The Insulin-Like Growth Factor System Modulates Retinal Pigment Epithelial Cell Tractional Force Generation

Abstract

purpose. The goal of this study was to determine the influence, if any, of the insulin-like growth factors (IGFs) on retinal pigment epithelial (RPE) cell tractional force generation and the contributions of vitreous insulin-like growth factor–binding proteins (IGFBPs) toward control of growth factor activity.

methods. IGF effects on RPE were evaluated in tissue culture assays that involved incubation on three-dimensional collagen matrices with responses measured as progressive reduction in matrix thickness. IGFBP effects were evaluated by using the same system, exposing cells to a non–IGFBP-binding growth factor analogue (R³IGF-I) or IGFBPs alone or in combination with native growth factors.

results. RPE cells generated tractional forces in response to IGF-I and -II with IGF-I being the more potent stimulus. Differential RPE responses to R³IGF-I reflected minor amounts of endogenous IGFBP production. IGFBP-2, -3, and -5 were effective inhibitors of both ligands, whereas IGFBP-6 reduced cell responses to IGF-II only. IGFBP-direct effects on the cells were binding-protein–specific, in that IGFBP-1 had detectable stimulatory effects, and IGFBP-3, -4, -5, and -6 inhibited RPE responses.

conclusions. IGF-I and -II are potent promoters of RPE cell tractional force generation in vitro. The effects of the six high-affinity IGFBPs on RPE responses are generally inhibitory and protein-specific. IGF ligands and binding proteins are known to be present in the vitreous, the environment that drives RPE responses in proliferative vitreoretinopathy (PVR), suggesting that the IGF system plays a potentially important role in the pathophysiology of this fibrocontractive disease.

Proliferative vitreoretinopathy (PVR) is essentially an anomalous wound-healing response in the eye ^{1 2} developing in 5% to 11% of rhegmatogenous retinal detachments. ^{3 4} It is one of the most common causes of failure to correct rhegmatogenous retinal detachment and complicates as many as one-third of all surgical repairs. ^{4 5 6 7 8} The management of this condition is further complicated by the fact that PVR can result in detachment of otherwise successfully reattached retinas or even cause new breaks, necessitating additional corrective surgeries. The most recent, comprehensive review suggests that despite significant advancements in vitreoretinal microsurgical techniques, the incidence of PVR in primary retinal detachments has failed to decline over the past 20 years. ⁹ Although anatomic success with current treatment modalities is reported in 60% to 80% of patients, the functional prognosis is disappointing, with only 40% to 80% of these cases recovering ambulatory vision. ^{4 10} The absence of more effective treatment options is, in part, attributable to an incomplete understanding of PVR’s pathogenesis. 

It is widely accepted that PVR is a dynamic phenomenon involving migration, proliferation, and connective tissue production by cells that gain access to the vitreous cavity. ^{5 7 11 12} The ensuing fibroproliferative biomass, consisting of different cell types enmeshed in an extracellular matrix scaffold, form scarlike tissues referred to as epiretinal or subretinal membranes. Tractional forces developing within these membranes are the result of activities of component cells interacting with the extracellular matrix. ¹³ Among the different cell types implicated in PVR, immunohistochemical and ultrastructural studies have consistently reported the presence of retinal pigmented epithelial (RPE) cells in fibrocellular scars. ^{14 15 16 17} Evidence in support of the causative role for RPE also comes from animal models that have demonstrated the capacity of these cells to cause tractional retinal detachment. ^{18 19 20 21 22}  

Studies of RPE tractional force generation in vitro have revealed a requirement for exogenous promoters in the form of growth factors. ^{18 23} Although several growth factors capable of stimulating RPE tractional force generation are also present in vitreous fluids, ^{24 25 26} it is still unclear which, if any, are responsible for driving PVR-associated matrix contraction. Studies with retinal Müller cells used as a target have revealed that the insulin-like growth factors (IGFs) can stimulate tractional force generation and are present in biologically active quantities in the vitreous fluids of patients with fibrocontractive diseases. ^{27 28 29 30} Of even greater relevance, similar studies with growth factor-neutralizing antibodies have attributed most vitreous contraction-promoting activity to IGF ligands. ^{27 31} Other laboratories examining RPE behavior in culture have demonstrated the capacity of IGFs to stimulate cell proliferation and migration. ^{32 33} However, at present, nothing is known about the influence of vitreous IGFs on RPE tractional force generation. 

