Search

The IGF system plays a crucial role in ovarian folliculogenesis, and changes in IGF-binding protein (IGFBP) levels modulate IGF bioavailability. Data from various mammalian models suggest a link between body size, IGF1 in serum and female reproduction parameters. Among the vertebrate species, the dog exhibits the widest span in body height. Height is known to be positively correlated with the concentration of serum IGF1. In this work, the ovarian physiology of 40 bitches exhibiting a wide span of height, and breed type was investigated. IGF1, IGF2, IGFBP3, estradiol (E(2)), and progesterone concentrations in plasma and preovulatory follicular fluid were quantified. A total of 455 follicles, 2-8 mm in diameter, were recovered at the preovulatory stage, measured, and punctured. Intrafollicular levels of IGF1 were positively correlated with plasma levels, and plasma IGF1 levels were positively correlated with both bitch height and weight. The concentrations were threefold higher in large dogs compared with small dogs. A positive correlation between intrafollicular and plasmatic IGFBP3 levels and a positive correlation between plasmatic IGFBP3 levels, and both height and weight of the bitches were observed. The number of preovulatory follicles and the diameter of the three largest follicles were positively correlated with bitch height. E(2) intrafollicular concentrations were higher in preovulatory follicles from small animals than in those from large animals. In conclusion, the strong variability in height between dogs appeared to be associated with dramatic differences in IGF1, and IGFBP3 levels, in both plasma and follicular fluid. These differences were associated with significant differences in some functional aspects of ovarian follicles.

Gram-negative bacterial infection or treatment of animals with bacterial lipopolysaccharide (LPS) induces a catabolic state with proteolysis, liver injury and an inhibition of the insulin-like growth factor-I (IGF-I) system. The purpose of this work was to elucidate the role of Kupffer cells in LPS-induced inhibition of the IGF-I/IGF-binding protein-3 (IGFBP-3) system. Adult male Wistar rats were either pretreated with the Kupffer cell inhibitor gadolinium chloride (10 mg/kg, i.v., 24 h prior to LPS exposure) or saline vehicle. Rats received two i.p. injections of 1 mg/kg LPS (at 17:30 and 08:30 h the following day) and were killed 4 h after the second injection. LPS administration induced a significant decrease in body weight and in serum concentrations of IGF-I and IGFBP-3 (P < 0.01), as well as in their gene expression in the liver. LPS-injected rats had increased serum concentrations of ACTH, corticosterone (P < 0.05), tumour necrosis factor-alpha (TNF-alpha) and nitrites (P < 0.01). Pretreatment of the animals with gadolinium chloride blocked the inhibitory effect of LPS on body weight, and on serum concentrations of IGF-I, IGFBP-3 and nitrites, as well as growth hormone receptor (GHR), IGF-I and IGFBP-3 gene expression in the liver. In contrast, gadolinium chloride administration did not modify the stimulatory effect of LPS on serum concentrations of ACTH, corticosterone and TNF-alpha. These results suggest that Kupffer cells are important mediators in the inhibitory effect of LPS on GHR, IGF-I and IGFBP-3 gene expression in the liver, leading to a decrease in serum concentrations of IGF-I and IGFBP-3.

