| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
III. Medical Department (K.K., D.F., Y.B., M.E., V.B., S.N., R.P.), and Interdisciplinary Center for Clinical Research Leipzig (K.K), University of Leipzig, D-04103 Leipzig, Germany
Correspondence: Address all correspondence and requests for reprints to: Professor Ralf Paschke, Universität Leipzig, Zentrum für Innere Medizin, Medizinische Klinik und Poliklinik III, Ph.-Rosenthal-Strasse 27, 04103 Leipzig, Germany. E-mail: pasr{at}medizin.uni-leipzig.de
 |
Abstract |
|---|
In reconstructing the line of events from early thyroid hyperplasia to MNG we will argue the predominant neoplastic character of nodular structures, the nature of known somatic mutations, and the importance of mutagenesis. Furthermore, we outline direct and indirect consequences of these somatic mutations for thyroid pathophysiology and summarize information concerning a possible genetic background of euthyroid goiter.
Finally, we discuss uncertainties and open questions in differential diagnosis and therapy of euthyroid and toxic MNG.
|
I. Definition and Epidemiology |
|---|
protein mutations in macroscopic toxic thyroid nodules both in solitary nodules and multinodular disease; the phenotype of patients with activating germline TSHR mutations; and a number of animal models of thyroid autonomy (reviewed in Refs.7, 8, 9, 10). Iodine deficiency is by far the best studied epidemiological risk factor for nodular thyroid disease. The prevalence of nodular thyroid disease (as well as goiter) is inversely correlated with the population’s iodine intake (11, 12). This has formerly been assessed clinically by palpation, nowadays considered highly inaccurate (13, 14, 15), but is also clearly documented by thyroid ultrasonography. Based on ultrasound investigation, a frequency of thyroid nodular disease as high as 30–40% (in women) and 20–30% (in men) of the adult population has been reported in iodine-deficient areas. Furthermore, even minor differences in the ambient iodine supply may be reflected in the different prevalence of thyroid abnormalities; Knudsen et al. (16) found a difference in goiter prevalence (15% in mild and 22.6% in moderate deficiency) and nodule size (increased in the moderate iodine deficiency group). The prevalence of thyroid nodules seems to increase with age (4, 17, 18). In a borderline iodine deficiency area, MNG was present in 23% of the studied population of 2656 Danish people aged 41 to 71 yr and increased with age in women (from 20 to 46%) as well as men (from 7 to 23%) (3). In contrast, the relation between age and thyroid volume is less coherent, whereby in iodine-deficient areas (except for severe deficiency), thyroid enlargement peaks around 40 yr with no tendency for further increase (19). Interestingly, similar observations have been made in an iodine-sufficient area. In the 20-yr follow-up of the Whickham Survey, the frequency of goiter decreased with age (goiter prevalence initially, 23% women and 5% men; at 20-yr follow-up of the same patients, 10% women and 2% men) (20).
Thyroid nodules are found with higher frequency in enlarged thyroid glands, although all clinicians will agree that they may also be present in an otherwise normal thyroid gland (4, 18, 21). The correlation between iodine supply and prevalence of nodular thyroid disease can similarly apply to TMNG. The high frequency of thyroid autonomy, which accounts for up to 60% of cases of thyrotoxicosis in iodine-deficient areas is largely due to TMNG (
50%, solitary toxic nodules 10%) (12, 22). Prevalence of thyroid autonomy correlates with increased thyroid nodularity and increases with age (4, 22). In contrast, thyroid autonomy is rare (3–10% of cases of thyrotoxicosis) in regions with sufficient iodine supply (22, 23). Correction of iodine deficiency in a population results in decrease of thyroid autonomy as demonstrated by the impressive 73% reduction in prevalence of TMNG only 15 yr after the doubling of iodine content of salt in Switzerland (12, 24). Although goiter and euthyroid and toxic nodular thyroid disease share the common and important epidemiology of iodine deficiency, it needs to be stressed that most epidemiological conclusions are derived from cross-sectional studies. Thyroid nodules (and goiter) also occur in individuals without exposure to iodine deficiency, and not all individuals in an iodine-deficient region develop a goiter. Moreover, there is a strong clustering of goiter in families (see Section VIII).
Screening has been performed for other "environmental factors" (19). Smoking has been proposed as a risk factor for goiter (25), and nodules were also found with higher prevalence in goiters of smokers compared with nonsmokers. The impact of smoking on thyroid disease could be due to increased thiocyanate levels in smokers exerting a competitive inhibitory effect on iodide uptake and organification (19, 26). The association is more pronounced in iodine deficiency (26). Radiation is another environmental risk factor not only for thyroid malignancy but also for benign nodular thyroid disease. An increased prevalence of nodular disease has been associated with exposure to radionuclear fallouts and therapeutic external radiation, and this theory is being discussed by some authors to explain occupational exposure to low-level radiation (27, 28, 29, 30, 31). Furthermore, several studies suggest that thyroid volume is also significantly correlated with body weight and body mass index. In agreement with this, a recent study has shown that in obese women, weight loss of more than 10% may result in a significant decrease in thyroid volume (32).
