A Dominant-Negative p65 PAK Peptide Inhibits Angiogenesis
PAK1 is a protein kinase downstream of the small GTPases Rac and Cdc42 that previous work has implicated in endothelial cell migration via modulation of cell contraction. The first proline-rich region of PAK that binds to an SH3 domain from the adapter protein NCK was responsible for these dominant-negative effects. To test the role of PAK in angiogenesis, we prepared a peptide in which the proline-rich region was fused to the polybasic sequence from the HIV Tat protein to facilitate entry into cells. We show that the short peptide selectively binds NCK, whereas a mutant peptide does not. Treatment of cells with the PAK peptide but not the control peptide disrupts localization of PAK. This peptide specifically inhibited endothelial cell migration and contractility similarly to full-length dominant-negative PAK. In an in vitro tube-forming assay, the PAK peptide specifically blocked formation of multicellular networks. In an in vivo chick chorioallantoic membrane assay, the PAK peptide specifically blocked angiogenesis. These results, therefore, suggest a role for PAK in angiogenesis.
Angiogenesis is a fundamental process in development and wound healing, as well as a component of pathologies such as diabetic retinopathy and cancer.1,2⇓ It involves movement of endothelial cells out of existing vessels, migration into new areas, proliferation, and formation of new capillaries. Migration of endothelial cells in response to angiogenic growth factors such as basic fibroblast growth factor (bFGF) and vascular endothelial growth factor (VEGF) plays an important role in this process.
Previous work identified p65 PAK1 as a potential modulator of endothelial cell migration.3 PAK is a 65-kDa serine/threonine kinase that serves as a direct effector of Rac and Cdc42.4 PAK’s downstream functions are complex, but it is clearly implicated in regulation of the actin cytoskeleton and cell migration. We reported that expression of dominant-negative PAK strongly inhibited growth factor-induced endothelial cell movement.3 Inhibition appeared to be mediated by a decrease in contractility, and we suggested that PAK played a critical role in myosin-dependent contraction and detachment of the rear of the cell. This study also mapped the inhibitory effect of dominant-negative PAK to a single proline-rich sequence near the N-terminus. This sequence is known to bind the SH3 domain from the adapter protein Nck, and this interaction has been proposed to target PAK to the cell membrane.5
The goal of the present study was to test whether the dominant-negative sequence that inhibited endothelial cell migration could also block angiogenesis. Toward this goal, we developed a synthetic peptide containing the PAK proline-rich sequence fused to the polybasic sequence from the HIV tat protein.6 This polybasic sequence has been found to mediate transport of peptides and proteins through the cell membrane into the cytoplasm and would therefore enable us to analyze effects of PAK inhibition without microinjection or transfection. We report here the verification of this reagent and its effects on in vitro and in vivo models of angiogenesis. The data identify PAK as an important component in these systems and a potential target for angiogenesis inhibitors.
Materials and Methods
Human microvascular endothelial cells (HMEC-1; Clonetics, San Diego, Calif)7 were maintained in endothelial growth medium (EGM; Clonetics) supplemented with an additional 8% fetal bovine serum (FBS; Gemini, Irvine, Calif) for a final concentration of 10%. For some experiments, cells were transferred to endothelial basal medium (EBM) supplemented with 0.2% FBS (Clonetics). Human umbilical vein endothelial cells (HUVECs) were purchased from Clonetics (San Diego, Calif) and maintained in EGM with a final concentration of 10% FBS. Cells were maintained at 37°C in a humidified incubator containing 7% CO2.
Type I rat tail collagen was purchased from Upstate Biotechnology (Lake Success, NY). Recombinant bFGF was purchased from Collaborative Biomedical Products (Becton Dickinson Labware). Recombinant VEGF was from Genentech. 10× medium 199 (M199) and phorbol ester (PMA) were from Sigma Chemical Co. ITS (insulin, transferrin, and selenium-A) and trypsin were from Gibco-BRL. The PAK and control peptides were synthesized by the Scripps peptide synthesis facility.
