| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Stazione Zoologica Anton Dohrn (M.D.F., R.D.L.) and Department of Cellular and Molecular Biology and Pathology (R.D.L.), University of Naples "Federico II," 80121 Naples, Italy
Correspondence: Address all correspondence and requests for reprints to: Professor Roberto Di Lauro, Laboratory of Animal Genetics, Stazione Zoologica Anton Dohrn, c/o CEINGE Via Comunale Margherita 482, 80145 Naples, Italy. E-mail: rdilauro{at}unina.it
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Thyroid Gland DevelopmentIV. ConclusionsReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Thyroid Gland DevelopmentIV. ConclusionsReferences |
|---|
Many human disorders appear to be derived from alterations of these initial morphogenetic events. For instance, severe malformations in foregut-derived organs (such as esophageal atresia, tracheoesophageal fistula, lung anomalies, and congenital stenosis of the esophagus and trachea) are common anomalies occurring in one in 2000 to one in 5000 live births (5).
This review will focus on the formation of the thyroid gland, the anterior-most organ that buds from the gut tube. We will concentrate on the events leading to the establishment, migration, organization, and functioning of the thyroid follicular cell (TFC), i.e., the cell type responsible for thyroid hormone biosynthesis, which also represents the most numerous cell population in the thyroid gland. Morphogenesis of TFCs is often disturbed in newborns, resulting in a set of conditions collectively known as thyroid dysgenesis (TD). TD is present in 85% of congenital hypothyroidism (CH), a condition that affects one of 3500 newborns worldwide. Because familial occurrence of TD is rare, this condition was considered a sporadic disease resulting mainly from nongenetic causes such as environmental factors or stochastic events during embryogenesis. This article will review data demonstrating that TD can be a genetic and inheritable condition. The genes known to be responsible for TD, the related molecular mechanisms, and how these affect thyroid function will be summarized. Because the number of TD patients bearing a mutation responsible for the condition is very small, a discussion on other genetic mechanisms that might be responsible for this condition will also be presented.
|
II. Thyroid Gland Development |
|---|
|
TopAbstractI. IntroductionII. Thyroid Gland DevelopmentIV. ConclusionsReferences |
|---|
The cells of the thyroid anlage and the ultimobranchial bodies migrate from their respective sites of origin and ultimately merge in the definitive thyroid gland. In the merging process, both the thyroid anlage and the ultimobranchial bodies disappear as individual structures, and the cells contained in them disperse in the structure of the adult thyroid gland. The cells originating from the anlage continue to organize the thyroid follicles, whereas the C cells scatter within the interfollicular space. Interestingly, in some animals the ultimobranchial structures remain distinct from the rest of the thyroid gland (9).
It was suggested that cells of the ultimobranchial bodies could differentiate toward TFCs (10). Indeed, there are reports of patients with ectopic lateral thyroid tissue and no detectable thyroid tissue in the normal median position (11). In support of a lateral origin of the TFCs is also the presence of seemingly colloid-containing follicles in human patients (12) and mutated mice (13) with persistent ultimobranchial bodies. However, mouse embryonic ultimobranchial bodies, transplanted under kidney capsules, never give rise to the typical TFCs, whereas the ventral thyroid rudiment, in the same condition, does (14). These latter data, together with the absence of evidence showing TFC-specific markers (such as thyroglobulin (Tg), thyroperoxidase (TPO), and others related to the mechanism of thyroid hormone biosynthesis; see Section II.A.5) in the ultimobranchial bodies, suggest that only cells originating in the median thyroid anlage will differentiate into functioning, i.e., TH-producing, TFCs.
In this review we will focus on morphogenetic and differentiation events related to TFCs only; for simplicity we will refer to them as either TFCs or "thyroid cells." Exhaustive analysis of rodent thyroid development has been a useful tool in understanding the mechanism underlying the gland’s morphogenesis. We feel that these results can be extended to other species, including humans (Table 1
), because, when studied, no major differences have been found; those known will be highlighted.
|
, and it is called, in developmental biology terms, "specification" or "determination" because, as a consequence of it, a group of cells are specified or determined to undertake a definite developmental program (see Ref.15 for a discussion on these two terms). Morphological and biochemical changes (to be described later in this review) accompany thyroid cell specification, making them clearly distinct from their neighbors. The first visible change is a thickening of the endodermal epithelium in the foregut, which is referred to as thyroid anlage. Thickening in a restricted region of a cell layer, a common event in the initiation of organogenesis, has been suggested to be an essential event in the generation of signals required for the continuation of organogenesis (16). The thyroid anlage is first identified in mouse embryos at embryonic day (E) 8–E8.5 in the ventral wall of the primitive pharynx, caudal to the region of the first branchial arch (17). This median thickening deepens, forms first a small pit (thyroid bud, E8.5–E9) (Fig. 1) and then an outpouching of the endoderm, which is adjacent to the distal part of the outflow tract of the developing heart. The signal that induces this primary event in the origin of the TFCs is not known. The ventral pharyngeal endoderm lies in close apposition to the heart mesoderm; the role of the endoderm in the specification and differentiation of myocardial cells is well known (18). Conversely, an influence of the developing heart on thyroid organogenesis has not been demonstrated, although a recent article demonstrates defects in foregut secondary to defects in heart organogenesis (19). Interestingly, cardiac malformations represent the most frequent birth defects associated with TD (20, 21). Along the same line, the thyroid has an asymmetric shape, the right lobe being larger than the left, but the lobe size is reversed in patients with dextrocardia (22). Thus, although there are no data demonstrating a role from neighboring cells in the specification of the thyroid anlage, the heart primordium is the most attractive candidate for such a role.
|
By E10, the thyroid primordium appears as a flask-like structure with a narrow neck that rapidly becomes a diverticulum. A small hole at the site of origin in the pharyngeal floor (the foramen cecum) is the remnant of the anlage, connected with the migrating thyroid primordium by a narrow channel (the thyroglossal duct). At E11.5 the thyroglossal duct disappears, and the thyroid primordium loses its connections with the floor of the pharynx and begins to expand laterally (Table 1
). Two days later the thyroid primordium reaches the trachea, which has extended ventrocaudally starting from the primitive laryngotracheal groove.
The molecular mechanisms involved in the translocation of the thyroid primordium have not been completely elucidated. Analysis of mice deprived of the transcription factor Foxe1 demonstrates that the TFC precursors themselves have an important role in the migration process (23), thus suggesting an active migration rather than a passive transport due, for example, to remodeling. However, an active migration of thyroid cell precursors has been recently questioned by the finding that these cells do not appear to undergo a classical epithelial to mesenchymal transition during morphogenesis, in contrast to what was observed in many other migrating cells (24). It is conceivable that the final location of the thyroid is due both to active migration of the precursors, perhaps by a novel mechanism, and to other morphogenetic events occurring in the neck region and in the mouth (25).