Elucidating how mitogens facilitate epithelial/stromal interactions is critical given that mitogens regulate mammary gland development and function. IGF-I is a potent mammary cell mitogen that is locally produced in the mammary gland. Since IGF-binding proteins (IGFBPs) regulate IGF-I bioavailability, we characterized the cell-type-specific production of IGFBP in primary bovine mammary epithelial (BME) and fibroblast (BMF) cells. Cells were treated with IGF-I and mRNA levels were analyzed via quantitative real-time (qRT)-PCR and Northern blot analysis. Media conditioned by cells treated with IGF-I for 48 h were analyzed via ligand blotting with 125I-labeled IGF-I and -II and immunoblotting with specific IGFBP antibodies. A reciprocal regulation of IGFBP-3 and -5 by IGF-I was observed between the two cell types. IGF-I induced large dose-dependent increases in IGFBP-3 mRNA and protein levels in BME cells, while IGFBP-5 protein was barely detectable and mRNA levels were detectable only by qRT-PCR. In BMFs, IGF-I induced large increases in IGFBP-5 mRNA and protein while IGFBP-3 mRNA was only slightly increased by IGF-I treatment and the protein was difficult to detect. IGFBP-6 mRNA was detected by Northern blot analysis in both cell types but was not regulated by IGF-I. In BME cells, IGFBP-6 protein levels were readily detectable under basal conditions and were increased by IGF-I. Interestingly, IGFBP-6 protein could not be detected in media conditioned by BMFs. IGFBP-4 mRNA was readily seen by Northern blot analysis in BMFs, however qRT-PCR was required to detect IGFBP-4 mRNA in BME cells. IGF-I increased IGFBP-4 mRNA levels by 2-fold in both cell types. IGFBP-4 protein was only detectable in media conditioned by BME cells when stimulated by IGF-I. In contrast, IGFBP-4 was present in media conditioned by untreated BMFs but was not consistently increased by IGF-I treatment. This was explained by the finding that IGF-I stimulated proteolysis of IGFBP-4, as evidenced by the appearance of two immuno-responsive fragments of 18 and 14 kDa. This proteolysis was specific to IGFBP-4, and was not observed in BME cells. We confirmed the protease to be pregnancy-associated plasma protein A (PAPP-A) by immunoblotting with an antibody against human PAPP-A/proMBP (pro form of eosinophil major basic protein) complex. In vitro immuno-neutralization experiments showed that blocking PAPP-A prevented the ability of IGF-I to stimulate IGFBP-4 proteolysis. IGFBP-2 mRNA and protein levels were observed under basal conditions in both cell types, with no significant regulation by IGF-I. The analysis of cell-type-specific regulation of the IGF system in both primary mammary epithelial cells and stromal cells will assist in the characterization of the mechanisms behind the role of the IGF system in normal mammary physiology and ultimately breast cancer.

The insulin-like growth factors, IGF-I and IGF-II, and their binding proteins play an important role in the growth and development of the central nervous system. In the brain, colocalization of IGFs and IGFBPs often occurs, suggesting that IGFBPs can modulate IGF action. In one strain of our human (h)IGF-II transgenic mice, which carry an hIGF-II transgene driven by the H-2Kb promoter, we found overexpression of hIGF-II in the brain, as measured by Northern blot analysis. To clarify the localization and influence of the hIGF-II transgene on different components of the GH-IGF axis in the brain, we studied the expression pattern of the hIGF-II transgene, endogenous IGF-I and IGF-II, and IGFBP-2, -3 and -5 in the brain of prepubertal 4-week-old mice, using nonradioactive in situ hybridization. We found that the hIGF-II transgene is exclusively expressed in neurons of the piriform cortex, the cerebral cortex, the medulla oblongata and the granular layer of the cerebellum. In general, this pattern is comparable to the expression pattern of endogenous IGF-I, with a few exceptions: there is no expression of IGF-I in the granular layer of the cerebellum, whereas the Purkinje cells of the cerebellum and thalamus both express IGF-I but no hIGF-II transgene. This hIGF-II transgene expression pattern contrasts markedly with endogenous IGF-II expression, which is mainly located in nonneuronal cells such as the meninges and choroid plexus, and in some nuclei of the medulla oblongata. The hIGF-II transgene affects neither endogenous IGF-I and IGF-II expression, nor the expression of IGFBP-3, which is located in the choroid plexus. Although the hIGF-II transgene is expressed in neuronal structures similar to IGF-I and IGFBP-5, it is not able to regulate IGFBP-5 expression, as has previously been reported for IGF-I. In the medulla oblongata, the IGFBP-2 expression level showed 10-fold upregulation by the transgene, suggesting a modulating role for IGFBP-2 at the hIGF-II transgene action in this region.

