Autoimmunity in Patients with Resistance to Thyroid Hormone

Abstract

Context: Resistance to thyroid hormone (RTH) is an inherited syndrome most often caused by thyroid hormone receptor β (TRβ) gene mutations. Given that autoimmune thyroid disease (AITD) is prevalent in the general population, its coexistence with RTH has been presumed coincidental. It was recently proposed that chronic TSH stimulation in RTH may induce an autoimmune response, thereby increasing the chance of their coexistence.

Objective: The aim was to examine the prevalence of AITD in a large cohort with RTH compared with their unaffected first-degree relatives.

Subjects and Methods: Among 130 families, 330 individuals with RTH confirmed by the presence of TRβ gene mutations and 92 unaffected first-degree relatives were tested for thyroglobulin and thyroperoxidase antibodies. The presence of AITD was based on at least one of the two antibodies being positive. Data were analyzed according to genotype, gender, age, and familial association. A large homogeneous family was analyzed separately.

Results: Individuals with RTH had an increased likelihood of thyroid autoantibodies (odds ratio = 2.36; P = 0.002). In males, the odds of having AITD were higher in individuals with RTH compared to unaffected first-degree relatives (odds ratio = 2.91; P = 0.042). Although female subjects with RTH had an odds ratio of 1.95 for having thyroid autoantibodies, the difference was not statistically significant (P = 0.097). Antibody prevalence at different ages was not affected by genotype.

Conclusions: Individuals with RTH due to TRβ gene mutations have an increased likelihood of AITD compared to unaffected relatives, but the prevalence of thyroid autoantibodies with advancing age is not affected by genotype. These novel findings demonstrate that the association between RTH and AITD is not coincidental.

Individuals with resistance to thyroid hormone have an increased likelihood of autoimmune thyroid disease throughout their lifetime, but the prevalence of thyroid autoantibodies with advancing age is not affected by genotype.

Resistance to thyroid hormone (RTH) is an inherited syndrome characterized by reduced sensitivity to thyroid hormone (TH). Elevated serum TH levels in the setting of a nonsuppressed TSH level reflects reduced response of target tissues to circulating TH (1,2). RTH affects 1 of 40,000 live births and is mostly inherited as an autosomal dominant trait. Common clinical manifestations include goiter, sinus tachycardia, growth delay, and attention deficit disorder (1,2,3). Eighty-five percent of cases are attributed to mutations in the TH receptor β (TRβ) gene.

In contrast, autoimmune thyroid disease (AITD), defined by the presence of thyroid autoantibodies, is relatively common in the general population. National Health and Nutrition Examination Survey III data revealed thyroglobulin antibodies in 10.4% and thyroperoxidase antibodies in 11.3% in a cohort without known thyroid disease (4). The majority of these individuals, however, never manifest clinical sequelae of thyroid dysfunction.

Reports have described isolated cases with both RTH and AITD (5,6,7,8,9,10). Given the high prevalence of AITD in the general population, this coexistence has been presumed coincidental (6,7,8). The presence of thyroid autoantibodies in an individual with RTH often leads to misdiagnosis, and patients subsequently undergo unnecessary medical and surgical treatment (6,9).

Gavin et al. (11) recently proposed that chronic TSH stimulation in RTH activates intrathyroidal lymphocytes leading to thyroid damage and autoimmune hypothyroidism. Given the reports of coexistent RTH and AITD and this new proposed causal mechanism, our study aimed to assess whether there is an increased prevalence of AITD in individuals with RTH and whether such a relationship increases with advancing age.

Subjects and Methods

Subjects

A total of 422 individuals from 130 unrelated families, identified between 1981 and 2008, were included in the study. There were 128 single probands, a family with twin probands, and one with no identifiable proband. Common reasons for obtaining thyroid function tests in probands included goiter, anxiety, tachycardia, attention deficit disorder, and growth delay. In addition to probands, laboratory and genetic analysis identified 202 subjects with RTH, including probands’ parents, siblings, and progeny. Thus, 330 individuals with RTH were studied.

Unaffected first-degree relatives (n = 92) served as controls. To eliminate genetic and environmental bias, first-degree relatives were defined as all siblings and progeny of individuals with RTH who were susceptible of inheriting the defective TRβ gene and thyroid autoantibodies. This definition encompassed all individuals at equal risk of inheritance, and thus parents and siblings of probands with de novo mutations were excluded.

Criteria for inclusion were a proband with RTH due to a TRβ gene mutation and thyroid autoantibody analyses performed in our laboratory using the same method. All family members underwent the same laboratory and genetic analyses as performed on the proband.

