Insulin-Like Growth Factor Binding Proteins Modulate Müller Cell Responses to Insulin-Like Growth Factors

Abstract

purpose. Müller cells are consistently identified in diabetic fibrocontractive ocular tissues and, in response to insulin-like growth factor I, generate tractional forces of the type that cause retinal detachment. Recent studies suggest that diabetes-associated increases in vitreous insulin-like growth factor activity cannot be attributed to simple increases in concentration alone, suggesting that more complex biochemical changes in vitreous growth factor control mechanisms are involved. The goal of this study was to evaluate the contributions of vitreous insulin-like growth factor–binding proteins (IGFBPs) toward control of growth factor activity.

methods. Native and recombinant IGFBPs effects were evaluated on IGF-I– and -II–stimulated Müller cells in tissue culture assays that involved cell incubation on three-dimensional collagen gels and that monitored progressive matrix condensation. IGFBP degradation by Müller cell–secreted proteases was assessed in Western ligand blots, and direct stimulatory effects were evaluated by incubating cells with IGFBPs alone.

results. IGFBP direct stimulatory effects on Müller cells were significant, but relatively modest, and IGFBP modulation through Müller cell–secreted proteases was undetectable. In contrast, IGFBP inhibitory effects on IGF-I and -II were highly variable and, in some cases, profound. IGFBP-3 effectively inhibited IGF-I and -II stimulation with detectable effects at concentrations equimolar to the growth factor. IGFBP-1, -2, -4, and -5 were of intermediate effectiveness as inhibitors, 3- to 11-fold less active than IGFBP-3. IGFBP-6 had virtually no inhibitory effects on IGF-I, but was moderately effective against IGF-II.

conclusions. IGFBP effects on IGF-I– and -II–stimulated Müller cells are primarily inhibitory with only modest direct stimulatory effects of limited physiologic relevance. IGFBP-2 and -3, the major binding proteins identified in vitreous, most likely function as the vitreous growth factor sink and control ligand activity through sequestration.

Diabetes affects nearly 16 million people in the United States alone and remains the leading cause of vision loss in adults between the second and seventh decades of life. ^{1 2 3} The risk of development of diabetic retinopathy increases with disease duration, as retinopathic changes are apparent in 98% of individuals with type 1 and 78% with type 2 diabetes within 15 years of diagnosis. During this same period, 33% of persons with type 1 and 17% of those with type 2 diabetes will also have proliferative diabetic retinopathy (PDR), a complication that typically begins with neovascularization of the optic disc, retina, or iris and can progress to fibrocontractive processes from which emerge tractional forces that cause retinal detachment and blindness. 

A substantial body of information suggests that Müller cells, the principal retinal glia, play causal roles in this disorder. Müller cells can respond to retinal injury, insult or changes in vitreous growth factors through altered expression of cytoskeletal proteins and enzymes, cellular hypertrophy, proliferation, and migration. ⁴ Electrophysiological and immunohistochemical studies of the retina in diabetes and diabetes-related animal models revealed Müller cell responses in advance of overt retinopathy. ^{5 6 7 8 9} Immunohistochemical studies of diabetic epiretinal membranes consistently identify Müller cells as a component of these proliferative tissues, ^{10 11} and studies of Müller cells in animal models and tissue culture revealed the capacity of this cell type to generate tractional forces and cause retinal detachment. ^{12 13}  

Müller cell proliferation is stimulated by a fairly broad panel of mitogens, including basic fibroblast growth factor, epidermal growth factor, nerve growth factor, platelet-derived growth factor (PDGF), insulin-like growth factor (IGF)-I. and thrombin. ^{14 15 16} However, Müller cell tractional force generation is induced by a more restricted panel of growth factors including IGF-I and members of the PDGF family. ^{12 17} Consistent with the latter profile of growth factor responsiveness, vitreous concentrations of IGF-I increase in diabetes by 150% to 300% over control cells, to concentrations well above our observed threshold of Müller cell sensitivity. ^{18 19 20 21 22 23 24 25}  

