Enhancement of migration by protein kinase Cα and inhibition of proliferation and cell cycle progression by protein kinase Cδ in capillary endothelial cells

Activation of protein kinase C (PKC) induces angiogenesis, migration, and proliferation of endothelial cells (EC), but can also prevent growth factor-induced EC proliferation. To determine whether these disparate effects are mediated by substrates of individual PKC isoenzymes, PKCα and PKCδ were overexpressed in rat microvascular EC. Basal and stimulated migration were enhanced in PKCα EC compared with either PKCδ or control EC. Serum-induced growth of PKCδ EC was decreased, while that of PKCα cells was similar to control EC. Phorbol ester markedly inhibited PKCδ EC growth but enhanced growth of PKCα and control EC. To determine possible causes for this altered proliferation, the effect of PKCδ on adhesion, mitogen-activated protein kinase activity, and cell cycle progression was measured. Adherence of PKCδ EC to vitronectin was significantly enhanced. Serum-induced extracellular signal-regulated kinase-2 activity was increased equally in both PKCα and PKCδ EC above that of control, while extracellular signal-regulated kinase-1 activity was similar in all EC. Cell cycle analysis suggested that PKCδ EC entered S phase inappropriately and were delayed in passage through S phase. Thus, PKCα may mediate some proangiogenic effects of PKC activation; conversely, PKCδ may direct antiangiogenic aspects of overall PKC activation, including slowing of the cell cycle progression.

The formation of new blood vessels and the repair of those damaged by disease or injury depend upon endothelial migration and proliferation (1,2). Several external agents that promote or inhibit proliferation and migration have been identified (3,4), but the intracellular messengers that mediate these processes are less clear.
Activation of the serine-threonine kinase protein kinase C (PKC) 1 by phorbol esters induces migration, proliferation (5), and tube formation of cultured endothelial cells (6,7) and causes angiogenesis in vivo (7)(8)(9). In addition, chemical inhibitors of PKC or the down-regulation of PKC by prolonged treatment with phorbol esters abrogates the proliferative effects induced by growth factors and mitogens (10,11) and also enhances endothelial permeability (12,13) and alters the expression level of several fibrinolytic enzymes and their inhibitors (14). In contrast, treatment of endothelial cells with direct activators of PKC alters some responses that are usually associated with stimulation by physiologic agonists (15) and, under some conditions, can prevent growth factor-induced proliferation (16). This apparent paradox might be explained by the fact that the PKC family is composed of related but structurally distinct isoenzymes, each a product of separate genes and with discrete cofactor requirements, substrate specificity, and tissue distribution (16 -18). Since phorbol esters activate multiple isoenzymes of PKC, the possibility is raised that each PKC isoenzyme may selectively mediate separate, and perhaps opposing, effects within stimulated endothelial cells.
Preliminary studies in our laboratory revealed that rat capillary endothelial cells expressed several isoenzymes of PKC, including PKC␣, -␦, -, -, and -. Of these isoenzymes, previous investigations have found that overexpression of PKC␣ or PKC␦ in various cultured cells could affect their proliferation (19 -21). Overexpression of PKC␣ in fibroblasts promoted their proliferation, while proliferation was inhibited in human breast cancer cells and other cell lines (20,(22)(23)(24); similar alterations of cell growth have been observed in cells overexpressing PKC␦ (20,21,25). Thus, the possibility exists that activation of either of these isoenzymes mediates the inhibitory component of PKC activation on endothelial growth, while the other promotes one or more processes essential to endothelial repair and angiogenesis. Such an effect would be presumably mediated by isoenzyme-specific substrates, of which a few have been identified (e.g. elongation factor eEF-1␣ (26)). In the present study, we began by testing the effect of overexpression of PKC␣ and PKC␦ on endothelial migration and proliferation, both of which are essential processes for angiogenesis and wound healing. When initial experiments revealed an inhibitory effect of PKC␦ on endothelial proliferation, we then examined potential underlying causes.

