help button home button Endocrine Society JCEM
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Samollow, P. B.
Right arrow Articles by Barmada, M. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Samollow, P. B.
Right arrow Articles by Barmada, M. M.
The Journal of Clinical Endocrinology & Metabolism Vol. 89, No. 7 3276-3284
Copyright © 2004 by The Endocrine Society

Genetic and Environmental Influences on Thyroid Hormone Variation in Mexican Americans

Paul B. Samollow, Graciela Perez, Candace M. Kammerer, David Finegold, Patrick W. Zwartjes, Lorena M. Havill, Anthony G. Comuzzie, Michael C. Mahaney, Harald H. Göring, John Blangero, Thomas P. Foley and M. Michael Barmada

Department of Genetics (P.B.S., P.W.Z., L.M.H., A.G.C., M.C.M., H.H.G., J.B.), Southwest Foundation for Biomedical Research, San Antonio, Texas 78245; Department of Human Genetics (G.P., C.M.K., D.F., M.M.B.), Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania 15261; and Division of Pediatric Endocrinology (D.F., T.P.F.), Department of Pediatrics, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15213

Address all correspondence and requests for reprints to: Paul B. Samollow, Southwest Foundation for Biomedical Research, P.O. Box 760549, San Antonio, Texas 78245-0549. E-mail: pbs{at}darwin.sfbr.org.


   Abstract
Top
 Abstract
Introduction
 Subjects and Methods
 Results
 Discussion
 References
 
Thyroid hormones play major roles in the regulation of a wide range of metabolic and physiologic processes, but the genes and environmental factors that affect normal, quantitative variation in thyroid hormone concentrations are largely unknown. Using quantitative genetic methods, we evaluated the effects of genes and environmental factors on thyroid hormone variation in 586 women and 425 men from 27 randomly ascertained Mexican-American families from the San Antonio Family Heart Study. Data were available on free and total T4, free and total T3, TSH, thyroglobulin, and T4-binding globulin, as well as on covariates, including sex, age, weight, lifestyle habits, physical activity, and others. These covariates accounted for 2–18% of total phenotypic variation, whereas genes accounted for 26–64% of the variation. Overall, free T3 had the highest heritability, which is noteworthy because it is the most biologically active thyroid hormone and accounts for the vast majority of metabolic and physiologic effects of thyroid hormones. Our results indicate that genes account for a substantial portion of variation in human thyroid hormone levels, and suggest that further studies to identify the genes involved in this variation could reveal important insights into the processes that govern thyroid-mediated metabolism.


   Introduction
Top
 Abstract
 Introduction
Subjects and Methods
 Results
 Discussion
 References
 
DIRECTLY OR INDIRECTLY, the actions of thyroid hormones impinge on virtually every level of physiologic integration. In early development and childhood, thyroid hormones are vital to proper skeletal and neural development (1, 2). In adults, thyroid hormones influence protein metabolism, bone deposition, lipolysis and lipogenesis, glucose balance, metabolic rate, thermogenesis and temperature regulation, cardiac output, blood pressure, and more (2). Normal variation in thyroid hormone phenotypes spans a continuous range between the frank clinical endpoints of hypothyroidism and hyperthyroidism (conditions that are strongly associated with a spectrum of morbid syndromes affecting both physical and chemical systems). These same systems are likely to be influenced by differences in levels of thyroid hormones that fall within the normal range as well. If so, then such variation in thyroid hormone levels could be intimately connected with variation in health-related characteristics such as body fat distribution, metabolic efficiency, bone metabolism, and lipoprotein metabolism. Because these variables are risk factors in susceptibility to diseases such as cardiovascular disease, diabetes, obesity, and osteoporosis, genetic variation in thyroid hormone regulation may be an important factor contributing to differential susceptibility to a variety of common diseases. Thus, determining the intrinsic and epigenetic factors that underlie interindividual variation in thyroid hormone phenotypes is important. The value rests not only in understanding the causes of normal and abnormal variation in thyroid function, but also in the potential for developing more effective diagnostic and therapeutic approaches for the control of thyroid variation that may lead to diminished health or lifespan from conditions not directly related to overt thyroid disease.

Several relationships between thyroid hormone levels and health-related risk factors in clinically normal individuals are known. For example, elevated serum TSH concentrations in clinically normal individuals are correlated with multiple alterations in serum lipoprotein concentrations (3, 4, 5, 6, 7), whereas reduced levels of TSH are associated with increased risk of atrial fibrillation, atherosclerosis, and cardiac death in normal middle-aged and elderly persons (8, 9, 10, 11, 12). In addition, variation in serum levels of T3 explains approximately 20% of variation in serum high-density lipoprotein cholesterol and apolipoprotein A-I, which are risk factors for cardiovascular disease, in a randomly ascertained (normal) Mexican-American population (13). These and other examples reinforce the concept of relationships between normal thyroid hormone variation and health-related physiologic processes.

Increased knowledge of the genetic regulation of the thyroid neuroendocrine system can lead to a clearer picture of the genetic pathways of thyroid hormone production and control, and perhaps to a better understanding of the molecular mechanisms of thyroid abnormalities. In recent years, important advances have been made in elucidating the molecular basis of thyroid hormone synthesis and secretion, and in the molecular characterization of Mendelian disorders resulting in gross abnormalities of thyroid physiology (recently reviewed by Ref.14). By contrast, our knowledge of the genetic regulation of normal, quantitative variation in thyroid hormone levels among euthyroid individuals is extremely limited. The few studies that have examined the genetics of normal thyroid hormone variation have focused primarily on correlations among twin pairs (15, 16, 17) or have comprised family studies of variation of a single hormone level (17, 18, 19). An exception is the recent study by Peeters et al. (20) showing associations between plasma thyroid hormone levels and three single-nucleotide polymorphisms (SNPs) in two thyroid hormone pathway genes in 156 randomly ascertained, unrelated individuals. Until now, however, there have been no large family-based studies of the genetic control of multiple thyroid hormone levels, or of the relative contributions of genes and environmental factors to interindividual variation in these quantitative phenotypes. Consequently, we have little understanding of how genes that are involved in thyroid hormone production and metabolism interact to modulate serum concentrations of thyroid hormones in the general population. To better understand the contribution of genetic diversity to thyroid hormone variation and the potential effects of such variation on human health outcomes, we conducted a study of the inheritance of normal variation in seven thyroid related phenotypes in a large set of Mexican-American families from San Antonio, TX.