While these observations are compelling evidence supporting the role of IGF-I–induced Müller cell responses in diabetic retinopathy, an interesting paradox emerges when the relationship between vitreous growth factor concentrations and biological activities are considered. In general, Müller cells are not stimulated to generate tractional forces by normal human vitreous, a result easily accepted from a pathogenic point of view. ²⁶ However, immunochemical studies of normal and nondiabetic vitreous samples report IGF-I concentrations that are also above the threshold of Müller cell sensitivity, suggesting that IGF-I activity is controlled or attenuated in normal vitreous. ^{19 20 21 23 24 25} Another recently completed study from this laboratory examined Müller cell tractional force generation in response to 68 diabetic vitreous samples, revealing 700% to 900% increases in growth factor activity over controls. ²⁷ The magnitude of this increase in biological activity grossly exceeds the increases in IGF-I concentrations reported above, leading to additional speculation about the role of diabetes-associated disturbances in the putative vitreous IGF-I control mechanism. 

In addition to the two low-molecular-weight ligands IGF-I and -II, the IGF system contains at least six high-affinity IGF-binding proteins (IGFBPs) and IGFBP function-modulating proteases. ^{28 29 30} IGFBPs vary in size, tissue distribution, ligand affinity, and biological activity, and, depending on the experimental system used, can inhibit or potentiate growth factor activity through growth factor binding and cell-direct mechanisms. At least two IGFBPs are known to be present in normal vitreous and, based on IGFBP effects reported in other systems, are probably involved in the putative vitreous IGF-I control mechanism. ^{28 31} There are also reports of diabetes-associated disturbances in the normal complement of vitreous IGFBPs ^{20 23 24 32 33} potentially arising through changes in local production, ^{28 31 34 35} effusion from plasma through changes in vascular permeability, ^{24 36 37} or degradation by IGFBP-specific proteases present in vitreous. ^{38 39}  

While compelling as a hypothesis, evidence in support of the proposed IGFBP function in vitreous is still circumstantial, as the effects of IGFBPs on Müller or other ocular cells are unknown. When one considers the wide range of IGFBP effects reported in other cell types, it may be that IGFBP species inhibit or potentiate IGF activity and may have direct stimulatory or inhibitory effects on ocular cells. ^{29 30 40 41} With this in mind, the goal of this study was to extend our understanding of the role of vitreous IGFBPs by defining biological activities and enable interpretation of disease-related changes. Using Müller cells and their capacity to generate tractional forces in response to IGF-I as a biologically relevant target assay, we systematically examined the effects of the six high-affinity IGFBPs on growth factor activities, evaluating direct and indirect effects as well as the capacity of Müller cells to modulate IGFBP biological activity through secreted proteases. 

Methods

Isolation and Culture of Porcine Müller Cells

Müller cells were isolated from normal porcine retina and maintained in culture by using previously published methods. ¹² For the experiments described herein, cells were used at culture passages 4 through 10. In the authors’ opinions, the methods used to secure animal tissue were humane and complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham. 

Extracellular Matrix Contraction

Müller cell responses to IGF-I and -II and the influence of IGFBPs were evaluated with a previously described assay of Müller cell extracellular matrix contraction. ¹⁷ Recombinant human IGF-I and IGF-II and native IGFBP-1 and recombinant IGFBP-2 were obtained from GroPep Ltd. (Adelaide, Australia). Recombinant IGFBP-1, -3, and -5 were obtained from Upstate Biotechnology, Inc. (Lake Placid, NY). Recombinant IGFBP-4 and -6 were from Austral Biologicals (San Ramon, CA). IGFBPs were suspended in phosphate-buffered saline containing 1 mg/mL crystalline BSA and growth factors were dissolved in 12 mM HCl containing 1 mg/mL crystalline BSA. To permit protein–protein interactions, IGFBP growth factor preparations in culture medium were incubated at 37°C for 60 minutes before assay. 