Construction of Rat Fat Pad Epididymal Endothelial Cell (RFPEC)
Cell Lines Overexpressing PKC␣ and PKC␦-The RFPEC were a generous gift from R. D. Rosenberg (MIT) (27,28) and were propagated in M199 medium (Life Technologies, Inc.) supplemented with 2 mM L-glutamine, penicillin (10 units ml Ϫ1 ), streptomycin (10 units ml Ϫ1 ), and amphotericin B (250 ng ml Ϫ1 ). To obtain the RFPEC that stably express vector (control), PKC␣, or PKC␦, pcDNA-Neo (Invitrogen, Inc.), pcDNA-bPKC␣, or pcDNA-hPKC␦ constructs, respectively, were transfected into early passage RFPEC cells by the calcium phosphate precipitation technique. Following selection for resistance to Geneticin, a number of vector-transfected, PKC␣-transfected, and PKC␦-transfected clones of endothelial cells (designated as control EC, PKC␣ EC, and PKC␦ EC) were isolated and expanded, and the mRNA was examined by Northern blot analysis for expression of the respective transcripts.
Determination of Protein Kinase C Activity-Cells were removed from subconfluent cultures of stably transfected RFPEC by trypsin and washed with PBS. Total cell counts were determined using a Coulter counter. Equivalent numbers of cells were pelleted, resuspended in homogenization buffer (50 mM Tris, pH 7.5, 5 mM EDTA, 10 mM EGTA, 50 mg ml Ϫ1 N ␣ -p-tosyl-L-chloromethyl ketone, 100 mg ml Ϫ1 N-tosyl-Lphenylalanine chloromethyl ketone, 100 mg ml Ϫ1 phenylmethylsulfonyl fluoride, 2 mg ml Ϫ1 leupeptin, 0.3% ␤-mercaptoethanol), and briefly sonicated. The cytosolic and cytoskeletal (the latter defined by its lack of solubility in 1% Triton X-100) fractions were obtained by differential centrifugation. PKC kinase activity was determined by measuring phosphorylation of a PKC-specific peptide substrate, based on a conserved region of the epidermal growth factor receptor, using the protein kinase C enzyme assay system purchased from Amersham Life Science, Inc.
RNA Analysis-The EC were removed from the culture dish with trypsin and washed with PBS. The cell pellet was resuspended in lysis buffer (140 mM NaCl, 1.5 mM Mg 2 Cl, 0.5% Triton X-100, 15 mM Tris, pH 8.3), vortexed for 30 s, and incubated on ice for 10 min. The nuclei were pelleted, and the supernatant was transferred to a fresh tube. An equivalent volume of buffer containing 25 mM EDTA, 300 mM NaCl, 2% SDS, 200 mg ml Ϫ1 proteinase K, and 200 mM Tris, pH 7.5, was added to the supernatant and incubated at 65°C for 1 h. Total RNA was extracted with phenol/chloroform, precipitated with ethanol, and resuspended in DEPC-treated double distilled H 2 O. For Northern transfer analysis, 20 g of total RNA was loaded per lane, subjected to electrophoresis on a 2% formaldehyde-agarose gel, and transferred to Gene-Screen Plus membrane according to the manufacturer's recommendations. The blot was UV-cross-linked (Stratalinker, Stratagene) and hybridized with random primed cDNA probes at 65°C for 3 h in Quik-Hyb solution (Stratagene). Blots were washed under high stringency conditions. Audioradiographs were analyzed using a scanning densitometer.
Immunohistochemistry-Stably transfected cells were propagated on coverslips placed in 12-well dishes and grown in the absence of serum for 24 h prior to immunofluorescence analysis of the PKC isoenzymes. The cells were washed with PBS and fixed in 2% paraformaldehyde in PBS for 10 min. The cells were then washed with PBS supplemented with 1% bovine serum albumin (PBS-BSA) and rendered permeable in 0.1% Triton X-100, PBS-BSA solution for 10 min. The cells were then washed and incubated in 10% goat serum, PBS-BSA solution for 30 min at room temperature. The solution was removed, and the cells were incubated in the presence of an optimal concentration of primary antibody (a 1:50 dilution of the PKC␣, PKC␦ (Santa Cruz Biotechnology), or histone (Accurate Biochemicals, Inc.) antibodies) in 2% goat serum, PBS-BSA solution for 1 h. The cells were washed with 0.05% Triton X-100 in PBS-BSA. The biotinylated anti-rabbit or anti-mouse IgG (Vector Laboratories) was added at a 1:200 final dilution and incubated at room temperature for 1 h. The cells were washed, and streptavidin-Texas Red (Amersham Life Sciences) was added at a 1:200 dilution and incubated for 30 min at room temperature. Cells stained for the Golgi apparatus were treated with 0.5 M BODIPY TR ceramide (Molecular Probes) for 1 h. Cells were washed once, and the coverslips were placed face-down on a slide with FluorSave (Calbiochem). The cells were visualized under ϫ 40 magnification by fluorescence microscopy.