IGF-I has been implicated as a factor that may predispose one to prostate cancer and to benign prostatic hypertrophy (BPH). We established murine IGF-I transgenic mice under the control of rat probasin promoter and analysed the histology of the murine IGF-I-overexpressing prostate. Immunohistochemically, IGF-I was expressed in prostatic epithelial cells or basement membranes of the ventral, dorsal and lateral lobes in a line of IGF-I transgenic mice, but not in their control littermates. The anterior lobe did not express IGF-I. IGF-binding protein-3 (IGFBP-3), inhibitory to the mitogenic action of IGF-I, was detected in epithelial cells of prostatic ventral lobes, but not in those of the dorsal, lateral or anterior lobes of IGF-I transgenic mice. In controls, IGFBP-3 was not detected in epithelial cells of any prostatic lobe. Macroscopic prostatic size and the appearance of IGF-I transgenic mice were comparable with those of their control littermates of the same age. With a computed morphometric analysis, epithelial glands and intraglandular lumens in the prostatic lobes except the ventral lobe were smaller at 17 Months of age than at 14 Months both in IGF-I transgenic mice and controls. Glands and intraglandular lumens in the ventral prostatic lobes of IGF-I transgenic mice expressing more IGF-I protein in the prostate than controls were dense and enlarged similar to cysts compared with those of non-transgenic littermates without showing epithelial growth. Glands and lumens in the dorsal and lateral lobes of the IGF-I transgenic mice were also larger than controls at 14 and/or 17 Months of age. Glands in the anterior prostatic lobe of the IGF-I transgenic mice were not morphologically or morphometrically different from those of non-transgenic littermates. In conclusion, IGF-I transgenic mice under the control of rat probasin promoter showed more dense and enlarged epithelial glands in their prostatic ventral, dorsal and lateral lobes.

Insulin-like growth factor (IGF) binding protein (IGFBP)-3 has been shown to be a growth inhibitory, apoptosis-inducing molecule by virtue of its ability to bind IGFs, in addition to previously demonstrated IGF-independent effects. The recent discovery of the interaction between nuclear IGFBP-3 and 9-cis retinoic acid receptor-alpha (retinoid X receptor alpha RXRalpha), a nuclear receptor, and its involvement in the regulation of transcriptional signaling and apoptosis represents an important paradigm shift in the understanding of IGFBP function. RXRalpha is required for the apoptosis-inducing effects of IGFBP-3. IGFBP-3 and RXR ligands are additive in inducing apoptosis in cancer cells. IGFBP-3 has direct effects on gene transcription, as RXR response element reporter signaling was enhanced and the all-trans retinoic acid receptor response element reporter signaling was inhibited. Accumulating evidence further confirms IGF-independent functions of this multifunction binding protein. Other binding proteins, in addition to other members of the IGF axis, have now been described in the nucleus and are postulated to have effects on transcriptional events. Investigation into these new interactions will expose new protein partners in the interface between the nuclear receptor and growth factor pathways and reveal new targets to be exploited in the treatment of cancer and other diseases.

It was recently discovered that the streptozotocin (STZ)-diabetic mouse model is characterised by GH hypersecretion in contrast to the STZ-diabetic rat, the former thus mimicking the changes in GH in human type 1 diabetes. Inhibition of circulating and renal IGF-I by long-acting somatostatin analogues reduces renal and glomerular growth and urinary albumin excretion in diabetic rats. The aim of the present study was to examine renal and glomerular growth in early experimental diabetes in mice along with changes in the GH/IGF-I axis following treatment with the somatostatin analogue octreotide. Balb/C(a) mice were randomised into non-diabetic controls, placebo-treated and octreotide-treated diabetic (50 microg/day) mice and examined 7 and 14 days after induction of diabetes. There was no effect of octreotide treatment on body weight, glycaemic control or food intake. However, octreotide treatment significantly inhibited renal and glomerular growth by the end of the study period when compared with placebo treatment. In addition, octreotide prevented an increase in kidney IGF-I by day 7. GH hypersecretion was observed in the diabetic groups but octreotide treatment reduced GH levels compared with placebo treatment by day 14. No significant differences in serum or kidney IGF-binding protein-3 levels were observed between placebo- and octreotide-treated diabetic mice. In conclusion, this new diabetic mouse model mimicking human type 1 diabetes is characterised by GH hypersecretion and the somatostatin analogue octreotide is able to prevent renal and glomerular growth, probably mediated through changes in circulating GH and local kidney IGF-I levels.