4. Completion of organogenesis.
By E15–E16 the thyroid lobes expand considerably, and the gland exhibits its definitive shape: two lobes connected by a narrow isthmus. The mechanisms leading to proliferation of the precursors and to the formation of the lobes remain to be elucidated. Surprisingly, TSH signaling, the best known growth stimulus for adult thyroid cells, does not appear to be involved (26, 27). Also the formation of the two symmetrical lobes does not have a mechanistic explanation. However, impairment of this process occurs, as in the cases of hemiagenesis (see Section III.E). Furthermore, in some animal models (13, 28) the isthmus is absent, and the two lobes remain separated (see Section II.B.2.b).
By E15.5, the first evidence of follicular organization appears with many small follicles disseminated within the gland. At this time calcitonin-producing C cells, derived from the ultimobranchial bodies that have fused with the primitive thyroid at around E14, can also be detected among follicles (29).
Reciprocal interactions between mesenchymal and endodermal cells control the morphogenesis of several other endoderm-derived organs (1, 30, 31, 32), but in thyroid organogenesis a role of neighboring tissues remains to be demonstrated. Previous reports have shown that follicular cells explanted from a developing chick thyroid required fibroblasts obtained from the capsule of the thyroid gland to organize a correct histological pattern in vitro (33). More recent papers have reported that mutations of genes expressed in the surrounding tissues and not in the thyroid bud itself impair the correct organogenesis of the gland (28, 34), thus suggesting that also in the case of the thyroid, cellular interactions are required for normal organogenesis (see Section II.B.2.b).
5. Functional differentiation and onset of hormonogenesis.
The differentiative program of TFCs is completed only when the gland reaches its final location. As a result, TFCs express a series of proteins that are typical of TFCs and that are essential for thyroid hormone biosynthesis. Because there will be no further differentiation event in the life of TFCs, this last differentiation could be called terminal or functional differentiation. Genes typical of this stage appear according to a given temporal pattern: Tg, TPO, and TSH receptor (Tshr) genes are expressed by E14.5 (35); sodium/iodide symporter (NIS) is detected by E16 (26). T4 is first detected at E16.5 (36) (Table 2). The onset of functional differentiation only after completion of migration raises the question of whether a time- or space-dependent signal is responsible for it. In contrast to the hypothesis that the signal is space related is the observation that patients with sublingual thyroid do produce, albeit in low amounts, thyroid hormones (37), thus indicating that the normal location of the TFCs is not an absolute requirement for the onset of functional differentiation. Along the same lines, mice deprived of the transcription factor Foxe1 (23) (see Section II.B.1.c) show a sublingual thyroid that expresses Tg. Taken together, these observations strongly suggest that the normal final location of TFCs is not a requirement for the onset of functional differentiation. However, the precise timing in the start of the gene expression program necessary for thyroid hormone biosynthesis indicates that a genetic mechanism must be responsible for such a control. The players of such a mechanism remain to be identified.
|
In the evolution of chordates, a bona fide thyroid characterized by nonencapsulated follicles first appears in agnathan vertebrates such as the lamprey. However, the development of the thyroid in the lamprey follows a more complex pattern. The pharyngeal anlage, in the ventral part of the pharynx, is the origin of a structure present only during larval life, the endostyle (39). The epithelium of the endostyle differentiates into many types of cells, a group of which displays both peroxidase activity and the ability to trap iodine. These cells are fated to become follicular cells only after metamorphosis, when they will form classic thyroid follicles (40).
Invertebrate chordates, such as amphioxus and tunicates, have a structure in the ventral pharynx that is similar to the lamprey endostyle and has been given the same name. The tunicate endostyle is thought to have an important role in filter feeding. The homology between lamprey endostyle and thyroid stimulated, solely on the basis of morphological observations, the proposal that the tunicate endostyle could be the primitive antecedent of the vertebrate thyroid gland (41, 42). The finding of both iodine-concentrating activity (43) and TPO in a group of cells of the endostyle of these animals (44, 45) strongly supports such a proposal. As will be discussed below, the homology between endostyle and thyroid has gained further support from studies demonstrating that similar genes are involved in the morphogenesis of mouse thyroid and of the chordate endostyle (46, 47, 48, 49, 50).
B. Molecular aspects
The molecular basis of thyroid gland development began to be investigated with the discovery that the transcription factor Titf1/Nkx2-1 (see below for nomenclature), identified as responsible for the thyroid-specific expression of Tg and TPO, is expressed not only in functioning thyroid cells but also in their precursors (35). Subsequently, the transcription factors Foxe1, Pax8, and Hhex were also found to be expressed both in mature thyroid cells and in their precursors. The expression of these factors in the thyroid anlage, at the very beginning of thyroid morphogenesis, immediately suggested that these genes might play an important role in the organogenesis of the thyroid gland. These factors are also present in other embryonic tissues, but all four are coexpressed only in the thyroid anlage (Fig. 2). Thus, the small number of cells in the primitive pharynx fated to become TFCs already at E8.5 are univocally characterized by the simultaneous expression of Titf1/Nkx2-1, (35) Foxe1 (51), Pax8 (52), and Hhex (53). When the thyroid diverticulum forms and begins its migration, the expression of these factors is restricted to the thyroid primordium as they are never expressed in the thyroglossal duct (35). For the rest of its life, a thyroid cell will be hallmarked by the simultaneous presence of Titf1/Nkx2-1, Foxe1, Pax8, and Hhex (Fig. 3). As will be described in detail below, the expression of these four factors is required for the early stages of thyroid morphogenesis. Other genes, both thyroid enriched and ubiquitous, that have been demonstrated to impair the development of the thyroid will be discussed below.
|
|
The Titf1/Nkx2-1 protein is encoded by a single gene in mice and humans. Mouse Titf1/Nkx2-1 is located on chromosome 12, whereas the human TITF1/NKX2-1 is on 14q13 (55) (Table 3). The distribution of Titf1/Nkx2-1 protein and of the corresponding mRNA has been exhaustively studied in rodents. In the primitive pharynx, Titf1/Nkx2-1 is present exclusively in the thyroid anlage, and its appearance coincides with specification of the anlage. Titf1/Nkx2-1 remains expressed in the TFC during all stages of development and in adulthood. Titf1/Nkx2-1 is also present in the trachea and lung epithelium (Figs. 2 and 3) and in selected areas of the forebrain, including the developing posterior pituitary (35). After birth and in adult organisms, Titf1/Nkx2-1 is still present in the thyroid and lung epithelium and in the posterior pituitary (55), whereas its expression in the brain is restricted to the periventricular regions and some hypothalamic nuclei (62). The expression of Titf1/Nkx2-1 in the hypothalamus is reduced in the adult compared with the embryonic brain; however, it increases transiently but markedly before the first endocrine manifestations of puberty (63). Interestingly, Titf1/Nkx2-1 mRNA was also identified in parafollicular C cells (64) and in the epithelial cells of the ultimobranchial body (65), which, even if of neuroectodermal origin, end up in the thyroid gland in close proximity with the TFCs.