Western Ligand Blots

Experiments to evaluate IGFBP degradation by Müller cell–secreted proteases in culture were performed using the same procedures and reagents as for contraction assays, except that cell cultures containing IGFBP-3 and -5 were prepared in polypropylene vessels to reduce adsorptive loss of proteins. Samples of recombinant IGFBPs or culture media aliquots were combined with nonreducing Laemmli sample buffer, ⁴² heated to 100°C for 3 minutes, and separated on 10% sodium dodecyl sulfate-polyacrylamide gels at 20 mA. Proteins were transferred to nitrocellulose membranes (Hybond ECL; Amersham Pharmacia Biotech, Piscataway, NJ) for 2 hours at 100 V and blocked with 3% BSA in Tris-buffered saline (TBS, 0.15 M NaCl, 0.02 M Tris-HCl; pH 7.8) for 60 minutes at room temperature. IGFBPs were detected with 20 ng/mL biotinylated IGF-II (GroPep, Ltd.) in TBS with 0.05% Tween-20 (TBST) for 3 hours at room temperature. Bound IGF-II was detected with 10 ng/mL horseradish peroxidase-conjugated avidin (NeutraAvidin; Pierce, Rockford, IL) in TBST for 60 minutes at room temperature. Chemiluminescence development was with a commercial substrate (Super Signal West Femto; Pierce), used according to the manufacturer’s instructions, and the membranes were exposed to autoradiograph film (Hyperfilm; Amersham). 

Results

Müller Cell Reactivity to Insulin-like Growth Factor System Ligands

As Müller cell extracellular matrix contraction in response to IGF-II had not been characterized previously, assays were performed with Müller cells incubated in various concentrations of this ligand and, for comparison, IGF-I. Kinetic studies revealed that Müller cells were equally responsive to 1 nM IGF-I or 1 nM IGF-II, with matrix contraction evident within 2 to 3 hours, yielding a cumulative 60% reduction in gel thickness within 24 hours (Fig. 1A) . In contrast, cell incubation without added growth factor resulted in gel contraction of less than 10%. Dose–response profiles generated after 24 hours of incubation revealed that IGF-I at lower concentrations consistently stimulated greater Müller cell responses than comparable concentrations of IGF-II (Fig. 1B) . Regression analysis were performed to calculate the concentration of each ligand yielding half-maximum gel contraction (C₅₀), revealing that IGF-II was approximately one-third as potent as IGF-I (Table 1) . 

Direct Effects of IGFBP on Müller Cells

Native IGFBP-1 (IGFBP-1n) and recombinant IGFBP-1 through -6 were obtained from commercial suppliers, and ligand reactivities were evaluated by Western ligand blot analysis. Major growth factor–binding species were present at positions consistent with protein monomers and dimers at the reported relative migrations for these IGFBPs (Fig. 2A) . ⁴³ As the capacity of Müller cells to modulate IGFBP biological activity through secreted proteases is of potential relevance to vitreous changes in diabetes and represents a confounding variable in IGFBP biological activity measurements, potential degradation was evaluated by adding recombinant IGFBP-1 to -6 to extracellular matrix contraction assays prepared with and without cells. Western ligand blots of media samples and densitometry of the exposed films revealed that the maximum IGFBP loss under assay conditions was 37% of IGFBP-2, with no evidence of increased IGFBP proteolysis in the cultures containing Müller cells (Fig. 2B) . Müller cell responses to IGFBPs alone were evaluated in extracellular matrix contraction assays in which the cells were incubated in various concentrations of each binding protein without other growth factor stimuli. Twenty-four hour dose–response profiles revealed that IGFBP-2 and -6 induced modest, but statistically significant, matrix contraction (Fig. 3) , conclusions supported by regression analyses of the dose–response profiles in which positive slopes and linear correlations of more than 0.9 were evident for only IGFBP-2 and -6 (Table 2) . These data were compared to the ligand dose–response profiles presented in Figure 1 to calculate the growth factor equivalence as a measure of biological relevance, revealing that the stimulatory effects of IGFBP-2 and -6 are functionally equivalent to relatively modest 1- to 2-pM concentrations of IGF-I. 