   Subjects and Methods
Top
 Abstract
 Introduction
 Subjects and Methods
Results
 Discussion
 References
 
Study population

We used data on 1,011 Mexican Americans (586 women and 425 men) from 27 extended families from the San Antonio Family Heart Study (SAFHS), a population-based prospective family study of atherosclerosis and its risk factors. These data were collected during the baseline phase of the SAFHS between 1991 and 1996. Probands for the SAFHS families were identified from a low-income neighborhood using a house-to-house recruitment procedure. Eligibility criteria for study probands were that they be 40–60 yr of age and have at least six first-degree relatives (excluding parents) living in the San Antonio area. All first-, second-, and third-degree relatives of the proband and the proband’s spouse were invited to participate; the invitation was extended regardless of the proband’s (or relatives’) medical history, including thyroid disease. However, information on disease status and medication use, including use of thyroid supplements/suppressors and other medications that impinge on thyroid function, was recorded for each participant subsequent to entering the study. Details of the sampling and recruitment procedures for the SAFHS have been described previously (21).

Participating subjects received a medical examination in the morning, after a 12-h fast. Fasting blood samples were collected between 0900 and 1100 h for subsequent biochemical analyses, and a 2-h glucose tolerance test was then performed after ingestion of a 75-g glucose equivalent load. Diabetes was diagnosed using the plasma glucose criteria of the World Health Organization (22) and self-report of current use of antidiabetic medications. The basic medical examination also included measurement of height and weight (both without shoes). Pregnant women were not eligible to participate in the blood drawing procedure or medical exam but were rescheduled for examination at least 3 months after their pregnancy. The Institutional Review Board at the University of Texas Health Science Center at San Antonio approved all procedures, and all subjects gave informed consent.

All prescription and nonprescription drug use was noted. Using this information, we identified 16 individuals taking thyroid supplements or suppressors and other drugs that alter thyroid hormone metabolism or production, such as lithium, furosemide, diphenyldantoin, and propanolol. These individuals were removed from the data set.

Covariates

A questionnaire was administered to obtain information about each subject’s medical history, medication use, dietary habits, physical activity patterns, and smoking and alcohol consumption behaviors. A reproductive history questionnaire that included questions about menstrual cycles and current use of oral contraceptives and estrogens was administered to all women. Women were considered to be menopausal if more than one year had elapsed since their last menstrual period or if they had undergone surgical menopause, defined as having both ovaries removed.

Physical activity was assessed using a modified version of the Stanford 7-d Physical Activity Recall Instrument (23). Subjects reported the weekly number of hours they slept and engaged in moderately strenuous, heavy, and very heavy physical activities. Examples of activities corresponding to each category were provided to assist the subject’s responses. Light physical activity is defined as the difference between the total possible hours of weekly activity (i.e. 7 d x 24 h/d = 168 h) and the number of hours accounted for by sleep and moderate, heavy, and very heavy activity. Each category of physical activity was scored in metabolic equivalents, or METS (one MET equals the energy expenditure of 1 kg body weight/h), and expressed on a per-week basis.

Thyroid phenotypes

Thyroid hormone phenotypes include serum concentrations of total T4 (TT4), free T4 (FT4), total T3 (TT3), and free T3 (FT3). Thyroid-related "hormone" phenotypes include serum concentrations of TSH, thyroglobulin (TG), and T4-binding globulin (TBG). Neither TG (thyroid hormone substrate and storage molecule) nor TBG (major thyroid hormone transport molecule) are true hormones; they are grouped with TSH as thyroid-related hormones for convenience of classification and discussion. These seven phenotypes are, herein, collectively referred to as thyroid phenotypes or traits.

As described above, blood samples were drawn from participants in the SAFHS, between 0900 and 1100 h, after an overnight fast. Serum samples were separated from cellular fractions and frozen in small aliquots at –80 C. Thyroid and thyroid-related hormone concentrations were determined in thawed serum sample aliquots by means of the following commercial RIA kits: TT3, DPC Coat–A-Count Total T3 kit (no. TKT35; Diagnostic Products Corp., Los Angeles, CA); FT3, DPC Coat–A-Count Free T3 kit (no. TKF35); TT4, DPC Coat–A-Count Total T4 kit (no. TKT45); FT4, DPC Coat–A-Count Free T4 kit (no. TKF45); TSH, Nichols TSH 100T kit (no. 40–2160; Nichols Institute Diagnostics, San Juan Capistrano, CA); TBG, DiaSorin GammaDab [125I] TBG kit (no. CA-1572; DiaSorin, Inc., Stillwater, MN); TG, DPC Thyroglobulin double antibody kit (no. KTGD1). All samples were also tested for the presence of antibodies against TG (TGAb) using the DPC IMMULITE Anti-TGAb kit (no. LKTG5). [We note that several different assay methods are used by the research and clinical diagnostic communities for the determination of serum total and free thyroid hormone concentrations. The accuracy and precision of these methodologies vary to some degree. Because all of our determinations for each hormone were performed using the same method for all study subjects, such variation will have no effect on our estimates of heritability (h2).]