|
). At E10.5 the thyroid primordium already appears much smaller in size in comparison with wild type, and at E11 no thyroid cells are detectable; these data and the presence of apoptotic cells suggest that Titf1/Nkx2-1 is required to prevent the initiation of an apoptotic process (67). Hence, Titf1/Nkx2-1 is an essential requirement for the survival of thyroid cell precursors, but it is not required for their initial formation. The Titf1/Nkx2-1–/– mouse has been of great relevance because it demonstrates two important concepts: the first is that athyreosis might be due to the inability of the TFC precursors to survive rather than to lack of their specification; the second consists in the proof of the concept that athyreosis can be due to a single, heritable genetic lesion.
|
Titf1/Nkx2-1 is also implicated in epithelial/mesenchymal signaling. In Titf1/Nkx2-1–/– mice a reduction of the number of cartilage rings of the trachea (68) was observed. Because there is no Titf1/Nkx2-1 in the cartilage, it is very likely that the tracheal rings defect is the result of a defective signaling process, which is normally controlled by Titf1/Nkx2-1 in the tracheal epithelium and which is necessary for the normal development of tracheal cartilage. These interactions represent a necessary step in the morphogenesis of trachea and lungs (70). Supporting this hypothesis are studies carried out on defective anatomical structures in Titf1/Nkx2-1–/– mice, showing that the expression of some well-known signaling molecules is controlled by this transcription factor. Bone morphogenetic protein (Bmp)4, a TGFß-related peptide growth factor, expressed in the growing tip of the branching lung epithelium in a normal embryo, is undetectable in the lungs of Titf1/Nkx2-1–/– embryos (68); hence, Titf1/Nkx2-1 controls, directly or indirectly, Bmp4 expression, and it is very likely that the absence of this growth factor is responsible for the alteration in lung morphogenesis in the mutant. However, a more complex model is required to explain the role of Titf1/Nkx2-1 in the development of the pituitary, a gland that derives from the fusion of two buds, one posterior (derived from the hypothalamus) and another anterior (Ratke’s pouch, derived from the oral ectoderm). Titf1/Nkx2-1 is exclusively detected in the posterior bud. This developing posterior pituitary expresses two growth factors, Bmp4 and fibroblast growth factor (Fgf)8, and is adjacent to Ratke’s pouch, which migrates upward from the mouth cavity and expresses Fgfr2, an Fgf receptor. In mice deprived of Titf1/Nkx2-1 (71), Bmp4 is still expressed whereas Fgf8 expression is abolished in the posterior pituitary and, presumably as a consequence, apoptosis is observed in the anterior bud. Later in development, no pituitary, either anterior or posterior, is present, thus showing that Titf1/Nkx2-1 is required both for the development of the posterior bud and for controlling the expression of a signaling molecule, perhaps Fgf8, that is essential for the survival of the anterior portion. Interestingly, expression of Bmp4 is Titf1/Nkx2-1 independent in the posterior pituitary, whereas it is strictly dependent on Titf1/Nkx2-1 in the lung. Notably, in Titf1/Nkx2-1 null mice apoptosis is also observed in thyroid precursor cells (67). The finding that Fgfr2 is expressed in thyroid cells (72) suggests that an Fgf-dependent mechanism, regulated by Titf1/Nkx2-1 and shared by both thyroid and pituitary cells, is required for the survival of the cells.
The mechanisms responsible for the expression of Titf1/Nkx2-1 are not clear. Because the initiation of Titf1/Nkx2-1 expression is coincident with specification, elucidation of these mechanisms might shed some light on the specification process itself. In mice, an inductive signaling by the axial mesendoderm (73) could be implicated in the initial activation of Titf1/Nkx2-1 in the prosencephalic neural plate. Sonic hedgehog (Shh) is relevant for the ventralizing signal, and Titf1/Nkx2-1 is expressed in patterns that are either coincident or adjacent to domains of Shh expression. Analysis of Shh–/– embryos has revealed that Titf1/Nkx2-1 is indeed regulated by this factor only in the forebrain. Indeed, in Shh-deficient embryos no Titf1/Nkx2-1 is observed in the brain, whereas normal levels of the protein are detected in the thyroid and lung anlage (74). On the basis of in vitro studies, it was proposed that the zinc finger factor Gata6 regulates the transcription of Titf1/Nkx2-1 (75). However, analysis of Gata6–/– chimeric lungs has demonstrated that Titf1/Nkx2-1 is localized normally in epithelial cells of both wild-type and mutated lungs (76).
In conclusion, Titf1/Nkx2-1 appears as a protein that is highly regulated in a strict cell-specific manner, because it regulates different genes in different cell types. Furthermore, some genes appear to depend on Titf1/Nkx2-1 for expression in some cells, whereas they are independent from it in others, as in the case of Bmp4 in lung and posterior pituitary. In the thyroid cells themselves, Titf1/Nkx2-1 clearly plays radically different roles during development, as it controls survival at the beginning of organogenesis and the expression of TFC-specific genes in adult life. This latter role cannot be investigated in knockout mice because thyroid cells disappear before the onset of functional differentiation. A conditional knockout of the gene encoding Titf1/Nkx2-1 is necessary to address the role of this transcription factor in the adult thyroid gland. Other issues that remain to be explored are the identification of the effectors of Titf1/Nkx2-1 action (Titf1/Nkx2-1 target genes) as well as those that control Titf1/Nkx2-1 gene expression. These latter genes, given the remarkable restriction of Titf1/Nkx2-1 gene expression to the emerging thyroid and lung bud, might give important clues about the players responsible for patterning the foregut.
Orthologs of Titf1/Nkx2-1 have been isolated from chicken (77), Xenopus (78), and zebrafish (79). In these organisms, the distribution of Titf1/Nkx2-1 is similar to that of the mouse. Of special interest is the expression of Titf1/Nkx2-1 mRNA in the cartilaginous fish lamprey (80), in the cephalochordate amphioxus (47), and in the urochordate Ciona intestinalis (46), organisms in which the emergence of the thyroid gland during evolution can be traced (see Section II.A.6). In addition to the brain, in these organisms Titf1/Nkx2-1 mRNA is present in the endostyle, the primitive antecedent of the thyroid. It should be stressed, however, that in the endostyle of C. intestinalis Titf1/Nkx2-1 mRNA is not present in the same cells that concentrate iodine and that contain a peroxidase similar to TPO. Thus, it is likely that the primordial function of Titf1/Nkx2-1 was to specify the most anterior part of the gut and not to regulate the expression of genes necessary for thyroid hormone biosynthesis. Titf1/Nkx2-1 would have been recruited for this function later during the evolution of chordates. Along these lines, it is of great interest that a Titf1/Nkx2-1 ortholog has been found in the pharynx of hemicordates (81), which lack a true endostyle. Finally, in Drosophila a gene has been identified, scro, with a homology to vertebrate Titf1/Nkx2-1 higher than the other NK2 genes (82). The presence of scro in the brain and pharynx of Drosophila embryos strengthens the hypothesis that the primitive function of Titf1/Nkx2-1 was to contribute to the specification of the anterior part of the gut.
b. Pax8.
