Third Medical Department (K.K., R.P.) and Interdisciplinary Center for Clinical Research (K.K.), University of Leipzig, D-04103 Leipzig, Germany
Address all correspondence and requests for reprints to: R. Paschke, M.D., Third Medical Department, University of Leipzig, Ph. Rosenthal Strasse 27, D-04103 Leipzig, Germany. E-mail: pasr{at}server3.medizin.uni-leipzig.de
The etiology of thyroid autonomy and autonomously functioning thyroid nodules has been reviewed, with emphasis on histological (1, 2), molecular (3), and clinical aspects (4) as well as transgenic animal experiments that model thyroid autonomy (5). Somatic TSH receptor (TSHR) mutations have been discovered as the most prevalent molecular event in the etiology of autonomously functioning thyroid nodules (6). However, the mechanisms that induce TSHR mutations or determine their variable clinical phenotype are currently unknown. The purpose of this review is therefore to integrate new findings concerning the early molecular etiology of thyroid autonomy with previous molecular, histological, and in vitro results and to examine their possible clinical and therapeutic consequences.
Diagnosis of thyroid autonomy
Autonomously functioning thyroid tissue is characterized by the ability to function without TSH. For clinical diagnosis, radioiodine or 99mTc-pertechnetate imaging of the thyroid gland is performed to determine increased radionuclide uptake of autonomous thyroid epithelia compared with surrounding parenchyma (7, 8). In patients without complete scintigraphic suppression of the surrounding parenchyma, thyroid hormones are administered before suppression scintigraphy to differentiate autonomous from normal thyroid tissue. Histological examinations of tissue corresponding to scintigraphically autonomous areas show different architectural patterns and histological features of follicular cells compared with normal quiescent tissue (1, 9, 10, 11). Thyroid autonomy is most frequently found in toxic multinodular goiters, a heterogeneous clinical disorder that includes the presence of autonomously functioning nodules in a goiter with or without additional nodules. Additional nodules can show normal or decreased (cold nodules) uptake on scintiscan. The variable scintigraphic pattern includes circumscript areas of increased uptake with suppression of the surrounding thyroid tissue or an uneven (patchy) uptake with areas of increased uptake but no suppression of the surrounding parenchyma. Epidemiological studies (12) suggest that toxic multinodular goiters mostly originate from euthyroid goiters in iodine-deficient regions (for details, see below). Three distinct categories of thyroid autonomy have been described: 1) benign thyroid tumors with a follicular differentiation presenting as adenoma (encapsulated tumor) or adenomatous nodules (not encapsulated but circumscribed), 2) malignant thyroid epithelial tumors causing hyperthyroidism, and 3) microscopic areas in euthyroid goiters that show strong 125I labeling under TSH suppression (13). This latter form of autonomy might also be the histological equivalent of disseminated autonomy.
Thyroid epithelial cell function
In normal thyroid tissue TSH stimulates all aspects of thyroid physiology, i.e. iodine uptake and metabolism, thyroid hormone synthesis, and release (14). In the human thyroid TSH signaling is mediated through cAMP and phospholipase C (15). To date, the cAMP aspect of TSH signaling is much better studied and understood. Early work by Pisarev et al. demonstrates that cAMP produces goiter (16). Moreover, in contrast to a previous report (17), cAMP stimulates the proliferation of many endocrine tissues, including thyroid epithelial cells in vitro (18, 19, 20). Studies with goitrogen-treated rats exhibiting elevated TSH levels also suggested a link between proliferation and TSH signaling in vivo (21). More recently, the introduction of transgenic models has added the aspect of chronic in vivo stimulation of the cAMP cascade to study thyroid epithelial cell proliferation in vivo (22, 23, 24). Targeted expression of a dominant negative cAMP response element-binding protein to the thyroid glands of transgenic mice provide the first evidence that the cAMP effect on thyroid epithelial cell proliferation in vivo is a direct one (25).
Studies of interactions between TSH and growth factors have been
focused especially on insulin and insulin-like growth factors (IGFs)
(26). Numerous in vitro and in vivo
experiments in different species show a permissive effect of
insulin/IGF-I on TSH signaling (27, 28, 29, 30). Moreover, a
cooperative interaction of TSH and insulin/IGF-I is very likely
required for the signaling of both TSH and insulin/IGF-I in thyroid
tissue (31). IGF-I is an autocrine factor in the thyroid
that is regulated by TSH in the level of expression, receptor
signaling, and intracellular signal transduction (32, 33, 34, 35).