A large Azorean family with a homogeneous genetic background was excluded from the overall analysis in order not to bias the results (12). All members of this family descended from a late 19th-century founder couple. The prevalence of AITD in this family was analyzed separately.

All studies were approved by the Institutional Review Boards of The University of Chicago and the Hospital Divino Espírito Santo. All family members were offered complete testing at no cost, and written informed consent was obtained from all subjects.

Laboratory studies

The following laboratory determinations were obtained in all subjects: total T₄ and T₃, free T₄ index, and TSH.

Antibodies to thyroglobulin and thyroperoxidase were determined by an agglutination method (Fujirebio, Inc., Tokyo, Japan). To eliminate bias, the same agglutination method was used for all samples because it was the available method when sample collection began 30 yr ago. Agglutination at a dilution of 1:20 or greater was defined as a positive test. Subjects were considered to have AITD in the presence of a positive test to at least one of the two antigens. Because titer value does not correlate with disease severity, subgroup analyses according to titer were not performed.

DNA from peripheral blood was used for TRβ gene sequencing as previously described (13). De novo mutations were confirmed when both parents of an affected proband were homozygous for the wild-type allele.

Statistics and data analysis

Evaluation of AITD risk among subjects with RTH and their unaffected first-degree relatives was performed using odds ratios (ORs) from generalized estimating equations models. This technique allows the fitting of logistic regression models that account for the correlation within families (14). Initial models were fitted with genotype, age, and gender as predictive variables. Subsequent models examined the effects of genotype and age in males and females separately. Finally, interaction terms between genotype and age were added to determine whether the prevalence of thyroid autoantibodies with advancing age is affected by genotype.

For the Azorean family, Fisher’s exact test was used for analysis. Logistic regression was not used due to the small sample size and the rule that 10 events are required per covariate in a logistic regression model.

Stata v. 10.1 (Stata Statistical Software, release 10; StataCorp LP, College Station, TX) was used for all statistical analysis. A P value <0.05 was considered statistically significant. Results are presented in terms of ORs and the corresponding 95% confidence intervals (CIs).

Results

In total (excluding the Azorean family), 422 individuals from 130 unrelated families were analyzed. Individuals with RTH totaled 330, of whom 191 were female and 139 were male. Of the 92 unaffected first-degree relatives, 48 were female and 44 were male (Table 1). Mean age at testing was not statistically different in unaffected individuals (21 ± 17 yr) compared with those with RTH (26 ± 16 yr); overall age range was 0–90 yr. There was also no significant age difference between genders (mean age, 27 ± 18 yr for females and 23 ± 19 yr for males).

Table 1.

Study subjects by gender, genotype, and antibody status

	RTHa		Unaffectedb
	Ab+c	Ab−d	Ab+c	Ab−d
Unrelated families
Females (n = 239)	50	141	6	42
Males (n = 183)	27	112	3	41
Total (n = 422)	77	253	9	83
Azorean family
Females (n = 35)	3	15	6	11
Males (n = 35)	3	17	2	13
Total (n = 70)	6	32	8	24

^aSubjects with RTH due to TRβ gene mutations. 

^bFirst-degree relatives unaffected by RTH. 

^cSubjects with thyroid autoantibodies. 

^dSubjects without thyroid autoantibodies. 

Risk of AITD according to genotype

The overall gender- and age-adjusted risk for AITD among all 422 subjects was significantly higher in individuals with RTH (OR = 2.36; P = 0.002; 95% CI, 1.37–4.06). Among male subjects (n = 183), the odds of AITD were higher in individuals with RTH compared with unaffected male first-degree relatives (OR = 2.91; P = 0.042; 95% CI, 1.04–8.14). Although female subjects with RTH had an OR of 1.95 for having AITD, the difference was not statistically significant (P = 0.097; 95% CI, 0.89–4.31).

Schuppert et al. (15) reported increased serum thyroid autoantibody levels in patients with nontoxic goiter compared with nongoitrous individuals. To eliminate selection bias from goiter, we analyzed a subset of subjects (n = 325) whose probands presented for reasons other than goiter. The overall risk of AITD in this cohort was higher in subjects with RTH (OR = 3.26; P < 0.001; 95% CI, 1.75–6.10). The odds of AITD in males (n = 138) was higher in subjects with RTH compared with unaffected male relatives (OR = 4.36; P = 0.024; 95% CI, 1.22–15.64). Females (n = 187) also showed an increased odds of AITD among subjects with RTH (OR = 2.61; P = 0.042; 95% CI, 1.03–6.61). Thus, the exclusion of families whose proband presented with goiter did not abolish the overall increased risk of AITD in subjects with RTH.

