Apoptotic Cells Initiate Endothelial Cell Sprouting via Electrostatic Signaling

Abstract

Angiogenesis, the development of new blood vessels from preexisting vessels, is crucial to tissue growth, repair, and maintenance. This process begins with the formation of endothelial cell sprouts followed by the proliferation and migration of neighboring endothelial cells along the pre-formed extensions. The initiating event and mechanism of sprouting is not known. We demonstrate that the phenotypic expression of negative-charged membrane surface in apoptotic cells initiates the formation of directional endothelial cell sprouts that extend toward the dying cells by a mechanism that involves endothelial cell membrane hyperpolarization and cytoskeleton reorganization but is independent of diffusible molecules.

INTRODUCTION

Vascular expansion is a multistep process that includes activation of pre-existing endothelial cells (EC), formation of sprouts, migration of EC along the sprouts, the formation of tubular structures and, finally, vessels with a distinct blood-transporting lumen (1, 2). Although still unclear, the initiation of angiogenesis has been attributed to changes in the net balance between proangiogenic and antiangiogenic factors in the microenvironment (3, 4) that can be induced by severing the blood supply, by wounding, or by rapid tissue expansion typical of tumor growth. Since oxygen can only diffuse ∼120 μm from capillaries (5), injury and hypoxia invariably lead to apoptosis (6). Because cell death is an integral component of angiogenesis (7, 8), it is possible that apoptotic cells play a fundamental role in signaling and/or initiating angiogenic EC responses.

To determine the potential role of apoptotic cells in vascular expansion, we co-cultured EC with autologous apoptotic EC or xenogeneic apoptotic tumor cells. Surprisingly, nonproliferating EC extended sprouts exclusively toward apoptotic cells. This process was initiated by the negative charge of apoptotic cell surface and involved EC membrane hyperpolarization and cytoskeleton reorganization.

MATERIAL AND METHODS

Cell Lines and Reagents. Rat kidney EC were a gift from S. Adler (New York Medical College) and R. Johnson (University of Florida) (9). The K-1735 mouse melanoma cells were obtained from M. L. Kripke (The University of Texas M. D. Anderson Cancer Center) (10). Stable GFP-expressing K-1735 cell lines were generated as described previously (11). Cationized and anionized ferritin, rabbit anti-ferritin IgG, and cytochrome-c (cytc) from horse heart were purchased from Sigma (St. Louis, MO). Rabbit anti-caspase 3 IgG was purchased from Cell Signaling Technology (Beverly, MA). Monoclonal PCNA antibody was from DAKO (Glostrup, Copenhagen, Denmark). Polyclonal antibodies to EGFR, VEGFR-2, and phosphorylated EGFR were from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal antiphosphorylated VEGFR-2 was from Oncogene (Cambridge, MA). The TUNEL assay kit was obtained from Promega (Madison, WI). Alexa Fluor 594 labeled anti-rabbit IgG, CellTracker Red CMTPX, and membrane potential sensitive fluorescent dye, di-8-ANEPPS, were purchased from Molecular Probes (Eugene, OR). Annexin V apoptosis detection kit was from BD (San Diego, CA), and protein A conjugated Sepharose beads were purchased from Amersham Biosciences (Uppsala, Sweden). AEE788 was obtained from Novartis Pharma (Basel, Switzerland). Compartmental culture dishes with removable separating walls were made of medical-grade silicon Elastomer (Maxzon Scientific, Houston, TX). Ion channel blockers, amiloride for sodium channel, protopine for calcium channels, and glyburide and charybdotoxin for potassium channels, were purchased from EMD Biosciences (San Diego, CA).

Co-cultures of Viable and Apoptotic Endothelial Cells. EC were grown to confluency in 60-mm culture dishes in Dulbecco's minimal essential medium (DMEM) with 10% fetal bovine serum (FBS) in a 5% CO₂ incubator at 37°C. Confluent EC cultures were overlaid for 2 min at 37°C with 0.25% trypsin in a 0.1% EDTA solution. Floating cells were washed off with serum-free DMEM. The remaining EC were subsequently cultured in serum-free medium. Morphological changes to the cells in the monolayer margin adjacent to the denuded areas were monitored every 30 min by microscopy using an inverted light microscope. Once EC sprouts developed, the cultures were washed, and the cells were incubated with annexin V or fixed with paraformaldehyde (15 min at 20°C) and subjected to TUNEL assay and PCNA staining following the manufacturer's protocols.

