Division of Endocrinology, Department of Pediatrics, University of North Carolina School of Medicine (J.J.V.W.), Chapel Hill, North Carolina 27599; and the Division of Endocrinology and Metabolism, Department of Internal Medicine, University of Cincinnati School of Medicine (E.P.S.), Cincinnati, Ohio 45229
Address all correspondence and requests for reprints to: Dr. Judson J. Van Wyk, Department of Pediatrics, University of North Carolina School of Medicine, C.B.# 7220, Chapel Hill, North Carolina 27599.
![]() |
Introduction |
---|
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.
![]() |
Overview of IGF ligands, receptors, and biological actions |
---|
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.
![]() |
IGF-binding proteins (IGFBPs) |
---|
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, 2, 3)-IGF-I] have markedly reduced affinity for binding proteins and in many systems are far more active than native IGFs (19).
![]() |
Role of IGFs in skeletal growth |
---|
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 Greens 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).
![]() |
Evidence from gene deletion studies on the role of IGFs in skeletal growth |
---|
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).
![]() |
Synthesis and biological actions of IGFs in skeletal tissue |
---|
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 Jansens 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.
|
![]() |
Interactions among IGFs, GH, and sex steroids in pubertal growth |
---|
There is poor correlation between circulating levels of sex steroids
and IGF levels at puberty, possibly due to direct actions of sex
steroids on the growth plate. Androgen receptors (49) and both and
ß estrogen receptors (50) have been localized to human growth plate
chondrocytes, and dihydrotestosterone and estradiol have direct actions
on human chondrocytes. Another factor contributing to the poor
correlation between circulating levels of estradiol and IGF-I may be
the biphasic action of estrogen on growth and GH/IGF expression. Low or
physiological doses of estrogen augment statural growth (51), whereas
the high doses of estrogen used to treat acromegaly or tall stature in
girls suppress IGF-I levels and reduce acral and statural growth (52).
New insights on the roles of sex steroids and IGFs in pubertal growth
have come from the identification of three rare patients whose
epiphyses failed to fuse as the result of disrupted estrogen action.
Smith et al. reported a tall, 28-yr-old, normally
masculinized man whose epiphyses had not yet fused and who was
unresponsive to large doses of exogenous estrogen. His lack of bone
maturation was secondary to a loss of function mutation of estrogen
receptor (53). Two men with p450 aromatase deficiency have been
described with a similar phenotype (54, 55). Failure of aromatization
caused profound estrogen deficiency and produced skeletal lesions
similar to those in the patient with estrogen insensitivity. Treatment
of the latter two patients with exogenous estrogens led to epiphyseal
fusion and improvement in their skeletal lesions. In all three of these
patients the IGF-I levels were appropriate for skeletal age.
These experiments of nature show that estrogen is responsible in both males and females for the terminal phases of epiphyseal differentiation leading to fusion, and that IGF-I is unable to effect epiphyseal fusion in the absence of estrogen. These men, however, were well virilized and attained their midparental height at an appropriate age. Although it is tempting to attribute their estrogen-independent growth during childhood to their normal androgen levels, it is equally possible that their normal growth was attributable to IGF-I. The latter possibility is supported by the relatively normal statural growth of individuals with primary hypogonadism and an intact GH/IGF axis (56). The ability of IGF-I to promote linear growth without unduly accelerating skeletal maturation has important therapeutic implications.
![]() |
Potential therapeutic usages and frontiers for further study |
---|
The most obvious clinical use for IGF-I is to treat children who are unable to respond to GH itself. The growth responses to IGF-I in these children have been impressive, although not as sustained as those in GH-deficient children treated with GH itself. Underwood et al., who have treated 10 such children for longer than 5 yr, have encountered no serious side-effects, although the children regularly have developed mild enlargement of abdominal viscera and lymphoid tissue (23). Similar findings have not been observed in children treated with GH.
Except in primary IGF-I-deficient states, there are significant disadvantages to the systemic administration of IGF-I or IGF-II for organ-specific purposes. IGF receptors are present in virtually every cell in the body, and exposure to pharmacological concentrations of exogenously administered IGFs are likely to produce unpredictable and unwanted side-effects. Enthusiasm for treating insulin-dependent diabetes, for example, waned when concerns arose that IGF-I might be exacerbating diabetic retinopathy.
Despite the ubiquity of the IGFs and their receptors, the body is able to selectively manipulate the actions of the IGFs by taking advantage of the many layers of regulation that govern their actions. The challenge, therefore, is not only to increase our understanding of the roles that IGFs play under varying pathological circumstances, but to learn how to manipulate these effects in a selective and beneficial manner.
The most obvious approach is to modify access of the appropriate ligand to the target tissue. Because IGF-II is more abundant than IGF-I in blood and skeletal tissue, future studies would profit from comparing the effects of IGF-I and IGF-II with one another and with those of modified forms that do not interact with the IGFBPs.
The six major binding proteins and the proteolytic enzymes that degrade them offer nearly inexhaustible opportunities for selectively modulating the availability of biologically active IGFs to specific targets. A recent example of this approach was a study in transgenic mice in which the smooth muscle hyperplasia induced by selectively overexpressing IGF-I was partially abrogated by concomitantly overexpressing the inhibitory IGFBP-4 (58). Binding proteins might also be manipulated through the hormones that regulate their production and/or degradation or by using modified IGFs that do not attach to binding proteins.
A different approach to capitalizing on the many actions of IGFs in skeletal tissue is to modify the effects of other hormones whose actions are mediated or impacted by IGF. An example is the use of estrogen antagonists that selectively block the effects of estrogen on epiphyseal fusion while retaining their beneficial effects on bone mass. Case reports suggest that tamoxifen, while promoting bone mineralization, may be an estrogen antagonist at the level of the chondrocyte maturation. This is supported by anecdotal reports that tamoxifen treatment can lead to normal statural growth without stimulating bone age advancement in patients with elevated estrogens secondary to McCune-Albright syndrome. In an experimental study, the accelerating effect of estrogen on bone age advancement in mice was blocked by Faslodex, an estrogen receptor blocker (59).
Gene therapy may ultimately prove to be a viable technique for achieving the goals outlined above. Although the enormous potential of gene therapy has proven far more difficult to realize than originally envisioned, it is likely that in coming years we will gain expertise in targeting designer genes to fulfill specific objectives at discrete anatomical sites. To successfully target a modified gene to a specific area or function in the growth plate, it will be important to identify unique promoters or cell-specific fingerprints for the multiple chondrocyte stages in growing bone. Realization of the enormous potential of targeted gene therapy to treat skeletal disorders, however, requires far more knowledge than we now possess on how IGFs function in collaboration with other hormones and growth factors to influence each stage of skeletal growth.
Received August 4, 1999.
Revised September 15, 1999.
Accepted August 12, 1999.
![]() |
References |
---|