City, CA). Each exon was sequenced completely in both directions, from at least two independent PCR products.
T3 Binding affinity: of wild-type and mutant receptors. The complementary DNA (cDNA) coding for human TRß1 was subcloned from pHE-A1 (gift from Dr. William Chin) into the multiple cloning site of pcDNAI (Invitrogen, San Diego, CA). Both the E460K and A317T mutations were made using QuickChange (Promega Corp., La Jolla, CA) site-directed mutagenesis kit. The presence of the expected mutation was confirmed using the direct sequencing with the ABI/Prism system and the SP6 primer for the E460K mutation or a specific cDNA primer sequence (CCAGAGTGGTGG) for the A317T mutation. Wild-type and mutant thyroid hormone receptors were generated by in vitro transcription (13) and translation (14), and T3-binding affinity was assessed by filter-binding studies using labeled T3 (15).
Transfection studies. A 1.3-Kb rat ßTSH subunit promoter, linked to a luciferase reporter, was kindly provided by Dr. Margaret Shupnik (16). JEG-3 or Neuro-2a (ATCC catalog no. HTB36 and CCL no. 131, respectively) cells were divided to a concentration of approximately 2 million cells/100-cm2 plate, 24 h before transfection using standard medium (DMEM with penicillin/streptomycin) containing FBS that had been stripped of endogenous thyroid hormone (17). Four hours after the incubation media had been refreshed, 10 mg of reporter, 2 mg of a TK-CAT (17) transfection efficiency control, and 10 mg of either wild-type, E460K, or A317T TRß1 expression plasmid or the empty pcDNAI vector were cotransfected using the calcium-phosphate transfection technique (Life Technologies/BRL, Gaithersburg, MD). Eighteen hours later, media were changed to either media with stripped FBS (-T3) or media in which T3 was added back (+T3) to a final concentration of 100 nmol/L. After an additional 24 h of incubation, cells were harvested in lysis buffer with 1 mmol/L fresh dithiothreitol (17) and luciferase (17) and CAT activity (18) subsequently assayed.
During the course of our work, Maia et al. (19) reported a potential cryptic element within the luciferase cDNA that mediated receptor-dependent inhibition of luciferase activity in a JEG-3, COS, and CV-1 cell-line. However, in the Neuro-2a cells used in these experiments, we demonstrated that such negative regulation of luciferase expression failed to occur.
Statistics. Data were analyzed using Systat 5.2.1 (SPSS, Inc., Chicago, IL). The relationship between lnTSH and T4I in patients and controls was evaluated by regression analysis. Comparisons between patient groups and transfection groups were made by ANOVA. Significance between groups was calculated by post hoc testing using the Bonferroni adjustment.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Sequencing of exons 9 and 10 from genomic DNA, using dideoxysequencing, confirmed the presence of a single heterozygous base-pair substitution in each family. In unrelated families A and B, the mutation resulted in a substitution of lysine for glutamic acid at amino acid position 460 (E460K). This mutation has been previously reported (20). In family C, a threonine was substituted for alanine at amino acid 317 (A317T). This mutation has also been previously reported (21).
Comparison of lnTSH/T4 index relationship in patients with THR to patients presumed to exhibit normal pituitary-hypothalamic sensitivity to thyroid hormone
We have previously described the TSH/FT4I relationships in three clinical groups: 1) patients under treatment for Graves’; 2) patients with spontaneous hypothyroidism who were also under treatment with L-T4; and 3) patients with thyroid cancer who had undergone total thyroidectomy and were at various stages of incremental replacement with L-T4. Although the average daily dose of L-T4 varied between groups that were on thyroid hormone therapy, we did not detect any difference in the TSH/FT4I relationships in these groups.
To exclude the possibility that our previously published analysis of
the TSH/FT4I ratio (22) was distorted by the
inclusion of patients who had been sampled only at a single time point
or had not been challenged by a sufficient increment in
L-T4, we recalculated this data base
by omitting isolated values as well as data points of individuals who
had not been challenged with at least a 2-fold increase in
T4 dose. Results of this recalculation are
depicted in Fig. 1
and confirm the
results of previous studies of the TSH/FT4
relationship (22, 23), which had pointed to an exponential decrease in
TSH with progressive increases in serum T4 in the
ranges studied.
|
demonstrates that TSH values
corresponding to serum T4I values 
5.0
were well represented by the exponential function.
Thus, TSHT4I = TSH5.0
e-(T4I)
, where TSHT4I is
the concentration of TSH at a given level of serum
T4I, and TSH5.0 is the
concentration of TSH when the T4I is 5.0. The
function can be regarded as a measure of the sensitivity of
pituitary TSH suppression to the ambient level of thyroid hormone. It
is apparent, from this equation, that the larger the value of , the
steeper the slope of the lnTSH plotted as a function of
T4I. Moreover, if a hypothetical mutation of the
receptor gene completely obliterated TSH suppression by
T3, there would be no fall in the TSH level in
response to increasing levels of hormone. Under such conditions, the
value of would be 0.
The data depicted in Fig. 1 allow us to assess the slope : 1) in
patients under treatment for thyroid disease; 2) in the two unrelated
families (A) and (B) expressing the same mutation (E460K); and 3) in
patient C1with the mutation A317T. The linear regression was
significant for all three groups of patients (P <
0.023). The regression coefficient for the normal controls was 0.55;
for the patients with the E460K mutation, 0.76; and for the patient
with the A317T mutation, 0.78. Post hoc analysis showed that
the controls differ from each of the mutations, with P
< 0.000; whereas the mutations differ from each other, with
P = 0.049. The slopes of the two unrelated patients
with the same mutation did not differ significantly from each other. As
would be expected, the slope of the curve generated by patients with
RTH is less than those who are assumed to be normally sensitive to
thyroid hormone. When all of the patients with E460K mutation are
combined, the ratio of
(normal)/(E460K
resistant) is approximately 2.5.
