Human IGF-1 can now be produced in practically unlimited amounts using recombinant DNA technology, and the number of studies of its possible clinical applications is rapidly growing. For example, diabetic patients with extreme insulin resistance had substantial improvement in metabolic control during administration of IGF-1 [9, 10]. More importantly, IGF-1 seemed to improve metabolic control in type II diabetic patients [11].

We describe three main issues. First, we analyze whether using IGF-1 as a substitute for insulin is appropriate, because major differences exist in the kinetic mechanisms of insulin and IGF-1. Second, we review data on the metabolic effects of IGF-1 in normal and diabetic animals. Finally, we describe the metabolic effects of IGF-1 in humans and discuss the potential clinical benefits and risks of IGF-1 therapy in patients with diabetes.

Evolutionary Origin of Insulin-like Growth Factor-1

The concept that IGF-1 mimics insulin action can be defended on the basis of IGF-1 phylogenesis. Chan and colleagues [12] found that insulin and IGF molecules emerged from a common ancestor protein. A gene encoding a polypeptide with a deduced sequence that contains features of both insulin and IGF was detected by DNA cloning in the primitive cephalochordate Branchiostoma californiense [12]. The investigators hypothesized that IGF-1 and IGF-2 genes evolved from the insulin gene by rearrangement of introns (a nucleotide sequence in the DNA of a gene that generally does not code information for protein synthesis and is absent from the mature messenger RNA made from that gene) and subsequent gene duplication, around 300 to 600 million years ago [13]. This intriguing hypothesis explains the known structural similarity between insulin and IGF genes and the homology of these molecules in vertebrates, especially in mammals [14-16].

Kinetic Mechanisms of Insulin-like Growth Factor-1

The production and targeting to sites of action of insulin and IGFs are different, indicating that the biologic roles assigned to these substances have evolved in different directions [13]. Insulin controls the use of body fuels (amino acids, glucose, and fatty acids). It is produced and stored by specialized, sensing cells in the pancreatic islets that are located in a position that enables secretion directly into the portal system [17]. Thirty to 70% of secreted insulin reaches the microcirculation of the major peripheral organs of insulin action, which are skeletal muscle and adipose tissue [18]. Insulin crosses the vascular barrier with high efficiency, ensuring that most insulin molecules reach specific receptors [19]. The amount of insulin bound is controlled to some extent by receptor affinity that decreases in a curvilinear manner (negative cooperativity) with increased occupancy [20] and by insulin receptor down-regulation [21].

Most circulating IGF-1 is produced by hepatocytes [22, 23] and produced by various other cells to act locally in an autocrine or paracrine manner [24, 25]. The regulation of IGF-1 production is to some extent tissue- or organ-specific. In the liver, production is controlled by pituitary growth hormone [23]. In the extrahepatic tissues, its generation is probably controlled by the concerted action of growth hormone [26] and other pituitary hormones (for example, thyroid-stimulating hormone in the thyroid follicular epithelium and follicle-stimulating hormone and luteinizing hormone in the gonads) as well as by locally produced growth factors (for example, platelet-derived growth factor, epidermal growth factor, and fibroblast growth factor) [27]. None of the cellular sources of IGF-1 can store preformed IGF-1 [28]. The various sources of IGF-1 production, the apparent lack of any known forms of intracellular storage, and the existence of both local and endocrine effects suggest that, unlike insulin, IGF-1 is more like a cytokine than a hormone [29]. The other apparent difference between insulin and IGF-1 is that IGF-1 exists in two forms, a free and a bound form with high affinity to and specificity for various soluble insulin-like growth factor binding proteins (IGFBPs). Our knowledge of all of these proteins is incomplete. Four discrete IGFBPs have been identified in human serum, and a fifth has been identified in human cerebrospinal fluid [29-32]. Together they bind most of the IGF-1 in the circulation, leaving less than 10% of the total serum concentration of IGF-1 in the free form [33]. Under normal conditions, IGFBP-3 is the most abundant binding protein in adult human serum [34, 35] but not in human lymph [36]. After complexing with IGF-1, this protein binds an additional 85-kd acid-labile subunit, and the 150-kd complex circulates in the serum with a half-life of 12 to 15 hours [37, 38]. The 150-kd complex is a major storage form of Ireleased after the complex has been broken down by specific proteases [39]. The production of IGFBP-3 is increased in response to increases in growth hormone [40], insulin [41], IGF-1 [42], and a protein-rich diet [42]. Apart from its functions as an IGF-1 storage and cargo protein, IGFBP-3 can bind to cells and modulate IGF-1-stimulated cell growth [43, 44] and metabolism [44] in vitro.

