This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jain, S. Articles by Geffner, M. E. Search for Related Content PubMed PubMed Citation Articles by Jain, S. Articles by Geffner, M. E.

Insulin-Like Growth Factor-I Resistance¹

Memorial Sloan-Kettering Cancer Center and Cornell University Medical College (D.W.G.), New York, New York; Division of Epidemiology and Department of Anthropology (R.B.), The University of Illinois at Chicago, Chicago, Illinois; and Division of Endocrinology and Metabolism (M.E.G.), UCLA Children’s Hospital, Los Angeles, California 90095-1752

	Abstract

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 

I. Introduction

II. The IGFs and Their Receptors

III. Genetically Engineered Mouse Models

IV. Malnutrition

V. Conditions Associated with Genetic IGF-I Resistance in Humans

A. African Efe Pygmies

B. Leprechaunism and other genetic disorders associated with severe insulin resistance

C. Deletions of the distal arm of chromosome 15

D. Rare examples of "idiopathic" short stature associated with elevated circulating levels of IGF-I

VI. Conditions Associated with Acquired IGF-I Resistance in Humans

A. HIV-1 infection

B. Chronic renal insufficiency

C. Mauriac syndrome

VII. Conclusions

	I. Introduction

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 
INSULIN-LIKE growth factor type I (IGF-I) resistance as a cause of short stature in human beings has not been widely recognized. With the wider applicability of techniques of molecular biology and a greater understanding of the components of the human "GH-IGF-I axis," new genetic abnormalities associated with growth retardation have been identified, involving the proximal portion of this axis, as well as IGF-I itself (1, 2). Studying the clinical syndrome of IGF-I resistance has further contributed to the evolution of our knowledge regarding this complex system. This article reviews our current understanding of the GH-IGF-I axis and discusses animal models and human conditions associated with IGF-I resistance.

	II. The IGFs and Their Receptors

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 
The existence of IGFs (somatomedins) was first postulated in the 1950s when Salmon and Daughaday (3) demonstrated the existence of a GH-dependent sulfation factor. Concurrent research had identified nonsuppressible insulin-like activity in normal serum (NSILA) (4). A third line of investigation described multiplication-stimulating activity (MSA), a mitogenic factor produced by fetal Buffalo rat liver cells that had the mitogenic and metabolic characteristics of sulfation factor and NSILA (5). In the mid-1970s, two candidate somatomedins were purified (6, 7). Later, they were shown to resemble proinsulin and were thus renamed insulin-like growth factors (IGFs) (8).

IGF-I is a 70-amino acid peptide that has 50% sequence homology with insulin and 70% homology with IGF-II (8). Both IGF-I and IGF-II are encoded by single large genes, IGF-I on the long arm of chromosome 12, and IGF-II on the short arm of chromosome 11. Like insulin, both IGFs have A and B chains that are linked by disulfide bonds. Although the liver is the major source of the IGFs, there is great variability of tissue expression at various stages of development.

Most of the actions of IGF-I and IGF-II are mediated by the type 1 IGF receptor (IGF1R). For a comprehensive review of the molecular and cellular biology of the IGF1R, the reader is referred to the 1995 Endocrine Reviews article by LeRoith et al. (9). Like the insulin receptor, with which it has a high degree of homology, the IGF1R is a member of the tyrosine-kinase class of cell-surface receptors (10). It consists of two ß-half-receptors, linked by disulfide bonds between the -subunits to form the mature ₂ß₂-heterotetrameric holoreceptor. The -subunit, containing cysteine-rich regions, contains the site of IGF-I binding while the intracellular region of the ß-subunit possesses tyrosine kinase activity. Multiple forms of the - and ß-subunits have been described in various tissues and cell lines, the etiology and functions of which are unclear (9). The IGF1R binds insulin and IGF-II with 100- to 1000-fold and 2- to 15-fold lower affinity, respectively, than IGF-I.

The human IGF1R cDNA has been expressed (10) and the 5'-flanking region of the gene has been cloned and characterized (11). There are several atypical features, including an unusually long 5'-untranslated region (>1 kb from the ATG). With a numbering system starting with the transcription start site designated "0," the putative promoter region is -1 to -517. The sequence is 72% GC-rich and lacks identifiable TATA and CAAT boxes. There is an initiator motif at -6 to +11, which contains the single transcription start site. There is also a consensus sequence for the transcription factor AP-2 and five sites for Sp1 (GC boxes), as well as a sequence that is weakly homologous to the binding site of an epidermal growth factor (EGF) receptor transcription factor. Deletion analysis of the promoter and upstream region indicates that a HindII-SstI fragment (-517 to +439) has significant promoter activity (11). Studies of the rat IGF1R promoter demonstrate that Sp1, a ubiquitous member of the family of zinc-finger transcription factors, can regulate expression of the IGF1R gene by acting both on GC boxes in the 5'-flanking region of the promoter and on a CT element (a homopurine/homopyrimidine motif) in the 5'-untranslated region (12). Mutations or deletions of the promoter sequence could result in reduced IGF1R gene expression. Nonsense mutations or stop codons in the coding sequence could result in decreased IGF1R numbers. These alterations in gene structure could diminish the binding of IGF-I to its receptor, thereby causing IGF-I resistance.

Upon ligand binding, the receptor autophosphorylates and initiates a series of cytoplasmic kinase activations which, in turn, activate nuclear transcription factors. A defect in IGF-I action would result if proper ligand binding did not occur. This could happen if either IGF-I or the IGF-1R were altered (as in the case of a mutation in the coding sequence or if autoantibodies against either component interfered with normal binding). As a result of the close homology between IGF-I and insulin receptors, there is formation of hybrid insulin/IGF-I receptors. Such hybrids have been isolated from various tissues and cultured cell lines and may be common in vivo (9). These hybrids bind both IGF-I and insulin with affinities comparable to pure IGF1R (13, 14). Although the IGF-I half-receptor can bind ligand, dimerization is necessary for trans-autophosphorylation and cellular response. Thus, the presence of kinase-defective insulin receptors could result in relative IGF-I resistance, especially if an increased number of defective insulin receptors compete with normal IGF1Rs for dimerization.

IGF1R action involves components of several intracellular signaling pathways (9). Ligand binding induces IGF1R autophosphorylation, resulting in activation of intrinsic tyrosine-kinase activity. The predominant substrate of IGF1R is the insulin receptor substrate-1 (IRS-1), a "docking" protein that can bring together and regulate the activity of some intracellular proteins (15). IRS-1 activates Ras via a Grb2-mSos signaling pathway. Ras, in turn, activates a cascade of protein kinases leading to mitogen-activated protein (MAP) kinase (16, 17). MAP kinases can regulate a variety of cellular and nuclear proteins, including transcription factors. IRS-1 also interacts with Grb-S, Nck, and Syp, which are other proteins involved in growth-factor signaling pathways (9). The phosphorylation cascade activated by the IGF1R tyrosine kinase also activates Shc, which associates with Grb2 and subsequently activates Ras. Other less well characterized substrates of the IGF1R include a cytoskeletal protein, other MAP kinases, and some nuclear protein products such as c-jun. These may be direct or indirect targets. IGF-I signaling may be negatively modulated by the protein kinase A signaling pathway and by CD45 phosphotyrosine phosphatase. Disruption of an intracellular second-messenger pathway (i.e., by a mutation of the gene encoding a key component) could result in resistance to IGF-I and other growth factors whose receptors share the pathway for signal transduction. Alternatively, resistance to a hormone that competes for shared second messengers could cause secondary IGF-I resistance by shunting away necessary intermediaries.

