FETAL THYROID PHYSIOLOGY

Thyroid hormone is essential for the growth and maturation of many target tissues, including the brain and skeleton. As a result, abnormalities of thyroid gland function in infancy and childhood result not only in the metabolic consequences of thyroid dysfunction seen in adult patients, but in unique, adverse effects on the growth and and/or maturation of these thyroid hormone-dependent tissues as well. In most instances, there are critical windows of time for thyroid hormone-dependent effects and so the specific clinical consequence of thyroid dysfunction depends on the age of the infant or child. For example, newborn infants with congenital hypothyroidism frequently have hyperbilirubinemia, and delayed skeletal maturation, reflecting immaturity of liver and bone, respectively, and they are at risk of permanent mental retardation if thyroid hormone therapy is delayed or inadequate; their size at birth, however, is normal. In contrast, hypothyroidism that develops after the age of three years (when most thyroid hormone-dependent brain development is complete) is characterized predominantly by a deceleration in linear growth and skeletal maturation but there is no permanent effect on cognitive development. These clinical consequences find their determinants in tissue-specific differences in the timing of maturation of thyroid hormone metabolism and thyroid hormone responsiveness discussed in this chapter. In general, infants with severe defects in thyroid gland development or inborn errors of thyroid hormonogenesis present in infancy whereas babies with less severe defects or acquired abnormalities, particularly autoimmune thyroid disease, present later in childhood and adolescence. In the newborn infant, thyroid function is influenced not only by the neonate’s own thyroid gland but by the transplacental passage from the mother of factors that affect the fetal thyroid gland.

In the last several decades, there have been exciting advances in our understanding of fetal and neonatal thyroid physiology, and screening for congenital hypothyroidism has enabled the virtual eradication of the devastating effects of mental retardation due to sporadic congenital hypothyroidism in most developed countries of the world. In addition, advances in molecular biology have led to new insights regarding the early events in thyroid gland embryogenesis and mechanisms of thyroid action in the brain and are beginning to help unrave the molecular basis for many of the inborn errors of thyroid hormonogenesis and thyroid hormone action. However, new questions and new challenges remain. The survival of increasingly small and premature fetuses has resulted in a growing number of neonates with abnormalities in thyroid function and a continuing controversy as to which of these infants require therapy. Similarly, despite the increasing number of genes that have been associated with thyroid hormone-dependent brain development, the critical mechanism(s) of thyroid hormone action on the developing central nervous system (CNS) are only beginning to be understood. This chapter will focus on current concepts regarding the ontogenesis of thyroid function, metabolism and action in the fetus and neonate, with special reference to the effects of thyroid hormone on the developing brain and will review the major disorders of thyroid gland function in infants and children.

Ontogenesis of thyroid function and metabolism in the fetus and infant

Much of our knowledge regarding the maturation of thyroid function in the fetus and neonate is derived from studies in rats and sheep. These two animal species are excellent models of human thyroid gland development and differ mainly in the timing of events. Thus the rat thyroid gland is much less mature at birth than its human counterpart. As a result, significant maturation of the thyroid gland and of the hypothalamic-pituitary-thyroid axis in this species occurs in the first 2 or 3 weeks after birth in the absence of placental or maternal influence, as compared with the third trimester in human infants. In contrast, the neonatal sheep thyroid is relatively more mature than is its human counterpart so that this species is a model of thyroid gland development in the presence of maternal and placental influence.

The ontogeny of mature thyroid function involves both hypothalamic-pituitary-thyroid gland organogenesis and maturation as well as the maturation of each of the component systems for thyroid hormone action. These include the pathways for thyroid hormone metabolism and thyroid hormone action. The placenta plays a key role in the transfer of hormones and factors other than T4 that impact on thyroid function. In the first half of pregnancy, maternal T4 provides an important source of hormone for the developing fetus.

Thyroid gland embryogenesis

The rodent model

Embryogenesis. The thyroid gland is derived from the fusion of a medial outpouching from the floor of the primitive pharynx, the precursor of the thyroxine (T4)-producing follicular cells,and bilateral evaginations of the fourth pharyngeal pouch, which give rise to the parafollicular, or calcitonin (C)- secreting cells (1). As illustrated in Table 15-1, commitment towards a thyroid-specific phenotype as well as the growth and descent of the thyroid anlage into the neck results from the coordinate action of a number of novel, recently-cloned transcription factors. These include the thyroid transcription factors (TTF)-1 and TTF-2, and Pax 8 all of which are expressed in the rat prior to or just following the first appearance of the thyroid diverticulum on fetal day 9.5-10 (2-6). TTF-1, a homeodomain-containing transcription factor belongs to a large family of genes that impart important patterning information in embryogenesis (2). TTF-1 is expressed in both follicular cells and C-cells and targeted disruption of the TTF-1 gene by homologous recombination results in mice completely devoid of thyroid tissue (7). In contrast, the thyroid glands of mice lacking Pax 8, a paired domain-containing transcription factor, are reduced in size, lack follicles and are composed almost exclusively of C-cells (8). Other homeodomain-containing or Hox genes, (Hoxa-3, and the paralogous gene Hoxb-3), appear to be important in the expression of Pax 8 and TTF-1, respectively (9,10). Each of these transcription factors is also expressed in a limited number of other cell types, suggesting that it is the specific combination of transcription factors and possibly non-DNA binding cofactors acting coordinately that determine the specific phenotype of a cell.