Cell Migration Assay-Stably transfected cells were grown to confluency and then incubated in serum-free M199 for 24 h prior to the start of the assay. The cells were removed from the culture dish with trypsin, washed once in PBS, and resuspended in serum-free M199 at a final concentration of 1 ϫ 10 6 cells/ml of medium. Chemotactic agents were placed in the lower wells of a 48-well microchemotaxis chamber (Neuro Probe, Inc.) at indicated concentrations. An 8-m porous, polyvinylpyrrolidone free, polycarbonate membrane (Poretics Corp.), precoated with collagen type I, was placed between the upper and lower wells of the chemotaxis chamber. The cell suspension was then added to the upper wells of the chamber at a density of 5 ϫ 10 4 cells/well. Chemotaxis was assayed over 4 h at 37°C in a CO 2 incubator, under both unstimulated (basal) and stimulated (25 ng of hepatocyte growth factor/scatter factor per ml of medium) conditions. The membrane was removed from the chamber, fixed in 70% ethanol for 20 min at Ϫ20°C, and stained in hematoxylin overnight. The upper surface of the stained membrane was scraped using a cotton swab, leaving only the cells that migrated to the undersurface. Migration was assessed by counting the number of cells on the lower surface of the membrane at a ϫ 200 magnification by light microscopy.
Proliferation Assay-Endothelial cells were seeded at equivalent densities in six-well culture dishes and allowed to adhere overnight in complete medium. Following a wash with PBS, the cells were incubated for 24 h in M199 without serum. The cells were subsequently stimulated by the addition of enriched serum concentrations (either 1 or 15%) in the presence or absence of 1 M phorbol 12-myristate 13-acetate (PMA). At the indicated times following stimulation, the cells were treated with trypsin, and the cells were counted with a Coulter counter.
Cell Adhesion Assay-The determination of endothelial cell adhesion was performed as described previously (29) with some modification. Briefly, Corning 96-well enzyme-linked immunosorbent assay plates were coated with 1 mg of purified human vitronectin resuspended in Ca 2ϩ , Mg 2ϩ -free PBS (pH 7.4) for 1 h at 37°C. The plates were rinsed with the same buffer, coated with 1% heat-denatured BSA in the same buffer, and incubated at room temperature for 30 -60 min. Actively growing (50 -80% confluent) cultures of RFPEC were removed from the culture plate with trypsin, washed, and resuspended in M199 supplemented with 0.5% BSA. Cells were plated in each matrix-coated well at 2.5 ϫ 10 4 cells/well and incubated at 37°C for 60 -90 min. The unadhered cells were removed, and the wells were gently washed. Adherent cells were detected by incubating in the presence of 6 mg ml Ϫ1 pnitrophenylphosphate (Sigma) in 50 mM sodium acetate (pH 5), 1% Triton X-100 for 1 h at 37°C, followed by the addition of 1 N NaOH, and the absorbance was determined at 405 nm using an enzyme-linked immunosorbent assay plate reader.
Flow Cytometry-The cells were seeded, synchronized by serum deprivation for 72 h, and stimulated as described for the proliferation assay. DNA flow cytometric analysis of stably overexpressing EC cells was performed using a technique described by Tennenbaum et al. (30). The endothelial cells were removed from the culture plate by treating with 0.003% trypsin in sample buffer (3.4 mM sodium citrate, 0.1% Nonidet P-40, 1.5 mM spermine HCl, 0.5 mM Tris-Cl, pH 7.6). The reaction was stopped by the addition of 0.05% trypsin inhibitor, 0.01% ribonuclease A in sample buffer. The cells were subsequently treated with ice-cold 0.042% propidium iodide, 0.116% spermine HCl in sample buffer. The cells were kept on ice and in aluminum foil until analyzed. Flow cytometry was performed on a FACStar plus flow cytometer at an excitation of 488-nm wavelength and 630DF22 emission, and the data were analyzed using the Verity MODFIT software.