The type 2 IGF receptor has no structural homology to the IGF1R or the insulin receptor. However, it is identical to the cation-independent mannose-6-phosphate receptor and is thus referred to as the IGF-II/Man-6-P receptor (18). The human IGF-II/Man-6-P receptor gene is located on chromosome 12 (19) and encodes a single peptide chain (18). The receptor has different binding sites for IGF-II and Man-6-P or Man-6-P-containing glycoproteins. The receptor does not have tyrosine kinase activity. As yet, the mechanism of signal transduction has not been elucidated. The IGF-II/Man-6-P receptor also binds IGF-I with lesser affinity.

In plasma, IGFs are complexed with binding proteins (IGFBPs) that extend their serum half-life, transport them to target cells, and modulate their interactions with membrane receptors (20, 21, 22). Six IGFBP cDNAs have been cloned and sequenced (21, 22) with a seventh currently being characterized (23). They have significant sequence homology (24). IGFBP-3 is the primary serum IGFBP and is associated with a supporting protein, acid-labile subunit. Multiple studies have shown IGFBP-3 levels to be GH-dependent (25, 26). One study using transgenic mice supported IGFBP-3 being IGF-I-dependent (27). Mice that were GH deficient had a 15.7-fold decrease in serum IGF-I and a 5.5-fold decrease in serum IGFBP-3, whereas those GH-deficient mice that expressed IGF-I had serum levels of IGF-I and IGFBP-3 that were 69% and 64% of those in normal sera, respectively. In mice with IGF-I overexpression, serum IGFBP-3 was increased 2.9-fold. There continues to be debate as to whether IGFBP-3 is IGF-I-dependent in humans. Gargosky et al. (28) have described GH-receptor-deficient adults who have high GH levels, and low IGF-I, IGF-II, and IGFBP-3 levels. Administration of IGF-I to these patients failed to increase IGFBP-3 levels. It is not clear whether this finding can be generalized to normal human subjects.

True to the somatomedin hypothesis, most of the actions of GH are mediated by IGF peptides. Circulating GH stimulates hepatic IGF-I production, which is reflected in serum IGF-I levels. In a variety of tissues, there is also IGF-I production that may have autocrine and paracrine effects. Although GH and the IGFs also have a variety of actions that are independent of each other, the relative roles of GH and IGF-I are not completely understood. Although both IGF-I and GH stimulate growth and protein synthesis, the quantitative growth effect of GH is greater with differential effects on various organs (29). The metabolic effects of GH and IGF-I also differ somewhat in that the former is "diabetogenic" and the latter has blood glucose-lowering effects. IGF-I action in vitro includes effects on cell-cycle progression, proliferation, death, differentiation, and a variety of cell-specific functions (1).

	III. Genetically Engineered Mouse Models

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 
The development of embryonic stem cell technology has allowed investigators to study the effects of targeted removal of one or more genes from an animal. Mouse "knockout" studies have led to a greater understanding of the pre- and postnatal effects of IGF-I, IGF-II, and the IGF1R on growth and development. The IGF1R-deficient mouse provides a model of primary IGF-I resistance. Animals with this mutation, in conjunction with IGF-I or IGF-II gene deletion, allow further investigation into the interactions between IGFs and the IGF1R.

There are various strategies of gene targeting that involve homologous recombination between an introduced mutated gene and an endogenous chromosomal allele. Replacement vectors are constructed using subcloned genomic fragments of the gene of interest and transcriptionally competent neo and tk casettes (30, 31). The casettes replace critical coding regions of the gene and either produce multiple stop codons or abolish function of the protein product. Embryonic stem cells are transfected by electroporation and selected by single or double antibiotic resistance before they are expanded. After DNA analysis to identify clones with homologous recombination, approximately five to 15 clones are injected into blastocysts, and the embryos are replanted into pseudopregnant female mice. Chimeric animals that incorporated the embryonic stem cells into their germline transmit the mutant DNA to their offspring. Heterozygotes are then inbred to yield animals homozygous for the disrupted gene.

Animals that were heterozygous for the IGF1R mutation [IGF1R(-/-)] had no discernible difference in phenotype from wild-type animals (32). Those animals with complete absence of the IGF1R [IGF1R(-/-)] were significantly smaller at birth than were wild-type animals (32, 33, 34, 35). This is in contrast to the observation that absence of GH in mutant animals or experimental ablation of the pituitary gland does not affect prenatal growth (36, 37, 38, 39). Finally, the IGF1R(-/-) mice had severe growth retardation with approximately 45% of normal birth weight and had 100% neonatal mortality (32). The pups were born alive and made visible efforts to breathe but become cyanotic and died in a few minutes. Apparently, air never reached the alveoli, although histopathological examination revealed no particular abnormalities. There were no structural blockages, no abnormality of the alveolar epithelium, and no difference in immunostaining for surfactant apoprotein compared with wild-type animals. There was, however, significant generalized hypoplasia of all tissues, including respiratory muscles, which may explain the respiratory failure at birth. The skin was translucent with abnormally thin stratum spinosum. Examination of the nervous system revealed significantly increased cellular density in the mantle zone, thought to be secondary to a reduction in the surrounding neuropil (i.e., neuronal fibers and cytoplasm of neuroglial cells). The timing of the first appearance of several ossification centers was analyzed with the finding that IGF1R(-/-) animals had a delay of 2–4 embryonic days in development of ossification centers of cranial and facial bones.

Studies of double-mutant animals have provided further insight into the interaction of IGFs with IGF1R (32). The genotypes studied were IGF-I(-/-)/IGF-II(p-), IGF-I(-/-)/IGF1R(-/-), and IGF-II(p-)/IGF1R(-/-). The IGF-I(-/-)/IGF-II(p-) animals were homozygous for absence of IGF-I and carried a paternally derived mutant IGF-II. Because the IGF-II gene is paternally expressed, these animals are phenotypically indistinguishable from mice with absence of the IGF-II gene [IGF-II(-/-)] (34). The IGF-II(p-)/IGF1R(-/-) animals had absence of both IGF-II and IGF1R. The IGF-I(-/-)/IGF1R(-/-) animals were homozygous for absence of both IGF-I and IGF1R. All three genotypes had 100% neonatal mortality. The IGF-I(-/-)/IGF-II(p-) and IGF-II(p-)/IGF1R(-/-) mice had approximately 30% of normal birth weight and the IGF-I(-/-)/IGF1R(-/-) mice had approximately 45% of normal birth weight. The observation that IGF1R(-/-) and IGF-I(-/-)/IGF1R(-/-) had similar phenotypic abnormalities supports the conclusion that IGF-I interacts primarily through the IGF1R. There is indirect evidence that IGF-II interacts with IGF1R, as the IGF-I(-/-) phenotype with normal IGF-II and IGF1R was less severely affected than either the IGF1R(-/-) or the IGF-I(-/-)/IGF1R(-/-) phenotypes.

In summary, the major conclusion from studying animals with absence of IGF1R is that the IGFs, independent of GH, are essential for normal embryonic growth and possibly development in mice. The second significant conclusion from the study of IGF1R-deficient mice is that the IGF1R gene is an essential gene in mice. Viability of the IGF-II(p-) strains suggests that the critical element necessary for life in those animals is the interaction between IGF-I and the IGF1R. IGF-I interacts primarily with the IGF1R, but may also have in vivo interactions with insulin/IGF-I hybrid receptors. The variable survival of the animals with absence of IGF-I [IGF-I(-/-)] [5–68% vs. the 100% mortality seen with IGF1R(-/-)] indicates the presence of compensatory mechanisms, perhaps involving the action of IGF-II through the IGF1R.