Gaining a complete understanding of IGF’s effects on RPE cell behavior is complicated by the potential influence of the insulin-like growth factor binding proteins (IGFBPs). In addition to the two well-characterized ligands, IGF-I and -II, the IGF system contains at least six high-affinity IGFBPs that can inhibit or potentiate growth factor activity through either growth factor sequestration or cell-direct mechanisms. ^{34 35 36 37 38} Extracellular matrix contraction studies exploring Müller cell responses revealed that IGFBPs modulate IGF-stimulated tractional force generation. ³⁹ In addition, there is strong evidence of disease-associated disturbances in vitreous IGFBP levels and that growth factor–binding protein interactions play an important role in determining the net changes in vitreous IGF biological activities. ^{40 41 42 43 44} Based on the reported effects of IGFBPs in these other systems, we hypothesize that IGFBPs may also regulate RPE responses to IGF and constitute a previously unrecognized regulatory system of RPE activity in PVR. With this in mind, the goal of the present study was to investigate the effects of IGF system components on the RPE’s biological activity. Using RPE tractional force generation as a highly sensitive and pathogenically relevant target assay, we systematically examined the effects of IGF ligands and the cell-direct as well as the IGF-binding–dependent effects of the six high-affinity IGFBPs. 

Methods

Isolation and Culture of Porcine RPE Cells

The methods used to procure animal tissues complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Review Board at the University of Alabama at Birmingham. RPE cells were dissociated from porcine eyecups and maintained in culture as previously described. ⁴⁵ Briefly, eyes were enucleated from anesthetized animals and transported to the laboratory in sterile ice-cold normal saline. After removal of the anterior segment of the eye and the vitreous, the posterior eyecup was incubated (Leibovitz L-15 medium; Invitrogen Corp., Carlsbad, CA) at room temperature for 30 minutes, and the retina was detached with forceps and scissors to expose the RPE monolayer. RPE cells were harvested by serial 30-minute incubations at 37°C in 0.25% trypsin (Sigma-Aldrich, St. Louis, MO) in L-15 and collected by repeated gentle trituration with a 1-mL sterile pipette. To facilitate dissociation, trypsin-released RPE cells were incubated with 1% DNase enzyme (DNase I; Sigma-Aldrich) in L-15 for 2 minutes at room temperature. The cells were further purified by density centrifugation on a cushion composed of a single-density gradient (Percoll 40%; GE Healthcare Biosciences Corp., Piscataway, NJ) prepared with 0.01 M Na₂PO₄ and 0.15 M NaCl (pH 7.4). After centrifugation at 500g for 3 minutes, the RPE cells were recovered as a pellet. Cells suspended in growth medium composed of Dulbecco’s minimum essential medium (DMEM; Invitrogen) containing 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; Mediatech, Inc., Herndon, VA), 1% antibiotic-antimycotic (Invitrogen), and 10% fetal bovine serum (Invitrogen) were introduced into 100-mm tissue culture dishes (BD Biosciences, Franklin Lakes, NJ) and incubated at 37°C in a humidified atmosphere composed of 5% CO₂ and 95% air. RPE isolates were routinely characterized by cytokeratin 18 content by our published immunochemical methods. ⁴⁵ For routine culture, growth medium was changed weekly, and the cells were harvested for subculture or experimentation by using 0.05% trypsin and 0.02% EDTA (Invitrogen). Based on our published studies of RPE tractional force generation in vitro, the primary cultures were routinely expanded to passage 5 before use in contraction assays. ¹⁸  

