Department of Pediatrics, University of Massachusetts Medical School, Worcester, Massachusetts 01655
Address all correspondence and requests for reprints to: Rosalind S. Brown, M.D., Department of Pediatrics, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655. E-mail: . Rosalind.Brown{at}umassmed.edu
No type of human transformation is more distressing to look at than an aggravated case of cretinism. The stunted stature, the semi-bestial aspect, the blubber lips, retrousse nose sunken at the root, the wide open mouth, the lolling tongue, the small eyes half-closed with swollen lids, the stolid, expressionless face, the squat figure, the muddy dry skin, combine to make the picture of what has been termed the "pariah of nature." Not the magic wand of Prospero or the brave kiss of the daughter of Hippocrates ever effected such a change as that which we are now enabled to make in these unfortunate victims, doomed heretofore to live in hopeless imbecility, an unspeakable affliction to their parents and their relatives.
[Sir William Osler, 1897]
Remarkable progress in our therapy and knowledge of congenital hypothyroidism (CH) has been made since the demonstration by Murray in 1891 that thyroid extract could ameliorate many of the features of untreated cretinism, an advance described in dramatic detail in the above quotation from Sir William Osler. At this time, despite the striking improvement in the clinical features observed in affected patients, the associated developmental delay proved to be less amenable to therapy, and indeed some cognitive delay was thought to be inevitable. It was not until the 1970s that the importance of the timing of postnatal treatment in obviating the mental retardation was demonstrated convincingly. In a study by Klein et al. (1), 78% of infants with CH treated before 3 months of age but 0% treated after 6 months of age had an intelligence quotient (IQ) above 85, the mean IQ of the early treated group being 89, compared with an IQ of 54 in those treated late. Unfortunately, only one third of patients were recognized clinically within the first 3 months, and even fewer (10%) in the first month of life. The subsequent development by Dussault and Laberge (1A ) of a sensitive and specific RIA for the measurement of T4 in dried whole blood eluted from filter paper paved the way for the modern era of newborn screening for CH prior to the development of clinical manifestations.
The legacy of newborn screening for CH
There is no doubt that newborn screening programs have been a resounding success in improving the therapy and, consequently, the outcome of babies with CH. Clearly, the major objective of screening, the eradication of mental retardation, has been achieved. Even the minor neuro-intellectual sequelae that were reported in the most severely affected infants, particularly in the early days of screening when a lower initial replacement dose was used, seem to be preventable as long as postnatal treatment is started sufficiently early and is adequate (2). An additional benefit of newborn screening has been the elucidation of the incidence of CH as well as the prevalence of its various causes. We know now, for example, that nonendemic CH, occurring in 1 in 3,000 to 1 in 4,000 neonates, is one of the most common preventable causes of mental retardation throughout the world, and that it is, in fact, four to five times more common than phenylketonuria, for which newborn screening programs were initially developed. In the vast majority (>90%) of babies, a permanent abnormality of the thyroid gland is discovered. Of these infants, 1015% have an abnormality in one of the steps required for thyroid hormone synthesis (thyroid dyshormonogenesis) whereas 8085% are found to have an abnormality of thyroid gland development (thyroid dysgenesis). In fewer than 10% of neonates, the abnormality is found to be transient. Often, transient CH, more common in premature infants, is due to maternally transmitted iodine, drugs, and immunoglobulins. The syndrome of transient CH due to maternal TSH receptor blocking antibodies is particularly important to be aware of because the picture may mimic the much more common thyroid agenesis (3). CH secondary to pituitary or hypothalamic disease is the least common cause, occurring in just 1 in 50,000 to 1 in 100,000 babies.