Therefore, local autocrine regulation through IGF-I signaling most
likely acts in combination with endocrine regulation through systemic
TSH/thyroid hormones to control thyroid epithelial cell growth and
function. There are several reasons for maintaining a local regulatory
system (IGF-I signaling) in addition to the endocrine control of
thyroid epithelial cells. 1) A second system would allow the regulation
of other cells in the thyroid that do not respond to TSH signaling
(e.g. fibroblasts), but need to be regulated with respect to
thyroid epithelial cell growth and function in concert with thyroid
epithelial cells. 2) It would allow the integration of multiple
mitogenic signals. 3) A local system would also allow accumulation of
the endocrine signal so that a cooperative response by a number of
epithelial cells is necessary to induce growth stimulation. This
accumulation of local growth factors [e.g. IGF-I or other
growth factors produced in thyroid epithelial cells, such as
transforming growth factor- (36)] in addition to an
endocrine signal has been suggested by Dawson and Wynford-Thomas
(37) to be the basis for a mechanism that protects against
tumor development in the thyroid (for details see below).
Constitutive activation of the cAMP cascade
In 1989, Dumont et al. (38) hypothesized
that relative overactivity of the cAMP system could account for some of
the benign thyroid tumors and that molecular defects resulting in
signaling through the cAMP cascade might be responsible for autonomous
function of thyroid epithelial cells. The first clinical evidence for
this reasoning came from the identification of somatic mutations in the
Gs protein that conferred constitutive
activation to the cAMP cascade resulting in GH-producing pituitary
adenomas (39). However, these mutations only provided a
molecular explanation for a limited number of benign thyroid tumors
(40). Therefore, signaling proteins other than
Gs
had to be considered. Screening for the
molecular etiology of thyroid tumors was stimulated by in
vitro mutagenesis studies. Constitutive activation (independent of
the ligand) of the ß-adrenergic receptor after substitution of a
single amino acid established a new model for G protein-coupled
receptor activation (41). In vivo, a
constitutive activation of the TSHR as the result of somatic point
mutations was subsequently identified in hyperfunctioning adenomas of
the thyroid (42). Moreover, the clinical relevance of TSHR
mutations has been demonstrated in the context of germline mutations
causing autosomal dominant nonautoimmune hyperthyroidism
(43). Whereas a hyperfunctioning thyroid adenoma harboring
a constitutively activating somatic mutation of the TSHR develops over
a long period and generally rather late in life, the same mutation
present in the germline expressed in every thyroid cell results in
severe neonatal hyperthyroidism and goiter with frequent relapses
during antithyroid therapy and after partial thyroidectomy (reviewed in
Refs. 6 and 44). The principle of
constitutively activating mutations was expanded to a number of G
protein-coupled receptors as the genetic basis of endocrine disease
(reviewed in Refs. 45 and 46).
Frequency of TSHR or Gs
mutations
Studies investigating the prevalence of TSHR or
Gs mutations in autonomously functioning
thyroid nodules have reported variable results, from 882% for
activating TSHR mutations (47, 48, 49, 50, 51, 52, 53). Similarly, the
gsp mutation frequency ranges from 875% (40, 49, 50, 51, 54). A direct comparison of these studies is difficult
due to differences in the extent of mutation detection. Reports of a
low prevalence of TSHR mutations are mostly based on a partial screen
of the TSHR gene (47, 48, 49, 55). However, a partial screen
of the TSHR in these studies does not completely explain the low
frequency of TSHR mutation detection compared with studies based on a
full screen of the receptor for the following reason; less than 25% of
all known mutations are located in the parts of the TSHR gene not
screened in these studies [see TSH Receptor Database at
http://www.uni-leipzig.de/
innere/TSH (56)].
Therefore, these studies miss statistically only 25% of all TSHR
mutations located in the full coding sequence. Besides different
mutation frequencies in different cohorts, which is discussed below,
technical limitations are likely to have compromised some mutation
detections (57). Different screening methods have been
used, e.g. direct sequencing (47, 48, 50, 51, 52)
or single strand conformation polymorphism (55), which is
less sensitive in detecting point mutations (58).