The molecular structure of IGFBP-1 and the sequence of its encoding gene have been established [45, 46]. The levels of this protein in serum appear to be inversely related to prevailing insulin levels [47-50]. Insulin-like growth factor binding protein-1 reaches maximal concentration in the serum during the night when insulin levels are at their nadir [48]; IGFBP-1 apparently serves as a shuttle transporter of IGF-1 from the serum to the interstitial fluid, and it also controls the concentration of free IGF-1 at its site of action [29]. In addition, IGFBP-1 has been shown to modulate the growth-promoting effect of IGF-1 [29]. An increase of IGFBP-1 is associated with growth inhibition of normal tissues [29] and may slow progression of those tumors in which IGF-1 acts as a potent mitogen [51].

The structure of IGFBP-2 and its corresponding gene have also been determined [52]. Its levels appear to be down-regulated by growth hormone [40] and insulin [53] and increased by IGF-1 [54]. The physiologic role of IGFBP-2 is poorly understood, but it may serve as a shuttle transporter of IGF-1 between intravascular and interstitial spaces of target organs. It is the predominant form of IGFBP in cerebrospinal fluid [55]. The regulation of IGFBP-4 is presently unknown, although its molecular structure and the localization of the encoding gene in humans have been reported [56]. Little is known about IGFBP-5. It has been detected in human cerebrospinal fluid and appears to have selective affinity for IGF-2 [57].

Kinetic data for IGF-1 indicate that, despite the common evolutionary origin and the molecular similarity of insulin and IGF-1, their different physiologic assignments result in the development of different systems of generation and delivery to their target organs. Currently used methods of treatment with exogenous IGF-1 rely on doses high enough to substantially increase the level of free IGF-1, but this approach is not physiologic and sometimes produces serious side effects. For example, IGF-1 use in the reversal of catabolic states is substantially compromised by its hypoglycemic effects [58]. In contrast, stimulation of growth (anabolism) may be a side effect when IGF-1 is used to correct hyperglycemia refractory to insulin.

The manipulation of IGFBPs may hold the key to future therapy with IGF-1. At present, little is known about how specific sets of IGFBPs modify IGF-1's spectrum of biologic activity. Despite information derived from in vitro experiments, the nature of these effects in vivo is still largely untested. Clinical states associated with high susceptibility to the hypoglycemic effect of IGF-1 provide some information. For example, both growth hormone deficiency and Laron-type dwarfism are characterized by increased susceptibility of affected persons to the hypoglycemic effect of insulin and IGF-1 [59], a phenomenon usually explained by an absent counter-regulatory effect of growth hormone. These two disorders are also characterized by similar alterations in IGFBPs. The concentrations of both the 150-kd complex and its constituent IGFBP-3 are decreased, although the levels of IGFBP-2 are markedly increased [40, 60, 61]. These changes are especially apparent in Laron dwarfs. Interestingly, increases in IGFBP-2 are observed in patients with extrapancreatic (non-insulin mediated) tumor hypoglycemia [54]. The hypothetical role of specific types of IGFBP in regulating the mode of IGF-1 biologic activity (for example, growth-promoting effect versus hypoglycemic action) is presented in Figure 1.