PAK or control peptides containing an N-terminal cysteine were dissolved in coupling buffer (50 mmol/L tris pH 8.0, 5 mmol/L EDTA) and 0.5 mL at 3 mg/mL peptide added to 2 mL of activated thiol-Sepharose (Pharmacia). Samples were incubated at 4°C overnight and then 1% β-mercaptoethanol added to block unreacted sites. Beads were rinsed and stored at 4°C. Endothelial cell cultures in 15-cm dishes were extracted in 1.0 mL of ice cold lysis buffer (20 mmol/L pH 7.4 sodium phosphate, 100 mmol/L NaCl, 1% Triton X-100, 10 mmol/L NaF, 2 mmol/L EDTA, 1 mmol/L sodium orthovanadate, 1 mmol/L PMSF, and 10 μg/mL each of leupeptin and aprotinin). Extracts were centrifuged for 10 minutes at top speed in a microcentrifuge and the supernatants removed. Clarified extract (0.5 mL) was incubated with 10 μL of packed beads for 1 hour rotating at 4°C. Beads were rinsed 4×0.5 mL lysis buffer and dissolved in SDS sample buffer. Samples were analyzed by Western blotting on 10% gels and probed with anti-Nck (Santa Cruz), anti-Src, anti-Abl, or anti-Crk (Santa Cruz). The transfers were rinsed and probed with HRP-protein A as secondary, then developed with ECL (Amersham).
To assay migration, HMEC-1 cells were treated with various concentrations of either PAK or mutant peptide before plating. Cells were plated on coverslips coated with 2 μg/mL of fibronectin (FN). Dishes containing coverslips were prepared as described3 and maintained at 37°C under CO2 as described.8 Cells were viewed with a Nikon DiaPhot Microscope equipped with a SenSys-cooled CCD video camera linked to a Silicon Graphics workstation running the Inovision ISEE software program. At the end of the experiment, images of cells were outlined and the centroid (cell center) calculated. Displacement of the centroid was then used to determine movement over time. Statistical analysis was performed using Microsoft Excel statistics software. Approximately 50 cells were filmed and tracked per experiment, with at least 3 independent experiments for both the PAK peptide and the mutant peptide. Results were expressed as mean±standard error of the mean. The Student’s t test was used to determine whether the observed differences were statistically significant (P<0.05).
To detect entry of peptides into cells, cells exposed to the FITC-tagged peptide were fixed for 10 minutes in 2% formaldehyde (Polysciences) in phosphate buffer saline (PBS), then washed twice in PBS, and mounted in Immunofluore mounting medium (ICN Immunobiologicals). Slides were viewed using a Bio-Rad 1024 MRC Scanning Confocal Microscope.
To visualize cell contractility, flexible rubber substrates were generated as described.9 Briefly, silicone (dimethyl polysiloxane), viscosity 10 000 centistokes (Dow Corning Co), was applied to the coverslip at the bottom of 35-mm tissue culture dishes. The silicone was then coated with a thin layer of gold-palladium using a Hummer VI sputter coater, which polymerized a thin layer of silicon rubber. The surface was additionally coated with 2 μg/mL of fibronectin (FN). Endothelial cells were plated on the rubber substrate and allowed to adhere. Cells that had wrinkled the substrate beneath were then visualized for 4 hours, then scored by time-lapse phase contrast microscopy. After this period, they were treated with cytochalasin D (1 μg/mL) to release all wrinkles that were due to actin-dependent cell tension.
HMECs were plated on FN-coated coverslips in the presence or absence of 20 μg/mL peptides. After 30 minutes, cells were injected into the nuclei with cDNA coding for human HA-tagged wild-type PAK1 at 50 μg/mL together with cDNA for GFP at the same concentration.3 After an additional 1.5 hours, cells were fixed in 2% formaldehyde, permeabilized with 0.2% Triton X-100, and stained using anti-HA monoclonal antibody (BABCO) followed by Cy5 goat anti-mouse secondary (Jackson Immunoreagents). Images were taken using a Bio-Rad 1024 confocal laser scanning microscope.
