VEGF-Induced Retinal Angiogenic Signaling Is Critically Dependent on Ca²⁺ Signaling by Ca²⁺/Calmodulin-Dependent Protein Kinase II

Abstract

Purpose.: The authors conducted an in vitro investigation of the role of Ca²⁺-dependent signaling in vascular endothelial growth factor (VEGF)-induced angiogenesis in the retina.

Methods.: Bovine retinal endothelial cells (BRECs) were stimulated with VEGF in the presence or absence of 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester (BAPTA-AM; intracellular Ca²⁺ chelator), U73122 (phospholipase C (PLC) inhibitor), xestospongin C (Xe-C), and 2-aminoethoxydiphenyl borate (2APB) (inhibitors of inositol-1,4,5 triphosphate (IP₃) signaling). Intracellular Ca²⁺ concentration ([Ca²⁺]_i) was estimated using fura-2 Ca²⁺ microfluorometry, Akt phosphorylation quantified by Western blot analysis, and angiogenic responses assessed using cell migration, proliferation, tubulogenesis, and sprout formation assays. The effects of the Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) inhibitor KN93 were also evaluated on VEGF-induced Akt signaling and angiogenic activity.

Results.: Stimulation of BRECs with 25 ng/mL VEGF induced a biphasic increase in [Ca²⁺]_i, with an initial transient peak followed by a sustained plateau phase. VEGF-induced [Ca²⁺]_i increases were almost completely abolished by pretreating the cells with BAPTA-AM, U73122, Xe-C, or 2APB. These agents also inhibited VEGF-induced phosphorylation of Akt, cell migration, proliferation, tubulogenesis, and sprouting angiogenesis. KN93 was similarly effective at blocking the VEGF-induced activation of Akt and angiogenic responses.

Conclusions.: VEGF increases [Ca²⁺]_i in BRECs through activation of the PLC-IP₃ signal transduction pathway. VEGF-induced phosphorylation of the proangiogenic protein Akt is critically dependent on this increase in [Ca²⁺]_i and the subsequent activation of CaMKII. Pharmacologic inhibition of Ca²⁺-mediated signaling in retinal endothelial cells blocks VEGF-induced angiogenic responses. These results suggest that the PLC/IP₃/Ca²⁺/CaMKII signaling pathway may be a rational target for the treatment of angiogenesis-related disorders of the eye.

Angiogenesis, the formation of new capillaries from preexisting blood vessels, is a fundamental process involved in normal tissue growth and development, reproduction, and wound healing. ¹ However, the dysregulated growth of new blood vessels plays a key role in a variety of sight-threatening disorders such as retinopathy of prematurity, age-related macular degeneration, and diabetic retinopathy. ² Vascular endothelial growth factor (VEGF) is a hypoxia-inducible growth factor ³ that is considered to be the main mediator of ischemia-induced retinal angiogenesis. ² VEGF exerts its biological effects on endothelial cells through two major receptor tyrosine kinases, VEGFR-1 and VEGFR-2. ⁴ It is now well recognized that VEGFR2 is principally responsible for the angiogenic effects of VEGF. ⁵ The binding of VEGF to VEGFR2 leads to a stimulation of several signaling pathways implicated in angiogenesis, including the PI3K/Akt pathway, the Ras/MAPK pathway, the FAK/paxillin pathway and the RhoA/ROCK pathway. ⁶  

Most endothelial cell functions depend to various degrees on changes in intracellular calcium concentration ([Ca²⁺]_i). In addition to the pathways indicated, VEGF-induced activation of VEGFR-2 stimulates phospholipase C (PLC)-dependent inositol 1,4,5-trisphosphate (IP₃) production, thereby increasing endothelial cell [Ca²⁺]_i. ^{7 –9} The VEGF-induced increase in [Ca²⁺]_i is a consequence of the release of Ca²⁺ from the endoplasmic reticulum by IP₃ receptors, followed by extracellular Ca²⁺ influx. ^10,11 Recent studies have identified members of the canonical transient receptor potential (TRPC) subfamily of cation channels as the most likely candidates for VEGF-induced Ca²⁺ influx in endothelial cells. ¹⁰ Until now, our understanding of the involvement of the PLC-IP₃ Ca²⁺ signaling pathway in the regulation of VEGF-dependent angiogenesis has been limited. However, the inhibition of VEGF-induced proliferation in human umbilical vein endothelial cells (HUVECs) by the drug carboxyamidotriazole has been attributed to its ability to inhibit IP₃ formation and to block Ca²⁺ channels. ¹¹ Dominant-negative suppression of TRPC6 channel function has also been shown to attenuate the VEGF-mediated [Ca²⁺]_i signal, migration, proliferation, and sprouting angiogenesis in human microvascular endothelial cells (HMVECs). ¹²  