In an attempt to address the complex and clinically challenging question of the causes of muscle wasting in patients with cachexia, we have developed a primary adult human skeletal muscle cell model. The cultured cells were characterised by immunocytochemistry using antibodies to the myofibrillar protein constituents desmin and titin. Myotube formation was confirmed biochemically by a fourfold increase in the activity of the muscle-specific enzyme creatinine kinase, and myoblast withdrawal from the cell cycle, which is essential for terminal differentiation, was associated with progressive retinoblastoma protein dephosphorylation. Having successfully confirmed the phenotype of these adult human muscle cells, we assessed their interaction with the insulin-like growth factor (IGF) system. IGF-I is known to stimulate myoblast survival, proliferation and differentiation in cell lines, and, like insulin, is a potent anabolic agent in the regulation of protein metabolism. We have shown that IGF-I stimulated both replication and differentiation of myoblasts, whilst fibroblast growth factor-2 stimulated replication but inhibited differentiation. Examining the IGF system during the process of terminal differentiation, we found that both myoblasts and myotubes expressed insulin, IGF-I and insulin-IGF-I hybrid receptors, with the levels of all three receptor types increasing on differentiation. The cells also produced a wide range of IGF binding proteins (IGFBPs) including IGFBP-2, IGFBP-4 and abundant IGFBP-3, which has not been shown to be produced by any other skeletal muscle cell line examined to date. Both insulin and IGF-I had anabolic effects on myotube protein metabolism at physiological concentrations. Insulin was more potent than IGF-I: use of the IGF analogue long R(3)IGF-I demonstrated that the effects of exogenous IGF-I on protein metabolism were not affected by the high levels of endogenous IGFBP production. In summary, we have developed and characterised a clinically relevant in vitro model with which to address the aetiology of muscle wasting associated with chronic catabolic conditions, and we anticipate that future work will enable the development of novel, effective therapeutic interventions.

IGF-I and IGF-II are potent mitogens, postulated to exert autocrine/paracrine effects on growth regulation in human lung cancer. Their proliferative effects are modulated by IGF-binding proteins (IGFBPs), which are found in conditioned medium (CM) of lung cancer cell lines. The biological role of the IGFBPs, which are ontogenetically and hormonally regulated, is not fully understood. Both inhibitory and stimulatory effects on cell growth have been demonstrated. Exogenous IGFBP-3 has been consistently shown to block IGF action, inhibiting cell growth in vitro. In order to evaluate the action of endogenously produced IGFBP-3 on cell growth in lung cancer, we stably transfected the non-small cell lung cancer cell line NCI-H23 with human IGFBP-3 cDNA (resulting in NCI-H23 pOPI3/BP-3) or with the vector pOPI3CAT as control (resulting in NCI-H23 pOPI3CAT). RT-PCR confirmed expression of IGFBP-3-specific mRNA in NCI-H23 pOPI3/BP-3, but not in N! CI-H23 or NCI-H23 pOPI3CAT. Western ligand blot and Western immunoblot analysis of CMs yielded strong signals of the characteristic 40-44 kDa human IGFBP-3 protein in NCI-H23 pOPI3/BP-3. An IGFBP-3 ELISA demonstrated a 20-fold increase in IGFBP-3 protein expression in NCI-H23 pOPI3/BP-3 as compared with NCI-H23. The growth of NCI-H23 pOPI3/BP-3 in serum-containing medium was significantly slower (1.7-fold) than that of NCI-H23 or the vector-transfected control NCI-H23 pOPI3CAT. While the proliferation rate of parental and vector-transfected cells could be stimulated by IGF-I, IGF-II, IGF-I analog Long R(3) IGF-I or insulin, that of NCI-H23 pOPI3/BP-3 could not. Xenotransplantation in nude mice resulted in marked tumor growth after the injection of NCI-H23 or NCI-H23 pOPI3CAT, but absent or minimal growth for the IGFBP-3-transfected cell line. These data suggest that IGFBP-3 is a potent inhibitor of cell growth in human lung cancer c! ell lines and may impair tumorigenicity in vivo.

