Activation of Cytosolic Phospholipase A2-α as a Novel Mechanism Regulating Endothelial Cell Cycle Progression and Angiogenesis*

Release of endothelial cells from contact-inhibition and cell cycle re-entry is required for the induction of new blood vessel formation by angiogenesis. Using a combination of chemical inhibition, loss of function, and gain of function approaches, we demonstrate that endothelial cell cycle re-entry, S phase progression, and subsequent angiogenic tubule formation are dependent upon the activity of cytosolic phospholipase A2-α (cPLA2α). Inhibition of cPLA2α activity and small interfering RNA (siRNA)-mediated knockdown of endogenous cPLA2α reduced endothelial cell proliferation. In the absence of cPLA2α activity, endothelial cells exhibited retarded progression from G1 through S phase, displayed reduced cyclin A/cdk2 expression, and generated less arachidonic acid. In quiescent endothelial cells, cPLA2α is inactivated upon its sequestration at the Golgi apparatus. Upon the stimulation of endothelial cell proliferation, activation of cPLA2α by release from the Golgi apparatus was critical to the induction of cyclin A expression and efficient cell cycle progression. Consequently, inhibition of cPLA2α was sufficient to block angiogenic tubule formation in vitro. Furthermore, the siRNA-mediated retardation of endothelial cell cycle re-entry and proliferation was reversed upon overexpression of an siRNA-resistant form of cPLA2α. Thus, activation of cPLA2α acts as a novel mechanism for the regulation of endothelial cell cycle re-entry, cell cycle progression, and angiogenesis.

The vascular endothelium consists of a monolayer of endothelial cells that lines the luminal surface of all blood vessels in vivo. The endothelium actively participates in a variety of key vascular processes such as the regulation of vascular tone and blood fluidity. In addition, the endothelium regulates the formation of new blood vessels by the process of angiogenesis in development, tissue repair, and tumor vascularization (1,2). The mature endothelium consists of contact-inhibited conflu-ent monolayers of cells that reside in the G 0 phase of quiescence. Upon loss of cell-cell contacts, endothelial cells re-enter the cell cycle and proliferate. This entry of endothelial cells into the cell cycle from G 0 is a critical component of the angiogenic response and the formation of new capillaries from pre-existing blood vessels (1,2). Thus, the inhibition of endothelial cell proliferation has great potential for the treatment of diseases involving unwanted blood vessel formation.
The phospholipase A 2 (PLA 2 ) 3 family of enzymes hydrolyze the sn-2 group of glycerophospholipids to concomitantly release free fatty acids and lysophospholipids (3). The PLA 2 family represents a diverse family of enzymes that can be divided into three main groups as follows: the group IV cytosolic PLA 2 (cPLA 2 ), the group VI Ca 2ϩ -independent PLA 2 (iPLA 2 ), and the secretory PLA 2 enzymes (4). The cPLA 2 group of enzymes consists of at least six members (cPLA 2 ␣, -␤, -␥, -␦, -⑀, and -), of which cPLA 2 ␣ is the most extensively characterized. cPLA 2 ␣ is Ca 2ϩ -sensitive and translocates to intracellular membranes upon agonist stimulation and cytosolic Ca 2ϩ elevation utilizing an N terminal Ca 2ϩ -dependent lipid binding (C2) domain (5)(6)(7). Upon membrane binding, cPLA 2 ␣ preferentially cleaves phospholipids containing arachidonic acid (AA) at the sn-2 position to liberate free AA (3). As such, cPLA 2 ␣ is seen as the rate-limiting enzyme in receptor-mediated AA release (8). Proliferating, nonconfluent endothelial cells release much greater levels of arachidonic acid and prostaglandin than quiescent confluent cells (9 -11), which has been attributed to elevated cPLA 2 ␣ activity. In quiescent confluent cells, cPLA 2 ␣ is inactivated upon sequestration at the Golgi apparatus and is subsequently released and activated in proliferating cells (11,12). Despite this, the actual function of this differential regulation of cPLA 2 ␣ activity has not been defined.
Here we identify a novel role for cPLA 2 ␣ activation in the regulation of endothelial cell cycle progression. Upon the loss of cell-cell contacts and the induction of endothelial cell prolifer-ation, activation of cPLA 2 ␣ is required for the induction of cyclin A expression and efficient progression through G 1 and S phases. Our work and work by others have previously shown that the activity of iPLA 2 also influences the progression of endothelial cells through S phase (13)(14)(15). Here we demonstrate that cPLA 2 ␣ and iPLA 2 work cooperatively to influence endothelial cell cycle progression with cPLA 2 ␣ providing a stimulation-and Ca 2ϩ -dependent source of lipid metabolites required for controlling endothelial cell cycle progression in response to monolayer disruption or growth factor stimulation.
AA Release Assay-This assay was performed as described previously (11). Briefly, HUVECs were cultured to the required density in either 6-well or 60-mm 2 dishes and labeled for 24 h with 0.5 Ci/ml [ 3 H]AA in growth media. Subconfluent, confluent, or wounded cells were washed with PBS and incubated with cPLA 2 ␣ inhibitors for 15 min or overnight as indicated. For measurement of passive AA release, media were collected following overnight incubation (18 h), cleared by centrifugation, and assayed for radioactivity by liquid scintillation counting. For ionophore-stimulated release, cells were incubated with 5 M A23187 in HEPES/Tyrode Buffer (X) containing 2 mM CaCl 2 , supplemented with 0.3% (w/v) fatty acid-free bovine serum albumin for 15 min. Cleared supernatant containing released AA was assayed for radioactivity by liquid scintillation as were total cell lysates prepared by lysis of cells in 0.5% Triton X-100. AA release under each condition was expressed as a percentage of total cellular radioactivities.