The downstream signaling pathways through which Ca²⁺ mediates the angiogenic effects of VEGF have also yet to be fully elucidated. The ubiquitous Ca²⁺ binding protein calmodulin (CaM) is involved in many aspects of Ca²⁺ regulation in various cell types, including endothelial cells. ^13,14 When activated by Ca²⁺, it has no intrinsic enzymatic activity itself but acts by binding to and activating other proteins, such as Ca²⁺/CaM-dependent protein kinases (CAM kinases), protein phosphatases, and phosphodiesterases. ¹⁵ Accumulating evidence from pharmacologic studies suggests that Ca²⁺/CaM plays a significant role in angiogenic signaling in vitro and in vivo, ^16,17 but the specific target proteins involved in mediating these effects have not been identified. One protein kinase of particular interest is Ca²⁺/CaM-dependent protein kinase II (CaMKII), a multifunctional enzyme that catalyzes the phosphorylation of many proteins within the cell. ¹⁸ CaMKII is a widely recognized effector of Ca²⁺/CaM signaling and has been reported to modulate a number of cellular processes in endothelial cells, including ion homeostasis, nitric oxide production, and vascular permeability. ¹⁹  

The present study was designed to evaluate the importance of the PLC-IP₃ Ca²⁺ signaling pathway in VEGF-induced retinal angiogenic signaling using a range of cell-based assays and an in vitro model of sprouting angiogenesis. The role of CaMKII has also been explored using KN93, a selective inhibitor of this kinase. 

Materials and Methods

Cell Culture

Bovine retinal endothelial cells (BRECs) were isolated and cultured as previously described. ²⁰ Briefly, bovine retinal homogenate was subjected to enzyme digestion consisting of pronase, DNase, and collagenase. This mixture was agitated at 37°C for 30 minutes and filtered through a 53-μm nylon gauze followed by centrifugation at 400g for 5 minutes. Finally, the pellet was suspended in Iscove's modified Dulbecco's medium (IMDM) containing growth supplements. The isolated vessel fragments were then transferred onto 1.5% gelatin-coated dishes (40 mm) and incubated at 37°C. During culturing, BRECs migrated from the vessel fragments and proliferated to form a monolayer over the course of 7 days. BRECs were used from passages 1 and 2 and were serum starved for 5 hours in medium containing 0.5% porcine serum (PS) before experimentation. In all experiments, BRECs were preincubated for 30 minutes with test compounds before VEGF was added. All experiments were repeated a minimum of three times using different batches of cells. 

[Ca²⁺]_i Measurements

BRECs, plated on gelatin-coated coverslips, were incubated with 5 μM fura-2 AM for 20 minutes at room temperature to load them with the Ca²⁺ indicator. They were then washed and superfused with Hanks solution at 37°C in a perfusion bath mounted on the stage of an inverted microscope (Eclipse TE2000; Nikon, Tokyo, Japan) and were alternately illuminated by 340/380 nm light from a dual monochromator (5-nm bandwidth) and light chopper (Cairn Research Ltd., Faversham, UK). Emitted fluorescence was measured from the side port of the microscope through an adjustable rectangular window, a filter (510 nm), and a photon-counting photomultiplier tube in the light path. Fluorescence equipment was controlled by acquisition software (Acquisition Engine, v1.1.5; Cairn Research Ltd., Kent, UK), which was also used for analysis of the fluorescence data. At the end of each experiment, background fluorescence was quantified by incubating the cells with 10 mM MnCl₂ in Ca²⁺-free solution. Changes in the ratio of the background-corrected fluorescence emitted at each excitation wavelength (R = F340/F380) were used as a measure of changes in the [Ca²⁺]_i. ²¹  

Western Blot Analysis

Protein extracts prepared from BRECs were subjected to SDS-PAGE on 10% gels and were transferred to a polyvinylidene difluoride membrane (Westran; Roche Diagnostics Limited, West Sussex, UK). Membranes were probed with rabbit monoclonal antibodies to either phospho-Akt (Ser-473; 1:2000) or total Akt (1:10,000), and mouse polyclonal antibodies to β-actin (1:50,000; Cell Signaling Technology, Danvers, MA). After washing, membranes were incubated with appropriate goat secondary antibodies against rabbit or mouse IgG (IRDye 680 and IRDye 800; Li-COR, Cambridge, UK). Membranes were then scanned and analyzed using ImageJ software (developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). Densitometry data were normalized to the β-actin loading controls. 