The 140 kDa ternary complex of insulin-like growth factor-binding protein-3 (IGFBP-3), IGFs and an acid-labile subunit (ALS) has previously been shown to be decreased in diabetes mellitus in humans and rats. We have studied IGF-I levels and ternary complex formation in normal and diabetic cats. Total IGF-I concentrations, measured by RIA using des(1-3)-IGF-I as tracer were (+/-s.e.m.) 54+/-13 nmol/l in eight normal and 227+/-57 nmol/l in eight diabetic cats (P<0.01). The size-distribution of IGFBPs in the cat circulation was determined by incubation with (125)I-IGF-II and Superose 12 chromatography. In normal animals 26+/-2% of the (125)I-IGF-II were in a 140 kDa form compared with 48+/-5% in diabetic cats (P<0.01). When samples from normal and diabetic animals were co-incubated 52+/-3% were at 140 kDa. A similar shift was seen when normal cat and normal human serum were co-incubated. A 2-fold increase in the 140 kDa form in diabetic cats was confirmed first by size-fractionating samples and then performing a ligand-binding assay with (125)I-IGF-I or -II and charcoal separation. SDS-PAGE and Western ligand blotting demonstrated a 45 kDa doublet (presumably IGFBP-3) and 30-35 kDa forms. There were no apparent differences between normal and diabetic profiles on SDS-PAGE, suggesting that a proportion of IGFBP-3 which circulates 'free' in normal cats forms a ternary complex in the diabetic circulation. We conclude that (i) in contrast to humans and rats, ALS is the limiting factor for ternary complex formation in normal cats, (ii) ALS concentrations increase in feline diabetes mellitus and, by promoting ternary complex formation, this leads to an increase in total IGF-I concentrations, and (iii) total IGF-I concentrations may not be reliable in the diagnosis of acromegaly in diabetic cats.

Hyper-adrenal activity and increased glucocorticoid hormone release are associated with disruptions in reproductive function and adverse effects on the ovary. The aim of this investigation was to determine whether elevated glucocorticoid hormone levels can influence ovarian IGF-I synthesis and action in vivo. To elevate endogenous glucocorticoid levels, gilts were treated with ACTH during the luteal phase of the oestrous cycle (days 9-13) while the control group received saline. The gilts were subsequently ovariectomized on either day 14 or day 18 of the oestrous cycle. Follicular fluid (FF) was collected from individual follicles; IGF-I and steroid hormone concentrations were determined by radioimmunoassay, and IGF-binding protein (IGFBP) expression was assessed by Western ligand blotting. Granulosa cells were also recovered and placed in culture to determine IGF-I, progesterone (P(4)) and oestradiol-17beta (E(2)) production levels. The cells were cultured in serum-free medium for 5 days and supplemented with: (a) media alone, (b) IGF-I, (c) FSH and androstenedione (A(4)), or (d) IGF-I with FSH and A(4). The FF from ACTH-treated gilts was characterized by elevated (P<0.05) cortisol levels on day 14 and lower (P<0.05) E(2) values on both day 14 and day 18. Lower (P<0.05) IGF-I concentrations were also measured in the FF of ACTH-treated gilts collected on day 18. This altered hormone profile in FF was associated with impaired IGF-I and steroid hormone synthesis by granulosa cells. IGF-stimulated P(4) production (P<0.01) by cells recovered from ACTH-treated gilts on day 14 was lower (P<0.05). By day 18, IGF-I, P(4) and E(2) production by cells from the ACTH group were all significantly (P<0. 05) lower. These results demonstrate that increased glucocorticoid concentrations can disrupt subsequent ovarian IGF-I synthesis and IGF action in vivo and can, potentially, impair follicle maturation.