Stably transfected control, PKC␣, and PKC␦ EC were removed with trypsin from the culture plate and resuspended in PBS at 5 ϫ 10 6 cells ml Ϫ1 for the flow cytometric analysis of the surface expression of ␤ 3 or ␤ 5 integrins. Fifty-microliter aliquots of the cell suspensions were incubated in the presence of optimal concentrations of the primary antibody (1 g of rabbit anti-rat ␤ 3 IgG fluorescein isothiocyanate-conjugated (Pharmingen, Inc.) or 0.5 l of rabbit anti-human ␤ 5 polyclonal antibody (Chemicon, Inc.) per 50 l reaction) at room temperature for 15 min. The cells were pelleted at 1000 ϫ g for 5 min at 4°C. When necessary, the cells were resuspended in 50 l of PBS and incubated with appropriate volumes of the secondary antibody (1.2 mg ml Ϫ1 donkey antirabbit IgG fluorescein isothiocyanate-conjugated (Jackson Immunochemicals, Inc.)) for 15 min at room temperature. The cells were washed twice with PBS and resuspended in 200 l of PBS. Fluorescence bound to cells was detected with a FACStar plus flow cytometer set at a 488-nm excitation wavelength and 530DF30 emission.
MAP Kinase Activity Assay-Vector control, PKC␣, and PKC␦ EC were rendered quiescent for 24 h prior to the assay. The cells were stimulated with M199 supplemented with 15% FBS or 1 M PMA for the indicated times. The cells were then harvested by removing the medium, washing once with ice-cold PBS, and incubating in radioimmune precipitation buffer (50 mM Hepes, pH 7.4, 150 mM NaCl, 5 mM EGTA, 5 mM EDTA, 20 mM NaF, 20 mM sodium pyrophosphate, 1% Triton X-100, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride) on a rocking incubator at 4°C for 30 min. The lysed cells were scraped from the culture dish and transferred to a microcentrifuge tube. Large cellular debris was removed from the protein suspension by centrifugation at 15800 ϫ g for 5 min at 4°C. The cleared protein extracts were transferred to a fresh microcentrifuge tube, and total protein concentration was determined by means of the bicinchoninic acid assay (Pierce). MAP kinase was immunoprecipitated using extracellular signal-regulated kinase (ERK)-1 or ERK-2 antibody (Santa Cruz Biotechnology) from extract containing equal amounts of protein.
After the sample volumes were adjusted with radioimmune precipitation buffer, the ERK-1 or ERK-2 antibody (1 g of antibody/250 g total protein) was added. The extracts were incubated on a rocking incubator at 4°C overnight. The immune complexes were pelleted with protein A-agarose (Life Technologies, Inc.), washed three times in radioimmune precipitation buffer, and suspended in 50 l of kinase buffer (10 mM Tris-Cl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 ). ERK-2 activity was assayed by incubating 20 l of each sample with 20 l of the reaction mixture (8 mg of myelin basic protein (Sigma), 0.5 Ci of [␥-32 P]ATP (specific activity 3000 mCi/mmol) (DuPont NEN), and 10 M ATP) for 30 min at 25°C. The reaction was quenched with 15 l of 4 ϫ Laemmli buffer. The phosphorylated myelin basic protein was then resolved on a 15% SDS-polyacrylamide separating gel with a 4% polyacrylamide stacking gel and visualized by autoradiography.

Isolation and Characterization of Overexpressing Cell
Lines-To investigate the role of PKC␣ and PKC␦ isoenzymes in relation to angiogenesis, stably transfected microvascular RFPEC were established that overexpressed the eukaryotic expression vector pcDNA-Neo containing the complete cDNA sequence of PKC isoenzyme PKC␣ or PKC␦ or, as a control, the expression vector without an inserted gene (PKC␣ EC, PKC␦ EC, and control EC, respectively). The stably expressing RF-PEC cell lines were selected by neomycin resistance and screened by Northern blot analysis for gene expression (Fig. 1).
The stably transfected RFPEC displayed the cobblestone morphology typical of endothelial cells and were not visibly altered by transfection or overexpression. Immunoblot analysis demonstrated increased protein production of the corresponding PKC isoenzymes. The enzymatic activity of PKC was determined in several cell lines. Following this initial analysis, two clones of each type (PKC␣ EC and PKC␦ EC) were chosen for further study on the basis of similar levels of total kinase activity. Table I summarizes the PKC activity of the control and the two selected PKC␣ and PKC␦ EC lines, revealing total PKC activity that was increased in both the cytosolic and cytoskeletal fractions and was comparable between PKC␣ and PKC␦ EC.