Migration Assay

BRECs were seeded onto 40-mm Petri dishes, grown to 80% confluence, and serum starved for 5 hours, and a scratch wound was made across the center of each plate using a sterile 200-μL pipette tip. The cells were then gently washed, the medium was changed, and wound closure for various treatments was monitored by taking digital photographs immediately after injury and 16 hours after wounding using a camera (E5400; Nikon) attached to a phase-contrast microscope (Eclipse TS100; Nikon). Assays were performed in the presence of 5-fluorouracil (1 mM) to prevent proliferation of the endothelial cells. Cell migration was determined by measuring the surface area of migrated cells into the wound area using imaging software (NIS-Elements BR 3.0; Nikon). 

Proliferation Assay

Endothelial cell proliferation activity was assessed using a commercially available BrdU-ELISA kit (Roche Molecular Biochemicals, Mannheim, Germany). BRECs were seeded at a density of 1 × 10⁴ cells/well into 96-well plates in IMDM containing 10% PS. After 12 hours, the cells were serum starved for 5 hours and then stimulated with VEGF in the presence or absence of test compounds, and BrdU labeling reagent (final concentration, 10 μM) was added. After 16 hours of incubation, the cells were fixed and DNA denatured by the application of solution (FixDenat, 100 μL/well; Roche) that was then thoroughly washed out, and peroxidase-labeled anti-BrdU was added and incubated for 30 minutes at room temperature. Tetramethylbenzidine substrate solution (100 μL) was added and incubated for approximately 30 minutes until the color development was sufficient for photometric detection. The reaction was stopped with the addition of 25 μL/well of H₂SO₄ (1 M). The absorbance of the samples was measured in a microplate reader (TECAN, Reading, UK) at 450 nm. 

Tubulogenesis Assay

For the assessment of tube-like structure formation, 40-mm Petri dishes were coated with 0.5 mL basement membrane matrix (Matrigel; BD Biosciences, Oxford, UK) diluted 1:1 with IMDM and incubated at 37°C for 30 minutes. Once the basement membrane matrix had hardened, BRECs were plated at a density of 1 × 10⁶ cells in IMDM containing 10% PS for 12 hours and subsequently serum starved for an additional 5 hours. Cells were then preincubated with test compounds for 30 minutes followed by the addition of VEGF and were further incubated for 24 hours. Images were captured with a digital camera attached to a phase-contrast microscope (Eclipse TS100; Nikon). Software was used to measure tube lengths (NIS-Elements; Nikon). 

Angiogenic Sprouting Assay

An in vitro angiogenic sprouting assay was performed as previously described. ²² Briefly, 1 × 10⁶ BRECs were resuspended in 25 μL IMDM and mixed in a 1:1 ratio with growth factor–reduced basement membrane matrix, after which 50-μL aliquots were spotted onto 40-mm Petri dishes. After basement membrane matrix polymerization, blobs were covered in IMDM with 10% PS and further incubated for 12 hours. After serum starvation for another 5 hours, the medium was aspirated off and a second layer of basement membrane matrix diluted 1:1 with IMDM was layered evenly over the primary spots, creating duplex cultures. In each treatment, drugs were added to both the secondary basement membrane matrix layer and the surrounding culture media. The number of angiogenic sprouts invading the second layer of the duplex culture was counted after 24 hours and photographed using phase-contrast microscopy. A phase-dark line demarcated the boundary of the primary and secondary gel layers. 

Trypan Blue Exclusion Assay

BRECs were seeded onto 12-well plates at a density of 1 × 10³ cells for 12 hours. After serum starvation for 5 hours, they were treated with test compounds for a further 24 hours. They were then collected by trypsinization, washed in PBS, and stained with 0.2% trypan blue, and the proportion of viable cells was determined from hemocytometer counts. 

Drugs and Solutions

Reagents used in this study were obtained from the following sources: 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester (BAPTA-AM) and U73433 (Enzo Life Sciences International, Inc., Plymouth Meeting, PA); U73122 and cyclopiazonic acid (CPA; Tocris, Ellisville, MO); xestospongin-C (Xe-C; Cayman Chemical, Ann Arbor, MI); 2-aminoethoxydiphenyl borate (2APB; Calbiochem, San Diego, CA); and fura-2 AM (Tocris). All other chemicals and solutions were obtained from Sigma-Aldrich (Poole, UK). 

