Homozygous Mutation in Human Serum Albumin and Its Implication on Thyroid Tests

Introduction

Familial dysalbuminemic hyperthyroxinemia (FDH) is an autosomal dominant condition that is caused by heterozygous gain-of-function mutations in the human serum albumin (HSA) gene that result in mutant HSA with increased binding affinity for T4 (1 –4). Herein, the first report of thyroid test abnormalities in an individual homozygous for mutant HSA in a large consanguineous family of Anatolian Turkish descent is given.

Patients

The homozygous individual was an uncle of the proband, a 23-year-old woman who presented to one of the authors (A.K.) for evaluation of hirsutism. She was otherwise asymptomatic. Thyroid tests showed a high free thyroxine (T4) of 2.82 ng/mL (reference range [RR] 0.58–1.25 ng/mL) and triiodothyronine (T3) of 5.36 pg/mL (RR 2.3–4.2 pg/mL), with a normal thyrotropin (TSH) of 2.1 mIU/L (RR 0.4–4.0 mIU/L), raising suspicion for resistance to thyroid hormone beta (RTH-β).

Results

Written consent was obtained in accordance with an Institutional Review Board–approved protocol. As shown in Figure 1A, 9/14 relatives tested had elevated serum total T4 (TT4) with normal free T4 index (FT4I) and total T3 (TT3), except individual IV-2. These results, in the presence of non-suppressed TSH, suggested FDH. Four individuals (II-5, III-4, IV-2, and IV-5) had mildly elevated TSH consistent with subclinical hypothyroidism, and two (II-5 and IV-5) had positive thyroid autoantibodies. Individual IV-2, a 10-year-old boy with obesity, had FT4I and TT3 at 8% and 32% above the adult reference range, respectively (Fig. 1A), consistent with RTH-β. However, sequencing of exons 7–10 of THRB was negative for a mutation.

OPEN IN VIEWER

FIG. 1.

(A) Pedigree of the family with thyroid test results aligned below each individual family member's symbol on the pedigree; abnormal values are in bold. When more than one determination was obtained from samples collected at different times, results were averaged. Males and females are represented by squares and circles, respectively. A single arrow indicates the proband, and a double arrow indicates the individual homozygous for the HSA mutation. The asterisk indicates an individual who had a thyroidectomy and is on replacement levothyroxine. Subjects who were not available for testing are in gray. Roman numerals to the left of the pedigree indicate generation, and Arabic numerals indicate distinct individuals of a given generation. (B) Radioautograph of serum with¹²⁵I-labeled thyroxine (T4) added prior to separation by isoelectric focusing. Sera were from a spouse and cousin to individual III-2 in lane 2 (WT, lane 1) and three siblings homozygous (HOMO, lane 2), heterozygous (HET, lane 3), and WT (lane 4) for the HSA mutation. (C) Radioautograph of non-denaturing polyacrylamide gel electrophoresis (PAGE) of serum with added ¹²⁵I-labeled T4 from the same individuals as in (B) (see legend to Fig. 2 for gel conditions). TBG, T4-binding globulin; TTR, transthyretin; HSA, human serum albumin.

Sequencing of DNA from peripheral leukocytes revealed a mutation in the HSA gene (c.653G>A; Fig. 2A), resulting in substitution of the normal arginine-218 with histidine (p.R218H) in the nine hyperthyroxinemic family members. Interestingly, individual III-2, born to heterozygous parents, was homozygous for the R218H mutation (Fig. 2A). His TT4 was 37% and 161% above the mean TT4 of heterozygous and wild-type (WT) family members, respectively (Fig. 2B). Isoelectric focusing (IEF) of this individual's serum, visualized with radiolabeled T4, identified the mutant HSA protein (Fig. 1B, lane 2). Notably, labeled T4 bound the mutant protein to the exclusion of WT HSA present in sera of a heterozygous sibling (Fig. 1B, lane 3).

OPEN IN VIEWER

FIG. 2.