View larger version (47K): [in this window] [in a new window]   Figure 1. Kinetics of insulin-like growth factor-1 in vivo. 1. The circulating 150-kd complex consisting of insulin-like growth factor-1 (IGF-1), binding protein-3 (BP-3), and a 85-kd acid-labile subunit, is the major storage form of IGF-1 in the body. 2. IGF-1 free or bound to other binding proteins (BP) can cross the vascular barrier to reach the intercellular space. 3. Free IGF-1 interacts with IGF type I, type II, or insulin receptors. 4. Binding proteins associated with cells have a modulatory effect on cell growth and metabolism. 5. Numerous cell types can generate IGF-1 to exert paracrine or autocrine effects. 6. Numerous cells can also produce binding proteins.

In Vitro Studies

Insulin-like growth factor-1 can affect target cells by acting on a specific type I receptor, the insulin receptor, and on the IGF type II receptor [62] (see Figure 1). The affinity of IGF-1 for IGF receptors does not show negative cooperativity; therefore, receptor occupancy increases in a linear manner with the concentration of ligand until binding reaches saturation [63]. Although IGF receptor down-regulation by IGF-1 is well documented in vitro [64, 65], this process can be prevented in vivo by the buffering action of IGFBPs, restricting availability of the free ligand for target cell receptors [65]. The IGF type I receptor and the insulin receptor share substantial homology in their quaternary structure and the amino acid sequence of their and ß subunits [66]. In contrast, the IGF type II receptor has a distinct structure and is homologous with the mannose-6-phosphate receptor [67]. It has greater affinity for IGF-2 than for IGF-1 and probably does not mediate acute metabolic effects.

The close similarity between the insulin receptor and the IGF type I receptor can be explained on the basis that their genes emerged from a common ancestor [68]. According to the old concept, the metabolic effects of IGF-1 (for example, glucose-lowering) are mediated by its effect on insulin receptors [69]. However, under certain circumstances the growth response may be mediated by the insulin receptor and the metabolic response by the IGF type I receptor [69]. For example, in the human hepatoma (Hep-G2) cell line, glycogen synthesis is stimulated by the IGF-1 and the insulin receptors [70]. Additionally, in fibroblasts expressing a defective tyrosine kinase in the ß subunit of the insulin receptor, glucose transport sensitivity after stimulation by insulin is decreased, but its sensitivity to IGF-1 stimulation remains unaltered [71]. These data suggest that the signaling pathway initiated by IGF-1 binding to its receptor is different and functionally independent, at least in its initial steps, from that initiated by insulin. Additional support for this divergence comes from studies of myocytes with and without down-regulated insulin receptors. In myocytes, IGF-1 stimulation of deoxyglucose uptake is unchanged. On the other hand, the cells with down-regulated insulin receptors are less sensitive to stimulation of deoxyglucose uptake by insulin [72]. These examples provide evidence that, under certain conditions, the metabolic consequences of insulin resistance can be corrected by IGF-1 acting on its own specific (type I) receptor. The emerging question then is whether IGF-1 is able to overcome insulin resistance in liver, muscle, and fat, tissues that are normally the major targets of insulin action.

Stimulation of muscle glucose uptake by serum extracts rich in IGFs (nonsuppressible insulin-like activity) was shown for the first time more than 25 years ago [73, 74]. This effect was later confirmed with pure IGF-1 in animal and human muscle preparations. In rat soleus muscle, IGF-1 was equipotent to insulin in stimulating glucose transport [75]. The preferential pathway of glucose disposal during IGF-1 stimulation was a nonoxidative pathway involving glycogen synthesis and then glycolysis of glycogen into lactate, whereas [75] glucose oxidation remained unchanged. In human abdominal rectus muscle, insulin and IGF-1 are equipotent in stimulating 2-deoxyglucose uptake [76].