The molecular era: thyroid dyshormonogenesis
Analogous to the pivotal role of newborn screening in improving the outcome of affected babies, the major advances in molecular genetics that have occurred in the last decade or two have been of paramount importance in enabling the characterization of numerous genes that are essential for thyroid hormonogenesis and thyroid gland development. This, in turn, has permitted the elucidation of the molecular basis for many of the causes of CH. Nowhere has this been more successful than in the defects of thyroid hormonogenesis, a group of disorders that are distinguishable clinically from the much more common thyroid dysgenesis by the finding of a normally placed (or eutopic) thyroid gland that may be of normal size or enlarged at birth. That this should be so is not surprising in view of their frequent autosomal recessive mode of inheritance, consistent with a single gene mutation. For example, a variety of homozygous and compound heterozygous mutations in the genes for the sodium-iodide symporter, thyroid peroxidase, and thyroglobulin have been shown to underly iodide trapping defects, abnormal organification of iodide, and defective thyroglobulin synthesis or storage, respectively (4). Pendreds syndrome (sensorineural deafness, goiter, and partial organification defect) has been linked to a mutation in the gene for pendrin, a recently discovered porter of iodide on the apical surface of the thyroid follicular cell. Mutations in the newly cloned gene for THOX-2, an NADPH oxidase, have been associated recently with a total organification defect in several patients (5).
Thyroid dysgenesis: a genetic disease?
In contrast to thyroid dyshormonogenesis, the elucidation of the underlying etiology of most cases of thyroid dysgenesis is much less well understood. Thyroid dysgenesis, the term used to describe abnormalities in thyroid gland development, includes both the complete absence of thyroid tissue (agenesis) and a thyroid of decreased size (hypoplasia) with or without associated ectopy. Although most cases of thyroid dysgenesis are sporadic, the 2% familial occurrence, the higher prevalence of thyroid dysgenesis in babies of certain ethnic groups (e.g. Hispanics) than in others (e.g. African Americans), the higher frequency in female vs. male infants, and the increased incidence in babies with Down syndrome, all suggest that genetic factors might play a role in some cases. Furthermore, it has been argued that the apparent familial occurrence may be underestimated because affected children from previous generations usually were not fit enough to reproduce. The transcription factors NKX2.1 (also known as TTF-1, TITF-1, or T/ebp), TTF-2, and PAX-8 would seem to be obvious candidate genes in the etiology of thyroid dysgenesis in view of their important role in thyroid organogenesis and/or migration as well as in thyroid-specific gene expression. To date, however, abnormalities in these genes have been found in only a small proportion of patients with isolated thyroid dysgenesis unassociated with other dysmorphic findings. For example, no germ line mutations in the NKX2.1 gene were found in a total of 76 CH patients studied by two different groups of investigators in Italy (6, 7). Similarly, germ line mutations of the PAX-8 gene were found in only 3 of 145 Italian and German CH patients with thyroid dysgenesis, in one of whom the abnormality was familial with an autosomal dominant mode of inheritance (8). In these and other reported cases, a heterozygous loss-of-function mutation of the PAX- 8 gene has been identified and affected patients have had thyroid hypoplasia with or without ectopy.
Another potential candidate gene to explain the development of isolated thyroid hypoplasia is the TSH receptor. Because this gene is only expressed after the thyroid gland has migrated into the neck, however (9), loss-of-function mutations in this gene could only explain the finding of hypoplasia but not ectopy. Both homozygous and compound heterozygous loss-of-function mutations of the TSH receptor have been described in some patients with TSH resistance (10, 11). An autosomal recessive mode of inheritance has been noted in most reported cases. Most affected babies have had a normal or hypoplastic gland; in rare cases no thyroid gland at all has been discernible on thyroid imaging, a picture indistinguishable from thyroid agenesis. A discordance between the results on imaging and the serum thyroglobulin concentration has been reported in some but not all of these patients. Unfortunately, like the aforementioned transcription factors NKX2.1, TTF-2, and Pax8, TSH receptor gene mutations do not seem to be a common cause of thyroid hypoplasia, even in consanguineous families with two or more children affected by CH (12). TSH resistance due to an inactivating mutation of the stimulatory guanine nucleotide regulatory protein of adenylate cyclase (Gs) (13) is an even rarer cause of CH.
In contrast to the relative failure to identify germ line mutations of NKX2.1, TTF2, and Pax8 in patients with isolated thyroid dysgenesis, emerging evidence suggests that heterozygous mutations in these genes may be a much more important cause of abnormal thyroid gland development when thyroid dysgenesis is associated with other dysmorphic findings. This is consistent with the important role of these transcription factors not only in the thyroid but in nonthyroid tissues as well during embryonic development. For example, heterozygous deletions of NKX2.1 have been reported in a number of patients with CH, unexplained neonatal respiratory distress, and neurological manifestations (14, 15, 16), reminiscent of the findings of abnormal thyroid, lung, pituitary, and forebrain development in mice with a targeted disruption of this gene. Similarly, a homozygous missense mutation in the TTF-2 gene has been associated with the syndrome of thyroid agenesis, bifid epiglottis cleft palate, kinky hair, and choanal atresia (17).