IGFBP Modulation of Müller Cell Responses to IGF-I and -II

IGFBP effects on Müller cell responses to IGF-I and -II were evaluated by combining moderate stimulatory concentrations of IGF-I and -II with 10-fold molar excesses of each IGFBP in extracellular matrix contraction assays. The kinetics of Müller cell responses and variations in cell morphology observed with IGF-I and IGFBP-1 are presented in Figures 3 and 4 to illustrate the binding protein’s effects. Müller cells exposed to 0.1 nM IGF-I were stimulated to contract the collagen matrix in a time-dependent fashion, resulting in a 48% reduction in gel thickness within 24 hours, while cells incubated without growth factor had a lower response of approximately 10% (Fig. 4) . Cells incubated with 0.1 nM IGF-I and a 10-fold molar excess of IGFBP-1 had an intermediate response of 32%, suggesting that IGFBP-1 modestly inhibits IGF-I activity under these conditions. Müller cell morphologies after 8 hours of incubation reflected the different stimuli, in that cells incubated without growth factor were attached, but remained rounded and lacked overt evidence of contractile activity (Fig. 5A) . In contrast, cells exposed to IGF-I alone extended processes over the gel surface, from which lines of matrix under tension were evident (Fig. 5B) . Cells exposed to IGF-I and IGFBP-1 were of intermediate morphology, in that the cells had a higher degree of spreading than growth factor–free control cells, but were less active than the cells incubated with growth factor alone (Fig. 5C) . The effects of IGFBP-2 to -5 were highly variable, in that IGFBP-2 and -4 significantly reduced Müller cell responses to IGF-I with complete inhibition by IGFBP-3 (Fig. 6A) . In contrast, the effects of IGFBP-5 and -6 were not significant. Similar 10-fold molar excesses of IGFBP-1 to -4 attenuated Müller cell responses to 0.1 nM IGF-II, but in this case, IGFBP-5 and -6 inhibition was comparable to that of the other binding proteins (Fig 6B) . Together, these data indicate that the effects of IGFBPs on Müller cell responses to IGF-I and -II are variable and, in some cases, ligand specific. 

To quantify and compare IGFBP-specific activities, we performed extracellular matrix contraction assays with intermediate stimulatory concentrations of IGF-I and -II (0.1 nM) and various concentrations of each IGFBP. Included in these studies were evaluations of IGFBP-1 in native and recombinant forms, as variations in native IGFBP-1 phosphorylation reportedly alters the effects of this binding protein. ⁴⁴ For clarity, 24-hour dose-dependent effects for three of the seven IGFBPs tested are presented in Figure 7 , including the IGFBPs with maximum and minimum effects and one with an intermediate effect. IGFBP-3 was by far the most effective inhibitor of IGF-I, with detectable effects at concentrations equimolar to those of IGF-I (Fig. 7A) . Comparable IGFBP-1 effects were observed at concentrations approximately 10-fold higher than that of IGFBP-3, and IGFBP-6 inhibition was observed at only the highest concentration tested. As with IGF-I, IGFBP-3 was the most potent inhibitor of IGF-II (Fig. 7B) . However, in this case, IGFBP-1 appeared to be even less effective against IGF-II, whereas IGFBP-6 inhibition was pronounced. To enable direct comparisons of IGFBP activities, data from individual assays were normalized to internal positive and negative controls and then used to calculate IGFBP dose-inhibition profiles for IGF-I (Fig. 7C) and -II (Fig. 7D) . Regression analyses of these data yielded percent inhibition per mole of IGFBP and these values were normalized to the disparate concentrations of growth factor stimuli to calculate the molar ratio of binding protein to growth factor necessary to achieve 100% inhibition (Table 3) . IGFBP-3 was determined to be the most effective inhibitor of IGF-I and -II with complete inhibition of at 2.5 and 1.1 molar excesses, respectively. Of potential physiologic relevance is the observation that IGFBP-2 to -6 were all substantially more effective inhibitors of IGF-II than of IGF-I, particularly IGFBP-5 and -6 which showed no substantial effect on IGF-I, but effectively inhibited IGF-II. Finally, the effects of native versus recombinant IGFBP-1 were not significantly different on either ligand. 

Discussion

The primary goal of this study was to evaluate the effects of individual IGFBPs on IGF-I and -II stimulation by using Müller cell responsiveness as a biologically relevant target. With regard to direct effects independent of growth factor binding, IGFBP stimulation of Müller cells was measurable, but relatively modest compared with that of low concentrations of biologically relevant growth factors. In contrast, IGFBP inhibitory effects on IGF-I and -II-stimulation were variable and, in some cases, profound. IGFBP-3 was an extremely potent inhibitor of IGF-I and -II with measurable effects at concentrations equimolar with the growth factor. IGFBP-5 and -6 effects were also of particular interest, in that these binding proteins lack effects on IGF-I, but are effective inhibitors of IGF-II stimulation. The remaining binding proteins can be characterized as inhibitors of intermediate effectiveness on IGF-I and -II, with the general observation that each binding protein seemed to inhibit IGF-II activity more effectively than that of IGF-I. From these data, we conclude that the net effect of IGFBPs on Müller cells appears to be inhibition of IGF-I and -II activities and that this effect is most likely achieved through growth factor sequestration. 