Receptor affinity for T3
The finding of statistically significant differences in the slopes
of the TSH/T4I relationships among all three
groups poses a problem as to the molecular mechanisms responsible for
the observed diversity. One possibility is that differences in slope
could reflect differences among these receptors in their ability to
bind T3. To test this hypothesis, we compared the
T3-binding affinity of the wild-type and mutant
receptors generated by in vitro transcription and
translation. When normalized to the wild-type
T3-binding affinity (Ka),
the E460K mutation results in a binding affinity 49% of wild-type and
the A317T mutation has a binding affinity 10% of wild-type. Previous
reports had suggested values of 25% (20) and 13–22% (21),
respectively. Comparing the calculated slopes of the
lnTSH/T4I relationships with the
T3-binding affinity of the receptor for
T3 is consistent with a linear relationship (Fig. 2). Given the inherent variability in the
methods involved, these findings are consonant with the hypothesis that
the slope of the lnTSH/T4I curve is strongly
influenced by the strength of pituitary sensitivity to
T3 at the level of the thyroid hormone
receptor.
|
4)
domain, which has been shown to exhibit the potential of independent
transactivation (24). We therefore considered the possibility that this
segment might also play a ligand-independent role in stimulating target
gene expression. To test this hypothesis and to gain further insight
into the mechanisms of RTH exhibited by these receptors, we
cotransfected thyroid hormone receptors (wild-type, E460K, or A317T)
and a rat ßTSH-luciferase reporter construct into Neuro-2a
cells and measured the response of the reporter construct.
As illustrated in Fig. 3, transfection of
wild-type thyroid hormone receptor in the absence of
T3 resulted in T3-induced
reporter activity above that generated by the empty vector. In the
presence of T3, however, both the E460K and the
A317T mutated receptors transactivate the reporter activity only to
levels comparable with that of the wild-type receptor. These findings
thus did not support the hypothesis that either of the two unliganded
mutant receptors studied could superinduce ßTSH transcription above
the level attained by the unliganded wild-type receptor.
|
also shows that, in the presence
of a cotransfected thyroid hormone receptor, the transcriptional
activity of the TSH reporter was significantly decreased by thyroid
hormone in a dose-dependent manner. When the data at 1 and 10 nmol/L
were subject to ANOVA, we found that the A317T response was
significantly less than that of the wild-type receptor; and the
response of the E470K, at 10 nmol/L, was also less than that of the
wild-type. To characterize better the dose/response relationship, all
the data were fitted to a four-parameter logistic-log regression
equation to calculate the ED50 (dose that yields
50% of the maximum response) to T3 (25). We
found that the ED50 correlated very well with the
receptor-binding affinity. Thus, the wild-type receptor had the lowest
ED50 at 0.424 nmol/L T3,
the E470K mutation had an intermediate value at 0.652 nmol/L
T3, and the A317T mutation required the highest
dose of T3 at 3.6 nmol/L to suppress
transcription by 50% of the maximum response. The relative response to
each of these receptors also correlated with the in vivo
dose/response data depicted in Fig. 1. Despite the differences in
affinity for T3, all three receptors suppressed
the TSH transcriptional activity to the same level, at very high
concentrations of thyroid hormone.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
One of the requirements essential in the diagnosis of RTH is the demonstration that the pituitary shares this resistance. To the best of our knowledge, there have been no reports of patients with selective peripheral resistance at the molecular level but with normal pituitary sensitivity. A higher level of free T4 is required to achieve a given level of TSH in a patient with RTH than in a subject with normal sensitivity to thyroid hormone.
A major question, first posed by Hayashi et al. (6), is whether the reduction in the affinity of T3 binding to the receptor, as determined in vitro, correlates with the pituitary resistance as assessed in vivo. The latter was measured by the response to the incremental administration of liothyronine (T3) over a period of 3 days. They concluded that, among the mutations studied, the majority of cases showed a positive correlation between in vitro and in vivo measurements. However, a subset failed to show a significant correlation.
The data generated in Fig. 1 in the present paper suggest that the
laboratory diagnosis of pituitary RTH can also be made by assessing the
slope of the lnTSH as a function of the free T4
concentration or the FTI. Although only two mutations were studied by
us, one of which appeared in two unrelated families, there are, to the
best of our knowledge, no reports defining receptor mutations that are
clinically significant but show normal pituitary suppressibility.
Moreover, from a practical clinical point of view, administration of
graded doses of LT4 to the patient with THR is
often useful in choosing the optimal thyroid hormone dose to suit the
overall metabolic needs of the patient.
Caution should be exercised to make certain that the patient is
fully equilibrated on a given dose of
L-T4 and that the possibility of a
pituitary TSH-producing tumor is excluded by a serum measurement of the
-subunit of pituitary glycoprotein.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received April 19, 1999.
Revised July 22, 1999.
Revised October 28, 1999.
Accepted March 11, 2000.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. R. Huber, B. Sandler, B. L. West, S. T. Cunha Lima, H. T. Nguyen, J. W. Apriletti, J. D. Baxter, and R. J. Fletterick Two Resistance to Thyroid Hormone Mutants with Impaired Hormone Binding Mol. Endocrinol., April 1, 2003; 17(4): 643 - 652. [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 |
), family A with the E460K mutation (
), family B with the E460K
mutation (
), and Patient C with the A317T mutation (
). The
regression line for normal subjects with an FTI 