Pax8 (paired box gene 8) is a member of a family of transcription factors characterized by the presence of a 128-amino acid DNA binding domain (paired domain) (52). Within this family, composed of nine members (65), Pax8, on the basis of a higher sequence similarity, forms a subfamily with Pax2 and Pax5 (83, 84). The gene encoding Pax8 (called Pax8 in mice and PAX8 in humans) is located on chromosome 2 in both species (Table 3) (52, 85). The molecular properties of Pax8, as well as its role in controlling TFC-specific gene expression, have been recently reviewed (61). Hence, we will concentrate on the role of the Pax8 protein in the organogenesis of thyroid and other cell types, as deduced by its distribution and, chiefly, by the phenotype of knockout animals. It is relevant to mention that whereas the coexpression of Pax8 and Titf1/Nkx2-1 only in thyroid cells has suggested that these factors can cooperate in the stimulation of TFC-specific genes (86), no direct evidence was offered in support of this hypothesis. However, a recent paper has provided the demonstration that Pax8 and Titf1/Nkx2-1 directly interact in vivo in thyroid cells (87).
In the endoderm Pax8 mRNA is present only in the thyroid anlage (52) (Fig. 2). Like Titf1/Nkx2-1, Pax8 is detected in the developing thyroid from E8.5, i.e., at the time of specification. Expression of Pax8 is maintained in TFCs during all stages of development (Fig. 3) and in adulthood. In the nervous system (52), Pax8 mRNA is transiently expressed in the myelencephalon and through the entire length of the neural tube. No signals are detected in the brain at later stages of the development, nor are they present in the adult brain. In the excretory system, Pax8 mRNA is present in the nephrogenic mesenchyme, which gives rise to the epithelial structures of nephrons as a consequence of the instructive interactions of the growing nephric duct and ureter. Indeed, Pax8 mRNA is expressed in the nephrogenic cord at E10.5, in mesenchymal condensations at E13, in the cortex of the metanephros at E16, and in the adult kidney.
Analysis of Pax8–/– mice (65) revealed the role of this transcription factor during embryonic life. Whereas no phenotype has been detected in heterozygous Pax8+/– mice, homozygous Pax8–/– mice are born at the expected Mendelian frequency but show growth retardation and die within 2–3 wk. These mice do not display any apparent defects in the spinal cord, midbrain/hindbrain boundary, or kidneys. On the contrary, in Pax8–/– mice the thyroid gland is severely affected because neither follicles nor TFCs (Fig. 4) can be detected, and the rudimentary gland is composed almost completely of calcitonin-producing C cells. Hypothyroidism is the cause of death of the mutated animals: the administration of T4 to Pax8–/– mice allows the animals to survive. A detailed study during the early steps of thyroid morphogenesis shows that in Pax8 null embryos the thyroid diverticulum is able to evaginate from the endoderm, but Pax8 is required for further development. In the absence of this transcription factor, at E11.5, the thyroid primordium appears much smaller than wild-type primordium, and at E12.5 the follicular cells are essentially undetectable. Thus, like Titf1/Nkx2-1, Pax8 seems to be required for the survival of thyroid cell precursors and not for their specification. Furthermore, in the thyroid anlage of Pax8–/– mice the expression of Foxe1 and Hhex is strongly down-regulated (88). In addition to these important roles in morphogenesis of the TFC component of the thyroid gland, it has been shown, in cell culture systems, that Pax8 is a master gene for the regulation of the thyroid-differentiated phenotype (89). In conclusion, Pax8 not only is required for the survival of the thyroid precursor cells but also holds a specific upper role in the genetic regulatory cascade, which controls thyroid development and functional differentiation. These functions of Pax8 in thyroid development are consistent with the findings that in other organs Pax genes have a relevant role both in initiating and maintaining the tissue-specific gene expression program (90, 91).
Surprisingly, different members of the Pax2/5/8 family have been identified in the thyroid of Xenopus and zebrafish. In Xenopus, Pax2 is the only Pax gene expressed in the thyroid (92). In zebrafish, both Pax8 and Pax2.1 are expressed in the thyroid, and the latter certainly has an important role in thyroid development because thyroid follicles, and Pax8 expression, are absent in Pax2.1 mutant zebrafish embryos (38).
Our knowledge of Pax8 expression in the protothyroid structures of nonvertebrates is still scarce. Pax8, Pax2, and Pax5 originated from a common ancestral gene (93) that duplicated during the evolution of chordates, after the separation of the cephalochordates lineage. In ascidians the ancestral Pax-2/5/8, homologous to the vertebrate Pax2, Pax5, and Pax8 genes, is detected in the primordial pharynx (94) and in neural tube cells. Its expression in the endostyle has not yet been studied. In amphioxus Pax-2/5/8 mRNA is present in the developing endostyle (48), suggesting that this ancestral gene has been co-opted in thyroid specification in cephalocordates.
c. Foxe1.
Foxe1 (formerly called TTF-2 for thyroid transcription factor-2) was originally identified as a thyroid-specific nuclear protein that recognizes a DNA sequence present on both Tg and TPO promoters under hormone stimulation (54, 95). Rat Foxe1 cDNA was cloned and characterized (51) as a member of a winged helix/forkhead family of transcription factors. The official name for the mouse genetic locus is Foxe1 (FOXE1 for the human locus) (59, 60). In this review we will use Foxe1 throughout and suggest that the name TTF-2 for the protein be abandoned. Foxe1 is located on mouse chromosome 4 (51), and FOXE1 is located on human chromosome 9q22 (23, 96) (Table 3). The molecular properties of Foxe1, as well as its role in differentiated thyroid cells, have been recently reviewed (61). Hence, we will focus on the role of Foxe1 in the organogenesis of the thyroid as deduced by its distribution and, chiefly, by the phenotype of knockout animals.
