| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Medicine, Rajavithi Hospital (T.S., S.N., S.K.), Department of Pediatrics, Siriraj Hospital, Mahidol University (K.A., S.L.), Bangkok, Thailand; Department of Medicine (N.H.S., S.R.) and Pediatric (S.R.) and the J. P. Kennedy Jr. Mental Retardation Research Center (S.R.), The University of Chicago, Chicago, Illinois 60637
Address correspondence and requests for reprints to: Thongkum Sunthornthepvarakul, M.D., Rajavithi Hospital, Bangkok 10400, Thailand. E-mail: thongkum{at}health.moph.go.th
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The molecular basis of FDH was identified in 1994 by us (7) and by others (8). A mutation in the HSA gene, resulting in the replacement of the normal arginine-218 with histidine (R218H), produces an HSA with 10- to 15-fold higher affinity for T4 than the normal molecule and only a 4-fold increase in the affinity for T3 (7, 9). The identical mutation was found in all 22 Caucasian families with FDH so far tested (7, 8, 10). In 1997, Wada et al. (11) reported the first Oriental family with FDH caused by a different mutation in the same amino acid, replacing the normal arginine-218 with proline (R218P). Although this variant HSA has 83-fold higher affinity for T4 than normal molecule, similar to the common R218H variant, the lesser increased in T3-binding affinity accounts for the modest increase of this hormone in serum of affected subjects.
We now report a new mutation in the HSA gene in a Thai family that replaces the normal leucine-66 with a proline (L66P), resulting in an HSA with 40-fold higher affinity for T3 but only 1.5-fold increase in the affinity for T4. The condition is also dominantly inherited and presents clinically as familial dysalbuminemic hypertriiodothyroninemia (FDH-T3) when T3 is measured by radioimmunoassay. More importantly, serum T3, and to a lesser extent T4, levels were variably low or even undetectable when measured by a variety of standard assays that use nonisotopic conjugates of these iodothyronines.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The proposita, a second child to unrelated parents, was born 6
months after the institution of routine screening for neonatal
hypothyroidism in the Bangkok Christian Hospital by measurement of
thyrotropin (TSH) and T4 in serum from cord blood. Because
the T4 concentration of 87.5 nmol/L, measured by
enzyme-linked immunosorbent assay (ELISA) (Enzymun-test, Boehringer
Mannheim Immunodiagnostics, Mannheim, Germany), was below the cut-off
value of 90.1 nmol/L, the tests were repeated at 6 days (see Table 1
for these and subsequent results of
thyroid function tests). Although serum T4 and
T3 concentrations were now within the normal range, the
serum TSH of 6.1 mU/L was considered to be elevated for her age. No
treatment was given. At 2 months of age, serum T3 measured
by radioimmunoassay (RIA) was more than 9.24 nmol/L (normal range,
1.08–2.70), and TSH had returned to normal (1.5 mU/L). At 4 months of
age, the serum T4 concentration was 61.8 nmol/L (normal
range, 57.9–154.4), and T3 was extremely low at 0.20
nmol/L (normal range, 1.23–2.77). Both were measured by ELISA. Despite
normal growth and development and a normal serum TSH concentration of
1.7 mU/L (normal range, 0.23–4.0) and a normal serum free
T4 level of 14.4 pmol/L (normal range, 11.6–24.5) measured
by Enzymun-test, her family physician began treatment with
levothyroxine (L-T4) at 4 months of age. She was first
given 25 µg L-T4 per day, increasing the dose every month
to reach 75 µg/day at 6 months of age. She was referred to Siriraj
and Rajavithi Hospitals for further investigations. At 8 months of age,
while receiving L-T4, her serum TSH was suppressed but the
level of T4, and in particular T3, remained low
when measured by ELISA or an enzyme-linked chemiluminescence assay
(CHEM) (Access, Sanofi Diagnostics Pasteur, Chaska, MN) but not by RIA.