Both TTF-1 and Pax 8 also regulate thyroid-specific gene expression. However, the mRNAs for both thyroglobulin, thyroid peroxidase , and the thyrotropin (TSH) receptor do not appear until 5-5.5 days later (fetal day 15-15.5) after migration is complete, indicating that additional factors must be involved in the expression of these genes (2,11). TTF-2, a forkhead domain-containing binding protein, is down regulated between fetal day 13 and fetal day 15 in mice (equivalent to fetal day 14 and fetal day 16 in the rat) (4). This has led to the suggestion that in addition to its role in the commitment to a thyroid-specific phenotype, TTF-2 acts both to promote migration and to repress thyroid-specific gene expression in thyroid follicular cell precursors. This interpretation is supported by findings in TTF-2-null mutant mice that develop one of 2 phenotypes. Either the thyroid gland fails to develop, or a sublingual thyroid gland is formed. As predicted, the sublingual gland demonstrates evidence of thyroid differentiation, at least as indicated by thyroglobulin expression (12). In addition, caudal migration of the thyroid appears to result indirectly as a consequence of the growth and expansion of adjacent tissues (12a).

Terminal maturation. At fetal day 15, despite early evidence of thyroglobulin, thyroid peroxidase and TSH receptor gene expression, the thyroid gland is difficult to distinguish from the surrounding structures. At this time, neither iodine organification nor thyroid hormonogenesis is present and only a primitive follicular structure can be discerned. Thus, TTF-1 and Pax 8 are necessary but not sufficient for the expression of the fully differentiated thyroid phenotype. On fetal day 17, TSH receptor gene expression is significantly upregulated and this is accompanied by significant growth and by rapid development in both structural and functional characteristics (13). Expression of both thyroglobulin and thyroid peroxidase mRNA is also increased at this time, thyroid follicles first appear on morphological examination, thyroid peroxidase function can be demonstrated and there is evidence of thyroid hormonogenesis (13-15). The aforementioned findings suggest that the TSH receptor plays an important role only at this later stage of development, but is not involved earlier in gestation. In support of this hypothesis, hyt/hyt mice which have a loss of function (Pro556-Leu) mutation in the transmembrane domain of the TSH receptor, have severe hypothyroidism, and hypoplastic but normally located thyroid glands with a poorly developed follicular structure (16). Similar findings have been described in babies born to mothers with potent TSH receptor blocking antibodies and in some babies with loss of function mutations of the TSH receptor (17,18). The upregulation of TSH receptor gene expression on fetal day 17 is coincident with the first appearance of pituitary TSH in the fetal circulation and implies that, unlike in the adult, the TSH receptor in the fetus is upregulated by TSH (19). Following fetal day 17 there is continuing maturation of thyroid gland morphology and function, accompanied by increasing TSH receptor expression.

Human thyroid gland development

Information regarding thyroid gland development in the human is much more limited and is summarized in Figure 15-1. Embryogenesis is largely complete by 10 to 12 weeks gestation. At this stage, tiny follicle precursors can be seen, iodine binding can be identified and thyroglobulin detected in follicular spaces (20-24). Thyroid hormones are detectable in fetal serum by gestational age 12 weeks with both thyroxine T4 and triiodothyronine (T3) being measurable (24-26). However, as discussed in further detail below, it is likely that a fraction of the hormones detectable at this early stage is contributed by the mother through transplacental transfer. Thyroid hormones continue to increase gradually over the entire period of gestation as does serum thyroxine-binding globulin (TBG )(Figure 15-1) TBG is present at levels of 100 nmol/L at gestational age 12 weeks and progressively increases up to the time of birth, reaching concentrations of 400 to 500 nmol/L(25,26). The serum TBG concentrations are higher in the infant then in adult humans (~300 nmol/L) as a consequence of placental estrogen effects on the fetal liver. In addition to the increase in total T4, however, there is also a progressive increase of free T4 concentrations between 18 and 36 weeks gestation, indicating a maturation of the hypothalamic pituitary thyroid axis (Figure 15-1)(25,26). The increased total T4/TSH and free T4/TSH ratios also indicate an increased ability of the thyroid gland to respond and suggest that the TSH receptor is probably upregulated at this time.

Figure 15-1. Maturation of fetal thyroid gland development and of thyroid hormone secretion in the human infant.

While thyroglobulin can be identified in the fetal thyroid as early as the 5th week, and is certainly present in follicular spaces by 10 to11 weeks, maturation of thyroglobulin secretion takes much longer and it is not known when circulating thyroglobulin first appears in the fetal serum (not shown). By the time of gestational age 27 to 28 weeks, however, thyroglobulin levels average approximately 100 ng/ml, much higher than in the adult and they remain approximately stable until the time of birth (27).

Iodide concentrating capacity can be detected in the thyroid of the 10 to 11 week fetus, but the capacity of the fetal thyroid to reduce iodide trapping in response to excess iodide (the Wolff-Chaikoff effect) does not appear until 36 to 40 weeks gestation. Thus the premature fetus is more sensitive than the fullterm neonate to the thyroid-suppressive effects of iodine exposure.

The hypothalamic-pituitary axis

TSH is detectable at levels of 3 to 4 mU/L per liter at gestational age 12 weeks and increases moderately over the last two trimesters to levels of 6 to 8 mU/L at the time of delivery (Figure 15-1)(26). When TRH is given to mothers during the second or third trimester, serum TSH rises in the fetal circulation(28). This has been demonstrated as early as 25 weeks gestation in humans (Figure 15-1). Thus, the fetal thyrotroph responds to TRH as early as 25 weeks gestation and in fact, the fetal TSH increment after TRH is greater than is the paired-maternal response. This can be explained either on the basis of enhanced TSH release or impaired TSH degradation, perhaps due to immaturity of the hepatic glycoprotein metabolic clearance system.