Automated counting, standard curve determination, and sample estimation for the seven thyroid phenotypes were performed on an LKB-Wallac 1282 Compu-{gamma} counter (Wallac Oy, Turku, Finland). TGAb were measured on a DPC IMMULITE-1000 Immunofluorescence Analyzer. Intra- and interassay coefficients of variation (CV) for the seven thyroid phenotypes were: TT3, 5.9%, 8.7% (intraassay CV, interassay CV); FT3, 5.3%, 9.7%; TT4, 4.3%, 6.7%; FT4, 5.0%, 6.9%; TSH, 2.5%, 9.5%; TG, 7.5%, 10.9%; TBG, 2.5%, 7.8%. Individuals with elevated TGAb (>20 IU/ml), which could interfere with the TG assays, were eliminated from the TG data analyses.

Before statistical and genetic analysis, the distribution of each thyroid phenotype was examined for normality. TSH and TG were square-root transformed to reduce skewness and kurtosis. The analysis of each thyroid phenotype was restricted to those individuals for whom all covariate data were complete.

Statistical analyses

The overall aim of our analyses was to determine the extent to which genes and measured environmental factors contribute to variation in serum thyroid and thyroid-related hormone levels. We used quantitative genetic methods to model the total variation in a trait as a function of the mean trait value, effects attributable to the measured covariates, and the proportions of the remaining variation that could be attributed to the residual genetic and unmeasured environmental effects, respectively. For each thyroid trait, we estimated sex-specific effects of age (linear and quadratic), diabetes status, height, weight, smoking status (current smoker vs. not), alcohol consumption (current drinker vs. not), and physical activity level (in METS). In women, we also estimated effects associated with the following variables: menopausal status (postmenopausal vs. not), oral contraceptive use (current user vs. not), hormone replacement therapy (current user vs. not), ovarectomy, and hysterectomy.

To determine which covariates had significant effects on the thyroid phenotypes, we performed multivariate linear regression using a stepwise (forward and backward) procedure. The correlation between covariates was examined to determine whether there were colinearity effects, and collinear variables were removed in a backward-stepwise procedure. All multivariate regressions were compared to derive a consensus set of covariates that had significant effects on the thyroid phenotypes. Because our primary objective was to estimate the proportion of unexplained variance that could be explained by genetic factors, rather than the rigorous exclusion of specific covariates, we used a liberal threshold of P ≤ 0.1 for significance in these analyses. All analyses were performed individually for each thyroid phenotype, such that a custom set of covariates was determined for each phenotype. Residuals from the multivariate regressions were then used as covariate-adjusted trait values to estimate correlations and residual additive genetic heritabilities. Total additive genetic h2 was calculated as the proportion of variance attributable to the additive genetic effects of multiple genes (polygenes) divided by the total variance (i.e. 1.00 plus the proportion of variance due to covariates).

After determining which covariate effects were significant, and regressing out their effects, genetic effects were estimated using a pedigree-based likelihood approach (24, 25). Only additive polygenic effects were estimated so that we defined residual h2 as the proportion of the total residual trait variance ({varsigma}2T) attributable to the additive effects of polygenes ({varsigma}2G). In this model, the additive genetic variance plus the environmental variance equals the total residual phenotypic variance. Estimation of the additive genetic h2 follows basic quantitative genetic theory, which models the phenotypic covariances between two individuals in a pedigree as a function of their degree of biologic relatedness. Maximum likelihood methods were used to estimate the values of the parameters (including the h2) that resulted in highest likelihood obtained across all of the pedigrees. A separate analysis was also performed in which the covariate effects were estimated simultaneously with the additive genetic effects. Both analyses (estimating the covariate and genetic effects simultaneously, and regressing out the covariate effects first then estimating the genetic effects) produced similar results. These analyses were conducted using the SOLAR software program (Ref.26 ; see also http://www.sfbr.org/sfbr/public/software/solar/index.html).

The final models included all covariates that were significantly associated with the thyroid and thyroid-related traits, as well as an additive genetic effect modeled as a random effect. The relative proportions of the variance explained by the measured covariates and genes were calculated as the variance attributable to that particular component divided by the total phenotypic variance. The residual variance that was not accounted for by the measured environmental and additive genetic components is the variance attributable to unmeasured environmental factors (including error) and nonadditive genetic effects (such as genotype x environment interaction, dominance, and others).


   Results
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
Discussion
 References
 
Thyroid and thyroid-related hormone phenotypes were available for a total of 1,011 individuals in 27 two- and three-generation pedigrees that ranged in size from 6–83 individuals, with a median size of 33 individuals. Detailed descriptions of the characteristics of the SAFHS families have been presented elsewhere (21). Briefly, there were 586 women and 425 men in the sample, with a mean age of 39.3 and 38.9 yr, respectively (range, 16–94 yr), and mean body mass index of 29.9 and 27.8 kg/m2, respectively (range, 16.1–62.6 kg/m2). Approximately 15% of the family members were diabetics, 16% of the women and 33% of the men were smokers, 27% of the women were postmenopausal, and 11% of the women were taking oral contraceptives.

Serum concentrations (±SD) of the seven thyroid and thyroid-related hormones in the combined SAFHS population sample were: TT3, 1.94 ± 0.42 nmol/liter; FT3, 5.31 ± 1.74 pmol/liter; TT4, 95.2 ± 24.8 nmol/liter; FT4, 17.5 ± 4.8 pmol/liter; TSH, 1.53 ± 1.47 mIU/liter; TBG, 31.11 ± 8.76 mg/liter; and TG, 8.40 ± 5.15 µg/liter. The mean TSH level was comparable with that reported for the Mexican-American "disease-free" population (1.43 ± 0.03 mIU/liter) in the large National Health and Nutrition Examination Survey (NHANES) III study (27), which is the population subset in that study that is most similar in composition to the SAFHS sample. TT4 level was not reported for the NHANES III disease-free Mexican-American population; but in the "reference" population (a subset of the disease-free population), mean TT4 concentration was 116.3 ± 0.7 nmol/liter, which is somewhat higher than that in the SAFHS population.