Proliferation ELISA-HUVEC proliferation rates were assessed using a 5-bromo-2Ј-deoxyuridine (BrdUrd) incorporation-based ELISA (Roche Diagnostics). Cells were seeded at 1 ϫ 10 3 cells per well (0.55 ϫ 10 3 cells/cm 2 ) in 96-well plates, grown for 24 h, incubated for 16 h with inhibitors and BrdUrd, and then processed according to manufacturer's instructions. Proliferation was also determined after 48 h of growth by assessing viable cell numbers using the MTS-based CellTiter AQ ueous nonradioactive cell proliferation assay (Promega) according to the manufacturer's instructions.
Microscopy and Quantification-Phase contrast images were acquired using an Olympus CK2 inverted microscope (10ϫ lens) linked to an Olympus OM-1 camera. Deconvolution fluorescence microscopy was performed using an Olympus IX-70 inverted fluorescence microscope (63 ϫ 1.5 oil immersion lens) and DeltaVision deconvolution system (Applied Precision Inc.). Individual optical sections of 0.2 m were generated from 15 iterative cycles of deconvolution. Some images were collected using a Zeiss LSM510 META Axiovert 200 M confocal microscope. Multiple comparisons were performed using oneway analysis of variance and Tukey's post-test analysis with GraphPad Prism software.
Flow Cytometry-Subconfluent HUVECs were cultured for 16 h in the presence or absence of inhibitor. Cells were then harvested, fixed in 70% ice-cold ethanol, and incubated with propidium iodide (50 g/ml) and RNase A (20 g/ml) for 3 h. DNA content was then assessed using a FACSCalibur flow cytometer (BD Biosciences) and the percentage of cells in each phase of the cell cycle analyzed using Modfit software (Verity Software House). For some experiments G 0 -synchronized confluent HUVECs, serum-starved overnight, were seeded at subconfluent density in the presence or absence of inhibitors for 14 h. Cells were then either chased in fresh media containing inhibitors for various time points prior to processing as above, or siRNA-treated cells were pulsed with media supplemented with 10 M BrdUrd for 30 min prior to harvesting, fixation, and analysis for BrdUrd incorporation.
Biochemistry-Lysate preparation and Western analysis were performed as described previously (11). Immunoprecipitations were performed overnight at 4°C using protein G-Sepharose (Upstate Biotechnology, Inc.), 3 g of antibody, and 500 g of total protein in 1% Nonidet P-40 lysis buffer (50 mM Tris/HCl, 150 mM NaCl, 0.5 mM EDTA, 0.5 mM EGTA, 1% Nonidet P-40, pH 7.4). Samples (20 g of protein or total bead volume) were resolved for 60 min at 30 mA/gel on 10% SDS-PAGE mini-gels using a discontinuous buffer system (18). For immunoblotting, protein was transferred onto nitrocellulose membranes for 3 h at 300 mA (19). Membranes were blocked in 5% (w/v) nonfat milk in PBS for 30 min and then incubated overnight with primary antibody (1:500) at room temperature. After incubation with horseradish peroxidase-conjugated antigoat IgG (1:3000) for 1 h, immunoreactive bands were visualized using a West Pico enhanced chemiluminescence (ECL) detection kit (Pierce). Images were captured on a Fuji Film Intelligent dark box II image reader. Band intensities were determined densitometrically using Aida (Advanced Image Data Analyzer) 2.11 software.
Immunofluorescence Microscopy-Immunofluorescence was performed as described previously (11). Briefly, cells grown on coverslips were fixed in 10% (v/v) formalin in neutral buffered saline (HT50-1-128; Sigma) for 5 min at 37°C. After permeabilization with 0.1% (v/v) Triton X-100 for 5 min, cells were refixed (5 min), washed with PBS, and then incubated in 50 mM ammonium chloride for 10 min. Following PBS washes, nonspecific binding sites were blocked with 5% (v/v) donkey serum for 3 h. Cells were incubated overnight with primary antibody followed by the appropriate secondary antibodies. Finally, coverslips were mounted on microscope slides in Fluoromount-G mounting medium (Southern Biotech). For analysis of BrdUrd incorporation, cells were fixed in ice-cold 80% ethanol for 20 min followed by a 20-min fixative step in 100% ethanol on ice. Cells were then incubated with 2 N HCl, 0.5% Triton X-100 for 20 min at room temperature followed by a wash in 100 mM sodium tetraborohydrate. Cells were then processed as above with the indicated antibodies.
Differentiation and Migration Assays-To assess HUVEC differentiation, 1 ϫ 10 5 cells were seeded in 24-well dishes coated with 100 l of Matrigel (BD Biosciences). Cells were incubated in the presence or absence of inhibitor for 16 h prior to imaging. Tubule length was quantified using ImageJ software (rsb.info.nih). For migration assays, HUVECs (5 ϫ 10 4 cells) were seeded in serum-free media to the top chamber of 24-well modified Boyden chambers (3 m pores; Transwell-Costar Corp.). Cells were allowed to migrate toward serumcontaining media for 16 h in the presence of absence of inhibitor. Migrated cells were fixed, stained with DAPI, and then counted.
Angiogenesis Assay-Co-cultures of HUVECs seeded on a bed of human fibroblasts (TCS Cellworks) were cultured for 7 days in the presence or absence of inhibitor. Tubules were fixed, stained, and imaged by phase contrast microscopy. Tubule length was quantified using ImageJ software.