The maturation of the negative feedback control of thyroid hormone synthesis, on the other hand, occurs by approximately mid-gestation and elevated serum TSH concentrations have been noted in infants as early as 28 weeks. Conversely, at this time, the administration of 700 µg of T4 into the amniotic fluid 24 hours before Cesarean section causes an elevation in fetal T4 and a reduction in cord serum TSH(29)(Figure 15-1). In addition, the neonatal surge in TSH (see below) is suppressed by such treatment indicating that the hypothalamic-pituitary feedback has begun to mature. Similar inferences can be made from observations that TSH is reduced in the cord serum of infants with neonatal thyrotoxicosis due to the transplacental passage of thyroid stimulating antibodies from mothers with Graves' disease.

Serum levels of TRH are higher in the fetal circulation than in maternal blood, the result both of extrahypothalamic TRH production (placenta and pancreas) and the decreased TRH degrading-activity in fetal serum (30). The physiological significance of these increased levels of TRH in the fetal circulation is not known.

Maturation of peripheral thyroid hormone activation and inactivation processes.

As discussed in Chapter 3, there are three iodothyronine deiodinases involved in the activation and inactivation of thyroid hormone (Figure 15-2). All three are coordinately regulated during gestation and function to closely regulate the supply of T3 to developing tissues. The seleno-enzyme, Type 1 iodothyronine deiodinase (D1), which in adult life catalyzes the bulk of T4 to T3 conversion but which also catalyzes the inactivation of the sulfated conjugates of T4, is low throughout gestation. As a consequence, circulating T3 concentrations in the fetus are quite low and at birth are on the order of 50 to 60 ng/dl (~ 1 nmol/L)(26). The Type 2 deiodinase (D2), another activating enzyme that converts T4 to T3, is detectable by 7 weeks’ gestation. The Type 3 or inner ring deiodinase (D3), an inactivating deiodinase that converts T4 to the biologically inactive isomer reverse T3, is also expressed in fetal brain by 7 wks gestation (31). D2 and D3 are the major isoforms present in the fetus and are especially important in defining the level of T3 in the brain and pituitary.The highest concentration of D2 is in brain, pituitary, placenta and brown adipose tissue. The maturation of D2 activity in brain is tightly linked to thyroid hormone receptor ontogeny both temporally and spatially, further emphasizing the importance of this pre-receptor level of control. D3 is present in many fetal tissues, most prominently the brain, utero-placental unit, skin, and gastrointestinal trac (32,33). This is consistent with the key role of D3 in protecting fetal tissues against high maternal T4 concentrations present either in the placenta or in amniotic fluid.

Figure 15-2. Interrelationships of maternal, placental, and fetal thyroid hormone metabolism. I, II, and III refer to type 1, type 2, and type 3 iodothyronine deiodinases. SO4 indicates a sulfation pathway; T2, 3, 3’-diiodothyronine; T2S, 3,3’-diiodothyronine sulfate. (From Burrow et al. (5), with permission.)

In the presence of hypothyroidism, D2 activity increases while D3 decreases These coordinate activities have been found to be critically important in defending the rat fetus against the effects of fetal hypothyroidism providing that maternal T4 levels are maintained at normal concentrations (34). Despite the low levels of circulating T3, brain T3 levels are 60-80% those of the adult by fetal age 20-26 weeks (24,35). Thus, whereas the physiological interrelationships between the various deiodinases in the fetus and placenta (Figure 15-3) (28,34-36) seem designed to maintain circulating T3 concentrations at a reduced level, specific mechanisms have evolved for maintaining brain T3 concentrations so that normal development can procede. The physiological rationale for the maintenance of reduced circulating T3 concentrations throughout fetal life is still unknown, but it has been suggested that its function may be to avoid tissue thermogenesis and potentiate the anabolic state of the rapidly growing fetus(37).

Figure 15-3. Structures of thyroxine (T4), triiodothyronine (T3), and reverse triiodothyronine (rT3). Sulfation at the 4’ hydroxyl position produces the sulfate conjugates of T4, T3, and rT3. (From Burrow et al. (5), with permission).

As would be predicted from the fact that D1 activity in the fetus is low, the concentrations of the specific substrates metabolized by this deiodinase, namely reverse T3 and the sulfate conjugates of T4, reverse T3, T3 and 3'T2 (Fig.15-3), are markedly elevated in the fetal circulation and to a considerable extent in amniotic fluid(38-45). This occurs because D3, the inner ring deiodinase, has a very low capacity for their deiodination (44). Thus, the bulk of the T4 metabolites found in the fetus are biologically inactive, although there is some evidence that T3 sulfate can suppress TSH, presumably after desulfation by one or more tissue sulfatase enzymes. Whether the high levels of reverse T3 plays any biological role is not known (37).

Role of the Placenta

Contributions of the maternal thyroid to fetal thyroid economy.

In the human infant under normal circumstances, the placenta has only limited permeability to thyroid hormone and the fetal hypothalamic-pituitary- thyroid system develops relatively independent of maternal influence.The relative impermeability of the human placenta to thyroid hormone is due to the presence of D3 which serves to inactivate most of the thyroid hormone presented from the maternal or fetal circulation. The iodide released in this way can then be used for fetal thyroid hormone synthesis.

Interest in the potential role of maternal T4 in the fetal thyroid economy, however, has been reawakened in recent years with the recognition that in infants with the congenital absence of thyroid peroxidase, hence, the complete inability to synthesize T4, T4 is present in cord serum at concentrations between 25 and 50% of normal (46).Similar results are obtained in retrospective studies of cord serum in infants with sporadic congenital athyreosis. This T4 disappears rapidly from the newborn circulation with a half-life of approximately 3 to 4 days.