The relative activities of the IGFBPs are, in some respects, consistent with the equilibrium constants reported for IGFBP association with IGF-I and -II. ⁴⁰ In general, IGFBP affinities for IGF-II are higher than for IGF-I which is consistent with our observations. Similarly, the affinity of IGFBP-6 for IGF-I is approximately 25 times lower than that for IGF-II which likely accounts for the paucity of IGFBP-6’s effect on IGF-I stimulation. However, one important difference detected in this study was the exceptional effectiveness of IGFBP-3 in relation to that of the other binding proteins. Based on reported equilibrium constants, IGFBP-3’s affinity for IGF-I is higher than that of IGFBP-1 and -2, but lower than IGFBP-4 or -5. Clearly, these values are not consistent with the biological effects measured in this study. Similarly, IGFBP-3’s affinity for IGF-II was approximately equal to that of the other binding proteins, except for IGFBP-2, with an approximately 10-fold lower affinity. When one considers the potential mechanisms that might account for IGFBP-3’s effects, is seems most likely that there are unique direct effects on Müller cells not detectable in the current experimental design. Several previously published studies with different models provide evidence of at least two plausible mechanisms to account for this difference. Competitive ligand binding studies revealed that IGFBP-3 can interact directly with the IGF1R and inhibit IGF-I binding to its receptor, thus modulating growth factor effect without sequestration. ⁴⁵ Yet another, potentially more relevant, series of studies reported IGFBP-3 differential inhibition of several unrelated growth factors. In addition to IGF-I, exogenous IGFBP-3 inhibited chick embryo fibroblast mitosis in response to fibroblast growth factor and transforming growth factor-β, but had no effect on PDGF stimulation. ^{46 47} Although the mechanism is incompletely characterized, these studies represent compelling evidence of IGFBP-3 direct inhibitory effects that are independent of growth factor affinity. Müller cell experiments with IGF analogues lacking binding protein affinity are needed to elucidate activities of these types and are currently under way. 

Although speculative, it is possible to interpret the effects of individual IGFBPs on vitreous growth factor activities. As described in the introduction, studies from a number of different laboratories reported the presence of IGF-I and -II in normal vitreous with disease-associated increases of 150% to 300% yielding IGF-I concentrations ranging from 0.15 to 0.89 nM and IGF-II from 3.1 to 6.0 nM. Several laboratories have also reported binding protein and growth factor concentrations in the same control and diabetic populations. Two studies from the same laboratory reported detecting IGFBP-1 in vitreous fluids from nondiabetic and diabetic surgical cases at mean concentrations of 0.4 and 1.6 ng/mL, respectively. ^{20 48} This translates into 16 and 64 pM, which would be sufficient to attenuate 2% to 3% of vitreous IGF-I and less than 1% of vitreous IGF-II activities. These values suggest that vitreous IGFBP-1 is unlikely to play a significant role in modulating vitreous growth factor activity in normal or diabetic vitreous. This does not appear to be the case with IGFBP-2. Meyer-Schwickerath et al. ²³ measured IGF-I, IGF-II, and IGFBP-2 concentrations in the same population of cadaveric control and diabetic samples and reported mean binding protein concentrations of 85 and 260 ng/mL, which translates into vitreous molarities of 2.74 and 8.39 nM. ²³ In this case, there is sufficient binding protein to attenuate 44% and 39% of IGF-I activity or 12% and 21% of IGF-II activities in normal and diabetic vitreous, respectively. The results achieved with IGFBP-3 are even more promising when one considers the higher levels of biological activity observed. In the same study, IGFBP-3 measurements showed mean levels of 83 and 250 ng/mL or 1.9 nM and 57 nM in the control and diabetic vitreous samples. This is sufficient to control 414% and 347% of IGF-I or 51% and 86% of IGF-II activities in normal and disease states. Similar analysis of IGFBP-3 and growth factor concentrations reported by Spranger et al. ²⁵ yielded similar results including 306% and 296% inhibition of IGF-I or 23% and 40% inhibition of IGF-II activities in diabetic and control samples. Waldbillig et al. ³² detected IGFBP-4 in canine vitreous as well as increases in galactose-fed animals. However, these investigators were unable to detect IGFBP-4 in Western blot analysis of human vitreous leaving the role of this binding protein uncertain. Similarly, we are unaware of any studies reporting the presence or absence of IGFBP-5 or -6 in human vitreous. 