Extracellular Matrix Contraction Assay

Native type I collagen isolated from rat tail tendons by limited pepsin digestion and sequential salt precipitation was used for preparation of collagen gels. ⁴⁶ Acid-soluble collagen in 12 mM HCl was adjusted to physiologic pH and ionic strength with 10× PBS (0.1 M Na₂HPO₄, 1.5 M NaCl) and 0.1 M NaOH, while being maintained on ice. Aliquots (0.2 mL) at 1.8 mg/mL of collagen were added to the center of 12-mm diameter circular scores at the bottom of 24-well tissue culture plates (Corning Inc., Corning, NY) and allowed to polymerize in a humidified atmosphere composed of 5% CO₂ and 95% air at 37°C for 90 minutes. The surface of the polymerized gels were prewetted with 40 μL of contraction medium composed of DMEM with reduced sodium bicarbonate (2.7 g/L) and 1 mg/mL crystalline BSA (Sigma-Aldrich) and incubated for 30 minutes under the conditions just described. Passage 5 or 6 RPE cells were released from subconfluent cultures with trypsin-EDTA treatment, washed with growth medium to inactivate the trypsin, washed again with contraction medium, and then counted electronically (Beckman Coulter, Inc., Miami, FL). Except as otherwise indicated, aliquots (30 μL) of contraction medium containing 5000 cells were applied to the top of the wetted gel surface and incubated at 37°C for 60 minutes to permit cell attachment. After incubation, a 0.68-mL sample of contraction medium containing the experimental additives was added to each well, providing sufficient volume to cover the gel completely. Immediately after flooding, the thickness of each hemispherical gel was measured by an inverted, phase-contrast microscope (TMS model; Nikon Instrument Group, Melville, NY) equipped with a z-axis digitizer (LaSico, Los Angeles, CA) by adjusting the plane of focus from a reference point on the bottom of the well to the cell layer on the gel surface while recording the distance of stage movement. Before contraction, the hemispherical gels were approximately 2.5 mm thick at the center and varied in thickness by <10%. This initial gel thickness served as a reference against which all subsequent gel measurements were compared. The percentage of contraction, referred to throughout the article, was calculated by dividing the remaining gel height by the original height and then subtracting this percentage from 100%. Photomicrographs documenting RPE morphologies on collagen gels were taken at 8 hours of incubation with an inverted phase-contrast microscope equipped with a digital camera system (RETIGA EXi; QImaging Corp., Burnaby, BC, Canada) and assembled into composite photographs with image-management software (Photodeluxe; Adobe Systems, Inc., San Jose, CA). 

Other Reagents

Recombinant human growth factors IGF-I and -II, their analogues R³IGF-I and R⁶IGF-II, and IGFBP-2 were obtained from GroPep, Inc. (Adelaide, SA, Australia). Recombinant IGFBP-1, -3, and -5 were from Upstate Biotechnology, Inc. (Lake Placid, NY). Recombinant IGFBP-4 and -6 were from Austral Biologicals (San Ramon, CA). All growth factors and IGFBPs were reconstituted according to the manufacturer’s instructions. 

Results

RPE Response to IGF System Ligands

To characterize the effects of IGF system ligands on RPE extracellular matrix contraction, cells placed on collagen gels were incubated with 10 nM IGF-I or -II, and the gel thicknesses were measured periodically over 24 hours. Evaluation of contraction kinetics revealed that RPE cells responded to both growth factors, with measurable matrix contraction evident within 2 hours of incubation and an overall reduction in gel thickness of 30% to 40% occurring at 24 hours (Fig. 1) . 

RPE morphology at 8 hours of incubation reflected the different stimuli. Cells incubated without growth factor were attached and extended short processes over the gel surface, but generally remained rounded with only minimal evidence of matrix contraction (Fig. 2A) . In contrast, IGF-I- and -II-stimulated cells were well spread, extended multiple processes, and induced lines of tension in the underlying collagen matrix (Figs. 2B 2C) . 

Contraction assays performed with various concentrations of growth factor revealed that IGF-I and -II drive matrix contraction by RPE in a dose-dependent manner (Fig. 3A) . However, IGF-I concentrations in the dynamic region of the profile (1.5 × 10⁻¹⁰–2.5 × 10⁻⁹M) consistently stimulated larger RPE responses than similar concentrations of IGF-II. For comparative analyses, normalized curves of IGF-I and -II were generated from the 24 hour dose–response profiles by first subtracting the background contraction achieved with the negative control and then dividing by the extent of matrix contraction achieved by the positive control wells containing stimuli (Fig. 3B) . Regression analyses performed to calculate the concentration of each growth factor ligand needed to stimulate half-maximum gel contraction revealed that IGF-I is approximately twice as potent as IGF-II (Table 1) . 