To ensure that overexpression of PKC␣ and PKC␦ isoenzymes in the endothelial cells did not cause abnormal subcellular localization, we assessed the intracellular location of these PKC isoenzymes by immunofluorescence in both quiescent and PMA-treated EC. Experiments in PMA-treated EC were performed because activation of some PKC isoenzymes is associated with their redistribution into distinct subcellular locations (31). In quiescent control EC, PKC␣ could be detected primarily within the cytoplasm and nucleus ( Fig. 2A). Staining of these cells with antibodies directed against histone proteins or with a fluorescently tagged ceramide confirmed the nuclei and Golgi apparatus structures (Fig. 2, I and J), respectively. Following a 10-min incubation with PMA, PKC␣ could still be seen in the nucleus, but also at the periphery of the cell along the plasma membrane (Fig. 2B). Interestingly, PKC␣ translocated primarily to regions of the plasma membrane in which there was cell-cell contact. A similar cellular distribution of PKC␣ was noted in both the PKC␣ EC (Fig. 2, C and D) and in the PKC␦ EC (data not shown). Immunofluorescent staining for PKC␦ demonstrated the nuclear and cytosolic location for this isoenzyme in serum-starved control EC (Fig. 2E). PKC␦ redistributed to the plasma membrane and the nuclear membrane upon activation with PMA (Fig. 2F). A similar pattern of staining for PKC␦ was noted in the PKC␣ (data not shown) and PKC␦ (Fig. 2, G and H) EC. Thus, constitutive overexpression of PKC␣ or PKC␦ did not affect normal cellular localization in either stimulated or quiescent endothelial cells, although enzymatic activity was increased.
Effect of PKC Isoenzymes on Endothelial Cell Migration-To determine the role of PKC␣ and ␦ in endothelial cell migration, the respective stably transfected cell lines were seeded in a microchemotaxis chamber, and the number of endothelial cells that migrated through the polycarbonate membrane was determined as described under "Experimental Procedures." When hepatocyte growth factor (HGF or scatter factor), a powerful stimulus for migration and angiogenesis (32,33), was utilized as the agonist, PKC␣ EC traversed the membrane at a significantly greater rate than did PKC␦ or control EC (Fig. 3), suggesting a migratory response mediated by PKC␣ to this stimulus. The basal rate of migration (i.e. that occurring in the absence of any chemotactic agent) of PKC␣ was also consistently greater than that of the control EC or PKC␦ EC (Fig. 3), further implicating a specific role for PKC␣ in enhancing endothelial cell migration. Thus, both basal and agonist-stimulated endothelial migration differed between PKC␣ and PKC␦ EC, and PKC␣ EC migration was enhanced from the response seen in control EC.
Effect of PKC␣ and PKC␦ on Endothelial Cell Proliferation-Stimulation of quiescent PKC␣ EC with low (1%) concentrations of serum induced a growth rate similar to that of the vector control EC (Fig. 4A). In contrast, PKC␦ EC exhibited much less proliferation in response to serum stimulation than did either the control or PKC␣ EC. Proliferation of the control EC and PKC␣ EC in response to PMA was mildly enhanced   (Fig. 4B), but the growth rate of PKC␦ EC was significantly inhibited further by PMA, with an inhibition of 46.2 Ϯ 8.8% below that of non-PMA-treated PKC␦ EC at 72 h (p Ͻ 0.05) (Fig. 4C). This inhibition of serum-induced growth to serum stimulation was noted in both clones of PKC␦ EC tested; neither of the PKC␣ EC clones exhibited altered growth. Qualitatively similar responses were seen when quiescent cells were stimulated with higher (15%) serum concentrations in the presence or absence of PMA (data not shown). Thus, overexpression and stimulation of one isoenzyme (PKC␦) blocked endothelial proliferation, a response not seen when PKC␣ was overexpressed to a similar degree.