	IV. Malnutrition

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 
Significant and prolonged malnutrition is characterized by diminished weight gain and impaired linear growth. Although the exact mechanism for the growth failure associated with malnutrition is unknown, roles for GH resistance, IGF-I deficiency, and IGF-I resistance have been promulgated based mostly on data derived from experiments performed on protein-restricted rats (40).

In studies of Sprague-Dawley rats either fasted for 3 days or fed a reduced protein diet, GH receptor and IGF-I mRNA levels were coordinately decreased in liver (41). Other studies of protein-restricted animals suggested maintenance of normal GH-receptor function with the emergence of a post-GH-receptor defect. In fasting rats, serum GH and IGF-I levels fall (42), whereas there are normal or increased serum GH and reduced IGF-I levels in human fasting, kwashiorkor, and marasmus (40). The reduced serum IGF-I levels reproducibly found in states of chronic human undernutrition have been linked to concomitant GH resistance (40), reduced amino acid availability (43), alterations in IGFBPs (increased IGFBP-1 and -2) (44), and enhanced serum clearance and degradation of IGF-I (40).

Initial studies of protein restriction in rats demonstrated a state of apparent GH resistance (45). The low serum IGF-I levels that ensued were not restored to normal after physiological GH replacement, which successfully normalized IGF-I levels in hypophysectomized rats. Moreover, these physiological GH doses, as well as a regimen of supraphysiological GH replacement that increased serum IGF-I levels 2-fold over baseline (but to levels still 50% of rats fed a normal protein diet), both failed to promote growth. Since IGF-I deficiency is readily apparent after 12 h of protein restriction, at which time there are only minimal effects upon GH receptors, the induced GH resistance is thought to be postreceptor in origin (46). The specific mediator of this phenomenon is unknown. One proposed mechanism involves the effects of amino acid restriction on hepatic protein synthesis, as reflected by a specific reduction in IGF-I mRNA synthesis (43).

The above studies suggest that insufficient dietary protein may block the actions of IGF-I in addition to those of GH. To test this hypothesis, Philipps et al. (47) administered recombinant IGF-I to neonatal rats who were intentionally malnourished by increasing litter size. After body weight reductions of 20–25%, IGF-I injections failed to stimulate any increase in somatic growth of these animals (47). Similar experiments were conducted by Thissen et al. (42) in which 4-week-old protein-restricted rats were infused with recombinant IGF-I for 1 week. Although this treatment regimen normalized serum IGF-I concentrations, carcass growth was not stimulated. Of note, the infusion of IGF-I into well nourished hypophysectomized rats at a 50% lower dose induced a significant increase in carcass growth. Furthermore, combined infusion of GH and IGF-I also failed to stimulate significant carcass growth (40). Analogous studies in obese humans subjected to dietary restriction and GH treatment for 11 weeks showed loss of an initially observed augmentation of nitrogen balance despite maintenance of persistent elevations in plasma IGF-I concentrations (48). The authors of this latter study concluded that the loss of the anabolic action of GH under these conditions may have resulted from the development of IGF-I resistance.

Taken collectively, the above observations indicate that protein malnutrition is associated with combined GH and IGF-I resistance. In addition, serum IGF-I levels are low despite IGF-I resistance. This apparent paradox may arise directly from coexisting GH resistance or result from the ability of undernutrition to override an increased serum IGF-I level, which might be predicted based on standard negative feedback, as usually occurs with ligand in the setting of reduced receptor activity and/or reduced downstream postreceptor function.

	V. Conditions Associated with Genetic IGF-I Resistance in Humans

Top Abstract I. Introduction II. The IGFs and... III. Genetically Engineered... IV. Malnutrition V. Conditions Associated with... VI. Conditions Associated with... VII. Conclusions References

 
A. African Efe Pygmies
The African Pygmies are a unique population of short-statured humans whose existence dates back to early Paleolithic times. Their stunted height should be thought of as a variation (or adaptation), rather than as a defect or abnormality, since being small is likely to be the result of an evolutionary adaptation for such populations, which have experienced, perhaps for thousands of years, cyclical undernutrition (i.e., subsisting in a hot, humid rain forest with low-lying, dense vegetation), and have not received significant gene flow from other populations (49, 50, 51). Early studies of Mbuti and Babinga Pygmies concluded that their prepubertal growth was similar to that of other Africans and that their short adult stature was due primarily, if not solely, to absence of an adolescent growth spurt (52, 53, 54). These studies, however, were carried out on subjects of estimated age and employed cross-sectional data. More recent investigations in which African Pygmy children from the Efe tribe of nomadic hunter-gatherers in northeast Zaire were followed from birth longitudinally to 5 yr of age noted that the growth retardation of Pygmy children is present at birth and, during the first 5 yr of life, becomes progressively greater (55, 56). The Efe children show a trend of increasingly negative mean SD score compared with an American reference population, from -2.71 (SD = 0.93) at 6 months to -4.16 (SD = 0.46) at age 5 yr and Efe children’s stature declines steadily in comparison to other African rural populations (56). The resultant mean adult stature of male Efe Pygmies (143 cm) is the shortest of any human population known, including other Pygmy populations of Africa (55).

The exact mechanism of short stature in African Pygmies is unknown although early evidence implicated resistance to the growth-promoting action of GH as the cause (57) on the basis of 1) low serum IGF-I levels in adult Pygmies (58); 2) failure of short-term GH administration in the field to Pygmies to induce the expected rise in serum IGF-I levels, nitrogen retention, increased insulin secretion, increased urinary excretion of calcium, and decreased urinary excretion of phosphate (59); 3) reduced secretion of IGF-I by Epstein-Barr-virus (EBV)-transformed B lymphocyte cell lines established from six adult African Pygmies in response to in vitro GH stimulation compared with that secreted by control cell lines (60); 4) low serum levels of high-affinity GHBP (61, 62), a fragment of the extracellular portion of the GH receptor which is presumed to be proportionate to the number of functioning tissue-bound GH receptors in Pygmies (63); and 5) a restriction fragment length polymorphism present in 0.17–0.33% of GH receptor genes in two Pygmy populations, but present in only 0.02% of nonPygmy Africans (57).

Despite these observations leading to the notion that GH resistance was the underlying mechanism responsible for the short stature of the African Pygmy, it should be noted that African Pygmies bear no particular morphological or biochemical resemblance to individuals with genetically mediated GH resistance due to proven mutations in the GH receptor gene, i.e., Laron dwarfs (64), other than that both groups are very short. It seemed reasonable, therefore, to explore other mechanisms for Pygmy short stature, specifically IGF-I resistance, since GH resistance, at least in theory, could be an epiphenomenon resulting from underlying IGF-I resistance (65). At the same time, it became important to find alternative explanations for the aforementioned evidence linking GH resistance to stunted Pygmy growth that would also be consistent with our own hypothesis. For example, failure of short-term GH treatment to induce expected metabolic changes in adult Pygmies (59) could be explicable by the requirement of IGF-I to mediate these actions of GH. Diminished proliferation of Pygmy EBV-transformed B lymphocyte cell lines in response to GH (60) could be explained by resistance to, along with blunted secretion of, IGF-I compared to normal. Although the low serum levels of high-affinity GHBP reported in a sample of African Pygmies (61, 62) suggested that these subjects have at least a partial deficiency of GH receptors (63), it is possible that unintended inaccuracies in estimating age (56, 66) and nutritional repletion account for observed differences in serum GHBP levels between Pygmies and controls. Furthermore, there is significant overlap between GHBP levels in Pygmies and controls (62). Despite the disproportionate representation of the aforementioned restriction fragment length polymorphism in the GH-receptor gene in Pygmies (57), more recent studies have found a normal sequence of the GH receptor gene, including its 3'- and 5'-extensions, in these Pygmies (62).