Together, these observations provide reasonably clear evidence that IGF system ligands can drive RPE tractional force generation. However, the data presented in Figure 1indicate that RPE cells incubated without exogenous growth factors still generated significant, albeit lower, amounts of gel contraction. To explore this further, assays measuring RPE cell-number–dependent responses with and without exogenous stimulation were performed (Fig. 4) . The results obtained after 24 hours of incubation revealed a cell-number–dependent response without added stimuli and a consistently increased response to added IGF-I. Whereas these data indicate that an optimal growth-factor–induced response was achieved with approximately 4000 cells per gel, this result also indicated that RPE tractional force generation in vitro is not completely dependent on exogenous stimulation. 

Influence of Endogenous IGFBP Production

To explore the potential influence of endogenously produced IGFBPs, we stimulated RPE on collagen gels with various concentrations of two growth factor analogues reported to have generally reduced affinities for IGFBPs. ^{47 48} The effects of the analogues were compared to native ligands in the same experiment, and the data were normalized to internal controls, to enable direct comparisons. Data obtained after 24 hours of incubation revealed the expected dose-dependent response to increasing concentrations of R³IGF-I (Fig. 5) . R³IGF-I at concentrations between 4.0 × 10⁻¹¹ and 6.2 × 10⁻¹⁰M was found to be more active than IGF-I at equimolar doses. In contrast, no statistically significant differences were detected between IGF-II and its analogue, R⁶IGF-II (data not shown). Regression analyses performed to calculate the concentration of each growth factor yielding half-maximum gel contraction revealed that IGF-I is approximately half as active as its non-IGFBP-binding analogue (Table 1) . These observations indicate that RPE cells in culture produce a relatively small pool of functional IGFBPs that can influence ligand-induced matrix contraction. 

IGFBPs Modulate RPE Responses to IGF-I

To examine the overall effects of IGFBPs on growth factor–stimulated cells, RPE on collagen gels were incubated with an intermediate, stimulatory concentration of IGF-I (5.0 × 10⁻¹⁰M) combined with a 10-fold molar excess of each individual IGFBP. When compared with cells incubated with growth factor alone, significant inhibition was observed with three of the IGFBPs (Fig. 6) . The presence of IGFBP-2 and -3 caused 85% and 96% inhibition, respectively, whereas IGFBP-5 reduced IGF-I stimulation by 48%. In contrast, IGFBP-1, -4, and -6 failed to modulate IGF-I effects to any significant degree. 

RPE cell morphologies after 8 hours of incubation also reflected the different conditions. As before, cells exposed to IGF-I alone extended processes over the matrix, from which lines of matrix under tension were visible (Fig. 7A) . In contrast, cells exposed to IGF-I and a 10-fold molar excess of IGFBP-2 were of an intermediate morphology (Fig. 7B) , in that the cells were clearly less active than the stimulated cells, but showed a higher degree of spreading than unstimulated cells (Fig. 2A) . Repeated phase-contrast microscopic visualizations of the cells throughout the experiment provided no indication of IGFBP-2 cytotoxicity. 

To explore the mechanism of IGFBP inhibition, we examined the potential for cell-direct effects in two additional experiments. To examine potential cell-direct inhibition, we incubated RPE cells attached to collagen gels with IGFBPs as just described, but in this case stimulated with an intermediate concentration of the non–IGFBP-binding analogue R³IGF-I (5.0 × 10⁻¹⁰M). IGFBP-3, -4, -5, and -6 showed a modest, but nonetheless statistically significant, inhibitory effect on R³IGF-I-stimulated RPE contraction (Fig. 8A) . IGFBP-1 and -2 had no effect. The potential for direct stimulatory effects were evaluated by exposing RPE cells to IGFBPs alone (5.0 × 10⁻⁹M), without other growth factor stimuli. In this case, IGFBP-1 induced weak, but statistically significant, matrix contraction, whereas the rest of the IGFBPs had no substantial effect on RPE contractility (Fig. 8B) . Together, these data indicated that IGFBP effects are, for the most part, dependent on growth factor binding, as the growth-factor–independent effects are relatively modest by comparison. 

