July 2008
|
The Physiology of Human Growth: A Review
The complex processes of prenatal, infancy, childhood, and adolescent growth are controlled by genetic, nutritional/environmental, and hormonal factors that vary with the growth phases.
BY Arlan L. Rosenbloom, MD
Human growth physiology can be considered to include the dynamic period beginning with cleavage of the zygote and ending with completion of adolescence, marked by the end of long bone growth. The skeletal infrastructure dictates the magnitude of linear growth; chondrocytes in the cartilage growth plate proliferate, enlarge, and ossify, with ultimate fusion of the distal epiphyseal and central metaphyseal regions. Genetic, nutritional/environmental, and hormonal factors that vary with the growth phases control this complex process. The phases of growth are prenatal, infancy, childhood, and adolescence.
DEFINITION AND NATURAL HISTORY OF HUMAN GROWTH
Prenatal growth. The most dramatic period of growth is the development of the microscopic zygote into the 51-cm (20-in) newborn. After completion of organogenesis in the first trimester, rapid acceleration in the second trimester reaches a peak velocity of 2.5 cm per week. Maternal size and nutritional status are the most important influences on fetal growth. Genetic factors have little influence on fetal growth, except for transmitted or new mutations affecting skeletal growth, such as achondroplasia, or affecting key hormonal mechanisms.1,2 The intrauterine endocrine milieu is a complex interaction of fetal, placental, and maternal substrates, precursors, and hormones affecting growth, including, in addition to the insulin-like growth factors (IGFs), fetal insulin production in response to maternal glycemia, human placental lactogen, and sex steroids. Both IGF-I and IGF-II are essential for fetal growth, and their production in utero is independent of growth hormone (GH). Thyroid hormone is absolutely essential for postnatal growth, but its absence in utero, as with inborn errors of thyroidogenesis or thyroid aplasia, does not affect fetal growth. Production of testosterone by the male fetus, under the influence of maternal gonadotropins, begins at approximately 10 weeks of gestation and is essential to male genital differentiation. The mini-adolescence characterized by elevated testosterone levels near term promotes penile growth and accounts for the observation that newborn males have slightly greater lean mass and less fat mass than females and are on average 0.9 cm longer and 150 g heavier.3
Growth in infancy. Infancy is characterized by a rapidly changing growth rate. The infant shifts from a growth velocity primarily determined by maternal factors to one determined by genetic endowment. Although midparental height standard deviation score or percentile may estimate what this endowment consists of, it is only reliable if the parental heights are actually measured and if their own childhoods were devoid of factors that might have impaired growth. Linear growth is a stepwise process rather than being continuous, as implied by the smoothed growth charts derived from cross-sectional data, a phenomenon which is especially striking in infancy.4 Growth velocity in the first year of life declines from 20 cm per year in the first few months to 10 to 12 cm per year by 1 year of age, by which time length has increased 50% and weight threefold from birth. The influence of parental genetic endowment on infant growth is demonstrable in the shift of growth channels that occurs for about two thirds of normal infants during the first 6 to 18 months of life, with equal numbers shifting upward and downward.5
The effect of the mini-adolescence of the male fetus continues for the first 3 to 6 months after birth, when males grow more rapidly than females. Although it was earlier thought that growth in the first 6 months of life was independent of GH, it is now recognized that GH deficiency and GH-receptor deficiency, resulting in severe IGF-I deficiency, affect postnatal growth from the outset.6
Growth in childhood. Growth velocity averages 10 to 13 cm per year in the second year of life and 7.5 to 10 cm per year in the third year. From age 3 years to puberty, growth is stable at 5 to 6 cm per year, although there may be a slowing to as little as 2 cm per year for a time preceding the adolescent growth spurt. This slow growth is especially noticeable in children with constitutional delay in growth and maturation, and it is frequently accompanied by reduced GH responses to stimulatory testing, misdiagnosis of GH deficiency, and inappropriate treatment with recombinant human (rh)GH. Childhood growth is also characterized by a rapid change in body proportions, when the legs grow faster than the trunk, and both grow much faster than the head in proportion to overall body length. The ratio of upper body to lower segment (measured as the distance from the top of the symphysis pubis to the floor or measuring board with the legs straight) goes from 1.7 at birth, to 1.4 at 2 years, to unity by age 10.3
Growth in adolescence. At the appropriate biologic age as reflected in osseous maturation, suppression of the hypothalamic-gonadotropin axis of childhood begins to be lifted and results in a slow increase in the levels of sex hormones that result in adolescence. Although girls start their adolescence, signaled by breast budding, at an average 6 months earlier than do boys whose signal is testicular enlargement, the adolescent growth spurt is 2 years earlier in girls. Thus, the adolescent growth spurt is early in female maturation and late in male maturation. This timing, giving boys a longer period of slow growth, accounts for some of the greater adult height of males, along with the growth effects of testosterone. The pubertal growth spurt accounts for more than 20% of adult stature and 50% of adult bone mass accrual. Growth is completed when, under the influence of estrogen, either secreted by the ovary or converted by aromatization from testosterone in males, fusion of the epiphyses occurs. In addition to sex hormones, there are large increases in insulin, GH, and IGF-I that contribute to adolescent growth, which, along with normal thyroid function, are essential to the adolescent growth spurt.7