Effects of PKC␣ and PKC␦ Overexpression on Endothelial Cell Adhesion-In order to better understand the cause of the decreased proliferation in PKC␦ EC, we next asked whether PKC would lessen adhesion to extracellular matrices, prevent mitogen-activated protein kinase (MAP kinase) activation, or alter cell cycle progression in EC. To determine whether integrin-mediated endothelial adhesion, an event that is required for endothelial proliferation and angiogenesis (29,34,35), was lessened by PKC␦, we examined the adhesion of subconfluent cultures of control, PKC␣, and PKC␦ EC seeded on vitronectincoated plates. Rather than being diminished, the ability of the PKC␦ EC to adhere to the extracellular matrix was significantly enhanced above that seen with the control EC (Fig. 5), with a mean increase in adherence of 32.4% (p Յ 0.005). In contrast, increased PKC␣ expression did not significantly alter the ability of the endothelial cells to adhere to vitronectin. Preincubation of these cells with a synthetic peptide, GRGDSP, which corresponds to the vitronectin protein sequence that directly interacts with the integrin receptor binding domain, abolished adherence of these cells, thus demonstrating that the base line adherence of the cells, plus that enhanced in PKC␦ EC, was specific for the integrin receptors. To determine whether the enhanced adhesion resulted from increased expression of integrin receptors on the cell surface, we analyzed the cellular surface expression level of ␣ v ␤ 3 and ␣ v ␤ 5 by immunofluorescence using flow cytometry. Neither the overall cellular surface expression level of ␤ 3 nor that of ␤ 5 was significantly enhanced in PKC␦ EC as compared with PKC␣ or control EC (data not shown). Thus, enhanced adhesion in PKC␦ EC most probably resulted from increased affinity modulation of the integrin receptors. These results, therefore, demonstrate that the reduction in cell growth in PKC␦ did not result from impaired adhesion to extracellular matrices.
Serum-induced ERK-2 Activation in PKC␣ and PKC␦ EC-We next tested the possibility that impaired PKC␦ EC growth resulted from impaired activation of one of the MAP kinases, ERK-1 or ERK-2, that are known to be activated following overall PKC activation (36,37) in EC.
Serum stimulation of vector (control) EC that had been rendered quiescent demonstrated a rapid increase in ERK-2 activity by 10 min, with a gradual diminution of the kinase activity by 2-4 h after stimulation (Fig. 6, A and B). ERK-2 activity also increased within 10 min following serum stimulation of the PKC␣ and PKC␦ EC; however, the activity was enhanced above control and remained elevated above the basal kinase activity even at 4 h following stimulation. Phorbol ester treatment of the stably transfected cells resulted in similar levels of ERK-2

FIG. 2. Intracellular location of PKC␣ and ␦ in vector, PKC␣, and PKC␦ EC. In vector control EC (A, B, E, and F), locations of both PKC␣ (A and B)
and PKC␦ (E and F) are shown under quiescent (A and E) and PMA-stimulated (B and F) conditions. In PKC␣ EC (C and D), the location of PKC␣ under quiescent (C) and PMA-stimulated (D) conditions is shown. In G and H, the location of PKC␦ in PKC␦ EC is shown both before (G) and after (H) stimulation with PMA. Vector control EC were also stained for histone protein (I) to identify nuclei; with fluorescently labeled ceramide (J) to identify the Golgi apparatus; and with secondary antibody alone (anti-mouse and anti-rabbit antibodies; K and L) as negative controls. activity in PKC␣ or PKC␦ EC as compared with control, with the overall ERK-2 activity diminishing at a more rapid rate in control EC (i.e. 4 h) (Fig. 6, A and C). Similar analyses demonstrated very low overall ERK-1 activity in all cells tested. Serum stimulation resulted in mild enhanced ERK-1 activity; however, there were no noticeable differences between the con-trol, PKC␣, and PKC␦ EC (data not shown). Thus, PKC␦ and PKC␣ appeared to be equally effective in activating ERK-2␣, and thus the decrease in proliferation in PKC␦ EC appeared to result from a mechanism independent of ERK-1 or ERK-2 activity.