We detected no individuals meeting the stringent NHANES III criteria for clinical or subclinical hyperthyroidism (TSH < 0.1 mIU/liter). However, based on the commonly used laboratory reference of 0.4 mIU/liter as a normal lower limit for TSH, 3.1% of the SAFHS population sample (2.2% of men and 3.8% of women) would be classified as clinically (TT4 ≥ 169.9 mIU/liter) or subclinically (TT4 < 169.9 mIU/liter) hyperthyroid. This is similar to the 2.1% reported for the NHANES III disease-free Mexican-American population (Ref.27 ; see Table 3Go). Elevated TSH indicative of clinical or subclinical hypothyroidism (>4.5 mIU/liter) was less common in the SAFHS population than in the NHANES III disease-free Mexican-American population (1.4 vs. 3.8%, respectively). Based on criteria for clinical hypothyroidism (TSH > 4.5 mIU/liter and TT4 < 57.9 nmol/liter) and subclinical hypothyroidism (TSH > 4.5 mIU/liter and TT4 ≥ 57.9 nmol/liter) used in the NHANES III study, this between-population difference is attributable entirely to subclinical hypothyroidism rates (1.2 vs. 3.6%, respectively); the clinical hypothyroidism rates were identical (0.2%) in both populations. We have no explanation for the lower incidence of subclinical hypothyroidism in the SAFHS population relative to the NHANES III population. As has been reported for several populations (Table 3Go in Ref.27), elevated TSH was more common in women than in men (2.0 vs. 0.6%, respectively) in the SAFHS sample.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Statistically significant1 covariate effects in Mexican-American men and women

 
We also detected 46 individuals (27 women and 19 men) with elevated TG antibodies among the 647 individuals for whom we had antibody data (393 women and 254 men). This proportion (7.1%) is slightly lower than that observed in the NHANES III Mexican-American disease-free subpopulation (8.2%) (27). Overall, then, clinically important thyroid function indicators (TSH and TT4), rates of overt and subclinical hyper- and hypothyroidism, and evidence of thyroid autoimmune disease in our study population were generally similar to those reported for the NHANES III Mexican-American population.

Considering these data by sex and age (Table 1Go) reveals some general patterns in thyroid and thyroid-related hormone levels between males and females and among four different age segments of the population. For example, mean TSH level increases with increasing age in both sexes, although this increase is greater in women; and TSH levels are higher in women than in men in all but the oldest age group. Mean TT4 and TBG are higher in females across all age groups, whereas FT4 is slightly higher in males of all age groups. FT3 exhibits little sex-specific disparity at any age. Mean levels of TT4, FT4, TT3, and FT3 tend to decrease with age in both sexes. This trend is especially obvious in the oldest age groups.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Thyroid and thyroid-related hormone levels in Mexican Americans

 
Analyses of covariates

Statistical analyses of the possible covariates affecting the thyroid phenotypes (Table 2Go) in the SAFHS cohort overall revealed that females had significantly lower FT3 and FT4 (as indicated by the negative regression coefficients), but higher TBG and TG, than males. In addition, FT3 and FT4 significantly decreased with increasing age, although TG and TBG increased. As revealed by the significant sex x age and sex x age2 interactions, TSH concentrations significantly increased with increasing age in younger women and then decreased in older women; but in men, there was no significant relationship. These interaction effects are the cause of the apparent paradox that females have higher mean TSH than males in our age groups (Table 1Go), but that the regression coefficient for sex is negative (Table 3Go). As can be seen from Table 1Go, TSH in females decreases rapidly with decreasing age and, at age zero (e.g. birth), would be predicted to be lower in females than in males. The value of the regression coefficient for sex (Table 2Go) represents the effect of sex at the intercept (where age equals zero).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Covariate effects on thyroid phenotypes in Mexican Americans

 
In comparison with men and women without type 2 diabetes, type 2 diabetic men had significantly higher TT3, whereas type 2 diabetic women had lower TT3. Postmenopausal women had higher FT3 than premenopausal women, and women on oral contraceptives had highly significantly increased levels of FT4, TT3, TT4, and TBG. TT3, TT4, and TBG decreased with increasing alcohol consumption, and TT4 levels decreased with increasing physical activity. As reported by Sepkovic et al. (28), we detected an effect of current smoking on TT4 levels, but it was opposite in sign to the effect reported in that study. Specifically, TT4 levels were higher, rather than lower, in SAFHS smokers. Unlike Sepkovic et al. (28), we did not observe any effect of current smoking on TT3 levels. One of the reasons for these discrepancies might be that we were analyzing data across all ages and both sexes in a different population and only used self-report data on smoking. Although some of these covariates, such as oral contraceptives, had highly significant effects on thyroid phenotypes, as a group the statistically significant, measured covariates accounted for an average of only 9% [range, 2% (for TSH) to 18% (for TBG), Table 4Go] of the total variation in serum levels of the seven thyroid and thyroid-related hormones.


View this table:
[in this window]
[in a new window]
  		TABLE 4. Genetic and environmental contributions to variation in thyroid phenotypes in Mexican Americans

 
We also analyzed the effects of external covariates, including age and sex, in men and women separately (Table 3Go). As expected, we observed covariate effects in these sex- specific analyses that were similar to those that were observed overall. For example, age and/or age2 had significant effects on all thyroid phenotypes. Increased alcohol consumption decreased TT4 and TBG in both sexes and increased TT3 in women but not in men. In addition, increased physical activity decreased TT4 in women but not in men.