Insulin-like growth factor-1 receptors are present in human, but not in rat, adipocytes [77-79]. In both rodent and human adipose tissues, stimulation of glucose uptake by IGF-1 is mediated by the insulin receptor [80]. In contrast, lipoprotein lipase activity in adipose tissue is more sensitive to stimulation by IGF-1 than by insulin [81]. Interestingly, the effect of IGF-1 on insulin receptor autophosphorylation in human adipocytes is quantitatively smaller than that of insulin, even though glucose transport stimulation was equivalent [80]. Thus, complete insulin receptor autophosphorylation may not be necessary to stimulate glucose transport [73, 82].

The existence of IGF-1 receptors in rat [83-85] and human [83, 86] liver is well documented. The fact that IGF-1 receptors are present in the liver is interesting because of the known similarity in molecular structure between IGF-1 and proinsulin. Proinsulin has been shown to have a relatively greater capacity to suppress hepatic glucose production [87] compared with its stimulation of peripheral glucose uptake [87]. Whether the preferential action of proinsulin on the liver is due to its interaction with IGF-1 receptors is not known; however, proinsulin has less affinity for the IGF-1 receptor than does insulin [86].

Studies in Animal Models

In both awake [88] and anesthetized normal rats [89], IGF-1 exerts a hypoglycemic effect with a potency about 50 times less than that of insulin. The same difference in potency between insulin and IGF-1 on glucose disposal has been observed using the euglycemic clamp technique [89]. However, equivalent hypoglycemic doses of insulin and IGF-1 produce a similar disposition of glucose in muscle glycogen and lipids in adipose tissue [89]. These data suggested that IGF-1 effects are mediated by insulin receptors [89]. On the other hand, in normal rats [88] and depancreatized dogs [90] given doses of insulin and IGF-1 equivalent in their glucose-disposal effects, IGF-1 produced smaller effects on adipose tissue lipolysis and hepatic glucose production than did insulin. In addition, in the rat, IGF-1 was more selective than insulin in stimulating glycogen synthesis in the liver, skeletal, and heart muscle [88], suggesting that IGF-1 had no stimulatory effect on oxidative glucose disposal. Because only a single dose of IGF-1 and insulin was used by both groups of investigators [88, 90], no conclusions about the effect of dose specificity on the observed phenomena could be made.

The effects of IGF-1 administration in 90% partially pancreatomized rats were compared with those in sham-operated controls [91]. In the control rats, infusions of IGF-1 in doses about 30 times higher than insulin produced an equal effect in total body glucose disposal during euglycemic clamping. When the same protocol was used in partially pancreatomized rats, insulin-mediated glucose disposal was decreased, although IGF-1-mediated glucose disposal was similar [91]. In control rats, no additive effect of the two hormones was observed, although in the pancreatomized rats, IGF-1 co-infused with insulin restored glucose use to normal [84]. Similarly, in the diabetic BB rat (model of insulin-dependent diabetes), the metabolic actions of insulin but not those of IGF-1 were decreased [92]. In another model of insulin resistance (obese Zucker rats) IGF-1 could not overcome this defect [93]. Resistance to the hypoglycemic action of IGF-1 was also observed in obese mice [94, 95].

Obese Zucker rats and mice are hyperinsulinemic, whereas endogenous insulin levels are decreased in partially pancreatomized and diabetic BB rats. As noted above, insulin produces substantial changes in the IGFBPs. Thus, marked differences in IGFBP-1 levels can be expected in relation to insulin concentrations. Expression of IGFBP-1 and IGFBP-2 is markedly increased in the liver and kidney in rats with diabetes induced by the selective destruction of pancreatic ß cells by streptozotocin [96]. Further, in diabetic animals, the expression of a protein corresponding in mass to IGFBP-3 is decreased in the liver [96]. Changes in binding proteins may augment the hypoglycemic action of IGF-1. It is possible that a postreceptor defect in insulin action is located in a signal pathway common to insulin and IGF-1 in hyperinsulinemic rats and mice but not in insulinopenic diabetic rats and dogs.