Microinjection of Cytc into EC. Confluent EC were partially trypsinized as described above. Detached cells were removed by washing. The remaining cells were then cultured for 5 h in serum containing DMEM to allow adhesion of any loosely detached cells. The border of a denuded area was marked on the bottom surface of the culture dish and photographed. Microinjection needles with an inner diameter of ∼0.1 μm were pulled from glass capillaries with a horizontal electrode puller (Brown Micropipette Puller; Sutter Instrument Company, Novato, CA). Cytc at 10 mg/ml in water containing 0.5 μM CellTracker Red CMTPX (a concentration that does not induce cell death when injected alone) was loaded into the microinjection needles. Cytc was injected into the cytosol at 80-100 hPa within 0.3 s on the stage of an inverted microscope with a Narishige pressure injector (Model IM-5B' Narishige) and micromanipulator (Leitz) as described previously (12). Injection into the cytosol was verified by fluorescence microscopy. Bright-field images were recorded very 5 min.

Cell Culture Manipulation and Treatment with the VEGFR/EGFR Inhibitor, AEE788. Microinjection needles were used to produce scratches on the inner bottom surface of the culture dish between a sprouting EC and attracting apoptotic cell. Briefly, an area of interest was located by microscopy and several scratches were then produced on the plastic surface. The ability of EC to produce directional sprouts was then monitored. Microinjection needles were also used to detach partially attached apoptotic cells that were aspirated with a pipette. In separate studies, EC were subjected to partial trypsinization and serum starvation while being continuously mixed in an orbital shaker to preclude the formation of potential chemotactic gradients. The ability of these cultures to sprout was then monitored. For AEE788 treatment, partially trypsinized EC were cultured in serum-free medium with AEE788 (0.1 μM) for 12 h in 60-mm culture dishes (n=10). Sprouts were then counted and compared to controls incubated with vehicle alone. Apoptotic cells were identified by TUNEL. Student's t test was used to compare between groups. P<0.01 were considered statistically significant.

Characterization of Cell Surface Charge with Cationized and Anionized Ferritin. The surface charge of apoptotic and sprouting EC was determined by the ability of the cells to bind cationized (electron negative cells) or anionized (electron positive cells) ferritin. The ability of ferritin to inhibit sprouting was determined by incubating the cultures in the presence of ferritin (100 μg/mL). The relative charge of sprouting and apoptotic cells was determined by adding cationized or anionized ferritin subsequent to sprout formation. In this case, sprouting cultures were fixed with 4% paraformaldehyde (15 min at 20°C) followed by three 10-min washes with phosphate-buffered saline (PBS). The cells were then incubated overnight at 4°C with PBS containing 100 μg/ml cationized or anionized ferritin. The cultures were washed and then incubated in blocking buffer (3% bovine serum albumin in PBS for 1 h at 20°C). Binding was determined immunohistochemically with rabbit anti-ferritin (1/100 dilution for 2 h at 20°C) followed with fluorescein-conjugated anti-rabbit IgG under the same conditions.

Initiation of Sprouting with Negatively Charged Sepharose Beads and Phospholipid Vesicles. Charged Sepharose beads were produced by antibody-mediated coupling of anionized or cationized ferritin. Briefly, protein A Sepharose (30 mg) was generated by incubating the beads with ferritin antibody (15 ng in 1 mL PBS) overnight. IgG not bound was removed by washing 5X with PBS. The antibody-conjugated beads were then incubated with ferritin (1 mg at 4°C for 3 h). Excess ferritin was removed by washing. The beads were then air-dried. Negatively charged multilamellar vesicles were produced by drying 1 mg of phosphatidylserine (PS) with 1 mg of phosphatidylcholine (PC) (Avanti Polar Lipids, Alabaster, AL) under nitrogen. The dried lipids were then hydrated by vortexing in PBS containing 25% sucrose (w/v). Small vesicles were removed by sedimentation at 1 g for 1 h. Control neutral vesicles were composed of PC exclusively. To determine whether negatively-charged beads or vesicles could substitute for apoptotic cell-dependent EC sprouting, anionized or cationized beads and neutral phospholipid vesicles and negatively-charged vesicles were carefully added to the medium of denuded EC cultures. The denuded areas containing a single particle were selected for continuous monitoring.