Heritabilities of thyroid and thyroid-related hormone phenotypes

We considered two measures of the genetic component of phenotypic variation. Total h2 is the proportion of total phenotypic variation that is attributable to additive polygenic effects and is an indicator of the importance of genetic differences, relative to all nongenetic effects, in determining interindividual phenotypic differences. Residual h2 is the proportion of phenotypic variation attributable to additive polygenic effects after removing the effects of significant covariates. Measures of residual h2 enable comparison of the relative importance of genetic effects between different phenotypes in the absence of measured covariate effects.

Residual heritabilities of the seven thyroid phenotypes (Table 2Go) ranged from 0.31 ± 0.06 for TBG to 0.67 ± 0.06 for FT3, indicating that interindividual variation in all of these hormone traits is moderately to strongly influenced by genetic differences among individuals. Women exhibited higher residual heritabilities for FT3, FT4, TSH, TG, and TBG, and lower residual heritabilities for TT3 and TT4 than men (Table 3Go).

Compared with the impact of genetic variation, the measured environmental covariates contributed relatively little to hormone level differences between individuals (Table 4Go). The proportion of phenotypic variation explained by the covariates for both sexes combined ranged from 0.022 (for TSH) to 0.176 (for TBG), whereas total heritabilities ranged from 0.260 for TBG to 0.635 for FT3. Thus, the variation in measured environmental covariates contributes an average of 28.9% (range, 6.6–67.7%) as much to total variation in the hormone levels as do genetic factors. However, the influence of measured external covariates on hormone level variation was not equal between the sexes. For all hormones except FT3, the proportion of total phenotypic variance attributable to the measured covariates was higher in females than in males, suggesting that thyroid and thyroid-related hormone regulation in males is less influenced by extrinsic factors than it is in females. For FT3, the opposite appears to be true.

TSH exhibited moderate h2 in the overall population (total h2 = 0.317, Table 4Go); however, when considered by sex, the estimated h2 in women was almost twice that in men (0.572 vs. 0.293, respectively). Because TSH tends to increase in women with increasing age, and some of this increase is conjectured to be attributable to an increase in autoimmune thyroiditis, we reanalyzed the TSH data, overall and in men and women separately, after excluding elevated TSH values (TSH > 4.5 mIU/liter). The h2 estimates, using data on all individuals with TSH ≤ 4.5, were similar in all three groups: 0.20 ± 0.07 overall, 0.25 ± 0.13 in men, and 0.20 ± 0.12 in women. These results suggest: 1) that h2 of nonelevated TSH is similar in men and women; and 2) that increased h2 in women, when all TSH values are included, may be attributable to a heritable autoimmune component.

Correlations among thyroid and thyroid-related hormone phenotypes

We also estimated the residual phenotypic correlations among the seven thyroid phenotypes after removing the effects of significant covariates (Table 5Go). As expected from the known biological relationships among these hormones, there are moderate correlations among TT3, FT3, TT4, and FT4, ranging from 0.17 between TT3 and FT4 to 0.39 between TT4 and FT4. TBG was moderately correlated with TT3 and TT4, whereas TG and TSH had little to no phenotypic correlation with the other hormones.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Residual phenotypic correlations among thyroid phenotypes in Mexican Americans

 

   Discussion
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
References
 
Interest in identifying genes and environmental factors that influence thyroid hormone function is increasing, especially given recent reports that subclinical alterations in thyroid hormone levels can have detrimental effects on cardiac function (8, 9, 12, 29, 30); alter serum lipoproteins, triglycerides, and other risk factors for cardiovascular disease (4, 5, 6, 7, 31); and increase risk of atherosclerosis and myocardial infarction (11), coronary and peripheral artery disease (5, 31), reduced bone mineral density and osteoporosis (32, 33, 34, 35), and cognitive dysfunction (36, 37). To date, most genetic studies of thyroid hormone function have focused on molecular and biochemical analyses of Mendelian disorders (14). Although these disorders are individually rare, they have provided useful insights regarding the molecular basis of the pathophysiology of the thyroid system. Recently, Peeters et al. (20) reported that SNPs in the DIO1 (iodothyronine deiodinase 1) and TSH receptor genes are associated with differences in plasma TSH and iodothyronine levels in clinically normal individuals. This study shows that simple genetic polymorphisms can affect normal variation in serum concentrations of thyroid or thyroid-related hormones. However, with the exception of a study of genetic and environmental factors that affect T4 (assumed by us to be TT4) in small Mennonite families (18), these studies do not indicate the extent to which genes and measured environmental factors influence variation in thyroid and thyroid-related hormone levels in the general population. Our study addresses this information gap by estimating the relative effects of genes and external environmental factors on normal population level variation in thyroid and thyroid-related hormone phenotypes in a large set of randomly ascertained Mexican-American families.

The present study yields two major findings. First, normal phenotypic variation in serum concentrations of the primary thyroid hormones (T3 and T4), the chief regulator of thyroid hormone synthesis and secretion (TSH), the main thyroid hormone substrate molecule (TG), and the major thyroid hormone serum transport protein (TBG) in this Mexican-American population are all moderately to strongly influenced by genetic, or more precisely, familial differences among individuals. Our methods do not distinguish shared environmental (household) or genotype x environment interaction effects from our estimates of h2. To the extent that such effects occur in our dataset and mimic additive genetic effects, they will be included in our estimates of h2. However, given that we have analyzed large, multigenerational families, each of which is subdivided among many individual households, such effects are likely to be minimal. Second, the genetic contribution to this variation is considerably greater than the effects of several physiologic and extrinsic factors (collectively referred to as environmental covariates) that are widely viewed as important determinants of thyroid hormone regulation. Specifically, the additive effects of genes (total heritabilities) range from 26–64% of the total interindividual variation observed in this population, whereas the effects of environmental covariates such as sex, age, diabetes status, and a broad range of lifestyle covariates such as smoking and alcohol consumption habits, physical exercise, oral contraceptive use, and other factors account for only 2–18% of the total phenotypic variation.