There is also accumulating evidence that maternal-fetal T4 transfer occurs in the first half of pregnancy, when fetal thyroid hormone levels are low(47) . Low concentrations of T4, presumably of maternal origin, have been detected in human embryonic coelomic fluid as early as 6 weeks gestation and in fetal brain as early 10 weeks gestation prior to the onset of fetal thyroid function. Furthermore, both D2 and D3 activity as well as TR isoforms are present in human fetal brain from the mid first trimester, indicating that the machinery to convert T4 to T3 and to respond to T3 are present.

The critical role of maternal T4 as a source of T3 in the fetal central nervous system has been particularly well studied in the rat. In a series of careful studies, Morreale de Escobar et al have demonstrated that in the iodine-deficient or chemically-thyroidectomized fetal rat, the concentration of T3 in the CNS can be maintained at normal levels by the transplacental passage of maternal T4 (32,48-51). As noted above, it seems likely that when fetal thyroid function is normal, the net flux of T4 from mother to fetus may be relatively limited(52). However, the euthyroxinemic mother and hypothyroxinemic fetus permits significant bulk transfer of T4 to the fetal circulation. This can occur both at the level of the placental maternal capillary interface and via uptake of thyroid hormone from the amniotic fluid through the immature epidermis. T4 uptake by the fetus can also occur via fetal ingestion of amniotic fluid, thus permitting the intestinal absorption of significant quantities of the T4 therein. Instillation of T4 into the amniotic cavity causes significant increases in fetal serum T4 and suppression of TSH in the term fetus (29). This approach has been advocated as a method of delivery of T4 to an athyreotic infant e.g.,following inadvertent radioiodine administration during pregnancy(53). While the T4 concentrations in amniotic fluid appear modest, the fraction of T4 free in amniotic fluid is approximately ten-fold higher than that of serum and thus the free T4 concentration in amniotic fluid is approximately equal to that in fetal serum at 20 weeks gestation. It has been shown on numerous occasions in both animals and humans that maternal amniotic fluid iodothyronine concentrations reflect those in the maternal circulation(42).

Placental permeability to thyroid hormone is greatest in the rat and least in sheep. This difference in placental permeability may be due to the fact that the sheep placenta contains components of both maternal and fetal origin, whereas humans and rats have hemochorial placentas consisting of exclusively of fetal tissue.

Placental permeability to maternal TRH, TSH, and to other factors

As noted, the placenta is freely permeable to TRH and to iodide which is essential for fetal thyroid hormone synthesis (28). The placenta is also permeable to certain drugs, hormones and to immunoglobulins of the IgG class. Thus, the administration to the mother of excess iodide , drugs (especially propylthiouracil or methimazole), or the transplacental passage of TSH receptor antibodies from mothers with severe Graves’ disease or primary myxedema may have significant effects on fetal and neonatal thyroid function.

Maternal TSH does not cross the placenta. Similarly, thyroglobulin is undetectable in the serum of athyreotic infants, indicating the absence of any transplacental passage of this large protein (54).

In summary, these results illustrate that human prenatal thyroid physiology is the consequence of a complicated series of interactions between the pituitary-hypothalamic thyroid axis and various pathways for activation and inactivation of thyroid hormones. Maternal thyroid function can play a critical role in the fetus. Accordingly, maintaining normal T4 concentrations during gestation is of greater importance to the fetus than has been recognized previously. The potential for T4 transfer from amniotic fluid to the fetus provides a pathway by which levothyroxine can be administered to fetus in the rare circumstances where this is indicated (54a).

Thyroid Function In The Full-Term Neonate, The Infant, And During Childhood

The full-term neonate

Marked changes occur in thyroid physiology at the time of birth in the full term newborn. (Figure 15-4) One of the most dramatic changes is an abrupt rise in the serum TSH which occurs within 30 minutes of delivery. Peak TSH concentrations in the full term infant can reach 60 to 70 mU/L (20). This causes a marked stimulation of the thyroid and an increase in the concentrations of both serum T4 and T3 (55,56). These consist of an approximate 50% increase in the serum T4 and an increase of three to four fold in the concentration of serum T3 within 24 hours (Figure 15-3). Studies in experimental animals suggest that the increase in TSH is a consequence of the relative hypothermia of the ambient extrauterine environment (20). However, while a significant portion of the marked increase in T3 from its low basal levels in cord serum can be explained by the abrupt increase in TSH, there is also an increase in D1 activity at the time of delivery. Thus, serum T3 concentrations rise to levels which are typical of those expected in adults. As a consequence, the elevated concentrations of the other substrates of D1, reverse T3, and T3 sulfate, decrease relatively rapidly during the newborn period(38). D2 has also been identified in human brown adipose tissue and the acute increase in T3 in this tissue at birth is required for optimal uncoupling protein synthesis and thermogenesis (57,58).

Figure 15-4. Postnatal TSH, T4, T3, and rT3 secretion in the full-term and premature infant in the first week of life (modified from Fisher DA (37) ).

Infants and Children

Following the acute perturbations of the neonatal period there is a slow and progressive decrease in the concentrations of T4, free T4, T3 and TSH during infancy and childhood (20). The serum concentration of reverse T3 remains unchanged or increases slightly. The most important aspect of thyroid physiology in the infant and child, however, is the markedly higher T4 turnover in this age group relative to that in the adult. In infants, T4 production rates are estimated to be on the order of 5 to 6 µg/kg per day decreasing slowly over the first few years of life to about 2 to 3 mg/kg/day at ages 3 to 9 years(20). This is to be contrasted with the production rate of T4 in the adult which is about 1.5 µg/kg/day. Serum thyroglobulin levels also fall over the first year of life reaching concentrations typical of adults by about 6 months of age. The size of the infant thyroid gland increases quite slowly. The thyroid gland of the newborn weighs approximately 1 gram and increases about 1 gram per year until age 15 when it has achieved its adult size of about 15 to 20 g. In general, the size of the thyroid lobe is comparable to that of the terminal phalanx of the infant or child's thumb.