We have studied changes in the IGF axis in an ovine model of myocardial infarction (MI), in order to determine the relationship between time-based changes in post-infarct myocardium and IGF levels. IGF localization was studied by immunocytochemistry, production by in situ hybridization, and specific binding by radioligand studies. In surviving tissue, IGF-I peptide localized to cardiomyocytes, with strongest immunostaining at 1 and 2 days post-infarct in the immediate border area adjoining the infarct, where IGF-I mRNA also increased, reaching a maximum at 2 days. Binding of radiolabelled IGF-I in surviving tissue was initially lower than that seen in cardiomyocytes in control myocardium, subsequently increasing to become significantly greater by 6 days post-infarct. In necrotic tissue, IGF-I peptide was still detectable in cardiomyocytes at 0.5 days post-infarct, but had cleared from this area by 1 day, becoming detectable again at 6 days post-infarct in macrophages and fibroblasts infiltrating the repair zone. IGF-I mRNA was not detected in necrotic tissue until 6 days, when probe hybridized to macrophages and fibroblasts. Within the necrotic zone, high levels of radiolabelled IGF-I binding to a combination of receptors and binding proteins were observed in cardiomyocytes in islands of viable tissue located close to the border. Weak immunostaining for IGF-II was observed in cardiomyocytes of the surviving tissue. IGF-II mRNA was not detected in either surviving or necrotic areas. Binding of radiolabelled IGF-II was predominantly to macrophages in both surviving and infarct areas, although as with IGF-I, high levels of binding of radiolabelled IGF-II to a combination of receptors and binding proteins were observed in islands of viable tissue close to the border within the necrotic area. We conclude that, following MI, surviving cardiomyocytes at the infarct border show marked changes in IGF-I localization, production, and specific binding, indicating that the IGF axis is directly involved in post-infarct events, possibly in the maintenance of cardiac function by the induction of hypertrophy and in cell survival by decreasing apoptotic cell death, which has been demonstrated in other cell types.

