Cobbold Laboratories, The Middlesex Hospital, University College, London introduction
Postnatal growth can be considered to consist of at least three distinct phases: infancy, childhood, and puberty. The infancy component is largely a continuation of the longitudinal growth process observed in utero. This displays a peak growth velocity around 27-28 weeks of gestation with a decline in growth rate during the last trimester of pregnancy. Birth, in a sense, is incidental to this declining growth rate, which continues during the first 3 years of life, reaching a plateau at or around the fourth year of life and remaining at this level until the commencement of the pubertal growth spurt. This plateau is interrupted in a large number of children by a "juvenile" or "mid-growth" spurt of small magnitude, which occurs between 6 and 8 years of age. The factors influencing these distinct growth periods are different. We know little of the factors influencing fetal and early infant growth but know from animal experiments that nutrition plays a key role. With the appearance of the growth hormone (GH) receptor in the growth plate at around 6 months of postnatal life, the GH-dependent growth assumes greater importance and during the childhood years, growth is largely dependent on the GH secretory status of the individual. The final step in the growth process, the pubertal growth spurt, comprises a 50% contribution from sex steroids and 50% contribution from GH.
This chapter initially outlines our understanding of factors involved in regulating fetal growth with particular emphasis on the insulin-like growth factor (IGF) axis. This allows us to compare and contrast the role of the IGF axis in the regulation of fetal and postnatal growth. In the understanding of the endocrinology of growth, it is important to realize that, during postnatal growth, a large number of factors influence the growth process, but the majority of these operate through modulation of the GH axis. This is not to say that all growth failure in childhood is due to GH deficiency but rather that GH acts as a final common pathway for the integration of all these signals. For example, patients with celiac disease grow poorly due to malabsorption, but in addition their GH response to a number of provocative stimuli is blunted. They are not, per se, GH insufficient, as the GH secretion returns to normal once the underlying abnormality in the gastrointestinal tract is rectified.
Any consideration of the growth process, particularly in the postnatal period, requires an understanding of the physiology of GH secretion. GH secretion takes place in a pulsatile manner and it is important to understand the part played by this pattern in the generation of growth in the human.
endocrinology of growth GH Secretion—Cellular and Molecular
The pituitary gland develops as an outpouching of the stomatodeum—Rathke's pouch. The process takes place between 30 and 35 days postconception and is tightly regulated by a series of homeobox genes. Close apposition between this structure, which is destined to form the anterior pituitary, and the base of the hypothalamus takes place; and this leads to descent of neural tissue with the pouch to form the posterior pituitary. Stalk vascular cannulization completes the process. Differentiation of the anterior pituitary mass into the recognizable cell types is influenced in part by the homeobox genes involved in development and cell-specific homeo-box gene expression. Somatotroph cell differentiation utilizes in addition the expression of two genes: Prop-1 and Pit-1. Enlargement of the somatotroph cell number requires the induction of the growth hormone releasing hormone (GHRH) receptor by Pit-1. This allows the hypothalamic peptide to stimulate somatotroph cells, leading to synthesis and release of GH. In addition, GHRH stimulation also leads to somatotrophic hyperplasia.
The human GH gene is located on chromosome 17, along with two genes for human somatomammotropin (Figure 5-1). Pituitary GH is coded for by the hGHN gene and transcription leads to the synthesis of GH with a molecular weight of 22 KD. The excision of the second intron of hGHN leads to an alternative splicing site, resulting in deletion of the message for amino acid residues 32-46, the 20 KD GH variant. This forms 10% of the circulating GH.
The synthesis of GH is regulated largely by the levels of GHRH impinging on the anterior pituitary somatotropes. GHRH acts on the somatotrope by binding to its own specific receptor, which activates a secondary messenger system via cyclic AMP synthesis. This receptor is characterized by seven transmembrane loops and internal coupling to the G(guanine)-protein system (Figure 5-2). In their resting state, the G-proteins exist as heterodimeric complexes with a, p, and y subunits. In practice, the p and y subunits associate with such a high affinity that the functional
units are Ga and G^. After association of the G-protein complex with the occupied receptor, conformational changes in the a subunit lead to an increased rate of dissociation of GDP, which is replaced by GTP. This guanine nucleotide exchange in turn causes the a subunit to dissociate from the heterotrimeric complex. The liberated a subunit, together with its activating GTP, then binds to a downstream catalytic unit adenylate cyclase.