Foxe1 mRNA is detected at E8.5 in all the endodermal cells of the floor of the foregut, including the thyroid anlage. Hence, at variance with Titf1/Nkx2-1 and Pax8, the expression of which in the pharynx is strictly limited to the thyroid anlage, Foxe1 has a wider domain of expression. However, the expression of Foxe1 is limited posteriorly because no Foxe1 mRNA is present in the lung (Fig. 2). It has also been noted that at E8.5 Foxe1 seems to be much more evident in a region of the pharynx posterior to that of the thyroid precursor cells (51). Expression of Foxe1 in the thyroid cell precursors is maintained during development (Fig. 3) and persists in adult TFCs. The initial report of a discontinuity in expression has been retracted (97). A detailed analysis of Foxe1 expression (98) shows that Foxe1 is present, in addition to the thyroid anlage, in the epithelium lining both the anterior pharynx and the pharyngeal arches but is absent in the pouches. Caudally, Foxe1 is detected along the entire foregut including the future esophagus. According to this expression pattern, at later stages of development Foxe1 is expressed in the tissues derived from the pharyngeal arches and pharyngeal wall: thyroid, tongue, epiglottis, palate, and esophagus. In the adult, Foxe1 is still present in the thyroid, whereas the expression in the esophagus is faint. In ectoderm-derived structures, at an early stage of development, Foxe1 is present in the posterior stomatodeum, in the buccopharyngeal membrane, and in the cells of the roof of the oral cavity indenting to constitute Rathke’s pouch, which will form the various components of the anterior pituitary. At later stages, Foxe1 mRNA expression in the pituitary is down-regulated (51), whereas it appears in the secondary palate, in the definitive choanae, and in the whiskers and hair follicles (98). In humans, FOXE1 (formerly called FKHL15) mRNA is also detected in adult testis (99) and several other tissues (96). However, in the latter case, additional mRNAs of different size have been reported.
The generation of Foxe1 null mice (23) has allowed the elucidation of the role of this factor in thyroid development. Homozygous Foxe1–/– mice are born at the expected ratio but die within 48 h. These mice display no thyroid in its normal location and an absence of thyroid hormones. Furthermore, the mice show a severe cleft palate, probably responsible for the perinatal death, and elevated TSH levels in the bloodstream. This compensatory response reveals normal pituitary functions, as confirmed also by the finding that there are no differences in the levels of the other pituitary hormones between wild-type and Foxe1–/– mice (23). Studies of the early stages of thyroid morphogenesis demonstrate that the budding of the thyroid primordium does not require Foxe1, because the anlage can be easily detected by the expression of Titf1/Nkx2-1 and Pax8, and a normal primordium is formed. However, at E9.5 in Foxe1 null embryos, thyroid precursor cells are still on the floor of the pharynx, whereas in wild-type embryos they are detached from the pharynx cavity and begin to descend. At later stages of development, in the absence of Foxe1, mutant mice exhibit either a small thyroid remnant still attached to the pharyngeal floor or no thyroid gland at all (Fig. 5). The variable expressivity of the phenotype could be due to stochastic events during thyroid morphogenesis. In addition, it is possible that either the individual genetic background or sex-related factors are responsible for the variability of the phenotype of the Foxe1 null mice. It is worth noting that the nonmigrating thyroid cells are able to complete their differentiative process as tested by the synthesis of Tg. Hence, Foxe1 plays an essential role in promoting migration of TFC precursors, whereas both Titf1/Nkx2-1 and Pax8 seem to be relevant in the survival and/or differentiation of these cells. Furthermore, the data that found that in 50% of Foxe1 null mice the thyroid disappears indicate that this gene, too, is implicated in the control of the survival of thyroid cells at a step different from those controlled by Titf1/Nkx2-1 and Pax8. The role of Foxe1 in the adult gland is still a matter of study. Functional studies in cell cultures have demonstrated that Foxe1 can act as a promoter-specific transcriptional repressor (100). Because the Foxe1 null mice die at birth, only the creation of an animal model with a thyroid-specific, conditional knockout of Foxe1 will permit elucidation of the role of this factor in the physiology of the gland.
|
d. Hhex.
Hhex (hematopoietically expressed homeobox) is a homeodomain-containing transcription factor that was first identified in hematopoietic cells (103, 104). The genomic locus encoding Hhex is called Hhex in mice (located on chromosome 19) (105) and HHEX in humans (located on chromosome 10q23.32) (106) (Table 3). The gene is split into four exons and codes for a protein 271 amino acids long in mice and 270 amino acids long in humans. Hhex is characterized by the presence of a proline-rich region, localized at the N terminus of the protein, probably involved in regulating the transcription of the target genes (107).
During early mouse development, Hhex mRNA is expressed in the primitive endoderm and then in the presumptive definitive endoderm cells (53). At later stages, the endodermal expression of Hhex is localized in the ventral gut and, from E8.5 onward, marks the primordium of several organs derived from the foregut, such as thyroid, liver, thymus, pancreas, and lungs (Fig. 2) (53, 108). Among these organs, both developing and adult thyroid (Fig. 3) express Hhex at the highest level (53).
The analysis of Hhex–/– embryos (109) demonstrates that this factor is essential for thyroid morphogenesis in accordance with the finding that Hhex is an early marker of thyroid cells. In Hhex null embryos, at E9.5 the thyroid primordium is absent or hypoplastic, still connected to the floor of the pharynx; notably, no expression of Titf1/Nkx2-1 and Foxe1 mRNA is observed in the thyroid bud (109). A more detailed study (88) has shown that in the absence of Hhex, the thyroid anlage is properly formed and expresses Titf1/Nkx2-1, Foxe1, and Pax8; at later stages, the expression of all these transcription factors is down-regulated. The possibility that the direct cause of the impaired development of the thyroid is the absence of the other factors cannot be excluded; the role of Hhex could then be to maintain the expression of Titf1/Nkx2-1, Foxe1, and Pax8 mRNA in the thyroid anlage. However, the relationship among these factors could be more complex. In both Titf1/Nkx2-1 and Pax8 null embryos Hhex mRNA is undetectable in the thyroid remnant (our unpublished data). These observations indicate that Titf1/Nkx2-1 and Pax8 are both required to maintain the expression of Hhex. This regulatory network between transcription factors seems to be in place in differentiated TCFs also, as reported in a recent paper (110) showing that Titf1/Nkx2-1 regulates the activity of Hhex promoter in thyroid cell lines.
The role of Hhex in the adult thyroid cannot be studied in Hhex null mice because thyroid cells disappear at an early stage. Some reports have indicated that Hhex may act as a transcriptional repressor in cell cultures (107) as well as in vivo (111, 112). Consistently, in thyroid cell lines, the overexpression of Hhex partly inhibits Tg promoter activity; furthermore, the level of Hhex in these cells is down-regulated by TSH (113). These observations led to the hypothesis that the control of Hhex levels could be another mechanism by which TSH regulates the expression of Tg.
In vertebrates, Hhex orthologs have been identified in chicken (103), Xenopus (114), and zebrafish (112). In all these organisms Hhex mRNA is expressed in the developing thyroid gland (114, 115, 116).