Moreover, the quality of the examined tissue and, hence, the integrity
of the DNA were highly variable, as DNA extracted from
paraffin-embedded tissue (54, 55) is highly degraded
(59) and therefore more difficult to examine than DNA
extracted from shock frozen tissue (47, 48, 50, 51, 52, 53, 60).
Nevertheless, even studies comparable for their mutation detection
methods and screening at least exon 10 of the TSHR, which encodes the
entire transmembrane region, have reported frequencies ranging from
2082% in cohorts similar with respect to iodine supply
(50, 51, 52, 53, 60). Therefore, the tissue identification and
tissue asservation should be considered. Specimens collected from
autonomously functioning thyroid nodules rarely consist of pure
adenomatous tissue. Due to contamination with blood, connective and
surrounding healthy tissue, as well as degenerate areas within the
nodule, the amount of mutated allele may be decreased to less than
50%, making the detection of a somatic mutation difficult by
fluorescent or radioactive direct sequencing (61). A
comprehensive study of our group using the more sensitive denaturing
gradient gel electrophoresis (61, 62, 63) to screen for TSHR
mutations in 75 consecutive toxic thyroid nodules, including specimens
that have been studied with direct sequencing previously (52, 53) revealed a constitutive TSHR mutation in six cases that were
mutation negative with direct fluorescent sequencing (64).
The frequencies of TSHR mutations and Gs
mutations in this latter study were 57% and 3%, respectively. Most
studies on somatic mutations in autonomously functioning thyroid
nodules (40, 42, 49, 50, 51, 54) have concentrated on thyroid
adenomas, which are characterized by a capsule of connecting tissue
around the thyroid lesion (65, 66). However, a large
number of toxic thyroid nodules do not share the feature of
encapsulation, which does not necessarily imply a different etiology.
Indeed, TSHR and Gs
mutations have also been
detected in adenomatous thyroid nodules (52, 60), which
further underlines that clonality of the lesion is the definite
criterion and that histological analysis alone does not allow
identification of the clonal origin. However, the investigation of
autonomously functioning adenomatous thyroid nodules for somatic
mutations requires special attention to avoid sampling errors. It is
necessary to clearly identify the autonomously functioning nodule
macroscopically by ultrasound and scintigraphy and to histologically
evaluate the tissue used for mutation detection.
Likelihood of other somatic mutations
A frequency of 60% for somatic mutations in autonomously
functioning thyroid nodules detected with denaturing gradient gel
electrophoresis and direct sequencing (64) raises
the question of the molecular etiology of mutation-negative nodules.
Mutation-negative autonomously functioning thyroid nodules could
comprise polyclonal lesions that do not evolve from a single mutated
cell. Heterozygous polymorphisms in X-chromosome-linked markers
(67) can be used to decide the clonal origin of thyroid
tissue from female patients. Using a PCR approach to amplify the
X-linked human androgen receptor, 50% of mutation-negative cases from
female patients show a monoclonal origin when tested for X-chromosome
inactivation (64, 68). However, as discussed for the
detection of somatic mutation (see above), monoclonal origin could be
concealed by contamination with blood and connective and surrounding
healthy tissue. As the phenotype of thyroid autonomy (adenoma and
hyperthyroidism) seems to be mainly regulated by constitutive cAMP
activation (69), it is likely that further candidate genes
for the development of toxic thyroid nodules are located in this
signal cascade. This hypothesis is also supported by the finding of
resistance to TSH without the presence of loss of function mutations in
the TSHR (70). Different components of the complex G
protein-signaling pathway have been cloned in the past years and
include 20 different mammalian G protein -subunits, 4 ß-subunits,
and 6
-subunits (71, 72), as well as 16 phospholipases
C (73), 8 adenylyl cyclases (74), and more
than 20 phosphodiesterases (75, 76, 77). Despite the
heterogeneity in putative signaling proteins, components downstream of
the TSHR have only been defined at the G protein level to date
(78, 79). In addition to activating mutations in
-subunits of G proteins, mutations could occur in adenylyl cyclases
or more downstream effector molecules, such as protein kinase A and its
substrates or in molecules activating other signaling cascades
(e.g. tyrosine kinase cascade through IGF-I), which can act
synergistically with cAMP accumulation. Alternatively, overexpression
of signaling proteins such as the TSHR,
Gs
-subunits, adenylyl cyclases, or downstream
effector molecules could also be a cause of autonomously functioning
thyroid nodules. The possible pathophysiological significance of the
degree of Gs
expression in pathological
conditions has recently been demonstrated in animal models. In mice,
for example, long-term cardiac overexpression of the
Gs
-subunit leads to myocardial hypertrophy
(80), whereas blockade of the Gq
protein signaling pathway resulted in a reduction of myocardial
hypertrophy (81). Regarding the level of adenylyl cyclase
activity, the overexpression of adenylyl cyclase VI has been shown to
result in a proportional increase in ß-adrenergic receptor-dependent
cAMP accumulation in embryonic cardiac rat myocytes (82).