Premature infants

Thyroid function in the premature infant reflects, in part, the relative immaturity of the hypothalamic-pituitary-thyroid axis that is found in comparable gestational age infants in utero. Thus, in cord blood samples obtained by cordocentesis, there is an increase in the TSH, TBG, T4 and T3 concentration in fetuses with increasing degrees of maturity (Figure 15-5) (26). More recent studies performed in umbilical cord and fetal blood samples suggest that this increase in TSH, free T4 , sulfated T4 and reverse T3 is not gradual but peaks during the late second and early third trimester, with a subsequent decrease to term (58a). Following delivery, there is a surge in T4 and TSH analogous to that observed in term infants, but the magnitude of the increase is less in premature neonates and there is a more dramatic fall in the T4 concentration over the subsequent 1-2 weeks (Figure 15-6). This decrease in the T4 concentration is particularly significant in very low birth weight infants, (<1.5 kg, approximately equivalent to <30 weeks gestation) in whom the serum T4 may occasionally be undetectable (59,60). In most cases, the total T4 is more affected than the free T4, a consequence of abnormal protein binding and/or the decreased TBG in these babies with immature liver function. In addition to the aforementioned changes in T4 and TSH concentrations, the serum reverse T3 tends to stay higher and serum T3 reduced for a longer period in the premature newborn, reflecting the greater immaturity of D1 (Figure 15-3)(61).

Figure 15-5. Postnatal changes in T4 secretion in the premature infant according to gestational age. (From Mercado et al (59), with permission).

Figure 15-6. Maturation of thyroid hormone effects in the human fetus and neonate. The left edge of the bars indicate the approximate time the effects of thyroid hormone become manifest (From Fisher DA (67), with permission).

The causes for the decrease in T4 observed postnatally in premature infants are complex. In addition to the clearance of maternal T4 from the neonatal circulation, preterm babies have decreased thyroidal iodide stores (a problem of particular significance in borderline iodine-deficient areas of the world), they are frequently sicker than their more mature counterparts, are less able to regulate iodide balance (62), and they may be treated by drugs that affect neonatal thyroid function. In addition, since the capacity of the immature thyroid to adapt to exogenous iodide is reduced, there is an increase in sensitivity to the thyroid-suppressive effects of excess iodide found in certain skin antiseptics and drugs to which these babies are frequently exposed (see below).

Despite the reduced total T4 observed in some preterm babies, the TSH concentration is not significantly elevated in most of these infants. In some babies, transient elevations in TSH are seen, the finding of a TSH concentration >40 mU/L being more frequent the greater the degree of prematurity. Frank et al found, for example, that the prevalence of a TSH concentration >40 mU/L in very low birth weight, (<1.5 kg), i.e., very premature, infants was 8-fold higher and in low birth weight,(1.5 kg-2.5 kg) neonates 2-fold higher than the prevalence in term babies(63). Whereas in some cases, an elevated TSH concentration may reflect true primary hypothyroidism, in other instances, this increase in TSH seen in the preterm infants at several weeks of age may reflect the elevated TSH observed in adults who are recovering from severe illness(64). Such individuals may develop transient TSH elevations which are associated with still reduced serum T4 and T3 concentrations.These have been interpreted as reflecting a "re-awakening" of the illness-induced suppression of the hypothalamic pituitary axis. (See Chapters 4 and 5) As the infant recovers from prematurity associated illnesses such as respiratory distress syndrome (RDS), a recovery of the illness-induced suppression of the hypothalamic pituitary thyroid axis would also occur.

Somewhat surprisingly, given the relative immaturity of the thyroid gland, serum thyroglobulin concentrations may be higher in the premature than in the full term infant. For example, median serum thyroglobulin concentrations of 102 µg/L were observed in 45 premature infants, as opposed to levels of 73 ng in the full term infant (65). This could be explained on the basis of increased secretion of a poorly iodinated thyroid hormone precursor or impaired clearance of this glycoprotein from the circulation by the immature liver. However, because of the attenuated TSH burst at the time of birth, the serum thyroglobulin concentrations tend not to rise as greatly in the postpartum period in the premature as in the full term infant.

Small-for-gestational-age (SGA) infants

SGA infants have significantly higher TSH and lower total and free T4 values than do infants of normal weight (66). This can be related to the severity of the malnutrition in these infants, presumably in part due to fetal hypoxemia and acidemia. This may be a reflection of impaired placental perfusion and chronic starvation. This pattern of reduced T4 and elevated TSH differs from the response to starvation in older individuals and healthy adults in whom TSH is reduced. The explanation for the relatively higher TSH in such infants is not known.

Ontogenesis Of Thyroid Hormone Action In The Fetus And Infant

The ontogenesis of thyroid hormone-mediated responsiveness, like thyroid hormone metabolism, is temporally and spatially regulated and tightly coordinated. Whereas thyroid hormone-mediated effects in the pituitary, brain and bone can be detected prenatally, thyroid hormone-dependent action in brown adipose tissue, liver, heart, skin, and carcass are apparent only postnatally (Figure 15-7)(67). Because of the pivotal importance of thyroid hormone on brain development, there have been numerous studies evaluating the ontogeny and mechanism of action of thyroid hormone action on this model system.

