Common Variants in the G Protein β3 Subunit Gene and Thyroid Disorders in a Formerly Iodine-Deficient Population

Introduction

G protein–mediated signaling is of key importance in the control of thyroid function. The thyroid-stimulating hormone (TSH) receptor, a G protein–coupled receptor, activates all four classes of G proteins to regulate downstream effectors, including phospholipase C, adenylylcyclases, and kinase pathways, which finally mediate the TSH signal into iodine uptake, thyroid hormone synthesis and secretion, and thyroid cell proliferation (1). Somatic gain-of-function mutations in the TSH receptor gene are one cause for sporadic hyperfunctional thyroid adenomas, while rare cases of germline mutations result in familial and sporadic nonautoimmune hyperthyroidism (2,3). Similarly, somatic gain-of-function mutations in the gene for the G protein α_S subunit (GNAS) have been identified in hyperfunctional thyroid adenomas, autonomous multinodular goiter, thyroid carcinomas, and other endocrinopathies (2 –6). The relevance of G protein–coupled mechanisms for thyroid disorders is further underscored by reports of altered expression or function of regulators of G protein signaling proteins (7) and G protein receptor kinases (8). Together, a complex scenario for G protein–mediated control of thyroid function and proliferation emerges where stimulatory signals are counter-regulated by inhibitory mechanisms (9).

G proteins are composed of a Gα subunit and a Gβγ dimer (1). G protein–coupled receptor stimulation leads to separation of the activated Gα subunit from the Gβγ dimer that in turn regulates different effectors including adenylylcyclases, ion channels, phospholipases, and kinase cascades (1). A common variant (825C>T, rs5443) in the G protein β3 subunit gene (GNB3) has been associated with preferential generation of Gβ3 splice variants (Gβ3s, Gβ3s2) conferring enhanced signal transduction (10,11). GNB3 alleles have been associated with complex phenotypes comprising hypertension, atrial fibrillation, obesity, metabolic abnormalities, functional gastrointestinal disorders, depression, and altered drug responses, although not all elements have been replicated (11 –16). Although such GNB3-associated phenotypes bear some resemblance to pleiotropic symptoms of thyroid disorders, evidence for a role of GNB3 variants to the pathogenesis of thyroid diseases is limited to data from a collection of histological specimens. In this series of 361 consecutive thyroid nodules, a significant increase in GNB3 825CC genotypes was observed in 80 adenomas compared with healthy Caucasian volunteers. The genotype distributions of samples from follicular and papillary thyroid carcinomas, however, did not differ from healthy individuals (17). An extended investigation of this series resulted in the perception that GNB3 825CC genotypes predominate in or favor oncocytic architectures or tumor subtypes (18). However, functional data of the investigated tumors were missing in these studies.

So far, only a fraction of ∼50% of thyroid adenomas is pathogenetically explained by somatic mutations in the TSH receptor or in GNAS (2,3). Although rare cases of germline mutations in familial nonautoimmune hyperthyroidism have been identified (3,5), the contribution of common germline variants for the pathogenesis of thyroid disorders has not extensively been studied. Against the background of an association of GNB3 variants with thyroid nodules, the importance of G protein signaling for thyroid function and the occurrence of complex GNB3 variant phenotypes that resemble components of thyroid disorders, we wondered whether GNB3 germline variants are involved in the pathogenesis of thyroid diseases. Although the reports by Sheu et al. (17,18) focused on overt pathologically verified thyroid tumors, it remains elusive how representative their data are for the general population. Development and growth of thyroid adenomas is a continuing process, from subclinical abnormalities to manifest diseases. Here, we analyzed in the Study of Health in Pomerania (SHIP) whether GNB3 variants are associated with subclinical and overt thyroid disorders. SHIP is a cross-sectional population-based survey from a formerly iodine-deficient German region that has been designed to assess general health issues with thyroid disorders being a major focus (19).

Subjects and Methods

Based on the population registry of 212,157 inhabitants in a region in northeast Germany (West Pomerania), a final representative sample of 4310 participants aged 20–79 was included into SHIP. The design of the study and the recruitment procedures have been detailed previously (19). The study region is a formerly iodine-deficient area with a high prevalence of iodine deficiency–related disorders including goiter, thyroid nodules, and decreased serum TSH levels. All participants were of German citizenship. The study followed the recommendations of the Declaration of Helsinki and was approved by the local Ethics Committee of the University of Greifswald. All participants gave written informed consent.