Metabolic Effects of Insulin-like Growth Factor-1 in Humans

Bolus intravenous injections of 13.3 nmol (100 µg)/kg of IGF-1 or 1 nmol (0.15 IU)/kg of insulin produce nearly identical hypoglycemic responses in humans [97]. Both hormones produce the same depression of fatty acids, but decreases last substantially longer after insulin injection. A profound and prolonged hypoglycemic effect of 75 µg/kg of IGF-1 given as an intravenous bolus occurs in patients with Laron-type dwarfism [98]. Fasting hypoglycemia was produced during an 11-day constant subcutaneous infusion of 16 to 24 µg/kg of IGF-1 in a 9-year-old Laron dwarf [98]. In contrast, a subcutaneous infusion of either 7 or 14 µg/kg of IGF-1, begun 24 hours before and continued during oral glucose and meal tolerance tests in normal adult volunteers, did not change fasting plasma glucose levels or glucose tolerance. Similar glycemic profiles, however, were achieved with IGF-1 infusions at substantially lower insulin and C-peptide levels. Levels of growth hormone were substantially depressed by IGF-1. Thus, IGF-1 administration may improve insulin sensitivity [33] by suppression of insulin and growth hormone secretion. However, enhanced insulin sensitivity after IGF-1 administration is only a speculative conclusion, because no data were presented indicating improvement in insulin action that could be clearly distinguished from the summation of hypoglycemic effects of IGF-1 and insulin. In Laron dwarfs, decreases in insulin levels during administration of IGF-1 were associated with postprandial hyperglycemia, probably due to a decreased insulin secretory response [98].

When glucose concentrations were clamped within the normal fasting range, IGF-1 infusions in the range of 5 to 30 µg/kg per hour caused a dose-response decrease in the levels of insulin, fatty acids, triglycerides, ß-hydroxybutyrate, and leucine; the infusions increased exogenous glucose consumption [99]. Infusions of IGF-1 also had a sparing effect on protein catabolism as determined by the rates of leucine turnover and oxidation [93]. These investigators concluded that IGF-1 had metabolic effects qualitatively similar to those of insulin [99]. In a similar manner, when a primed (20 µg/kg bolus), constant (24 µg/kg per hour) infusion of IGF-1 with euglycemic clamping is used, IGF-1 behaves like exogenous insulin by decreasing C-peptide levels, by inhibiting hepatic glucose production, by stimulating peripheral glucose uptake, and by inhibiting lipolysis and protein catabolism [94]. Like insulin, it also stimulates oxidation of glucose and fatty acids. The close similarity in IGF-1 and insulin actions indicates that IGF-1 acts on the insulin receptor or that the metabolic effects mediated by activation of the IGF receptor occur through a similar cascade of intercellular events [100]. Moreover, exogenous IGF-1 not only decreases plasma levels of insulin and C-peptide but to a lesser degree also depresses glucagon levels, probably because of the direct inhibitory effect of IGF-1 on pancreatic and ß cells [99, 100].

Although less potent than insulin, IGF-1 can produce hypoglycemia, an effect that limits its application in reversing the catabolic state or stimulating linear growth [58]. The hypoglycemia induced by a bolus intravenous injection of IGF-1 produces the same counter-regulatory responses of glucagon, growth hormone, cortisol, and catecholamines seen with insulin-induced hypoglycemia [97]. During hypoglycemia induced by primed, constant infusions of IGF-1, the secretory responses of glucagon and growth hormone were suppressed. However, the secretory response of cortisol was normal and that of catecholamines was increased, when compared with hormonal counter-regulation during hypoglycemia of an identical rate of fall and nadir as that induced by a primed, constant insulin infusion [101]. After IGF-1 administration, glucose recovery was more sluggish than after insulin, because endogenous glucose production does not increase [101]. When infusions of IGF-1 and insulin were stopped, the insulin levels returned promptly to baseline values, although IGF-1 levels remained increased. When hypoglycemic clamp studies were done with a primed, constant IGF-1 infusion, only a suppression of glucagon response was observed, although other counter-regulatory hormones responded in a similar manner to those seen during hypoglycemic hyperinsulinemic clamping experiments [101].