Hydrolysis of the GTP bound to Ga due to its intrinsic GTPase activity liberates the Ga subunit from the catalytic subunit and allows reassociation of Ga GDP with the Gpr This newly reformed heterotrimer then returns to the G-protein pool in the membrane. In this way, an individual G-protein complex is recycled, so that it can respond to further receptor occupation by ligand.
Gsa activates membrane bound adenylyl cyclase, which catalyses the conversion of ATP to the potent second messenger cAMP. This cyclic nucleotide in turn activates a cAMP-dependent protein kinase (PKA), which modulates multiple aspects of cell function. PKA phosphorylates a transcription factor called CREB (cAMP response element binding protein). This is then translocated to the nucleus, where it binds to a short palindromic sequence in the promoter region of the GH genes, the process that leads to transcription and synthesis of GH. The transcription of the GH gene is regulated in turn by a number of other hormones such as thyroxine and cortisol.
Prior to consideration of the endocrine regulation of different stages of human growth, it is worth considering the physiology of the GH-insulin-like growth factor axis. Figure 5-3 shows this particular pathway as a classic closed loop feedback system. Figure 5-4 shows a typical GH profile in a 9-year-old boy generated by taking blood samples for GH measurement every 20 minutes.
Frequent sampling is essential to define clearly the true heights of the peaks. If sampling were too infrequent, the true peak heights might be underestimated or peaks missed altogether. The profile is characterized by episodes of GH release, generating peak GH concentrations, interspersed with periods when GH secretion is effectively switched off and GH concentrations are undetectable. This appears to be the predominant pattern in males, whereas in females in varying species, although the peak concentrations tend to be similar, the most striking difference is that there is an elevation in the trough concentrations, so that at all times concentrations are detectable. One further point to note is that the pulses occur at fairly frequent intervals of one every 3 hours, suggesting that most of the GH signal is contained in the amplitude of the pulses.
In both rodents and humans, evidence supports the concept of an inverse relationship between the secretion of the two hypothalamic peptides: somatostatin (SS) and GH releasing hormone. GHRH is involved in both the release and synthesis of GH while SS inhibits GH release. Normal GH pulsatility requires endogenous GHRH, although GH responses to exogenous GHRH are variable and reveal varying periods of responsiveness and refractoriness. There are several possible explanations for this. First, the phenomenon may be intrinsic to the GH-secreting cells. Second, acute downregulation of the GHRH receptors or their intracellular signaling systems may take place. This appears to be an unlikely explanation as down-regulation takes place only at very high GHRH levels, certainly well above those usually encountered physiologically. Third, it could be by depletion of the read-
ily releasable pool of GH. The final and most likely explanation is that the pattern reflects variation in endogenous SS tone imposing an ultradian rhythm in GHRH responses. Evidence for this comes from the observation that continuous GHRH administration leads to pulsatile GH release, implying modulation by another factor, somatostatin. Although SS readily suppresses pulsatile GH secretion in rats and humans, its effects are short-lived and rapid release of GH takes place on SS removal. This rebound secretion can be detected in vitro but is even more pronounced in vivo. In general, GHRH administration alone leads to the gradual attenuation of the GH response with time. SS withdrawal can produce GH rebound secretion in the human subject, but for regular repeatable GH release to take place, the combination of SS withdrawal coupled with GHRH administration is the most efficacious.
09 12 15 18 21 24 03 06 09 Clock time (hours)
figure 5-4 Twenty-four-hour serum GH concentration profile.
The close relationship between GHRH and SS acting as integrators of other signals, such as sleep, is complicated by the recent discovery of a family of GH secretagogues, which differ substantially from GHRH in both their structure and receptors. These GH secretagogues behave in a similar manner in terms of their physiology as GHRH, but the important difference is that they act synergistically with GHRH to generate GH release (Figure 5-3). Recent work has identified the endogenous "third factor" as Ghrelin, which is present in the stomach. The precise physiology of this substance remains to be determined. A number of endogenous agents affect GH pulsatility, including opioids, calcitonin, and glucagon, but these have not been used to manipulate GH secretory patterns to alter growth and their physiological relevance to the control of endogenous pulsatility remains unclear.
There is good evidence that GH feeds back to inhibit its own release, and this may have a bearing on the temporal control of GH pulsatility. Experiments indicate that exogenous GH acts directly on the hypothalamus rather than on the pituitary. The most likely mode of action of GH is through increased secretion of SS into the portal blood, but there is evidence that GH also leads to an inhibition of GHRH production. In addition, GH can also feed back indirectly by the generation of IGF-1, which in turn inhibits GH synthesis and release chiefly at the pituitary level.
