PDGF- and Insulin/IGF-1–Specific Distinct Modes of Class I_A PI 3-Kinase Activation in Normal Rat Retinas and RGC-5 Retinal Ganglion Cells

Abstract

purpose. To compare PDGF- and insulin/IGF-1–induced class I_A PI 3-kinase/Akt survival signaling in normal retinas and retinal ganglion cells (RGCs).

methods. Normal rat retinas and RGC-5 cells were used for (1) immunohistochemical and immunoblot studies to detect PDGF receptor (PDGFR) subtypes and (2) immunoprecipitation, immunoblot, and in vitro lipid kinase assays to determine basal and PDGF-induced class I_A PI 3-kinase/Akt survival signaling, in comparison with insulin or IGF-1 responses. Furthermore, RGC-5 cells were exposed to broad-spectrum (LY294002) or p110 isoform-selective (PI-103) PI 3-kinase inhibitors (versus Akt inhibitor) to assess the consequent effects on Akt phosphorylation, caspase-3/PARP cleavage, apoptotic phenotype, and cell viability, as a function of serum trophic factors.

results. PDGFR-α and -β immunoreactivity was observed in rat retinal Müller cells and in the RGC layer and blood vessels, respectively. In addition, PDGFR-α and -β protein expression was observed in RGC-5 cells. Both retinas and RGC-5 cells exhibited a similar pattern of subunit-specific basal class I_A PI 3-kinase activity, which was stimulated in a temporal and signal-specific manner by PDGF and insulin/IGF-1. Furthermore, RGC-5 cells showed PDGFR-α/β tyrosine phosphorylation that induced the p85α regulatory subunit to activate p110α/β-associated class I_A PI 3-kinase, which in turn enhanced Akt phosphorylation. Exposure of serum-deprived RGC-5 cells to PI 3-kinase or Akt inhibitors increased susceptibility to apoptotic phenotype as revealed by caspase-3 and PARP cleavage.

conclusions. The present findings provide direct evidence of two distinct modes of retinal class I_A PI 3-kinase activation that occurs in response to PDGF receptor and insulin/IGF-1 receptor stimulation. PDGF-induced PI 3-kinase/PIP3/Akt axis may provide new therapeutic approaches to ameliorate cell death in diabetic retinopathy and other retinal neurodegenerations.

Previous studies have shown that intraretinal platelet-derived growth factor (PDGF) has beneficial and detrimental effects toward cell survival/maintenance under physiological conditions and aberrant phenotypic changes under pathologic conditions, respectively. ^{1 2 3 4 5 6 7 8 9 10} From a broad perspective, PDGF exerts autocrine and paracrine effects that facilitate the survival of retinal ganglion cells (RGCs), ¹¹ neuroretina, ^{11 12 13 14 15} and retinal microvasculature. ^{2 3 4 5 6 7 9 10 11 14 15 16} Notably, PDGF has its principal origin in RGCs during development and in the mature retina, ^{1 7 11} and the diminution in retinal PDGF expression is associated with RGC death. ¹¹ In addition, it is clear that retinal PDGF ligands and/or PDGF receptor (PDGFR) subtypes undergo marked alterations in expression profile and localization pattern, which are intimately associated with retinal diseases, including diabetic retinopathy ^{4 5 7 8} and neurodegenerative and microvascular disorders. ^{5 11 15} Although subtype-specific retinal PDGF ligands and PDGFR have been extensively characterized in health and disease, ^{1 2 7 11 17 18} it remains unknown whether PDGF interaction with its specific receptors leads to sequential activation of downstream survival signaling pathways in the normal retinas and retinal cells, including RGCs. 

Retinal expression and dysregulation of all four PDGF polypeptide chains (PDGF-A, -B, -C, and -D) encoded by four different genes have been reported. For more than a decade, retinal layer–specific changes in endogenous PDGF-A and -B mRNA/protein or the consequences of ectopically expressed or gene-targeted retinal PDGF-A and -B have been studied extensively. ^{1 2 4 7 13 14 19} Since the discovery of PDGF-C and -D, ^{20 21} recent reports have highlighted the implications of altered retinal PDGF-C or -D levels on retinal intercellular communication and survival. ^{22 23} Active PDGF ligands exist in the form of homodimers or heterodimers that include PDGF-AA, -AB, -BB, -CC, and -DD. ²¹ These PDGF dimers bind with two different PDGFR subtypes such as PDGFR-α and/or -β, thereby facilitating its homodimerization to PDGFR-αα or -ββ or heterodimerization to PDGFR-αβ. ²¹ PDGF-BB homodimer is unique in that its interaction results in the formation of all three dimeric forms of the activated PDGFR. ²¹ In view of the complexity of retinal PDGF ligands, ^{1 2 16} PDGF receptors, ^{1 2 7 14} and PDGF ligand-receptor interactions, it is critically important to understand the pan-PDGF receptor-mediated activation of downstream survival signaling. In this regard, it is pertinent to examine PDGF-BB-induced activation of key retinal survival kinases that include phosphoinositide 3-kinase (PI 3-kinase) and its downstream effector, protein kinase B (Akt). 

Previous observations from our laboratory and other investigations have underscored the importance of PI 3-kinase activation toward retinal survival in vivo and in vitro based on the use of the broad-spectrum PI 3-kinase inhibitors LY294002 and wortmannin. ^{24 25 26 27 28 29 30 31 32 33 34} Although LY294002 treatment inhibits Akt activation and thus induces retinal apoptosis, recent studies involving cancer cell PI 3-kinase inhibition with the p110 subunit–specific PI 3-kinase inhibitor PI-103 ^{35 36 37} have revealed diminished Akt phosphorylation without the induction of apoptosis. ³⁶ To ascertain the causal link between PI 3-kinase activation and retinal cell survival, further studies are clearly needed, to compare the effects of LY294002 and PI-103 on the retinal cell apoptotic phenotype. 

The neuroprotective effects of PDGF have been demonstrated in the context of cerebellar neuronal cell survival. ^{38 39 40} The cell-permeable tyrosine phosphorylated PDGFR peptide activates PI 3-kinase and Akt signaling to induce cell survival. ⁴⁰ In addition, Stitt et al. ⁹ have shown that advanced glycation end-product–mediated retinal pericyte apoptosis is reversed by PDGF-BB-induced increases in Akt phosphorylation. The increases in Akt phosphorylation may occur through differential activation of p110α and/or p110β subunit–specific PI 3-kinase by PDGF ^{35 41 42 43 44} or insulin, ^{35 37 41 42 43 44} as observed in adipocytes or aortic endothelial cells. Thus, it is important to compare the effects of retinal PDGF and insulin/IGF-1 survival signaling in terms of PI 3-kinase isoform activation and its direct correlate with Akt phosphorylation. PI 3-kinase is categorized into three major classes that include class I_A/_B, class II, and class III isoforms. Class I_A PI 3-kinase is a heterodimer that consists of regulatory (p85α, p85β, p55γ, p55α, or p50α) and catalytic subunits (p110α, p110β, or p110δ). ⁴⁵ Class I_A PI 3-kinase catalyzes the phosphorylation of phosphatidylinositol 4,5-bisphosphate (PIP2) lipid substrate to form phosphatidylinositol 3,4,5-trisphosphate (PIP3) lipid products in vivo. Of importance, class I_A PI 3-kinase isoforms constitute the major source for the generation of PIP3 lipids, which facilitate the activation of serine/threonine protein kinase, Akt. ^{46 47 48} Elucidation of basal and PDGF-mediated retinal PI 3-kinase subunit–specific activity would therefore constitute the essential first step toward determining absolute increases in key lipid kinase survival signaling in the retinas or retinal cells. 