Serum free T4 concentration was now high at 47.6 pmol/L and
in agreement with the low TSH level. For this reason and because the
same discrepancy in laboratory test results was found in her father and
older sister, treatment with L-T4 was discontinued. One
month later serum TSH level had returned to normal, but the
T3 concentration, measured by RIA, remained high (Table 1).
|
Tests of thyroid function
Serum total T4 and T3 concentrations were measured by two RIAs (DPC double antibody, Diagnostic Products, Los Angeles, CA, and Amerlex-M, Amersham International, Amersham, UK), three CHEM (Access, Sanofi Diagnostics Pasteur, Chaska, MN; Immulite, Diagnostic Products, and Amerlite, Johnson & Johnson Clinical Diagnostics, Amersham, UK) and two ELISA (Enzymun-Test, Boehringer Mannheim Immunodiagnostics, Mannheim, Germany and Magia, E. Merck Diagnostica, Darmstadt, Germany). Serum free T4 was measured by RIA (Amerlex-M, Amersham International, Amersham, UK) and serum free T3 was measured by fluoroimmunoassay (Delfia, Pharmacia Diagnostics, Uppsala, Sweden). Serum TSH concentration was determined by an IRA (Incstar Corporation, Stillwater, MN), by an enzyme-immunological method (Boehringer Mannheim Immunodiagnostics, Indianapolis, IN) and by a CHEM (Sanofi Diagnostics Pasteur).
Tests to assess defects of T3 and T4 binding to serum proteins
The presence of autoantibodies to T4 and T3 were assessed by polyethylene glycol precipitation of 125I-T4 or 125I-T3 (Dupont, NEN Research Products, Boston, MA) added to the serum samples as described (12). Binding of T4 and T3 to HSA was determined in 1:50 diluted serum by precipitation of added 125I-T4 or 125I-T3 with HSA antibody (Incstar, Stillwater, MN) (13). The binding affinities of HSA for T4 and T3 were determined as described for the measurement of binding of these iodothyronines to thyroxine-binding globulin (TBG) (14). Methodological differences were lesser serum dilution (1:40) and the addition of higher concentrations of the unlabeled iodothyronines to saturate the low capacity binding sites of TBG. The affinity constants (Ka) were calculated from the slope of the best fit line by the method of Scatchard (15).
Assessment of the interference in the T4 and T3 immunoassays
Serum samples from an affected member and from a normal member of the family were serially diluted (1:2, 1:4, and 1:6) with the appropriate zero calibrators, and T3 and T4 were assayed by RIA (DPC double antibody) and by CHEM (Access). Also, T4 (66.3, 132.6, 265.2 nmol/L) and T3 (1.72, 3.44, 8.96 nmol/L) were added to serums of an affected and a normal member of the family, and the samples were assayed by RIA (DPC double antibody), CHEM (Access) and ELISA (Enzymun-test). Finally, the effect of the blocking agent, 8-anilino-1-naphthalene-sulphonic acid (ANS), was assessed by its addition in excess (8 mg/mL of serum or 200 and 800 µg/assay tube, respectively, for the T4 and T3 determinations) to serum of two affected and two normal family members before measurement of T3 and T4 by RIA (DPC double antibody), CHEM (Access) and ELISA (Enzymun-test).
Preparation of genomic DNA and linkage study using the Sac I polymorphism in the HSA gene
Genomic DNA was isolated from peripheral-blood leukocytes (16). The SacI (±) polymorphism at codon 532, near the carboxyl terminus of the HSA, was determined in genomic DNA from all members of this family. For this purpose, a DNA fragment containing the polymorphic site in exon 13 was amplified by the polymerase chain reaction (PCR), as described (7).
Preparation of RNA, complementary DNA, and DNA sequencing
Total RNA was extracted by the acid guanidinium thiocyanate technique (17, 18) from mononuclear cells isolated from 30 mL heparinized blood by centrifugation in Ficoll-Paque (Pharmacia, Piscataway, NJ).