Figure 15-7. Brain neurologic development relative to thyroid function in the rat and human. TH, thyroid hormones; dpc, days postconception; dpn, days postnatal. (From Porterfield and Hendrich, with permission.)

Thyroid Hormone and Central Nervous System Development

Chronological Correlations of Thyroid and CNS Development in Humans and Experimental Animals: the role of maternal T4

The best studied model for thyroid-hormone-related events in central nervous system development is the rat. Many of the concepts developed from this animal model appear to be applicable to the human. However, a major advantage of the rat as an experimental model is that at the time of birth the rat central nervous system, like the thyroid, is relatively immature compared with the human full term infant. Therefore, events which take place in the latter third of gestation in the human can be studied in the rat over the first 2 to 3 weeks of life. In order to compare the two species, however, it must be recognized that the differences in the timing of these critical steps in brain maturation need to be correlated with the timing of the development of thyroid function in the two species.

The various phases of fetal and neonatal thyroid hormone physiology and the correlations in the rat and human of those events with the critical stages of central nervous system maturation are depicted in Fig. 15-8(68). The major difference between the rat and human in the development of the thyroid and levels of circulating thyroid hormones at various stages is the relatively long period of intrauterine life of the rat during which the maternal thyroid is the only source of fetal hormone available to the fetus. Thyroid function in the rat fetus does not begin until approximately four days prior to birth. Thus, in the rat, a substantial portion of cerebral neurogenesis and migration takes place during the time when thyroid-hormone dependent processes would be markedly influenced by maternal thyroid status. On the other hand, most neuronal differentiation and synaptogenesis in the human has been initiated just at the end of the first trimester and myelinogenesis is well underway at the time of birth whereas this does not occur until much later in the rat. It is thus apparent that the impact of congenital hypothyroidism will be different in the two species since the human fetus is dependent for a much longer period on its intrinsic thyroid function than is the rat. While it is not possible to make a precise determination of when thyroid hormone-dependent central nervous system developmental processes begin, there is a critical period in neurogenesis after which reversal of hypothyroidism cannot normalize brain function. In the rat, this period extends from at least 18 to 19 gestational days but may be as early as 14-16 days(68-70). In the human, studies of infants with congenital hypothyroidism show that normal or near normal intellectual development can be achieved in most if not all infants as long as adequate and sustained levothyroxine replacement is instituted within a few weeks of birth(69,71-74). However, as noted above, accumulating evidence suggests that maternal thyroid hormone may even play a role early in gestation, prior to the onset of fetal thyroid gland development (47,75,76).

Figure 15-8. Cord blood concentrations of T4, FT4, T3, and TSH in premature infants according to gestational age. Lines indicate the mean, 5th, and 95th percentile. Hatched areas represent lower limits of sensitivity for the assays. (From Thorpe-Beeston et al (26), as modified by Franchi (60), with permission).

Endemic cretinism provides a critically important clinical model for elucidating the thyroid hormone dependent events in the human central nervous system because both fetal and maternal thyroid function are deficient. A specific neurological syndrome has been described. The major features of this syndrome are mental disability, deaf-mutism, pyramidal tract disturbance, extrapyramidal dysfunction with a spastic diplegia or quadriplegia and a characteristic gait disorder. These findings have been particularly well described in studies in China and Indonesia by Boyages and Halpern(77,78). The most common manifestation of pyramidal dysfunction is hyperreflexia of the patellar (but not achilles) reflexes. Extrapyramidal features of neurological endemic cretinism include rigidity, dystonic posturing, and other lesions indicating a disorder of the basal ganglia. Supporting these observations have been recent MRI studies demonstrating abnormalities in the corpus striatum and substantia nigra in three adult Chinese endemic cretins(78). The clinical features suggest the abnormalities typical of cerebral palsy. Delong has attributed the onset of damage to these systems to about 14 weeks gestation with persistence through to the third trimester(79). This is consistent with the earlier findings of Pharoah et al that that iodized oil will prevent the development of the neurological manifestations of cretinism providing it is given before the second trimester of pregnancy(80). An impairment of thyroid hormone-dependent events in the second trimester can also explain the abnormalities in the cochlea (developing during the 10th to 18th week), the impairment of cerebrocortical function (developing during the 14th to 18th week) and abnormalities in the basal ganglia (12^th to 18th week)(68,77). None of the neurological features of severe endemic cretinism are found in infants with sporadic congenital hypothyroidism who receive early and adequate postnatal treatment. Even infants who displayed clinical manifestations of hypothyroidism at birth, presumably indicating a severe deficiency of endogenous thyroid function and a failure of maternal T4 to compensate for this, do not have evidence of extrapyramidal or pyramidal disease, and impaired hearing, when found, is much milder and less frequent (81-83, 83a, 83b). This would appear to provide unequivocal evidence that the neurological damage sustained by infants with endemic cretinism can be prevented by maternal T4 as in the example of infants with congenital athyreotic hypothyroidism. Consistent with this interpretation, significant developmental delay despite early and adequate postnatal therapy has also been reported in other models of maternal-fetal hypothyroidism, such as materno-fetal POU1F1 (formerly called Pit 1) deficiency (84) and TSH receptor blocking antibody-induced congenital hypothyroidism (85). It is not known as yet whether neurological abnormalities similar to those found in endemic cretinism occur in these other disorders nor whether treatment prior to the second or third trimester will normalize the impairment seen in cognitive function.