Results

From the total SHIP population of 4310 individuals, a group of 423 participants (82 men and 341 women) reported a history of thyroid disorders or current use of antithyroid medication (ATC classification H03). Frequencies of GNB3 825CC, CT, and TT genotypes were 50.1%, 39.4%, and 10.4% in subjects with and 47.1%, 42.7%, and 10.2% in subjects without history of thyroid disorders (p = 0.463), respectively. The SHIP study design did not allow in such cases to contact the attending physicians to clarify diagnoses. Another 58 persons (32 men and 26 women) refused interview or thyroid ultrasound examination or were uncertain regarding the history of previous thyroid disorders. For 401 further individuals (203 men and 198 women), genotyping was not successful or DNA specimens were lacking. This resulted in a total study population of 3428 subjects (1800 men and 1628 women) who were included into the present analysis.

In all, 1615 (47.1%) subjects carried the CC genotype, 1463 (42.7%) the TC genotype, and 350 (10.2%) the TT genotype. This distribution was consistent with Hardy–Weinberg equilibrium.

Further sociodemographic information on the study population including data on smoking habits, use of oral contraceptives, and hormone replacement therapy are presented in Table 1. GNB3 carrier status did not affect any of these variables. Likewise, urine iodine and thiocyanate excretions were not different between genotype groups. As quantitative functional and morphological thyroid parameters differ between both genders, all further analyses were stratified for men and women.

Table 2a summarizes the available information on thyroid functional parameters. TSH concentrations were classified as normal, high, and low. Categorized and mean TSH levels were not affected by the GNB3 genotype status.

The results of the ultrasound examinations are detailed in Table 2b. There was no association between thyroid volume and GNB3 genotype, neither in men nor in women. In all, 690 men (38.3%) and 542 women (33.3%) fulfilled the ultrasound criteria for the diagnosis of goiter. Again, we did not observe any difference in the GNB3 genotype distribution between subjects with and without goiter. We further analyzed the genotype-dependent occurrence of thyroid nodules. Three hundred men (16.8%) and 366 women (22.6%) were diagnosed of thyroid nodules. Again, there was no association between the GNB3 genotype status and presence or absence of nodules. In particular, we did not observe an enrichment of the CC genotype in subjects with thyroid nodules. Further, multivariable analyses addressing the risk for thyroid nodules or goiter by GNB3 genotype modified by age, gender, body mass index, smoking status, urine iodide, and thiocyanate excretion did not alter the conclusions of the crude analyses presented so far. In women, further inclusion of use of oral contraceptives and menopausal hormone therapy did not substantially affect these results.

Discussion

The well-characterized functional GNB3 rs5443 variants have been linked to a complex phenotype that parallels in part the pleiotropic symptoms of thyroid disorders (12 –16). Two recent reports have associated the GNB3 825C allele with the occurrence of thyroid adenomas and a trend toward an oncocytic histological morphology in thyroid carcinomas (17,18).

In contrast to other ubiquitously expressed Gβ subunits, Gβ3 exhibits a restricted expression pattern that includes thyroid and pituitary gland (GEO profiles; www.ncbi.nlm.nih.gov/projects/geo/; dataset 181, probe 225_at). We have analyzed whether GNB3 825C>T alleles are associated with subclinical and manifest disorders in SHIP. The study population is of German descent, and problems of multiethnic population admixture—potentially affecting genetic association studies—are negligible in this area of low migration. As with other large population-based studies, the cross-sectional design prevented a nonstandardized history taking. Its observational character further prohibited contacts to attending physicians and clinics to prevent confounding effects on future reexaminations. Hence, we documented “thyroid disorders” in the standardized interviews but abstained from further subclassifications based on patients' statements only. In total, we excluded 369 participants with a reported history of overt or potential thyroid disorders. This was necessary because former radio-iodine, surgical, or pharmaceutical treatments would significantly influence the actual status of thyroid volume and nodules, and current use of medication would probably have highly relevant effects on actual thyroid function. Of note, the GNB3 genotype distribution of this group did not differ from the genotype distribution of the whole study population making it unlikely that an exclusion of this group introduced a relevant bias into our study. Although this exclusion of patients may limit our conclusions to some extent, thyroid disorders develop over time. The large sample size (n = 3489) that covers an age range from 20 to 79 is particularly suited to detect early functional and morphological thyroid alterations.

We may have underestimated smaller thyroid nodules by using a 5 MHz ultrasound probe. Since ultrasound observers were blinded for the genetic analyses, we can exclude a differential bias across the genotype groups. We cannot unequivocally rule out, however, that an association between the GNB3 variant and small thyroid nodules may have escaped detection.

Our main finding is that common GNB3 variants do not affect functional or morphological thyroid parameters. Under population-based conditions there was no observable effect of this genetic variant on any key thyroid parameter investigated. This notion does not exclude a potential influence of GNB3 alleles on the final course of overt thyroid diseases, for example, by modulation of morphological and cytological parameters. In such a concept, GNB3-associated morphological alterations of thyroid nodules could precipitate in preferential surgical treatment, contributing to the accumulation in pathological series (17,18). Together, our findings argue against a major contribution of GNB3 variants to the pathogenesis of common thyroid disorders in the general population.