However, thyroid tissue studies quantifying adenylyl cyclase activity
in autonomously functioning thyroid nodules are controversial
(83, 84, 85). In addition to increased cAMP synthesis,
impaired cAMP degradation could be the cause of the higher cAMP
accumulation. However, down-regulation of the cAMP inhibitory pathways,
e.g. by inactivating mutations in the phosphodiesterases,
have not been reported in autonomously functioning thyroid nodules. In
contrast, the activities of different phosphodiesterase isoforms have
been found to be dramatically induced in autonomously functioning
thyroid nodules harboring a constitutively activating mutation
(86). As for cAMP metabolism, an altered turnover of the
TSHR could also play a role in aberrant TSH signaling, leading to
increased cAMP accumulation. Therefore, defects in receptor
desensitization and internalization (G protein receptor kinases,
ß-arrestins) may have to be considered. Changes in the expression and
activity of the receptor kinases have been associated with various
disorders, including the development of cardiomyopathy
(87, 88, 89, 90) and rheumatoid arthritis (90). These
examples underline the importance of components of the desensitization
apparatus for the function of the corresponding G protein-coupled
receptor. The TSHR G protein-coupled receptor kinases 2 and 5 as well
as ß-arrestin-1 have been suggested to play a key role in TSHR
desensitization in rat FRTL5 cells (91, 92). However, in
autonomously functioning thyroid nodules increased ß-arrestin-2
expression is predominant compared with that in surrounding tissue
(93). Moreover, G protein-coupled receptor kinase 3 and 4,
but not G protein-coupled receptor kinase 2, expression is
significantly increased in autonomously functioning thyroid nodules
(94). Further investigations are therefore mandatory to
evaluate the role of signaling proteins downstream of the TSHR as
candidate genes for the development of autonomously functioning thyroid
nodules.
Somatic TSHR mutations in toxic multinodular goiter
Similar to solitary toxic thyroid adenomas and adenomatous nodules somatic mutations in the TSHR have also been detected in hot nodules of toxic multinodular goiters (60, 95, 96, 97). Scintigraphically, these toxic multinodular goiters showed hot nodules, with suppression of the surrounding thyroid tissue. In contrast, Gabriel et al. (98), studying 24 cases of toxic multinodular goiters, did not detect any somatic mutation in the transmembrane and C-terminal regions of the TSHR. Toxic multinodular goiter is a clinical disorder rather than a disease entity. Therefore, different pathophysiologies of the clinical disorder toxic multinodular goiter may, in fact, reflect different and/or overlapping etiologies, possibly without common molecular features (e.g. TSHR mutations). Moreover, wrong classification due to insufficient TSH binding-inhibiting Ig assay sensitivity as another possibility is illustrated by the detection of TSHR antibodies with sensitive in vitro assays in hyperthyroid patients that show a scintiscan compatible with toxic multinodular goiters (99, 100). Therefore, the lack of somatic TSHR mutations (98) might also be explained in the context of different etiologies of toxic multinodular goiters. However, especially in areas with iodine sufficiency, it raises concern about the exclusion of Graves disease, the quality of sampling in terms of tissue identification, and tissue asservation in patients who show scintigraphic patterns similar to those initially presented by Kraiem (99). To identify the molecular events leading to toxic multinodular goiters requires meticulous dissection of both functional and morphological aspects of the tissue under investigation before tissue sampling. Without ultrasound and histological examination, circumscribed areas of a scintiscan in toxic multinodular goiters do not necessarily correspond to well defined nodular lesions. This can severely compromise the outcome of a mutation screening. In this context, the finding of the absence of somatic TSHR mutations in toxic multinodular goiter from iodine-sufficient areas (98) does not necessarily suggest a different genetic cause from those in iodine-deficient regions, but needs to be interpreted with caution. This is also true for a germline genetic variation in codon 727 of the TSHR gene (101), which has been associated with toxic multinodular goiter (98). Similar studies in geographic areas with iodine deficiency did not reveal such an association (102, 103). Moreover, the putative predisposition for toxic multinodular goiter 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 (102).