With the development of HTLV T lymphocyte and EBV B lymphocyte immortalization technologies, we have been able to acquire hormonally responsive cell populations that reflect the genetic complement of the individuals from whom they arose. This technology has demonstrated fidelity insofar as detection of genetically mediated resistance to GH in four patients with Laron dwarfism and to insulin in a patient with leprechaunism (67, 68). We thus sought to establish cell lines from Efe Pygmies as a means to confirm or refute the presence of GH resistance and also to quantify responsiveness to IGF-I. From two separate expeditions to Zaire, we were able to establish eight T- and three B cell lines from adult Efe Pygmies as well as three T cell lines from neighboring Lese farmers. It should be pointed out that there is known admixture between the Lese farmers and Efe Pygmies, which leads to maternal transmission of Efe genes into the Lese gene pool (69) and that the heights of the Lese are intermediate between those of the Efe and urban black Africans.

With this background, we performed in vitro clonal proliferation (clonogenic) assays in which T cell lines (70, 71) from Pygmy, Lese, and American controls were incubated with recombinant human IGF-I (in a concentration range of 7–250 ng/ml); recombinant human GH (25–500 ng/ml); recombinant human insulin (7–250 ng/ml); or PBS at pH 7.4. After 7–10 days, T cell colonies containing a minimum of eight cells were enumerated using an inverted microscope. For studies with B cell lines (72), proliferative responses of Pygmy and control cell lines were compared after stimulation with recombinant human IGF-I (5–250 ng/ml); the phorbol ester, phorbol 12-myristate 13-acetate (10–100 nmol/liter); or PBS at pH 7.4. After 4–7 days, B cell colonies containing a minimum of seven cells were counted.

In these studies (70, 71), T cell lines from eight Efe Pygmies were found to be completely resistant to the growth-promoting actions of IGF-I concentrations as high as 250 ng/ml and GH concentrations as high as 500 ng/ml, whereas control cell lines demonstrated a typical bimodal response to IGF-I (Fig. 1A) and unimodal response to GH (Fig. 1B). Interestingly, T cell lines from neighboring Lese farmers responded to both IGF-I and GH in an intermediate manner with respect to the Pygmies and American controls, consistent with their aforementioned genetic admixture. Furthermore, the observed resistance to IGF-I was not related to the local secretion of IGFBPs as the T cell line of one Pygmy so studied showed no clonal proliferation in response to the IGF-I analog, [Q³,A⁴,Y¹⁵,L¹⁶]IGF-I, which has reduced affinity for IGFBPs (70).

View larger version (25K): [in this window] [in a new window]   Figure 1. IGF-I, GH, and insulin responsiveness of T cell lines from Pygmies, Lese, and American controls. The unstimulated number of T lymphoblast colonies (referred to, on the ordinate, as colonies) formed in incubation mixtures without added growth factor is defined as 100%. The abscissa represents the concentrations of added growth factor (ng/ml). The data are presented as the mean ± SE. A, IGF-I (upper panel). The Pygmy T cell lines showed no clonal responsiveness to IGF-I concentrations as high as 250 ng/ml. The responses of the Lese T cell lines to IGF-I were intermediate between those of the Pygmies and American controls. B, GH (middle panel). The Pygmy T cell lines showed no clonal responsiveness to GH concentrations as high as 500 ng/mL. The responses of the Lese T cell lines to GH were intermediate between those of the Pygmies and American controls. C, Insulin (lower panel). There were no statistically significant differences between overall insulin response curves of the Pygmy, Lese, and American control T cell lines. [Reproduced with slight modifications with permission from M. E. Geffner et al.: J Clin Endocrinol Metab 80:3732–3738, 1995 (71 ). © The Endocrine Society.]

To expand upon our finding of IGF-I resistance, we next chose to verify that this unique finding was present in a second tissue, namely, B lymphocytes (72). Indeed, the Efe B cell lines showed essentially no responsiveness to the entire IGF-I concentration range, whereas control B cell lines demonstrated a typical bimodal stimulation of 50% above baseline for each peak (Fig. 2). In contrast, Pygmy and control B cell lines responded with comparable stimulation to all concentrations of phorbol 12-myristate 13-acetate, a phorbol ester which does not activate the IGF1R (75), thus confirming the specificity of the observed resistance to IGF-I.

View larger version (20K): [in this window] [in a new window]   Figure 2. IGF-I responsiveness of B cell lines from Pygmies and American controls. The unstimulated number of B lymphoblast colonies (referred to, on the ordinate, as colonies) formed in incubation mixtures without added growth factor is defined as 100%. The abscissa represents the concentrations of added IGF-I (ng/ml). The data are presented as the mean ± SE. The Pygmy B cell lines showed essentially no clonal responsiveness to IGF-I concentrations as high as 250 ng/ml, whereas the American control T cell lines had a typical bimodal response to IGF-I. [Reproduced with slight modifications with permission from A. B. Cortez et al.: Biochem Mol Med 58:31–36, 1996 (72 ).]

We next undertook three additional sets of experiments to expand upon our observation of IGF-I resistance in Pygmy tissues. Our first approach centered around the ability of GH to induce resistance to the in vitro growth-promoting action of insulin. Accordingly, we performed studies in which American control, Lese, and Pygmy T cell lines were stimulated with insulin alone and after preincubation with either GH or IGF-I (82). Preincubation with GH induced subsequent resistance to the mitogenic action of insulin with T cell lines from multiple American controls (Fig. 3A) and an African Lese (Fig. 3B), but failed to reduce the proliferative response of a Pygmy T cell line after insulin stimulation (Fig. 3C). A similar pattern of response to insulin of each group was found after IGF-I preincubation (Fig. 4, A–C). Furthermore, the ability of either GH (Fig. 5A) or IGF-I (Fig. 5B) to induce insulin resistance in American control and Lese T cell lines was completely abrogated by preincubation with IR-3 antibody against the IGF1R. These observations suggest that the mechanism of GH-induced resistance to the mitogenic action of insulin is mediated by IGF-I in T cell lines of normal subjects, and that failure of either GH or IGF-I to induce insulin resistance in the Pygmy cell line is due to its inability to respond to IGF-I. This interpretation is consistent with in vivo studies showing that resistance to the metabolic action of insulin in acromegaly is more closely correlated to circulating IGF-I than to GH levels (83).

View larger version (25K): [in this window] [in a new window]   Figure 3. Insulin responsiveness of T cell lines from American controls, Lese, and Pygmies after GH (50 or 100 ng/ml) preincubation. The unstimulated number of T lymphoblast colonies (referred to, on the ordinate, as colonies) formed in incubation mixtures without added insulin is defined as 100%. The abscissa represents the concentrations of added insulin (ng/ml). The data are presented as the mean ± SE. A, American control T cell lines (upper panel). Clonal responsiveness to insulin was completely abrogated after preincubation with GH. B, African Lese control T cell line (middle panel). Clonal responsiveness to insulin was also completely blocked after preincubation with GH. C, Pygmy T cell line (lower panel). Clonal responsiveness to insulin was not significantly reduced after preincubation with GH. [Reproduced with slight modifications with permission from M. E. Geffner et al.: Diabetes 43:68–72, 1994 (82 ).]

View larger version (25K): [in this window] [in a new window]   Figure 4. Insulin responsiveness of T cell lines from Pygmies, Lese, and American controls after IGF-1 (8 ng/ml) preincubation. The unstimulated number of T lymphoblast colonies (referred to, on the ordinate, as colonies) formed in incubation mixtures without added insulin is defined as 100%. The abscissa represents the concentrations of added insulin (ng/ml). The data are presented as the mean ± SE. A, American control T cell lines (upper panel). Clonal responsiveness to insulin was completely abrogated after preincubation with IGF-I. B, African Lese control T cell line (middle panel). Clonal responsiveness to insulin was also completely blocked after preincubation with IGF-I. C, Pygmy T cell line (lower panel). Clonal responsiveness to insulin was not significantly reduced after preincubation with IGF-I. [Reproduced with slight modifications with permission from M. E. Geffner et al.: Diabetes 43:68–72, 1994 (82 ).]