Co-culture of EC and Apoptotic K-1735 Melanoma Cells in a Compartmental Dish. A culture dish was separated into two compartments by a wall of silicon Elastomer that was 5-mm tall and 1.5-mm wide. EC (1 × 10⁶) were plated into one compartment, and GFP-expressing K-1735 cells (1 × 10⁵) were plated into the other. The cultures were then incubated with DMEM containing 10% FBS. Ten hours later, the compartment with K-1735 cells was washed three times with serum-free medium and refed with serum-free DMEM. Apoptosis of K-1735 cells were induced by incubating cells for an additional 48 h. The silicon barrier was then removed and the serum-free DMEM was replaced with DMEM containing 10% FBS. The morphological changes in the EC at the boundary were monitored by light and fluorescence microscopy. Once EC sprouts were visible, the cells were fixed for 5 min with 4% paraformaldehyde and stained for caspase 3 using the manufacturers' protocol.

Dual Wavelength Imaging Membrane Potential Changes. Di-8-ANEPPS (final concentration at 1 μM) was added into the medium of denuded EC cultures and incubated for 20 min in a cell culture incubator. After washing with medium, the cells were monitored with an inverted fluorescent microscope equipped with 450 nm (green emission) and 510 nm (red emission) excitation filters. Two-second exposures at each excitation wavelength were recorded with a CCD camera. Images from 450 nm and 510 nm were merged. The morphologies at corresponding time points were also taken with normal light imaging.

Treatment of ECs with Ion Channel Blockers. Prior to the addition of PS vesicles, denuded EC cultures were incubated with media containing different ion channel blockers, amiloride (10 μM), protopine (10 μM), glyburide (10 μM), and charybdotoxin (50 nM) for 20 min. Images of cells from randomly selected areas (5 for each sample) were taken before and after the addition of the vesicles. The morphological changes were monitored continuously for 40 min.

Scanning Electron Microscopy. The surface morphology of normal, sprouting, and apoptotic EC was visualized by scanning electron microscopy (Model S520; Hitachi Denshi).

RESULTS

Non-proliferating EC Sprout Toward Apoptotic Cells. Culturing mildly trypsinized confluent rat glomerular EC monolayers resulted in the generation of sprouts that extended from a small number of cells. These sprouts extended from EC at the margin of the monolayer to rounded and partially detached residual cells that remained in the denuded areas (Fig. 1A and 1B). To determine whether these ″attracting″ cells were indeed apoptotic, the cells were assessed for exposure of PS and DNA fragmentation by their ability to bind annexin V (Fig. 1C-1F) and TUNEL (TdT-mediated dUTP nick end labeling) staining (Fig. 1G and 1H), respectively. Fluorescence microscopy showed that the attracting cells were apoptotic. Further examination of the cultures by scanning electron microscopy revealed that, in contrast to the rough surface of the non-sprouting EC, attracting cells had a relatively smooth surface (Fig. 1I). To test whether the correlation between sprouts and apoptotic cells is significant, we counted the number of sprouting cells and the number of sprouting cells coupled with apoptotic cells in 5 randomly selected areas. The average number of sprouting cells was 37, and all 37 had sprouts pointing to apoptotic cells. We did not find any endothelials with sprouts that did not point toward apoptotic cells. The distance between the attracting and sprouting cells was typically <200 μm.

Fig. 1.

Apoptotic cells initiate EC sprouting. (A, B) EC were cultured to 100% confluency and then partially trypsinized to produce denuded areas. The inner periphery contained partially detached cells, where EC sprouted (arrows) toward cells with typical apoptotic morphology (arrowheads). (C-F) The sprouts (arrow) extending toward the attracting cell (arrowhead) that was positive for annexin V binding (white arrows crossing images, green color). (G, H) TUNEL staining revealed that some attracting cells (arrowhead) are TUNEL-positive (green color). (I) Scanning electronic microscopy shows that the surfaces of the attracting cells (arrowheads) are smoother than the relatively rough surfaces of the neighbor cells (double-head arrow). Bar = 50 μM.

To unequivocally determine whether apoptotic cells were indeed the initiating stimulus for sprout formation, single cells along the monolayer margin were triggered into apoptosis by microinjection of cytc into the cytosol. Figure 2A and 2B show that within 20 min after injection of cytc (red fluorescence), a directional sprout pointing towards the cytc-positive cell began to form from an opposing EC. Surprisingly, apoptotic cells were also required for the maintenance of existing sprouts. This can be seen from results showing that removal of the sprout-initiating apoptotic cell caused the sprout to retract back into the main cell body of the sprouting EC (Fig. 2C and 2D).

Fig. 2.