ENVIRONMENTAL FACTORS IN GROWTH
During the 150 years preceding the mid-20th century, there was a secular trend in the pace of maturation and adult size of individuals in the Western countries for which such data are available. A century and a half ago, average males did not reach adult height until 23 years of age as opposed to the current 17 years, and the age of menarche has declined from 17 years to 12.5 years. The most apparent explanation for this phenomenon is the improvement in nutrition and reduction in childhood disease frequency and duration with attendant salutary effects on the GH/IGF-I axis. This secular trend appears to have leveled off in the last 50 years.3 For much of the world, undernutrition remains the most common cause of short stature. Overnutrition with obesity increases the rate of growth, accelerates skeletal maturation, and may advance pubertal onset in girls, but in contrast to the permanent effects of long-term childhood malnutrition or chronic disease, it is not typically associated with an effect on adult height.7
Growth differences in preschool children are more influenced by socioeconomic factors than by racial or genetic factors.8 Seven-year-old boys in families in the upper socioeconomic classes from eight different countries had very similar heights corresponding to the 50th percentile in the United States, demonstrating that size differences between ethnic or geographic groups result from environmental factors, rather than genetics.9
GENETIC CONTROL OF GROWTH
It is estimated that 70% to 90% of adult stature is genetically determined, if nutritional and socioeconomic factors are equal. In addition to the genetic factors affecting the production of and response to insulin, thyroid hormone, sex steroids, and the GH/IGF-I axis (discussed later), extensive genetic control of growth through the expression of numerous genes acting on the growth plate is being increasingly recognized.
Fibroblast growth factors (FGFs) interact with various FGF receptors to regulate growth and development of enchondral bone and longitudinal growth and fusion of long bones. The gene for FGF receptor 2 (FGFR2) is expressed by the earliest chondrocytes and induces the expression of a transcription factor needed for differentiation of the chondrocytes, as well as male genital development, SRY (sex-determining region Y)-box 9 (SOX9). FGFR3 stimulates proliferation of immature cells and limits division of proliferating chondrocytes. The gain of function mutation of FGFR3 is associated with achondroplasia.7 Prehypertrophic chondrocytes produce a protein referred to as Indian hedgehog that coordinates proliferation and differentiation of chondrocytes and osteoblasts as well as the process of bone formation; this protein is self-regulated by its control of parathyroid hormone-related protein, gain of function or loss of function mutations of which result in specific chondrodysplasias with stunting.
The short stature homeobox gene (SHOX) is found on the pseudoautosomal region of the short arms of the X and Y chromosomes and does not undergo inactivation in normal females; short stature in those with absence of the pseudoautosomal region of one X chromosome (ie, Turner syndrome) is attributed to the need for both alleles. Loss of function mutations in one SHOX allele or its isolated deletion has been described in children with otherwise unexplained short stature and in Leri-Weill dyschondrosteosis; loss of both alleles results in another form of osseous malformation, Langer mesomelic dysplasia. In contrast, overdosage of SHOX alleles in girls with triple X syndrome results in very tall stature and may explain the tall stature of other multiple X and Y syndromes. Increased stature has also been associated with variants in the melanocortin-4 receptor and catecholamine O-methyltransferase affecting estrogen metabolism. The suppressor of cytokine signaling (SOCS2) inhibits GH signaling by competitive binding to the GH receptor. Overgrowth occurs in mice that have had this gene knocked out. Overexpression can also lead to overgrowth, suggesting the kind of dual effect that is emerging for a variety of gene products.7
A recently recognized regulator of skeletal growth is C-type natriuretic peptide (CNP). Homozygosity for mutation of the CNP receptor-B resulting in loss of function causes the skeletal dysplasia referred to as Maroteaux-Lamy type (acromesomelic dysplasia). Heterozygous carriers of the mutation were found to be significantly shorter than noncarriers, and it was estimated that approximately 3% of children with idiopathic short stature might be heterozygous for this mutation.10
HORMONAL CONTROL OF GROWTH
Insulin, thyroid hormone, and sex steroids are important components at various stages of growth. Absence of insulin effect in utero results in severe growth failure in leprechaunism or Donohue syndrome. Thyroid hormone, as mentioned, does not affect intrauterine growth but is essential subsequently for differentiation and proliferation of chondrocytes and their ability to respond to growth factors. There are several congenital or genetic defects affecting embryogenesis and migration of the thyroid gland, thyroid-stimulating hormone (TSH) production, the TSH receptor, thyroid hormone production, and conversion to active triiodothyronine. In addition, acquired hypothyroidism due to autoimmune thyroiditis can completely stop a child's growth. Sex steroid effects on bone maturation are via the estrogen receptor; mutations of this receptor or of the aromatase that controls the conversion of testosterone to estrogen result in prolonged bone growth.