In conclusion, it has been shown that, in mice, the thyroid anlage, although distinguished by early expression of Titf1/Nkx2-1, Foxe1, Pax8, and Hhex, does not require these factors for the initial steps of morphogenesis (88). These results have also been obtained in zebrafish, in which nk2.1a, pax2.1, and hhex are required relatively late in thyroid development (117).
In the thyroid, as in many endoderm-derived organs, the genes expressed in the budding regions are known, but the genes required for bud formation have not yet been identified. In the future, the identification of the genes controlling Titf1/Nkx2-1, Foxe1, Pax8, and Hhex expression could provide information on how thyroid precursor cells differentiate themselves from their neighbors in the floor of the primitive pharynx.
2. Genes involved in the late stages of thyroid organogenesis
a. Tshr.
Tshr is a protein 765 amino acids long both in humans and in mice and belongs to the superfamily of G protein-coupled receptors. TSH binds to the extracellular portion of the receptor, a long amino-terminal extracellular domain that includes a succession of leucine-rich repeats. The COOH portion of Tshr forms the transmembrane and intracellular domains involved in transducing signals (118, 119, 120). Tshr, localized on chromosome 14q31 in humans (121, 122) and chromosome 12 in mice (123) (Table 3), spreads over 60 kb and is split into 10 exons. The extracellular amino-terminal domain is encoded by nine exons, whereas the transmembrane domain and cytoplasmic tail are encoded by a single large exon (124). It is interesting to note that other G protein-coupled receptors (such as adrenergic or muscarin receptor genes) are devoid of intron; therefore, Tshr seems to have evolved from an intronless protoreceptor fused to a set of duplicated genes coding for a leucine-rich sequence (124). The expression of Tshr mRNA is detected in rat thyroid at E15 (35, 125) (corresponding to E13.5–E14 in mice) and strongly increases by E17. Hence, Tshr mRNA is detected in the developing thyroid after the completion of the migration of the primordium, before the first evidence of follicular organization in the gland. The way the activation of Tshr regulates both proliferation and functioning of adult thyroid cells has already been exhaustively reviewed (126, 127). Hence, we will focus on the role of TSH/Tshr pathway during thyroid organogenesis. The analysis of thyroid development (26) in mice carrying spontaneous (128) or induced (27) alterations in the Tshr gene has provided a powerful tool in the exploration of the role of the TSH/Tshr pathway during embryonic life. Both the Tshrhyt/hyt mice, characterized by a loss-of-function mutation in the Tshr gene (129), and the Tshr null mice display a severe hypothyroidism, associated with thyroid hypoplasia in adult life. However, at birth, in both these mutants, the size of the thyroid does not appear to be affected, and the gland displays only some alterations in its structure (27, 130, 131). A detailed analysis performed at the end of the organogenesis, at E17, has revealed that in the absence of a functional Tshr, the size and the follicular structure of the thyroid are not affected, and the amount of Tg does not change, whereas the expression of both TPO mRNA and NIS is strongly down-regulated (26). These data indicate that, during embryonic life, the TSH/Tshr signaling is required to complete the differentiative program of the TFC, but, unlike what happens during adult life, this signaling is not relevant in controlling the growth of the gland.
b. Hoxa3 and Eya1.
The Hox genes belong to a large gene family (39 in both mice and humans) distributed in four different chromosomal complexes (132). The Hox genes encode a class of transcription factors containing a homeodomain DNA binding domain related to Drosophila antennapedia. In all organisms, the role of these factors is to regulate, during their development, the regionalization of the embryo along its major axes. Furthermore, the analysis of Hox null mutant mice has revealed that these genes are involved in the morphogenesis of several structures.
Some genes of the Hox family, expressed in the foregut during embryonic life, could be involved in the development of the thyroid. In particular, Hoxa3 is detected in the floor of the pharynx, in the developing thyroid (133, 134), and in the mesenchymal, endodermal, and neural crest-derived cells of the fourth pharyngeal pouch (13). The generation of a mutated mouse in which the gene had been disrupted has confirmed the role of this gene in thyroid organogenesis. Hoxa3 null mice (135), in addition to the absence of thymus and parathyroid, show thyroid hypoplasia. A more detailed analysis of Hoxa3–/– mice (13) revealed a variable expressivity and penetrance of the thyroid phenotype: the isthmus was either absent or displaced cranially, the number of follicular cells was reduced, and one lobe of the gland was absent or hypoplastic. Furthermore, the embryos show severe alterations in the development and migration of the ultimobranchial bodies, which do not fuse with the thyroid primordium (persistent ultimobranchial bodies), and a reduced or absent C cell population in the thyroid. The phenotype of mice carrying various mutant combinations in Hoxa3 and its paralogs Hoxb3 and Hoxd3 (136) was also analyzed. Both Hoxb3–/– and Hoxd3–/– single mutant mice have a thyroid gland that appears normal. However, both the double mutants Hoxa3–/– Hoxb3–/– and Hoxa3–/– Hoxd3–/– mice show a 100% penetrance of the thyroid and ultimobranchial body phenotype. The finding that the exacerbation of the thyroid defect corresponds to an increased severity of the ultimobranchial bodies alteration suggests the hypothesis that the defects observed in TFCs could be secondary to defects in the ultimobranchial bodies. Hence, it is conceivable that Hox3 paralogs do not play a direct role in the morphogenesis of the thyroid but could have an important role in the normal development and migration of the ultimobranchial bodies (136). This hypothesis is supported by the study of the phenotype of mouse embryos deprived of a functional Eya1 gene (28). At an early stage of embryonic life, Eya1 is expressed in the pharyngeal arches’ mesenchyme, in the pouches’ endoderm, and in the surface ectoderm of the clefts. Later, it is clearly evident in the thymus, parathyroid, and ultimobranchial bodies but is not detected in the developing thyroid. In Eya1 null mice, the thyroid phenotype is almost identical to the phenotype displayed from Hoxa3 mutants. Indeed, the embryos show persistent ultimobranchial bodies, hypoplasia of the lobes, absence of the isthmus, and a reduced number of follicular cells. Because Eya 1 is not expressed in the thyroid diverticulum, it is possible that the defects in follicular cells are due to the lack of fusion of the ultimobranchial bodies to the thyroid lobes. Consistent with this hypothesis is the observation that the early stages of thyroid development are not impaired (28) because ultimobranchial bodies touch the thyroid diverticulum and fuse with it by E14 (29). In the late stages of the organogenesis of the gland, interactions between the neural crest-derived cells of the ultimobranchial bodies and the epithelial cells of the thyroid diverticulum could be required. This hypothesis is confirmed by the observation that mice carrying mutations in the Pax3 (137) or Endothelin-1 (34) gene show defects in the thyroid similar to those observed in mice deprived of Hoxa3 or Eya1. Both Pax3 (138) and Endothelin-1 (139) are expressed in the pharyngeal arch and are implicated in the development of neural crest-derived structures. These interactions between the ultimobranchial bodies and thyroid diverticulum, critical for a correct morphogenesis, seem to be a unique feature of the mammalian thyroid because in chicken and fish the ultimobranchial bodies remain as bilateral structures and do not merge with the thyroid diverticulum (9). Indeed, zebrafish carrying mutations in sucker gene, homologous to mammalian Endothelin-1 (140), do not display any thyroid phenotype (K. Rohr, personal communication).