Nodular disease is more frequent (5- to 15-fold) (33, 34) in women, but the reasons for this are poorly understood. Thus, at present, one can only speculate as to a genetic susceptibility for thyroid disease (for details, see Section VIII) and/or a direct impact of steroid hormones. In fact, a growth-promoting effect of estrogen has been described in vitro in rat FRTL-5 cells and thyroid cancer cell lines and has been proposed as a possible contributing, constitutional effect of gender (35, 36). In addition, 17ß-estradiol has been suggested to amplify growth factor-induced signaling in normal thyroid and thyroid tumors (36). Interestingly, the use of oral contraceptives, which antagonize the physiological hormonal cycle, has been reported to be associated with a decrease in goiter (but not nodules, although this may represent an age artifact of the studied population). On the other hand, pregnancy-related thyroid enlargement was clearly related to iodine deficiency (19), and in one German study (37), increased MNG prevalence with parity was only observed in those women who had not taken iodine supplementation during an earlier pregnancy.
In summary, the development of nodular disease is influenced by multiple environmental components interacting with constitutional parameters of gender and age. However, whether these factors actually result in goiter or nodular thyroid disease is a different matter, ultimately decided by the genetic background of the individual patient (see Section VIII).
|
II. Clinical Aspects of MNG and TMNG |
|---|
Alternatively, patients may present with symptoms suggestive of hyperthyroidism, the clinical presentation of which varies considerably with age. In a series of 84 French patients with overt hyperthyroidism, classical signs of thyrotoxicosis, e.g., nervousness, weight loss despite increased appetite, palpitations, tremor, and heat intolerance were more frequently observed in younger patients (
50 yr) (42), whereas atrial fibrillation and anorexia dominated in the older age group (
70 yr). In addition, subclinical hyperthyroidism, defined by low or suppressed TSH with normal free T4 and free T3 levels is more commonly observed in older patients with TMNG (43). In fact, the incidental finding of low or suppressed TSH levels on routine investigation in iodine-deficient regions for other conditions is frequently a first indicator for presence of thyroid autonomy (4). Subclinical hyperthyroidism is more than just a low TSH status, because it is associated with increased prevalence of atrial fibrillation and bone density loss (43). In addition, an increased cardiovascular mortality rate in patients with low serum TSH levels has been described in a 10-yr cohort-study in the United Kingdom (44). The management of euthyroid and toxic multinodular thyroid disease has recently been extensively reviewed by Hegedüs et al. (33).
A. Differential diagnosis
Very rarely, TMNG occurs as an autosomal, dominantly inherited disease caused by activating germline mutations in the TSHR gene (45). A positive family history of recurrent hyperthyroidism and goiter with absence of typical diagnostic features of Graves’ disease, persistent neonatal thyrotoxicosis, and relapsing nonautoimmune thyrotoxicosis in childhood are highly suggestive of the condition. So far, more than 150 patients (10 families and 11 children with sporadic occurrence of TSHR germline mutations) have been reported in the literature (http://www.uni-leipzig.de/innere/TSHR). Thyroid ablation is advocated as the first-line treatment (surgery and/or radioiodine) to prevent relapses. Molecular analysis for germline TSHR mutations offers the possibility for family screening, preclinical diagnosis, and genetic counseling (7, 46).
In iodine-deficient areas, the distinction between thyroid autonomy and Graves’ disease can be complicated by absence of extrathyroidal signs of autoimmune thyroid disease and "atypical" diagnostic findings. In this regard, several possibilities may be encountered. First is the erroneous classification of Graves’ disease as TMNG due to presence of thyroid nodules observed in 10–15% of Graves’ disease patients or a patchy scintiscan appearance compatible with TMNG (47, 48). Second is the failure to detect TSHR antibodies (TRAB) in Graves’ disease using less sensitive assays. This is illustrated by the detection of TRAB with highly sensitive second-generation assays and/or bioassays in up to 56% of patients with scintiscan appearance of TMNG and up to 22% of patients with diffuse uptake and absence of eye disease and negative TRAB results in older assays (erroneously classified as "diffuse thyroid autonomy") (47, 48, 49). Third is confusion of familial occurrence of (autoimmune) hyperthyroidism with hereditary thyroid autonomy, which might clinically masquerade as Graves’ disease. In this scenario, absence of TRAB is highly suggestive of familial nonautoimmune hyperthyroidism due to a constitutively activating TSHR germline mutation (46).