Disorders along the GH-IGF-I pathway preceding the IGF-I receptor result in IGF-I deficiency and may be congenital or acquired. Congenital GH deficiency (GHD) is associated with structural malformations of the central nervous system, hypothalamus, or pituitary. IGF-I deficiency/resistance may result from genetic defects involving critical factors in the embryologic development of the pituitary or in the cascade from hypothalamic stimulation of GH release to completion of IGF effects on growth. Acquired abnormalities affecting the GH/IGF axis range from damage to the hypothalamic-pituitary region from trauma, tumors, infection, autoimmune disease, or radiation, to a broad spectrum of conditions characterized by catabolism.
Embryology of the pituitary. Pituitary differentiation in the embryo is in response to an orchestration of transcription factors appearing and disappearing in precise sequence. At 3 weeks' gestation, the ectodermal stomodeum of the embryo develops an outpouching anterior to the buccopharyngeal membrane. This outpocketing is the Rathke pouch, which usually separates from the oral cavity and will give rise to the adenohypophysis (anterior lobe) of the pituitary. An evagination of the diencephalon then gives rise to the neurohypophysis of the pituitary. Rarely, the primitive oral cavity origin of the pituitary results in a functional pharyngeal adenohypothesis.11 Secretion of pituitary hormones can be detected as early as week 12 in the fetus, and some of these hormones are found within the pituitary by 8 weeks' gestation.12
Differentiation of the primordial pituitary requires a cascade of factors be expressed in critical temporal and spatial relationships. These include extracellular signaling factors from the adjacent diencephalon that initiate anterior pituitary development from the oral ectoderm and transcription factors that control pituitary cell differentiation and specification. Several homeodomain transcription factors directing embryologic development of the anterior pituitary have been found to have mutations that result in congenital defects affecting the synthesis of GH and additional pituitary hormones.13 The human mutations that cause isolated GH deficiency or multiple pituitary hormone deficiency and associated features are summarized in Table 1.
The homeobox gene (HESX1) expressed in embryonic stem cells is important in the development of the optic nerve, as well as the anterior pituitary. HESX1 inhibits PROphet of Pit-1 (PROP1)-mediated gene effects (see later) and mediates forebrain development.14 HESX1 has also been referred to as the Rpx or Rathke pouch homeobox gene. Some mutations that have been described account for a small subset of the cases of septo-optic dysplasia with variable GH and other pituitary deficiencies.15
PITX2 is a paired-like homeobox gene expressed in the fetal pituitary and adult gland, thought to be required for pituitary development shortly after formation of the committed Rathke pouch. There have been at least eight mutations in PITX2 resulting in Rieger syndrome, which includes anomalies of the anterior chamber of the eye, dental hypoplasia, protuberant umbilicus, and mental retardation, but it is uncertain whether pituitary hormone deficiencies are associated.16
LHX3 accumulates in the Rathke pouch and the primordium of the pituitary and is thought to be involved in the establishment and maintenance of the differentiated cell types.17 Mutations of this transcription factor result in deficiencies of all pituitary hormones except adrenocorticotropin, and cervical spine rigidity indicates extrapituitary function for this factor in some families.18 LHX3 and LHX4 belong to the LIM family of homeobox genes expressed early in the Rathke pouch, with expression persisting into adulthood. This has suggested a maintenance function for anterior pituitary cells. Four patients in two unrelated families have been identified with LHX3 mutations with a hormonal phenotype similar to PROP1 deficiency, including marked pituitary enlargement in one patient (see later).18 There has been only one report of a mutation within LHX4.19
The Sonic hedgehog signaling pathway, mediated by three GLI genes, has been identified in a variety of tissues and has been implicated in complex disorders of pituitary development. Mutations of GLI2 are associated with holoprosencephaly.20 Penetrance is variable, with all affected patients having pituitary gland dysfunction.
X-linked hypopituitarism results from duplications of Xq26-27, a region that includes the SOX3 gene, for which a polyalanine expansion has been described in a pedigree with X-linked mental retardation and GH deficiency.21-23 Mutations that result in either overdosage or underdosage of SOX3 are associated with infundibular hypoplasia and variable hypopituitarism.24
PROP1 represses HESX1 expression and is required for initial determination of pituitary cell lineages, including gonadotropes and those of Pit1. At least 10 recessive mutations have been described in PROP1 that result in GH, prolactin (PRL), TSH, gonadotropin and, in some families or as the patients age, adrenocorticotropic hormone (ACTH) deficiency.25,26 Patients with PROP1 gene mutations may have pituitary gland enlargement originating from the intermediate lobe.27,28 Eleven recessive and four dominant mutations have been reported affecting the Pit1 gene (currently referred to as POU1F1), with resultant GH, PRL, and TSH deficiency.13,25 POU1F1 gene defects are associated with variable pituitary hypoplasia.29
Somatotroph development is also dependent on hypothalamic GH-releasing hormone (GHRH). Mutation in the gene encoding the GHRH receptor results in severe GH deficiency.30-32
Functional anatomy of the anterior pituitary (adenohypophysis). The adenohypophysis receives hormonal modulating signals from the hypothalamus, transmitted from ventromedial and infundibular nuclei axons, which terminate in the hypophyseal portal system. These signals result in production of corticotropin (ACTH) by 8 weeks' gestation, thyrotropin (TSH) by 15 weeks' gestation, somatotropin (GH) by 10 to 11 weeks' gestation, PRL by 12 weeks, and the gonadotropes LH and follicle-stimulating hormone (FSH), by 11 weeks. There are at least three distinct hormone-producing cell populations classified by staining characteristics.12 Fifty percent of the cells are chromophobes, 40% are characterized as acidophils, and the remainder are basophils. Acidophils secrete GH or PRL. Basophils secrete TSH, LH, FSH, or ACTH. Some basophils have a positive periodic acid-Schiff base reaction: these cells secrete the glycoproteins LH, FSH, or TSH. Although chromophobe cells are known to produce ACTH in the rat pituitary, the role of these cells in the human pituitary remains unclear.
Anterior pituitary hormones enter the portal venous system to drain into the cavernous sinus, enter the general circulation, and ultimately exert long-distance influence over their respective target organs. TSH promotes growth of the thyroid and production of thyroxine. LH and FSH stimulate gonadal maturation and hormonal cycling. GH exerts indirect growth effects through the elaboration of IGF-I in the liver and epiphyses, direct growth effects on chondrocyte proliferation, and direct metabolic effects primarily in adipose tissue.
The blood supply of the pituitary is subject to interruption during periods of severe hypotensive stress and hypoxia, resulting in Sheehan syndrome of hypopituitarism, originally described after intrapartum hypotension but possible in any hypovolemic crisis or increased intracranial pressure episode, as in hypopituitarism following recovery from cerebral edema complicating diabetic ketoacidosis.33 The internal carotid arteries supply the vascular branches that bathe the pituitary. The hypophyseal portal vessels, which originate from capillary beds in the median eminence and infundibular stem, supply the adenohypothesis.34
GH/IGF-I/IGF BINDING PROTEIN AXIS
GH. Human GH is a single-chain, 191-amino acid, 22-kDa protein, containing two intramolecular disulfide bonds.35 Release of GH from the anterior pituitary somatotrophs is controlled by the balance between stimulatory GHRH and inhibitory somatostatin from the hypothalamus. This balance is regulated by neurologic, metabolic, and hormonal influences; numerous neurotransmitters and neuropeptides are involved. These include vasopressin, corticotropin-releasing hormone, thyrotropin-releasing hormone, neuropeptide Y, dopamine, serotonin, histamine, norepinephrine, and acetylcholine, which respond to various circumstances that affect GH secretion such as sleep, nutritional state, stress, and exercise. Other hormones including glucocorticoids, sex steroids, and thyroxin also influence GH secretion. These various influences are important in the evaluation of GH secretion, which may be abnormal despite normal somatotroph function. Stimulation of GH release by GHRH is via specific GHRH receptors. In addition to the GHRH receptor defects noted earlier, four autosomal recessive disorders, an autosomal dominant mutation, and an X-linked form of isolated GH deficiency have been described. Most children with mutation of the GH1 gene resulting in total absence of GH treat injected rhGH as a foreign protein and develop resistance after a few months of treatment due to the inactivating antibodies they develop.
A number of synthetic hexapeptides, referred to as GH-releasing peptides (GHRPs), have been developed that act on other receptors to stimulate GH release.36,37 The naturally occurring ligand for the GHRP receptor, ghrelin, has been isolated and cloned.38 Ghrelin is unique among mammalian peptides in its requirement of a posttranslational modification for activation. This involves addition of a straight-chain octanyl group conferring a hydrophobic property to the N terminus, which may permit entry of the molecule into the brain. Similarly to synthetic GHRPs, ghrelin binds with high affinity and specificity to a distinct G protein-coupled receptor.39 Unlike GHRH, ghrelin is synthesized primarily in the fundus of the stomach,38 as well as in the hypothalamus, heart, lung, and adipose tissue, and its receptor is more widely distributed than that of GHRH.40 Ghrelin has widespread metabolic effects in addition to inducing GHRH release and being synergistic with GHRH in the stimulation of GH release through the serine-3 residue of ghrelin. Ghrelin stimulates release of PRL, ACTH, cortisol, and aldosterone and increases food intake and weight gain.41
Approximately 75% of the circulating GH is in the 22-kDa form. Alternative splicing of codon 2 results in a deletion of 11 amino acids and formation of a 20-kDa fragment accounting for 5% to 10% of secreted GH. Other circulating forms include deamidated, N-acetylated, and oligomeric GH. About 50% of GH circulates in the free state, and the rest is bound principally to GH-binding protein (GHBP). Because the binding sites for the radioimmunoassay of GH are not affected by the GHBP, both bound and unbound GH are measured.42
GHBP and the GH receptor. A high-affinity GHBP was identified in rabbit and human serum in the mid-1980s,43 and separate reports in 1987 found this binding protein to be absent in the sera of patients with GH resistance,44,45 who were identified by high circulating GH concentration with a phenotype of severe GH deficiency. The recognition that circulating GHBP in rabbit serum corresponded to liver cytosolic GHBP was followed by the purification, cloning, and sequencing of human GHBP.46 The human GHBP was found to be structurally identical to the extracellular hormone-binding domain of the membrane-bound GH receptor (GHR). The entire human GHR gene on chromosome 5 was subsequently characterized.47 The GHR was the first to be cloned of a family of receptors that includes the receptor for PRL and numerous cytokine receptors. Members of this family share ligand and receptor structure similarities, in particular, the requirement that the ligand bind to two or more receptors or receptor subunits and interact with signal-transducer proteins to activate tyrosine kinases.48
In humans, GHBP is the proteolytic product of the extracellular domain of the GHR. This characteristic permits assaying circulating GHBP as a measure of cellular-bound GHR, which usually correlates with GHR function. The GH molecule binds to cell surface GHR, which dimerizes with another GHR so that a single GH molecule is enveloped by two GHR molecules.49 The intact receptor lacks tyrosine kinase activity, but it is closely associated with JAK2, a member of the Janus kinase family. JAK2 is activated by binding of GH with the GHR dimer, which results in self- phosphorylation of the JAK2 and a cascade of phosphorylation of cellular proteins. Included in this cascade are signal transducers and activators of transcription (STATs), which couple ligand binding to the activation of gene expression and mitogen-activated protein kinases. Other effector proteins have also been examined in various systems. This is a mechanism typical of the growth hormone/PRL/cytokine receptor family.48,50 In human GH-GHR transduction, STAT5b appears to be the most important cellular protein activated. Five distinct homozygous mutations have been described in seven patients from six families, resulting in severe growth failure and variable immune incompetence, indicating the importance of STAT5b in cytokine function, as well as its primacy in GH-GHR transduction.51
The GH receptor in humans is also synthesized in a truncated form (GHRtr) lacking most of the intracellular domain. Although the quantity of this GHRtr is small relative to the full-length GHR, release of GHBP from this isoform is increased.52 Some of the changes in body composition that occur with GH treatment in GH deficiency may be related to changes in the relative expression of GHR and GHRtr.53
More than 50 mutations in the GHR have been described in the approximately 250 known patients with GH insensitivity, which result in a clinical picture identical to that of severe GH deficiency but with elevated serum GH concentrations.2,51 The report of the characterization of the GHR gene included the first description of a genetic defect of the GHR, a deletion of exons 3, 5, and 6;47 recognition that the exon 3 deletion represented an alternatively spliced variant without functional significance resolved the dilemma of explaining deletion of nonconsecutive exons. In contrast to the alternatively spliced variant lacking exon 3, the first mutation of this exon has been described in a typical GHR-deficient patient with heterozygosity for a nonsense mutation in exon 4, and family studies indicate that heterozygosity for the exon 3 mutant has no effect. This study also raises questions of the origin and function of the exon 3-deleted variant. More recently, this isoform, present in either the homozygous or heterozygous state, was found to be associated with 1.7 to 2 times more growth acceleration from GH administration during 2 years of treatment of children with short stature who had been small for gestational age or had idiopathic short stature. In addition to the original exon 5, 6 deletion, another deletion of exon 5 has been described, along with numerous nonsense mutations, missense mutations, frame shift mutations, splice mutations, and a unique intronic mutation resulting in insertion of a pseudoexon. A number of other mutations have been described that are either polymorphisms or have not occurred in the homozygous or compound heterozygous state.54
The point mutations that result in severe GH insensitivity when present in the homozygous state or as a compound heterozygote are all associated with the typical phenotype of severe GH deficiency. All but a few of the defects result in absent or extremely low levels of GHBP. Noteworthy is the D152H missense mutation that affects the dimerization site, thus permitting the production of the extracellular domain in normal quantities but failure of dimerization at the cell surface, which is necessary for signal transduction and IGF-I production. Two defects that are close to (G223G) or within (R274T) the transmembrane domain result in extremely high levels of GHBP. These defects interfere with the normal splicing of exon 8, which encodes the transmembrane domain, with the mature GHR transcript being translated into a truncated protein that retains GH- binding activity but cannot be anchored to the cell surface.
As noted, all these homozygous defects and the compound heterozygotes, whether involving the extracellular domain or the transmembrane domain and whether associated with very low or unmeasurable GHBP, result in a typical phenotype of severe GH deficiency. In contrast, the intronic mutation present in the heterozygous state in a mother and daughter with relatively mild growth failure (both with standard deviation score [SDS] for height -3.6), and resulting in a dominant negative effect on GHR formation, is not associated with other phenotypic features of GH deficiency. This splice mutation preceding exon 9 results in an extensively attenuated, virtually absent intracellular domain. Japanese siblings and their mother have a similar heterozygous point mutation of the donor splice site in intron 9, also resulting in mild growth failure compared to GHRD but with definite, although mild, phenotypic features of GHD. GHBP levels in the Caucasian patients were at the upper limit of normal with a radiolabeled GH binding assay and in the Japanese patients twice the upper limit of normal, using a ligand immunofunction assay. These heterozygous GHR mutants transfected into permanent cell lines have demonstrated increased affinity for GH compared to the wild-type full-length GHR, with markedly increased production of GHBP. When cotransfected with full-length GHR, a dominant negative effect results from overexpression of the mutant GHR and inhibition of GH-induced tyrosine phosphorylation and transcription activation. Naturally occurring truncated isoforms have also shown this dominant negative effect in vitro.54
A novel intronic point mutation was discovered in a highly consanguineous family with two pairs of affected cousins with GHBP-positive GHI, severe short stature but without the facial features of severe GH deficiency or insensitivity. This mutation resulted in a 108-bp insertion of a pseudoexon between exons 6 and 7, predicting an in-frame, 36-residue amino acid sequence in a region critically involved in receptor dimerization.55
Of approximately 250 reported cases of typical GHR deficiency, ethnic origin is predominantly Middle Eastern, Mediterranean, and South Asian. Nearly 50% are Oriental Jews as described in the original report, or known descendants of Iberian Jews who converted to Catholicism during the Spanish Inquisition. The latter group comprises the largest cohort (n >70) and the only genetically homogenous group. All but two subjects are homozygous for the E180 splice site mutation, which was also found in one Israeli patient of Moroccan heritage and recently in seven affected children from six families not previously known to be related, from northeast Brazil56 and two siblings from Chile (F. Cassorla, personal communication, May 2008). Most of the other defects appear to be highly family specific, with three nonsense mutations (R43X, C38X, R217X) and the intron 4 splice mutation the only ones thus far described that appear in disparate populations on different genetic backgrounds, indicating mutational hotspots.54
IGF-I. Most of the growth effect that gives GH its name is actually an effect of IGF-I production.57,58 IGF-I is a 70-residue single-chain basic peptide, and IGF-II is a slightly acidic 67-residue peptide. Their structure is similar to proinsulin, A and B chains connected by disulfide bonds and a connecting C-peptide, but unlike the posttranslational processing of insulin, there is no cleavage of the C-peptide. The two IGFs share approximately two thirds of their possible amino acid positions and are 50% homologous to insulin.59,60 The connecting C-peptide is 12 amino acids long in the IGF-I molecule, eight amino acids long in IGF-II, and has no homology with the comparable region in the proinsulin molecule. The IGFs also differ from proinsulin in having carboxy terminal extensions. These similarities and differences from insulin explain the ability of IGFs to bind to the insulin receptor and insulin's ability to bind to the type 1 IGF receptor, as well as the specificity of IGF binding to the IGF-binding proteins (IGFBPs).
IGFBPs. Hepatic IGF-I circulates almost entirely bound to IGFBPs (<1% are free). The IGFBPs are a family of six structurally related proteins with a high affinity for binding IGF. At least four other related proteins with lower affinity for IGF peptides have been identified and are referred to as IGFBP-related proteins.61 The principal BP, IGFBP-3, binds approximately 90% of circulating IGF-I in a large (150- to 200-kD) ternary complex consisting of IGFBP-3, an acid labile subunit (ALS), and the IGF molecule. ALS and IGFBP-3 are produced in the liver as a direct effect of GH. The ALS stabilizes the IGF–IGFBP3 complex, reduces the passage of IGF-I to the extravascular compartment, and extends its half-life.62 The remainder of bound IGF is in a 50-kD complex with mostly IGFBP-1 and IGFBP-2. IGFBP-1 concentrations are controlled by nutritional status as reflected in insulin levels, with the highest IGFBP-1 concentrations found in the fasting, hypoinsulinemic state. The circulating concentration of IGFBP-2 is less fluctuant and is partly under the control of IGF-I; levels are increased in IGF-I deficiency due to GH insensitivity but increase further with IGF-I therapy of such patients.63
The IGFBPs modulate IGF action by controlling storage and release of IGF-I in the circulation and influencing its binding to its receptor, facilitate storage of IGFs in extracellular matrices, and exert independent actions. IGFBPs 1, 2, 4, and 6 inhibit IGF action by preventing binding of IGF-I with its specific receptor. The binding of IGFBP-3 to cell surfaces is thought to decrease its affinity, effectively delivering the IGF-I to the type 1 IGF receptor. IGFBP-5 potentiates the effects of IGF-I in a variety of cells. Its binding to extracellular matrix proteins allows fixation of IGFs and enhances binding to hydroxyapatite. IGFs stored in such a manner in soft tissue may enhance wound healing. IGF-independent mechanisms for IGFBP-1 and IGFBP-3 proliferative effects have been demonstrated in vitro, and nuclear localization of IGFBP-3 has been reported. In addition to IGFBP phosphorylation and cell surface association determining the influence of IGFBPs, specific protease activity, particularly affecting IGFBP-3, is also important in the modulation of IGF action in target tissues. The proteolytic activity may alter the affinity of the binding protein for IGF-I, resulting in release of free IGF-I for binding to the IGF-I receptor.64
IGF receptors. IGF binding involves three types of receptors: the structurally homologous insulin receptor, type 1 IGF receptor, and the distinctive type 2 IGF-II/mannose-6-phosphate receptor. Splice variants and atypical forms occur but have not been found to have physiologic significance. Insulin/IGF-I hybrid receptors, however, are ubiquitous and may be the most important receptor for IGF-I in some tissues.64
The type 1 IGF-I receptor and insulin receptor are heterotetramers consisting of two alpha subunits that contain the binding sites and two beta subunits containing a transmembrane domain, an ATP binding site, and a tyrosine kinase domain comprising the signal transduction system.64 The IGF-I receptor is able to bind IGF-I and IGF-II with high affinity, but the affinity for insulin is approximately 100-fold less. Although the insulin receptor has a low affinity for IGF-I, IGF-I is present in the circulation at molar concentrations that are 1,000 times those of insulin. Thus, even a small insulin-like effect of IGF-I could be more important than that of insulin itself, if it were not for the IGFBPs that control the availability and activity of IGF-I. In fact, intravenous infusion of rhIGF-I can induce hypoglycemia, especially in the IGFBP-3-deficient state.63 It is not known why IGF-II and M6P share a receptor. This receptor differs from the type 1 receptor in binding only IGF-II with high affinity, IGF-I with low affinity, and insulin not at all.64
The role of the GH/IGF-I axis in growth. The growth effect of GH has at least three components, and their relative contributions are a subject of continuing investigation. The most familiar of these components are IGF-I, IGFBP-3, and acid labile subunit (ALS), because they are synthesized by the liver and secreted into the circulation, allowing them to be measured as circulating concentrations. The other GH effects are not directly measurable but inferred from much animal and some human data; they are epiphyseal prechondrocyte differentiation and enhancement of local (autocrine/paracrine) production of IGF-I, thereby stimulating clonal expansion of the differentiating chondrocytes.2,57,58
