Clinical Endocrinology Branch, National Institute of Diabetes Digestive and Kidney Diseases, Bethesda, Maryland 20892-1758
Address correspondence and requests for reprints to: Derek Le Roith, M.D., Ph.D., Chief, Clinical Endocrinology Branch, National Institute of Diabetes Digestive and Kidney Diseases, Room 8D12, Building 10, Bethesda, Maryland 20892-1758. E-mail derek{at}helix.nih.gov
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
The IGF System (Fig. 1![]() |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
In the local tissue environment, the IGFBPs modulate the interaction of
the IGF with their receptors. IGFBPs with high affinity for the IGF
have been shown to inhibit their interaction with the receptors.
However, regulatory mechanisms exist that cause the slow release of the
IGF, thus enabling prolonged interaction with the receptors and
resulting in enhanced biological responses. The affinity of the IGFBP
is regulated by protease activity, with proteolytic cleavage being one
mechanism potentially involved in releasing IGF from an inactive
complex. Other mechanisms allowing the release of the IGF from the
complex include the state of phosphorylation of the IGFBPs and the
partitioning of the IGFBPs from extracellular matrix to cell surface
association that decreases the affinity of the IGFBP for the ligands.
Finally, there is increasing evidence that certain IGFBPs may affect
cellular responses in an IGF-I/IGF-IR-independent manner. The exact
mechanism for this latter response is yet to be defined, although
interactions with cell surface receptors such as integrins and
transforming growth factor-ß (TGF-ß) type 5 receptor have been
reported (figure 1).
Regulation of the IGF system
Determining the distribution and ontogeny of the expression of the various ligands, binding proteins, and receptors has been the basis for a large number of studies over the past 2 decades. The advent of modern molecular biology and the characterization of the genes for most of the components of the IGF system stimulated this process. Most interesting, in terms of growth regulation, has been the identification and evaluation of the interaction between GH and the IGF-I system. In the 1970s, Daughaday et al. (1) coined the phrase "somatomedin hypothesis." The hypothesis emerged from the seminal findings that, for GH to stimulate 35S-sulphate incorporation into the cartilage of growing bones, a circulating factor was required that virtually disappeared from the circulation following hypophysectomy and returned after GH treatment to these rats (8). The production of "somatomedin," or IGF-I as it later became known, by the liver was enhanced by GH and entered the circulation whereby it reached its target tissue (i.e. the "endocrine form" of IGF-I). Isaksson et al. (9) modified this hypothesis in the 1980s after the discovery by DErcole et al. (10) that IGF-I was produced by most, if not all, nonhepatic tissues. Furthermore, GHR were also detected in multiple nonhepatic tissues, indicating that the actions of GH was not confined to the liver. The modified somatomedin hypothesis suggested that GH affects postnatal longitudinal growth by two mechanisms: 1) by stimulating hepatic IGF-I production, which produced IGF-I acting as an endocrine hormone; and 2) stimulating GHR present in the growth plate, thereby increasing IGF-I synthesis that acted locally in an autocrine/paracrine manner.
Proof of the modified hypothesis has awaited more modern technology, which has now become available (vida infra). The hypothesis stating the liver as the primary source of circulating IGF-I was based on calculations of hepatic production rate and by the observation that liver IGF-I expression far exceeds that in most nonhepatic tissues. Verification of this hypothesis was, however, technically difficult, and it is only with recent advances in molecular biology that experiments testing this hypothesis have become feasible. In particular, the discovery and validation of the Cre-LoxP system for engineering tissue- or inducible-gene knockouts has allowed, for the first time, scientists to examine tissue-specific functions of a given gene. To determine the relative contribution of the circulating hepatic "endocrine" form of IGF-I and the autocrine/paracrine form of IGF-I in postnatal growth and development, liver IGF-I gene was deleted in mice using the bacteriophage Cre-Lox/P system of tissue-specific tissue gene deletion. Mice with lox/P sequences flanking exon 4 of the IGF-I gene were created by homologous recombination and then mated with mice that express the recombinase enzyme Cre in the liver. Successful matings resulted in Cre recombining the region between the lox/P sequences only the liver, resulting in a marked reduction of IGF-I expression in the liver (11).
