Insulin-Like Growth Factors and Skeletal Growth: Possibilities for Therapeutic Interventions

In this, her last issue as editor-in-chief of JCEM, Dr. Maria New has invited some of her colleagues in pediatric endocrinology to contribute short commentaries on issues of current interest. It is a pleasure to honor Dr. New on this occasion, both for her leadership in taking our journal to a new level of excellence and for her many scientific contributions that have enhanced our understanding and treatment of adrenal disorders. The central focus of pediatric endocrinology has been the hormonal control of growth and development, and in recent decades it has been possible to correlate the serum concentrations of nearly every known hormone with each phase of normal growth and with most types of aberrant growth. The insulin-like growth factors (IGFs) or somatomedins, as they are frequently called, have been an important focus of pediatric endocrinologists because of their pivotal role in skeletal growth. The present essay addresses the roles of these peptides in the growth of the skeleton and considers how we might take advantage of their many actions for therapeutic purposes. The two known somatomedins, IGF-I and IGF-II, are structurally homologous to proinsulin. They were initially identified on the basis of three unique properties: their mediation of the skeletal growth-promoting actions of GH, their mitogenic properties, and their mimicry of the actions of insulin. These peptides were isolated in Zurich by Rinderknecht and Humbel on the basis of their insulin-like activity, but were renamed IGF-I and IGF-II when it became apparent that their growth-promoting properties were more important than their insulin-like activities (1). Somatomedin C was isolated in Chapel Hill from human blood fractions by Van Wyk et al. guided by assays based on the stimulation of proteoglycan and DNA synthesis in cartilage as suggested by William Daughaday. Somatomedin C was found on sequence analysis to be identical to IGF-I as reported by the Zurich group (2). As no peptide comparable to IGF-II was isolated under the somatomedin rubric, it is now conventional to use IGF-I and IGF-II when referring to the specific peptides and the term somatomedins or IGFs when referring to these peptides generically (3). The IGFs are more primitive hormones than insulin and have a much broader spectrum of action. Insulin presumably evolved from the IGFs to fulfill a more specialized function when the need arose to store energy during periods of fasting. The ancestral relationship of insulin to IGFs to insulin is similar to that of PTH to PTH-related protein (PTHrP), with PTHrP having a more ubiquitous spectrum of activities than PTH. Both of the somatomedins are developmentally regulated and subject to precise tissue-specific expression. Methods for the quantitation of IGF-I were developed before comparable methods became available for IGF-II (4, 5), and to this day much less is known about the physiological roles of IGF-II than those of IGF-I. Blood levels of IGF-II are considerably higher than those of IGF-I, and the concentrations of IGF-II in extracts of skeletal tissue are higher than those of IGF-I. Blood levels of IGF-I are more stringently regulated by GH and nutritional factors than are those of IGF-II, but IGF-II is more insulin like than IGF-I. The idea has been perpetuated that IGF-II is the fetal hormone and IGF-I is the somatomedin of postnatal life. It is now clear, however, that both IGF-I and IGF-II are important for normal fetal growth and that both have distinctive functions postnatally. Most of the IGF-I circulating in blood comes from the liver, where the expression of the IGF-I gene is regulated by GH (6). The IGFs are also produced in peripheral tissues, where the expression of their genes is regulated by many hormones. IGFs are required for the proliferation of most cell types, and they promote cell survival by inhibiting programmed cell death (7). They also regulate a vast number of highly differentiated cell functions. The nature of the responses to IGFs is dependent on physiological circumstance. It is pertinent to the role of IGFs in skeletal growth that sparsely plated chondrocytes respond to IGF-I with increased DNA synthesis, but not with increased sulfate incorporation into proteoglycans. On the other hand, when the same cells reach confluence, IGF-I stimulates matrix synthesis but has no effect on DNA synthesis. Such observations imply that IGFs may act as both a mitogen in the proliferative zone and a stimulant of proteoglycan synthesis in the hypertropic zones of the growth plate, with the specific effect being dependant on the cellular and hormonal environments. The biological importance of the somatomedins is often underestimated because their effects, compared with those of other growth factors, are often weak or difficult to demonstrate. This is because the IGFs characteristically exercise a permissive or modulating role on biological processes rather than serving as the primary agonist. This was illustrated in the interactions of IGF-I with pituitary hormones in the gonads and thyroid. Adashi et al. showed in rat granulosa cells that the effect of FSH plus IGF-I was nearly 10-fold greater than the effect of FSH by itself, even though IGF-I by itself was seemingly inactive (8). Similarly, whereas