There is little doubt that growth factor inhibition is the major effect of the IGFBPs known to be present in human vitreous. However, when one considers the sum of IGFBP inhibitory effects and the reported vitreous concentrations of IGF-I and -II, it is also clear that there are insufficient quantities of binding protein to control IGF-I and -II activities in normal or diabetic vitreous. This conclusion is inconsistent with a recent study by Simo et al. ⁴⁸ reporting that free IGF-I concentrations in diabetic vitreous representing approximately 10% of the total growth factor, indicating that 90% is sequestered. Although their study did not examine levels of free IGF-II, as IGFBPs affinities for IGF-II are consistently higher than IGF-I, it seems reasonable to predict that free IGF-II concentrations would be even lower than IGF-I. In short, these data strongly suggest that the capacity of diabetic vitreous to sequester growth factors is higher than that indicated by our measurements, and this is also likely to be true for normal vitreous. Possible explanations for this discrepancy are numerous and, at the least, include the essentially untested potential for biologically active quantities of IGFBP-4, -5, or -6 in vitreous. Other possibilities include indirect IGFBP effects similar to those mentioned for IGFBP-3 or even synergistic effects between binding proteins. Yet another confounding element is that vitreous IGFBP-3 is an approximately 30-kDa protein thought to be generated from a full-length 45-kDa protein by local proteases. ³⁹ Although not specifically examined, it is possible that the biological activity of this truncated version mimics the ∼30-kDa fragment of plasma IGFBP-3 associated with pregnancy and that growth factor affinities are somewhat lower than that of the intact protein. ⁴³  

In sum, although several important questions were addressed in this study, it is also clear that our understanding of the complex interactions of IGF system components in vitreous is incomplete. Additional studies are needed to define fully the role of individual IGFBP species in modulating growth factor activities in normal and disease states. 

 Supported by the Juvenile Diabetes Research Foundation, International (New York, New York), International Retinal Research Foundation (Birmingham, Alabama), National Eye Institute EY13258 and Research to Prevent Blindness.

 Submitted for publication January 20, 2004; revised February 10, 2004; accepted March 4, 2004.

 Disclosure: J.L. King, None; C. Guidry, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Müller cell responsiveness to IGF system ligands. Müller cells attached to collagen gels were incubated in various concentrations of IGF-I (○) or IGF-II (•) or without growth factor stimuli (□). (A) Kinetics of gel contraction in cultures incubated with 1 nM IGF-I or -II or without growth factor stimuli. (B) Dose–response profiles obtained after 24 hours of incubation. Each data point represents the mean and standard deviations of results obtained from triplicate cultures under each condition.

Table 1.

 

Comparison of Insulin-Like Growth Factor Activities

Ligand	C50	r corr
IGF-I	5.26 × 10−11	(0.985)
IGF-II	1.65 × 10−10	(0.990)

 C₅₀ refers to the calculated concentration of ligand necessary to induce a half-maximum response (35.4%) using a maximum response of 65% gel contraction less a background response of 5.9%. These data were calculated from regression analyses using a minimum of six data points at two concentrations for each ligand.

Figure 2.

 

IGFBP characterization and degradation by Müller cells. Commercial preparations of native IGFBP-1 (1n) and recombinant IGFBP-1 through -6 were separated by electrophoresis, transferred to nitrocellulose membranes, and probed in Western ligand blots with biotinylated IGF-II (A). Left: positions of prestained molecular weight standards. Recombinant IGFBP-1 to -6 were added to extracellular matrix contraction assay media (1 μg/mL) and incubated with and without Müller cells. At the times indicated (B), aliquots were removed for subsequent analysis by Western ligand blotting as in (A).

Figure 3.