IGF-I has been proposed as a key regulator of mammary epithelial cell (MEC) growth and differentiation. As IGF-I bioactivity is modulated by specific, high-affinity binding proteins (IGFBP), the forms of IGFBP that are secreted by the bovine MEC line, MAC-T, were identified. Media conditioned by MAC-T cells contained four forms of IGFBP that were identified, by western blotting with specific antibodies, as IGFBP-2, -3, -4 and -6. The amounts of IGFBP-3 in conditioned media were relatively low under basal conditions when analyzed by ligand blotting with 125I-IGF-II, but were increased dramatically relative to serum-free controls by exposure to IGF-I (100 ng/ml) or IGF-II (100 ng/ml) for 24 h. These increases in IGFBP-3 protein corresponded with dose-dependent increases in IGFBP-3 mRNA, with IGF-II eliciting a smaller response than was elicited by IGF-I at each concentration. Leu-IGF-I, which has reduced affinity for the IGF-I receptor but normal affinity for IGFBPs, failed to increase IGFBP-3 protein and mRNA levels, whereas B-chain IGF-I (normal affinity for the receptor but reduced affinity for IGFBPs) elicited the response, thus implying an IGF-I receptor-mediated event. Time-course studies indicated that IGFBP-3 mRNA was increased fourfold by 3 h of IGF-I treatment, with maximal increases of eightfold above serum-free controls observed between 8 and 13 h of treatment. By 24 h of treatment, IGFBP-3 mRNA levels had declined and were approximately threefold above controls in cells exposed to IGF-I. Amounts of messenger RNA of IGFBP-6 and IGFBP-2 were not increased by IGF treatment. However, retinoic acid (10(-6) M) stimulated both IGFBP-2 and IGFBP-6 protein and mRNA levels, but it decreased IGFBP-3 mRNA levels relative to controls. The combination of retinoic acid plus IGF-I had no additional effect on IGFBP-6 or -2 above that observed with retinoic acid alone, whereas IGF-I together with retinoic acid attenuated the decrease in IGFBP-3 observed with retinoic acid alone. Protein kinase A-mediated pathways were also shown to alter IGFBP synthesis. Forskolin, which increases cAMP, increased IGFBP-3 protein and mRNA levels. The combination of IGF-I plus forskolin resulted in greater increases in both protein and mRNA than were observed with either treatment alone. In contrast, forskolin decreased IGFBP-6 mRNA relative to controls, but had no effect on IGFBP-2. The decrease in IGFBP-6 was less marked when cells were treated with a combination of IGF-I and forskolin. Forskolin had no effect on IGFBP-2 mRNA levels. In summary, the ability of IGF-I specifically to regulate IGFBP-3 synthesis represents a mechanism whereby IGF-I may regulate its own bioactivity. In addition, the differential regulation of IGFBP-2, -3 and -6 by retinoic acid (which inhibits proliferation) and IGF-I (which stimulates proliferation) suggests that these forms of IGFBP have different roles in regulating mammary epithelial cell physiology.

The presence of IGFs and their associated binding proteins (IGFBPs) in human follicular fluid is well documented. Furthermore, most of the constituents of the IGF system in follicular fluid have been found to vary, either in total amount or by proteolytic cleavage, depending on the health status of the follicle. In this study we have examined the acid-labile subunit (ALS) and found that levels in follicular fluid (mean 146 nmol/l) were almost 50% of those in the circulation. This amount of ALS was considerably greater than that found in other extracirculatory fluids (20.9 for synovial fluid and 31.4 nmol/l for skin blister fluid). As in the circulation, ALS levels were in molar excess and did not vary between atretic and dominant follicles. Although the source of ALS is probably from blood (conditioned medium from ovarian cell cultures had no measurable ALS) it would appear that this glycoprotein is not merely diffusing from the circulation as the capillary endothelium becomes more permeable in dominant follicles and this is not reflected in the level of ALS. Analysis of the distribution of IGF-I, IGF-II and IGFBP-3 in fluid from healthy and atretic follicles revealed that the majority of these growth factors (> 80% of total IGF-II) were in the 150KDa complex, indicating that the ALS present was functional, in that it formed the ternary complex with a molecule of IGFBP-3 and IGF. No free IGF-II was found in any of the follicular fluids analysed nor was there any increase in the amount of unsaturated IGFBP-3 in atretic follicles. In summary, we have shown that the majority of IGF measured in follicular fluid, whether from healthy or atretic follicles, is bound in the ternary complex.