3. Other genes
a. Fgfr2.
The Fgf family includes at least 22 peptide growth factors that bind and activate specific tyrosine kinase receptors (Fgfr) (141). During embryonic life, the interactions between these growth factors and their receptors are implicated in the genetic pathways regulating cell differentiation and proliferation (142). In particular, one of the receptors, the Fgfr2-IIIb isoform (Table 3), is expressed in many types of epithelial cells and is activated by Fgfs (Fgf1, Fgf3, Fgf7, and Fgf10) that are present in the surrounding mesenchyme (143). In many cases it has been shown that the activation of Fgfr2-IIIb mediates the epithelium-mesenchyme cross-talk required for the development of different organs (143, 144, 145). Both mutated mice expressing a soluble dominant negative form of Fgfr2-IIIb receptor (146) and mice deficient for the same isoform (145) show absence of the thyroid. Furthermore, in Fgf10 null mice the thyroid is missing (144). These data strongly suggest that the interaction of Fgf10 with its receptor Fgfr2-IIIb is relevant for thyroid organogenesis. However, these reports indicate that the thyroid primordium is absent at E13, without indicating the stage at which thyroid morphogenesis is impaired. It is possible that Fgf10/Fgfr signaling is required for the progression of already established differentiative programs. Indeed, Fgf10 can constitute a critical mitogenic activity for precursor cells in the developing pituitary (147) and in the pancreas (148) and can act as a survival factor against apoptosis in limb bud growth (145) or hair follicles development (144). It has been reported that Fgfr2 mRNA is detected in the thyroid primordium starting from E11.5 (72). This finding suggests that thyroid precursor cells become competent to respond to Fgf10 after the budding of the primordium, as also shown in the thymus (149). Thus, whereas in lung morphogenesis Fgf10/Fgfr interactions are necessary for the inductive signaling required for lung bud formation itself (150, 151), in the thyroid Fgfr activation appears to be essential only after budding and initiation of migration.
b. Nkx2-6, Nkx2-3, and Nkx2-5.
In addition to Titf1/Nkx2-1, other genes of the Nkx2 family, such as Nkx2-6, Nkx2-3, and Nkx2-5, are expressed in the endodermal layer of the developing pharynx, including the thyroid anlage, as well as in other tissues. However, at E8.5 the expression of Nkx2-6 withdraws from the midline region and becomes restricted to the pharyngeal pouches only (152, 153). Consistent with this expression pattern, in the absence of Nkx2-6 the thyroid does not show any apparent phenotype (154). Unlike Nkx2-6, during embryonic life Nkx2-3 mRNA is detected along the entire pharynx from the oral floor to the lung bud. Nkx2-3 is present at high levels in the thyroid diverticulum and persists in the developing thyroid but becomes undetectable at birth (153). Although Nkx2-3 is strongly expressed in the thyroid, in Nkx2-3 null mice the gland appears histologically normal (153, 155).
Nkx2-5 mRNA (Table 3) is present in the ventral side of the pharynx and in the thyroid anlage at an early stage of development (153, 156). However, in the thyroid primordium its expression wanes around E12.5 (our unpublished results). Because of the early mortality of Nkx2-5–/– embryos, it is not easy to identify the role of this factor in thyroid morphogenesis. It was observed that, at E9.5, in Nkx2-5 null embryos the thyroid bud is present even if its size is smaller than that of a wild-type thyroid bud. Furthermore, the expression of Titf1/Nkx2-1, Foxe1, and Pax8 in the thyroid primordium is not impaired (our unpublished data).
In the developing thyroid, the specific role of Nkx2-3 or Nkx2-5, if there is one, has not yet been identified. Because the expression domains of Nkx2-3 and Nkx2-5 overlap, these factors could have redundant functions during organogenesis. Indeed, an essential role of Nkx2-6 in the development of the pharynx is revealed only in embryos deprived of Nkx2-5 (157). These data suggest that in the thyroid, also, the absence of one of these factors could be compensated by a different member of the family.
c. Hepatic nuclear factor 3ß (Hnf-3ß).
Hnf-3ß, now called Foxa2, is a member of the forkhead family of transcription factors and was originally identified as a transcription factor regulating the expression of liver-specific genes (158). Hnf-3ß has a wide and early expression in embryonic tissues. In particular, it is also expressed in the invaginating foregut endoderm (159) and then in the endoderm-derived structures including the developing thyroid (160) (Fig. 2). The expression of Hnf-3ß is down-regulated during thyroid development (160, 161) before TFCs accomplish their terminal differentiation (Fig. 3). However, Hnf-3ß is detected in cultured thyroid cells and in the adult thyroid gland (162). It is hard to identify the relevance of Hnf-3ß during thyroid morphogenesis because the disruption causes an embryo-lethal phenotype at a stage preceding that of thyroid bud formation (163).
III. Molecular Pathology of Thyroid Development Disorders
A. Genetics of TD
CH is the most frequent endocrine disorder in newborns, with an incidence of about one in 3500 live births (164) in iodine-sufficient regions. With the exception of rare cases due to hypothalamic or pituitary defects, CH is characterized by elevated levels of TSH in response to reduced thyroid hormone levels.
In 15% of cases, the disease is caused by inborn errors in the mechanisms required for thyroid hormone synthesis (165, 166); these cases show classical Mendelian recessive inheritance and very frequently lead to enlargement of the gland (goiter), presumably as a consequence of the elevated TSH levels. In the remaining 85%, CH is due to disturbances in the gland’s organogenesis, which result in a thyroid gland that is absent (thyroid agenesis or athyreosis), hypoplastic (thyroid hypoplasia), or located in an unusual position (thyroid ectopy). All these entities are grouped under the term "thyroid dysgenesis" (TD) (167). The relative proportions of the main TD phenotypes vary in different studies, depending on the methodology used to detect the presence of the gland. According to 99Tc scintigraphy (20, 168, 169), more sensitive but exclusively dependent on metabolic activity, ectopy of the thyroid is the most frequent type of dysgenesis (48–61% of cases), whereas athyreosis is the cause of 15–33% of cases of TD. The scintigraphic analysis could either not reveal the frequency of thyroid hypoplasia (168, 169), or most likely underestimated it at only 5% of CH patients (20). It is conceivable that these studies included in athyreosis some nonfunctioning thyroids and counted as hypoplasia only the smallest glands. In a study based on ultrasound diagnosis, which is independent from metabolic activity but not as sensitive as scintigraphy (170, 171), athyreosis was found in 48% of CH patients, ectopy in 18%, and hypoplasia in 17%. This technique most likely included in athyreosis some very small, ectopic thyroids. From these data one could conclude that the most frequent form of TD is thyroid gland ectopy and that a combination of ultrasound sonography and 99Tc scintigraphy might resolve a precise relative proportion of these phenotypes. However, cost-benefit considerations should be carefully evaluated before performing 99Tc scintigraphy in cases of CH.