IGFBP Modulation of IGF Effects on RPE Responses

To precisely quantify and compare the effects of exogenous IGFBPs on RPE cell responses to IGF-I and -II, extracellular matrix contraction assays were performed by incubating RPE cells with IGFBP-growth factor combinations consisting of intermediate stimulatory concentrations of IGF-I (5.0 × 10⁻¹⁰M) or -II (1.0 × 10⁻⁹M) with various concentrations of individual IGFBPs. For comparison of IGFBP-specific activities, data from individual experiments were normalized to internal positive and negative controls and then used to calculate IGFBP dose-inhibition profiles for IGF-I (Fig. 9A)and -II (Fig. 9B) . For the purpose of clarity, Figure 9contains the dose-dependent effects of only four of the six IGFBPs tested, including the maximally and minimally inhibitory IGFBPs as well as two IGFBPs with intermediate effects. The effects of recombinant IGFBPs on IGF-I-mediated RPE contraction were variable, in that IGFBP-2, -3, and -5 had dose-dependent inhibitory effects, whereas IGFBP-1, -4, and -6 had no inhibitory effects at all concentrations tested. Similar assays evaluating the effects of IGFBPs on IGF-II activity revealed that IGFBP-2, -3, and -5 had significant inhibitory effects while IGFBP-1, -4, and -6 had modest or no effects. 

Regression analyses of the IGFBP dose-inhibition profiles enabled calculation of the concentration of each IGFBP yielding half-maximum inhibition (IC₅₀). These values were further normalized to the disparate concentrations of growth factor stimuli to calculate the molar excess (x-fold) of binding protein to growth factor necessary to achieve 100% inhibition (Table 2) . Based on these data, IGFBP-2, -3, and -5 are substantially more effective inhibitors of IGF-II than of IGF-I, as was evident from the greater molar excess concentrations of these binding proteins required for complete inhibition of IGF-I. IGFBP-3 proved to be the most effective inhibitor of IGF-I, whereas IGFBP-5 was the most potent inhibitor of IGF-II. Of potential physiological importance is the observation that IGFBP-6 had no effect on IGF-I, but substantially inhibited IGF-II. 

Discussion

The goal of this study was to examine the influence, if any, of the IGF system on RPE tractional force generation, an activity relevant to PVR pathobiology. Studies to determine RPE responsiveness to IGF ligands indicated that both IGF-I and -II are extremely potent contraction promoters, with IGF-I approximately twice as active as IGF-II. The latter finding is consistent with IGF ligand effects in other experimental systems and is most likely a result of higher IGF-I affinity for its principal cell-surface receptor (IGF-IR). ^{38 49} In addition, these findings have physiologic relevance in that the effective concentration ranges are comparable to the IGF ligand levels reported in normal and pathologic vitreous. ^{41 50 51 52} When considered with the previously reported increases in vitreous IGF activity associated with PVR, ²⁷ these observations constitute strong circumstantial evidence of IGF system involvement in RPE tractional force generation. 

The effects of the IGFBPs on RPE tractional force generation are also of interest, in that we observed ligand-specific and IGFBP-specific effects. The net effect of the IGFBPs on the IGF ligands are generally inhibitory, with more effective attenuation of IGF-II than -I. The latter observation is consistent with reports that the IGFBP affinities for IGF-II are higher than for IGF-I and suggests that the mechanism of action is attenuation of the growth factors through physical association or sequestration. ^{34 36 37} IGFBP-1 was modestly stimulatory, but the mechanism of this effect is as yet unknown. IGFBP-1 direct effects in other system have been attributed to binding protein interactions with the α₅β₁ integrin. ^{53 54 55} However, this study provided no information to indicate whether a similar mechanism is involved in this model. Overall, IGFBP-2 and -3 were the most effective inhibitors of both ligands with measurable effects at concentrations equimolar to those of the growth factors. IGFBP-5 and -6 were somewhat different in that both were reasonably effective inhibitors of IGF-II, but had limited or no effects on IGF-I. Except for the modest cell-direct effects mentioned earlier, IGFBP-1 and -4 did not modulate RPE responses to either growth factor. From this, we conclude that the IGFBPs modulate IGF activities primarily through growth factor binding. 