In the present study, we used normal rat retinas and RGC-5 cells in culture to examine the PDGFR subtype expression and its tyrosine phosphorylation. In addition, we determined the basal and PDGF-BB-induced class I_A PI 3-kinase isoform activity and Akt phosphorylation in parallel with insulin and/or IGF-1 response. To confirm the causal link between retinal PI 3-kinase activation and cell survival, we determined the effects of selective and nonselective PI 3-kinase/Akt inhibitors on Akt phosphorylation, caspase-3 activation, and poly(ADP-ribose) polymerase (PARP) cleavage. The results strongly suggest that class I_A PI 3-kinase is critically important for retinal cell survival and that PDGF has the potential to exhibit neuroprotective effects by activating class I_A PI 3-kinase/Akt prosurvival signaling. 

Materials and Methods

Recombinant human platelet-derived growth factor-BB (PDGF-BB) and insulin-like growth factor-1 (IGF-1) were purchased from R&D Systems (Minneapolis, MN). Human insulin (Novolin R) was from Novo Nordisk Pharmaceuticals, Inc. (Princeton, NJ). The primary antibodies for PDGFR-α (cat no.: 3164), PDGFR-β (28E1) (cat no.: 3169), phospho-PDGFR β (Tyr751), phospho-IGF-1 receptor/insulin receptor, phospho-Akt (Ser473), phospho-Akt (Thr308), Akt, cleaved caspase-3, and cleaved PARP were purchased from Cell Signaling (Beverly, MA). The primary antibodies for PDGFR-α (cat no. 07-276), Thy-1 (clone OX-7), phosphotyrosine (clone 4G10), p85α (N-SH2, clone UB93-3), p85(pan), p110α, and p110β were purchased from Upstate Biotechnology/Chemicon/Millipore (Temecula, CA). The primary antibodies for phosphotyrosine (PY99), IRS-1, IRS-2, insulin receptor-β, and IGF-1 receptor were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The primary antibody for β-actin was purchased from Novus Biologicals (Littleton, CO). The primary antibody for vimentin was purchased from Neomarkers (Freemont, CA). The primary antibody for agrin was purchased from Stressgen (Ann Arbor, MI). Silica gel 60 TLC plates were purchased from EMD Biosciences (San Diego, CA). [γ-³²P]ATP (specific activity: 4500 Ci/mmol) was purchased from MP Biomedicals (Solon, OH). All other chemicals were from Fisher Scientific (Fair Lawn, NJ) or Sigma-Aldrich (St. Louis, MO). 

Animals and Ex Vivo Retinal Explant Studies

Male Sprague-Dawley rats (Charles River, MA), weighing approximately 150 to 175 g, were housed in a 12-hour light–dark cycle with standard food and water ad libitum. The rats were maintained in the Penn State College of Medicine animal facility in conformity with the Institutional Animal Care and Use Committee guidelines and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Rats were subjected to ketamine/xylazine anesthesia (4 mg/kg–0.4 mg/kg) and killed by decapitation. 

For in vivo experiments, retinas were removed and snap frozen in liquid N₂ followed by storage at −80°C. For ex vivo explant studies, the retinas were maintained in Dulbecco’s modified Eagle’s medium in a humidified atmosphere of 95% air and 5% CO₂, as described. ^{49 50} After stimulation with PDGF (30 ng/mL; 1.2 nM) or insulin (30 nM), control and treated retinas were washed with ice-cold PBS and snap frozen in liquid N₂ and stored at −80°C. 

Preparation of Retinal Homogenates

Control and treated retinas were homogenized with a microtube disposable pestle and cordless motor (Kontes; Fisher Scientific), with buffer A (consisting of 50 mM Tris-HCl [pH 7.5], 0.1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 50 mM sodium fluoride, 10 mM sodium β-glycerophosphate, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, protease inhibitor cocktail [Sigma-Aldrich], 0.1% β-mercaptoethanol, and 1 μM microcystin-LR). Aliquots of the supernatants of retinal homogenates were used for immunoprecipitation/in vitro kinase assays or treated with Laemmli sample buffer for immunoblot analysis. 

Immunohistochemistry

Immunohistochemical analyses were performed using frozen sections of rat retina, as described. ^{49 50} The primary antibodies used were PDGFR-α and -β. For colocalization studies, the primary antibodies for Thy-1 (marker for ganglion cell layer and inner plexiform layer), vimentin (marker for Müller cells), and agrin (marker for blood vessels) were also used. 

Cell Culture and Treatments

The rat retinal ganglion cell line (RGC-5) was obtained as a gift from Neeraj Agarwal (University of North Texas Health Science Center, Fort Worth, TX). RGC-5 cells have been shown to express ganglion cell markers and exhibit the characteristics of ganglion cells in culture. ^{51 52} RGC-5 cells (passages up to 10) were maintained in culture using Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA) along with 10% fetal bovine serum (HyClone, Logan, UT) and antibiotic/antimycotic solution (Sigma-Aldrich) in a humidified atmosphere of 95% air and 5% CO₂, as described. ⁵¹ After they attained confluence, the RGC-5 cells were trypsinized and seeded onto 100-mm/60-mm Petri dishes or six-well plates. The subconfluent cells were then subjected to serum deprivation (24 hours) and exposed to PDGF (1 ng/mL [0.04 nM], 10 ng/mL [0.4 nM], and 30 ng/mL [1.2 nM]), IGF-1 (30 ng/mL [4.0 nM]), or insulin (1 nM, 10 nM, and 30 nM), as described in the legends to the respective figures. Control and treated cells were washed with ice-cold PBS and snap frozen using buffer A (for immunoprecipitation/in vitro kinase assays) or lysed with Laemmli sample buffer (for immunoblot analysis). 