Recombinant human VEGF₁₆₅ was used at a concentration of 25 ng/mL. Hanks solution contained 140 mM NaCl, 6 mM KCl, 1.3 mM MgCl₂, 2 mM CaCl₂, 5 mM d-glucose, 10 mM HEPES (pH 7.4 with NaOH). Ca²⁺-free solution had the same composition as the Hanks solution except that CaCl₂ was replaced by 1 mM ethylene glycol bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA). Stock solutions of test compounds were prepared in dimethylsulfoxide (DMSO). The DMSO concentration in the assays never exceeded 0.1% (vol/vol), a concentration that was found not to influence the results of any of the assays. 

Statistical Analysis

All data are expressed as mean ± SEM. Statistical analyses were performed using graphing and statistics software (Prism V4.02; GraphPad Software, San Diego, CA). Distribution normality was assessed using the D'Agostino and Pearson omnibus normality test. Differences among treatment groups were evaluated for statistical significance using one-way ANOVA followed by the Bonferroni post hoc test. P < 0.05 was considered significant. 

Results

VEGF Elevates [Ca²⁺]_i in Retinal Endothelial Cells through the PLC-IP₃ Signaling Pathway

Stimulation of BRECs with VEGF caused a biphasic increase in [Ca²⁺]_i with an initial transient peak followed by a sustained plateau phase that remained above the original baseline (Fig. 1A). When BRECs were exposed to Ca²⁺-free solution (1 mM EGTA), the initial peak was little affected, but the plateau phase was completely abolished (Fig. 1B). When BRECs were pretreated with a membrane-permeant form of the Ca²⁺ chelator, BAPTA (1 μM BAPTA-AM), the application of VEGF failed to elicit any increase in free [Ca²⁺]_i (Fig. 1C). Pretreatment of BRECs for 30 minutes with the PLC inhibitor U73122 (10 μM), but not its inactive analog, U73433 (10 μM), also abolished the VEGF-induced [Ca²⁺]_i response (Figs. 1D, 1E). To determine whether IP₃ receptor activation plays a role in the VEGF-induced increase in [Ca²⁺]_i, the effects of the IP₃ receptor antagonists, Xe-C (10 μM) and 2APB (25 μM), were evaluated. In the presence of either inhibitor, only a small plateau increase in [Ca²⁺]_i was observed (Figs. 1F, 1G). Summary data for these experiments showing the effects of each inhibitor on the VEGF-induced peak and plateau changes in [Ca²⁺]_i are presented in Figure 1H. The interpretation of data using 2APB as an IP₃ receptor antagonist can be confounded by its off-target effects, particularly on store-operated Ca²⁺ entry (SOCE). ²³ Although it has previously been reported that 2APB concentrations of approximately 2 μM are most appropriate for the selective inhibition of IP₃ receptors, ²⁴ we found that concentrations lower than 25 μM failed to fully block the peak of the VEGF-induced [Ca²⁺]_i response (data not shown). Experiments were, therefore, performed to determine the effects of 25 μM 2APB on SOCE in BRECs. The effects of 2APB on SOCE was assessed by measuring the initial rate of increase in [Ca²⁺]_i when 2-mM Ca²⁺ Hanks solution was reapplied extracellularly after store depletion in Ca²⁺-free solution containing 5 μM CPA. ²⁵ Treatment of BRECs for 30 minutes with 2APB had no significant effect on the rate of SOCE (Figs. 1I, 1J). 

Figure 1.

 

Effects of VEGF (25 ng/mL) on [Ca²⁺]_i.. Typical traces showing changes in [Ca²⁺]_i (expressed as the ratio of fura-2 fluorescence excited at 340 nm and 380 nm; R340/380) in BRECs after the addition of (A) VEGF alone (B) VEGF in Ca²⁺-free EGTA solution, and (C–G) VEGF after preincubation with the intracellular Ca²⁺ chelator BAPTA-AM, the PLC inhibitor U73122 and its inactive analog U73433, and the IP₃ receptor blockers Xe-C and 2APB. (H) Summary data of the peak and plateau fura-2 ratio values for the various experimental groups (n = 6–9 cells). Post hoc analysis. ***P < 0.001 indicates the level of statistical significance for peak responses versus VEGF in normal Hanks solution. ###P < 0.001 denotes the equivalent statistical comparisons for the plateau phase of the VEGF response. (I) Representative traces showing the effect of 25 μM 2APB on SOCE induced by 5 μM CPA in a BREC. (J) Bar graph demonstrating that the rate of SOCE was unaffected in BRECs after exposure to 2APB (n = 6 cells).