Effect of PKC␦ and PKC␣ on EC Cell Cycle Progression-To determine if an alteration in cell cycle progression might explain the decrease in cell growth in PKC␦ EC, cell cycle analysis was performed on serum-deprived and stimulated PKC␦, PKC␣, and control EC. After 72 h of serum starvation, 19.8 Ϯ 4.2% of control EC and 17.1 Ϯ 0.4% PKC␣ EC were in S phase, with 74.2 Ϯ 5.5% and 77.3 Ϯ 0.9% in G 0 /G 1 , respectively (Fig.  7A). In contrast, 26.3 Ϯ 2.2% of the PKC␦ cells were in S phase, with 66.5 Ϯ 1.9% in G 0 /G 1 . These data indicate that an abnormally high percentage of PKC␦ EC entered S phase inappropriately, i.e. under conditions of serum deprivation. Stimulation with serum caused control EC and PKC␣ EC to reenter the cell cycle normally (Fig. 7, B and C). PKC␦ EC, on the other hand, after an initial increase in the percentage of cells in G 2 /M phase at 6 h, followed by an increase in cells in G 0 /G 1 at 12 h after serum, returned to a very high percentage of cells in S phase up to 60 h (Fig. 7D). The prolonged time that a high percentage of PKC␦ EC could be found in S phase suggested that these cells required an abnormal amount of time to complete S phase.

DISCUSSION
The two major findings of this study are that overexpression of two different PKC isoenzymes normally expressed in microvascular endothelial cells exert distinct effects on endothelial proliferation, migration, and adhesion to extracellular matrix and that PKC␦-mediated inhibition of endothelial growth results from a defect in S phase of the endothelial cell cycle. The observation that overexpression of PKC␦, but not PKC␣, prevents proliferation of microvascular endothelial cells, while PKC␣ enhances their migration in response to HGF (scatter factor), suggests that these isoenzymes phosphorylate different substrates in these cells with different physiologic effects. The disparity between the effects of overexpression of the two isoenzymes is heightened by PKC-activating phorbol esters in that treatment of PKC␣ EC exerts a mitogenic effect similar to that in control cells, while PKC␦ EC were even more strongly inhibited by similar treatment. Thus, these data suggest that activation of each of these isoenzymes by angiogenic stimuli, such as HGF or those contained in serum, may mediate distinct aspects of several processes that are required for vascular repair and angiogenesis. Stimulation of endothelium with phorbol esters, which activate both PKC␣ and PKC␦, as well as several other isoenzymes expressed in endothelium (5,38), has been noted to have both stimulatory and inhibitory effects on endothelial proliferation and angiogenesis (5,10,39,40) that can be temporally dispersed within the same cells (41). The results from this study suggest that PKC␦ might mediate those aspects of PKC activation that are inhibitory for endothelial repair and angiogenesis, while an important component of the proangiogenic effect, endothelial migration, might be mediated by PKC␣.
Repair of small defects in endothelium are accomplished largely by endothelial migration, rather than proliferation (3,42). Migration is an important component of angiogenesis as well (3). The present study's finding that endothelial migration is enhanced in PKC␣ EC might result from enhancement of cytoskeletal reorganization in response to stimuli, which is a necessary component of cell locomotion; overall PKC stimulation has been associated with promotion of cytoskeletal reorganization of endothelial cells (3,43). Our results suggest that PKC␣, but not PKC␦, may be at least one mediator of migration response to HGF (scatter factor), a powerful angiogenic agent that is present along with its receptor in a substantial amount in the vasculature (4,32,44,45).
We considered several possible explanations for the PKC␦mediated inhibition of endothelial growth, including a reduction in endothelial adhesion to matrix, a failure to activate downstream mediators such as ERK-1 or ERK-2, and a defect in progression of endothelial cells through the cell cycle. Of these explanations, only the latter appears to be the case. Regarding adhesion, both endothelial cell growth and migration are thought to require attachment of the cell to matrix via its integrin surface receptors (42,46,47). Blocking the vitronectin integrin receptors ␣ v ␤ 3 or ␣ v ␤ 5 inhibits neovascularization in the cornea or chick chorioallantoic membrane models (35), suggesting the importance of these two integrins for endothelial cell proliferation. In this study, microvascular endothelial cells in which PKC␦ was overexpressed demonstrated markedly enhanced adhesion to vitronectin or fibrinogen matrices, and thus a decrease in integrin mediated adhesion was not the cause of the decrease in proliferation of PKC␦ EC. This enhancement of adhesion could result from either a PKC-mediated alteration of the number of these receptors or an increase in avidity by either a direct effect on integrin conformation (so-called "inside-out" signaling) (48) or as an amplification step following cell adhesion ("outside-in") that prevents detachment (48). Overall PKC activation has been linked with promotion of integrin avidity for soluble fibrinogen and solid matrices in several cell types (49,50). Neither PKC␣ EC nor PKC␦ EC demonstrated increased expression of these receptors when compared with control; thus, a PKC␦-mediated effect on affinity of these integrins for vitronectin is a likely explanation for our results.