Immunoprecipitation of PDGFR-α, PDGFR-β, INSR, IGF-1R, Tyrosine Phosphorylated Proteins, IRS-1/2, or PI 3-Kinase Regulatory or Catalytic Subunits

Retinal tissue homogenates or RGC-5 lysates obtained using buffer A were sonicated (15 seconds × 4) and centrifuged at 14,000 rpm (4°C) for 10 minutes. The respective supernatants were then used for protein assays (Coomassie protein reagent; Pierce, Rockford, IL). The aliquots of supernatants (60 μg protein) were subjected to immunoprecipitation (4°C, overnight) with 2 μg each of anti-PDGFR-α, anti-PDGFR-β, anti-INSR, anti-IGF-1R, anti-phosphotyrosine, anti-IRS-1/2, anti-p85α, anti-p85(pan), anti-p110α, or anti-p110β primary antibody preconjugated (2 hours at 4°C) to G Sepharose (GammaBind; GE Healthcare, Piscataway, Nj). The immune complexes were then washed with buffer A and TNE buffer (consisting of 10 mM Tris-HCl [pH 7.4] 150 mM NaCl, 5 mM EGTA, and 0.1 mM sodium orthovanadate) before PI 3-kinase assays, as described. ^{49 50}  

In Vitro PI 3-Kinase Assays

PI 3-kinase assays were performed as described, with slight modifications. ^{49 50} After immunoprecipitation of proteins using specific primary antibodies, the respective immunocomplexes were subjected to PI 3-kinase assays by incubation at 35°C for 10 minutes in the presence of 50 μL TNE buffer (pH 7.4), phosphatidylinositol substrate (20 μg/assay), and γ-[³²P]ATP (10 μCi/assay). The reactions were stopped by adding 20 μL of 6 N HCl and 160 μL of CHCl₃/CH₃OH (1:1). Subsequently, the assay tubes were vortexed for 20 seconds and centrifuged at 14,000 rpm (room temperature) for 5 minutes. The phospholipid-containing lower organic phase from the respective reaction tubes was recovered and spotted onto silica gel thin-layer chromatography (TLC) plates (that were preheated to 100°C for ∼1 hour). The TLC plates were then subjected to ascending chromatography by using the freshly prepared solvent mixture CHCl₃/CH₃OH/H₂O/NH₄OH (60:47:11.3:2). Phosphatidylinositol 3-phosphate (PI3P) spots thus resolved were visualized and quantified by autoradiography and phosphorescence imager analyses (Phosphorimager; Molecular Dynamics, Sunnyvale, CA), respectively. As negative control experiments, mock immunoprecipitations were performed using lysis buffer, which revealed negligible formation of ³²P-labeled PI3P. 

Immunoblot Analysis

Retinal tissue and RGC-5 cell lysates (10 μg protein each) were electrophoresed with precast 4% to 12% minigels (NuPage; Invitrogen), and the resolved proteins were transferred to nitrocellulose membranes (Hybond C; GE Healthcare). The membranes were blocked and probed with the respective primary antibodies. After extensive washing, the immunoreactivity was detected by using specific HRP-conjugated secondary antibodies followed by enhanced chemiluminescence (GE Healthcare), as described. ⁴⁹ The protein bands were quantified by using a calibrated densitometer (GS-800; Bio-Rad, Hercules, CA). 

Cell Viability

RGC-5 cells (with or without FBS) were exposed to PI 3-kinase or Akt inhibitors as described earlier, and cell viability was assessed with a cell-viability assay (Live-Dead assay; Invitrogen-Molecular Probes, Eugene, OR). In brief, the control and treated cells in 24-well plates were incubated with 1 μM calcein-acetoxy methyl ester and 4 μM ethidium homodimer-1 for 30 minutes at room temperature in the dark. The green fluorescence of live cells and red fluorescence of dead cells were simultaneously determined by using a fluorescence plate reader. 

Statistical Analyses

Results are expressed as the mean ± SEM. Statistical analyses of the data were performed by repeated-measures one-way ANOVA followed by Bonferroni t-test. P < 0.05 was considered statistically significant. 

Results

PDGF Receptor Subtype Expression, Tyrosine Phosphorylation, and Association with PI 3-Kinase p85α Regulatory Subunit in Rat Retinas and RGC-5 Cells

Previous in vivo studies have shown that retinal ganglion cells or retinal neuronal cell bodies express and secrete constitutively active PDGF to exert paracrine survival effects on retinal astrocytes or glial cells. ^{7 11 13 19} To date, the potential autocrine or exogenous effects of PDGF on retinal/RGC survival signaling has not yet been examined. As an initial step to address this question, we examined the PDGFR subtype expression in intact rat retinas and the RGC-5 cells. As shown in Figures 1A 1B 1C 1D , immunohistochemical analyses of normal rat retinal cryosections revealed PDGFR-α immunoreactivity in RGC layer and Müller cells, and PDGFR-β immunoreactivity in blood vessels. Thy-1, vimentin, and agrin were used as markers for RGC layer, Müller cells, and blood vessels, respectively. Immunoprecipitation/immunoblot studies of the supernatants of retinal tissue homogenates showed PDGFR-α and -β expression, PDGF stimulation of PDGFR-α and -β tyrosine phosphorylation, and the consequent increases in p85α recruitment (Figs. 1E 1F) . In comparison, immunoblots of RGC-5 cell lysates with specific primary antibodies revealed the expression of both PDGFR-α and -β isoform proteins (Fig. 1G) . In addition, RGC-5 cells showed PDGF-induced PDGFR-α and -β tyrosine phosphorylation and the resultant p85α recruitment (Fig. 1H 5A 5B) . Together, these data demonstrate that PDGF stimulation of normal retinas and RGC-5 cells resulted in enhanced interaction of tyrosine phosphorylated PDGFR-α and -β subtypes with PI 3-kinase p85α regulatory subunit. 

Basal Class I_A PI 3-Kinase Activity in Rat Retinas and RGC-5 Cells

Before investigating the potential for PDGF to induce retinal PI 3-kinase/Akt survival signaling, we compared the ability of in vivo rat retinas and RGC-5 cells to exhibit increases in basal class I_A PI 3-kinase activity. To address this question, we determined the basal lipid kinase activity in normal rat retinas and serum-deprived (24 hours) RGC-5 cells. Figure 2shows the representative TLC spots that illustrate the relative increases in PI3P formed from basal class I_A PI 3-kinase activity associated with the following immunoprecipitates (IP): PDGFR-α (A) and PDGFR-β (B), INSR (C), IGF-1R (D), pTyr (E), IRS-1/2 (F), IRS-1 (G), IRS-2 (H), p85α (I), p85(pan) (J), p110α (K), and p110β (L). Visual investigation of the in vivo retinal and in vitro RGC-5 PI 3-kinase TLC spots in parallel reveals comparable increases in the overall spectrum of basal PI 3-kinase activities between the retinas and RGC-5 cells. In particular, the amounts of PI3P recovered from anti-p85 and anti-p110 immunocomplexes were noticeably higher within retinas and RGC-5 cells than that observed with anti-phosphotyrosine or anti-PDGFRα/β immunocomplexes. Similar findings of increased basal p85- or p110-associated PI 3-kinase compared with phosphotyrosine-associated PI 3-kinase activity have been observed in several tissues and cell types, including skeletal muscle, liver, embryonic fibroblasts from p85 knockout mice and smooth muscle cells. ^{53 54 55} Nevertheless, variable affinities of the immunoprecipitating antibodies preclude direct comparison of PDGFRα/β, INSR, IGF-1R, IRS-1/2, or subunit-specific basal PI 3-kinase activity. 