View larger version (35K): [in this window] [in a new window]   Figure 5. Insulin responsiveness of normal T cell lines after IGF-I receptor antibody [IR-3 (5 x 10-4 g/liter)] and either GH or IGF-I preincubation. The unstimulated number of T lymphoblast colonies (referred to, on the ordinate, as colonies) formed in incubation mixtures without added insulin is defined as 100%. The abscissa represents the concentrations of added insulin (ng/ml). The data are presented as the mean ± SE. IR-3 is a monclonal antibody that blocks the IGF1R. A, GH preincubation (upper panel). GH significantly reduced clonal responsiveness of both American and African Lese control T cell lines to all concentrations of insulin. Under conditions of combined preincubation with GH and IR-3, the profile of clonal responsiveness to insulin normalized. B, IGF-I preincubation (lower panel). IGF-I significantly reduced clonal responsiveness to all concentrations of insulin. Under conditions of combined preincubation with IGF-I and IR-3, the profile of clonal responsiveness to insulin normalized. [Reproduced with slight modifications with permission from M. E. Geffner et al.: Diabetes 43:68–72, 1994 (82 ).]

Our third set of validation experiments involved the quantification of T cell line proliferation in response to estradiol, testosterone, 1,25-(OH)₂-vitamin D₃, and T₃ (89). We chose these representative steroid and thyroid hormones without certainty as to whether their growth-promoting actions upon T cells were mediated by local IGF-I production. We quantified colony formation of American control T cell lines in the presence and absence of IR-3 and Pygmy T cell lines in response to physiological and supraphysiological concentrations of estradiol, testosterone, 1,25-(OH)₂-vitamin D₃, and T₃. There were no statistically significant differences in response curves for any of the four hormones comparing control clonal responses in the presence or absence of IR-3, and no statistically significant difference in responsiveness between control and Pygmy T cell lines. From these data, we conclude that 1) normal T cell lines proliferate in response to estradiol, testosterone, 1,25-(OH)₂-vitamin D₃, and T₃, hormones that bind to cytoplasmic/nuclear receptors; 2) these responses are not mediated through local IGF-I since they are not blocked by pretreatment with antibody to the IGF1R; and 3) Pygmy T cell lines, which are genetically resistant to IGF-I, grow equivalently to control T cell lines in response to estradiol, testosterone, 1,25-(OH)₂-vitamin D₃, and T₃, as would be predicted, since the growth-promoting actions of these hormones are IGF-I-independent in our system.

Armed with the above information pointing toward IGF-I resistance as the basis for short stature in the Efe Pygmy, we next undertook experiments to determine the molecular basis for our observation of reduced in vitro clonal responsiveness of T cell lines to IGF-I (90). In so doing, we found markedly decreased cell-surface expression of IGF1Rs on T cell lines established from Efe Pygmies (confirmed by affinity cross-linking) (Fig. 6), with normal IGF-I binding affinity compared with controls. Furthermore, the Pygmy IGF1Rs were not autophosphorylated in an in vitro kinase assay (Fig. 7). In addition, in response to stimulation by physiological concentrations of IGF-I, the Pygmy IGF1Rs did not transmit downstream signals, as measured by protein tyrosine phosphorylation quantified by Western blot analysis and inducibility (above baseline) of the early response genes, c-jun and c-fos, quantified by Northern blot analysis (Fig. 8). On Northern blotting, there was a significantly reduced (2–13% of control) steady-state level of IGF1R mRNA (Fig. 9). This represents decreased production of the mRNA since, in an RNase protection assay, there was equivalent stability of the IGF1R transcripts from Pygmy and American control cells after actinomycin D treatment. The nucleotide sequence of the full-length IGF1R cDNA in one Pygmy subject showed no significant variation from normal. Overall, four polymorphisms were found in Pygmy cDNA, although none were found in cDNA from all Pygmy subjects and not in control cDNA. Thus, no reproducible genetic alteration in the protein-coding sequence of the IGF1R was detected.

View larger version (73K): [in this window] [in a new window]   Figure 6. Receptor cross-linking with 125I-labeled IGF-I. American control cells (J-RN) and Pygmy 1 were incubated with 125I-labeled IGF-I in the presence and absence of excess unlabeled IGF-I and cross-linked with disuccinimidyl suberate. Cross-linked complexes were resolved by SDS-PAGE. Arrows indicate the 135-kDa receptor protein cross-linked with 125I[IGF-I]. Numbers on the left indicate the positions of molecular mass standards (in kilodaltons). [Reproduced by permission from Y. Hattori et al.: J Clin Endocrinol Metab 81:2257–2263, 1996 (90 ). © The Endocrine Society.]

View larger version (56K): [in this window] [in a new window]   Figure 8. IGF-I-induced signal transduction. A, Phosphorylation of intracellular proteins (upper panel). Control 1 (J-RN) and Pygmy 1 cells were incubated with different concentrations of IGF-I or with 584 ng/ml of insulin for 5 min at 37 C. Protein tyrosine phosphorylation was detected by Western immunoblot analysis with an antiphosphotyrosine monoclonal antibody. A 160-kDa phosphorylated protein, perhaps representing IRS-1, was detected and is indicated by the arrow. Numbers on the left indicate the positions of molecular mass markers (in kilodaltons). B, Induction of c-fos and c-jun mRNA (lower panel); 2.2-kb c-fos and 2.7-kb c-jun bands were visualized by Northern blot analysis. The mRNA levels were quantitated by PhosphorImager, and the mRNA level without IGF-I stimulation of each cell was defined as 1.0. [Reproduced by permission from Y. Hattori et al.: J Clin Endocrinol Metab81:2257–2263, 1996 (90 ). © The Endocrine Society.]

View larger version (49K): [in this window] [in a new window]   Figure 9. IGF1R mRNA expression. A, Northern blot analysis of IGF1R (upper panel). mRNA levels were quantitated by scoring each band by PhosphorImager and cumulating the values. Numbers on the left indicate the positions of molecular mass markers (in kilobases). The scores indicated below the figure show IGF1R mRNA levels adjusted to that of ß-actin and calculated by considering the value for Pygmy 1 cells as 1.0. B, Northern blot analysis showing selectively decreased expression of the IGF1R gene in Pygmy cells (lower panel). The HTLV-II-transformed Laron dwarf T cell line was established from a subject with Laron dwarfism (68 ). Expression of insulin-receptor and GH-receptor mRNAs are also quantitated for comparison. The positions of the 28S and 18S ribosomal RNAs are indicated to the left, and mRNA levels are expressed relative to the level in control 1 (J-RN) cells, as quantitated by PhosphorImager. [Reproduced by permission from Y. Hattori et al.: J Clin Endocrinol Metab 81:2257–2263, 1996 (90 ). © The Endocrine Society.]

B. Leprechaunism and other genetic disorders associated with severe insulin resistance
In 1948, Donohue (91) described a child born to a related couple who appears to be the first reported patient with leprechaunism. In 1954, with Uchida, Donohue described a sibling of the index case; both children manifested intrauterine and postnatal growth retardation; generalized lipodystrophy; facial hirsutism; and an aged face with thickened lips, large wide-set eyes, and large low-set ears (91, 92). Beginning in the 1970s, additional clinical features of this syndrome, dubbed leprechaunism or Donohue syndrome, were described, including hyperinsulinemia, severe insulin resistance, and abnormal glucose homeostasis; acanthosis nigricans; clitoromegaly and ovarian hyperandrogenism in females; gonadotropin-independent precocious puberty; and myocardial hypertrophy (67, 93, 94). The condition is inherited in an autosomal recessive fashion, affects approximately one in 4 million live births, and is usually associated with death in the first year of life (95, 96).