Mice in which exon 4 of the IGF-I gene was deleted in liver had an
approximately 80% reduction in circulating IGF-I levels. A
several-fold increase in serum GH levels was also observed, most likely
due to the loss of negative feedback from circulating IGF-I on
pituitary GH synthesis and release. Interestingly, the levels of
IGFBP-1 and IGFBP-3 were also markedly reduced. However, despite the
loss of liver IGF-I and the marked reduction in circulating IGF-I,
postnatal and peripubertal growth was normal. Therefore, it would seem
that although pubertal growth spurts induced by GH are entirely IGF-I
dependent (J. L. Liu, unpublished data), it is not
dependent on liver IGF-I production. Rather, autocrine/paracrine IGF-I
expressed at the sites of GH action maybe sufficient for this function
(Fig. 2). The role for circulating IGF-I,
in addition to controlling GH secretion, still remains to be
established. Thus, yet again, the somatomedin hypothesis has been
revisited, found to be incomplete, and, therefore, is in need of
revision.
|
IGF in fetal and postnatal development
Fetal tissues produce both IGF-I and IGF-II. In rodents
IGF-II synthesis declines rapidly postnatally and is virtually
undetectable in most adult tissues and serum. IGF-I, on the other hand,
increases in the circulation during the peripubertal stages under the
direct influence of circulating GH. The importance of both growth
factors in fetal growth and development has been convincingly
demonstrated by gene-deletion experiments. Using homologous
recombination, the genes for IGF-I and IGF-II were deleted in mice (12, 13). In both instances, fetal development was affected. At birth the
animals with the IGF-I gene deleted were smaller and often died
immediately after birth due to underdevelopment of their respiratory
muscles. Those that survived this period grew more slowly postnatally
than their normal littermates, and both males and females were
infertile. At birth they were 60% of the normal weight, and
throughout the pubertal growth spurt showed minimal growth, becoming
30% of the weight of wild-type adult animals.
Mice with the IGF-II gene deleted were 50% normal birth weight, but
grew at a normal rate postnatally, in parallel with their normal
littermates, and, thus, did not show catch-up growth. Their development
was otherwise normal. These results can be interpreted to have shown
that although both IGF-I and IGF-II are important for fetal growth,
IGF-I is essential for normal postnatal growth and development in the
mouse. Additionally, when GH was administered to the IGF-I knockout
mice there was no change in growth measured as body weight and femoral
length. Thus, the effects of GH on these parameters in the peripubertal
stage of development is totally dependent on IGF-I. This result is in
stark contrast to the ability of IGF-I overexpression to correct the
dwarfism associated with GH deficiency.
The function of IGF-II is less clear, although at a paracrine level it is presumably able to emulate the mitogenic functions of IGF-I by activation the IGF-IR. The absence of IGF-II in juvenile mice and rats has meant that experiments examining its functions are difficult, relying on larger experimental animals, such as sheep and pigs. However, quantitative trait loci affecting skeletal and cardiac muscle mass and fat deposition have been mapped to the IGF-2 gene locus in pigs, suggesting an important role for IGF-2 in postnatal development in some mammals.
In vivo metabolic effects of IGF-I and IGF-II
Whereas the IGF are primarily growth factors, when administered
either by acute bolus injection or chronic infusion many insulin-like
effects on metabolic functions have been reported. These responses can
differ depending on whether IGF-I is given acutely or chronically.
However, it should be noted that these effects are pharmacological.
Acute or chronic administration of rhIGF-I to animals and humans
results in a lowering of blood glucose (14). The exact mechanism
whereby this effect is brought about is still incompletely understood.
IGF-I inhibits insulin and C-peptide release from the ß-cell of the
pancreas, acting through IGF-IR expressed on these cells. IGF-I, in
contrast to insulin, inhibits glucagon secretion by the -cells of
the pancreatic islets, and the lowered circulating glucagon levels
reduce hepatic glucose output. On the other hand, the effect of IGF-I
on inhibiting hepatic glucose output is less than that for insulin
because hepatocytes express IR but not the IGF-IR. IGF-I may stimulate
glucose uptake by the muscle either via the IR or via the IGF-IR, both
of which are expressed in skeletal muscle. In regard to protein
metabolism, although insulin prevents proteolysis in muscle, IGF-I both
inhibits proteolysis and enhances protein synthesis.
IGF-I administered chronically by sc injection causes circulating
FFA to rise, probably due to inhibition of insulin that is
antilipolytic. As noted earlier, IGF-II seems to activate the IR. The
significance of this result is unclear, although it is interesting to
note that in the human circulating IGF-II levels are actually higher
than that of IGF-I (500700 ng/mL compared with 200300 ng/mL for
IGF-II and IGF-I, respectively). The possibility that IGFs in the
circulation may contribute to the regulation of glucose homeostasis by
insulin, perhaps by regulating the availability of IGF-I through
regulating IGFBP levels in the circulation, has been discussed. For
example, insulin is a potent inhibitor of hepatic IGFBP-1 production
(15). IGFBP-1 inhibits IGF metabolic activities, whereas transgenic
mice expressing high levels of IGFBP-1 have been reported to have
defects in glucose homeostasis.