Induction and removal of apoptotic attracting cells. (A) Cytc microinjection into viable EC (arrowhead) together with red fluorescent dye (arrowhead in the inset). (B) Twenty min after microinjection of cytc, an EC formed a sprout toward the injected cell (arrow). (C, D) Removal of the attracting apoptotic cell (arrowhead) resulted in retraction of the extending sprout within 15 min (arrows). (E) Confluent EC were co-cultured with apoptotic GFP-labeled K-1735 melanoma cells. An EC sprout (insert, enlarged view of the sprout) directed toward the apoptotic K-1735 melanoma cell (arrowhead, dashed line). (F) The apoptotic status of the K-1735 cell (yellow color) was confirmed by caspase 3 staining with Texas Red conjugated secondary antibody and merged with the green GPF image of K-1735 cells. Bar = 50 μM.

The ability of xenogeneic apoptotic tumor cells to induce sprouting was also determined. EC and GFP-expressing K-1735 melanoma cells were co-cultured in separate compartments of a compartmentalized tissue culture dish. Apoptosis was induced in the melanoma side of the plate by serum starvation for 48 h. No EC sprouting was observed as long as the two compartments were separate. Removal of the barrier resulted in the formation of EC sprouts within 12 h. Importantly, while some intermixing between cell populations was unavoidable, only apoptotic (caspase-positive) melanoma cells attracted sprouts (Fig. 2E and 2F). We also cultured epithelial origin cells, MCF7 (breast cancer cells), and Du145 (prostate cancer cells) in the presence of apoptotic cells. We did not observe formation of sprouts.

Initiation of Sprouting by Electrostatic Signaling. Several distinct mechanisms could be responsible for apoptosis-dependent EC sprouting. These include a cell-expansion proliferation-dependent mechanism, activation of VEGF and/or EGF receptors known to participate in angiogenesis, or specific chemotactic factors released by dying cells (13). To determine whether sprouting was dependent on EC proliferation, mixed viable/apoptotic EC cultures were stained with proliferative cell nuclear antigen (PCNA) antibody. Figure 3A shows that sprout-producing EC were non-proliferating (PCNA-negative) and closest to, but not adjacent to, the attracting cells.

Fig. 3.

Non-proliferating EC sprout independent VEGF and EGF pathways. (A) PCNA staining (arrowheads, brown color) of control endothelial cells with sprouts showing that the cells that sprout are PCNA-negative (arrow, blue color stained with Hematoxylin) and nearest to, but not adjacent to, the attracting cells. (B) Western blot analysis showing inhibition of VEGFR (pVEGFR) and EGFR (pEGFR) phosphorylation with AEE788. Total VEGFR (tVEGFR) and EGFR (tVEGFR) served as controls. (C) PCNA staining of endothelial cells treated with AEE788 showing more sprouts (arrowheads) and less proliferating cells as compared with control (A). (D) Statistical analysis of the number of sprouts in control and AEE788-treated EC (P<0.001). (E) Statistical analysis of the number of proliferating cells and apoptotic cells (identified with TUNEL assay) in control and AEE788-treated EC (P<0.001). (F) Co-localization of sprouts (arrows) with TUNEL-positive cells (arrowheads, green color) in AEE788-treated EC. Bar = 50 μM.

To determine whether VEGF or EGF regulate sprouting, trypsinized EC monolayer was incubated in the presence of AEE788, a potent competitive inhibitor of both EGFR and VEGFR phosphorylation (14). Figure 3 shows that while both EGFR and VEGFR phosphorylation were effectively inhibited (Fig. 3B), sprout formation was not inhibited (Fig. 3C). Surprisingly, AEE788 decreased the fraction of PCNA-positive proliferating EC (Fig. 3C), and increased the density of sprouts (Fig. 3D) and the fraction of TUNEL-positive apoptotic cells (Fig. 3E and 3F).

To test whether concentration gradients of diffusible compounds might be responsible for directional sprouting, the formation concentration gradient was prevented by incubating trypsinized EC on a horizontal shaker. As shown in Figure 4A, culturing cells with constant shaking did not prevent the formation of directional sprouting. To rule out that there was no invisible pre-existing physical cell-cell contact between the apoptotic cell and sprouting EC, scratches were made on the surface of the culture dish between attracting apoptotic and sprouting EC. Scratches did not disturb the extension of the sprout (Fig. 4B and 4C), and similar to the results shown in Figure 2, the sprouts retracted only in response to removal of the attracting apoptotic cell (Fig. 4D).