In the late 1970s, studies of cultured fibroblasts from a child with leprechaunism showed markedly decreased insulin binding to insulin receptors and a severe reduction in the ability of insulin to stimulate glucose incorporation into the patient’s cells (97). This led to the suspicion that abnormalities in the insulin receptor might be responsible for this syndrome. Since the cloning of the human insulin receptor cDNA in 1985 (98, 99), more than 50 various mutations in the insulin-receptor gene (typically either homozygotes or compound heterozygotes) have now been described. These mutations can be stratified into five subgroups resulting in either decreased levels of insulin-receptor mRNA, impairment of intracellular transport and posttranslational processing, defects in insulin binding, impairment of receptor-associated tyrosine-kinase activity, or accelerated insulin-receptor degradation (100).

The frequent occurrence of acanthosis nigricans, ovarian hyperandrogenism in females, and myocardial hypertrophy in infants with leprechaunism led to the hypothesis that the massive hyperinsulinemia resulting from the underlying genetically mediated insulin resistance results in certain tissue-specific mitogenic effects mediated through a presumably intact homologous IGF1R-effector mechanism (93). The available data in leprechaun tissues are contradictory as to the normalcy of the IGF-I signaling pathway.

The presence of normal IGF-I-mediated functions was first supported by in vitro observations in leprechaun Winnipeg/NIH. In fibroblasts from this subject, IGF1R affinity was reduced by 75% compared with that seen in control fibroblasts. This defect was associated with reduced MSA-induced glucose incorporation, whereas MSA-stimulated methyl aminoisobutyric acid (AIB) uptake was normal (101). Rechler concluded that leprechaun Winnipeg/NIH possesses a defect in a component of the insulin receptor-effector mechanism that is common both the insulin receptor and the IGF1R, and that this abnormality is specifically involved in coupling both receptors to glucose incorporation, but is spared in their coupling to methyl-AIB uptake. In fibroblasts from another leprechaun patient, Sasaoka et al. (102) found reduced insulin binding and normal IGF-I binding, normal responsiveness of glucose incorporation and AIB uptake to insulin with a shift to the right in the dose-response curve (decreased sensitivity), and normal responsiveness and sensitivity of glucose incorporation and AIB uptake after IGF-I stimulation. These investigators reported similar findings in a 23-yr-old patient with the Kahn type A syndrome of insulin resistance, which usually occurs in young females in association with signs of virilization or accelerated growth (103). In fibroblasts from leprechaun Minn-1 with a non-sense mutation in one of two insulin-receptor alleles at exon 14, mRNA levels for the insulin receptor were markedly decreased whereas IGF-I binding, and IGF1R gene mRNA content and transcription rate, were normal (104).

In studies from our laboratory employing an HTLV-II-transformed T lymphoblast cell line from another infant with leprechaunism, we showed absent in vitro colony proliferation in response to physiological concentrations of insulin, but normal colony proliferation in response to supraphysiological concentrations of insulin and physiological concentrations of IGF-I (67). We deduced that these responses were mediated through an intact IGF1R-effector mechanism, based on studies employing IR-3, a monoclonal antibody directed against the IGF1R (67). More recent studies of fibroblasts from leprechaun Richmond reported by Jospe et al. (105) showed a non-sense mutation in the extracellular domain of the insulin-receptor gene resulting in the complete absence of cell-surface insulin receptors. However, both supraphysiological concentrations of insulin and physiological concentrations of IGF-I stimulated DNA synthesis normally in fibroblasts, an effect, in both cases, mediated through an intact IGF1R-effector mechanism, based on studies employing IR-3 (105, 106). Normal IGF-I-induced hexose uptake into fibroblasts of a Dutch patient with leprechaunism has also been described despite a homozygous mutation in the -chain of the insulin receptor and markedly reduced 2-deoxyglucose uptake by the patient’s fibroblasts (107). In a patient with leprechaunism from Japan with a homozygous mis-sense mutation in the tyrosine kinase domain of the insulin-receptor gene, there was a marked diminution in glucose uptake, IRS-1 tyrosine phosphorylation, thymidine incorporation into DNA, and glycogen synthase activation in response to insulin, yet glucose uptake and IRS-1 tyrosine phosphorylation were normal in response to IGF-I (108). Thus, this leprechaun patient provides further evidence for preservation of a normal IGF1R-effector mechanism in the setting of a major genetic abnormality involving homologous insulin receptors.

The presence of dual defects in insulin and IGF-I action, as well as in EGF action (109, 110), has been suggested as causative in the severe pre- and postnatal growth failure that affects all infants with leprechaunism. Moreover, Kaplowitz and D’Ercole (109) reported that stimulation of [³H]AIB uptake by fibroblasts from a patient with leprechaunism in response to insulin, EGF, MSA, and somatomedin-C was uniformly decreased. In addition, these investigators demonstrated subnormal stimulation of [³H]thymidine incorporation by the leprechaun fibroblasts in response to EGF, somatomedin-C, and fibroblast growth factor. Desbois-Mouthon et al. (111) described a leprechaun patient with severe resistance to both insulin and IGF-I. This patient had two mis-sense mutations located in the gene encoding the ß-subunit of the insulin receptor. Mutant insulin receptors from this patient showed normal insulin binding, but elevated levels of in vivo autophosphorylation and in vitro exogenous tyrosine kinase activity in the absence of insulin, and no further stimulation with the addition of insulin. Moreover, these insulin receptors were unable to promote stimulation of either metabolic or mitogenic pathways. Although IGF-I binding and IGF-I-stimulated receptor kinase activity were normal, the ability of IGF-I to stimulate glycogen and DNA synthesis was significantly diminished compared with that seen in control fibroblasts. In a patient with the type A syndrome of insulin resistance and no discernible alteration in splicing or primary insulin-receptor structure, Knebel et al. (112) reported abnormal insulin binding due to reduced receptor affinity for insulin, abnormal metabolic and mitogenic responsiveness to insulin, and markedly impaired inducibility of c-fos mRNA in response to both insulin and IGF-I.

Another approach toward documenting the presence or absence of IGF-I resistance in states of genetically mediated insulin resistance is to study the in vivo effects of IGF-I administration to affected individuals. In a series of 11 patients from Japan with extreme insulin resistance, including patients with type A insulin resistance, congenital generalized lipodystrophy, and leprechaunism, short-term IGF-I treatment resulted in a fall in blood glucose in each patient to a degree comparable to that seen in normal individuals (113). This hypoglycemic action was maintained for as long as 6 months. Similar partial beneficial metabolic effects in response to IGF-I administration were reported in two Swiss patients over a 10-day period (114) and in another Japanese patient for up to 9 months (115), both of whom had the type A syndrome of insulin resistance. In contrast, IGF-I treatment of two infants with leprechaunism failed to stimulate any apparent glucose-lowering or nitrogen-sparing effects (116). Lastly, neither GH nor IGF-I administration to a child with the Rabson-Mendenhall syndrome (characterized by severe insulin resistance, abnormal glucose tolerance, typical coarse facial features, marked growth retardation, advanced dentition, abdominal protuberance, acanthosis nigricans, and virilization in females) had any significant growth-promoting effect (117). Thus, these in vivo experiments involving IGF-I treatment in conditions characterized by severe insulin resistance, similar to the in vitro data presented above, show variable findings insofar as concomitant IGF-I resistance is concerned.