These observed IGFBP effects on RPE tractional force generation also have potentially important implications for PVR pathobiology. At least two IGFBPs are known to be present in normal vitreous including IGFBP-2 and -3, the most effective inhibitors of IGF-I and -II in our study. ^{56 57} Although IGF-I and -II are also present in normal vitreous, their biological activity is normally undetectable when Müller cells are used in a target assay. ^{27 28 41 42 51 52 58} However, in PVR, vitreous contraction-stimulating activity attributable to these ligands increases. ³¹ Based on our current findings, the increase in IGF activity could arise through increased growth factor concentration, loss of the growth factor sink, or both. Unfortunately, no information is currently available about PVR-associated changes in vitreous IGFBP concentrations, and so it is not yet possible to determine which of the two pathogenic mechanisms are contributory. 

Several other observations merit mention, as they suggest that RPE may be a biologically relevant source of IGF system components. Contraction assays performed with the reduced IGFBP-affinity growth factor analogue R³IGF-I revealed that RPE in culture produce modest, but detectable amounts of functional IGFBPs. This is in accordance with previously published data showing IGFBP expression by human RPE cells ⁵⁹ and more recent work in this laboratory in primary porcine RPE cell cultures (Mukherjee S et al. IOVS 2004;45:ARVO E-Abstract 5355). Although of limited impact in these experiments, continuous production by intraocular cells would likely achieve higher concentrations and may alter vitreous growth factor bioavailability. We also observed that RPE cells are capable of generating significant tractional forces in the absence of added promoters. Based on the previously reported capacity of RPE cells to produce multiple contraction-promoting growth factors, ^{45 60 61} it seems likely that the basal RPE responses are indicative of an autocrine–paracrine stimulatory loop. This may involve IGF-I or -II ^{54 55 56} or other contraction promoting growth factors such as platelet-derived growth factor (PDGF). ^{26 62}  

In sum, our observations suggest that IGF system components in the vitreous may play an important role in driving RPE tractional force generation, a key cell-mediated activity in PVR that culminates in tractional retinal detachment. However, our understanding about these seemingly complex interactions is still limited and requires characterization of changes in the vitreous IGFBP profile as well as their concentrations in PVR. Further investigations are also needed to define the precise intraocular role of individual IGFBPs and to understand the complex IGF–IGFBP interactions in normal and disease states. 

 Supported by the International Retinal Research Foundation (Birmingham, AL), NIH EY013258, Research to Prevent Blindness and EyeSight Foundation (Birmingham, AL).

 Submitted for publication September 13, 2006; revised November 8, 2006; accepted January 18, 2007.

 Disclosure: S. Mukherjee, None; C. Guidr y, 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.

 

RPE matrix contraction in response to IGF-I and -II. Presented is the kinetics of matrix contraction by RPE cells attached to collagen gels and incubated in medium containing 10 nM IGF-I (○), 10 nM IGF-II (•) or no growth factor (□). Each data point represents the mean ± SD of results obtained from triplicate cultures under each condition.

Figure 2.

 

Morphologies of RPE cell contracting collagen matrices under the conditions described in the legend to Figure 1 . Phase-contrast photomicrographs were taken of RPE cells after 8 hours of incubation on collagen matrices in serum-free DMEM (A), 10 nM IGF-I (B), or 10 nM IGF-II (C). Scale bar, 200 μm.

Figure 3.

 

RPE dose-dependent responses to IGF-I and -II. RPE cells attached to collagen gels were incubated in medium containing the indicated concentrations of IGF-I (○) and or IGF-II (•). Presented are the 24 hour dose–response profiles (A) and the data normalized to internal controls to permit direct comparison (B). *Statistically significant differences between the concentrations in the experimental samples were assessed by paired Student’s t-tests, with all differences defined by P < 0.05. Each data point in these plots represents the mean ± SD of results obtained from triplicate cultures under each condition.

Table 1.

 

Comparison of Insulin-Like Growth Factor and Analogue Activities

	C50 (nM)	r corr
IGF-I vs. IGF-II
IGF-I	0.80	0.963
IGF-II	1.74	0.969
IGF-I vs. R3 IGF-I
IGF-I	0.56	0.960
R3 IGF-I	0.27	0.969

 C₅₀ refers to the concentration of ligand necessary to induce a half-maximum response and is calculated after subtracting the background response from the maximum response of gel contraction observed. For comparison of C₅₀ concentrations between the native ligands and the ligands with their respective analogues, dose-response experiments for the comparative groups were performed on the same day. These data points were calculated from regression analyses using a minimum of six data points for each ligand.