IGF-I is a relatively weak mitogen in rat thyroid cells, its mitogenicity is potentiated more than 30-fold after these cells are exposed to TSH (9). Both IGF-I and IGF-II produce their biological effects through type I receptors that are homologous with the insulin receptor (10). Insulin and the IGFs all cross-react with the type I and insulin receptors and with hybrid receptors containing subunits of each receptor. Type I receptors are present in most, if not all, tissues. The cytoplasmic domain of the type I receptors contains a tyrosine kinase that initiates a phosphorylation cascade through the Ras-Raf-mitogen-activated protein kinase and phosphoinositol 3′-kinase pathways (7). A variety of docking proteins, including insulin receptor substrate-1 and -2 and Shc, act as immediate substrates for the receptor tyrosine kinase. Type I receptors have been observed in proliferating chondrocytes from several species, with predominant expression in developing chondrocytes (11, 12). The IGF-II/mannose-6-PO4 receptor protects against toxic levels of IGF-II, but is not believed to mediate the actions of IGF. Secretion of IGFBPs accompanies the secretion of IGFs in most, if not all, tissues in which IGFs are made. The patterns of IGFBP secretion vary considerably between different tissues and in response to differing physiological circumstances, thus imparting a high degree of specificity to the actions of the IGFs. The specific endoproteases that degrade the several IGFBPs are likewise hormonally regulated and provide additional mechanisms for regulating IGF actions. Each of the six major IGFBPs has been identified in skeletal tissues. In osteoblast-like cells, IGFBP-3, -4, and -5 are probably the most important based on their relative abundance and biological potencies (13). The addition of exogenous IGFBPs to bone cells has produced both stimulatory and inhibitory effects on IGF actions (14). IGFBP-4 does not bind to cells, and it inhibits bone cell proliferation by competing with the IGF receptor for binding the IGFs (15). By contrast, IGFBP-3 and -5 enhance receptor binding and the anabolic effect of the IGFs on bone (16). IGFBPs may be responsible for the high levels of IGF-I and IGF-II in bone matrix. IGFBP-3 binds to proteoglycans prepared from a variety of cartilage tissues, and the bound IGFBP-3 sequesters IGF-I in extracellular matrix (17). Matrix proteoglycans do not bind IGFs in the absence of IGFBP-3. IGFBP-5 has strong affinity for hydroxyapatite, and in the absence of IGFBP-5, the IGFs do not bind to hydroxyapatite (18). The importance of binding proteins as modulators of IGF action suggests that the most promising possibilities for exploiting IGFs as therapeutic agents in skeletal disease may depend on finding ways to selectively manipulate IGFBPs. This might be accomplished by altering their expression or proteolytic degradation or by discovering substances that compete with IGFs for binding to the binding proteins. Another approach has been to modify IGF-I by truncating the amino-terminus. Such peptides [e.g. des(1–3)-IGF-I] have markedly reduced affinity for binding proteins and in many systems are far more active than native IGFs (19). An early issue was whether GH acts directly on skeletal tissues to stimulate growth, or whether its growth-promoting actions require the mediation of IGF-I as stipulated in the somatomedin hypothesis of GH action. Although it is now well established that IGFs are essential for normal skeletal growth, arguments have persisted concerning whether the IGF-I that mediates skeletal growth is derived from the peripheral circulation or is synthesized in the growth plate in response to GH and other hormones that stimulate its synthesis locally (20). In our opinion this is not a substantive issue, because both endocrine and paracrine modalities have been well documented in vivo. Targeted knockout of the IGF-I gene in the liver (but not in other tissues) by the Cre/loxP system does not impair the growth-promoting actions of GH, even though blood levels of IGF-I are substantially decreased (21). Controversies also remain, however, over whether IGF-I by itself can stimulate the proliferation of chondrocytes or can do so only after GH has first stimulated the differentiation of prechondrocytes into more mature cartilage cells, as postulated by Green’s dual effector theory of GH action (22). The ability of IGF-I to restore growth in dwarfed children who lack functional GH receptors suggests that IGF-I delivered from the circulation is sufficient to produce growth without the need for GH (23). Such responses, however, attenuate over time and are less than those observed with GH treatment of GH deficiency states. Ohlsson has postulated that GH is required for providing an adequate stem cell population of prechondrocytes (24). Definition of the roles of the somatomedins in skeletal growth has been obtained from gene knockout models in mice lacking the genes encoding IGF-I, IGF-II, or their respective receptors. Knockout studies confirmed that both IGF-I and IGF-II are essential for normal prenatal growth, as mice deficient in either IGF-I or IGF-II were only 60% of normal size at birth (25). Knockout of IGF-II did not alter postnatal growth, as the size ratio between normal and mutant littermates was maintained into adulthood. Unlike the IGF-II-deficient