This is not the first study to detect genetic influences on normal thyroid hormone phenotypes. Based on comparisons of phenotypic similarities between monozygotic and dizygotic twin pairs, Meikle et al. (16) estimated that genetic differences accounted for a substantial proportion of the variation between individuals in levels of TT4 (h2 = 38%) and FT4 (h2 = 44%) but were insignificant for TT3, TSH, and TBG. Similarly, Poehlman et al. (15) found evidence of significant correlations between twins in the responses of their TT4 and FT4 levels to physical training and estimated heritabilities of 0.42–0.71 for both of these responses. However, they too failed to find any significant h2 for T3 (TT3 or FT3). In the only family-based study of thyroid hormone variation, Martin and Crawford (18) estimated h2 of TT4 to be 31% in Mennonites living in Kansas and Nebraska and also estimated that the measured environmental factors accounted for 6% of the total variation. These results are similar to what we observe in our Mexican-American families. TBG variation is known to be strongly influenced by a broad spectrum of individually rare TBG structural gene variants (reviewed by Refs.14 and 38), and the high h2 (47%) of quantitative TBG variation in a small set of Swedish families is attributable to variation at this X-linked locus (19). Based on preliminary linkage analyses, the TBG variation in our study appears to be independent of the X-linked TBG locus (results not shown). Peeters et al. (20) found that normal variation in TSH and rT3 (a metabolite of T4) levels among 155 normal blood donors was correlated with the presence and zygosity level of SNPs in the TSH receptor and DIO1 genes, respectively, but found no effect of SNPs in the DIO2, DIO3, or TRß (T3 receptor ß) genes on any of the measured hormone levels (TT4, FT4, TT3, TSH, or rT3).

Thus, ours is the first large-scale study that examines variation in multiple thyroid and thyroid-related hormone phenotypes simultaneously in a large set of randomly ascertained families, and the first to consider the effects of a broad range of environmental covariates on these phenotypes. This study design enables the direct comparison of the relative importance of genetic vs. measured environmental factors as underlying causes of this phenotypic variation, and the relative magnitudes of genetic contributions to variation in the serum concentrations of the seven different thyroid and thyroid-related hormones, after correcting for significant extrinsic covariate effects. Similar to previous reports, our study found T4 variation to be significantly heritable (TT4 total h2 = 31%; FT4 total h2 = 35%); but in contrast to those reports, we also found substantial h2 for TT3 (total h2 = 30%), FT3 (total h2 = 64%), TSH (total h2 = 32%), TG (total h2 = 39%), and apparent non-X-linked variation in TBG (total h2 = 26%). Moreover, the effects of the measured covariates were smaller than the effects of genes for each of the seven phenotypes (Table 4Go). An important next step will be to determine whether the genetic components of variation among these various thyroid and thyroid-related phenotypes might be due to a common or overlapping set (or sets) of genes or are largely genetically independent. Analyses to discriminate between these possibilities are beyond the scope of the present report.

The residual heritabilities for the six thyroid phenotypes excluding FT3 range from 30.6–36.6%, compared with 67.2% for FT3 (Table 2Go). This indicates that inheritance is approximately twice as important in determining FT3 levels as it is for any of the other thyroid phenotypes. The high residual h2 of FT3 is due in part to its particularly high value in females (residual h2 = 0.773), but it is substantial in males as well (residual h2 = 0.482, Table 3Go). This observation is intriguing because FT3 is the most physiologically active of the thyroid hormones and, therefore, the most relevant measure of thyroid hormone action on cellular activity, metabolism, and regulation. Thus, it may be the most important measure with regard to potential effects of thyroid variation on health-related characteristics such as cardiac function, blood pressure regulation, serum lipid metabolism and profiles, bone metabolism and deposition, fat and protein metabolism and distribution, cognitive function, and more. The very strong h2 of FT3 detected by the present study implies that familial-based variation in FT3 levels could have important implications for understanding inherited differences in susceptibilities to a variety of abnormal physiologic conditions that contribute to frank disease.

The hypothalamic-pituitary-thyroid axis comprises the basic unit of neuroendocrine regulation of thyroid hormone synthesis and secretion. It comprises a tightly connected series of negative feedback systems that determine the levels of thyroid hormones in the blood and tissues that, in turn, regulate cellular metabolism within a relatively narrow window of acceptable function. The production of thyroid hormones resultant of hypothalamic-pituitary-thyroid interactions, balanced against the metabolic degradation and clearance of thyroid hormones and recycling of inorganic iodide, determines the concentrations of thyroid hormones in the serum and intracellular compartments of the body (39). The demonstration that seven central thyroid phenotypes, six of which are controlled by this tightly regulated feedback system (TBG is not), are moderately to strongly heritable suggests that genetic differences between individuals are important in determining interindividual differences in thyroid hormone levels, and that there is no single, so-called normal, composite thyroid hormone phenotype. Rather, thyroid hormone setpoints vary normally among healthy (euthryroid) individuals in a population, in large part due to normal genetic differences among those individuals. To the extent that these differences contribute to variation in other health-related characteristics, better understanding of the role of inherited differences in thyroid hormone homeostasis will broaden our comprehension of the connection between thyroid function, general health, and disease.


   Acknowledgments
 
We are deeply grateful for the cooperation of the families participating in the SAFHS. We also thank Theresa Scott for assistance with some of the analyses.


   Footnotes
 
This work was supported by Research Grants RO1-DK57003 (to P.B.S.) and PO1-HL45522 awarded by the National Institutes of Health.

Abbreviations: CV, Coefficient(s) of variation; FT3, free T3; FT4, free T4; h2, heritability; NHANES, National Health and Nutrition Examination Survey; SAFHS, San Antonio Family Heart Study; SNP, single-nucleotide polymorphism; TBG, T4-binding globulin; TG, thyroglobulin; TGAb, antibodies against TG; TT3, total T3; TT4, total T4.