Formation of 3-Dimensional Collagen Gels
Collagen gels were formed by mixing ice-cold gelation solution (10× M199, H2O, 0.53 mol/L NaHCO3, 200 mmol/L l-glutamine, type I collagen, 0.1 mol/L NaOH; 100:27.7:50:10:750:62.5 by volume) with cells in 1× basal medium at a concentration of 3×106 cells/mL at a ratio of 4 vol of gelation solution:1 vol of cells.10 The gels were allowed to form by incubation in a CO2-free incubator at 37°C for 30 to 60 minutes. The gels were overlaid with 1× basal medium consisting of M199 supplemented with 1% FBS, 1× ITS, 2 mmol/L l-glutamine, 50 μg/mL ascorbic acid, 26.5 mmol/L NaHCO3, 100 U/mL penicillin, 100 U/mL streptomycin, 40 ng/mL bFGF, 40 ng/mL VEGF, and 80 nmol/L PMA. At the end of the experiment, the medium aspirated and the gels fixed in 3.7% formaldehyde. Photographs were taken on a Nikon Eclipse TE300 inverted phase microscope using Hoffmann modulation contrast optics and a Polaroid digital microscope camera linked to a Macintosh G4 using the Improvision software package running OpenLab version 2.2.5. In some experiments, fixed cells were stained 30 minutes with 0.04% crystal violet in 2% methanol and imaged a Nikon DiaPhot microscope equipped with a SenSys-cooled CCD video camera linked to a Silicon Graphics workstation running the Inovision ISEE software program. Assays were performed using 48-well plates with n=3 wells for each experimental manipulation, and all experiments were repeated at least 3 times.
Quantification of Tube Formation
Gels were examined on a Nikon TE300 microscope equipped with Hoffman modulation optics and a cooled CCD camera (Optronics, Technical Instruments).10 Four fields were imaged using OpenLab, and the data were imported as TIFF files into Inovisions ISEE software program. The total length of each tube or the long axis of single cells or groups of adjacent cells was measured and raw data were imported into Microsoft Excel where the mean length was determined for each well, followed by calculation of the mean±SEM for each experimental group.
Chick Chorioallantoic Membrane Assays
Angiogenesis assays were performed essentially as previously described.11 Filter discs saturated with 1.0 μg/mL bFGF were placed on the chorioallantoic membranes (CAMs) of 10-day-old chick embryos followed immediately by daily topical addition of either PAK inhibitory peptide (20 μg) or control peptide (20 μg) in 20 μL of FBM. After 72 hours, filter discs and associated CAM tissue were harvested and quantified. Angiogenesis was assessed as the number of visible blood vessel branch points within the defined area of the filter discs. At least 20 CAMs were used for each treatment.
Previous work indicated that overexpression of residues 1 to 74 from the N-terminus of PAK1 in endothelial cells inhibited migration and contractility, whereas mutation of 2 key prolines eliminated this effect. We therefore synthesized a peptide in which 13 residues from the first proline-rich domain of PAK1 were fused to the HIV Tat protein polybasic sequence to facilitate entry into cells6 (Figure 1A). We also prepared a control peptide in which 2 prolines critical for SH3 binding were mutated.3 Two additional sets of peptides were synthesized, one set with an N-terminal fluorescein group to allow visualization and another lacking the tat sequence but with a C-terminal cysteine for coupling to agarose.
To test binding specificity, detergent extracts of human endothelial cells were incubated with peptide-agarose, followed by washing and analysis by Western blotting with anti-Nck. As controls, we also probed for 3 other SH3-containing proteins. We observed that PAK peptide bound Nck from the lysates, whereas no detectable Nck bound to the control peptide (Figure 1B). Little or no binding to c-src, c-Abl, or Crk were detected. This result confirms the ability of this proline-rich sequence to specifically bind Nck, presumably through its SH3 domain.
To test the ability of Tat fusions to enter cells, the fluorescent derivatives were incubated with endothelial cell monolayers, washed, and fixed. We observed some heterogeneity in the intensity of labeling, but uptake increased steadily between 0 and 30 minutes, so that by 30 minutes, essentially all of the cells showed strong cytoplasmic labeling with the fluoresceinated peptides. Some labeling of the matrix and nuclei was also noted (Figure 2). These results suggest that the Tat-PAK peptides may be useful for testing function in live cells.