mice, most of the IGF-I mutant mice died at birth, and those that survived showed severe postnatal growth retardation. Mice with gene knockout of the type I IGF receptor had even greater fetal growth deficits (45% of normal birth weight) than knockouts of either IGF-I or IGF-II alone (26, 27). Mice lacking the type I IGF receptor exhibited delayed appearance of their ossification centers and delay in epiphyseal maturation, and they usually died at birth due to generalized hypoplasia of all muscles, including those responsible for respiration. Experiments with combined deletions made clear that the type I receptor mediates the essential functions for IGF-I and most of those for IGF-II, at least during fetal development. Gene knockout studies revealed that a primary function of the IGF-II receptor (IGF-II/M6Pr) is to regulate IGF-II levels, which can be lethal if too high (28). A null mutation in the IGF-II/M6Pr gene resulted in large birth weights and lethality of nearly all mutant embryos. Simultaneous deletion of the IGF-II gene completely rescued the phenotype associated with the IGF-II/M6Pr gene knockout, because high levels of IGF-II were avoided (29). Genetic models of selective GH or IGF-I deficiency in humans confirm the rodent knockouts. A child reported by Woods et al., who had an inactivating mutation of the IGF-I gene, suffered severe intrauterine and postnatal growth retardation (30). In contrast, children who lack the GH gene or who are unresponsive to GH are essentially normal at birth, and their growth retardation is limited to postnatal life. From this we can conclude that IGF-I and IGF-II, but not GH, are essential for normal intrauterine skeletal growth. Immortalized T lymphocytes from a tribe of African Pygmies have a decreased number of type I IGF receptors, and those present are not phosphorylated and do not transmit a signal in response to IGF-I (31). Baserga and colleagues have shown that the proliferation of transgenic cells lacking the IGF type I receptor is markedly impaired, and that they do not respond to epidermal growth factor and other mitogens (32). Skeletal tissues are rich in IGFs. Baylink et al. found that extracts from demineralized human femora removed during hip replacement procedures contain a factor(s) that stimulates cell division in rat calvaria. The active ingredient that they provisionally called skeletal growth factor was subsequently purified and found on sequence analysis to be IGF-II (33). The formation of IGFs, their receptors, and their binding proteins in skeletal tissues is regulated by many hormones, including GH, estradiol, testosterone, bone morphogenic proteins, PTH, PTHrP, 1,25-dihydroxyvitamin D3,, and a variety of cytokines and growth factors. At least some of the actions of sex steroids on bone are dependent on their stimulation of IGF production. Estradiol stimulation of cell proliferation and collagen synthesis in rat osteoblasts can be blocked by anti-IGF-I antibodies (34), and testosterone-dependant growth of rat mandibular condyles is also blocked by an antibody to IGF-I (12). In female reproductive tissues, not only is the IGF-I gene responsive to estrogen (35), but, conversely, IGF-I can activate the estrogen receptor in the absence of estrogen (36). There is also cross-talk between the IGF and estrogen signal transduction pathways (37). Such interactions may provide an important key to manipulating IGFs for therapeutic advantage. The location of IGF production in the epiphyses is pertinent to their role in skeletal growth. Induction of IGF-I messenger ribonucleic acid by GH has been demonstrated in the rodent growth plate, and IGF-I transcripts have been localized to both proliferating and hypertropic chondrocytes. In situ studies, however, suggest minimal expression of IGF-I messenger ribonucleic acid in rodent epiphyses and high expression of the gene encoding IGF-II (11, 38). It is intriguing to speculate on how abnormal epiphyseal maturation in certain chondrodysplasia might be influenced by the interactions of IGF-I with PTHrP or with fibroblast growth factor (FGF; Fig. 1). PTHrP (or PTH) stimulates chondrocyte proliferation at least in part by stimulating increased production of IGF-I; indeed, PTH stimulation of osteoclasts is blocked by an anti-IGF-I antibody (39). The commitment of chondrocytes to further differentiation, however, is inhibited by PTHrP, an action that is opposite that of the IGFs (40). Knockouts of the PTH/PTHrP receptor display accelerated chondrocyte differentiation and disordered organization. Blömstrand chondrodysplasia, a form of short-limbed dwarfism, and Jansen’s metaphyseal chondrodysplasia are both caused by mutated PTHrP receptors in the chondrocyte (41, 42). Thus, a potentially fruitful area for further research is to determine how IGFs influence the effect of PTHrP on chondrocyte differentiation under normal conditions and in the chondrodysplasias resulting from disordered expression of the PTH/PTHrP receptor. Many of the chondrodystrophies are the result of mutant genes encoding the receptors for FGF or PTHrP. This figure illustrates that the actions of FGF and PTHrP on the growth plate in these diseases are opposite the actions of IGFs, suggesting possible therapeutic approaches to circumvent the growth impairment in these disorders. Other forms of chondrodystrophies, including