The small amount of HSA messenger RNA (mRNA) present in mononuclear cells served as a template for the synthesis of complementary DNA (cDNA) using MMLV (RNase H-) reverse transcriptase (Gibco BRL, Gaithersburg, MD) (19). To amplify specifically the mutant allele of the HSA associated with the Sac I (-) polymorphism (see Results), an allele specific antisense primer (5'-CCTTGGGCTTGTGTTTCACA-3') was used to synthesize the first cDNA strand. The latter was then used as a template to amplify by PCR four overlapping fragments that extended from exon 1 through exon 13 of the HSA gene. The distal cDNA fragment, covering exon 13 through exon 15, was similarly generated from the first strand synthesized using an antisense primer complimentary to the noncoding 3' sequence of the HSA (5'-AGACAGGGTGTTGGCTTTAC-3'), which was then amplified by PCR using an allele specific sense primer (5'-AAACTGCACTTGTTGAGCTT-3') and antisense primer (5'-TCTTATTCTCATGGTAGGCTGAG-3'). All amplified DNA fragments were sequenced by a 373 DNA Sequencer (Applied Biosystems, Perkin-Elmer, Foster City, CA). Sequencing primers and PCR conditions have been published (7).
Confirmation of the HSA mutation
To confirm the presence of the mutant nucleotide in genomic DNA and to identify the mutation in all family members, a mismatched oligonucleotide primer was synthesized that is complementary to sequences near but not overlapping the mutant nucleotide. It was designed so that the product of amplification would create a unique restriction site (Bsl I) only if the template contained the mutant nucleotide (endonuclease-digestion allele-specific primer method) (20). The primer sequences are 5'-CAGCAGTGTCCATTTGAAGA-3' (sense) and 5'-ctccacaattagaatccactt-3' (antisense, intronic, and mismatched nucleotides are underlined).
Following PCR amplification of the subjects’ genomic DNA, the products were digested with Bsl I and submitted to electrophoresis in 3% Nusieve/1% agarose (Amresco, Solon, OH). Partial cleavage of the DNA fragment indicated that the mutant nucleotide was present in one of the two alleles.
Statistical analysis
Numerical, grouped results are expressed as mean ± SD, and the statistical differences of means were determined by the unpaired Student t test. A P value less than 0.05 was considered significant. Linear regression analysis was performed by the least squares method, and the coefficient of correlation (r) value was obtained for the best-fit curve. The linkage analysis and calculation of the logarithm of the odds ratio (LOD) score were done as described before (21).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. The phenotype that characterizes the
affected individuals is a serum total T3 level that is
above the upper limit of normal when measured by RIA (mean ±
SD: 4.97 ± 1.52 nmol/L for affected vs.
1.49 ± 0.59 nmol/L for normal), while being very low when
measured by a CHEM and ELISA method. Similar, though less profound
method-dependent variation in serum total T4 concentration
also occurred, with low T4 levels measured by CHEM and
ELISA methods, but without significant increase when measured by RIA
(112.0 ± 25.7 nmol/L for affected and 99.1 ± 16.7 nmol/L
for normal). Because there were no significant differences between
normal first degree relatives and relatives by marriage for all tests,
values from both normal groups were combined in the statistical
analyses. All affected members had normal serum TSH concentrations and
the differences between means of the affected and normal subject were
not significant irrespective of the assay method (1.4 ± 0.7 for
affected and 1.2 ± 0.7 for normal). Serum free T4 concentrations
were normal with mean values of 13.80 ± 2.88 pmol/L for affected
and 14.97 ± 3.16 pmol/L for normal, and serum free T3
levels were also normal with mean values of 3.39 ± 0.97 pmol/L
for affected and 3.42 ± 0.87 pmol/L for normal (Fig. 1).