Migration and Contractility
To test whether these peptides influenced endothelial cell behavior similar to overexpression of the longer PAK constructs, endothelial cells on coverslips were incubated with the peptides and time-lapse imaging carried out. Analysis of these images showed that endothelial cells in the presence of 10 μg/mL PAK peptide migrated at a substantially slower rate, whereas the control peptide had no significant effect (Figure 3A). Endothelial cell migration rates were essentially constant over the time course of the experiment, which argues against toxicity or other deleterious effects (Figure 3B). Analysis of different concentrations showed that the half-maximal inhibition was observed at approximately 5 μg/mL (2 μmol/L), and was nearly maximal at 20 μg/mL (Figure 3C). Transwell assays in which cells migrated toward bFGF also showed substantial inhibition by the PAK but not control peptide (not shown). To test effects on cell survival and proliferation, cells were plated and maintained in the presence of PAK or control peptides at 20 μg/ml. Initially, 371±28 cells were present per low-power field. After 48 hours, cell number in untreated wells increased to 755±89 cells per field compared with 856±66 for the PAK peptide and 898±91 for the control peptide. These differences are not statistically significant. Additionally, cells appeared entirely viable (not shown). These results are inconsistent with inhibitory effects on either survival or cell proliferation.
To examine effects on contractility, cells were plated on elastic substrata that can be pulled into wrinkles by cell-generated forces.9,12⇓ Cells plated on these surfaces initially induced wrinkles, indicating the application of contractile forces to the substratum. Addition of PAK peptide at 10 μg/mL caused release of wrinkles, indicating a decrease in cell-generated tension (Figure 4A), whereas the control peptide had no effect. Quantitation of these results indicated a consistent and significant inhibition of contractility (Figure 4B). Taken together, these data show that soluble Tat-PAK peptides exert effects on cells that are essentially identical to the previously examined dominant-negative PAK expression constructs.
To gain some insight into the mechanism of inhibition, we measured PAK kinase activity but did not detect changes in PAK activity when peptides were added. We therefore examined PAK localization. Because of the lack of suitable reagents for localization of endogenous PAK, cells were transfected with wild-type HA-tagged PAK by microinjecting the expression vector DNA into nuclei. PAK peptide was added and after 1.5 hours cells were fixed and stained for the epitope tag. In untreated or control peptide–treated cells, PAK strongly localized to lamellipodial edges and in some examples to filaments within extensions (Figure 5). Treatment with the PAK peptide completely disrupted this pattern, resulting in a diffuse distribution throughout the cytoplasm. These results suggest that the peptide may not inhibit PAK activity but may work by causing loss of localization to downstream targets.
In Vitro Tube Formation
As an initial test of the peptide’s effect on endothelial cell behavior in 3-dimensional matrices, in vitro tube formation assays in matrigel were carried out. HUVECs in matrigel in the presence of bFGF, VEGF, and PMA aligned and formed endothelial tubes with lumens over approximately 2 days. Assays were set up in the presence of several concentrations of PAK or control peptides and examined at 48 hours. These were quantified by determining the length of endothelial cell tubes. As shown in Figure 6, the PAK peptide significantly inhibited tube formation in this system. Morphology of individual cells was not noticeably perturbed, nor was cell number greatly altered, ruling out toxicity or general disruption of the cytoskeleton under these conditions. These data suggest that PAK-regulated contractility and migration are required for formation of multicellular structures in a 3-dimesional matrix.
In Vivo Angiogenesis
To determine whether results obtained in vitro are relevant to in vivo angiogenesis, we employed a chick CAM model in which filters containing bFGF are placed on the chorioallantoic membrane of developing chick embryos.11 New blood vessels invade within 2 to 5 days, and angiogenesis can be readily detected. Filters without any additions, with control, or PAK peptides were implanted and examined 3 days later. Figure 7 shows that the PAK peptide substantially inhibited angiogenesis compared with untreated filters, and that the control peptide had essentially no effect. These results show that the dominant-negative PAK construct can block bFGF-induced angiogenesis in vivo.
Our previous work showed that expression of dominant-negative PAK constructs in human endothelial cells inhibited contractility and cell migration.3 Importantly, these effects occurred despite an increase in actin stress fibers, but correlated with a modest decrease in myosin light chain phosphorylation. The present study was designed to extend these results to 3-dimensional systems where endothelial cells form complex structures. We took advantage of results showing that the dominant-negative effect could be attributed to a single proline-rich sequence at the N-terminus of PAK. A synthetic peptide covering this region showed specific binding to Nck, suggesting that the short (13 aa) sequence should be sufficient. Coupling the PAK sequence to a region from the HIV tat protein promoted its rapid entry into endothelial cells, where it exerted dominant-negative effects that were essentially indistinguishable from those of overexpressed full-length dominant-negative PAK. The peptide showed no deleterious effects on endothelial cell number, suggesting that it did not influence survival or growth. This peptide therefore allowed us to analyze effects of dominant-negative PAK in more complex systems that are not readily amenable to transfection.