TSHR mutations in hot microscopic areas of iodine-deficient euthyroid goiters
Thyroid autonomy is very prevalent in geographic regions with
iodine deficiency (47% of all cases with hyperthyroidism) and rarely
occurs in iodine-sufficient regions (6% of all cases with
hyperthyroidism) (12). Moreover, studies that compare the
frequency of different thyroid pathologies before and after iodine
supplementation (104) suggest that the level of dietary
iodine is an important etiological factor in thyroid autonomy.
Baltisberger et al. (104) demonstrated that
iodine deficiency is associated with a significantly increased
frequency of thyroid autonomy, whereas iodine supplementation leads to
a significantly reduced frequency of thyroid autonomy. These studies
tempted us to speculate that iodine deficiency is probably involved in
the generation or propagation of somatic TSHR mutations. If this is
correct, early stages of thyroid autonomy already detectable in
euthyroid goiters in iodine-deficient regions should contain somatic
TSHR mutations. Scintigraphically, nonsuppressible thyroid areas have
been demonstrated in 40% of euthyroid goiters in iodine-deficient
regions (8). Moreover, microscopic foci of
autoradiographically increased uptake of 125I
have been demonstrated in euthyroid goiters of previously
iodine-deficient areas (1, 13, 105, 106). Therefore,
we recently studied microscopic autonomous areas with increased
125I uptake in euthyroid goiters studied by
autoradiography 20 yr ago (105) for the occurrence of
somatic TSHR mutations. The detection of constitutively activating TSHR
mutations in autoradiographically hot areas of euthyroid goiters, as
shown in Fig. 1 (107), marks
an important step in understanding the etiology of thyroid autonomy in
endemic goiters. Minute areas in euthyroid goiters were thus identified
as likely starting foci, probably leading to toxic thyroid nodules in
iodine-deficient goiters. Therefore, iodine deficiency leads not only
to euthyroid goiter, but also to thyroid autonomy. Our finding also
elucidates one possible molecular basis for the long-discussed
follicular heterogeneity in euthyroid goiters, although it does not
explain why heterogeneity, as defined by morphological or
immunohistochemical criteria, is also present without detectable
autonomy or nodular alterations (1, 10, 13, 106, 108).
|
TSHR mutations have been detected with a frequency of up to 80%
in toxic thyroid nodules (69). Our screening for TSHR
mutations suggests a much lower frequency in hot microscopic areas from
euthyroid goiters (107). These results imply that
autoradiographically labeled microscopic areas with constitutively
activating TSHR mutations preferentially develop into autonomously
functioning thyroid nodules. A likely explanation for this finding
could be a process of extreme clonal selection and expansion driven by
a constitutively activated TSHR. A constitutively activated TSHR
overrides the normal feedback control that limits thyroid epithelial
cell growth and function. TSHR mutations in hot microscopic areas from
euthyroid goiters demonstrate that TSHR mutations coincide with thyroid
hyperplasia. Indeed, TSHR mutagenesis might actually be initiated by
thyroid hyperplasia. This process is likely to include the following
sequence of events (see also Fig. 2). 1)
Iodine deficiency results in impaired thyroid hormone synthesis. In
animal models of iodine deficiency a significant reduction in
T4 levels is detectable (109, 110).