Even maternal hypothyroidism unassociated with any fetal thyroid abnormality has been associated in with an impairment in psychomotor development in the offspring, a result first noted by Man (86). Haddow et al detected a 4 point IQ deficit in 7 to 9 year old children whose mothers were hypothyroid in the first trimester of pregnancy (86a). In support of this observation, Pop et al have demonstrated that even babies born to women whose free T4 levels are in the lowest 10% of normal have a measurable impairment in psychomotor development at 10 months of age as compared with the rest of the population (86b). Because this effect was observed only at 12 weeks but not at 32 weeks gestation, the authors concluded that maternal thyroid hormone was most important in the first trimester of gestation, prior to the onset of fetal thyroid function. In contrast, Liu et al failed to demonstrate an IQ deficit in babies born to hypothyroid mothers as long as the hypothyroidism was corrected in the second trimester.

The weak effect of maternal hypothyroidism alone should be contrasted with the severe deficits observed in combined maternal-fetal hypothyroidism due to iodine deficiency. In the latter case, an effect is observed primarily when iodine is not administered before the second trimester suggesting that while a weak effect is observed earlier, most thyroid-hormone dependent brain development occurs later in gestation. Nonetheless, the incidence of maternal hypothyroidism during pregnancy (3 per 1000 in iodine-sufficient populations (86c)) is almost 10 times that of congenital hypothyroidism for which routine population screening is widespread. This has led some to recommend routine thyroid screening of all pregnant women preferably prior to the first trimester. (86d, 86e).

Ontogenesis of Thyroid Hormone Receptors in the Central Nervous System

The classical effects of thyroid hormone occur through interaction of T3 with its nuclear receptor(88). Accordingly, considerable attention has been given to the study of ontogenesis of T3 receptors in the central nervous system since such data are critically important for assessing the timing and localization of thyroid hormone-dependent processes. Little information is available for the human fetus. Specific nuclear T3 binding proteins have been identified in human fetal brain by 10 weeks gestation, and during the period 10 to 18 weeks there is an increase in the level of these receptors (35,89). This is subsequent to the appearance of fetal thyroid hormone in the circulation(22,25,26). Specific radioimmunoassay techniques have demonstrated a high degree of saturation of the receptors with T3 (24,35). More recent studies have demonstrated the presence of the mRNA encoding the α1, α 2, and β1 isoforms of the receptor as early as 7 to 8 weeks and receptor protein was first observed by immunocytochemistry as early as 7 weeks with increasing expression from the first to second trimester (47). Monocarboxylate transporter (MCT 8), an important membrane transporter of thyroid hormone into brain cells also is first expressed at 7 weeks (47).

In the rat, T3-binding nuclear proteins in the cerebral cortex have been identified by day 13 of fetal life and these increase progressively to the time of birth (90-94). At that time, the nuclear T3 binding capacity is about half that of the adult. Similarly, studies using specific antisera directed against the α1 isoform of the T3 receptor have shown a marked increase in this protein during the late gestational period, presumably in neurons(91). In the case of the brain, this is associated with a marked increase in the mRNA encoding this species. A number of investigators have compared the levels of mRNA encoding receptor mRNAs in the central nervous system(92,95-97). As discussed in Chapter 3, the ß1, ß2 and α1 receptors are the biologically active isoforms. However, the ß2 form of the receptor is of special interest with respect to central nervous system development since it appears to be localized to specific structures, including the pituitary, paraventricular nuclei, arcuate and ventromedial nuclei of hypothalamus as well as the median eminence(98). This relatively CNS specific ß receptor isoform has also been identified in the basal ganglia by day 19 of rat gestation. Given the fact that hypothyroid neonatal rodents develop deafness and cochlear dysfunction, the detection of ß1 and ß2 mRNA in the portions of the embryonic ear giving rise to the cochlea at 12.5 days of fetal life is quite intriguing(93,99). The ß2 isoform has been identified in the chick's retinal outer nuclear layer early in embryogenesis, though this decreases at later times in development(97). This is especially important since thyroid hormone has a marked influence on the development of the amphibian eye.

It is of interest that while the α1 and ß2 mRNA isoforms are dominant in the central nervous system, they do not undergo developmental changes. Since these are present prior to the increase in the ß1 mRNA, it has been speculated that in the central nervous system, as in the tadpole in general, T3-induced induction of ß1 thyroid hormone receptor may occur by its interaction with the α1 receptor protein in this tissue (95,100-102). While T3 receptors have been identified in both neurons and glia, their highest concentration is in neurons (103-106). In the rat, the neuronal nuclei express TRß1 whereas only c-erb-A2 protein was detected in astroglial nuclei(105). As is discussed below, specific thyroid hormone dependent processes have been identified in both neurons and glia. Thyroid hormone nuclear receptors appear during the time of neuroblast proliferation and, at least as judged by studies in the rat, increase markedly as this animal enters phase three of central nervous system development (Fig. 15-8). This is a logical pattern of development and implicates thyroid hormone receptors in these effects. Furthermore, while the quantities of mRNA encoding the α1 and α2 isoforms of the thyroid hormone receptor are much higher than those of TRß1, developmental changes appear to be most specifically associated with increases in TRß1 (and perhaps ß2) mRNAs suggesting that these receptors may be the critical inducers of thyroid hormone effects. Changes in the ß1 isoform have been especially well demonstrated by immunofluorescent studies in the cerebellar Purkinje cells, which develop relatively later in neonatal life than do cerebrocortical cells. These correlate well with specific effects of T3 directed transcription of certain Purkinje cell genes (91,107). Studies of the TR mRNAs during differentiation suggest the possibility that the α1 thyroid hormone receptor could play a role in the differentiation process once neuroblast development has occurred.

Until recently, one of the unexplained paradoxes has been the surprising lack of developmental abnormalities seen in mutant mice lacking TRα1, TRβ or both, in contrast to the severe abnormalities observed in hypothyroid animals. Emerging evidence suggests that the reason for the abnormal brain development observed after thyroid hormone deficiency but not TR deficiency is transcriptional repression by the unliganded TR. For example, when mutant mice lacking the TR α1 receptor were made hypothyroid no effects on cerebellar development were seen, contrary to findings in wild-type animals (108).