Fig. 4.

Absence of mediation of the signaling between an attracting and sprouting cell by diffusible molecules and electrostatic characterization of apoptotic and sprouting cells. (A) Continuous horizontal shaking did not prevent sprout formation (arrow) toward its attracting cell (arrowhead). (B, C) Scratching the surface of the culture dish between an attracting cell (arrowheads) and sprouting cell (arrows) did not disrupt the pre-existing sprout (C, 30 min after B). (D) Removal of the attracting cell led to shrinkage of the sprout (15 min after C). The sprouting EC bound anionized ferritin (E, F, arrow), whereas the apoptotic cell bound cationized ferritin (G, H, arrow). (I) At the early stage of the process, cationized ferritin completely blocked the formation of sprouts, and (J) anionized ferritin failed to block the formation of sprouts (arrow). (K, L) The addition of anionized ferritin regressed the sprout before it reached the apoptotic cell (arrows; 100 min between K and I) but did not regress an established sprout (arrows with diamond head). Bar = 50 μM.

Because apoptosis is associated with an increase in net negative cell surface charge (15, 16), the possibility exists that the initiation of EC sprouts could be dependent on specific electrostatic charge interactions between the attracting apoptotic cell and the sprouting EC. To dissipate the polarity differential between the apoptotic and sprouting EC, trypsinized EC cultures were treated with cationized or anionized ferritin, respectively. Binding of ferritin to the cells was determined by staining with ferritin antibodies. Figure 4 shows that anionized ferritin bound exclusively to sprouting EC (Fig. 4E and 4F). Cationized ferritin, on the other hand, bound only to attracting apoptotic cells (Fig. 4G and 4H). Interestingly, the addition of cationized (Fig. 4I) but not anionized (Fig. 4J) ferritin to the cultures prevented sprout initiation. In contrast, addition of anionized, but not cationized, ferritin to EC bearing established sprouts caused the extending sprout to retract back into the main cell body within 90-100 min. Ferritin-induced sprout reversal, however, was ineffective once a sprout reached the attracting apoptotic cell (Fig. 4K and 4L).

To test whether negative surface charge is the primary stimulus by which EC are triggered to sprout, we attempted to initiate charge-dependent sprouting in an apoptotic cell-free system. For this purpose, Sepharose beads coated with anionized (negative-charged) or cationized (positive-charged) ferritin, and negative-charged PS or neutral PC vesicles were added to trypsinized EC cultures. Figure 5A shows that negative-charged, but not positive-charged (Fig. 5B), Sepharose beads initiated sprouting. Similarly, negative-charged (Fig. 5C) but not neutral (Fig. 5D) vesicles initiated sprouting. Collectively, these results suggest that the negative-charged cell surface expressed in apoptotic cells provides the initiating stimulus that induces the formation of EC sprouts.

Fig. 5.

Induction of EC sprouting by negatively-charged Sepharose beads and negatively-charged phospholipid vesicles. (A) A single negatively-charged Sepharose bead (asterisk) located in the vicinity of a denuded area of partially trypsinized EC induced sprout formation (arrows). (B) Positively-charged beads failed to initiate sprouting. (C) Negatively-charged PS vesicles (asterisk) induced formation of a sprout from the nearby EC (arrow), and (D) neutral vesicles did not. Bar = 50 μM.

EC Membrane Hyperpolarization and Cytoskeleton Reorganization Triggered by PS Phospholipid Vesicles. Because EC membrane is rich in ion channels (17) and is polarized in resting cells (18), we tested whether distant static negative charge can alter the membrane potential of EC. Dual wavelength imaging analysis (19) of EC labeled with membrane potential sensitive dye, di-8-ANEPPS, revealed the hyperpolarization of EC cell exposed to PS vesicles. This was detected within 15 min after exposure to the vesicles (Fig. 6A to 6D) and preceded the appearance of sprouts (Fig. 6E). No transmembrane potential changes were detected in the control cells (Fig. 6F to 6I).

Fig. 6.

PS vesicles trigger EC membrane hyperpolarization prior to sprout formation. EC exposed to PS (asterisk) (A, C, and D) had hyperpolarized membranes as indicated by the increased red emission intensity (arrows) within 15 min after addition of the vesicles (B, D, arrows, note that the fluorescence shifted from green to yellow). Sprout formation was detected 30 to 40 min after exposure to the vesicles (A, C and E, arrows). Control EC did not show significant morphological changes during a 15-min monitoring (F, H), and their corresponding membrane potentials remained unchanged (G, I). Bar=20μm.