RESULTS
cPLA 2 ␣ Mediates Endothelial Cell Proliferation-We have previously demonstrated a requirement for both iPLA 2 -and cPLA 2 ␣-mediated AA release in the regulation of HUVEC proliferation (11,15). We sought to further examine the contribution of cPLA 2 ␣ to endothelial cell proliferation by measuring BrdUrd incorporation into DNA in the presence of concentrations of inhibitors that maximally block cPLA 2 ␣ activity. To initially define concentrations of inhibitors that maximally inhibited cPLA 2 ␣ activity, passive AA release generated by proliferating HUVECs was quantified following 18 h of incubation with varying concentrations of the cPLA 2 ␣-specific inhibitors pyrrolidine and Wyeth-1 (Wy-1) (20,21) or the cPLA 2 ␣/iPLA 2 inhibitor AACOCF 3 (22). Maximal inhibition of cPLA 2 ␣-mediated AA release was achieved with between 2.5 and 5 M pyrrolidine and 2.5 and 5 M Wy-1 with no further increase in inhibition observed at higher concentrations (Fig. 1A). AACOCF 3 also maximally inhibited AA release at 2.5 M, although the extent of this inhibition was much greater than with pyrrolidine or Wy-1 alone, consistent with the combined inhibition of both cPLA 2 ␣and iPLA 2 activities, which contribute to the pool of released AA by proliferating endothelial cells. The modest reduction in long term AA release following cPLA 2 ␣ inhibition may reflect both the assay conditions, where radiolabeled AA may be produced by non-cPLA 2 ␣-dependent mechanisms that cannot be controlled for (such as normal membrane turnover) and the contribution of the cPLA 2 ␣-generated pool to this process.
However, ionophore treatment selectively liberates AA derived from pools accessible to the Ca 2ϩ -dependent isoforms of PLA 2 and can thus be considered selective for the activation of cPLA 2 ␣ in HUVECs. Under these conditions, similar to the overnight release assay, short term (15 min) preincubation with 2.5 M pyrrolidine or Wy-1 was also sufficient to maximally inhibit Ca 2ϩ -induced AA release induced by the ionophore A23187 (Fig. 1B). AACOCF 3 appeared less effective at reducing stimulated AA release, and the varying extent of the inhibition by the various drugs may reflect the differing initial bioavailabilities of these structurally distinct compounds. A role for cPLA 2 ␣ in the regulation of endothelial cell proliferation was then confirmed upon incubation of HUVECs with 2.5 M pyrrolidine or Wy-1 ( Fig. 1, C and D). Maximal inhibition of cPLA 2 ␣ activity was sufficient to significantly retard HUVEC proliferation. Furthermore, incubation with 2.5 M pyrrolidine also significantly inhibited HDMEC proliferation (Fig. 1E) suggesting that a role for cPLA 2 ␣ in the regulation of cellular proliferation is common to other endothelial cell types. The antiproliferative effect of cPLA 2 ␣ inhibition could not be attributed to increased cell death as the concentration of inhibitors used did not affect cell viability (as assessed by trypan blue exclusion and annexin V binding) or caspase-3 activation and did not promote the cleavage of ␣II-spectrin or poly(ADP-ribose) polymerase (data not shown). However, varying levels of inhibitor cytotoxicity were observed at doses exceeding those used in this study (data not shown).
Specific analysis of the vascular endothelial growth factor A (VEGF-A) pathway showed that cPLA 2 ␣ activity is required for the induction of cellular proliferation by this important angiogenic stimulus. Indeed, VEGF-A-mediated cell turnover was reduced by 40 -50% in the presence of the various cPLA 2 ␣ inhibitors as determined by MTS-based assay of relative cell number after 48 h of growth (Fig. 1F). Furthermore, incubation with 5 M AACOCF 3 and the combined inhibition of cPLA 2 ␣ and iPLA 2 reduced HUVEC proliferation to levels similar to those found in the absence of growth factor and serum, suggest-   FEBRUARY 27, 2009 • VOLUME 284 • NUMBER 9 ing that cPLA 2 ␣ and iPLA 2 may function cooperatively during VEGF-induced endothelial cell proliferation. cPLA 2 ␣-mediated Proliferation Is Essential for Angiogenesis-Angiogenic tubule formation is a multistep process involving the proliferation, migration, and differentiation of endothelial cells (1,2). Consequently, conversion of endothelial cells from quiescence to a proliferative state is vital to angiogenesis (1). As cPLA 2 ␣ activity is required for endothelial cell proliferation (Fig. 1), we hypothesized that cPLA 2 ␣ activity plays a key role in angiogenesis. Endothelial cell tubule formation can be assessed using co-culture assays in vitro. In this assay, HUVECs were seeded on a bed of human dermal fibroblasts and cultured for 7 days. Under these conditions endothelial cells form tubules with patent lumens, reminiscent of mature capillaries (23). Tubule formation in this assay involves the combined proliferation, migration, and differentiation of endothelial cells. Incubation with pyrrolidine reduced tubule length to 32.8 Ϯ 4.8% of controls ( Fig. 2A). Wy-1 also showed a similar ability to reduce tubule formation ( Fig. 2A). Thus cPLA 2 ␣ activity regulates the formation of new blood capillaries by angiogenesis in in vitro assays.
To further examine the importance of cPLA 2 ␣ activity for endothelial cell proliferation during angiogenesis and the long term effects of pyrrolidine on cell viability, endothelial cell numbers were determined by counting over 7 days of culture (Fig. 2B). HUVECs seeded at a density of 3000 cells/cm 2 were given 24 h to settle and then were cultured for 7 days in the presence or absence of 2.5 M pyrrolidine. At 24-h intervals, cells were harvested and counted (Fig. 2B). Control cell numbers increased as a function of time, whereas cells grown in the presence of pyrrolidine remained at a constant density. Inhibition of growth was not because of cell death as incubation with pyrrolidine for 7 days did not affect cell viability, and endothelial cell morphology was unaffected by cPLA 2 ␣ inhibition (data not shown). Furthermore, this block in cell proliferation was entirely reversible as cells grown in the presence of pyrrolidine for 5 days recovered upon washout of inhibitor, with a 24-fold increase in cell density after 8 days of recovery (Fig. 2C). Similarly, culture in the presence of pyrrolidine also significantly blocks HDMEC growth (Fig. 2D). Thus, long term inhibition of cPLA 2 ␣ had a profound effect on cell growth and the ability of endothelial cells to form angiogenic tubules.