Impaired thyroid hormone synthesis causes thyroid hyperplasia
(111). 2) The resulting increase in thyroid cell mass
compensates for impaired thyroid hormone synthesis. 3) The increase in
mitotic activity and proliferation of thyroid epithelial cells
increases the chance of mutagenesis due to an increase in the number of
cell divisions and DNA replications, which leads to TSHR mutations.
Alternatively, iodine deficiency might increase radical generation and
the mutation rate by excessive
H2O2 production (112, 113). 4) Hyperplasia induces cell proliferation, which generates
small thyroid epithelial cell clones. A small cell clone from a
progenitor cell containing an activating TSHR mutation might
proliferate even without paracrine stimulation, because constitutive
activation of the cAMP cascade causes production of autocrine factors
[e.g. IGF-I (32, 33, 34, 35)]. In contrast to a
single constitutively activated cell in which production of an
autocrine factor might not be sufficient to cause
self-stimulation, a small cell clone might produce enough
autocrine factors for self-sustained proliferation. The constitutively
active clone might therefore overcome inhibitory paracrine influences
(e.g. lack of coproduction of an activating factor) from
surrounding normal cells, as suggested by Dawson and Wynford-Thomas
(37). In contrast to iodine deficiency, in areas
with adequate iodine supply where hyperplasia is not common, the
lack of coproduction of autocrine factors through surrounding
nonproliferating cells could prevent clonal expansion of thyroid cells
with activating TSHR mutations. Moreover, the frequency of TSHR
mutations might be lower in iodine sufficiency vs. iodine
deficiency due to a lower number of cell divisions. However, thyroid
hyperplasias caused by goitrogens [e.g. propylthiouracil in
animal experiments (114) or goitrogens in food
(115) and in drinking water (116)] instead
of iodine deficiency could also initiate the suggested line of events
and lead to autonomously functioning thyroid nodules that are diagnosed
in iodine-sufficient areas.
|
Nodular transformation could be the result of neoplastic growth,
defined as clonal expansion caused by an advantageous mutation. This
assumption is supported by data demonstrating a predominant monoclonal
origin of benign thyroid adenomas or adenomatous nodules (64, 68, 117, 118, 119, 120). Moreover, a systematic study in multinodular goiters
demonstrates a larger number of nodules of monoclonal vs.
polyclonal origin (121). All of these studies suggest a
chain of events leading to clonal lesions that is detailed here. For
autonomously functioning thyroid nodules the advantageous mutation
(e.g. in the TSHR or Gs protein)
leads to constitutive activation of the cAMP cascade, which stimulates
growth and function. However, a differential growth advantage of
affected cells is necessarily the result of increased cAMP signaling
over that of unaffected cells. This differential growth advantage of
cells with TSHR or Gs
mutations could be
increased by a relatively lower cAMP level in cells with wild-type TSHR
or Gs
protein. Currently, long-term changes in
TSH levels during goiter formation induced by iodine deficiency are not
well investigated. Animal models with a severe reduction of iodine in
the diet show a tremendous increase in TSH. This has also been shown in
humans living in areas of extreme iodine depletion (122).
In newborn populations in Europe, there is an inverse relationship
between the urinary iodine concentration and the frequency of serum TSH
levels above 50 µU/mL, which is the recall level for the suspicion of
congenital hypothyroidism (123). In contrast, studies by
Brabant et al. (124) suggest a short-term
reduction of serum TSH in adults in response to a moderate decrease in
iodine supply. Barbant et al. (124) hypothesize
that changes in sensitivity to TSH explain thyroid hypertrophy and a
more effective release of thyroid hormones. A similar conclusion has
been drawn from experiments in rats fed a low iodine diet
(125). Therefore, differences in the serum TSH response to
a decrease in iodine supply are probably modulated by the severity of
iodine depletion, as higher depletion leads to an increase in TSH
(reviewed in Refs. 126 and 127). The TSH
response to low iodine is likely to be different for newborns and
children (high increase) compared with adults or adolescents (low
increase to decrease), as studies of extreme iodine depletion in Zaire
also demonstrate an inverse relationship of the TSH level with the age
of the patients (122). Moreover, changes in serum TSH are
likely to depend on the duration of the iodine depletion, as older
people (long duration) show lower TSH levels. Finally, the work of
Fenzi et al. (128) and Berghout et
al. (129) suggests that the serum TSH level decreases
with goiter size and is below normal in patients with large goiters.