![]() |
Complications/Diseases |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
A single case of IGF-I gene deletion in a young boy has been described recently (16). Homozygous partial deletion of the coding region of the IGF-I gene abrogated IGF-I gene expression, and the patient demonstrated absent circulating IGF-I. Similar to the murine model described above, the young boy was growth retarded. He had a number of dysmorphic facial features, a small brain, and mild mental retardation. As expected, serum GH levels were markedly elevated, and this was associated with increased ALS levels. Both were normalized on rhIGF-I therapy that also increased IGFBP-3 levels. Other causes of IGF-I deficiency in children include primary and secondary GH deficiencies, GHR mutations, malnutrition, and many severe disease states. These will be discussed in other reviews in this volume.
IGF-I in type 1 diabetes
The GH/IGF-I axis is often abnormal in patients with poorly controlled type 1 diabetes. With inadequate insulin therapy and the resultant hyperglycemia, the patients develop a resistance to GH, as evidenced by reduced circulating IGF-I levels and elevated serum GH levels. The lack of insulin causes an elevation of IGFBP-1, and the lowered IGF-I levels cause a fall in IGFBP-3 in the circulation. If this situation occurs during the period of the pubertal growth spurt, patients will exhibit reductions in growth velocity, reversible with adequate insulin therapy and normalization of the GH/IGF-I axis. Normalization of the axis will also lower the GH levels, which might play a role in the deterioration of glycemic control by exercising an anti-insulin effect (17).
IGF-I has insulin-like effects when given acutely, particularly on blood glucose levels. Thus, rhIGF-I has been tested as a possible therapeutic agent in type 1 diabetes (as well as in type 2 diabetes). When administered sc at bedtime, rhIGF-I decreased GH secretion (18). Long-term administration resulted in an improvement in glycemic control and a reduction in insulin requirements (19). These effects are apparently due to inhibition of GH secretion and the effect of GH on antagonizing insulin, thereby improving insulin sensitivity in these cases. Although these results were initially very promising, in one study the effect was seen at 3 months of therapy and absent at 6 months after initiation of therapy. Furthermore, the results are tempered by the side effects experienced by the patients (usually at higher doses than required for adequate therapy), and the concern that long-term therapy may be associated with mitogenic responses, including retinopathy and tumor progression, although no long-term clinical trials have been carried out.
Use of rhIGF-I in patients with extreme severe insulin resistance was a particularly exciting idea because these patients fail to respond even to extremely high doses of insulin. The patients tested included those with the type A insulin resistance syndrome, the Rabson-Mendenhall syndrome, leprechaunism, and lipodystrophy, all due to a genetic mutation in the IR. Many of the patients treated with rhIGF-I demonstrated increased insulin sensitivity and a lowering of the blood glucose. Because many of these patients had mutations in the IR, it was concluded that the effects of the administered rhIGF-I were mediated through the IGF-IR. Proof of this hypothesis needs further investigation because some studies have shown no response in the most severe cases, such as, for example, in a case of leprechaunism due to a compound heterozygous mutation with a deletion of the IR from one allele (20, 21).
The role of the IGF system in the complications of type 1 diabetes
Multiple factors play a role in the development of microvascular complications of type 1 diabetes. Hyperglycemia is one of the most important factors since improvements in glycemic control has been shown unequivocally by the Diabetes Control and Complications Trial to delay or prevent microvascular events, such as diabetic retinopathy and nephropathy. In addition to hyperglycemia, growth factors, including the IGF system, have been invoked as being causative in the microvascular changes. Neovascularization seen in diabetic retinopathy may lead to blindness. GH therapy has resulted in deterioration of retinopathy in a few cases, and IGF-I has been invoked as the mediator of this effect. In a model of ischemia-induced retinopathy in mice, GH antagonists reduced this effect in parallel with a decrease in GH and IGF-I levels (22, 23). Similar results were obtained with a somatostatin analog that potently inhibits GH levels. When either GH or IGF-I were coadministered with the somatostatin analog, the neovascularization was restored. Whereas IGF-I is not a very powerful mitogen for endothelial cells by itself, it enhances the expression of other growth factors, such as fibroblast growth factor and vascular endothelial growth factor, which are powerful mitogens and have been shown to cause pathological changes in the vasculature (24). Thus, local retinal production of IGF-I in diabetics has been invoked as a causative agent in retinopathy. Similarly, the deterioration of the retinopathy seen in the early stages of insulin therapy and correction of hyperglycemia maybe due to an increase (normalization) of circulating IGF-I levels. These findings have led many investigators to send a message of caution in the use of rhIGF-I in patients with type 1 diabetes.