Received October 1, 2003.

Accepted March 19, 2004.


   References
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 References
 

  1. Green WL 1987 Physiology of the thyroid gland and its hormones. In: Green WL, ed. The thyroid. New York: Elsevier; 1–46
  2. Larsen PR 2003 The thyroid gland. In: Larsen PR, Ronenberg HM, Melmed S, Polonsky KS, eds. Williams textbook of endocrinology. 10th ed. Philadelphia: WB Saunders Company; 331–390
  3. Kung AW, Pang RW, Janus ED 1995 Elevated serum lipoprotein(a) in subclinical hypothyroidism. Clin Endocrinol (Oxf) 43:445–449[Medline]
  4. Duntas LH 2002 Thyroid disease and lipids. Thyroid 12:287–293[CrossRef][Medline]
  5. Mya MM, Aronow WS 2002 Subclinical hypothyroidism is associated with coronary artery disease in older persons. J Gerontol A Biol Sci Med Sci 57:M658–M659
  6. Luboshitzky R, Aviv A, Herer P, Lavie L 2002 Risk factors for cardiovascular disease in women with subclinical hypothyroidism. Thyroid 12:421–425[CrossRef][Medline]
  7. Ineck BA, Ng TM 2003 Effects of subclinical hypothyroidism and its treatment on serum lipids. Ann Pharmacother 37:725–730[Abstract/Free Full Text]
  8. Auer J, Scheibner P, Mische T, Langsteger W, Eber O, Eber B 2001 Subclinical hyperthyroidism as a risk factor for atrial fibrillation. Am Heart J 142:838–842[CrossRef][Medline]
  9. Sawin CT, Geller A, Wolf PA, Belanger AJ, Baker E, Bacharach P, Wilson PW, Benjamin EJ, D’Agostino RB 1994 Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N Engl J Med 331:1249–1252[Abstract/Free Full Text]
  10. Biondi B, Palmieri EA, Filetti S, Lombardi G, Fazio S 2002 Mortality in elderly patients with subclinical hyperthyroidism. Lancet 359:799–800
  11. Hak AE, Pols HA, Visser TJ, Drexhage HA, Hofman A, Witteman JC 2000 Subclinical hypothyroidism is an independent risk factor for atherosclerosis and myocardial infarction in elderly women: the Rotterdam Study. Ann Intern Med 132:270–278[Abstract/Free Full Text]
  12. Osman F, Gammage MD, Franklyn JA 2002 Hyperthyroidism and cardiovascular morbidity and mortality. Thyroid 12:483–487[CrossRef][Medline]
  13. Comuzzie AG, Blangero J, Mahaney MC, Sharp RM, VandeBerg JL, Stern MP, MacCluer JW 1996 Triiodothyronine exerts a major pleiotropic effect on reverse cholesterol transport phenotypes. Arterioscler Thromb Vasc Biol 16:289–293[Abstract/Free Full Text]
  14. Medeiros-Neto G, Knobel M, Degroot LJ 2002 Genetic disorders of the thyroid hormone system. In: Baxter JD, Melmed S, New M, eds. Genetics in endocrinology. New York: Lippincott Williams, Wilkins; 375–402
  15. Poehlman ET, Tremblay A, Nadeau A, Dussault J, Theriault G, Bouchard C 1986 Heredity and changes in hormones and metabolic rates with short-term training. Am J Physiol 250:E711–E717
  16. Meikle AW, Stringham JD, Woodward MG, Nelson JC 1988 Hereditary and environmental influences on the variation of thyroid hormones in normal male twins. J Clin Endocrinol Metab 66:588–592[Abstract/Free Full Text]
  17. Premawardhana LD, Lo SS, Phillips DI, Prentice LM, Rees Smith B 1994 Variability of serum thyroglobulin levels is determined by a major gene. Clin Endocrinol (Oxf) 41:725–729[Medline]
  18. Martin LJ, Crawford MH 1998 Genetic and environmental components of thyroxine variation in Mennonites from Kansas and Nebraska. Hum Biol 70:745–760[Medline]
  19. Kollind M, Iselius L, Pettersson T, Adamson U, Carlstrom A 1985 Genetics and clinical significance of thyroxine binding globulin deficiency, an analysis of seven families. Acta Endocrinol (Copenh) 109:83–89[Abstract/Free Full Text]
  20. Peeters RP, van Toor H, Klootwijk W, de Rijke YB, Kuiper GG, Uitterlinden AG, Visser TJ 2003 Polymorphisms in thyroid hormone pathway genes are associated with plasma TSH and iodothyronine levels in healthy subjects. J Clin Endocrinol Metab 88:2880–2888[Abstract/Free Full Text]
  21. Mitchell BD, Kammerer CM, Blangero J, Mahaney MC, Rainwater DL, Dyke B, Hixson JE, Henkel RD, Sharp RM, Comuzzie AG, VandeBerg JL, Stern MP, MacCluer JW 1996 Genetic and environmental contributions to cardiovascular risk factors in Mexican Americans. The San Antonio Family Heart Study. Circulation 94:2159–2170[Abstract/Free Full Text]
  22. World Health Organization Expert Committee 1980 Second report on diabetes mellitus. World Health Organization Technical Report Series No. 646. Geneva: WHO
  23. Blair SN, Haskell WL, Ho P, Paffenbarger Jr RS, Vranizan KM, Farquhar JW, Wood PD 1985 Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments. Am J Epidemiol 122:794–804[Abstract/Free Full Text]
  24. Lange K, Westlake J, Spence MA 1976 Extensions to pedigree analysis. II. Recurrence risk calculation under the polygenic threshold model. Hum Hered 26:337–348[CrossRef][Medline]
  25. Hopper JL, Mathews JD 1982 Extensions to multivariate normal models for pedigree analysis. Ann Hum Genet 46:373–383
  26. Almasy L, Blangero J 1998 Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 62:1198–1211[CrossRef][Medline]
  27. Hollowell JG, Staehling NW, Flanders WD, Hannon WH, Gunter EW, Spencer CA, Braverman LE 2002 Serum TSH, T(4), and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab 87:489–499[Abstract/Free Full Text]
  28. Sepkovic DW, Haley NJ, Wynder EL 1984 Thyroid activity in cigarette smokers. Arch Intern Med 144:501–503[Abstract/Free Full Text]
  29. Biondi B, Palmieri EA, Lombardi G, Fazio S 2002 Effects of subclinical thyroid dysfunction on the heart. Ann Intern Med 137:904–914[Abstract/Free Full Text]
  30. Biondi B, Palmieri EA, Lombardi G, Fazio S 2002 Subclinical hypothyroidism and cardiac function. Thyroid 12:505–510[CrossRef][Medline]
  31. Mya MM, Aronow WS 2003 Increased prevalence of peripheral arterial disease in older men and women with subclinical hypothyroidism. J Gerontol A Biol Sci Med Sci 58:68–69
  32. Bertoli A, Fusco A, Andreoli A, Magnani A, Tulli A, Lauro D, De Lorenzo A 2002 Effect of subclinical hypothyroidism and obesity on whole-body and regional bone mineral content. Horm Res 57:79–84
  33. Foldes J, Tarjan G, Szathmari M, Varga F, Krasznai I, Horvath C 1993 Bone mineral density in patients with endogenous subclinical hyperthyroidism: is this thyroid status a risk factor for osteoporosis? Clin Endocrinol (Oxf) 39:521–527[Medline]
  34. Mudde AH, Reijnders FJ, Kruseman AC 1992 Peripheral bone density in women with untreated multinodular goitre. Clin Endocrinol (Oxf) 37:35–39[Medline]
  35. Uzzan B, Campos J, Cucherat M, Nony P, Boissel JP, Perret GY 1996 Effects on bone mass of long term treatment with thyroid hormones: a meta-analysis. J Clin Endocrinol Metab 81:4278–4289[Abstract]
  36. Jackson IM 1998 The thyroid axis and depression. Thyroid 8:951–956[Medline]
  37. Jensovsky J, Ruzicka E, Spackova N, Hejdukova B 2002 Changes of event related potential and cognitive processes in patients with subclinical hypothyroidism after thyroxine treatment. Endocr Regul 36:115–122[Medline]
  38. Hayashi Y, Refetoff S 1995 Genetic abnormalities of thyroid hormone transport serum proteins. In: Weintraub BD, ed. Molecular endocrinology: basic concepts and clinical correlations. New York: Raven Press, Ltd.; 371–387
  39. Chopra IJ, Sabatino L 2000 Nature and sources of circulating thyroid hormones. In: Braverman LE, Utiger RD, eds. Werner and Ingbar’s the thyroid. 8th ed. Philadelphia: Lippincott Williams, Wilkins; 121–135