Thyroid autoantibody prevalence according to age and genotype

To determine whether RTH plays a role in the development of AITD with advancing age, the percentages of individuals with thyroid autoantibodies at different ages were examined in both females and males according to genotype. For females, age had a significant quadratic (nonlinear) effect on antibody prevalence (P = 0.0002) (Fig. 1). There was, however, no interaction between age and genotype (P = 0.573), suggesting that the prevalence of antibodies in females over time is not influenced by RTH.

Figure 1.

Thyroid autoantibody percentages per age group according to gender and genotype. Values plotted are the percentage of individuals with positive antibodies grouped into 8-yr intervals for subjects between 10 and 65 yr of age. In subjects aged less than 10 yr, age groups include subjects aged 1 month to 4 yr and aged 5–9 yr. Due to a very small number of elderly subjects, the final age category included all subjects aged 66 yr and older. Solid and dotted lines are fitted values from generalized estimating equation models.

For males, age had no significant linear or nonlinear effect on antibody prevalence (P = 0.174 and P = 0.398, respectively). In addition, despite the increased likelihood of AITD among males with RTH, the pattern of antibody prevalence over different age groups was similar for the two genotypes (P = 0.736) (Fig. 1).

Risk of AITD in the Azorean family

In a separate analysis, 70 individuals of a 260-member Azorean family met inclusion criteria. Individuals with RTH totaled 38, of whom 18 were female and 20 were male. Of the 32 unaffected first-degree relatives, 17 were female and 15 were male (Table 1). Mean age difference at testing of 26 ± 18 yr for individuals with RTH and 32 ± 19 yr for unaffected relatives (range, 0.5–80 yr) was not statistically significant. Fisher’s exact test showed no significant differences in the presence of thyroid autoantibodies between subjects with RTH and unaffected relatives overall (P = 0.381) or in males or females alone (P = 1.00 and P = 0.264, respectively).

Discussion

The coexistence of AITD and RTH is intriguing given their different epidemiological and genetic features. AITD is more common in women (4), whereas RTH has equal prevalence in both genders (1,2,3). The majority of cases of RTH are due to TRβ gene mutations (1,3), whereas AITD is the result of complex genetic, environmental, and hormonal factors (16,17). RTH is known to have a dominant negative effect (18) not seen with AITD.

We reviewed what we believe to be the largest database of patients with RTH. In comparing individuals with RTH and unaffected family members, we determined an increased prevalence of AITD among the RTH cohort. We conclude that this coexistence is a true association and not merely coincidental. The possibility of selection bias was considered due to the 3.6-fold difference in the number of subjects with and without TRβ gene mutation. Because 41 families had a single affected subject and 26 families consisted of a single affected child and an affected parent, only 237 subjects with RTH had a first-degree relative with no TRβ gene mutation. Taking into consideration that mothers with RTH have a 50% reduction in the birth of unaffected children (12), the 237 subjects with RTH having unaffected first-degree relatives are expected to have 177 unaffected siblings, a difference of less than 2-fold of the 92 examined. Under these circumstances, a selection bias should reduce the difference in the prevalence of AITD between subjects with RTH and unaffected relatives because the presence of AITD would motivate the latter group to participate in the study.

To examine the hypothesis proposed by Gavin et al. (11), the prevalence of thyroid autoantibodies was investigated at different ages. If chronic TSH elevation causes thyrocyte destruction, one would expect an increase in thyroid autoantibodies through time among individuals with RTH. This study demonstrates that the antibody percentages per age group were not influenced by genotype. Thus, TSH stimulation leading to intrathyroidal lymphocyte destruction and ultimately AITD is unlikely to be the mechanism of increased autoimmunity in individuals with RTH.

Further supporting this conclusion is the lack of correlation between RTH and AITD in the Azorean family. If TSH-activated lymphocytes damaged the thyroid glands of individuals with RTH, our data should have revealed similar findings in the Azorean subjects as in the other families. The fact that there was no association between RTH and AITD in the Azorean family may be due to a genetic predisposition for AITD enhanced by RTH not present in the genetic background of this homogeneous family.

The immune system may be activated via another mechanism in individuals with RTH. Fabris et al. (19) demonstrated a relationship between TH levels and thymic function, and Villa-Verde et al. (20) revealed TH receptors in murine thymic epithelial cells. It is plausible that elevated TH levels in RTH stimulate the immune system at the TRα level. Such is the presumed mechanism for tachycardia in RTH because increased levels of TH activate TRα on cardiac cells (1,2,3). To test this immune hypothesis, future research should focus on the autoimmune environment in individuals with RTH, as well as the antibody presence in mice carrying Trβ gene mutations.