In summary, based mostly on studies performed in individuals with leprechaunism and severe genetically mediated insulin resistance, associated IGF-I resistance is only variably found. To date, no specific mutation of the IGF1R has been described in any patient with leprechaunism or in any human disease. Thus, it would appear that those patients with defects in both insulin or IGF-I action may possess a dual genetic abnormality in a common pathway beyond the convergence of the two pathways at the site of IRS-1 tyrosine phosphorylation. Alternatively, the inhibition of IGF-I signaling may be acquired secondary to the presence of abnormal insulin receptors. For example, it has been suggested that mutant insulin receptors form hybrid oligomers with wild-type IGF1Rs, and, thus, interfere with normal IGF-I signaling (118). Another possibility is that mutant insulin receptors and wild-type IGF1Rs may compete for common substrates, such as IRS-1, or such substrates may be desensitized by chronic activation of mutant insulin receptors (111). Nevertheless, it remains unclear as to why some patients with leprechaunism maintain normal IGF1R-mediated functions whereas others do not.

C. Deletions of the distal arm of chromosome 15
There have been 10 cases of patients described in the literature with deletions of the distal long arm of chromosome 15 (119, 120, 121, 122, 123, 124, 125), as well as 33 patients with ring chromosome 15 which often results in terminal 15q deletions (126, 127, 128, 129, 130, 131, 132, 133, 134, 135). The majority of these patients had intrauterine growth retardation (IUGR) and postnatal growth failure in addition to other developmental abnormalities. Several genes have been mapped to 15q, including FES (feline sarcoma virus oncogene), FUR (furin membrane-associated receptor protein), CD13 antigen, CTSH (cathepsin H), and the IGF1R. The IGF1R gene has been assigned to 15q26.3. It has been proposed that absence of one copy of the IGF1R gene plays a role in the growth deficiency present in these syndromes.

1. Deletions of the distal long arm of chromosome 15 without ring chromosome. Six of the 10 patients had deletions spanning the region of the IGF1R. A patient described by Pasquali et al. had a 15q26qter deletion (119), while two siblings described by Kristoffersson et al. both had 15q24>qter deletions (120). One of the two siblings was monosomic while the other was trisomic for this deletion. The mother was found to carry a balanced translocation of a segment of chromosome 15 to the distal part of the short arm of chromosome 6 [46,XX, t(6;15)(p25;q24)]. Roback et al. described a patient with 15q26.1qter deletion and confirmed the loss of one copy of the IGF1R gene by Southern blot analysis (121). Most recently, Siebler et al. (122) described two separate cases of children with deletions of 15q26.1qter. Again, quantitative Southern blot analysis confirmed loss of one copy of the IGF1R gene in both patients. All these cases were associated with IUGR. Those who survived the neonatal period also had postnatal growth failure. Of the four other patients, two had IUGR [deletion 15q22[rarrowq24 (123), deletion 15q22q25 (124)], whereas the other two, with slightly more proximal deletions, did not [deletion 15q21 (125), deletion 15q21q24 (124)]. The patient of Fryns et al. (125) ultimately developed postnatal growth failure, whereas the patient of Formiga et al. (124) did not.

The patients that had deletions spanning the IGF1R gene had other clinical features in common. All had some craniofacial abnormalities, including microcephaly, triangular facies, hypertelorism, high-arched palate, abnormal ears, and micrognathia. All of these patients, except one of the siblings reported by Kristoffersson et al. (120), also had skeletal abnormalities including clinodactyly, proximal placement of digits, club feet, and scoliosis. Those that survived the neonatal period all had developmental delay and mental retardation.

Five of the 10 cases had pulmonary abnormalities and of these, three also had renal abnormalities. The patient of Clark et al. (123) had lung hypoplasia and cystic renal dysplasia. The patient of Roback et al. (121) had lung hypoplasia and small kidneys. The siblings described by Kristoffersson et al. (120) both had lung hypoplasia and diaphragmatic hernias and also had cystic kidneys. The infant described by Siebler et al. (122) had lung hypoplasia and a diaphragmatic hernia.

Data on serum hormone levels in these patients are limited. Lack of one copy of the IGF1R gene might be expected to cause primary IGF-I resistance through a quantitative deficiency of the receptor. With end-organ insensitivity, negative feedback by IGF-I on the pituitary gland would be ineffective, resulting in elevated GH levels and, subsequently, elevated IGF-I levels. The patient of Roback et al (121) had a reportedly normal IGF-I level, but GH levels were not mentioned. The 12.3-yr-old female patient of Siebler et al. (122) had the most comprehensive evaluation. At 10.4 yr of age, her bone age was 7.5 yr. Her basal plasma GH level was normal and showed a normal response to glucagon stimulation. It is not clear whether serial basal GH levels were obtained. Multiple IGF-I levels were essentially normal except for a single elevated measurement of 340 µg/liter at 2 yr of age (normal 11–206 µg/liter). The girl was treated with three courses of hGH lasting 1–2 yr each. In the first two courses, there appeared to be an initial, but unsustained, increase in height velocity. Although the patient was clinically somewhat resistant to GH, the basal IGF-I and GH levels were not thought to be consistent with IGF-I resistance.

The two cases of Siebler et al. (122), with confirmed lack of one copy of the IGF1R gene, had further studies of receptor expression and function in cultured fibroblast cell lines. Solution hybridization/nuclease protection assay results showed that the level of IGF1R mRNA was approximately 50% of that in control fibroblasts. Binding of [¹²⁵I]IGF-I to placental tissue obtained from one case was normal. However, the investigators showed reduced ligand binding to IGF1R on cultured fibroblasts from both patients. Although they were unable to distinguish decreased receptor number from decreased binding affinity, the authors felt that, in light of the decreased IGF1R mRNA levels, the former etiology was more likely. They also studied responsiveness to IGF-I by quantification of IGF-I-stimulated [-1-¹⁴C]methyl-AIB uptake and [methyl-³H]thymidine incorporation by the patients’ cultured fibroblasts. In these experiments, they were able to demonstrate that the maximal response to IGF-I was significantly decreased, compared with controls, in fibroblasts from one of the two patients when expressed as fold response over the basal value. However, when the data were expressed as net stimulation (maximal response minus basal), there was no significant decrease compared with controls. As the studies could not provide conclusive evidence for impaired biological responsiveness to IGF-I, the results suggest that growth retardation in these patients is not related to missing one copy of the IGF1R gene. The authors warn, however, that extrapolation of the cultured fibroblast data to other organ systems in situ may be inaccurate. It is possible that in certain tissues, IGF1R number is more tightly coupled to biological response. In such situations, decreased IGF1R number would result in IGF-I resistance with impaired growth response to IGF-I.

2. Ring chromosome 15. The patients with ring chromosome 15 generally have deletions of 15q26.2 or 15q26.3 bands, although other terminal bands may be involved, as well as deletions of the distal short arm of chromosome 15 (126, 127). In all 33 cases in the literature, there was postnatal growth failure (127, 128, 129, 130, 131). In 14 of the 17 cases reviewed by Butler et al. (127), birth length was below the fifth percentile. Phenotypic abnormalities included microcephaly (88%), mental retardation (95%), triangular facies (42%), hypertelorism (46%), brachydactyly (44%), various cardiac abnormalities (30%), cryptorchidism (30%), fifth finger clinodactyly (26%), and talipes equinovarus (15%).

The absence of one copy of the IGF1R gene was documented in three of five patients described by Francke et al. (132) and Peoples et al. (128). Those three who were hemizygous had severe growth failure, whereas those who had both copies of the IGF1R gene had only mild growth failure. Baba et al. (133) cited a case with one missing copy of the IGF1R gene. That child’s height at 4 yr of age was -5.8 SD. Tamura et al. (134) reported an 11-yr-old girl with ring chromosome 15, pre- and postnatal growth failure, and deletion of one copy of the IGF1R gene.