The secretion of GH into the circulation is pulsatile. The entry rate of endogenous GH is governed by the kinetics of GH release from the somatotropes and the removal from the circulation is largely determined by the amount of GH bound to its binding protein and internalized by the GH receptor on the target organ cells. The precise role of the binding proteins in humans, at least, is far from clear. There is correlation with GH status but only at the very extremes, and the effect of GH treatment is highly variable. Although GH binding protein might increase the amount of GH available for constant delivery of GH to the receptor, this is not at all clear. There is no evidence to suggest that the preferred mode of presentation of GH to the receptor is continuous. A body of evidence suggests that the pulsatile mode is most optimal. An alternative role for the binding proteins might be to buffer the system from overexposure to GH. Given the high affinity of GH for its binding protein this might be a more likely explanation.
The pulsatile signal appears to be important in determining a number of target organ effects. For example, in rodents, the pattern of hormone secretion, either pulsatile (male mode) or continuous (female mode), has an impact on the growth of the animal, expression of a number of liver enzymes, the determination of the level of GH binding protein in the circulation, as well as GH receptor expression. The tissue response is also variable in that the liver generates IGF-1 in response to GH irrespective of the mode of administration, whereas adequate expression of IGF-1 in the muscle is highly dependent on the pulsatile mode of administration.
Increasing evidence in humans suggests that the mode of GH secretion is important in determining target organ response. In humans, IGF-1 generation occurs best when GH is present in the pulsatile rather than the continuous mode at least in the physiological situation. This may not be the case when GH treatment uses the subcutaneous route, as the pharmacokinetics of subcutaneous GH tend toward a more continuous exposure. On the other hand, the pattern of fat distribution around the abdomen is influenced more by the trough concentrations of GH in humans.
The GH receptor together with those for the cytokines, such as the interleukins and erythropoietin, share a common major structural feature in that they have four long alpha-helixes arranged in an antiparallel fashion. As a consequence, this subgroup is commonly referred to as the cytokine/hemopoietic receptors. The structure of human GH with its receptor is a ternary complex consisting of a single molecule of the hormone and two receptors. After GH has bound to one molecule of receptor, this is followed by association of this complex with a second receptor molecule. The dimerization of the cytoplasmic region in the ternary complex is particularly important for signal transduction.
The GH receptor uses an unusual intracellular signaling system: Janus-associated kinase-2 (JAK-2). The JAK system is coupled to further intracellular proteins, the so-called STAT proteins (Figure 5-5). These are transcription factor proteins. They contain a crucial tyrosine residue located in the carboxy-terminal in a homologous position in all STAT proteins (residue 694), and phosphorylation of this is essential for STAT activation. STAT proteins have dual functions: signal transduction in the cytoplasm followed by activation of transcription in the nucleus. The family members of STAT proteins have been named in the order of their identification. GH induces tyrosine phosphorylation of STAT proteins 1, 3, 5a, and 5b, but STAT 5 is probably the major axis of the JAK-STAT cascade. Dimerization of the STAT proteins appears to be essential for their final translocation to the nucleus, where they activate immediate early response genes that regulate proliferation or more specific genes that determine the differentiation status of the target cell.
gh and igf-1 axis and its association with fetal growth
Given the paradigm depicted in Figure 5-1, it is interesting to contrast the situation in the fetus with that in the child and the adolescent. The fetus is subject, particularly in the latter half of pregnancy, to a fairly constant delivery of metabolites via the placenta. The flow of metabolites continues through the umbilical vein, and the distribution thereafter utilizes a circulatory pattern in which blood is predominantly diverted in its oxygenated form to the developing brain. Given that the fetus is highly dependent on this mode of delivery of substrate, perhaps it is not too surprising that the growth process differs. In addition, the growth of the individual organs also differs from that observed in postnatal life with different patterns of growth and development exhibited by many of the differing tissues.