Since alteration in membrane potential can trigger cytoskeleton reorganization (19), we tested whether EC sprouting was associated with reorganization of the cytoskeleton by staining with FITC-conjugated phalloidin. Reorganization of the cytoskeleton occurred during different stages of sprouting (Fig. 7). Non-sprouting EC exhibited a non-directional distribution (Fig. 7A), whereas in sprouting EC, the cytoskeleton polarized (Fig. 7B), elongated in a parallel manner (Fig. 7C), and finally concentrated at the tip of the sprout (Fig. 7D). Since Ca⁺⁺ flux changes have been shown to regulate cytoskeleton reorganization (21), we tested whether calcium channel blockers can inhibit sprout formation. Figures 7E and 7F show that pretreatment of EC cells with protopine completely inhibited PS vesicle-induced sprout formation.

Fig. 7.

Cytoskeleton re-organization and involvement of calcium signaling during sprouting. The cytoskeleton of non-sprouting ECs is non-directional (A), whereas the cytoskeleton of sprouting cells is polarized (B), extended in a parallel manner (C) at the early stage, and concentrated at the tip of sprout at the late stage (D). Control ECs formed sprouts (arrows) in response to PS vesicles (asterisk) (E), while the pre-incubation of EC with the calcium channel blocker, protopine, inhibited sprout formation (F). Bar=20μm.

DISCUSSION

Angiogenesis in development, wound healing, and neoplasia is dependent on the initiation of a defined sequence of events that include EC migration toward the region of (re)vascularization, EC proliferation and, finally, reorganization into blood carrying tubules. Although the critical initiating event for the generation of new blood vessels has been attributed to the production of diffusible growth factors that stimulate EC migration and proliferation, recent data suggests that endogenous electric fields may also participate in this process (22). Indeed, alterations in electric fields are associated with wounding where they persist until repair is complete (23).

Cells undergoing apoptosis undergo dramatic intracellular and membrane alterations. In particular, the normally asymmetric transmembrane distribution of membrane phospholipids reorganizes in such a manner that PS, normally localized exclusively in the cell's inner membrane leaflet, redistributes to the outer membrane leaflet. The expression of anionic phospholipid results in increasingly negative-surface charge (24) commonly identified by the ability of the cells to bind annexin V (25). The data presented here demonstrate that EC produce sprouts in direct response to cell-surface electrostatic charge on apoptotic cells. Studies carried out in the presence of the tyrosine kinase inhibitor, AEE788, a dual inhibitor of EGFR and VEGFR phosphorylation, revealed that sprout formation was independent of activation of the VEGFR and EGFR. In addition, continuous agitation to prevent formation of solute concentration gradients failed to affect sprout formation. Taken together, these data indicate that sprout formation towards attracting apoptotic cells is growth factor/growth factor receptor-independent.

All the attracting cells were annexin V-positive and bound cationized ferritin that, when added during the early stages of sprout formation, was inhibitory. This suggests that a net negative surface charge is required for the initiation of sprout formation. Additional evidence in support of the concept that cell sprouting can be initiated by electrostatic charge comes from experiments showing that negatively charged beads and PS-containing vesicles (Fig. 5) also initiated sprouting. In contrast to negative-charged apoptotic cells that bound cationized ferritin, sprouting EC bound anionized ferritin, suggesting that the sprouting cell has a strong positively-charged surface. While the source and nature of the positive surface charge on sprouting EC remains unclear, this finding is consistent with reports that EC derived from angiogenic macrovascular tissues elongate and migrate toward the cathode in a direct current electric field (26, 27).

Using dual wavelength imaging, we found that EC respond to distant negative charges by altering membrane potential and becoming hyperpolarized, a phenomenon similar to what occurs in neuronal cells following exposure to electrical fields (19). The nature of EC membrane hyperpolarization seems to be unrelated to potassium and sodium channels sine glyburide, carbdotoxin (potassium channel blocker), and amiloride (sodium channel blocker) were without effect (data not shown). However, preincubation EC with calcium channel blocker, but no other channel blockers (data not shown) did inhibit sprout formation, indicating that calcium signaling is critical to sprout formation. Ion channel blockers did not reverse pre-formed sprouts (data not shown).

In conclusion, the data presented here provide evidence that apoptosis is not only important for marking the cell for elimination by phagocytes, but also triggers a sequence of events important for angiogenesis and vascular remodeling.