|
III. Natural Course of MNG and TMNG |
|---|
). The conclusion, which is suggested by these data, is that both in an iodine-deficient and -sufficient setting a variable portion, but most likely not all, nodules will grow, and the speed of growth is highly heterogeneous. Thus, identification of nodules with an increased growth potential is a challenge. This may also be relevant to therapeutic management. In fact, one could speculate that discrepant results reported in various treatment studies may actually reflect this heterogeneity of proliferation (and/or the potential to taper it down by treatment) rather than an evidence-based treatment effect of iodine vs. iodine plus levothyroxine or levothyroxine alone. Furthermore, results of currently available studies (3, 52, 53, 55) do not allow conclusions as to whether nodule growth is associated with an increased risk of thyroid malignancy, and thus the question arises as to the benefits of nodule volume reduction. In the authors’ opinion, therapy, if efficient, is possibly better aimed at the primary prevention of the evolution of novel/further thyroid nodules in predisposed patients (56, 57), with the long-term prospective goal to reduce ablative thyroid treatment for cosmetic reasons, compression symptoms, and, importantly, thyroid malignancy. However, the proof of principle for any of these suggestions is still awaited.
|
|
IV. Clonal Origin of Thyroid Nodules |
|---|
Our investigations of the clonal origin of autonomously functioning thyroid nodules (AFTNs) and solitary cold thyroid nodules (CTNs) (both adenomas and adenomatous nodules) used a PCR approach to amplify the X-linked human androgen receptor from genomic DNA of female patients (84, 85, 86). After thorough screening for somatic mutations in these thyroid nodules (for details, see Section V), we could demonstrate that thyroid nodules with a somatic mutation are predominantly of clonal origin (84, 85, 86). This is not surprising, because it is in full agreement with the widely accepted paradigm in tumor biology that neoplasia originates from a single mutated cell (87). Moreover, we were especially interested in data that would elucidate the etiology of mutation-negative nodules. Interestingly, more than 50% of mutation-negative cases from female patients show a monoclonal origin when tested for X chromosome inactivation (84, 85, 86). This could indicate a neoplastic process with a mutation in a gene other than the TSHR, the Gs protein, or the ras family of oncogenes. Moreover, our finding of an overall frequency for the monoclonal origin of thyroid nodules at about 60–70% agrees with a number of other studies (67, 68, 70, 71) and further underscores that thyroid nodules predominantly result from a neoplastic process with somatic mutations as the starting point (8).
|
V. Hot Thyroid Nodules |
|---|
protein, at higher TSH concentrations an activation of the phospholipase C cascade by Gq has also been shown (88, 89). Moreover, there is evidence that the TSHR may be coupled to other members of the G protein family (88, 90). However, experimental data are frequently focused on the cAMP branch of TSH signaling. Early work by Pisarev et al. (91) demonstrated that cAMP elevation causes goiter. Also, in the thyroid gland and cultured thyroid epithelial cells as well as other endocrine tissues, it is widely accepted that cAMP stimulates proliferation (92, 93, 94, 95). More recently, transgenic models were studied to further understand TSHR signaling in more detail: chronic in vivo stimulation of the cAMP cascade stimulates epithelial cell proliferation in vivo (82, 96, 97). A dominant, negative cAMP response element binding protein blocks signaling downstream of cAMP and causes severe growth retardation and primary hypothyroidism (98). TSH/TSHR signaling generally controls iodine metabolism but only affects growth in the adult thyroid gland and not during embryonic development (99, 100).
Somatic point mutations that constitutively activate the TSHR were first identified by Parma et al. (101) in hyperfunctioning thyroid adenomas. However, in different studies, the prevalence of TSHR and Gs mutations in AFTNs has been reported to vary from 8 to 82% and 8 to 75%, respectively (101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112). These studies differ in the extent of mutation detection and the screening methods. Comparisons with respect to the obvious differences between the studies have been performed elsewhere (8, 113, 114). A comprehensive study of our group using the more sensitive denaturing gradient gel electrophoresis (115, 116, 117) revealed a frequency of 57% TSHR mutations and 3% Gs mutations in 75 consecutive AFTNs (86). These results raise the question of the molecular etiology of TSHR and Gs mutation-negative nodules. A possible answer is given by clonal analysis of these AFTNs that demonstrate a predominant clonal origin of thyroid nodules and imply a neoplastic process driven by genetic alteration (for details, see Section IV). In a recent study (118), we found that AFTNs without a TSHR mutation show an increased expression of the tumor suppressor protein p53-binding protein 2, which interacts with p53 and specifically enhances p53-induced apoptosis but not cell cycle arrest (119). From this finding, one could speculate that increased expression of this gene could increase apoptosis in AFTNs without a TSHR mutation and thus have a negative effect on the growth of the tumor. However, data on apoptosis in AFTNs do not allow a comparison with respect to the TSHR mutation status (120, 121). Furthermore, the AFTNs without a TSHR mutation differ from the nodules harboring a TSHR mutation in their increased expression of two genes that are involved in the signal transduction of G protein-coupled receptors RGS 6 and G protein-coupled receptor kinase 2. Members of the RGS family have been shown to modulate the function of G proteins by activating the intrinsic GTPase activity of the -subunits (122), whereas G protein-coupled receptor kinases play a role in the receptor desensitization (123). In general, a higher expression of these genes would rather restrict cAMP accumulation in AFTNs and could have a negative effect on functional autonomy. However, further experiments have to explore the importance of these genes in the etiology of AFTNs without TSHR mutations.
AFTNs with TSHR mutations lack a clear genotype/phenotype correlation (113). A similar finding is evident for germline TSHR mutations (124). Variable phenotypes associated with the same TSHR mutation could be the result of influences on signaling downstream of the TSHR. This modulation could have a number of targets, such as G protein coupling, receptor desensitization, and internalization or cross talk with other signaling cascades. Although our knowledge concerning these targets is far from complete, recent findings are very promising. First, the TSHR itself could be the subject of regulatory mechanisms that contribute to the etiology of AFTNs and the clinical phenotype. Voigt et al. (125) showed that ß-arrestins interact with the TSHR and are able to desensitize the receptor. Increased expression of ß-arrestin 2 in AFTNs could cause desensitization of the TSHR and thereby down-regulate constitutive activation. A similar result could be caused by increased TSHR internalization due to increased expression of G protein-coupled receptor kinases in AFTNs (118, 126). In addition to the direct effects on the TSHR protein, altered interaction with G proteins would be the next downstream level where modulation interferes with constitutive activation. For example, the mutated TSHR could show a shift of the coupling specificity for G proteins (127). Such a shift could allow a more efficient activation of other downstream cascades (e.g., JAK/STAT pathway) through protein kinase C in addition to cAMP and inositol phosphate (128, 129). Gene expression analysis in AFTNs with TSHR mutations in comparison with AFTNs without a TSHR mutation would support this hypothesis, demonstrating an increased expression of JAK 1, protein kinase Cß1, and 
mRNA (118). Evidence for other cascades (e.g., RAS/RAF/MEK/ERK/MAP pathway) to play a role in constitutive TSHR signaling in AFTNs is currently missing.
In addition to the intracellular signaling network connected to the TSHR, the extracellular action of different growth factors enhances the complexity of the signal flux into the thyroid cell. Growth factors such as IGF-I, epidermal growth factor, TGF-ß, and fibroblast growth factor stimulate growth and dedifferentiation of thyroid epithelial cells (130, 131). Studies that have been focused on insulin and IGF show a permissive effect of insulin and IGF-I on TSH signaling (132, 133, 134, 135, 136) and a cooperative interaction of TSH and insulin/IGF-I (137). Signal modulation of the TSHR that would define the etiology of AFTNs and the clinical phenotype could therefore take part at a number of stages and very likely involves genetic/epigenetic, sex-related, and or environmental factors.
B. Secondary/indirect effects of activating TSHR mutations
Because constitutively activating TSHR mutations disturb the coordinated signal transduction network of the thyroid drastically, subsequent changes in the signal transduction network can be expected. These alterations based on the constitutive activation of the TSHR signaling can be described as indirect or secondary effects of the activating TSHR mutations. The use of the microarray technique offers the advantage of a highly parallel analysis of gene expression to analyze changes between AFTNs and CTNs compared with surrounding tissue. This approach also allows the evaluation of which genes and groups of genes are most frequently affected in the molecular etiology of thyroid nodules, so that we may possibly deduce a molecular defect from the expression pattern. In a recent study using the Affymetrix GeneChip technology (Affymetrix UK Ltd., High Wycombe, United Kingdom), we could show a distinctly changed pattern of gene expression of the TGF-ß signaling pathway between AFTNs with and without TSHR mutations and their normal surrounding tissue (Fig. 1) (118). The type III TGF-ß receptor, mothers against decapentaplegic homolog (Smad) 1, 3, and 4, as well as p300, a transcriptional coactivator, showed a decreased expression in AFTNs, whereas the inhibitory Smad 6 and 7 showed an increased expression in AFTNs. These findings suggest inactivation of TGF-ß signaling in AFTNs due to constitutively activated TSHR (e.g., resulting from TSHR mutations). This assumption is supported by the findings of Gärtner et al. (138), who showed a decreased expression of TGF-ß1 mRNA after TSH stimulation of thyrocytes. Because TGF-ß1 has been shown to inhibit iodine uptake, iodine organification, and thyroglobulin (TG) expression (139, 140), as well as cell proliferation in different cell culture systems (141, 142, 143, 144), these and our novel findings suggest that inactivation of TGF-ß signaling is a major prerequisite for increased proliferation in AFTNs (120, 145).
|
|
VI. CTNs |
|---|
Hypotheses in studies that aim to understand the molecular or genetic causes of human cancer in general (150, 151) or AFTN (8) and thyroid carcinomas (152) in detail have often also been applied to studies of CTNs [e.g., ras mutations (153, 154); for review, see Wynford-Thomas (155)]. In contrast to thyroid carcinomas, where a number of genes have been implicated in the pathogenesis of these lesions (152, 156), and in contrast to AFTN where constitutively activating TSHR mutations are very prevalent genetic events (7, 8), knowledge concerning the molecular etiology of CTNs is limited.
A. Iodide transport and metabolism
With reference to their functional status (i.e., reduced iodine uptake), failure in the iodide transport system or failure of the organic binding of iodide has been detected as a functional aberration of CTNs long before the molecular components of iodine metabolism were known (for review, see Ref.157). Later, a decreased expression of the Na+/I– symporter (NIS) in thyroid carcinoma and benign CTNs suggested the molecular mechanism for the failure of the iodide transport (reviewed in Refs.158, 159, 160). Although the extent of decrease in NIS mRNA expression of CTNs varies in different studies (160, 161, 162, 163), reduction is in many cases very likely the result of hypermethylation in the NIS promoter (160). Moreover, in vitro studies suggest that reduced NIS mRNA expression could be caused by constitutive activation of RET or RAS genes (164, 165, 166). However, reduced NIS mRNA expression does not necessarily lead to reduced NIS protein expression (Fig. 2) (160). In contrast to other thyroid disorders with congenital iodide transport defects (for review, see Refs.167 and 168), no NIS gene mutation that would render this protein nonfunctional was detectable in CTNs (160). Therefore, the recently identified defective cell membrane targeting of the NIS protein is a more likely molecular mechanism that could account for the failure of the iodine uptake in CTNs (158, 160, 169). However, the ultimate cause of this defect is currently unknown.
|
B. Signaling proteins
In addition to a failure in metabolic proteins that might explain the development of CTNs on the molecular level, pathological changes of signaling molecules might reprogram the growth stimulus and lead to clonal expansion of thyroid epithelial cells. Although not a particular subject of this review, much can be learned from findings in thyroid carcinoma. In this regard, genetic changes (i.e., point mutations) that cause constitutive activation of the RAS/RAF/MEK/ERK/MAP pathway have been suggested as a key mechanism during tumor initiation or progression in thyroid follicular cells (for review, see Ref.155). So far, the only known molecular event that evidently causes such activation in thyroid carcinomas and CTNs is a mutation in one of the small RAS oncogenes (153). Recently, BRAF mutations, first detected in melanomas and with lower frequency in other cancers (176), have been detected in thyroid papillary carcinomas (177). They can also activate this pathway and might therefore also cause benign follicular lesions. Strikingly, both in colorectal and thyroid cancers, v-raf murine sarcoma viral oncogene homolog B1 (BRAF) mutations occur only in tumors that do not carry mutations in a RAS gene. In a recent study of 40 cold thyroid adenoma and adenomatous nodules, we detected ras mutations in only a single case (85). Moreover, in the same set of CTNs we did not detect point mutations in the mutational hot spots of the BRAF gene (178). This is in line with the lack of BRAF mutations in benign follicular adenoma in other studies (177, 179, 180). So far, only one study (181) detected a single BRAF mutation in a set of 51 follicular adenomas. Instead of RAS and BRAF mutations, there could be other molecular events that could constitutively activate the RAS/RAF/MEK/ERK/MAP pathway. Such candidate molecules include other members of the RAF gene family, like RAF-1, or downstream genes like ERK or MAPKs. However, mutations in these genes have not been reported in benign thyroid lesion thus far. Furthermore, molecular events that lead to activation of other cascades that exert synergy with MAPK signaling (e.g., cAMP signaling) or inactivation of independent cascades that restrict proliferation (e.g., TGF-ß signaling) could explain CTNs. In addition to the TSHR, several G protein isoforms like Gi2 (182) or Gq, G11 (183), and some candidate genes mediating downstream cAMP signaling like Epac and Rap1 (184) have been screened in CTNs for mutations. However, only a single somatic mutation in the Gi2 gene (182) was found in follicular adenomas.
C. Results of gene expression studies by arrays
Currently, expression profiling of signaling proteins using microarray methodology is a promising approach that may contribute to further understanding the molecular events that lead to the development of AFTNs (147, 185) or thyroid carcinoma (186, 187, 188). Functional characteristics of CTNs suggest that mechanisms initiating growth but not leading to hyperfunction need to be defined. As far as future screenings for genetic defects are concerned, expression profiling could describe the molecular mechanism and rule out a number of possible targets (e.g., because they are not expressed) or unmask alternative candidates. Moreover, as demonstrated for AFTNs, results of expression profiling might shift attention to other signaling cascades (for details on AFTNs, see Section V). For differentially expressed genes within these cascades, a sequencing approach might then be warranted. In addition, knowledge of the molecular signature of CTNs and benign thyroid tumors in general could be very helpful to define differences between benign and malignant thyroid disease with diagnostic or therapeutic relevance.