Studies in nondiabetic humans show that some of the metabolic effects of IGF-1 differ from those seen in studies of animals (in vitro and in vivo). In this respect, IGF-1 in humans seems to have a more profound effect on endogenous glucose production and fatty acid levels, indicating substantial action on the liver and adipose tissue. Glucose oxidation in humans may also be stimulated by IGF-1. The hypothesis that the effects of IGF-1 are mediated by insulin receptors is plausible but may not be entirely correct. More studies are needed to substantiate the role of IGF-1 receptors in the liver; the physiologic significance of IGF-1 stimulation of lipoprotein lipase in adipose tissue; and the consequences of depressed insulin, glucagon, and growth hormone levels on the rate of hepatic glucose production and on lipolysis in adipose tissue. Despite these questions, the results of studies in nondiabetic persons may justify trials of IGF-1 in clinical situations in which profound insulin resistance cannot be overcome by insulin.

In this respect, IGF-1 has a glucose-lowering effect in patients with type A insulin resistance [9]. This syndrome is encountered in women and is characterized by acanthosis nigricans, virilization, and extreme insulin resistance due to a structurally defective insulin receptor [102], although the IGF type I receptor appears to be intact. Use of IGF-1 (in two 100 µg/kg bolus injections) produced a slow hypoglycemic response (> 6 hours) with nadirs of 3.2 to 5.5 mmol/L and a decrease in the markedly increased insulin and C-peptide levels (Figure 2). In these patients, IGF-1, probably acting on its own receptors, was able to overcome the insulin-receptor defect [9]. A person with a severe form of insulin resistance frequently did not respond to intravenously administered insulin in doses as large as 3000 units per hour [10]. In this patient, substitution of large doses (100- to 500-µgrams bolus injections) of IGF-1 for insulin produced a prompt normalization of serum glucose levels that lasted as long as the total IGF-1 concentration in the serum was more than 1100 µg/L (0.14 µmol/L, Figure 3. The described effects of IGF-1 were probably mediated through an IGF-receptor-signaling pathway that bypassed the insulin receptor defect [10].

View larger version (24K): [in this window] [in a new window]   Figure 2. Effect of insulin-like growth factor-1 in syndromes of extreme insulin resistance. Recombinant human insulin-like growth factor-1 (IGF-1) was administered in two (at 0 and at 120 min) 100 µg/kg intravenous bolus injections. The effect of IGF-1 on plasma glucose, insulin, and C-peptide levels is shown in three patients with acanthosis nigricans and insulin resistance. Bars represent values from each patient, black squares represent means. Adapted from Schoenle and colleagues [9] with permission.

View larger version (38K): [in this window] [in a new window]   Figure 3. Effect of insulin-like growth factor-1 administration in a patient with insulin-dependent diabetes mellitus and severe insulin resistance. An intravenous bolus dose of insulin-like growth factor-1 (IGF-1), 500 µg/kg, was given at 1530 hours. The hatched area indicates the period in which the serum IGF-1 concentration was more than 1100 µg/L and the serum free insulin concentrations were not detectable and thus the period when changes to serum glucose could be attributed to IGF-1. Adapted from Usala and colleagues [10], with permission.