Concentration-Dependent Effects of PDGF and Insulin on RGC-5 Akt Phosphorylation, a Downstream Signal for Class I_A PI 3-kinase

As an initial step to determine agonist-induced PI 3-kinase/Akt signaling, we performed concentration-dependency experiments with RGC-5 cells. As shown in Figure 3 , acute challenge with PDGF (0.04–1.2 nM) for a fixed time (6 minutes) increased Akt phosphorylation as a function of concentration. Even with the lowest concentration, PDGF produced significant increases in both Akt(Ser473) and Akt(Thr308) phosphorylation and the maximum response was observed between 0.4 and 1.2 nM. In parallel, insulin induced significant increases in Akt phosphorylation with 1 nM concentration, and the maximum response was attained at concentrations between 10 and 30 nM. Although PDGF and insulin at lower concentrations showed synergism on both Akt(Ser473) and Akt(Thr308) phosphorylation (Fig. 3A) , statistical analysis of these data revealed no significant synergistic effects (Figs. 3B 3C) . 

To compare the individual effects of PDGF and insulin on PI 3-kinase/Akt activation in retinal tissues and RGC-5 cells in culture, we used 1.2 nM PDGF and 30 nM insulin for all subsequent experiments. 

Time Dependency of PDGF, Insulin, or IGF-1 Stimulation of Class I_A PI 3-Kinase/Akt Signaling in Rat Retinas and RGC-5 Cells

Next, we compared the temporal increases in agonist-specific class I_A PI 3-kinase activity in rat retinas and RGC-5 cells. Stimulation of ex vivo rat retinas with PDGF or insulin for 6, 20, and 60 minutes (Fig. 4A)led to a maximum increase in PI 3-kinase activity or Akt phosphorylation at the 6-minute time point (Fig. 4A , Table 1 ). PDGF stimulation produced significant increases in p85α-associated PI 3-kinase activity that was accompanied by increased Akt(Ser473) and Akt(Thr308) phosphorylation. As expected, aliquots of PDGF-stimulated retinal tissue homogenates did not show significant increases in IRS-1/2-associated PI 3-kinase activity. In contrast, insulin induced significant increases in p85α- and IRS-1/2-associated PI 3-kinase activity, indicating recruitment of the p85α regulatory subunit to the IRS-1/2 adapter molecules on insulin receptor stimulation and subsequently enhanced Akt(Ser473) and Akt(Thr308) phosphorylation. These data suggest that IRS-dependent and IRS-independent mechanisms contribute to activation of p85α-associated class I_A PI 3-kinase in the retina. 

Next, the time-dependent increases in PI 3-kinase/Akt signaling were examined in serum-deprived RGC-5 cells after stimulation with PDGF, IGF-1, or insulin (Fig. 4B) . In other studies, anti-p85 or anti-phosphotyrosine immunocomplexes have been used to determine PDGF-induced increases in PI 3-kinase activity in several cell types. ^{53 56 57} Hence, we used anti-phosphotyrosine immunocomplexes to determine PDGF-induced PI 3-kinase activity in RGC-5 cells. At the 6-minute time point, PDGF stimulation led to maximum increases in PDGFR-β tyrosine phosphorylation, phosphotyrosine-associated PI 3-kinase activity, and Akt(Ser473) and Akt(Thr308) phosphorylation. Insulin stimulation also led to maximum increases in insulin receptor tyrosine phosphorylation and IRS-1-associated PI 3-kinase activity after 6 minutes, with sustained increases in insulin-induced Akt(Ser473) and Akt(Thr308) phosphorylation for up to 60 minutes. IGF-1 stimulation showed PI 3-kinase/Akt activation profiles similar to that of insulin. 

Figure 4Cillustrates the temporal increases in PDGF- and insulin-induced Akt(Ser473) phosphorylation in rat retinas and RGC-5 cells. Comparison of PDGF and insulin stimulatory response at the 6-minute time point reveals that the elevations of corresponding Akt phosphorylations in rat retinas were significant but relatively lower than that observed with serum-deprived RGC-5 cells. It is conceivable from the previously published reports that the observed differences in the inducible Akt phosphorylation in intact retinas may be attributable to the intraretinal existence of preformed PDGF ligands. In addition, PDGF or insulin stimulatory response within the retina may include both agonist-responsive (e.g., RGC) or nonresponsive cell types that may contribute to the overall modest elevations in the whole retina. 

Together, the present data suggest that both retinas and RGC-5 cells possess functional PDGF receptors and insulin/IGF-1 receptors that activate class I_A PI 3-kinase/Akt survival signaling. 

Comparison of PDGF, Insulin, and IGF-1 Stimulation of Class I_A PI 3-Kinase Activity Associated with Proximal Signaling Components and p85/p110 Regulatory/Catalytic Subunits in RGC-5 Cells

To understand the distinct modes of class I_A PI 3-kinase activation that occurs in RGC-5 cells after stimulation with PDGF or insulin/IGF-1, we thought it is important to determine the PI 3-kinase activity associated with the proximal signaling components and regulatory/catalytic subunits. ^{53 54 55 56 57} To investigate the role of specific proximal signaling components, we examined agonist-induced PI 3-kinase activity associated with the tyrosine phosphorylated proteins (that include activated PDGFR-α/β, insulin receptors, or IGF-1 receptors) or IRS-1/IRS-2 adapter molecules (recruited by activated insulin or IGF-1 receptors). To decipher the contribution of regulatory and catalytic subunits toward PI 3-kinase activation by PDGF versus insulin/IGF-1, we determined agonist-induced PI 3-kinase activity associated with the regulatory (p85α or p85pan) or catalytic (p110α or p110β) subunits. 

Figures 5A and 5Bshow that PDGF stimulation of RGC-5 cells enhanced PDGFR-α- and -β-specific tyrosine phosphorylation and the recruitment of p85α regulatory subunit to the tyrosine phosphorylated PDGFR subtypes. In addition, PDGFR-α and -β activation resulted in enhanced PI 3-kinase activity by 11.4 ± 1.2- and 48.3 ± 3.4-fold, respectively (P < 0.05, n = 3). PI 3-kinase assays of the immunocomplexes of tyrosine phosphorylated proteins revealed robust increases in PDGF-induced PI3P formation (∼38-fold), compared with modest increases in IGF-1 and insulin responses by four- and threefold, respectively (Fig. 5C , Table 2 ). Notably, the differential activation of PI 3-kinase by PDGF and insulin has been demonstrated in adipocytes, wherein PDGF primarily induces activation of plasma membrane-associated PI 3-kinase but insulin preferentially activates microsomal PI 3-kinase. ⁴¹ In conformity with the previous observations in adipocytes, ⁴¹ the present study shows a preponderance of p85α recruitment and its tyrosine phosphorylation by PDGF unlike insulin/IGF-1. Thus, immunoblot studies of the respective immunocomplexes of tyrosine phosphorylated proteins (Figs. 5C 1F)showed (1) PDGFR-α and -β subtype tyrosine phosphorylation and marked p85α recruitment and tyrosine phosphorylation; (2) IGF-1 receptor, IRS-1, and IRS-2 tyrosine phosphorylation; and (3) insulin receptor, IRS-1, and IRS-2 tyrosine phosphorylation. In addition, immunoblot analysis studies using the respective immunocomplexes of p85α regulatory subunit (Fig. 5D)showed protein tyrosine phosphorylation in the molecular weight region corresponding to PDGF receptors in PDGF-treated RGC-5 cells and protein tyrosine phosphorylation in the molecular weight region corresponding to IRS in insulin-treated RGC-5 cells. These findings are in conformity with the previous observations in PDGF- or insulin-treated adipocytes. ⁵⁷ In addition, these data suggest that PDGF receptor proteins and IRS proteins that contain Tyr-Xaa-Xaa-Met (YXXM) motifs undergo tyrosine phosphorylation, ^{46 47 48} which in turn recruits SH2-domain containing p85α regulatory subunit for the respective PI 3-kinase activation in RGC-5 cells. As a sequel to this observation, PI 3-kinase assays and immunoblot analysis studies using the immunocomplexes of IRS-1 or IRS-2 adapter molecules revealed the inability of PDGF to induce IRS-1- or IRS-2-associated PI 3-kinase due to insufficient p85α recruitment (Figs. 5C 5D 5E 5F , Table 2 ). In contrast, IGF-1 and insulin induced proportionate increases in IRS-1- or -2-associated PI 3-kinase activity as a function of p85α recruitment. These data clearly suggest that RGC-5 cells possess the characteristic features of PDGF- and insulin/IGF-1-specific proximal signaling mechanisms that contribute to two distinct modes of PI 3-kinase activation. In other words, RGC-5 cell PDGF receptor α/β tyrosine phosphorylation recruits p85α to produce robust increases in PI 3-kinase activity, whereas insulin/IGF-1 receptor–mediated IRS-1/IRS-2 tyrosine phosphorylation promotes p85α recruitment to augment PI 3-kinase activity. 

It should be noted that anti-phosphotyrosine, anti-p85, and anti-p110 immunocomplexes have been used in other studies to determine PDGF, insulin, or IGF-1-induced increases in PI 3-kinase activity in several cell types. ^{53 54 55 56 57} In addition, PDGF and insulin have been shown to regulate downstream signaling events in aortic endothelial cells via use p110α and p110β catalytic subunits, respectively. ⁴² Hence, it is important to examine whether PDGF and insulin/IGF-1 differentially regulate p85- or p110-associated PI 3-kinase activity in RGC-5 cells. As shown in Figures 5G 5H 5I 5Jand Table 2 , PI 3-kinase assays and immunoblot studies using the immunocomplexes of p85α, p85(pan), anti-p110α, and antip110β revealed that PDGF and insulin/IGF-1 induced a similar elevation of subunit-specific PI 3-kinase activity, and p85α association with p110α or p110β catalytic subunit remained essentially the same after PDGF or insulin/IGF-1 challenge, which confirms a previous report that p85 regulatory subunit is bound to the p110 catalytic subunit in a stoichiometric ratio of 1:1. ⁵⁸ These data demonstrate that PDGF and insulin/IGF-1 produce independent but redundant increases in p85- and/or p110α/β-associated class I_A PI 3-kinase activity in RGC-5 cells. 

Thus, PI 3-kinase activity assays at the level of proximal signaling components and p85/p110 regulatory/catalytic subunits further support the view that PDGF (in addition to insulin or IGF-1) is critically important for class I_A PI 3-kinase survival signaling in retinal neurons. 

Comparison of the Effects of p110 Isoform-Selective Versus Broad-Spectrum PI 3-Kinase Inhibitors on Retinal Akt Phosphorylation and RGC Akt Phosphorylation/Apoptotic Markers

We and several other investigators have used broad-spectrum nonspecific PI 3-kinase inhibitors (LY294002 and wortmannin) to demonstrate that PI 3-kinase/Akt signaling is critical for retinal cell survival. ^{24 25 26 27 28 29 30 31 32 33 34} The recent development of cancer-therapeutic small molecules has led to further studies that involve treatment of cancer cell lines with p110-catalytic subunit–specific novel PI 3-kinase inhibitor (PI-103), to determine the physiological role of subcellular PI 3-kinase. ^{35 36 37} Unlike LY294002, PI-103, which inhibits Akt phosphorylation, does not induce apoptosis in cancer cells, thereby raising questions about the causal link between PI 3-kinase inactivation and apoptotic phenotype. ³⁶ These important observations emphasize the need to compare the effects of LY294002 and PI-103 on retinal cell apoptotic phenotype. 

As shown in Figure 6A , retinal tissue exposure to either PI-103 or LY294002 led to abrogation of PDGF- or insulin-induced phosphorylation of Akt(Ser473) and Akt(Thr308). These data suggest that retinal PI 3-kinase p110-catalytic subunit inhibition would attenuate downstream Akt activation. 

To determine the consequence of Akt inhibition on apoptotic phenotype, we maintained RGC-5 cells in complete medium (+FBS) or serum-free conditions (−FBS) and exposed them to increasing concentrations of PI-103 or LY294002. As positive controls, we used an Akt inhibitor (Akt-X) that causes selective inhibition of Akt phosphorylation without affecting upstream kinases including PI 3-kinase. ⁵⁹ As shown in Figures 6B 6C 6D , serum deprivation of RGC-5 cells decreased basal Akt(Ser473) phosphorylation without the induction of apoptotic markers, cleaved caspase-3, and PARP. However, RGC-5 exposure to PI-103, LY294002, or Akt-X led to complete inhibition of basal Akt(Ser473) phosphorylation, and marked induction of caspase-3 and PARP cleavage. The apoptotic effects of PI 3-kinase/Akt inhibitors were much more pronounced in serum-deprived conditions. Furthermore, live–dead cell assays revealed that PI-103 and LY294002 did not cause significant changes in cell viability at the indicated concentrations, although Akt-X at 30-μM concentrations induced cytotoxicity. These data from PI-103 inhibitor studies further confirm the importance of the PI 3-kinase p110-catalytic subunit in the maintenance of retinal cell survival. 

To determine the contribution of PI 3-kinase toward PDGF- or insulin-mediated retinal cell survival, PI-103-pretreated RGC-5 cells were exposed to the indicated concentrations of PDGF or insulin. As shown in Figures 6E and 6F , PI-103-induced increases in caspase-3 and PARP cleavage were significantly diminished but not completely reversed by either PDGF or insulin. These data clearly suggest that PDGF- or insulin-induced retinal neuronal survival is dependent, in part, on the functionally active PI 3-kinase. In addition, PDGF and superphysiological concentrations of insulin may also exert antiapoptotic effects through PI 3-kinase-independent signaling pathways. 