Recently, our group investigated 588 genes by cDNA expression arrays in three AFTNs, three CTNs, and corresponding normal surrounding tissue. In general, changes in the expression of several signal-transducing components were detected. Although this seems to reflect a disturbed signaling system, the results of that limited study did not allow us to identify specific signal transduction cascades that might be involved in nodular development (147). To gain a higher resolution, we compared gene expression for approximately 10,000 full-length genes between CTNs and their corresponding normal surrounding tissue (307). Regulation of gene expression in CTNs was most consistent for a group of several histone mRNAs. Increased expression of these histone mRNAs and of cell cycle-associated genes like cyclin D1, cyclin H/cyclin dependent kinase 7, and cyclin B most likely reflect a molecular setup for an increased proliferation in CTNs (189). In line with the low prevalence of ras mutations in CTNs (85), we find a reduced expression of ras-MAPK cascade-associated genes, which might suggest a minor importance of this signaling cascade.
D. Chromosomal aberrations
Loss of heterozygosity, microsatellite instability, and, more recently, gene rearrangements and chromosomal translocations as different forms of chromosomal aberration are considered important steps in carcinogenesis and have been investigated as potential markers to discern benign from malignant nodular disease. Findings of chromosomal aberrations and microsatellite instability in benign thyroid tumors, although sometimes sporadic, suggest that there is a difference in the extent of these DNA changes (190, 191). Alternatively, these results could stem from errors in histological characterization (192).
Gene rearrangements unique to thyroid adenomas have recently been the focus of study (reviewed in Ref.193). These studies led to the identification of the thyroid adenoma associated gene that encodes a death receptor-interacting protein (194).
Although also reported for thyroid follicular carcinoma (195), our finding of loss of heterozygosity at the TPO locus is characteristic for some CTNs (15%) but rather points to defects in a gene near TPO on the short arm of chromosome 2 (175). Moreover, after identification in a significant portion of follicular carcinomas (196), Pax-8/peroxisome proliferator-activated receptor 
gene rearrangement have also been reported for CTNs (197) but seem to be a rare finding (198, 199, 200) or attributable to histological misclassification of the thyroid nodules (192). Although the frequency of each of these DNA aberrations is rather low, these chromosomal changes need to be considered together in the further elucidation of the molecular etiology of CTNs.
|
VII. Multinodular Goiter |
|---|
A. Mutagenesis as the cause of nodular transformation
From animal models of hyperplasia caused by iodine depletion (79, 203, 204) we learn that, in addition to an increase in functional activity, a tremendous increase in thyroid cell number occurs. These two events very likely orchestrate a burst of mutation events. Although the enzymatic setup awaits further characterization (171), it is known that thyroid hormone synthesis goes along with increased H2O2 production and free radical formation (205), which may damage genomic DNA and cause mutations (206). As a consequence, the spontaneous mutation rate in the thyroid is almost 10 times higher than in other organs (e.g., compared with liver) (K. Krohn, unpublished observations). Together with a higher spontaneous mutation rate, a higher replication rate will more often prevent mutation repair and increase the mutagenic load of the thyroid, thereby also randomly affecting genes crucial for thyrocyte physiology. Mutations that confer a growth advantage (e.g., TSHR or Gs protein mutations) very likely initiate focal growth. Hence, AFTNs are likely to develop from small cell clones that contain advantageous mutations as shown for the TSHR in hot microscopic regions of euthyroid goiters (9).
B. Etiology
Epidemiological studies, animal models, and molecular/genetic data outline a general theory of nodular transformation. Based on the identification of somatic mutations and the predominant clonal origin of AFTNs, we propose the following sequence of events that could lead to thyroid nodular transformation in three steps (Fig. 3). In the first step, iodine deficiency, nutritional goitrogens, or autoimmunity cause diffuse thyroid hyperplasia. Then, at this stage of thyroid hyperplasia-increased proliferation together with a possible DNA damage due to H2O2 action causes a higher mutational load with a higher number of cells bearing a mutation. Some of these spontaneous mutations confer constitutive activation of the cAMP cascade (e.g., TSH-R and Gs mutations) that stimulate growth and function. Finally, in a proliferating thyroid growth factor expression (IGF-I, TGF-ß1, or epidermal growth factor) is increased. As a result of growth factor costimulation, all cells divide and form small clones. After increased growth factor expression ceases, small clones with activating mutations will further proliferate if they can achieve self-stimulation. They could thus form small foci, which will develop into thyroid nodules. This mechanism could explain AFTNs by advantageous mutations that both initiate growth and function of the affected thyroid cells as well as CTNs by mutations that stimulate proliferation only (e.g., ras mutations or other mutations in the RAS/RAF/MEK/ERK/MAP cascade). Moreover, nodular transformation of thyroid tissue due to TSH-secreting pituitary adenomas (207), nodular transformation of thyroid tissue in Graves’ disease (208), and in goiters of patients with acromegaly (209) could follow a similar mechanism, because thyroid pathology in these patients is characterized by early thyroid hyperplasia.
|
|
VIII. Pathogenesis and Genetic Etiology of Euthyroid Goiter |
|---|
A. Family and twin studies
Although lack of iodine is the most prevalent factor for simple goiter as well as endemic goiter, other causes are likely. Familial clustering of goiters and the female predominance of goiters are the two major arguments suggesting a genetic background for euthyroid goiters. Family and twin pair studies in endemic and nonendemic areas clearly demonstrated a genetic predisposition for goiter development. Within a Greek region, endemic goiter affects some families more than others (211). This could provide evidence for a genetic etiology, although environmental factors that differ between families must also be considered (212). The familial aggregation of goiters in Greece was confirmed in a subsequent study (213). The progeny of affected persons were more often affected by goiter than were the descendants of unaffected subjects. Likewise, other euthyroid goiter family studies in Greece, Slovakia, and Africa also lead to the conclusion that a genetic predisposition is present in the affected individuals (211, 214, 215). In addition, more rapidly growing goiters in a subgroup of schoolchildren (15–20%), despite iodine supplementation (214) and differences in thyroid volume of adolescent siblings with sufficient iodine intake in Slovakia (216), also support a genetic influence on thyroid growth. Moreover, studies in Greek populations (217) have shown the persistence of endemic goiters in certain regions despite iodine supplementation. Familial occurrence of euthyroid goiter in an iodine-replete region in Sweden was reported for 41% of the patients with goiter, with an even higher frequency of familial occurrence in those individuals with prepubertal development of the goiter (218). Although family studies are a reliable method to determine goiters in many family members over several generations, it is impossible to definitively conclude whether the members shared the same susceptible genetic makeup or the same environment. Therefore, twin studies are more informative in demonstrating a genetic component in the etiology of goiter. Several investigations have provided evidence that there is a predisposing genetic background for goiter in twins. In endemic as well as nonendemic areas, female monozygotic twins [80% (213, 219) and 42% (220), respectively] have a higher concordance rate for goiter than female dizygotic twins [40–50% (213, 219) and 13% (220)]. Twins of the same sex are supposed to share the same family environment. Therefore, the increased concordance was attributed to greater genetic similarity characterizing the monozygotic twins. Contribution of genetic susceptibility to the development of goiter was calculated to be 39% in endemic regions (219). Moreover, a study of 5479 monozygotic and dizygotic twins (220) performed by path analyses (structural equation modeling) suggests that the genetic predisposition to develop goiter is 82%, with 18% according to individual environmental factors in a nonendemic area. However, the previously reported twin studies show the importance of both hereditary and environmental factors (33).
B. Candidate loci
Because of their important role in thyroid physiology and hormone synthesis, TG, TPO, NIS, the pendrin gene (PDS), and TSHR are major candidate genes for familial euthyroid goiters.
Studies of hypothyroid goiters have identified several genetic defects in TG and TPO (172). Congenital goiter and hypothyroidism caused by qualitative and quantitative defects of the TG gene were described by Medeiros-Neto et al. (221). Other studies have also shown a link between the TG gene and congenital goiter and hypothyroidism (222, 223, 224, 225, 226, 227, 228, 229). Furthermore, an inherited abnormality in TG synthesis leading to a lower content of TG in the thyroid gland was reported by Yoshida et al. (230), and a single amino acid substitution in the TG protein (Leu2366Pro) causes endoplasmic reticulum storage disease as determined in the cog/cog mouse (231). Although TG was postulated to be a major candidate gene for euthyroid simple goiter, only one genetic variation associated with euthyroid goiter has been identified in the TG gene up to date. Corral et al. (232) found a G
T substitution at position 2610 of the TG cDNA. This resulted in replacement of histidine for glutamine at codon 870. This sequence alteration was located within exon 10 of the TG gene and was present in 25 of 26 members of three families affected by euthyroid goiter. However, Perez-Centeno et al. (233) found the same point mutation in TG exon 10 gene in only one of 36 patients with endemic euthyroid goiter. Hishinuma et al. (226, 229) found two novel cysteine substitutions in TG, which caused defects in the intracellular transport of TG in patients with a variant type of adenomatous euthyroid goiter. Gonzalez-Sarmiento et al. (234) identified a large heterozygous deletion within the TG gene in a study of 50 cases affected with nonendemic goiter. The deletion involved the promotor region and the exons 1 to 11 of the TG gene and was associated with euthyroid goiter.
Mutations responsible for dyshormonogenesis have also been described in the TPO gene. TPO catalyzes the oxidation of iodide to an iodination species that forms iodotyrosines and iodothyronines. Defects of TPO synthesis caused by a heterogenous spectrum of TPO mutations (235, 236, 237, 238, 239, 240) have been reported to result in reduced TPO activity in combination with total iodide organification defect. Hagen et al. (241) described an intelligent, euthyroid child with goiter. Together with her affected sister she showed no iodide peroxidation or thyroxine iodination activity. Likewise, a euthyroid woman with a recurrent goiter and partial iodide discharge was described by Pommier et al. (242). She had normal iodide peroxidation but deficient TG iodination. However, these are the only two examples for TPO mutation resulting in euthyroid goiters. Most of the previously reported homozygous or compound heterozygous mutations in the TPO gene lead to goiter and hypothyroidism (236, 243, 244).