Analysis of endothelial cells in vitro in a 3-dimensional ECM showed that the PAK peptide markedly blocked formation of extended multicellular structures without apparent deleterious effects on single cell morphology or number. Previous studies showed that these structures contain lumens and therefore resemble capillary tubes in vivo.10 The effect of the PAK peptide is therefore consistent with inhibition of cell migration. Analysis of angiogenesis in a chick CAM model also revealed significant inhibition by the PAK peptide. Taken together, these results strongly suggest that dominant-negative PAK inhibits endothelial cell migration and angiogenesis, most likely via effects on the cytoskeleton. However, it is likely that the peptide also enters accessory cells such as fibroblasts and pericytes; effects on these cells may also contribute to inhibition of angiogenesis in vivo.
Angiogenesis is a complex process that involves proteolysis of existing ECM, detachment and migration of endothelial cells out of mature blood vessels, and proliferation and assembly of cells into new blood vessels.1,2⇓ Contraction has been proposed to be a necessary element of individual cell migration because it promotes both forward movement of the cell body and detachment of the trailing edge.13 Additionally, contraction has been proposed to create tension within 3-dimensional matrices that serves to guide migration of endothelial cells toward one another to facilitate tube formation.14 Thus, inhibition could involve not only direct inhibition of cell movement but also perhaps interference with guidance mechanisms for establishing interacting cell-cell networks.
Whatever the mechanism, these data identify PAK as a potential target for angiogenesis inhibitors. In addition to proline-rich sequences that are critical for PAK targeting during its activation and/or effector functions, PAK also contains motifs involved in autoinhibition of kinase activity, as well as the kinase domain itself.15,16⇓ All of these sites represent potential interaction surfaces for small molecule inhibitors.
This work was supported by grant No. AHA120069Y to W.B.K. and United States Public Health Service grant No. P01HL57900 to M.A.S. and Nos. CA50286, R01 CA45726, and PO1 CA78045 to D.A.C.
Original received July 3, 2001; resubmission received January 11, 2002; revised resubmission received February 14, 2002; accepted February 15, 2002.
- ↵Kiosses WB, Daniels RH, Otey C, Bokoch GM, Schwartz MA. A role for p21-activated kinase in endothelial cell migration. J Cell Biol. 1999; 147: 831–843.
- ↵Schwarze SR, Ho A, Vocero-Akbani A, Dowdy SF. In vivo protein transduction: delivery of a biologically active protein into the mouse. Science. 1999; 285: 1569–1572.
- ↵Schwartz MA. Spreading of human endothelial cells on fibronectin or vitronectin triggers elevation of intracellular free calcium. J Cell Biol. 1993; 120: 1003–1010.
- ↵Bussolino F, Wang JM, Turrini F, Alessi D, Ghigo D, Costamagna C, Pescarmona G, Mantovani A, Bosia A. Stimulation of the Na/H exchanger in human endothelial cells activated by granulocyte and granulocyte-macrophage colony stimulating factor. J Biol Chem. 1989; 264: 18284–18287.
- ↵Harris AK, Wild P, Stopak D. Silicone rubber substrata: a new wrinkle in the study of cell locomotion. Science. 1980; 208: 177–179.
- ↵Schwarzbauer JE. Cell migration: may the force be with you. Curr Biol. 1997; 7: 292–294.
- ↵Davis GE, Camarillo CW. Regulation of endothelial cell morphogenesis by integrins, mechanical forces and matrix guidance pathways. Exp Cell Res. 1995; 216: 1113–1123.
- ↵Zhao ZS, Manser E, Chen XQ, Chong C, Leung T, Lim L. A conserved negative regulatory region in αPAK: inhibition of PAK kinases reveals their morphological roles downstream of Cdc42 and Rac1. Mol Cell Biol. 1998; 18: 2153–2163.
- ↵Zenke FT, King CC, Bohl BP, Bokoch GM. Identification of a central phosphorylation site in p21-activated kinase regulating autoinhibition and kinase activity. J Biol Chem. 1999; 274: 32565–32573.