The importance of IGF-I in normal intrauterine growth in humans has been demonstrated in a single patient with a homozygous partial deletion of the IGF-I gene, a patient with mutation of the IGF-I gene resulting in high circulating levels of an ineffective IGF-I, and in two patients with mutations of the IGF-I receptor, all having severe intrauterine growth retardation.2 Cord serum IGF-I and IGF-II concentrations correlate with birthweight and are significantly increased in large-for-gestational-age infants compared with appropriate-for-gestational-age newborns.65
Intrauterine IGF-I synthesis, however, does not appear to be GH dependent, because most patients with genetically determined severe IGF-I deficiency, due to GHRH defects, GH receptor deficiency (GHRD), or GH gene mutations, have normal or only minimally reduced intrauterine growth. SDS for length declines rapidly after birth, however, in these conditions, demonstrating the immediate need for GH-stimulated IGF-I synthesis for postnatal growth.2 Growth velocity in the absence of GH is approximately half normal, but it has sometimes been reported to be normal or supranormal.66 This apparent growth without GH has been described in patients after craniopharyngioma resection, in septo-optic dysplasia, in obese children with GH deficiency, in GH-deficient infants, and in patients who had undergone resection of a variety of central nervous system tumors.67 Normal or supranormal growth velocity has been attributed to hyperinsulinemia, increased leptin levels, or hyperprolactinemia. PRL levels are not consistently increased, however. Obesity or rapid weight gain is a frequent common denominator among these patients who demonstrate low GH levels to provocative stimuli, low IGF-1 and IGFBP-1, and low IGFBP-3 levels.
The metabolic and growth effects of GH and IGF-I are compared in Table 2. In addition to direct protein-sparing effects and synthesis and release of IGF-I from the liver, GH stimulates autocrine and paracrine production of IGF-I in other tissues, primarily bone and muscle. GH has a direct effect on differentiation of prechondrocytes into early chondrocytes, which in turn secrete IGF-I. This local IGF-I stimulates clonal expansion and maturation of the chondrocytes, resulting in growth.58 It is estimated that 20% of normal growth (ie, 40% of GH-stimulated growth) is the result of the direct effect of GH on maturing bone and the autocrine/paracrine production of IGF-I in this tissue. Treatment studies of children with GHRD compared to GHD patients support this hypothesis.2 IGF-II is considered an important growth factor in utero, but its role in extrauterine life is unclear; concentrations of IGF-II in serum parallel those of IGF-I.
The direct stimulation by GH of mitosis in cartilage precursor cells of the growth plate, which have GH receptors, and the stimulation of local production of IGF-I led to the hypothesis that autocrine/paracrine IGF-I was the main determinant of GH-dependent postnatal body growth and that hepatic or endocrine IGF-I served predominantly as a negative feedback regulator of GH secretion.58 Subsequent studies of mice with selective deletion of the hepatic IGF-I gene described unaffected growth.2 Deletion of the ALS gene in mice and its mutation in man results in very low circulating IGF-I and IGFBP-3 concentrations but only a 15% reduction in postnatal growth.62 It is uncertain if there was any growth effect in the first two patients reported with ALS mutations, one of whom reached a stature of -0.9 SDS and the other a height that was 0.4 SDS greater than mean parental height.2 Subsequent reports of ALS mutations in six patients have described more effect on growth from 0.8 SDS to 2.0 SDS less than target height.68,69
CONCLUSION
The understanding of the physiology of growth has progressed from auxologic observations, description of dysmorphic syndromes, and inferred hormonal dysfunction with growth failure or, far less commonly, excessive growth, to reasonably accurate measurement of hormones influencing growth, identification of many more growth factors, understanding of the control of bone growth, and definition of the molecular basis for normal and abnormal growth states. These developments span only a half century. With the rapidly accelerating capabilities for hormonal and molecular study, the complexity of the genetic, hormonal, and environmental factors and their interaction in the fundamental human process of growth will continue to be unraveled.
Arlan L. Rosenbloom, MD, is Adjunct Distinguished Service Professor Emeritus in the Division of Endocrinology, Department of Pediatrics, University of Florida College of Medicine. He may be reached at Rosenal@peds.ufl.edu; or phone: 352-334-1393.
|