We also assessed the role cPLA 2 ␣ in endothelial cell migration, and differentiation. HUVEC migration, as assayed using Boyden chambers (23), was not affected by incubation with pyrrolidine (Fig. 2E). In addition, pyrrolidine did not inhibit endothelial cell migration or differentiation, as studied using Matrigel assays (Fig. 2F). Here a fixed number of cells were grown on an appropriate substratum for 24 h. Sufficient cells are present . C, G 0 -synchronized HUVECs were re-seeded subconfluently and allowed to re-enter the cell cycle for 14 h in the presence or absence of 2.5 M pyrrolidine. Cells were then processed at various subsequent time points with or without inhibitor prior to staining with propidium iodide and analysis by flow cytometry as in A. The percentage of G 0 -G 1 (left) and S phase cells (right) was determined using Modfit and expressed as a percentage of total cell number (n ϭ 4; *, p Ͻ 0.05 versus respective control; **, p Ͻ 0.01 versus control 0 h). D, cells prepared as in C were incubated with media supplemented with BrdUrd (Brdu, 10 M) for 45 min, 14 h after re-seeding. This was followed by ethanol fixation, staining with fluorescein isothiocyanate-conjugated anti-BrdUrd antibody, and analysis by flow cytometry. BrdUrd incorporation was expressed as a percentage of total cells and compared with confluent monolayer uptake (n ϭ 4; *, p Ͻ 0.05; **, p Ͻ 0.01). E, subconfluent HUVECs grown in the presence or absence of 2.5 M pyrrolidine for 18 h were analyzed for Ki67 levels. Membranes were re-probed for GRASP55 to show equal loading. Relative immunoreactivities were determined using Aida densitometry software and plotted (n ϭ 3, Ϯ S.E.). *, p Ͻ 0.01 versus control. F, subconfluent HUVECs grown in the presence or absence of 2.5 M pyrrolidine for 24 h were analyzed for PCNA, cyclin A, and cdk2 levels by Western blotting. Relative immunoreactivities were determined using Aida densitometry software and plotted (n ϭ 3, Ϯ S.E.). All results are representative of at least three separate experiments. *, p Ͻ 0.01 versus control.
to allow migration and differentiation into tubules even in the absence of cell proliferation (23). In this assay, HUVEC tubule formation, branching number, and branch length were not affected by inhibition of cPLA 2 ␣ with either pyrrolidine or Wy-1 (Fig. 2F). As cPLA 2 ␣ does not play a role in HUVEC migration or differentiation (Fig. 2, E and F), inhibition of endo-thelial cell proliferation caused by blocking cPLA 2 ␣ activity (Fig. 1, C and F) must be sufficient to block angiogenic tubule formation in coculture assays (Fig. 2A). The role of cPLA 2 in mediating this proliferation defect was examined in further detail below. cPLA 2 ␣ Modulates Endothelial Cell S Phase Progression and Cell Cycle Residence-The role of cPLA 2 ␣ in the regulation of cell cycle progression was assessed by analyzing the cell cycle distribution of endothelial cells using flow cytometry. Proliferating endothelial cells were cultured in the presence or absence of 2.5 M pyrrolidine for 16 h. Cells were then stained with propidium iodide, and cellular DNA content was analyzed by FACS. As determined by fluorescence-activated cell sorter, the number of cells in S phase was markedly reduced upon cPLA 2 ␣ inhibition (Fig. 3, A  and B). As a result, significantly more cells resided in the G 0 -G 1 phases of the cell cycle, suggesting that cPLA 2 ␣ modulates G 1 to S phase progression. To examine the rate of progression of HUVECs through the cell cycle upon cPLA 2 ␣ inhibition, serum-starved, quiescent confluent cells (G 0 synchronized) were trypsinized, re-seeded to subconfluent density, and allowed to re-enter the cell cycle for 14 h (optimized for maximal cell cycle re-entry, data not shown) in the presence or absence of pyrrolidine (2.5 M). Cells were then incubated in fresh media with or without pyrrolidine for various time points (0, 3, 6, 9 h), harvested at the indicated times, ethanol-fixed, stained with propidium iodide, and analyzed by flow cytometry. As shown in Fig. 3C, treatment of cells with pyrrolidine resulted in a significant increase (ϳ10%) in the number of cells in G 0 -G 1 phase, despite a timedependent reduction in the total number of cells in G 0 -G 1 . Importantly, the distribution of cells in S phase was also altered upon cPLA 2 ␣ inhibition over the 9-h chase period (Fig. 3C). Pyrrolidine-treated cells displayed both reduced numbers of cells in S phase (ϳ5% reduction in total cell number) and a delay in reaching S phase peak (ϳ6 h post-chase) compared with control cells (S phase peak reached ϳ3-4 h . siRNA-mediated knockdown of endogenous cPLA 2 ␣ expression inhibits endothelial cell proliferation. A, HUVECs were transfected with either no siRNA (control), nontargeted control siRNA (mock), or cPLA 2 ␣-targeted siRNA (siRNA-1 and siRNA-2). 48 h after transfection, cells were seeded at a subconfluent density and allowed to proliferate for 16 h prior to lysis and analysis of cPLA 2 ␣ levels by Western blotting. Membranes were re-probed for GRASP55 to show equal protein loading. Relative amounts of cPLA 2 ␣ and GRASP immunoreactivity were determined using Aida densitometry software and then plotted (n ϭ 3, Ϯ S.E.). B, ability of transfected cells to proliferate was assessed using a colorimetric ELISA based on BrdUrd (BrdU) incorporation, as described under "Experimental Procedures" (n ϭ 6, Ϯ S.E.). All results are representative of three separate experiments. *, p Ͻ 0.05 versus mock. C, HUVECs were transfected with either no siRNA (control), nontargeted control siRNA (mock), or cPLA 2 ␣-targeted siRNA (siRNA-1 and siRNA-2). 48 h after transfection, cells were seeded at a lower density and allowed to proliferate for 16 h prior to lysis. The levels of cdk2, cyclin A, and PCNA expression in siRNA-transfected HUVECs were then analyzed by Western blotting, and the resulting band immunoreactivity was determined using Aida densitometry software (n ϭ 3, Ϯ S.E.). All results are representative of at least three separate experiments. *, p Ͻ 0.05 versus mock. D, G 0 -synchronized siRNAtreated (control or cPLA 2 ␣-targeted siRNA) HUVECs were re-seeded subconfluently and allowed to re-enter the cell cycle for 14 h. Cells were pulsed with BrdUrd-containing media, fixed, stained with fluorescein isothiocyanate-anti-BrdUrd, and analyzed by flow cytometry. The percentage of BrdUrd-positive cells was expressed as a percentage of the total cell number (n ϭ 4; *, p Ͻ 0.05 versus respective control). FEBRUARY 27, 2009 • VOLUME 284 • NUMBER 9 JOURNAL OF BIOLOGICAL CHEMISTRY 5789 post-chase). As the total number of cells in S phase under control conditions is quite small (10 -15% total cell number), this reduction following cPLA 2 ␣ inhibition represents a significant decrease in proliferative potential. Additionally, the total number of BrdUrd-positive cells at time 0 (14 h after release from G 0 ) was significantly reduced (ϳ33% reduction) upon treatment with pyrrolidine (Fig. 3D). Taken together, these results indicate that cPLA 2 ␣ participates in the regulation of endothelial S phase progression.