|
. Affected individuals had
significantly high serum T3 levels when measured by the two
RIA procedures, and very low to normal concentrations when measured by
the various CHEM and ELISA methods. Serum T4 levels in the
affected subjects were slightly high or normal when measured by RIA and
slightly to moderately low when measured by CHEM or ELISA
procedures.
|
. While recovery of T3 and
T4 from the serum of the affected subject measured by RIAs
was 98–113% and 93–101%, respectively, in CHEM assays recoveries
were 7–11% and 34–45%, respectively and in ELISA assays, 4–12%
and 66–75%, respectively. Recovery of T3 and
T4 from the normal serum sample did not differ with the
assay method and ranged, overall, from 92–116%. This variable
recovery of T3 and T4 from serum of affected
members of the family explains the wide range of values observed
depending on the test procedure. Values obtained by the DPC double
antibody RIA represent the true concentrations of these
iodothyronines.
|
The percent of 125I-labeled T3 and
T4 added to serum that was precipitated with antibody to
HSA is shown in Fig. 1
. A larger proportion of T3 in serum
of affected individuals precipitated with the HSA antibody and, with
one exception, the amount of hormone bound to HSA did not overlap with
that bound in serum samples of the unaffected family members. This
exception was an affected woman (II-11) who had high thyroxine-binding
globulin (TBG) because of pregnancy. HSA precipitable T3
was 5.3% during pregnancy and 33.9% two months after delivery. Mean
values were 29.2 ± 1.8 and 7.0 ± 1.0% (P
< 0.0001) for affected and normal family members, respectively. The
difference between the two groups in the percent T4
precipitable with HSA antibody was smaller (5.4 ± 0.8 and
4.3 ± 0.3%, respectively) but was also significant
(P = 0.01).
The association constant (Ka) of HSA for T3 in two affected
subjects was 1.5 x 106 and 1.6 x
106 M-1, or 40-fold that of normal relative of
3.9 x 104 M-1 and controls (3.8 x
104 and 4.3 x 104 M-1) (Fig. 3). As shown in Fig. 3, HSA with
increased Ka for T3 could be demonstrated in the affected
pregnant woman (II-11), with reduced proportion of T3
associated with HSA (Fig. 1), after saturation of the high affinity
binding sites on TBG. In contrast, the Ka of HSA for T4 in
an affected subject was 5.5 x 105 M-1,
or only 1.5-fold that of a normal (3.6 x 105
M-1) (graphed data not shown).
|
indicated that the phenotype of this family could be linked to the Sac
I (-) allele of HSA gene although, the LOD score was only 1.51,
because of the relatively small number of family members tested.
|
) was sequenced in its entirety and showed that the second nucleotide
of codon 66 (CTT), a thymine, was substituted by a cytosine
(CCT), resulting in the replacement of the normal
Leu66 by Pro66. Genotyping by the
endonuclease-digestion allele-specific primer method confirmed that the
nucleotide substitution in codon 66 of subject II-5 was also found in
one of the two alleles of each affected family members and that it was
not present in the normal relatives (Fig. 4C). |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Herein we report a new syndrome characterized by elevated serum total T3 but not T4 concentration when determined by procedures not subject to interference by the mutant HSA. As in the common form of FDH found in Caucasians (7, 8), and that recently described in a Japanese kindred (11), this familial dysalbuminemic hypertriiodothyroninemia (FDH-T3) was linked to the HSA gene. The mutant HSA, L66P, has 40-fold higher binding affinity for T3 than the common type HSA (Ka = 1.5 x 106 M-1 vs. 3.9 x 104 M-1), and only 1.5-fold higher affinity for T4 (Ka = 5.5 x 105 and 3.6 x 105 M-1). This contrast with FDH caused more distal mutations in codon 218, which has a higher binding affinity for T4 than for T3 (7, 8, 11). FDH-T3 is linked to the Sac I (-) intragenic polymorphism and FDH-218H with Sac I (+) polymorphism (7).
The identification of the key case was not straightforward. While studies were initiated because of a borderline low T4 value at birth, it was the falsely low T3 concentration measured by an ELISA method, and the failure of treatment with L-T4 to correct this abnormality, that led to the examination of family members. The apparent autosomal dominant mode of inheritance and male-to-male transmission were not compatible with a TBG defect, and the higher proportion of 125I-T3 associated with HSA pointed to a defect in the albumin gene. Failure to exclude linkage of the phenotype to the HSA gene and use of the polymorphic marker to specifically amplify the putative defective HSA gene led to the identification of the unique mutation.
Uncovering the explanation for the markedly discordant results of
iodothyronine concentrations, measured by different commonly used
clinical laboratory procedures, was challenging not only diagnostically
but also technically. Various mechanisms were considered, and the
likely reasons for the erroneous results were investigated. Table 3 outlines the constituents and the
principle of the seven methods of T3 and T4
measurement tested; the results are given in Table 2. False results
could not be traced to the nature of the antibody, the blocking agent,
or the method of separation of the antibody bound from free hormone.
While the identity of the blocking agents could not be obtained for all
assay methods, we found that addition of excess ANS did not play a
major role in the discrepant results. Iodothyronine recovery data
demonstrated that the RIA method provided the most accurate results.
Thus the common denominator for the low recovery of T3 and
to a lesser degree T4, which are associated with the mutant
HSA, is the use of iodothyronine conjugates. These iodothyronine
analogs, having variable reduction in the binding affinity for the
mutant HSA relative to the native molecule, bind preferentially to the
antibody and reduce the measured amount of iodothyronine. By contrast,
in the RIAs, the isotopically labeled tracer is chemically identical to
the native molecule and follows the same partition as the endogenous
iodothyronines. As a consequence the ratio of labeled tracer to
unlabeled iodothyronine is dependent on the amount of the endogenous
iodothyronine irrespective of the amount that bound to the mutant
protein or to the antibody, explaining the lack of interference by the
mutant HSA. Considering free T4 by RIA (Amerlex) and free T3 by
fluoroimmunoassay (Delfia), we found they were in the normal range,
with no interference from mutant HSA. Because the labeled iodothyronine
analogs used in the kits do not increase binding to mutant HSA, the
result would give authentic levels.
|
Serum T4 and T3 concentrations were normal when
measured by ELISA (Enzymun-test) in cord blood and at 6 days of age
(Table 1). It is possible that the mutant HSA is not expressed in early
life. Data clearly indicate that the mutant HSA was present in serum at
2 months of age. The slight reduction of serum T4
concentration at birth was thus fortuitous and it appears unlikely that
screening tests for neonatal hypothyroidism would identify this defect.
Thus, the clinical importance of FDH-T3 may be erroneous
diagnosis in the adult. While the prevalence of this anomaly is
unknown, the family under investigation has more than 300 members
living in various areas of Thailand, Singapore, and the USA.
Considering the dominant mode of inheritance and the apparent lack of
deleterious effect, this mutant HSA may be found with increasing
frequency as clinical laboratories move to the use of nonradioactive
analogs for the measurement of iodothyronines. The combination of
normal T4 and low T3 could easily be
misinterpreted as an alteration of nonthyroidal illness, and the
persistence of abnormal tests after recovery could result in
inappropriate treatment.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received January 5, 1998.
Accepted February 2, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Pannain, M. Feldman, U. Eiholzer, R. E. Weiss, N. H. Scherberg, and S. Refetoff Familial Dysalbuminemic Hyperthyroxinemia in a Swiss Family Caused by a Mutant Albumin (R218P) Shows an Apparent Discrepancy between Serum Concentration and Affinity for Thyroxine J. Clin. Endocrinol. Metab., August 1, 2000; 85(8): 2786 - 2792. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