A cascade of signaling events merging at the MAP kinase family of proteins, ERK-1 and ERK-2, is involved in many of the intracellular signaling pathways that lead to endothelial cell growth, migration, and adhesion (51). Activation of overall PKC leads to activation of MAP kinase; PKC␣, at least, has been shown to phosphorylate Raf kinase (40), an upstream mediator of MAP kinase (36). In our studies, both PKC␣ and PKC␦ enhance ERK-1 and ERK-2 kinase activity with a resultant prolongation of ERK-2 activity in stably transfected endothelial cells. The pathway by which PKC␦ blocks proliferation and cell cycle progression, however, must be distal or parallel to that leading to ERK-2 or ERK-1. In addition, our results suggest that effective activation of ERK-2 activity by PKC␦ is not sufficient for endothelial cell proliferation. This finding bolsters those of Hirai et al. (25), who found that PKC␦ inhibited growth of Ras-transformed NIH 3T3 cells despite activating AP-1, a component of the signaling pathway downstream from MAP kinase.
Inhibition of endothelial cell proliferation by PKC␦ appears to result from a specific defect in endothelial cell cycle progression, in which the cells enter S phase inappropriately and require additional time to complete this phase. Because of the distal nature of this defect and the intranuclear location of PKC␦, it seems likely that the isoenzyme interacts directly with one of the cyclin-cyclin-dependent kinase/inhibitor complexes that regulates entry and completion of S phase. Inappropriate S phase entry has been associated with apoptosis of vascular smooth muscle cells treated with basic fibroblast growth factor antisense oligonucleotides (52), but such apopto-FIG. 6. Analysis of ERK-2 activity in control (q), PKC␦ (Ç), and PKC␣ (Ⅺ) EC upon serum or PMA stimulation. All EC were rendered quiescent by serum deprivation and then stimulated with 15% serum or 1 M PMA. The cells were harvested, ERK-2 was immunoprecipitated, and kinase activity was assayed by the ability to phosphorylate myelin basic protein. The phosphorylated myelin basic protein was resolved on SDS-polyacrylamide gel electrophoresis, and a representative autoradiograph is shown in A. The audioradiographs were quantitated using a scanning densitometer, and the results are presented in B (serum-stimulated) and C (PMA-stimulated). sis was not found in PKC␦ EC. 2 Manipulation of PKC activity has not been previously associated with this S phase defect and only rarely with any abnormality of cell cycle function. Overexpression of PKC␦, but not PKC␣ or PKC, in Chinese hamster ovary fibroblasts causes an arrest in G 2 /M phase of the cell cycle, but only after activation with PMA (21); no defects were seen in S phase. In that study, inhibition of cell cycle progression by PKC␦ was attributed to an isoenzyme-specific effect; since the PKC enzymatic activity was much higher in the PKC␦-transfected cells than in the other transfectants, however, it is possible that the observed effect merely reflected an increase in overall enzyme activity. In this study, however, the similarities of enzymatic activity make it likely that the defect in cell cycle progression, as well as differences in adhesion and migration, resulted from interaction of PKC␦ with isoenzymespecific substrates. Thus, in addition to its isoenzyme-specific character, this arrest in S phase appears to be somewhat specific for endothelial cells.
Specific effects of individual PKC isoenzymes on proliferation are known to vary according to cell types. For example, overexpression of PKC␣ in murine fibroblast cells inhibits proliferation and does not lead to cell transformation (53,54), but in NIH 3T3 and human breast cancer cell lines, increases in PKC␣ expression altered the cell morphology, enhanced proliferation, and increased tumorigenicity (23,55). Thus, the tissuespecific effects of individual PKC isoenzymes are likely to be mediated by substrates or effectors with restricted tissue or subcellular expression. Even within an individual cell type, the substrates with which each isoenzyme interacts differ. A full understanding of the mechanism by which these two individual PKC isoenzymes mediate either enhanced adhesion or migration or decrease proliferation of endothelial cells will require identification of their selective downstream targets. Such identification, together with the assignment of selective endothelial functions to individual PKC isoenzymes provided by this study, would provide essential details critical to our understanding of reendothelialization and angiogenesis.