cPLA 2 ␣ and S Phase Progression
Ki67 is a nuclear protein expressed in the G 1 , S, G 2 , and mitotic phases of the cell cycle but not in the G 0 phase of quiescence (24,25). It is commonly used as a marker for identifying proliferating cells, and a decrease in its expression is indicative of reduced proliferative rates (26). Inhibition of cPLA 2 ␣ significantly reduced Ki67 expression in subconfluent HUVECs (Fig.  3E), indicating that fewer cells were in the G 1 -M phase. Slower passage through the cell cycle may have led to the accumula-tion of quiescent cells displaying reduced Ki67 levels. Thus, cPLA 2 ␣ activity plays a key role in G 1 to S phase progression and maintenance of endothelial cell cycle residence.
Transition from G 1 to S phase requires the assembly and activation of the DNA replication complex to initiate DNA synthesis. S phase entry and replication complex formation can be monitored by assessing cellular levels of the replication clamp proliferating cell nuclear antigen (PCNA). PCNA expression is low throughout the cell cycle until early S phase and the initiation of replication complex formation (27,28). PCNA expression was unaffected by cPLA 2 ␣ inhibition (Fig. 3F), suggesting that inhibition of S phase progression was after S phase entry and replication complex formation. Activation of pre-assembled replication complexes and initiation of DNA synthesis is mediated by the cyclin A-cdk2 complex in early S phase. Mammalian cells cannot synthesize DNA nor progress through S phase in the absence of cyclin A-cdk2 activity (29,30). Cyclin A expression increases in early S phase in conjunction with the accumulation of cdk2 in the nucleus to form active cyclin A-cdk2 complexes.
Upon cPLA 2 ␣ inhibition, cyclin A levels were reduced to ϳ26% of controls, potentially accounting for the block in G 1 -S phase progression and inhibition of DNA synthesis seen previously (Fig. 3F). Furthermore, inhibition of cPLA 2 ␣ also resulted in reduced cdk2 expression levels (ϳ29% of control) as often occurs when S phase progression is artificially blocked (31,32). Thus, in the absence of cPLA 2 ␣ activity endothelial cells enter S phase and express similar levels of PCNA relative to control cells; however, they are unable to proceed through S phase to G 2 -M as efficiently, demonstrating that cPLA 2 ␣ modulates the progression of endothelial cells through S phase.
siRNA-mediated Knockdown of Endogenous cPLA 2 ␣-The previous experiments are consistent with the hypothesis that cPLA 2 ␣ plays a central role in the control of endothelial cell cycle residence and S phase progression. However, to exclude the possibility that the effects of cPLA 2 ␣ inhibitors were nonspecific, we used RNA interference. Consistent with the inhibitor studies, knockdown of endogenous cPLA 2 ␣ by either 78 or 68% of control levels using two different cPLA 2 ␣-specific siRNA duplexes (Fig. 4A) significantly inhibited HUVEC pro-FIGURE 5. cPLA 2 is activated at the wound border. A, confluent HUVEC monolayers were wounded and allowed to recover for 16 h in the presence or absence of serum and growth factors. Cell lysates were prepared and subjected to immunoprecipitation (IP) of cPLA 2 . Bound proteins were separated by SDS-PAGE and Western-blotted for phosphorylated serine 505 and total cPLA 2 . Lysates were also assessed for phosphorylated and total ERK1/2. con, control. B, quantification of phosphoserine 505 and phospho-ERK levels from six independent experiments. p Ͻ 0.05, n ϭ 6. C and D, confluent HUVECs were loaded with [ 3 H]AA overnight, wounded, and allowed to recover for 16 h in the presence or absence of 2.5 M Wy-1. C, released [ 3 H]AA was assessed by liquid scintillation. D, recovered cells were also stimulated with A23187 (5 M, 15 min), and released [ 3 H]AA was measured. *, p Ͻ 0.05, n ϭ 3. liferation (Fig. 4B). Relative to nontargeting siRNA controls (mock), the rate of BrdUrd incorporation into HUVEC DNA was reduced by 44 or 43% upon transfection with siRNA-1 or siRNA-2, respectively (Fig. 4B). Furthermore, the siRNA-mediated knockdown of endogenous cPLA 2 ␣ also resulted in reduced cellular levels of cyclin A and cdk2 while having no effect on PCNA expression (Fig. 4C). Additionally, cell cycle progression of siRNA-treated cells was examined by measurement of BrdUrd incorporation using flow cytometry. Subconfluent siRNA-transfected cells, released from G 0 14 h previously, were pulsed with BrdUrd (10 M) for 45 min prior to fixation, allowing an indication of S phase cell numbers to be established. Under these conditions, cPLA 2 ␣ siRNA-treated cells showed a significant reduction (to ϳ20% of scrambled siRNA (mock) levels) in the total number of cells in S phase and in cells undergoing active cell cycle progression (Fig. 4D). Thus, cPLA 2 ␣ activity plays a central role in the regulation of endothelial cell proliferation and S phase progression by modulating the expression of cell cycle proteins and influencing cell cycle residence time.