This decrease in TSH is probably the consequence of increasing amounts
of autonomous tissue.
In a state with normal or reduced serum TSH (e.g. large
goiter or long duration of iodine deficiency), a mutation affecting the
TSH signaling pathway (i.e. cAMP cascade) resulting in
TSH-independent stimulation (e.g. through a mutation in the
TSHR or Gs protein) could induce differential
growth stimulation and induce clonal expansion. This clonal expansion
of thyroid cells beyond a certain cell mass would result in
hyperthyroidism. It is therefore very likely that endemic goiter in
iodine-deficient areas is the starting point for thyroid autonomy
(129). Nodular transformation is very likely the result of
a long process of clonal expansion, as the frequency of nodular goiters
increases with age (130). Moreover, subclinical
hyperthyroidism characterized by a reduced TSH level probably further
supports growth of the nodule, because the difference between lack of
TSH stimulation in normal thyroid tissue (decreased TSH level) and
growth stimulation by an advantageous somatic mutation (e.g.
causing constitutively activated cAMP) is increased. Nodular
transformation could therefore be the result of early thyroid
hyperplasia (e.g. as the result of iodine deficiency or
other goitrogens), which increases mutagenesis, followed by selection
of advantageous mutations (e.g. constitutively activated
TSHR or Gs
protein). The process of clonal
expansion and thyroid nodule formation is likely to require a longer
time period if the differential growth advantage of mutated cells is
small. To date, an estimate of the differential growth advantage is
difficult because the results of in vitro studies of TSHR
activation are difficult to extrapolate to in vivo
conditions (131). Finally, our hypothesis that nodular
transformations and growth are supported by low TSH levels is also
supported by the observation that thyroid hormone therapy does not lead
to a significant reduction of nodule size (132, 133).
Iodine deficiency, hyperplasia, mutagenesis, and selection of clones
could also explain cold thyroid adenomas or adenomatous nodules (see
Fig. 2) by advantageous mutations that only initiate growth, but not
function, of the affected thyroid cells (e.g. ras
mutations). Moreover, nodular transformation of thyroid tissue after
resection of TSH-secreting pituitary adenomas (134),
nodular transformation of thyroid tissue in Graves disease
(135), as well as goiters in patients with acromegaly
(136) could follow a similar mechanism (see Fig. 2
). The
etiology of the nodular thyroid pathology in these patients is
characterized by thyroid hyperplasia, which is also induced by iodine
deficiency.
Proliferation in toxic thyroid nodules
If TSHR or Gs mutations trigger a process
of increased proliferation leading to the manifestation of autonomously
functioning thyroid nodules and hyperthyroidism, a differential growth
advantage should be detectable as increased proliferation in late
stages of thyroid nodules. We and others have addressed this issue,
studying proliferation markers in nodular and surrounding thyroid
tissue (137, 138). In general, the expression of these
proliferation markers is significantly increased in most autonomously
functioning thyroid nodules compared with that in normal thyroid
tissue. These findings indicate that an increase in cellular
proliferation is still detectable in late stages of thyroid nodules.
However, the 2- to 3-fold increase in proliferation in toxic thyroid
nodules over surrounding tissue (137) or a slightly higher
increase over surrounding tissue when comparing with the area of
highest density of the marker within the nodule (138) does
not seem to be sufficient to explain the formation of autonomously
functioning thyroid nodules. It has been reported that human adult
thyroid cells divide about 5 times during their life (139, 140), and that mitotic activity decreases in the adult rat
(141). A 2- to 3-fold increase in thyroid epithelial cell
proliferation would lead to 7 additional cell divisions, which for a
single cell with a somatic mutation would translate into a cell mass of
about 128 cells given that the life span of the cell does not change.
Therefore, the average increase in proliferation markers in
autonomously functioning thyroid nodules vs. surrounding
tissue does not seem to be sufficient to generate an autonomously
functioning thyroid nodule with thousands of cells. However, peak
stimulation of thyroid cells and, hence, proliferation might be much
higher and is very likely reached long before the patient becomes
hyperthyroid. Indeed, a decline of mitotic activity in thyroid tissue
has been found after prolonged TSH stimulation in a rat model of
prolonged goitrogen administration (142). Such a decline
in TSHR signaling could be the result of receptor desensitization
[e.g. through activation of G protein-coupled receptor
kinases as well as ß-arrestin (91, 92, 93)] or uncoupling
of TSHR signal transduction.