This article has been cited by other articles:


Home page
 J. Clin. Endocrinol. Metab.Home page
G. Gelwane, N. de Roux, D. Chevenne, J. C. Carel, and J. Leger
Pituitary-Thyroid Feedback in a Patient with a Sporadic Activating Thyrotropin (TSH) Receptor Mutation: Implication That Thyroid-Secreted Factors Other Than Thyroid Hormones Contribute to Serum TSH Levels
J. Clin. Endocrinol. Metab., August 1, 2009; 94(8): 2787 - 2791.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
V. Panicker, S. G. Wilson, T. D. Spector, S. J. Brown, B. S. Kato, P. W. Reed, M. Falchi, J. B. Richards, G. L. Surdulescu, E. M. Lim, et al.
Genetic Loci Linked to Pituitary-Thyroid Axis Set Points: A Genome-Wide Scan of a Large Twin Cohort
J. Clin. Endocrinol. Metab., September 1, 2008; 93(9): 3519 - 3523.
[Abstract] [Full Text] [PDF]

Home page
 EndocrinologyHome page
L. C. Moeller, M. Alonso, X. Liao, V. Broach, A. Dumitrescu, J. Van Sande, L. Montanelli, S. Skjei, C. Goodwin, H. Grasberger, et al.
Pituitary-Thyroid Setpoint and Thyrotropin Receptor Expression in Consomic Rats
Endocrinology, October 1, 2007; 148(10): 4727 - 4733.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Physiol. Endocrinol. Metab.Home page
P. S. Hansen, T. H. Brix, I. Iachine, T. I. A. Sorensen, K. O. Kyvik, and L. Hegedus
Genetic and environmental interrelations between measurements of thyroid function in a healthy Danish twin population
Am J Physiol Endocrinol Metab, March 1, 2007; 292(3): E765 - E770.
[Abstract] [Full Text] [PDF]

Home page
 Eur J EndocrinolHome page
R. P Peeters, W. M van der Deure, and T. J Visser
Genetic variation in thyroid hormone pathway genes; polymorphisms in the TSH receptor and the iodothyronine deiodinases.
Eur. J. Endocrinol., November 1, 2006; 155(5): 655 - 662.
[Abstract] [Full Text] [PDF]

Home page
 EndocrinologyHome page
A. E. Baum, L. C. Solberg, P. Kopp, N. Ahmadiyeh, G. Churchill, J. S. Takahashi, J. L. Jameson, and E. E. Redei
Quantitative Trait Loci Associated with Elevated Thyroid-Stimulating Hormone in the Wistar-Kyoto Rat
Endocrinology, February 1, 2005; 146(2): 870 - 878.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Samollow, P. B.
Right arrow Articles by Barmada, M. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Samollow, P. B.
Right arrow Articles by Barmada, M. M.

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