Since the cloning and molecular characterization of the human NIS gene (245), several defects in this gene have been detected in patients with different phenotypes of thyroid diseases (246). A heterozygous T354P mutation results in congenital hypothyroidism and goiter (247, 248, 249, 250, 251). However, two studies (252, 253) reported that the homozygous T354P mutation is associated with euthyroidism and goiter. Moreover, other allelic variants producing deletion, missense, or truncation of the NIS protein have been described (254, 255).
Mutations in the PDS gene cause Pendred syndrome, which is characterized by congenital sensorineural hearing loss combined with goiter. In Pendred syndrome the thyroid enlargement typically begins in childhood and can vary between and within families (256, 257). Reported goiter sizes vary from small nodules to large MNG (258). Almost all affected individuals are clinically and biochemically euthyroid. Positive perchlorate tests suggest that the PDS defects impair the organification of iodide. Different PDS mutations, each segregating with the disease in the families in which they occurred, have been identified (259–263). Most of them are loss-of-function mutations that directly cause thyroid disease in Pendred syndrome. Therefore, the PDS gene is a candidate gene for development of euthyroid goiter.
Furthermore, the TSHR on chromosome 14q31 could be a candidate gene for euthyroid goiter according to its central role for thyroid function and growth. TSHR germline mutations have been found in rare cases of euthyroid familial goiter (7) (see also Section I.A). Moreover, a germline genetic variation in codon 727 of the TSHR gene (AspGlu) (264) has been associated with TMNG in iodine-deficient areas (265). However, the wild-type TSHR response to TSH was very low in this study, and this in vitro finding was not confirmed in another study (266, 267). Moreover, the putative predisposition for TMNG was based on functional analysis of the TSHR codon 727 variation, which revealed a higher cAMP response compared with the wild-type receptor. However, this in vitro finding was not confirmed in a recent study (266). Recently, Peeters et al. (268) reported that the homozygous C variant (coding for aspartic acid) of the D727E polymorphism was associated with a lower serum TSH level in 156 healthy blood donors. Although this finding supports a possible functional relevance of this polymorphism, the role of a lower TSH level in the development of MNG is unknown.
C. Linkage analysis
Over the past years, linkage analyses became a reliable method to identify novel susceptibility loci for both Mendelian and complex diseases in large families or affected sibling pairs by using genetic markers for microsatellite DNA or repetitive sequences in the entire human genome. Several different susceptible areas have been discovered in large families with non-Mendelian transmission of euthyroid familial goiter. The multinodular goiter 1 (MNG-1) locus on chromosome 14q31 was first reported by Bignell et al. (269) as the result of a genomewide linkage analysis of a large Canadian family with 18 patients affected with MNG. A maximum two point LOD score of 3.8 at D14S1030 and a multipoint LOD score of 4.88, defined by D14S1062 and D14S267, was calculated. In another study, a family with recurrent euthyroid goiters was investigated for linkage to the same candidate region (270). According to a dominant pattern of inheritance with full penetrance, indication for linkage was obtained by a maximal two-point LOD score of 1.5 at marker D14S1030 at the MNG-1 locus. Moreover, a maximum multipoint LOD score of 1.49 was obtained for the region between the TSHR and the MNG-1 candidate loci. The haplotype cosegregation of microsatellite markers confirmed the entire chromosomal segment between both loci on chromosome 14q31 as a positional candidate region for nontoxic goiter. Although the TSHR was a first-line candidate gene, sequence analysis of the TSHR alone revealed several previously reported polymorphisms. In the study by Bignell et al. (269), the TSHR was previously clearly excluded as a candidate gene. Another study reported the analysis of a three-generation, Italian pedigree, including 10 affected females and two affected males (271). An X-linked dominant pattern of inheritance was observed. The investigation of 18 markers spaced at 10-cM intervals on the X chromosome revealed evidence for linkage at marker DXS1226 with a significant LOD score of 4.73. These findings led to the conclusion that defects in the Xp22 region caused thyroid disease in this family. Moreover, the haplotype inspection reduced the critical interval to 9.6 cM between the markers DXS1052 and DXS8039.
In view of the different susceptibility loci for euthyroid goiter, a heterogenous mode of inheritance for euthyroid goiter is very likely. Linkage analysis for the thyroid candidate genes has been performed in four German and one Slovakian family (272). The candidate genes were also analyzed, assuming different recombination fractions for the microsatellite markers in two-point and multipoint analysis. Linkage a
Top
Abstract
I. Definition and Epidemiology