Activation of cPLA 2 ␣ by Release from the Golgi Is Critical to S Phase Progression-We have previously demonstrated that cPLA 2 ␣ activity is reduced in confluent monolayers of endothelial cells that are in the G 0 phase of quiescence (11), and we proposed that cPLA 2 ␣ is inactivated in confluent endothelial cells upon its sequestration at the Golgi apparatus and exclusion from its intracellular membrane substrate (12). Wounding of confluent monolayers provided the stimulus for cPLA 2 ␣ to be released from the Golgi apparatus and allow access its phospholipid substrate (11). To confirm that cPLA 2 ␣ is activated at the wound border following release from the Golgi apparatus, confluent monolayers were scratch-wounded in a grid pattern and recovered for 18 h prior to lysis. Following immunoprecipitation, the phosphorylation status of cPLA 2 ␣ on serine 505 was assessed by Western blotting as cPLA 2 ␣ can be activated downstream of multiple signal transduction cascades upon phosphorylation of serine 505 by ERK1/2. Wounding and recovery of HUVECs resulted in the growth factor-dependent phosphorylation of cPLA 2 consistent with the increased phosphorylation and activation of ERK1/2 (Fig. 5, A and B). Serum-starved HUVECs exhibited both reduced phospho-ERK levels and reduced phosphorylation on the critical Ser-505 residue of cPLA 2 . Importantly, the wounding and recovery of HUVEC monolayers also resulted in the increased release of AA during the recovery period (Fig. 5C). Recovered cells also displayed a dramatically increased capacity to release AA upon stimulation with the Ca 2ϩ ionophore A23187; this release was blocked by inhibition of cPLA 2 ␣ with Wyeth-1 (2.5 M ; Fig. 5D). These results are consistent with the release of cPLA 2 from the Golgi apparatus being critical for the activation of the enzyme.
Furthermore, subconfluent, actively proliferating endothelial cells display elevated levels of activated cPLA 2 ␣ compared with quiescent cells, as assessed by Western blotting of cPLA 2 ␣ immunoprecipitates for phosphorylated serine 505 (Fig. 6A). Again, the elevated phosphorylation status of cPLA 2 ␣ was dependent on the presence of growth factors in the media (Fig.  6, A and B). As reported previously (10), subconfluent endothelial cells also show a cPLA 2 ␣-dependent elevation in Ca 2ϩ -in- duced AA release compared with confluent HUVECs (Fig. 6C). Thus, cPLA 2 ␣ is consistently activated in proliferating, but not quiescent, endothelial cells, and this correlates with AA release.
As cPLA 2 ␣ modulates endothelial cell proliferation and S phase progression, we hypothesized that release from the Golgi and activation of cPLA 2 ␣ would be required for re-entry of quiescent endothelial cells back into the cell cycle. Upon mechanical wounding of confluent monolayers, cells at the wound borders enter into the cell cycle from G 0 , proliferate, and migrate into the denuded areas. After 24 h of recovery, HUVECs at the wound border (indicated by the diagonal line) begin to express PCNA and cyclin A at high levels as they enter and progress through the cell cycle (Fig. 7, A and B). As reported previously (11), cells at the wound border underwent relocation of cPLA 2 ␣ from the Golgi apparatus to the more diffuse staining of subconfluent HUVECs, indicating activation of cPLA 2 ␣ (Fig. 7, A-D). This activation of cPLA 2 ␣ would explain the elevated levels of AA released upon mechanical wounding of confluent endothelial cells shown in Fig. 6 (10). Importantly, release from the Golgi and activation of cPLA 2 ␣ were critical to the induction of cyclin A expression (Fig. 7, A-D). In confluent HUVEC monolayers mechanically wounded and recovered for 24 h in the presence of 2.5 M pyrrolidine, cPLA 2 ␣ had relocated in cells at the wound border, similar to un-inhibited cells (Fig. 7, A-D). Quantification of the number of cells expressing high levels of nuclear PCNA revealed that inhibition of cPLA 2 ␣ activity had no effect on PCNA up-regulation upon wounding (Fig. 7E), consistent with replication complex formation. However, the number of cells expressing high levels of nuclear cyclin A at the wound border fell from ϳ49 to 7% (i.e. 7-fold) in the presence of pyrrolidine (Fig. 7F). Highly expressing cells were easily distinguishable as their cyclin A nuclear fluorescence intensity was consistently ϳ4-fold than low expressing cells. Upon cPLA 2 ␣ inhibition, the number of cells expressing high levels of cyclin A at the wound border fell to the levels seen with quiescent confluent monolayers (Fig. 7, B and F).
A similar reduction in cyclin A-positive cells at the wound border was also found in cPLA 2 ␣ siRNA-and Wyeth-1-treated cells (Fig. 8A). Furthermore, only cells at the wound border exhibited BrdUrd incorporation that was significantly inhibited by Wyeth-1 (2.5 M) treatment (Fig. 8, B and C). We have previously demonstrated a role for iPLA 2 in the control of HUVEC proliferation and progression through S phase. Thus, we sought to examine whether there was a cooperative role for the two phospholipases A 2 in the control of S phase entry after monolayer wounding. BrdUrd incorporation was examined by immunofluorescence microscopy in wound-recovered HUVECs treated with siRNA to a scrambled sequence, cPLA 2 ␣ (siRNA-1), iPLA 2 (15), or cPLA 2 ␣ and iPLA 2 siRNA together. Consistent with previous findings, siRNA knockdown of cPLA 2 ␣ reduced the number of BrdUrd-positive cells at the wound border (Fig. 8D). Treatment with the iPLA 2 siRNA duplexes also resulted in an ϳ55% reduction in cells incorporating BrdUrd at the wound border. Importantly, the simultaneous knockdown of both cPLA 2 and iPLA 2 inhibited BrdUrd incorporation by ϳ80% in an additive manner (Fig. 8D). These results suggest that both Ca 2ϩ -dependent and Ca 2ϩ -independent phospholipase A 2 isoforms may cooperate to coordinate the response of endothelial cells to various proliferative stimuli.