 

IGFBP direct stimulation of Müller cell responses. Müller cells attached to collagen gels were incubated in the indicated concentrations of IGFBP-2 (○) or IGFBP-6 (•). Presented are the dose–response profiles obtained after 24 hours of incubations representing the means and standard deviations obtained from triplicate cultures under each condition.

Table 2.

 

Summary of IGFBP Direct Stimulation Activities

	% Contraction per Mole	r corr	IGF-I Molarity Equivalent
IGFBP-1	−6.06 × 108	0.44	—
IGFBP-2	1.16 × 1010	0.99*	2.28 × 10−12
IGFBP-3	−2.41 × 108	0.23	—
IGFBP-4	−1.80 × 107	0.01	—
IGFBP-5	5.79 × 108	0.49	—
IGFBP-6	1.19 × 109	0.91*	1.79 × 10−12

 Percent contraction per mole was calculated from regression analysis of 24-hour dose–response curves, using a minimum of four concentrations, each tested in triplicate.

*  Correlations above 0.9 (asterisks) are considered significant.

Figure 4.

 

Kinetics of Müller cell responses to IGF-I modulated by IGFBP-1. Müller cells attached to collagen gels were incubated in medium containing 0.1 nM IGF-I (○), 0.1 nM IGF-I with 1 nM IGFBP-1n (•), or no growth factor (□). Presented are the means and standard deviations of results obtained from triplicate cultures under each condition.

Figure 5.

 

Morphologies of Müller cells under the conditions described in Figure 3 . Phase-contrast photomicrographs were taken of Müller cells after 8 hours of incubation in medium without added growth factors (A), 0.1 nM IGF-I alone (B), or 0.1 nM IGF-I with 1 nM IGFBP-1n (C).

Figure 6.

 

Evaluation of IGFBP effects on ligand-stimulated Müller cells. Müller cells attached to collagen gels were incubated for 24 hours in medium containing 0.1 nM IGF-I (A) or 0.1 nM IGF-II (B) with 10-fold molar excesses of the indicated IGFBPs. Also presented are the responses of positive and negative control cultures incubated with growth factor alone (+) or without stimuli (−), respectively. Each bar represents the means and standard deviations of results obtained from triplicate cultures under each condition.

Figure 7.

 

IGFBP dose-dependent inhibition of IGF-I and -II stimulation. Müller cells attached to collagen gels were incubated with 0.1 M IGF-I (A) or 0.2 M IGF-II (B) and the indicated concentrations of native IGFBP-1 (○), IGFBP-3 (•), and IGFBP-6 (□). Presented are the averages and standard deviations of results obtained after triplicate cultures obtained after 24 hours of incubation under each condition. The percent inhibition for IGF-I (C) and IGF-II (D) is calculated by subtracting the basal responses of cultures without growth factor stimuli, dividing by the responses of cultures exposed to growth factor alone, and then subtracting from 100.

Table 3.

 

Summary of IGFBP Inhibition of IGF-I and IGF-II

	0.1 nM IGF-I % Inhibition per Mole	r corr	100% Inhibition (Molar Ratio)	0.2 nM IGF-II % Inhibition per Mole	r corr	100% Inhibition (Molar Ratio)
IGFBP
IGFBP-1n	6.49 × 1010	(0.95)	15.4	1.39 × 1010	(0.90)	35.9
IGFBP-1r	6.06 × 1010	(0.92)	16.5	1.53 × 1010	(0.90)	32.7
IGFBP-2	1.20 × 1011	(0.98)	8.3	1.66 × 1011	(0.93)	3.0
IGFBP-3	4.01 × 1011	(0.97)	2.5	4.97 × 1011	(0.98)	1.1
IGFBP-4	5.09 × 1010	(0.92)	19.6	6.59 × 1010	(0.90)	7.6
IGFBP-5	2.04 × 1010	(0.91)	50.0	1.18 × 1011	(0.91)	4.2
IGFBP-6	7.17 × 109	(0.76)	139.5	7.41 × 1010	(0.95)	6.7

 Percent inhibition/mole IGFBP was calculated from regression analyses of the data presented in Figure 7 , using a minimum of three binding protein concentrations, each in triplicate. 100% inhibition refers to the calculated molar ratio of IGFBP to growth factor necessary to achieve complete inhibition normalized to growth factor concentration tested by division.