The IGFs are believed to play an important role in the regulation of steroidogenesis and follicular maturation in the human ovary. The activities of the IGFs are regulated by a family of binding proteins (IGFBPs) which are subject to a number of potential post-translational modifications. The aim of this study was to investigate both the production and modification of the IGFBPs in follicular fluid and in medium conditioned by granulosa cells and theca from individual follicles at different stages of maturation. In follicular fluid from healthy, dominant follicles there was an increase in the amount of IGFBP-2, -3 and -4 present as lower molecular weight forms (23 kDa, 29 kDa and 16.5 kDa respectively) in comparison to that seen in atretic follicles from the same ovary. Furthermore for IGFBP-4, this fragmentation was confirmed to be attributable to the presence of a specific protease which could be inhibited not only by the addition of metal ion chelators or serine protease inhibitors, but also by the addition of other recombinant unsaturated IGFBPs, particularly IGFBP-3. IGF-I did not modulate the activity of the IGFBP-4 protease in solution but was able to prevent the inhibition seen with IGFBP-3. Analysis of granulosa cell conditioned medium from the same series of healthy and atretic follicles revealed that IGFBP-2 and -4 were the predominant IGFBPs with no fragments seen on immunoblotting. In contrast, IGFBP-3 in conditioned medium from theca of atretic follicles was always found as an intact doublet, but was found partially fragmented (30 and 32 kDa) in medium conditioned by theca from healthy, dominant follicles with the proportion of IGFBP-3 in this lower molecular weight or fragmented doublet increasing with follicular maturation. A similar situation was also found for IGFBP-4 with the progressive increase in the amount of the 15 and 16.5 kDa fragments. IGFBP-2 was always found to be intact. Finally, IGFBP production from stroma explants was also examined. This revealed a wide variation in IGFBP pattern between different ovaries, although there was a remarkable degree of consistency between different stroma explant cultures from the same ovary. Immunoblotting for IGFBP-3 revealed that, where present, it existed as both an intact and a lower molecular weight doublet and that IGFBP-2 was again always found to be intact. In conclusion we have demonstrated alterations in the proteolytic modification of the IGFBPs which differ in the various follicular compartments and are closely linked to the stage of follicular development.

In this study, we have found that IGF-binding protein-3 (IGFBP-3) in calf serum added to tissue culture medium is degraded by cultured FRTL-5 cells and a major 31 kDa fragment of IGFBP-3 is produced. When FRTL-5 rat thyroid cells were cultured in 6H medium (modified F-12M medium containing TSH, insulin, hydrocortisone, somatostatin, transferrin, and glycyl-histidyl-lysine) containing 5% calf serum, both 44-46 and 31 kDa IGFBPs were found in conditioned medium by ligand blot analysis using 125I-labelled IGF-II. However, predominantly the 44-46 kDa IGFBP was detected in unconditioned 6H medium containing 5% calf serum. When calf serum in the media was replaced by human serum similar results were obtained, and the 44-46 kDa and 31 kDa IGFBPs were recognized using a human IGFBP-3 antibody following Western blot analysis. FRTL-5 cells secreted only small amounts of an endogenous 29 kDa IGFBP, thought to be IGFBP-5. To separate the 31 kDa fragment of IGFBP-3 from the endogenous IGFBP-5, culture media were fractionated by concanavalin-A-Sepharose chromatography and aliquots of both flow-through and eluate from the column were analyzed by ligand blotting. A 31 kDa IGFBP was found in the eluate fractions from concanavalin-A-Sepharose chromatography following the separation of conditioned 6H medium supplemented with calf serum, suggesting that this species was an N-linked glycoprotein and could be derived from the degradation of serum IGFBP-3 by FRTL-5 cells. Using a modified zymographic assay, we examined whether the degradation of IGFBP-3 could depend on the cell membrane. Confluent FRTL-5 cells were washed with PBS and overlaid with liquid agarose solution. After the agarose had solidified, unconditioned 6H medium containing 5% calf serum was incubated with the cells at 37 degrees C for 16 h. Both 44-46 and 31 kDa IGFBP species were found in the overlying, conditioned medium by ligand blot. However, the 31 kDa IGFBP was not found in medium in the absence of FRTL-5 cells, and no IGFBP could be found in serum-free conditioned medium from agarose-covered FRTL-5 cells. This suggests that the 44-46 kDa IGFBP-3 in serum was degraded to yield a 31 kDa fragment, while any endogenous IGFBP-5 could not pass out of the agarose. The degradation of 44-46 kDa IGFBP-3 in the modified zymographic assay was inhibited by phenylmethylsulfonyl fluoride, EDTA, and aprotinin, but not by leupeptin. In summary, these results indicated that IGFBP-3 in calf serum added to culture medium could be degraded by FRTL-5 cells and that this may involve calcium-dependent serine proteases.