Figure 4.

 

Effect of the number of cells on RPE tractional force generation. RPE cells attached to collagen gels in the indicated numbers were incubated with 0.5 nM IGF-I (○) or without any growth factor stimuli (•). Each data point represents the mean ± SD of results obtained from triplicate cultures under each condition.

Figure 5.

 

RPE matrix contraction in response to IGF-I and R³IGF-I. RPE cells attached to collagen gels were incubated in medium containing the indicated concentrations of IGF-I (○) or R³IGF-I (•). Presented are the normalized curves generated from 24 hour dose–response profiles. *Statistically significant differences between the concentration values of the experimental samples were assessed by paired Student’s t-tests, with all differences defined by P < 0.05. Each data point in these plots represents the mean ± SD obtained from results of triplicate cultures under each condition.

Figure 6.

 

Evaluation of IGFBP effects on IGF-I-stimulated RPE matrix contraction. RPE cells attached to collagen gels were incubated for 24 hours in medium containing a 5-nM concentration of each IGFBP combined with 0.5 nM IGF-I. Also presented are the responses of representative positive and negative controls incubated with 0.5 nM IGF-I alone (+) or without any stimuli (−). Each bar represents the mean ± SD of the results obtained from triplicate cultures under each condition. *Statistically significant differences in contraction responses between the experimental samples was assessed by paired Student’s t-tests, with all differences defined by P < 0.05.

Figure 7.

 

Morphologies of RPE cells on collagen matrices in the presence of experimental additives as described in the legend to Figure 6 . Phase-contrast photomicrographs were taken of RPE cells after 8 hours of incubation in medium containing 0.5 nM IGF-I alone (A) or IGF-I with a 10-fold molar excess of IGFBP-2 (B). Scale bar, 200 μm.

Figure 8.

 

Evaluation of IGFBP cell-direct effects on RPE matrix contraction. RPE cells attached to collagen gels were incubated for 24 hours in medium containing a 5-nM concentration of each IGFBP combined with 0.5 nM R³IGF-I (A), or no added growth factor (B). Also presented are the responses of positive and negative controls incubated with 0.5 nM R³IGF-I alone (+) or no stimulus (−). Each bar represents the mean ± SD of the results obtained from triplicate cultures under each condition. *Statistically significant differences in contraction responses between the experimental samples was assessed by paired Student’s t-tests, with all differences defined by P < 0.05.

Figure 9.

 

IGFBP dose-dependent inhibition of IGF-I and -II stimulation. RPE cells attached to collagen gels were incubated with 0.5 nM IGF-I (A) or 1.0 nM IGF-II (B) and the indicated concentrations of IGFBP-1 (○), IGFBP-2 (•), IGFBP-3 (□), and IGFBP-5 (▪). Presented are the normalized dose-inhibition profiles obtained after 24 hours of incubation, representing the mean ± SD obtained from triplicate cultures under each condition.

Table 2.

 

Summary of IGFBP Inhibition of IGF-I and -II

IGFBP	0.5 nM IGF-I IC50 (nM)	r corr	100% Inhibition (X-Fold Molar excess)	1 nM IGF-II IC50 (nM)	r corr	100% Inhibition (X-Fold Molar excess)
IGFBP-1	NS			NS
IGFBP-2	2.19	0.97	12.38	2.52	0.98	6.44
IGFBP-3	1.37	0.96	10.68	2.41	0.99	6.94
IGFBP-4	NS			NS
IGFBP-5	5.21	0.91	21.4	2.29	0.98	5.36
IGFBP-6	NS			4.07	0.94	9.71

 IC₅₀ refers to the calculated concentration of ligand necessary to induce a half-maximum inhibition using the maximum response of gel concentration observed less the background response. 100% Inhibition refers to the calculated molar ratio of IGFBP to growth factor necessary to achieve complete inhibition. n/s indicates a statistically insignificant inhibitory effect of the indicated IGFBPs.