Discussion

The developmental and species-specific changes in the retinal PDGF ligand and PDGFR subtype mRNA, protein, and localization have been extensively characterized in previous studies. ^{1 2 4 6 7 8 9 10 11 12 13 14 15 16 18} In the present study, we compared retina- and RGC-specific PDGF receptor expression, tyrosine phosphorylation, and its direct functional relationship with class I_A PI 3-kinase survival signaling. In conformity with previous findings from postmortem and rodent retinas, ^{1 2 7 18} the present observations from rat retinas confirm the localization of PDGFR-α and PDGFR-β subtype immunoreactivity in RGC layer and Müller cells and in blood vessels, respectively. In addition, we have demonstrated the ability of PDGF to induce retinal PDGFR-α and -β tyrosine phosphorylation, a key event for subsequent downstream survival signaling. Furthermore, RGC-5 cells express both PDGFR-α and -β subtypes and exhibit PDGFR-α and -β tyrosine phosphorylation in response to PDGF, a neurotrophic survival signal. It should be noted that RGC-5 cells express PDGFR-β isoform, whereas the rat retinal ganglion cells in vivo do not exhibit detectable immunoreactivity for this isoform. However, recent studies of human embryonic retinal sections have demonstrated the expression of PDGFR-β isoform in the neural retina. ¹⁸ In this respect, our current observations from the RGC-5 cell culture model delineate the importance of PDGFR-β and PDGFR-α-mediated class I_A PI 3-kinase activation toward retinal neuronal cell survival. 

The functional relevance of PDGF in the retina has hitherto been studied in the context of RGC as the principal source of PDGF, ^{1 7 11} neuronal development, ^{1 12} neuron-astrocyte/glial cell interactions, ^{12 13 19} and pericyte recruitment and microvasculature integrity. ^{6 10 15 60} Although the downstream survival signaling events have not been elucidated in those studies, the present findings underscore the importance of class I_A PI 3-kinase as a key intermediary signaling event that promotes PDGF-induced retinal/RGC survival. 

The present study provides the first direct evidence of the spectrum of retinal PI 3-kinase subunit-specific activity under basal and PDGF- versus insulin/IGF-1-stimulated conditions (see Fig. 7 ). Of importance, the pattern of basal PI 3-kinase regulatory subunit-, catalytic subunit-, and adapter molecule-specific enzyme activity is similar between in vivo rat retinas and in vitro RGC-5 cells in culture. With regard to developmental regulation of basal PI 3-kinase expression profile, O’Driscoll et al. ⁶¹ recently demonstrated that mature mouse retinas, in comparison with postnatal retinas, exhibit decreased mRNA/protein expression levels of p85α regulatory subunit, increased mRNA and decreased protein expression levels of p110α, p110β, and p110δ catalytic subunits, in concurrence with diminished Akt phosphorylation. In conformity with these key observations on the existence of basal retinal Akt phosphorylation, the present study provides further evidence of retinal and RGC basal lipid kinase activity associated with upstream signaling components, including p85 regulatory and p110 catalytic subunits and IRS-1/2 adapter molecules. The sustenance of such basal class I_A PI 3-kinase activity would subserve retinal tissue and cellular survival in normal physiological conditions in the absence of acute challenges with extracellular stimuli. 

With regard to PDGF-induced prosurvival effects, Stitt et al. ⁹ have demonstrated that retinal pericyte apoptosis induced by advanced glycation end products is reversed by PDGF-induced Akt phosphorylation. However, in these studies, the likely intermediary role of class I_A PI 3-kinase has not been considered. In the present study, acute challenge with PDGF or insulin/IGF-1 produces distinct increases in class I_A PI 3-kinase activity and the resultant Akt phosphorylation in a temporal manner, as observed using ex vivo rat retinas and RGC-5 cells. Together, these data support the notion that mature retinal, RGC, and pericyte PI 3-kinase activity and Akt phosphorylation are subject to acute regulation by signal-specific survival factors, including PDGF. 

From a broad perspective, several investigators have demonstrated that retinal tissue/cellular PI 3-kinase activity and Akt phosphorylation/activity are regulated by visible light, ⁶² insulin, ^{24 63} IGF-1, ^{27 64 65} VEGF, ^{28 66} N-methyl-d-aspartate, ³¹ erythropoietin, ⁶⁷ and 17-β-estradiol. ⁶⁸ In addition, the prosurvival or antiapoptotic effects of retinal PI 3-kinase/Akt signaling have been extensively studied in several experimental models that include normal rats and mice, ^{31 61} diabetic rats and mice, ^{50 69} PI 3-kinase p65-regulatory subunit knockout mice, ⁷⁰ retinal generation (rd1) transgenic mice, ^{71 72} axotomized rat RGCs, ^{28 67 73} intravitreous injection and cellular treatment of PI 3-kinase inhibitors, ^{24 26 27} Akt isoform knockout mice, ⁷⁴ ex vivo rat retinas, ⁵⁰ bovine and rat rod outer segments, ^{62 75} and cell culture systems. ^{9 24 64 66} Although these studies have increased the understanding of in vivo retinal PI 3-kinase survival signaling in health and disease, the present approach provides new insights toward dissecting agonist-specific increases in PI 3-kinase subunit-specific activity and PIP3 levels in the retinas or retinal cells. For instance, acute PDGFR tyrosine kinase inhibition in RGC-5 cells abrogates PDGF-induced class I_A PI 3-kinase activity but preserves insulin-induced IRS-1-associated PI 3-kinase activity (Biswas SK, Zhao Y, Sandirasegarane L, unpublished observations, 2008). Such intriguing findings would allow development of new therapeutic strategies to restore apparent decreases in agonist-specific PIP3 levels that would promote retinal survival. 

Given that PDGF- versus insulin-specific class I_A PI 3-kinase activation has been extensively characterized in extraretinal tissues and cells, the present observations support the existence of two distinct, signal-specific modes of PI 3-kinase activation in the retinas and RGC-5 cells. PDGF-selective PI 3-kinase activation has been demonstrated by its ability to induce compartment-specific plasma membrane-recruited p85 tyrosine phosphorylation, ⁴¹ and p110α-associated lipid kinase activity ⁴² in adipocytes, ^{35 41 43} aortic endothelial cells, ⁴² and wild-type (versus p110α kinase-deficient) mouse model. ⁴⁴ In contrast, insulin-selective extraretinal PI 3-kinase activation in cell culture and animal models ^{35 37 41 42 43 44} has been attributed to (1) microsomal increases in lipid kinase activity and p85 levels, ⁴¹ (2) IRS-1/p110α-associated lipid kinase activity ⁴⁴ without an obligatory p85 tyrosine phosphorylation, ⁴¹ and (3) p110α- ^{35 44} or p110β-associated ^{42 43} or redundant p110α- and p110β-associated ³⁷ lipid kinase activity. From the results in the present study, it is clear that intact rat retinas respond to acute PDGF challenge by exhibiting temporal increases in p85α-associated and IRS-1-independent PI 3-kinase activity, unlike insulin. The specificity of retinal PDGF versus insulin/IGF-1 response is further supported by extensive characterization of RGC class I_A PI 3-kinase activity. In RGC-5 cells, PDGF and insulin/IGF-1 produce distinct, robust increases in phosphotyrosine- and IRS-1/IRS-2-associated class I_A PI 3-kinase activity, respectively. In addition, PDGF and insulin/IGF-1 produce independent but redundant increases in p85- and/or p110α/β-associated class I_A PI 3-kinase activity. Furthermore, the extent of subunit-specific PI 3-kinase activation in response to PDGF or insulin/IGF-1 is dependent on the preexisting basal PI 3-kinase activity. It is noteworthy that the previous studies on agonist-specific class I_A PI 3-kinase activation have been focused on determining its regulatory influence on adipocyte glucose transport, ^{35 43} in vivo glucose intolerance, ⁴⁴ and aortic endothelial cell actin reorganization. ⁴² It is clear from the present observations that class I_A PI 3-kinase is also critically important for retinal cell survival and that it is acutely regulated by survival factors, including PDGF. 