A small minority of cases of TD are not associated with reduced thyroid function. Hemiagenesis of the gland, for example, is a thyroid developmental anomaly that does not cause clinical symptoms by itself (172).
CH with TD occurs mostly as a sporadic disease. However, there is much evidence indicating that genetic factors are involved in the pathogenesis of this disorder.
TD is known to show a clear female prevalence (169, 173, 174, 175). One report (168) suggested that the female prevalence is significant for ectopy but not for athyreosis. Epidemiological studies have shown a different incidence of the disease in different ethnic groups (173, 176, 177), suggesting that genetic background plays a role in this condition. Furthermore, in populations where consanguineous marriages are common, the incidence of CH is increased (178). Other evidence in favor of the relevance of genetic factors in CH with TD is the finding of a small but significant proportion of familial cases. It has been reported that 2% of patients had an affected relative (175). This frequency is 15-fold higher than the frequency expected on the basis of chance alone, indicating that the involvement of genes required for correct thyroid morphogenesis is very likely in these familial forms. Interestingly, in some familial cases the affected members show either athyreosis or ectopy. This finding supports the hypothesis that athyreosis and thyroid ectopy could have common underlying mechanisms, as strongly suggested by the observation that mice deprived of Foxe1 gene products show either ectopy with a very small thyroid or no thyroid at all (23).
It is also reported that among the first-degree relatives of patients with sporadic CH with TD, there is a significantly higher rate of asymptomatic thyroid developmental anomalies than in normal populations. Indeed, the prevalence of these anomalies (including hemiagenesis or ectopy of the thyroid, thyroglossal duct cysts, pyramidal lobe) is less than 1% in the control population, whereas it is 8% in first-degree relatives of patients with CH (179). This suggests the hypothesis that both severe forms of TD and heterogeneous asymptomatic alterations could originate from the same genetic defects during thyroid morphogenesis. Interestingly, in asymptomatic anomalies there is no female preponderance (179).
The strongest argument against the notion of heritable TD is the finding that in 12 of 13 monozygotic twin pairs there is discordance for TD (180). This finding suggests that postzygotic events must be evoked in the pathogenesis of many cases of TD. However, it is most likely that the targets of such events are the same genes responsible for normal thyroid development. As a working hypothesis, we suggest that the function of genes involved in normal thyroid development can be interfered with by either mutations or epigenetic events. In both cases the result is TD, which would be inheritable, of course, only in the case of genetic mutations.
It should be stressed, however, that the demonstration that mutations in genes involved in thyroid development cause TD in animal models (Table 4) and that mutations in the same genes are associated with TD in patients shows unequivocally that TD can be a genetic and inheritable condition and offers us the tools to reveal the underlying molecular defects.
|
B. Athyreosis
The absence of TFCs in orthotopic or ectopic location is called athyreosis. This condition could be the consequence of lack of formation of the thyroid bud or could result from alterations in any step after the specification of the thyroid bud causing a defective survival/proliferation of the precursors of the follicular cells.
In athyreotic patients, the presence of cystic structures resulting from the persistence of remnants of the thyroglossal duct is frequently reported (182). This finding indicates that in these subjects some of the early events of thyroid morphogenesis have taken place but the cells fated to form the TFCs either did not survive or switched to a different fate. In many cases, scintigraphy failed to demonstrate the presence of thyroid tissue, but thyroid scanning by ultrasound reveals a very hypoplastic thyroid. The recent demonstration that TSHR pathway controls in vivo the expression of NIS (26) could explain some apparent athyreoses in patients with loss of function mutations in the TSHR gene. In some cases neither radioisotope scanning nor ultrasonography detect any thyroid tissue, but the presence of detectable levels of serum Tg suggest that some thyroid tissue, which is not revealed by current procedures, must be somewhere functioning (182).
The study of thyroid development in normal and mutated mouse embryos indicates that the simultaneous presence of Titf1/Nkx2-1 (67), Foxe1 (23), Pax8 (65), and Hhex (109) is required for thyroid morphogenesis. Hence, inactivating mutations in just one of them could be responsible for athyreosis in humans (Table 4). Indeed, the absence of thyroid was reported in patients with CH associated with FOXE1 defects and in one subject carrying a mutation in PAX8.
1. FOXE1 disease.
Bamfort-Lazarus syndrome (Online Mendelian Inheritance in Man 241850) (183, 184) is characterized by cleft palate, bilateral choanal atresia, spiky hair, and athyreosis. The finding that Foxe1–/– mice display thyroid defects and cleft palate has led to the hypothesis that FOXE1 could be a candidate gene for this syndrome. Even if defects in choanae and hair follicles have not yet been investigated in Foxe1 null mice, Foxe1 is expressed in both these structures (98). Indeed, two homozygous mutations in FOXE1 gene have been described in two pairs of siblings affected by this syndrome (99, 185) (Table 5). All the affected members carry homozygous missense mutations in conserved amino acids within the Foxe1 forkhead domain. The mutant proteins were tested in vitro and have shown a reduction in both DNA binding and transcriptional activity. Interestingly, the extrathyroid alterations appear less severe in the subjects that have a residual level of Foxe1 activity in vitro. In contrast, thyroid tissue is undetectable in all the patients. In mice, the absence of Foxe1 causes either athyreosis or defects in thyroid migration; in humans, ectopic thyroid associated with FOXE1 mutations has not yet been described. Because these mutations have been reported in only four patients, we must wait for more data before concluding whether FOXE1 defects in humans cause only athyreosis.
|
).
C. Ectopic thyroid
During embryonic life, the developing thyroid migrates from the thyroid anlage region to its definitive location in front of the trachea, leaving behind the foremen cecum. Occasionally, the thyroid primordium (or a portion of it) fails to descend along the normal pathway, and the gland develops in an abnormal position. The ectopic thyroid can be found in any location along the path of migration from the foramen cecum to the mediastinum. In the majority of cases, the ectopic thyroid appear as a mass in the dorsum of the tongue (lingual thyroid, usually functioning) (186). Sublingual ectopic tissues are rather less frequent (186); in this case, thyroid tissue is present in a midline position above (suprahyoid), below (infrahyoid), or at the level of the hyoid bone. Ectopic thyroid tissues within the trachea have also been reported (1