It is likely that the complexity of maturational control of thyroid hormone action involves developmental regulation of a myriad of factors that affect TR activity. These factors include corepressors and coactivators as well as transcription factors that compete with TRs for thyroid hormone response elements (TREs) on target genes, providing further levels of modulation (109).

Overview of Effects of Hypothyroidism on the Central Nervous System

At a functional level, thyroid hormone provides the induction signal for the differentiation and maturation of a diverse array of processes that lead to the establishment of neural circuits during a critical window of brain development. These processes include neurogenesis and neural cell migration (occurring predominantly between 5 weeks and 24 weeks), neuronal differentiation, dendritic and axonal growth, synaptogenesis, gliogenesis (late fetal to 6 months postpartum), myelination (2nd trimester to 24 months postpartum), and neurotransmitter enzyme synthesis. The absence of thyroid hormone appears to delay rather than eliminate the timing of critical morphological events or gene products, resulting in a disorganization of intercellular communication.

There are few data in humans regarding the pathological concomitants of the functional abnormalities the hypothyroid central nervous system. Some information regarding the anatomical defects is being generated by application of advanced radiological techniques to the evaluation of adult human endemic cretins. The radiological studies mentioned earlier have demonstrated a remarkably normal appearance of the brain with the exception of gliotic lesions of the globus pallidus and substantia nigra. In the few histo-pathological studies that have been performed, the brain has been found to be small, with poor development of cerebral and cerebellar cortex, basal ganglia and thalamus(77). Abnormalities have been identified in the hippocampus which show hypoplasia of pyramidal and granular cells, suggesting that hippocampal defects could be an explanation for the dementia of the endemic cretin, though whether such effects are produced during intrauterine life, or during the neonatal period is not clear. EEG studies in endemic cretins have demonstrated characteristic slowing of the alpha rhythm and other signs suggesting diffuse cerebral dysfunction.

Morphological studies of cortical neurons in the hypothyroid adult rat provide an example of the post-natal effects of thyroid hormone deficiency. The numbers of spines along the apical shaft of pyramidal cells are markedly reduced in hypothyroid rats and effects on pyramidal neurons could be demonstrated at all ages. It is possible to reverse or prevent these effects providing T4 is given within a few days of thyroidectomy. Partial reversal could be obtained if treatment were instituted at later time points. Similar studies of Purkinje cells in the cerebellar cortex were performed many years ago by Legrand et al(110). These showed striking abnormalities of the granular cell layer with disorganized and reduced numbers of dendrites, impairment of Purkinje cell migration to deeper layers of the cerebral cortex and impaired or absent synapses in certain cells of the cerebellar cortex.

Molecular Concomitants of Fetal and Neonatal Hypothyroidism On CNS Development

Consistent with a nuclear receptor-mediated mode of action, thyroid hormone stimulates numerous developmentally regulated genes, including genes for myelin, neurotrophins and their receptors, cytoskeletal components, transcription factors, extracellular matrix proteins and adhesion molecules, intracellular signalling molecules, as well as mitochondrial and cerebellar genes. In some cases these genes appear to be direct targets of thyroid hormone action as thyroid hormone response elements can be detected in the DNA regulatory region and/or the genes are stimulated in cell culture. In other cases, thyroid hormone control may occur secondarily as a consequence of effects on terminal differentiation. In addition, thyroid hormones regulate some genes at the level of mRNA stability or mRNA splicing. For further information about the molecular action of thyroid hormone in the brain, the reader is referred to several excellent reviews (47,109,111,112).

Inherited Thyroid Hormone Resistance and Brain Development

An extensive discussion of thyroid hormone resistance syndromes is presented in Chapter 16. However, this is an important model from the point-of-view of the effects of thyroid hormone on the central nervous system. In the autosomal dominant form of the disorder, attention deficit hyperactivity disorder has been identified in 70% of affected children as opposed to only 20% of their unaffected siblings(113). This gives a relative risk ratio of about 15 fold for the development of this syndrome in such patients. The relationship between this TRβ disorder and neuropsychiatric manifestations is particularly impressive given the similar genetic and environmental background of the individual families. The authors attributed this abnormality to a lack of expression of thyroid hormone effects, but it could also relate to the supra-normal levels of thyroid hormone present in these individuals.

Unfortunately, little is known of the specific functional or anatomical abnormalities leading to this syndrome. Recent studies have reported decreased hearing in some thyroid-hormone resistant patients, and the original patient of Refetoff who has no functional  receptor is deaf but has no labyrinthine dysfunction(114). This is quite intriguing given the evidence of early appearance of TRβ1 and TRβ2 mRNAs in the embryonic cochlear tissue of the rat fetus (93,99). It would point to hearing loss as a marker for early fetal hypothyroidism in humans as it is in rats(69).

In summary, it is clear that thyroid hormone dependent effects in the central nervous system are produced by a variety of mechanisms in both neurons and glia. While certain areas of the brain appear to be more susceptible to hypothyroidism than do others, particularly in the developing human central nervous system, the abnormality is more likely to be generalized rather than due to a disruption in a specific rate-limiting step. It is characterized by delays and concomitant disorganization in intercellular communication rather than in absolute deficiencies in function. On the one hand, it is surprising that cerebrocortical function can be as well maintained as it is in the presence of these diffuse abnormalities. On the other hand, the generality of the process and the critical nature of the chronology of CNS development makes identification of specific defects very difficult to unravel.

CONTINUE