Activation of the PLC-IP₃ Ca²⁺ Signaling Pathway Regulates VEGF-Induced Phosphorylation of Akt

Activation of the serine/threonine kinase Akt is known to orchestrate a number of signaling pathways involved in angiogenesis. ²⁶ To examine the possible involvement of the PLC-IP₃ Ca²⁺ signaling pathway in VEGF-induced angiogenic signaling, experiments were carried out to examine the effects of Ca²⁺ signaling inhibitors on the phosphorylation of Akt. It is well established that the phosphorylation of Akt at position Ser473 is essential for the activation of this kinase. ²⁷ BRECs were stimulated with VEGF for 16 hours in the absence or presence of BAPTA-AM, U73122, Xe-C, or 2APB, and Western blot analysis was performed on lysate supernatants for Ser473 phospho-Akt and total Akt. The application of VEGF to BRECs induced a 2.1-fold increase in Akt phosphorylation compared with control (unstimulated) cells (Fig. 2). Pretreatment of BRECs for 30 minutes with BAPTA-AM (1 μM), U73122 (10 μM), Xe-C (10 μM), or 2APB (25 μM) strongly inhibited the VEGF-induced phosphorylation of Akt (Fig. 2). None of the treatments modified total Akt expression. 

Figure 2.

 

Protein expression and phosphorylation of Akt. (A) Western blot analysis showing that VEGF-induced Akt phosphorylation (p-Akt) was blocked in cells treated with BAPTA-AM, U73122, Xe-C, or 2APB, whereas total Akt levels (t-Akt) were unchanged. β-Actin was used as a loading control (β-actin). (B, C) Summary data calculated from the integrated density of the protein bands and expressed as a ratio to β-actin. ***P < 0.001 vs. VEGF; ###P < 0.001, #P < 0.05 vs. control.

Disruption of the PLC-IP₃ Ca²⁺ Signaling Pathway Inhibits Multiple Aspects of VEGF-Induced Angiogenesis

Retinal angiogenesis involves several processes, including endothelial cell migration, proliferation, and tube formation. ² To clarify the role of the PLC-IP₃ Ca²⁺ signaling pathway in regulating VEGF-mediated angiogenic activity, we investigated the effects of BAPTA-AM, U73122, Xe-C, or 2APB on each of these processes. To address the general relevance of our findings, we also tested the effects of these agents on angiogenesis by assessing sprout formation using a three-dimensional in vitro basement membrane matrix duplex assay. 

Endothelial cell migration is an essential early event in angiogenesis. ²⁸ To examine whether blockade of the PLC-IP₃ Ca²⁺ signaling pathway inhibits the migration of BRECs, we undertook scratch-wound assays. VEGF treatment of BRECs led to almost complete wound closure over the 16-hour period studied (Fig. 3B). Exposure of BRECs to BAPTA-AM, U73122, Xe-C, or 2APB led to a significant inhibition of the VEGF-induced migration (Figs. 3C–F). 

Figure 3.

 

PLC-IP₃ Ca²⁺ signaling is essential for VEGF-induced migration of retinal endothelial cells. (A–F) Representative images of the migration scratch-wound assay on BRECs for the each of the experimental groups. Black dotted lines: wound edge at time 0. Scale bars, 100 μm. (G) Quantification of migrated area 16 hours after scratch. ***P < 0.001 vs. VEGF; ###P < 0.001, ##P < 0.01, #P < 0.05 vs. control.

The PLC-IP₃ Ca²⁺ signaling pathway has previously been shown to regulate VEGF-induced cell proliferation in other types of endothelial cells. ^8,29 However, the role of this pathway in regulating the proliferation of retinal endothelial cells has not been examined. In the present study, cell proliferation was determined using ELISA based on the measurement of BrdU incorporation during DNA synthesis. VEGF significantly increased DNA synthesis in BRECs compared with control cells (Fig. 4). BAPTA-AM, U73122, Xe-C, and 2APB prevented the VEGF-induced proliferation of BRECs (Fig. 4). 

Figure 4.

 

BrdU-ELISA cell proliferation assay. Bars show the mean values of BrdU absorbance for each treatment condition. Preincubation of BRECs with BAPTA-AM, U73122, Xe-C, or 2APB blocked the VEGF-induced increase in DNA synthesis. ***P < 0.001 vs. VEGF; ###P < 0.001, ##P < 0.01, #P < 0.05 vs. control.

To further assess the role of the PLC-IP₃ Ca²⁺ signaling pathway in the angiogenic activity of BRECs, we examined the ability of BAPTA-AM, U73122, Xe-C, and 2APB to inhibit basement membrane matrix-induced tubulogenesis. Twenty-four hours after seeding on basement membrane matrix, control cells showed a distinct tubulogenic response (Fig. 5A). VEGF enhanced tube formation by 1.9-fold compared with control cells (Fig. 5B), but this effect was totally abolished in the presence of Ca²⁺ signaling inhibitors (Figs. 5C–F). 

Figure 5.

 

VEGF promotes in vitro blood vessel formation through the PLC-IP₃ Ca²⁺ signaling pathway. (A–F) Representative images of BRECs cultured on basement membrane matrix in the absence or presence of VEGF or with VEGF plus inhibitors of the PLC-IP₃ Ca²⁺ signaling pathway. Addition of BAPTA-AM, U73122, Xe-C, or 2APB blocked VEGF-induced tubulogenesis. Scale bars, 100 μm. (G) Quantification of tube formation for each of the treatments. ***P < 0.001 vs. VEGF; ###P < 0.001, ##P < 0.01 vs. control.

Because BAPTA-AM, U73122, Xe-C, and 2APB prevented VEGF-induced migration, proliferation, and tube formation, we evaluated the effect of these agents on an in vitro model of sprouting angiogenesis. As expected, VEGF increased the number of angiogenic sprouts compared with control cells (Figs. 6A, 6B). The induction of sprout formation by VEGF was completely inhibited by treatment with BAPTA-AM, U73122, Xe-C, or 2APB (Figs. 6C–F). 

Figure 6.

 

Effects of inhibitors of PLC-IP₃ Ca²⁺ signaling pathway on sprout formation. (A–F) Representative phase-contrast images of the sprouting assay for BRECs cultured in the absence or presence of VEGF or with VEGF plus BAPTA-AM, U73122, Xe-C, or 2APB. Images shown are representative fractions from the whole circumference of the corresponding basement membrane matrix. Scale bars, 100 μm. (G) Quantification of the number of sprouts per basement membrane matrix. ***P < 0.001 vs. VEGF; ###P < 0.001, ##P < 0.01 vs. control.

To ensure that our findings could not be attributed to a loss of cell viability, we undertook trypan blue exclusion assays. Prolonged (24-hour) exposure of BRECs to BAPTA-AM, U73122, Xe-C, or 2APB had no significant effects on cell viability (P > 0.05 for all treatments vs. control). 

CaMKII Inhibitor Attenuates VEGF-Induced Phosphorylation of Akt and Angiogenic Activity in BRECs

The final series of experiments was undertaken to begin to elucidate the downstream signaling pathway through which the VEGF-induced [Ca²⁺]_i signal mediated its proangiogenic effects. Specifically, the effects of the CaMKII inhibitor, KN93 (10 μM), and its inactive analog, KN92 (10 μM), on VEGF-stimulated Akt phosphorylation and angiogenic activity were examined. 

As shown in Figure 7, incubation of BRECs with KN93, but not KN92, completely abrogated the effects of VEGF on the phosphorylation of Akt. Similarly, VEGF-induced migration, proliferation, tube formation, and sprouting angiogenesis were prevented after treatment with KN93, but not KN92 (Fig. 8). KN93 and KN92 had no effect on cell viability as determined using the trypan blue exclusion test (P > 0.05 for KN93 and KN92 vs. control). 

Figure 7.

 

CaMKII inhibitor KN93 blocks the VEGF-induced phosphorylation of Akt. (A) Western blot analysis showing that VEGF-induced phosphorylation of Akt (p-Akt) was blocked in BRECs treated with the CAMKII inhibitor KN93 but not the inactive analog KN-92. Total Akt was unaltered by treatment with either KN93 or KN92. (B, C) Mean data expressed as a ratio to β-actin calculated from the integrated density of the protein bands. ***P < 0.001 vs. VEGF; ###P < 0.001, #P < 0.05 vs. control.

Figure 8.

 

VEGF-induced cell migration, proliferation, tubulogenesis, and sprout formation are dependent on CAMKII signaling. (A) Representative images showing the effects of KN93 and KN92 on VEGF-stimulated cell migration using the scratch-wound assay. (B) Quantification of cell migration. (C) KN93, but not KN92, inhibited the increase in DNA synthesis elicited by VEGF as determined using the BrdU-ELISA. (D) Phase-contrast images showing the effects of CAMKII inhibition on VEGF-induced tubulogenesis. (E) Pooled data for the tubulogenesis assays. (F, G) Representative images and summary data showing that KN93 blocked VEGF-induced sprouting angiogenesis. Scale bars, 100 μm. ***P < 0.001 vs. VEGF; ###P < 0.001, ##P < 0.01, #P < 0.05 vs. control.

Discussion

Previous work has shown that activated PI3K, MAPK, FAK, and Rho pathways are important intracellular signaling steps for VEGF-induced angiogenesis. ^6,30 Using various pharmacologic inhibitors, we examined the role of the PLC-IP₃ Ca²⁺ signaling pathway and one of its major downstream targets, CaMKII, in modulating the VEGF-induced angiogenic behavior of retinal endothelial cells. Our data demonstrate that activation of the PLC/IP₃/Ca²⁺/CaMKII signaling cascade regulates a number of aspects of the angiogenic response and is an important mediator of VEGF-induced retinal angiogenesis in vitro. In most protocols, it was apparent that inhibitors of this second messenger system also reduced angiogenic responses to levels below those observed in control cells, suggesting that basal and VEGF-stimulated levels of angiogenic activity are regulated by the PLC/IP₃/Ca²⁺/CaMKII signaling pathway. 

The present study has provided the first detailed analysis of VEGF-induced Ca²⁺ signaling in retinal endothelial cells. We observed [Ca²⁺]_i responses to VEGF that were comparable to those previously described in other types of endothelial cells, including choroidal endothelial cells (CECs), ²⁹ HUVECs, ¹¹ and HMVECs. ¹² The VEGF-induced [Ca²⁺]_i signal in BRECs was characterized by a rapid but transient increase in [Ca²⁺]_i, followed by a sustained plateau phase during which [Ca²⁺]_i remained elevated above the original baseline. Removal of extracellular Ca²⁺ had no effect on the initial transient increase in [Ca²⁺]_i, suggesting that this component of the response arose as a consequence of Ca²⁺ mobilization from intracellular stores. In contrast, the sustained [Ca²⁺]_i elevation was abolished in the presence of Ca²⁺-free solution, implying that this resulted from the influx of Ca²⁺ from the extracellular medium. To demonstrate that the rise in [Ca²⁺]_i induced by VEGF was mediated by the activation of the PLC-IP₃ pathway, we examined the effects of the PLC inhibitor U73122 and the IP₃ receptor antagonists Xe-C and 2APB on this response. All three of these drugs blocked both phases of the VEGF-induced [Ca²⁺]_i signal, although a small, sustained increase in [Ca²⁺]_i was still observed in the presence of the IP₃ receptor inhibitors. This residual elevation of [Ca²⁺]_i could represent the activation of Ca²⁺ influx channels through a PLC-dependent, but IP₃-independent, pathway. Indeed, stimulation of PLC results in the generation not only of IP₃ but also of the lipid second-messenger molecule diacylglycerol, ³¹ which is known to activate several Ca²⁺-permeable channels within the TRP channel superfamily (e.g., TRPC3/6/7 and TRPA1). ³² Future work is clearly warranted to resolve the molecular identity of the Ca²⁺ influx channels activated by VEGF in retinal endothelial cells. 

It has previously been suggested that Ca²⁺ plays an important role in VEGF-induced angiogenesis. ³³ Arguably, the most direct method of assessing the role of Ca²⁺ in cellular responses is to buffer the intracellular free Ca²⁺ using chelating agents such as BAPTA-AM. It has been shown that BAPTA-AM, which is hydrolyzed intracellularly to release the Ca²⁺ buffer BAPTA, can inhibit the proliferation of CECs in response to VEGF, ²⁹ but its effects on other processes relevant to VEGF-induced angiogenesis have not been tested. In our study, treatment of BRECs with BAPTA-AM prevented the rise in [Ca²⁺]_i induced by VEGF and inhibited cell migration, proliferation, tube formation, and sprouting angiogenesis. Our data, therefore, provide some of the most concrete evidence to date that Ca²⁺ plays a pivotal role in mediating the proangiogenic actions of VEGF. Consistent with these findings, we also observed a block of VEGF-induced angiogenic responses using inhibitors of the PLC-IP₃ signaling pathway. Although the activation of PLC has previously been implicated in VEGF-induced endothelial cell proliferation, ⁴ migration, ³⁴ tubulogenesis, ⁷ and vascular sprout formation, ³⁴ the present study is the first to directly link IP₃ receptor activation to these processes. 

CaMKII is an important enzyme that participates in mediating the downstream effects of Ca²⁺ by CaM. ¹⁸ CaMKII has been most extensively studied in the nervous system, where it plays a key role in many processes, including synaptic plasticity, long-term potentiation, and learning and memory. ³⁵ Although less is known about endothelial-specific CaMKII signaling, data collected so far have demonstrated a potentially significant role of CaMKII in endothelial cell physiology and pathophysiology. ¹⁹ For instance, in pulmonary artery endothelial cells, CaMKII has been connected to thrombin-induced increases in monolayer permeability through activation of the MAPK ³⁶ and Rho ³⁷ signaling pathways. In the present study, we have revealed for the first time the involvement of CaMKII in modulating the angiogenic activity of endothelial cells. Pharmacologic blockade of CaMKII resulted in a marked decrease in VEGF-induced retinal endothelial cell migration, proliferation, tubulogenesis, and sprout formation. Collectively, our data strongly suggest that CaMKII acts as a key signaling molecule in the transduction of upstream Ca²⁺ signals into downstream angiogenic responses after VEGF stimulation. 

Our data have also provided insight into the possible molecular mechanism through which CaMKII regulates angiogenic signaling in retinal endothelial cells. A large number of studies have demonstrated that Akt signaling is essential for multiple steps of the angiogenic process, including cell migration, proliferation, survival, and vessel formation. ³⁸ We found that the inhibition of PLC/IP₃/Ca²⁺/CAMKII signaling completely abrogated the VEGF-induced phosphorylation of Akt in BRECs. It has been reported that VEGF-induced Akt phosphorylation in BRECs is also dependent on PI3K activation because it can be blocked with the PI3K inhibitor LY294002. ³⁹ Thus, our data, together with previous findings, suggest that VEGF-induced Akt phosphorylation in retinal endothelial cells requires activation of both the CAMKII and the PI3K pathways given that the blockade of either inhibits Akt activation. These findings are consistent with those of a recent study in bovine aortic endothelial cells showing that both the blockade of Ca²⁺/CaM and the inhibition of PI3K completely abrogated Akt phosphorylation mediated by VEGF. ⁴⁰ Other investigators have also reported that CAMKII is required for Akt phosphorylation in vascular smooth muscle cells after exposure to oxidative stress (H₂O₂) ⁴¹ and in fibroblast-like synovial cells during TRAIL-mediated apoptosis. ⁴² A future challenge will be to identify how the CAMKII and PI3K pathways converge to mediate Akt phosphorylation and whether CAMKII is capable of directly phosphorylating Akt. Of relevance, it has been shown that CAMKII is able to directly promote Akt activity, but the extent and rate of activation were considerably lower than those observed for 3-phosphoinsoitide-dependent protein kinase 1, one of the major Akt kinases of the PI3K pathway. ⁴³  

In summary, VEGF elevates [Ca²⁺]_i in retinal endothelial cells through activation of the PLC-IP₃ signaling pathway. Interventions that blocked Ca²⁺ signaling in retinal endothelial cells prevented VEGF-induced changes in cell migration, proliferation, tube formation, and sprouting angiogenesis. We have also provided the first evidence of the participation of CAMKII in the angiogenic process. A possible limitation of the present study is the reliance on the use of pharmacologic inhibitors to evaluate the role of the PLC/IP₃/Ca²⁺/CaMKII signaling pathway in VEGF-induced retinal angiogenic signaling. Future studies using molecular techniques would allow a more specific blockade of this pathway and identification of the relevant protein isoforms involved. Given recent advances in the treatment of proliferative retinopathies with anti-VEGF agents, ⁴⁴ our findings may provide a new focus for the development of more effective therapies for the treatment of neovascular eye disease. 

Footnotes

 Supported by Commonwealth Commission, United Kingdom.

Footnotes

 Disclosure: E. Banumathi, None; A. O'Connor, None; S. Gurunathan, None; D.A. Simpson, None; J.G. McGeown, None; T.M. Curtis, None