achondroplasia, hypochondroplasia, and thanatophoric dysplasia, are caused by gain of function mutations of the FGF receptor (43) that act to decrease chondrocyte proliferation and cellular hypertrophy (44). Studies in growth plate chondrocytes from a fetus with thanotropic dwarfism suggest that the receptor mutation led to premature terminal differentiation and increased apoptosis of chondrocytes in response to FGF (45). The enhanced apoptosis is again opposite the effects of IGF-I. IGF-I inhibits apoptosis in osteoblasts, and survival of these cells is blocked by an IGF-I receptor antibody. It remains to be demonstrated whether the antiapoptotic effects of IGF-I can block the effects of the mutant FGF receptor gene in these hereditary chondrodystrophies (46). The evidence that IGF-I plays a pivotal role in pubertal growth is substantial. In both sexes blood levels of both GH and IGF-I rise dramatically during the second decade, implicating the GH/IGF axis in the pubertal growth spurt (47), and in both sexes, estrogen is the primary stimulus for increased GH secretion and the rise in circulating IGF-I In this is presumably accomplished by of to estradiol, as do not a There is between circulating levels of sex steroids and IGF levels at due to actions of sex steroids on the growth receptors and both estrogen receptors have been localized to human growth plate and and have actions on human chondrocytes. Another factor to the between circulating levels of and IGF-I may be the action of estrogen on growth and GH/IGF expression. or physiological of estrogen growth whereas the high of estrogen to or in IGF-I levels and and growth New on the roles of sex steroids and IGFs in pubertal growth have from the of three epiphyses to as the result of estrogen action. et al. reported a epiphyses had not and who was unresponsive to large of exogenous lack of bone maturation was to a of function mutation of estrogen receptor with deficiency have been with a similar phenotype of caused estrogen deficiency and produced skeletal similar to those in the with estrogen of the two with exogenous led to epiphyseal and in their skeletal In all three of these the IGF-I levels were for skeletal These of nature that estrogen is responsible in both and for the terminal of epiphyseal differentiation to and that IGF-I is to effect epiphyseal in the absence of These however, were well and their at an Although it is to their growth during to their normal levels, it is possible that their normal growth was to IGF-I. The is by the relatively normal growth of with primary and an GH/IGF axis The ability of IGF-I to promote growth without skeletal maturation has important therapeutic The skeletal that be from or the actions of IGF-I or IGF-II forms of growth chondrodystrophies, of the and treatment of and many Many of the of IGFs in skeletal have been from therapeutic with GH, even though the effects of GH may be different from those of IGF-I. GH has a of effects that are and in These effects are opposite the anabolic effects of GH that are by IGF-I the growth that is an of high may be to IGF-I with less of than might be from treatment with The most use for IGF-I is to children who are to respond to GH The growth responses to IGF-I in these children have been not as as those in children with GH et al., who have children for than have no the children have developed of and tissue (23). have not been observed in children with in primary are to the of IGF-I or IGF-II for purposes. IGF receptors are present in every cell in the and to concentrations of IGFs are to produce and for for when arose that IGF-I might be the of the IGFs and their receptors, the is to selectively manipulate the actions of the IGFs by taking advantage of the many of that their actions. The is not only to our understanding of the roles that IGFs under circumstances, but to how to manipulate these effects in a selective and The most approach is to modify of the to the IGF-II is more than IGF-I in blood and skeletal studies from the effects of IGF-I and IGF-II with and with those of forms that do not with the IGFBPs. The six major binding proteins and the proteolytic that degrade nearly for selectively modulating the of active IGFs to specific A recent of this approach was a in transgenic mice in which the by selectively IGF-I was by the inhibitory IGFBP-4 proteins might also be through the hormones that regulate their production degradation or by IGFs that do not to binding proteins. A different approach to on the many actions of IGFs in skeletal tissue is to modify the effects of other hormones actions are or by IGF. An is the use of estrogen that selectively block the effects of estrogen on epiphyseal their effects on bone suggest that bone may be an estrogen at the level of the chondrocyte This is by that treatment can to normal growth without stimulating bone in with to In an the effect of estrogen on bone in mice was blocked by an estrogen receptor Gene may to be a for the Although the of gene has far more difficult to than it is that in we gain in genes to fulfill specific at a gene to a specific area or function in the growth plate, it be important to unique or for the chondrocyte in of the of gene to skeletal however, far more than we now on how IGFs function in with other hormones and growth factors to influence each of skeletal growth.