Studies in various animal models of diabetes have suggested a role for the IGF system in the renal complications of diabetes. In these animal models, IGF-I protein levels may be elevated in the kidney, causing hemodynamic changes including increased glomerular filtration rates, increased renal plasma flow, and an enlargement of the kidney typically seen in poorly controlled type 1 diabetes in humans. The effect on the vasculature is via the kallikrein and angiotensin systems (25). Persistent elevations in IGF-I levels may lead to changes in the microvasculature and basement membrane changes, which could explain the long-term nephropathic complication seen in patients with diabetes. The elevated IGF-I levels are apparently not associated with detectable elevations in IGF-I gene expression, but are apparently due to increased expression of IGFBPs, which trap circulating IGF-I and deliver it to the IGF-IR (26).
Cancer
Numerous studies have demonstrated a significant role for the IGF system in controlling tumor growth (27, 28, 29). Many tumors overexpress IGF-II, IGF-IR, and certain IGFBPs. IGF-II expression leads to an autocrine feedback loop stimulating the IGF-IR to stimulate cancer cell proliferation. The production of IGFBPs by the tumor cells often enhances the effect of the IGFs, thereby increasing cancer cell proliferation (30). Conversely, IGFBPs might also potentially act as tumor suppressors. For example, IGFBP-3 has been reported to stimulate apoptosis in an IGF-I-independent manner, possibly by binding to and activating the type V TGF-ß receptor (31).
IGF-II overexpression is found consistently in pediatric cancers, such as Wilms tumor (WT) and rhabdomyosarcoma (32, 33). In most normal human tissues, IGF-II is actively transcribed from the paternal allele, which is highly methylated in its promoter region. The maternal allele is silenced by a mechanism known as genomic imprinting. Loss of imprinting can lead to carcinogenesis. In the case of rhabdomyosarcomas and WT, for example, the silenced maternal IGF-II allele is activated. Furthermore, in cases of the Beckman-Wiedeman syndrome, there is evidence for unbalanced translocations with paternal duplication of chromosome 11, and other chromosomal abnormalities affecting the region of chromosome 11p, which contains the IGF-II gene locus (34, 35). The increased IGF-II gene expression may explain this overgrowth syndrome with increased incidence of WT. The WT gene product is a transcription factor that suppresses the expression of many growth factors and their receptors by direct binding to the promoter regions, thereby modulating their expression (36). When mutated in the DNA-binding domain, WT allows the overexpression of IGF-II and the IGF-IR. Mutations in WT are often found in cases of WT. The role of the IGF-IR in cancer has been studied extensively by a number of investigators. Overexpression of IGF-IR is a general phenomenon in cancer, and although not by itself an initiator of cancer development this receptor may enhance tumor proliferation and metastases. Reductions in IGF-IR levels using antisense technologies inhibit tumor growth (37, 38).
On rare occasions, the production of IGF-II by tumors may result in
hypoglycemia (Fig. 3). These tumors
express an incompletely processed form of IGF-II, called "big
IGF-II" (39). Big IGF-II cannot complex with IGFBP-3 and ALS in the
circulation and is shifted to a complex with IGFBP-2, allowing it to
reach the target tissues and interact with the IR. Big IGF-II retains
biological activity and stimulates skeletal muscle glucose uptake and
inhibits hepatic glucose output, resulting in hypoglycemia (40). Big
IGF-II inhibits GH hormone secretion, which in turn reduces circulating
IGF-I, IGFBP-3, and ALS levels. Removal of the tumor reverses these
effects, and the hypoglycemia is ameliorated (41).
|
In the diagnosis of growth retardation, GH levels after provocative stimuli are the standard test used. Because these measurements often fail to discriminate between normal and GH-deficient children, measurements of circulating total or free IGF-I were considered as an alternative. Although a number of studies have shown a good correlation between GH levels, IGF-I levels, and growth retardation, too many variables, such as nutritional status, complicate the use of IGF-I as an alternative (42). The addition of IGFBP-3 measurements increased the sensitivity, as did the increased levels of IGFBP-2 in GH-deficient children (43). Thus, a single parameter seems to be inadequate, and multiple parameters improve the sensitivity but still fail to improve specificity. In serum IGF-I, levels are increased in acromegaly and are used to monitor the effectiveness of therapy. In children with GH deficiency, however, growth rate is used to monitor response to GH rather than the levels of IGF-I in the circulation.
![]() |
Conclusions |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Received August 31, 1999.
Accepted September 20, 1999.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|