Therapeutic implications
Our hypothesis concerning nodular transformation also affects strategies of thyroid surgery for hyperthyroidism and strategies of radioiodine therapy for thyroid autonomy or euthyroid goiters. Even very small nodular structures that contain cell clones with an advantageous mutation left behind during partial resection of the thyroid could cause a recurrence. Indeed, a lower rate of recurrence has been reported for resections of the thyroid that completely removed nodular tissue (143) compared with partial resection (reviewed in Ref. 144). It is therefore necessary to remove all nodular thyroid tissue. This is especially true for the hilar region of the thyroid that is not removed in subtotal thyroidectomy (144). Moreover, total resection of nodular tissue is also the therapy of choice to deal with possible foci of malignant thyroid tissue within areas of nodular transformation (145). A reliable removal of all nodular tissue requires precise ultrasound and scintigraphy before surgery and careful morphological evaluation of the tissue during surgery. With regard to radioiodine treatment of thyroid autonomy, our hypothesis offers explanations for reports describing difficulties to equally determine the volume of thyroid autonomy in toxic multinodular goiters as a basis for the calculation of appropriate radioiodine doses (146). Finally, it is one explanation for the volume reduction of euthyroid goiters by radioiodine therapy (147).
Conclusion
Based on the identification of TSHR and
Gs mutations in autonomously functioning
thyroid nodules and especially in microscopic areas of euthyroid
goiters as well as on aspects proposed and reviewed recently (2, 4, 127), the sequence of events that very likely lead to thyroid
autonomy in iodine-deficient areas are as follows (1).
Early changes in response to iodine deficiency induce diffuse thyroid
hyperplasia in children (130) (2). Due to
increased proliferation during this stage of thyroid hyperplasia,
mutagenesis is increased, thus resulting in a higher number of cells
bearing a mutation. Some of these mutations confer constitutive
activation of the cAMP cascade, e.g. TSHR and
Gs
mutations (3). In a
proliferating thyroid most cells divide and form small clones. The same
is true for cells with a constitutively activated cAMP cascade. After
reduction of TSHR stimulation by suppressed TSH (e.g. as the
result of increased thyroid mass), small clones with constitutively
activated cAMP cascade gain increased growth advantage due to their
TSH-independent cAMP stimulation. They could thus form small hot foci
that will develop into solitary autonomously functioning thyroid
nodules or nodules within a multinodular goiter. A similar mechanism,
however, with mutations in genes that favor dedifferentiation
(e.g. ras oncogene) could be the origin of cold
thyroid nodules. Indeed, ras mutations are very common in
morphologically characterized thyroid adenomas or adenomatous nodules
(148). However, TSHR mutations are exclusively found in
hot thyroid nodules, whereas ras mutations are exclusively
found in cold thyroid nodules (49, 149, 150).
Hyperplasia is very likely a response to cope with impaired thyroid hormone synthesis and to avoid hypothyroidism as long as possible. However, as outlined above, it probably leads to nodular transformation later in live. As suggested by Dumont et al. (127), hypertrophy, including higher blood flow and enhanced iodide trapping, would be an alternative response to economize iodine utilization and avoid hypothyroidism. As goiter is not common to all people in iodine-deficient areas, this alternative response might occur and go undetected. According to our hypothesis, such a response should prevent thyroid autonomy and nodular transformation. However, as both responses avoid hypothyroidism, and thyroid autonomy and nodular transformation are pathologies that develop later in life and have no obvious effect on reproduction, evolution was very likely not required to decide between the two alternatives.
Footnotes
1 This work was supported by a grant from the Deutsche
Forschungsgemeinschaft (DFG/Pa423/10-1), Deutsche Krebshilfe
(10-1575-Pa1) and IZKF Leipzig, BMB-F, Interdisciplinary Center for
Clinical Research at the University of Leipzig 9504 (01KS 9504,
projects B5, B10 and B14).
Received November 10, 2000.
Revised February 26, 2001.
Accepted March 5, 2001.
References