To demonstrate the specificity of the cPLA 2 ␣ siRNA phenotype, the heterologous expression of an siRNA-resistant form of cPLA 2 ␣ (774 cPLA 2 ) using a vector containing an internal ribosome entry site-GFP sequence was found to rescue the prolif-eration deficit in siRNA-treated subconfluent HUVECs (Fig. 9A). A significant increase (ϳ6-fold) in the number of cyclin A-positive cells was evident upon restoration of cPLA 2 ␣ levels using the expression plasmid (Fig. 9B). Ki67 levels were also elevated under these conditions (Fig. 9A).
We have previously shown that the addition of exogenous AA to pyrrolidine-treated HUVECs was able to partially rescue the proliferation defect induced by cPLA 2 ␣ inhibition (11). Here we sought to assess whether AA could rescue other alterations in the cell cycle. Although re-expression of cPLA 2 ␣ was able to rescue the decrease in cyclin A and Ki67 levels induced by depletion of cPLA 2 ␣, we sought to examine whether the application of the cPLA 2 ␣ metabolites, AA or L-␣-lysophosphatidylcholine (LPC), could similarly rescue the expression of these key cell cycle proteins. Subconfluent proliferating HUVECs were treated with pyrrolidine together with either 10 M AA, 20 M AA, 20 M LPC, 20 M L-␣phosphatidylcholine (PC), or L-␣phosphatidylethanolamine for 48 h prior to analysis of cyclin A, cdk2, and Ki67 levels by Western blotting. As shown in Fig. 9C, AA unexpectedly reduced cyclin A expression while increasing the expression of cdk2 in control cells. However, addition of either PC, LPC, or L-␣phosphatidylethanolamine also increased the basal levels of cyclin A and cdk2, together with Ki67 indicating an increase in proliferation. Interestingly, only the addition of crude brain PC could completely reverse and/or prevent the cell cycle defects induced by cPLA 2 ␣ inhibition, restoring both cyclin A and cdk2 levels (Fig. 9D). Similar findings were seen with siRNA-treated HUVECs (data not shown). Why neither AA nor LPC was able to rescue the expression of these key S phase proteins remains unclear, but it may implicate a crucial role for the cPLA 2 ␣ protein itself or its modulation of cellular phospholipid content, as critical to the control of cell cycle progression.
Thus, under our experimental conditions, cell cycle re-entry of endothelial cells is at least partially dependent on cPLA 2 ␣ activation. The Ca 2ϩ -dependent translocation and activation displayed by cPLA 2 ␣ may allow it to preferentially respond to stimuli such as vessel damage or growth factors. Indeed, the contribution of cPLA 2 ␣ to regulating HUVEC proliferation appears more prominent in conditions where active detachment of cell-cell contacts is required by the cell prior to cell  FEBRUARY 27, 2009 • VOLUME 284 • NUMBER 9 cycle re-entry (i.e. the ϳ19% decrease in BrdUrd positive cells evident upon trypsinization and subconfluent re-seeding in Fig.  4D versus the ϳ55% decrease evident following wounding in Fig. 8C). Thus, the activation of cPLA 2 ␣ following monolayer disruption represents a novel mechanism mediating cyclin A-cdk2 expression and regulating S phase progression in endothelial cells. These results suggest that the Ca 2ϩ -dependent activation of cPLA 2 ␣ is an important step in mediating endothelial cell proliferation in damaged blood vessels as demonstrated by the disruption of angiogenic tubule formation upon inhibition of cPLA 2 ␣.

DISCUSSION
This study shows that cPLA 2 ␣ activation helps to regulate endothelial cell S phase progression and cell cycle entry. First, after mechanical wounding of confluent monolayers, endothelial cells at the wound borders release cPLA 2 ␣ from the Golgi apparatus resulting in its activation. Second, cPLA 2 ␣ activation is required for the elevation of cyclin A and cdk2 levels and subsequent S phase progression. Without active cPLA 2 ␣, the passage of endothelial cells through the early phases of the cell cycle (G 1 through S to early M) is retarded, which is sufficient to interfere with long term angiogenic tubule formation. Thus, cPLA 2 ␣ activation is a requirement for the efficient escape of quiescent endothelial cells from G 0 and entry into the cell cycle. Other studies have implicated arachidonic acid metabolites in the control of G 1 to S phase progression of cells (33)(34)(35); however, our findings provide the first evidence for the central role of cPLA 2 ␣ itself in cyclin A expression, S phase progression, and cell cycle entry. A role for the protein itself is further validated by the failure of AA or LPC, the two major products of cPLA 2 ␣ lipase function, to rescue cyclin A or cdk2 levels, despite previously being shown to partially rescue proliferation (11). Furthermore, in Chinese hamster ovary cells and neuroblastoma N2A cells, cPLA 2 ␣ activity is elevated during mid/late G 1 and following G 1 to S phase transition (36); and in rat thyroid cells, the cPLA 2 ␣-dependent production of glycerophosphoinositol is required for FIGURE 9. Overexpression of siRNA-resistant cPLA 2 ␣, but not AA addition, recovers cyclin A expression. A, HUVECs mock-or siRNA-1-transfected were electroporated with pIRES-GFP-774 cPLA 2 ␣ construct (774 cPLA 2 ) or empty vector and plated onto coverslips. 24 h later, cells were fixed and processed for GFP, cyclin A, and Ki67 expression analysis by confocal microscopy. DAPI, DAPI. B, cells positive for both cyclin A and GFP were quantified and expressed as a percentage of total GFP-positive cells. Images are representative of three separate experiments. ns, not significant. **, p Ͻ 0.01. C, subconfluent HUVECs were treated with the indicated concentrations of AA, PC, LPC, or L-␣-phosphatidylethanolamine (PE) for 48 h prior to lysis and analysis of Ki67, cPLA 2 ␣, cyclin A, cdk2, or ␣-tubulin levels by Western blotting. D, subconfluent HUVECs were treated and analyzed as in C, except pyrrolidine was added during the 48-h incubation period. Also shown is a control, mock-treated lysate. Images are representative of at least four separate experiments.