Normal serum GH and IGF-I levels have been reported by several investigators although none specified whether frequent basal sampling was obtained (126, 127, 133, 134, 135). Nuutinen et al. (136) described a child who did not have deletion of the IGF1R gene and whose IGF-I levels were subnormal. Although basal GH levels were not reported, the authors indicated a normal GH peak in response to clonidine stimulation (35.6 ng/ml), but a poor response to insulin-induced hypoglycemia (5.3 ng/ml). This child responded to GH therapy with sustained acceleration of growth and increase of IGF-I levels into the normal range. There are no published reports of GH treatment of any patient with ring chromosome 15 with deletion of the IGF1R gene. Neither have there been any studies on IGF1R expression or function.

In summary, some patients with ring chromosome 15 or deletions of the long arm of chromosome 15 have loss of one copy of the IGF1R gene. Although this may be associated with decreased IGF-I binding to its receptor, the exact mechanism has not been elucidated. Serum GH and IGF-I levels were not consistent with the pattern predicted for IGF-I resistance. Furthermore, there is, as yet, no evidence for altered tissue responsiveness to IGF-I which, if present, would represent a genetically regulated form of IGF-I resistance. Thus, there is no definitive evidence that lack of one copy of the IGF1R plays a causative role in the pre- and postnatal growth failure present in the majority of these patients.

D. Rare examples of "idiopathic" short stature associated with elevated circulating levels of IGF-I
The mouse "knockout" studies showed that isolated IGF-I deficiency caused IUGR and postnatal growth failure with variable survival. In humans with idiopathic short stature, there has only been one case described of a child with IUGR, short stature, dysmorphic features, mental retardation, and homozygosity for partial deletion of the IGF-I gene (137). In contrast, mouse "knockout" studies showed that the interaction between IGF-I and the IGF1R is critical for normal prenatal growth and development. If this finding is applicable to humans, one would expect that congenital abnormalities causing primary IGF-I resistance would be exceedingly rare. Indeed, there have been only a handful of patients described in the literature who presented with "idiopathic" short stature and elevated IGF-I levels.

Lanes et al. (138) described a boy with IUGR and short stature. At 11 2/12 yr of age, his height was below -5 SD, and he had a 4-yr bone age delay. He had no dysmorphic features and no evidence of hypoglycemia. Basal 24-h integrated GH concentration was normal, as were insulin-arginine- stimulated GH levels. Basal IGF-I values were elevated to 3–5 times normal values and did not increase with 5 days of GH therapy. The patient’s serum had normal IGF-I bioactivity, ruling out the presence of inhibitors or IGFBP abnormalities. The authors concluded that the findings were most consistent with an IGF1R or postreceptor defect. The lack of suppression of GH levels was attributed to ineffective negative feedback by IGF-I due to central IGF-I resistance. Momoi et al. (139) reported the case of a girl with IUGR and less severe postnatal growth failure than the patient described by Lanes et al. Her height tracked at -3 SD throughout childhood. She started menarche at 13 8/12 yr of age, but she had a final adult height of 134 cm (-5 SD). She had some unusual dysmorphic features, including malar hypoplasia, thin and stiff hair, blue sclerae, and upslanting palpebral fissures. She had no episodes of hypoglycemia. Her basal GH levels were normal, as were her stimulated levels after a variety of stimuli, including insulin, L-dopa, glucagon, and arginine. Basal IGF-I levels were approximately 2-fold elevated above normal. The patient’s serum had normal IGF-I bioactivity as measured by IGF-I-stimulated uptake of [³H]thymidine and [³⁵S]sulfate into chondrocytes. In both patients, fibroblast responsiveness to IGF-I was examined. Fibroblasts from the patient of Lanes et al. were subsequently studied by Heath-Monnig et al. and found to have normal responsiveness to IGF-I as measured by IGF-I-stimulated uptake of [³H]-AIB (140). Fibroblasts from the patient of Momoi et al. likewise responded normally to IGF-I, as measured by IGF-I- stimulated uptake of [³H]thymidine and [³H]-AIB (139). Both sets of authors concluded that the patient’s clinical and hormonal profiles were consistent with IGF-I resistance. The findings of normal IGF-I responsiveness of the patients’ fibroblasts are not incompatible with this diagnosis, as Momoi et al. (139) felt the fibroblast data could not be generalized to other organ systems. They suggested that tissue-specific unresponsiveness could be present.

Bierich et al. (141) published the case of a young girl with IUGR and profound postnatal growth failure. She grew only 11 cm in the first year of life. Her dentition was delayed and she had a chubby face with a well developed skull, small snub nose, receded chin, and a relatively obese trunk. During her second year of life, she was noted to have several episodes of early-morning drowsiness due to hypoglycemia. The frequency of these episodes declined with the addition of a midnight meal. The girl’s basal GH levels were elevated to 3- to 6-fold above normal and her basal IGF-I levels were elevated 6-fold. Her GH stimulation profile in response to arginine was abnormal with a small and delayed rise at 90 min. Her IGF-I level did not rise with GH stimulation. The patient’s serum had normal IGF-I bioactivity. However, there was a 50% diminution of the specific binding of IGF-I to the patient’s skin fibroblasts. The authors concluded that these findings were consistent with a genetically mediated defect of the IGF1R, although they did not study in vitro IGF-I action.

Heath-Monnig et al. investigated IGF-I responsiveness of fibroblasts from 11 normal control subjects and 9 patients with short stature, 5 of whom had normal GH concentrations and normal or increased IGF-I levels (140). One patient was found to have decreased [³H]-AIB uptake. The general shape of her response curve was normal, and the eventual degree of stimulation was similar to that of normal fibroblasts, but the ED₅₀ was 3.3 times that of the mean of normal fibroblasts. Interestingly, the patient’s mother’s fibroblasts also had decreased sensitivity to IGF-I. The proband had IUGR and significant, but not profound, short stature; her height was approximately -3.5 SD at 9 3/12 yr of age. Binding studies indicated normal affinity and number of IGF1R. Tollefsen et al. (142) later studied the responsiveness of the patient’s fibroblasts to [Q³, A⁴, Y¹⁵, L¹⁰]IGF-I, an IGF-I variant that has 600-fold reduced affinity for serum IGFBPs, and which is, thus, less likely to have its action modulated by endogenous IGFBPs (140, 142). In response to the variant IGF-I, the patient’s fibroblasts had a normal ED₅₀ of [³H]-AIB uptake, eliminating the possibility of a defect of the IGF1R. Ligand blot analysis showed that the patient’s fibroblasts secreted a significantly greater amount of several IGFBPs. The cells had a 10-fold increase in the amount of cell surface M_r 32,000 binding protein. Based on these findings, the authors concluded that abnormal production and/or cell association of IGFBPs caused resistance to IGF-I action in this patient.

In summary, cases of congenital IGF-I resistance associated with postnatal growth failure beginning immediately in the neonatal period and elevated IGF-I levels are extremely rare. Only one patient had documented tissue IGF-I resistance. The failure to demonstrate IGF-I resistance in two other patients with clinical and hormonal profiles consistent with this diagnosis may be attributable to the fact that only fibroblasts were studied. These patients may have tissue mosaicism, which allowed them to survive and have limited clinical manifestations. There has also been one patient described with congenital, secondary IGF-I resistance due to altered IGF-I bioavailability as a result of a genetically directed abnormality of IGFBP.

	VI. Conditions Associated with Acquired IGF-I Resistance in Humans