If the majority of patients with isolated thyroid dysgenesis do not have a defect in the genes for NKX2.a, TTF-2, PAX-8, or the TSH receptor, could they nonetheless be genetic in origin? For example, might they arise from a neomutation of a novel gene, such as an as yet unidentified cofactor required for gene expression or, possibly, elements involved in the signal transduction cascades that mediate thyroid embryogenesis? Alternatively, might some cases of thyroid dysgenesis be multigenic (several interacting genes are necessary to produce the phenotype)? The study by Perry et al. (18) in this issue of JCEM would seem to cast further doubt about the role of any of the aforementioned genetic mechanisms in most cases of isolated thyroid dysgenesis. These authors demonstrated that all 12 of 12 twins (5 monozygotic and 7 dizygotic) identified with CH because the initiation of newborn thyroid screening in Quebec and Brussels were discordant for thyroid dysgenesis. Phenotypic variability was excluded as a potential mechanism to explain this discordance by ultrasound examination in four of the monozygotic twin pairs. A similar discordance in the incidence of thyroid dysgenesis in monozygotic twins was noted in an additional six of seven cases reported in the medical literature, although the possibility of reporting bias may have influenced the latter results. Of particular concern was the finding that in a majority of monozygotic twins the diagnosis of CH was missed initially on newborn screening. The authors suggest that this frequent failure to diagnosis CH in monozygotic twins initially is most likely due to fetal blood mixing because the rise in serum TSH concentration between screening and diagnosis, a consequence of the clearance of in utero-acquired T4 from the circulation, was significantly greater in them than in matched singletons. This is a reasonable hypothesis both in view of their shared circulation anatomically, and in view of evidence of similar fetal blood mixing in other situations, such as congenital toxoplasmosis (19). Because a delay in diagnosis and therapy of CH may place monozygotic twins at risk of suboptimal cognitive development, the authors recommend that a second sample for CH screening at 14 d of life should be considered in all twins presumed to be monozygotic on the basis of the same sex, and this is an important practical suggestion.
Conclusions
As the fascinating story of CH unfolds, it is becoming increasingly evident that the clinical finding of thyroid dysgenesis may be due to multiple etiologic mechanisms. In a small number (probably <2%) of patients, isolated thyroid dysgenesis is due to an inherited or de novo mutation in one of the transcription factors important in thyroid gland development. Of these, heterogeneous mutations in the PAX-8 gene associated with thyroid hypoplasia with or without ectopy, and an autosomal dominant form of inheritance, seem to be most common. Alternately, thyroid hypoplasia or apparent aplasia has been associated with both homozygous and compound heterozygous mutations of the TSH receptor gene, and an autosomal recessive form of inheritance. The findings of Perry et al. (18), albeit in a small number of patients, are consistent with the larger body of research, all of which suggest that most cases of isolated thyroid dysgenesis, however, are sporadic in nature. Although at first glance the frequent discordance of thyroid dysgenesis in monozygotic twins may seem surprising, this finding also is consistent with the frequent discordance of birth defects in general in identical twins. Potential etiologic mechanisms that have been proposed to explain this discordance include epigenetic phenomena, early somatic mutations, or postzygotic stochastic events. For example, despite evidence of significant fetal blood mixing at birth, it remains possible that the microenvironments of monozygotic twins are not identical and that differences in the supply of blood, oxygen, or other nutrients may alter or disrupt a genetic cascade of development of the thyroid at a crucial time. In contrast to isolated thyroid dysgenesis, neomutations are more likely to be found in patients with thyroid dysgenesis associated with other dysmorphic features (syndromic thyroid dysgenesis).
The rapid pace of development in our understanding of the genes regulating thyroid embryogenesis and growth as well as of basic genetic mechanisms in general suggests strongly that the further important insights in the etiology of thyroid dysgenesis are likely to be reported in the years ahead. Stay tuned.
Acknowledgments
Footnotes
Abbreviations: CH, Congenital hypothyroidism; IQ, intelligence quotient.
Received July 15, 2002.
Accepted July 16, 2002.
References
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |