Decay-accelerating Factor Induction on Vascular Endothelium by Vascular Endothelial Growth Factor (VEGF) Is Mediated via a VEGF Receptor-2 (VEGF-R2)- and Protein Kinase C-α/ϵ (PKCα/ϵ)-dependent Cytoprotective Signaling Pathway and Is Inhibited by Cyclosporin A*

Vascular endothelial growth factor (VEGF)¹ represents a family of multifunctional glycoproteins centrally involved in vasculogenesis, angiogenesis, regulation of vascular permeability, and cytoprotection (1). In addition, VEGF may act as a proinflammatory cytokine regulating cellular adhesion, molecule expression, and T lymphocyte trafficking (2, 3). The importance of VEGF in angiogenesis is exemplified by the embryonic lethality observed in VEGF and VEGF receptor-1 and -2 (VEGF-R1 and -2) gene-targeted mice (4–6). The maintenance of quiescent vascular endothelium is dependent upon a variety of survival signals, many of which are regulated by VEGF (7). This was initially revealed by the demonstration that withdrawal of local VEGF synthesis results in retinal blood vessel regression (8). Subsequent studies have shown that VEGF enhances cytoprotection via induction of focal adhesion kinase (9) and the antiapoptotic genes Bcl-2 and A1 (10). In addition, VEGF increases endothelial nitric oxide synthase (eNOS) expression and local nitric oxide (NO) generation (11–14), contributes to the maintenance of an antithrombotic endothelial surface by increasing prostacyclin release (13, 15), enhances protection against complement-mediated injury (16), and may reduce leukocyte-EC interactions (17). These mechanisms may contribute to the capacity of VEGF to facilitate vascular repair during nephritis and thrombotic microangiopathy (18).

The actions of VEGF on vascular ECs are mediated via tyrosine-kinase signaling receptors VEGF-R1 (flt-1) and VEGF-R2 (flk-2/KDR), which are capable of interacting with a variety of downstream signaling pathways (1, 19). The formation of a VEGF-R2·c-Src complex results in activation of phospholipase C-γ (PLCγ), leading to the generation of diacylglycerol and inositol 1,4,5-triphosphate (13, 20). Protein kinase C (PKC) isozymes, singly or in combination, may subsequently participate and add functional diversity. Although the precise function of individual isozymes remains to be determined, evidence to date suggests that PKCα, -β, -δ, -ϵ, and -ζ contribute to VEGF-induced EC proliferation, prostacyclin secretion, and vascular permeability (20–23). VEGF may also utilize mitogen-activated protein kinase (MAPK) signaling pathways in conjunction with or independent from those associated with PKC (19). The p38, ERK1/2, and c-Jun NH₂-terminal kinase (JNK) MAPK pathways have all been associated with signaling downstream of VEGF-R2 (22–25).

The phosphoinositide 3-kinase (PI3K)/antiapoptotic kinase (Akt) pathway may also be activated following ligation of VEGF-R2 and is strongly associated with the cytoprotective actions of VEGF (20, 26, 27). This effector pathway induces expression of antiapoptotic genes including Bcl-2 and A1 (10, 28). Activation of PI3K/Akt favors cytoprotection over apoptosis through its ability to decrease p38 MAPK activity (29). In addition, PI3K activity has been implicated in VEGF-induced up-regulation of eNOS and generation of NO, which may in turn exert antiapoptotic effects by inhibiting the cysteine protease activities of caspases (30, 31).

Evidence is now emerging that endothelial injury, leading to widespread persistent global EC dysfunction, is common in systemic inflammatory diseases and predisposes a patient to premature atherosclerosis and cardiovascular mortality. Endothelial injury has also been linked to the obliterative vasculopathy and accelerated atherosclerosis seen following organ transplantation (32). Complement activation has an established role in the pathogenesis of inflammatory cardiovascular diseases including atherosclerosis, myocardial infarction, and accelerated arteriosclerosis following cardiac transplantation (33, 34). Innate mechanisms for the control of complement activation on the cell surface include the membrane-bound regulatory proteins DAF (CD55), membrane cofactor protein (MCP, CD46), and CD59 (35). DAF can be induced on the EC surface by treatment with C-reactive protein, proinflammatory cytokines, thrombin, VEGF, and basic fibroblast growth factor via distinct signaling pathways (16, 36–39). DAF is a glycosylphosphatidylinositol-anchored glycoprotein that acts to prevent the formation and accelerate the decay of C3 and C5 convertases (40). The cytoprotective importance of DAF is revealed by the increased susceptibility of DAF-deficient mice to glomerular injury in models of glomerulonephritis (41, 42).

The demonstration that VEGF is fundamental to the maintenance of vascular homeostasis highlights the importance of precisely defining the mechanisms underlying these functions. Herein we describe in detail the VEGF-R2- and PKC-dependent signaling pathway by which VEGF induces DAF expression on the EC surface, so enhancing cytoprotection against complement-mediated injury. We demonstrate dependence upon activation of PKCα and PKCϵ, PI3K-mediated constraint, and inhibition of VEGF-mediated cytoprotection against complement by the immunosuppressive drug cyclosporin A (CsA).

EXPERIMENTAL PROCEDURES

Monoclonal Antibodies and Other Reagents—Monoclonal antibodies 1H4 (anti-DAF) and RMAC8 (antiendoglin) were kind gifts from Dr. D. Lublin (St. Louis, MO) and Dr. A. d'Apice (Victoria, Australia), respectively. DETA NONOate was from Cayman Chemicals (Ann Arbor, MI), and CsA, thymeleatoxin, ingenol dibenzoate, Gö6976, GF109203X, and SU1498 were from Merck Biosciences Ltd. (Nottingham, UK). The PKCβ inhibitor LY379196 was a kind gift from Dr. K. Ways (Eli Lilly, Indianapolis, IN). N^G-Monomethyl-l-arginine (l-NMMA), U-73122, and LY294002 were from BIOMOL (Plymouth Meeting, PA). Myristoylated (myr) PKC peptide inhibitors (myr-ψPKC) (myr-Arg-Phe-Ala-Arg-Lys-Gly-Ala-Leu-Arg-Gln-Lys-Asn-Val) and PKCϵ V1–2 (myr-Glu-Ala-Val-Ser-Leu-Lys-Pro-Thr) were from Promega (Southampton, UK) and BIOMOL, respectively. myr-PKCθ (myr-Leu-His-Gln-Arg-Arg-Gly-Ala-Ile-Lys-Gln-Ala-Lys-Val-His-His-Val-Lys-Cys) and myr-PKCζ (myr-Ser-Ile-Tyr-Arg-Arg-Gly-Ala-Arg-Arg-Trp-Arg-Lys-Leu) were from Merck Biosciences. Recombinant human VEGF 165 was purchased from PeproTech EC Ltd. (London, UK). Placental growth factor (PlGF) was purchased from R&D Systems (Abingdon, UK). Other products were from Sigma.

Cell Isolation and Culture—Human umbilical vein ECs (HUVEC), isolated from umbilical cords as described previously (36), were cultured in 1% gelatin-coated tissue culture flasks (Costar, Cambridge, MA) in medium 199 (M199) (ICN Biomedicals Inc., Costa Mesa, CA) supplemented with 20% fetal bovine serum (Hyclone Laboratories Inc., Logan, UT), 100 IU/ml penicillin, 0.1 mg/ml streptomycin, 2 mm L-glutamine (all from Invitrogen), 10 units/ml heparin (Leo Laboratories, Prince Risborough, UK), and 30 μg/ml EC growth factor (Sigma).

Adenoviral Infection—Generation of adenovirus expression vectors for dominant-negative (DN) PKC isozymes has been described previously (43). DN-PKCα, DN-PKCϵ, and β-galactosidase adenoviruses were amplified in HEK-293A cells and purified using the BD Adeno-X™ purification kit (BD Biosciences). Viral titers were estimated by using the BD Adeno-X™ rapid titer kit. HUVEC were infected by incubation with the relevant adenovirus in serum-free M199 for 2 h at 37 °C. The medium was then changed to M199, 10% fetal bovine serum, and HUVEC were incubated overnight prior to addition of VEGF or carrier control for up to 48 h. Infection of HUVEC with a β-galactosidase control adenovirus demonstrated a transfection efficiency of 95%. Optimal multiplicity of infection (MOI) for the adenoviruses was determined by Western blotting.

Flow Cytometry—Flow cytometry was performed as described previously (36). Pharmacological antagonists were added 60 min prior to the addition of VEGF. In some experiments the results are expressed as the relative fluorescent intensity (RFI), representing mean fluorescent intensity (MFI) with test mAb divided by the MFI using an isotype-matched irrelevant mAb. Cell viability was assessed by examination of EC monolayers using phase contrast microscopy, cell counting, and estimation of trypan blue exclusion. In all experiments EC monolayers were treated with the appropriate vehicle controls.

Complement-mediated Cell Lysis—To estimate complement-mediated cell lysis, ECs in M199, 10% fetal calf serum were pretreated with VEGF, VEGF and CsA, or carrier controls for 48 h, then harvested with trypsin/EDTA, washed, opsonized with antiendoglin mAb RMAC8, and incubated with 5–20% rabbit serum (Serotec, Oxford, UK) in Veronal-buffered saline, 1% gelatin (VBS) for 75 min at 37 °C. Following further washing, ECs were resuspended in VBS, and propidium iodide (PI) (Sigma) was added to a final concentration of 50 μg/ml. ECs were analyzed by flow cytometry using the FL2 channel. Lysis was calculated in triplicate samples as the number of PI-positive cells expressed as a percentage of the total number of cells.

Western Blotting—HUVEC were lysed (4 mm EDTA, 50 mm Tris/HCl, pH 7.4, in 150 mm NaCl with 25 mm sodium deoxycholic acid, 200 μm sodium orthovanadate, 10 mm sodium pyrophosphate, 100 mm sodium fluoride, 1% Triton X-100, 1 mm phenylmethylsulfonyl fluoride, and 5% protease inhibitor mixture), and the protein content was determined using the Bio-Rad D_c protein assay. Cell lysates were subjected to SDS-PAGE on 12.5% gels, and separated proteins were transferred to Immobilon™-P transfer membranes (Millipore Corporation, Bedford, MA). The membranes were probed with rabbit polyclonal antibodies against PKC isozymes (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and against the phosphorylated and non-phosphorylated forms of p38 MAPK (New England Biolabs Ltd., Hitchin, Herts, UK). The blots were developed with an enhanced chemiluminescence substrate (Amersham Biosciences). Integrated density values for the test and control bands were obtained with an Alpha Innotech ChemiImager 5500 (Alpha Innotech, San Leandro, CA).

Statistical Analysis—Differences between the results of experimental treatments were evaluated by analysis of variance with the Bonferroni multiple comparison test using GraphPad 4.0 software (GraphPad Software, San Diego, CA). Differences were considered significant at p values of <0.05.

RESULTS

VEGF-induced Up-regulation of DAF Is Mediated via VEGF-R2 and PLCγ—DAF is expressed at a low level on resting ECs in culture. We have reported previously that VEGF induces a significant, unimodal increase in EC surface expression of DAF as measured by flow cytometry. This response is indirect, requiring the synthesis of one or more intermediate proteins, and is dependent upon an increase in steady-state DAF mRNA (16). DAF protein expression was significantly increased following 24 h of exposure to VEGF with maximal expression at 48–72 h (Fig. 1A). A dual approach was used to explore which VEGF receptor was involved in this response. First, preincubation of EC with SU1498, a selective inhibitor of VEGF-R2 tyrosine kinase, had no effect on basal DAF expression while abrogating VEGF-induced DAF up-regulation (Fig. 1B). Furthermore PlGF, which preferentially binds VEGF-R1, failed to induce DAF expression at concentrations up to 50 ng/ml (Fig. 1B). These data suggest that VEGF-R2 is the dominant receptor in the induction of DAF expression by VEGF.

Ligation of VEGF-R2 by VEGF results in the phosphorylation of PLCγ, a response that has been implicated in VEGF-dependent signaling pathways in EC (13, 44). To investigate the role of PLCγ, HUVEC were treated with the PLCγ antagonist U73122. This resulted in a dose-dependent inhibition of VEGF-induced DAF up-regulation (not shown), which was maximal following pretreatment with 10 μm U73122 (Fig. 1C).

VEGF-induced DAF Up-regulation Is Regulated by PKCα and PKCϵ —To explore the role of PKC in VEGF-induced DAF up-regulation, we studied the expression, activation, and inhibition of specific PKC isozymes. Initial immunoblotting experiments demonstrated that HUVEC expressed PKCα, -βII, -δ, -ϵ, -θ, and -ζ (data not shown). Activation of PKC with phorbol 12,13-dibutyrate (PBu), which activates both classical (cPKCα, -β, and -γ) and novel (nPKCδ, -ϵ, and -θ,) PKC isozymes, resulted in a significant increase in DAF expression (Fig. 2A). Furthermore, as seen in Fig. 2, A and B, both ingenol dibenzoate, which preferentially activates nPKC (45), and thymeleatoxin, which preferentially activates cPKC (46), were capable of inducing a significant dose-dependent increase in cell surface DAF expression.

Previous studies have demonstrated activation and translocation of PKCα, PKCβ, PKCδ, PKCϵ, and PKCζ in ECs exposed to VEGF (20, 21, 23, 47). Inhibition of cPKC and nPKC with GF109203X completely inhibited DAF induction by both PBu and VEGF (Fig. 2, C and D). However, Gö6976, a cPKC-specific inhibitor, only reduced DAF up-regulation by 30–50%, supporting additional involvement of nPKC isozymes. Subsequent experiments sought to establish which one of the cPKC isozymes, PKCα or -β, was involved in the response. As seen in Fig. 3A, both Gö6976 and a specific cell-permeable peptide antagonist of PKCα/β, myr-ψPKC, significantly inhibited VEGF-induced DAF expression. In contrast, LY379196, a PKCβ-specific antagonist, had no effect, thereby implicating PKCα in VEGF-induced DAF up-regulation. There was no significant effect of the antagonists alone on basal DAF expression (not shown).

Pretreatment of ECs with a specific peptide antagonist of PKCϵ (myr-PKCϵ) resulted in a minimal reduction in VEGF-induced DAF expression. However, a combination of myr-ψPKC and myr-PKCϵ completely abrogated the response (Fig. 3A). In contrast, similar myristoylated peptides specific for PKCθ and PKCζ and rottlerin, an antagonist of PKCδ, had no significant inhibitory effect (not shown). The role of PKCα and -ϵ was explored further using adenoviruses expressing PKCα and -ϵ dominant-negative constructs. Infection with each DN adenovirus led to a dose-dependent increase in the specific PKC isozyme immunoreactivity (Fig. 3B), and an MOI of 50 and 200 was used for further experiments with DN-PKCα and DN-PKCϵ, respectively. Infection with the DN-PKC and β-galactosidase adenoviruses had no effect on basal DAF expression (not shown). However, infection with DN-PKCα or DN-PKCϵ alone resulted in 25–30% inhibition of VEGF-induced DAF, whereas use of the dominant-negative adenoviruses in combination completely abrogated the response, thereby confirming the requirement for activation of both PKCα and -ϵ by VEGF for a maximal response (Fig. 3C).

VEGF-induced DAF Up-regulation Is Independent of Nitric Oxide Generation and Thrombin Release—Treatment of ECs with VEGF increases expression and phosphorylation of eNOS and local NO synthesis (13, 47, 48). To investigate the potential role of NO in VEGF-induced DAF expression, HUVEC were preincubated with the NO synthase inhibitor l-NMMA at concentrations up to 1 mm for 1 h prior to the addition of VEGF. In addition, HUVEC were exposed to the NO donor DETA NONO-ate at concentrations up to 25 μm for 48 h. l-NMMA had no effect on VEGF-induced DAF up-regulation (Fig. 4A), and prolonged exposure of ECs to NO, following addition of DETA NONOate, similarly had no effect (mean RFI ± S.E. for unstimulated ECs, 35.8 ± 1.6; and DETA NONOate-treated ECs, 38.9 ± 2.27).

In previous work we have shown that in contrast to VEGF, thrombin increases EC DAF expression through a direct transcriptional response (37). The functions of thrombin and VEGF are closely linked, with thrombin able to up-regulate VEGF receptors on ECs, thereby potentiating the activity of VEGF (49). Moreover, VEGF can accelerate EC tissue factor expression and thrombin generation (50), suggesting that they may act in concert to facilitate angiogenesis. These data raised the possibility that thrombin generation was responsible for VEGF-induced DAF up-regulation on ECs. To investigate this we used hirudin, which binds thrombin and inhibits its proteolytic effects on protease-activated receptor-1 (37). Although hirudin completely inhibited thrombin-induced DAF expression, it had no effect on VEGF-induced up-regulation, suggesting that the latter is thrombin-independent (Fig. 4B).

Inhibition of PI3K Enhances VEGF-induced DAF Expression—Activation of the PI3K/Akt pathway plays an important role in the cytoprotective antiapoptotic actions of VEGF (26, 28). To investigate the role of this pathway in VEGF-induced DAF expression, we used two inhibitors of PI3K, wortmannin and LY294002. Interestingly, inhibition of PI3K synergistically enhanced VEGF-induced DAF expression, suggesting that PI3K activity exerts an inhibitory effect on DAF regulation (Fig. 5A). This response was dose-dependent (not shown) with maximal up-regulation observed with 400 nm wortmannin and 25 μm LY294002 (Fig. 5A). We have shown that activation by VEGF of p38 MAPK, but not ERK1/2, is important in DAF regulation (16). We therefore explored the effect of PI3K inhibition on the phosphorylation of p38 MAPK using immunoblotting and found that inhibition of PI3K enhanced both basal and VEGF-induced phosphorylation of p38 MAPK (Fig. 5B). Densitometry revealed a 2.6-fold increase in p38 MAPK phosphorylation in response to both VEGF and LY294002 alone and a 4-fold increase in ECs exposed to VEGF in the presence of LY294002. Of note, this cross-talk between the PI3K and p38 MAPK pathways has been implicated in the enhanced induction of tissue factor by VEGF, as seen in HUVEC pretreated with wortmannin (51). To investigate the role of PKC in VEGF-induced p38 MAPK phosphorylation, ECs were treated with GF109203X prior to exposure to VEGF. Immunoblotting with quantification by densitometry revealed no inhibition of p38 MAPK phosphorylation (Fig. 5C), suggesting that the PKC and p38 MAPK pathways act in parallel.

Cyclosporin A Inhibits VEGF-induced DAF Expression— Having identified the signaling pathways involved in regulating DAF expression in VEGF-treated ECs, we sought to explore the effects of the immunosuppressive drug CsA on these pathways. VEGF is able to activate the calcium/calmodulin-dependent enzyme calcineurin, which in turn induces DNA binding and transcriptional activities of the nuclear factor of activated T lymphocytes (NFAT) and AP-1 in HUVEC (52). The immunosuppressive drug CsA targets this pathway and may inhibit VEGF-induced eNOS activation (53). Because EC toxicity is considered an important factor in CsA-induced vascular injury, we explored the effect of CsA on VEGF-induced DAF expression. Although CsA had no effect on basal DAF expression, it significantly inhibited the VEGF-mediated up-regulation in a dose-dependent manner (Fig. 6A). Moreover, pretreatment of ECs with CsA inhibited the phosphorylation of p38 MAPK by VEGF, an effect that was reversed by the presence of the PI3K antagonists LY294002 and wortmannin (Fig. 6B).

To address the functional significance of this inhibition of DAF up-regulation by CsA, ECs were treated with VEGF in the presence or absence of CsA and then opsonized with RMAC8, an IgG2a antiendoglin mAb. Following exposure to 10% rabbit serum for 45 min, EC lysis was measured by the uptake of propidium iodide. As seen in Fig. 6C, pretreatment of EC with VEGF for 48 h was cytoprotective, significantly reducing cell lysis. However, in the presence of CsA the beneficial effects of VEGF against complement activation were lost. The role of DAF in VEGF-induced cytoprotection was confirmed by complete reversal in the presence of an inhibitory DAF mAb 1H4.

DISCUSSION

DAF may exert potent cytoprotective and anti-inflammatory effects on vascular endothelium through its ability to reduce the deposition of C3 and C5b-9 and the generation of C5a, a powerful proinflammatory anaphylatoxin. We have shown previously that DAF expression on human vascular ECs may be regulated via distinct agonist-specific pathways, leading us to propose that DAF induction may be important in the maintenance of vascular integrity during inflammation, thrombosis, and angiogenesis (16, 36, 37). In vivo studies have also demonstrated up-regulation of DAF during glomerulonephritis (54), and a cytoprotective role in this setting was revealed by the increased susceptibility of DAF-deficient mice to glomerular injury (41, 42). Our observation that VEGF induced a marked increase in cell surface DAF expression on ECs and the emerging importance of VEGF in vascular cytoprotection (19, 55) led us to explore in more detail the signaling pathways involved in this response.

VEGF-R2 is recognized to be the major mediator of the proangiogenic and permeability-enhancing effects of VEGF (1). In contrast, the role of VEGF-R1 on adult ECs remains less well understood, although recent studies have demonstrated a role in VEGF-induced NO synthesis (14), in release of soluble mediators such as interleukin-6 by liver sinusoidal ECs (56), and in the regulation of vascular sprouting (57). The pleiotropic effects associated with ligation of VEGF receptors are reflected in the complexity of their downstream signaling pathways, with c-Src, PLCγ, PI3K, calcineurin, and PKC all implicated in VEGF-R2 signaling (7). The cytoprotective actions of VEGF have been most closely linked to VEGF-R2 and the PI3K/Akt pathway (10, 28).

In this paper we have identified a novel PKC-dependent cytoprotective pathway mediated via ligation of VEGF-R2 and activation of PLCγ. The role of PKC was confirmed by inhibition with the broad spectrum antagonist GF109203X. PKC is made up of three groups of isozymes, cPKC, nPKC, and atypical PKC. GF109203X inhibits cPKC and nPKC, of which PKCα, -βI, and -βII (cPKC) and PKCδ, -ϵ, and -θ (nPKC) may be expressed in ECs. Pretreatment of ECs with the cPKC antagonists Gö6976 and myr-ψPKC only partially inhibited the response to VEGF, whereas a PKCβ-specific inhibitor had no effect, suggesting a role for PKCα. Moreover, PKCδ and -θ antagonists had no significant effect (not shown), whereas a combination of myr-ψPKC and myr-PKCϵ completely abrogated VEGF-induced up-regulation of DAF. In view of concerns regarding the specificity of pharmacological antagonists, we sought to establish the role of PKCα and -ϵ by using validated DN constructs expressed in adenoviruses. This approach confirmed the role of these isozymes with complete inhibition of VEGF-mediated DAF up-regulation in ECs infected with adenoviruses expressing DN-PKCα and DN-PKCϵ.

Although there is some evidence for interspecies variability in the PKC isozymes activated by VEGF, both PKCα and PKCϵ are among those whose activation status is altered (20–23, 47). Moreover, there is emerging evidence for coordinated actions of these two isozymes in vascular smooth muscle spreading (58), Grb2-associated binder-1 tyrosine phosphorylation (59), and in the responses of ECs to mechanical strain (60). Intriguingly, the latter study demonstrated a temporal difference in the activation of the calcium-dependent cPKCα, for which activation occurred early, and the calcium-independent nPKCϵ, for which activation occurred later and was sustained. It has been argued that the VEGF-induced increase in eNOS expression and activity is PKC-dependent (13, 47, 48, 61), and activation of both PKCα and PKCϵ has been implicated in this response (47). This raised the possibility that NO might be an intermediate in the induction of DAF by VEGF. However, we were not able to demonstrate induction of DAF following treatment of ECs with the NO donor DETA NONOate, nor were we able to demonstrate inhibition of the response to VEGF by including l-NMMA.

The role of PI3K/Akt in VEGF-mediated EC cytoprotection led us to explore the role of this pathway in DAF expression. Treatment of ECs with the PI3K antagonist wortmannin or LY294002 enhanced both basal and VEGF-induced levels of DAF expression on the cell surface. A similar response has been reported in the regulation by VEGF of tissue factor, platelet-activating factor, and the cellular adhesion molecules, including E-selectin, VCAM-1, and ICAM-1 (2, 51, 62). We have reported previously that p38 MAPK, but not ERK1/2, is involved in the up-regulation of DAF by VEGF (16). In this study we found that inhibition of PI3K enhanced phosphorylation of p38 MAPK by VEGF, supporting earlier reports (29, 51) and providing an explanation for the increased levels of DAF seen. However, phosphorylation of p38 MAPK by VEGF was not inhibited by the PKC antagonist GF109203X, implying that VEGF activates parallel PLCγ/PKC and p38 MAPK pathways essential for DAF induction and analogous to those reported previously for the induction of platelet-activating factor by VEGF (62).

These data suggest that DAF expression is tightly controlled, with activation of PLCγ/PKC and p38 MAPK exerting a positive influence and with activation of PI3K/Akt exerting a negative influence. The balance between the pathways is likely to depend on the local microenvironment and on the type and duration of signaling events. Thus, in an inflammatory setting, when the risk of complement-mediated injury is increased, the up-regulation of DAF by VEGF may be enhanced by the local generation of tumor necrosis factor-α (TNFα) and thrombin (36, 37). This is supported by our recent demonstration of the up-regulation of DAF expression in the inflamed glomerulus, using in an in vivo model of glomerulonephritis (54). However, we recognize that cross-talk between the array of signaling pathways downstream of VEGF-R2 and indeed between VEGF-induced and other agonist pathways is highly complex and remains to be fully understood (7).

Persistent vascular injury resulting in accelerated allograft arteriosclerosis remains a common cause of late graft deterioration and death following cardiac transplantation. Moreover, a role for the membrane attack complex of complement in the pathogenesis of this process was revealed by the observation that C6-deficient rats are protected against accelerated arteriosclerosis following cardiac transplantation (34). CsA, although important for the prevention of allograft rejection, does not inhibit accelerated arteriosclerosis. Indeed, treatment with CsA may result in endothelial cell toxicity, which may in turn exacerbate vascular injury (63). This led us to consider whether CsA might interfere with the signaling pathways regulating DAF expression and whether this may contribute to CsA-mediated vasculopathy. CsA binds cyclophilin, and this complex targets the calcium/calmodulin-dependent protein phosphatase calcineurin (64), which regulates the phosphorylation state and binding activity of a variety of transcription factors including NFAT, nuclear factor-κB (NF-κB), and c-Jun NH₂-terminal kinase (52, 65, 66). The endothelial toxicity of CsA has been linked to inhibition of cyclophilin- and calcineurin-dependent pathways (53, 67). VEGF has been shown to play an important role in EC protection against CsA toxicity (55). Notwithstanding this, CsA may under certain circumstances inhibit actions of VEGF such as the induction of angiogenesis (68) and eNOS activation, which may in turn contribute to CsA-induced hypertension (53). Although treatment of ECs with CsA had no effect on basal DAF expression, it significantly inhibited VEGF-induced up-regulation. The concentrations of CsA used are consistent with those used in previous in vitro studies (53, 66, 67, 69). Although these may exceed plasma levels following therapeutic dosing (67), it has yet to be determined how plasma CsA concentrations relate to those achieved in vascular endothelium. A study of CsA tissue levels in the epidermis suggests that these may significantly exceed plasma concentrations (70).

Inhibition of VEGF-mediated DAF up-regulation by CsA rendered ECs sensitive to complement-induced cell lysis, a finding of potential relevance to CsA-mediated endothelial toxicity and the development of posttransplant vasculopathy. The mechanism by which CsA inhibits VEGF-induced DAF expression is likely to be multifactorial (71). We demonstrate that CsA inhibits VEGF-induced p38 MAPK phosphorylation, a response that could be reversed by inhibition of PI3K activity. Moreover, inhibition of EC p38 MAPK activation by CsA has been reported previously in intestinal microvascular ECs (71, 72). In addition to this, inhibition of calcineurin by CsA may lead to inhibitory effects at a transcriptional level. VEGF is known to trigger dephosphorylation, translocation, and transcriptional activity of EC NFAT, which is associated with AP-1 activation (52), and the activation of NF-κB (2). DAF expression has been associated with both AP-1 and NF-κB activity (37, 73). Thus, by targeting calcineurin, CsA may inhibit DAF transcription. However, the ability of CsA to inhibit PKC-mediated signaling pathways may also play a role. Although the effects on PKC activation may vary between cell types and are yet to be fully understood, CsA may inhibit both calcium-dependent cPKC, including PKCα (74), and calcium-independent nPKCϵ (75). The inhibition of DAF up-regulation by CsA had functional consequences, with VEGF-mediated cytoprotection against complement-induced cell lysis inhibited by the presence of CsA. This in turn may be an important contributory factor in CsA-mediated endothelial toxicity and may exacerbate the role of complement activation in posttransplant vasculopathy.

In conclusion, our data extend the role of PKC signaling in VEGF-mediated vascular cytoprotection. We have identified parallel linear PKCα/ϵ- and p38 MAPK-dependent pathways that integrate to induce DAF expression and that are negatively regulated by activation of PI3K and inhibited by CsA. Our data emphasize the importance of establishing a detailed understanding of VEGF receptor-mediated signaling pathways. This may in turn identify novel targets by which the endothelium can be therapeutically conditioned for the prevention and treatment of vascular inflammatory diseases involving complement activation.