In addition, we found that inhibition of cPLA 2 ␣ in other cell types (saphenous vein smooth muscle cells, MCF-7 carcinoma cells, and epithelial HeLa cells) dose-dependently blocked cellular proliferation (data not shown), suggesting that the role of cPLA 2 ␣ activity in S phase progression is common to other cell types. However, the precise roles of AA and its downstream products as well as lysophospholipids in regulating proliferation still require clarification. This is partially due to technical issues with lipid stability and also cellular requirements for lipid by-products to be produced close to their sites of action for efficient downstream activity (i.e. AA and cyclooxygenase and lipoxygenase enzymes) (38 -40). Despite this, inactivation of cPLA 2 ␣ by sequestration at the Golgi apparatus is unique to endothelial cells with this phenomenon not evident in other cell types examined (e.g. HeLa, Madin-Darby canine kidney cells, A549, saphenous vein smooth muscle, and EA.hy.926 cells; data not shown).
Endothelial cells undergo the contact-inhibition of proliferation at confluence and become quiescent. As cPLA 2 ␣ activity is a requirement for cell cycle re-entry, the existence of a readily releasable pool of cPLA 2 ␣ in quiescent endothelial cells could be utilized to allow rapid cell cycle re-entry upon the loss of cell-cell contacts. The role that serine phosphorylation plays in regulating both cPLA 2 ␣ activity and its release from the Golgi apparatus requires defining and could represent an important target for modulating phospholipid turnover. This is highlighted by the recent finding that phosphorylation of cPLA 2 ␣ on serine 727 disrupted binding to p11 and annexin A2 allowing the enzyme to access its phospholipid substrate (41). Another unique function of endothelial cells is their ability to proliferate, migrate, and differentiate to form new capillaries during angiogenesis. The induction of this process is important in wound healing and critical for solid tumor growth, resulting in antiangiogenic drugs being one of the most promising avenues of anti-cancer therapies (42)(43)(44)(45). The re-entry of quiescent endothelial cells into the cell cycle is critical to activate the angiogenesis program (1, 2), and we have shown that inhibition of cPLA 2 ␣ was sufficient to prevent angiogenic tubule formation, due solely to defects in the endothelial cell proliferation machinery. As such, targeting of endothelial cPLA 2 ␣, either alone or in combination with anti-iPLA 2 -VIA therapy, could represent a new approach to inhibit tumor neovascularization. Indeed, cPLA 2 ␣ activity and the generation of LPC were recently identified as crucial events in the protection of endothelial cells against radiation-induced apoptosis and may represent a new target for modulating the radiosensitivity of the endothelium (46).
We had previously demonstrated that inhibition of group VIA iPLA 2 (iPLA 2 -VIA), but not iPLA 2 -VIB or secretory PLA 2 , also blocked endothelial cell S phase progression and cell proliferation (15). Our work now suggests that both cPLA 2 ␣ and iPLA 2 -VIA play distinct roles in the regulation of endothelial cell cycle progression. Here we show that products of both enzymes are required for the efficient release of endothelial cells from contact inhibition and re-entry into the cell cycle. Thus, in vivo, cells could sense the products (i.e. AA, lysophos-pholipids, or as suggested by our results, a reduction in membrane phospholipids) of multiple PLA 2 s to determine their readiness for progression into S phase; this would provide a mechanism to coordinate the amount of phospholipid required by a cell to successfully progress though cellular division. Previous studies have linked phosphatidylcholine metabolism to cell cycle regulation, and its incorporation is an S phase-specific event (47)(48)(49). In a recent study, changes to membrane fluidity caused by alteration to iPLA 2 activities were hypothesized to be responsible for the growth defects seen upon iPLA 2 inhibition; inhibition of the Ca 2ϩ -independent isoform of PLA 2 led to an increase in membrane PC and resulted in G 1 phase arrest (13,14). Furthermore, it has been reported that excess PC can stimulate iPLA 2 -VIA activity, providing a mechanism for the ability of PC to recover the proliferation defect induced by cPLA 2 ␣ inhibition (50). Additionally, in contrast to cPLA 2 , the activity of iPLA 2 -VIA appears to be required for the migration of HUVECs as well as their proliferation, 4 which represents an important difference between their cellular activities and represents an area of future study. Defining how changes in phospholipase activities and the subsequent defects in phospholipid metabolism impact upon the cell cycle is crucial to utilizing this pathway as a therapeutic target.
However, it remains unclear whether cPLA 2 ␣ and iPLA 2 -VIA regulate cellular proliferation by distinct or overlapping coordinated mechanisms. Initial studies suggest that cPLA 2 ␣ and iPLA 2 -VIA have a synergistic effect on cell proliferation, and this may reflect both their different cellular localization (i.e. Golgi versus cytoplasm, respectively) and sites of their preferred substrates (i.e. endoplasmic reticulum/perinucleus versus plasma membrane, respectively), as well as their varied modes of activation. These variations may allow a cell to subtly coordinate its proliferative response to intracellular and extracellular cues by manipulating cellular lipids. The challenge will be to define the precise mechanisms by which cPLA 2 ␣ and other phospholipase A 2 enzymes regulate cell cycle progression.