Zenobi and colleagues [11] examined the metabolic effects of IGF-1 in type II diabetic patients. Patients received two daily, subcutaneous injections of IGF-1, 120 µg/kg, for 5 consecutive days. The treatment substantially decreased the hyperglycemic response to mixed meals and decreased fasting and postprandial insulin and C-peptide levels. A decrease in growth hormone and total triglyceride levels was also noted. The effects persisted for 3 days after IGF-1 administration was stopped. Based on these findings, the investigators concluded that IGF-1 administration improves insulin sensitivity and lipid metabolism in patients with type II diabetes [11] (Figure 4).

View larger version (38K): [in this window] [in a new window]   Figure 4. Effects of recombinant human insulin-like growth factor-1 in type II diabetic patients. Fasting glucose, insulin, C-peptide, and triglyceride levels in eight type II diabetic patients before (coarsely hatched bars, control), during (white bars), and after (grey bars) treatment with two subcutaneous doses of IGF-1 daily (120 µg/kg). *Statistically (P < 0.05) different when compared with control (days 1 to 5). Adapted from Zenobi and colleagues [11], with permission.

Hypertriglyceridemia is associated with insulin resistance and is commonly present in poorly controlled, type II diabetic patients [103-105]. Sensible insulin therapy in patients with type II diabetes not only substantially improves glycemic control and insulin action but also decreases hypertriglyceridemia [103-105]. Because a reduction in hypertriglyceridemia can also be achieved by using IGF-1, without increasing but actually decreasing insulin levels, it is still not clear whether an increase in IGF-1 levels is a better therapeutic option.

Indications for and Possible Hazards of Therapy Using Insulin-like Growth Factor-1

In most patients with diabetes, standard therapeutic approaches with diet, exercise, and insulin normally suffice. In those rare situations in which exogenous insulin is not effective because of nonfunctional or deficient insulin receptors, IGF-1 may be the therapy of choice. Short-term IGF-1 administration results in substantial metabolic improvement and can save lives.

The safety and efficacy of the long-term substitution of IGF-1 for insulin remains unknown, and therapeutic decisions should be weighed against potential risks. Insulin-like growth factor-1 stimulates cell growth but does not discriminate between normal and tumor cells [106]. In fact, growth of certain neoplasms is apparently IGF-1 dependent [107]. Thus, the notion that "IGF-I may prove to be a safe growth factor since it is a differentiation factor as well" [108] may not be entirely correct. The growth of arterial smooth muscle cells, one of the fundamental components of atherogenesis, is stimulated by IGF-1 [109]. In addition, IGF-1 administration is associated with increases in kidney size, renal blood flow, and glomerular filtration rate in normal human volunteers [110]. Despite the increase in glomerular filtration rate and renal hypertrophy after IGF-1 administration, microalbuminuria does not develop in normal persons [110]. In patients with diabetes, the increased glomerular filtration rate and renal hypertrophy precede the development of microalbuminuria and the progression to diabetic nephropathy [111]. Insulin-like growth factor-1 binds to and increases glomerular mesangial cell proliferation [112, 113], one of the most consistently observed structural changes in the diabetic kidney [114]. Thus, IGF-1 may accelerate the progression of diabetic nephropathy. Moreover, data [115-117] suggest that in certain patients with diabetes who have proliferative retinopathy, serum IGF-1 levels are increased. The proliferation of vascular endothelial cells is stimulated by IGF-1 in vitro [106, 118]. This process may stimulate both the healing of intimal lesions, a beneficial effect, and the proliferation of microvessels, an adverse effect [118]. Thus, IGF-1 administration may aggravate the progression of proliferative retinopathy.

Many reports suggest that IGF-1 may be a useful adjunct to the current therapy for diabetes and may even be the drug of choice in some metabolic emergencies. However, more data are needed to provide specific practical recommendations for long-term IGF-1 use. In addition, more information about the complex kinetic mechanisms of IGF-1 would help in devising systems in which the biologic effects of IGF-1 are more selective.

Abbreviations

IGF-1: insulin-like growth factor-1

IGFBP: insulin-like growth factor binding protein