(A) Representative chromatograms of sequences from family members WT (c.653G, p.R218), heterozygous (c.653G>A, p.R218H monoallelic substitution), and homozygous (c.653G>A, p.R218H biallelic substitution) for the HSA mutation. (B) Scatter and box plots of total T4 concentrations in WT, heterozygous, and homozygous individuals. Line represents the median, and whiskers represent the range, when applicable. (C) Quantification of relative T4 content in bands representing TBG, HSA, and TTR separated by non-denaturing PAGE shown in Figure 1C. The table below the graph shows the corresponding measured concentrations of T4, TBG, and HSA in sera from these individuals, and calculated quantity of T4 bound to HSA and TBG, adjusted for the serum concentrations of these proteins. PAGE was carried out as follows: 1 μL containing 20,000 counts per minute (cpm; or 16 pg) of ¹²⁵I-labeled T4 (specific activity 1250 μCi/μg) was added to 10 μL of serum from each individual and incubated for 30 min at room temperature. The added ¹²⁵I-T4 thus represented 0.6–2.2% the endogenous T4 measured in the serum samples. Following dilution in buffer (250 mM of Tris•Cl, pH 8.5), 6 μL was loaded for separation by 8% non-denaturing PAGE at 60 V for 4 h at room temperature. The gel was subsequently exposed to film at −80°C with an intensifying screen for 1–4 days and scanned into an electronic format, and relative quantification of the bands was performed using the publically available NIH ImageJ software.

Non-denaturing polyacrylamide gel electrophoresis of sera from individuals of each genotype with radiolabeled T4 was performed, as described in the legend to Figure 2, to determine the proportion of T4 bound to the three T4-binding serum proteins (Fig. 1C). The effect of sample preparation on the relative distribution of T4 among the T4-binding proteins was determined by applying different amounts of ¹²⁵I-labeled T4 to a WT serum (Fig. 3A), and by serial dilutions of sera from a WT and homozygous mutant individual (Fig. 3B). Measures varied by <15%, and the ratio of percent T4-bound to the mutant versus WT HSA ranged from 1.9 to 2.3. Quantification revealed that while the percent of transthyretin (TTR)-bound T4 is stable across genotypes, there is a stepwise increase in the percent of T4 bound to HSA in the WT (average 20%), heterozygous (36.4%), and homozygous (47.9%) individuals, and a concomitant reduction in T4-bound T4-binding globulin (TBG; Fig. 2C). Serum HSA and TBG were measured and used to estimate the relative quantity of T4 bound to each protein. Binding to HSA was six- and threefold higher in the homozygote and heterozygote relative to the WT. In contrast, the amount of T4 bound per milligram of TBG was similar in all genotypes (Table in Fig. 2C, rows 4 and 5).

OPEN IN VIEWER

FIG. 3.

Radioautographs and quantification of relative T4 associated with TBG, HSA, and TTR separated by non-denaturing PAGE in serum from (A) a WT individual with different amounts of added ¹²⁵I-labeled T4 and (B) serially diluted sera from a WT individual or an individual homozygous for the HSA mutation. (A) Sera are from aliquots containing 20,000 cpm (lane 1), 10,000 cpm (lane 2), or 5000 cpm (lane 3) of ¹²⁵I-labeled T4. (B) Sera was undiluted (lanes 1 and 4), diluted 1:2 (lanes 2 and 5), or diluted 1:4 (lanes 3 and 6), and 10,000 cpm (or 8 pg) of ¹²⁵I-labeled T4 was added to each sample. The relative amounts of T4 added with the ¹²⁵I-labeled T4 were 1.1–4.4% of the endogenous T4 for the WT and 0.3–1.2% for the mutant.

Discussion

FDH is a well-characterized condition, producing abnormal thyroid tests in euthyroid individuals (1 –4). The R218H mutation was first identified in a large Amish kindred (5) and is the most common of five HSA variants causing FDH (1). Substitution of R218 for a smaller residue electrostatically stabilizes the T4-binding site and thereby promotes increased affinity for T4 (6). Individuals with R218H have elevated serum TT4 and, to a lesser degree, reverse T3 due to increased mutant HSA binding affinity for these iodothyronines, but normal dialyzable free T4 and a non-suppressed TSH. Mutant HSA can interfere with indirect methods of measuring free T4 and T3, producing falsely elevated values (7), leading to inappropriate treatment and explaining the elevated free T4 and T3 levels in the proband at presentation.

To the authors' knowledge, this is the first report of a biallelic HSA mutation causing FDH. As HSA is present in 6000-fold molar excess relative to T4, T4 binds exclusively to mutant HSA, as shown by IEF, despite representing only half the total HSA in heterozygotes. Quantification of the T4 bound to HSA showed that the homozygote had a twofold increase in T4 binding per gram of HSA compared with heterozygotes (Fig. 2C), including subject IV-2 (1.3 μg T4/g HSA). Thus, biallelic expression of the mutant HSA requires twice the amount of bound T4 to maintain a normal free T4.