In conclusion, the present study revealed the neurotrophic role of retinal PDGFR activation through its potential stimulatory effects on class I_A PI 3-kinase subunit-specific activity using intact retinas and RGC-5 cells in culture. The definitive antiapoptotic role of retinal/RGC class I_A PI 3-kinase is also evident from the use of broad-spectrum and p110 isoform-selective PI 3-kinase inhibitors. Although dysregulated retinal PDGF ligands and PDGF receptors have been implicated in microvascular and neurodegenerative diseases, including diabetic retinopathy, ^{1 2 3 4 5 6 7 9 10 11 13 15} future studies are clearly warranted that would determine whether disruption of PDGF receptor tyrosine phosphorylation and/or downstream class I_A PI 3-kinase survival signaling contribute to diminished retinal vascular and neural cell viability. 

³  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by Grant 1-2005-1082 from the Juvenile Diabetes Research Foundation (JDRF), the JDRF Diabetic Retinopathy Center at Penn State University, and the Pennsylvania Lions Sight Conservation and Eye Research Foundation. TWG is the Jack and Nancy Turner Professor of Ophthalmology.

 Submitted for publication November 12, 2007; revised March 28, 2008; accepted June 11, 2008.

 Disclosure: S.K. Biswas, None; Y. Zhao, None; A. Nagalingam, None; T.W. Gardner, None; L. Sandirasegarane, 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.

 

Rat retinal and RGC-5 PDGF receptor subtype expression, tyrosine phosphorylation, and PI 3-kinase p85α regulatory subunit association. Cryosections of normal rat eyes fixed in 2% paraformaldehyde were subjected to immunohistochemical analyses as follows: (A) PDGFR-α immunoreactivity and colabeling with vimentin (Müller cells), (B) PDGFR-α immunoreactivity and colabeling with Thy-1 (RGC layer and the inner plexiform layer [IPL]), (C) PDGFR-β immunoreactivity and colabeling with agrin (blood vessels), and (D) PDGFR-β immunoreactivity and colabeling with Thy-1. No primary antibody was used as a negative control (Ab control). (E, F) Control and PDGF (1.2 nM)-stimulated (6 minutes) retinal explants were subjected to immunoprecipitation with anti-PDGFR-α or anti-PDGFR-β primary antibodies followed by immunoblot analysis with the appropriate primary antibody, to determine PDGFR subtype expression, tyrosine phosphorylation, and association with p85α. (G) Immunoblots of control RGC-5 lysates using anti-PDGFR-α or -β primary antibodies. To verify equal loading of samples, β-actin was used as an internal control. (H) Control and PDGF (1.2 nM)-stimulated (6 minutes) RGC-5 cells were subjected to immunoprecipitation/immunoblot studies as illustrated, to determine PDGFR-α and -β tyrosine phosphorylation. The data shown are representative of three separate experiments. (Figures 5A 5Bshow the association of tyrosine phosphorylated PDGFR subtype with p85α in RGC-5 cells).

Figure 2.

 

Comparison of rat retinal and RGC-5 basal class I_A PI 3-kinase activity associated with PDGFR-α, PDGFR-β, INSR, IGF-1R, tyrosine phosphorylated proteins, and regulatory/catalytic subunits. Rat retinal homogenates and serum-deprived (24 hours) RGC-5 cell lysates were subjected to immunoprecipitation with primary antibodies specific for PDGFR-α, PDGFR-β, INSR, IGF-1R, phosphotyrosine (pTyr), IRS-1, IRS-2, p85α, p85pan, p110α, and p110β. The respective immunocomplexes were then used for PI 3-kinase assays. The TLC spots illustrating PI3P formation are PDGFR-α (A), PDGFR-β (B), INSR (C), IGF-1R (D), pTyr (E), IRS-1/2 (F), IRS-1 (G), IRS-2 (H), p85α (I), p85(pan) (J), p110α (K), and p110β (L). The data are representative of three to four experiments in rat retinas and RGC-5 cells. Basal PI 3-kinase assays for all experiments were performed at the same time for rat retinas and RGC-5 cells.

Figure 3.

 

Concentration-dependent effects of PDGF and/or insulin on RGC-5 Akt phosphorylation. Serum-deprived (24 hours) RGC-5 cells were stimulated with increasing concentrations of PDGF (1 ng/mL [0.04 nM], 10 ng/mL [0.4 nM], and 30 ng/mL [1.2 nM]), insulin (1, 10, and 30 nM), or PDGF plus insulin for a fixed time (6 minutes). The cell lysates were subjected to immunoblot analysis for phospho-Akt(Ser473), phospho-Akt(Thr308), and total Akt. To normalize Akt phosphorylation and to verify equal loading of samples, β-actin was used as an internal control. (A) Representative immunoblots and the linear graphs (B, C) that compare PDGF- and/or insulin-induced increases in Akt(Ser473) and Akt(Thr308) phosphorylation in RGC-5 cells. *P < 0.05 compared with the respective controls, n = 3.

Figure 4.

 

Time-dependency experiments that compare PDGF- versus insulin/IGF-1-induced class I_A PI 3-kinase/Akt signaling in rat retinas and RGC-5 cells. (A) Ex vivo rat retinas were stimulated with PDGF (1.2 nM) or insulin (30 nM) for 6, 20, and 60 minutes. Aliquots of the respective tissue homogenates were used in assays, to compare the changes in p85α- or IRS-1/2-associated PI 3-kinase activity and Akt phosphorylation. (B) Serum-deprived (24 hours) RGC-5 cells were stimulated with PDGF (1.2 nM), IGF-1 (30 ng/mL [4 nM]), or insulin (30 nM) for 6, 20, and 60 minutes. The cell lysates were subjected to phosphotyrosine (pTyr)- or IRS-1-associated PI 3-kinase assays and immunoblot analysis for phospho-PDGFRβ, phospho-IGF-1R/insulinR, phospho-Akt, and Akt. To normalize Akt phosphorylation and to verify equal loading of samples, β-actin was used as an internal control. The TLC spots and the immunoblots shown in (A) and (B) are representative of those in three experiments. (C) Quantitative comparison of PDGF- and insulin-induced Akt(Ser473) phosphorylation as a function of time in rat retinas and RGC-5 cells. (*P < 0.05 compared with the respective controls, n = 3). Table 1shows PDGF- and insulin-induced PI 3-kinase activity in rat retinas at the 6-minute time point. *P < 0.05 compared with the basal PI 3-kinase activity, n = 4.

Table 1.