Studies in larger domestic animals show that pulsatile secretion of GH and other pituitary hormones is already demonstrable in fetal life and is sensitive to nutrition. Little is known about the evolution of GH secretion in the human fetus. Stud-
transcription is enhanced figure 5-5 Dimeric receptor (GH and receptor) signaling through JAK kinases.
transcription is enhanced figure 5-5 Dimeric receptor (GH and receptor) signaling through JAK kinases.
ies have demonstrated a gradual increase in circulating GH concentration during the first 12 weeks of pregnancy, reaching a peak at 20-24 weeks and declining toward birth. These early changes to serum GH concentrations appear to parallel the known development of the hypothalamic peptides GHRH and SS. The human fetal pituitary is able to respond to these two factors, and it is proposed that the GHRH effect predominates, with SS increasing in effect toward term. Even at term, GH levels are 20-30 times higher than those observed in childhood, but perhaps of greater importance is that the levels are continuously elevated and lack the pulsatile pattern observed in childhood and adult life. However, these high GH concentrations are not associated with elevated levels of IGF-1 in the fetus, implying that there is "relative resistance" to the effects of GH in the fetus. The effect may diminish toward term but it can be assumed that GH is not the predominant determinant of fetal growth. This is also borne out by experiments of nature in which the GH gene is deleted or where the GH receptor is nonfunctional. These individuals are normal size at birth when due account is taken of maternal size.
Because of the problems associated with accessibility to human fetal tissue, the role of endocrine factors in determining fetal growth is largely inferred from studies in animals. The most elegant series of these studies involves the use of transgenic animals in which the various components of the IGF axis (Table 5-1) have been knocked out. It must be appreciated that these studies reveal quite major effects of the whole gene and tell us that the peptide is of particular importance in the determination of size of the fetus. That they clearly are important comes from the observation that many of the knockout offspring die in the first few hours of life. Table 5-1 shows the type of knockouts that have been constructed, and from this, both IGF-1 and IGF-2 can be clearly seen to play important roles in the determination of body size in the mouse. It is likely that a similar situation pertains in the human, because there is a clear relationship between birth weight and levels of both these growth peptides; in addition, a boy with IGF-1 gene defect was born with low birth weight. Perhaps rather surprisingly, loss of the insulin gene did not appear to alter body size. This would, at first sight, appear contradictory to clinical observations of macrosomia associated with maternal hyperglycemia and the condition of hyperinsulinemic hypoglycemia of infancy, where excess fetal and neonatal insulin production leads to fetal overgrowth. It is likely that in these situations the effects of hyperinsulinemia in the fetus are mediated via the IGF receptors rather than a direct effect of insulin via its own receptor. The IGF receptor knockout studies indicate the importance of the Type 1 IGF receptor in mediating the growth effects of IGF-1 and IGF-2. All these studies demonstrate a pivotal role for the IGF family in the determination of fetal growth.
In the newborn, studies have revealed markedly amplified GH secretory episodes that occur throughout the day and night. Preterm infants have even higher secretory profiles than term babies. The high GH secretion at birth is sensitive to inhibition by dopamine and by stimulation by intravenous GHRH. The GH response to GHRH is in turn modified by the birth size of the baby, with greater responses seen in those of lower birth weight. As IGF-1 levels are lower in these babies, this
Human Growth and Development table 5-1 Results of IGF and Insulin Knockout Mice Studies
Insulin receptor Type l IGF receptor +IGF-l +IGF-2
might imply that the feedback effect of IGF-1 is also operative at this age. However, as IGF-1 levels are generally lower at birth and increase thereafter through childhood into adolescence, it is possible that elevated GH values may represent, in part, "immaturity" in this part of the feedback loop. Although the GH response to SS is blunted, the components for generating episodic secretion are clearly present and become more operative during the neonatal period.
endocrinology of prepubertal growth
After 2-3 months of life, clinical evidence suggests that GH is necessary for sustaining normal growth. The postnatal elevated GH levels observed subside, so that by 3-6 months of age, values approach those observed in childhood. During the prepubertal years, GH secretion gradually increases, primarily in terms of the amplitude of the GH pulses.
Although many reports demonstrate differences in GH secretion between tall normal and short stature children and between those with normal stature and short stature, the differences pertain more to the growth rates of the individuals than the stature observed (Figure 5-6). These differences in GH secretion are reflected in the serum levels of IGF-1 seen in these groups. IGF-1 values increase gradually from birth throughout childhood and relate well to the levels of GH secreted. Apart from pathological conditions, it remains difficult to relate any measure of GH pul-satility to the observed growth rate in individual children with short stature, probably because the variability in growth rate is so small.
The situation is complicated further by the interaction of body composition with GH secretion. Particularly in the area of short stature that is a heterogeneous condition, we can imagine a number of diagnoses impinging on the growth process that also influence the relationship between growth and GH. Generally speaking, height at its extremes can be related to GH secretory status. In situations where body mass index is controlled for, the more important relationship is that between GH secretion and growth rate. The relationship has been documented by several groups and is probably described as a curvilinear relationship (Figure 5-7).
Was this article helpful?