 

Comparison of Increases in PDGF- and Insulin-Induced Class I_A PI 3-Kinase Activity Associated with p85 and IRS-1/2 Signaling Components in Normal Rat Retinas

Primary Antibody	Basal	PDGF	Insulin
p85α	1	1.8 ± 0.2 (4)*	1.7 ± 0.1 (4)*
IRS-1/2	1	1.1 ± 0.1 (4)	2.4 ± 0.5 (4)*

 Ex vivo rat retinal explants were incubated in the absence (control) or presence of PDGF (1.2 nM) or insulin (30 nM) for 6 min. Control and treated retinas were then subjected to immunoprecipitation using the respective primary antibodies followed by PI 3-kinase assays as described in the Methods section. The increases (x-fold) in signaling component-specific PI 3-kinase activity were calculated after normalizing the basal PI 3-kinase activity to 1. Data are expressed as the mean ± SEM (n). The n value represents the number of experiments.

*   P < 0.05 compared with the basal PI 3-kinase activity.

Figure 5.

 

PDGF, IGF-1, or insulin activation of class I_A PI 3-kinase associated with specific receptors, tyrosine phosphorylated proteins (pTyr), IRS-1, IRS-2, p85α, p85pan, p110α, or p110β immunocomplexes in RGC-5 cells. (A–J) Serum-deprived (24 hours) RGC-5 cells were incubated in the absence (control) or presence of PDGF (1.2 nM), IGF-1 (4 nM), or insulin (30 nM) for a fixed time (6 minutes). Subsequently, RGC-5 lysates were subjected to immunoprecipitation with anti-PDGFR-α, anti-PDGFR-β, or the indicated primary antibodies followed by PI 3-kinase. The respective TLC spots for agonist-induced PI3P are shown in (A–C) and (E–J). The PI 3-kinase data from several independent experiments are shown in Table 2(n ≥ 4). Aliquots of the respective immunocomplexes were also subjected to SDS-PAGE followed by immunoblot analysis with anti-p85α and the indicated primary antibodies (A–J). These data are representative of two to three separate experiments.

Table 2.

 

Comparison of Increases in PDGF-, IGF-1-, and Insulin-Induced Class I_A PI 3-Kinase Activity Associated with Signal-Specific Components in RGC-5 Cells

Primary Antibody	Basal	PDGF	IGF-1	Insulin
pTyr	1	38.1 ± 3.9 (10)*	3.8 ± 1.6 (4)*	3.3 ± 0.7 (4)*
IRS-1	1	1.4 ± 0.1 (6)	4.3 ± 1.0 (6)*	7.2 ± 0.2 (8)*
IRS-2	1	0.8 ± 0.9 (8)	10.3 ± 1.6 (8)*	14.2 ± 4.3 (8)*
p85α	1	5.1 ± 0.9 (12)*	4.0 ± 0.6 (12)*	5.5 ± 0.7 (12)*
p85(pan)	1	2.4 ± 0.6 (6)*	2.5 ± 0.3 (6)*	1.9 ± 0.1 (6)*
p110α	1	1.6 ± 0.2 (4)*	1.9 ± 0.1 (4)*	1.8 ± 0.1 (6)*
p110β	1	2.0 ± 0.4 (5)*	1.4 ± 0.2 (5)	3.0 ± 0.6 (5)*

 Serum-deprived (24 hours) RGC-5 cells were incubated in the absence (control) or presence of PDGF (1.2 nM), IGF-1 (4 nM), or insulin (30 nM) for 6 min. Control and treated RGC-5 lysates were then subjected to immunoprecipitation using the respective primary antibodies followed by PI 3-kinase assays. The increases (x-fold) in signaling component-specific PI 3-kinase activity were calculated after normalizing the basal PI 3-kinase activity to 1. Data are expressed as the mean ± SEM (n). The n value represents the number of experiments.

*   P < 0.05 compared with the basal PI 3-kinase activity.

Figure 6.

 

Effects of PI-103 or LY294002 on retinal Akt phosphorylation and RGC-5 Akt phosphorylation and caspase-3/PARP cleavage. (A) Ex vivo retinal explants were pretreated with vehicle, p110 isoform-selective (PI-103), or broad-spectrum (LY294002) PI 3-kinase inhibitors for 3 hours. Subsequently, the retinal explants were exposed to PDGF (1.2 nM) or insulin (30 nM) for 6 minutes. The retinal tissue lysates were then subjected to immunoblot analysis for phospho-Akt(Ser473) or phospho-Akt(Thr308). (B, C) RGC-5 cells maintained under serum-containing (+FBS) or serum free (−FBS) conditions were treated with the indicated concentrations of PI-103 or LY294002 for 24 hours. The cell lysates were then subjected to immunoblot analysis of cleaved caspase-3, cleaved PARP, and phospho-Akt(Ser473). (D) As positive controls, RGC-5 cells were also treated with the Akt inhibitor (Akt-X) for 24 hours. (E, F) RGC-5 cells maintained under serum-free (−FBS) conditions were pretreated with 10 μM PI-103 for 30 minutes, and incubated for an additional 24 hours in the presence or absence of PDGF (1.2 nM) or insulin (30 nM). The cell lysates were then subjected to immunoblot analysis of cleaved caspase-3 and cleaved PARP (*P < 0.05 compared with control). To verify equal loading of samples, β-actin was used as an internal control. The data shown are representative of those in three separate experiments.

Figure 7.

 

A schematic representation of agonist-specific distinct modes of class I_A PI 3-kinase activation that mediate retinal cell survival. Class I_A PI 3-kinase is a heterodimer that consists of regulatory (p85α, p85β, p55γ, p55α, and p50α) and catalytic (p110α, p110β, and p110δ) subunits. PDGF challenge promotes retinal PDGF receptor tyrosine phosphorylation, recruits the p85 regulatory subunit, and induces p110 catalytic subunit kinase activity. Activated class I_A PI 3-kinase catalyzes the phosphorylation of phosphatidylinositol 4,5-bisphosphate (PIP2) lipid substrate, to form phosphatidylinositol 3,4,5-trisphosphate (PIP3) lipid products. In contrast, insulin and IGF-1 receptor activation leads to IRS adapter molecule tyrosine phosphorylation that provides docking sites for the p85 subunit for the subsequent activation of class I_A PI 3-kinase. PDGF- and insulin/IGF-1-induced PIP3 formation facilitates Akt Ser473/Thr308 phosphorylation, which in turn promotes retinal cell survival. The prosurvival role of retinal class I_A PI 3-kinase/Akt signaling is revealed by inhibitors of PI 3-kinase (LY294002 and PI-103) and Akt (Akt-X). All three inhibitory agents induce the expression of apoptotic markers, which include cleaved caspase-3 and PARP.

The authors thank Neeraj Agarwal (Department of Cell Biology and Genetics, University of North Texas Health Science Center, Fort Worth, TX) for the gift of RGC-5 cells and Rhona Ellis and Wade Edris (Microscopy and the Histology Core) and Wendy Holtry (JDRF Animal Core; both at Penn State College of Medicine) for providing technical assistance.