| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Immunology and Oncology, Centro Nacional de Biotecnología, Consejo Superior de Investigaciones Científicas/Universidad Autónoma de Madrid, Campus de Cantoblanco, E-28049 Madrid, Spain
1Correspondence: Department of Immunology and Oncology, Centro Nacional de Biotecnología/CSIC, Universidad Autónoma de Madrid, Campus de Cantoblanco, E-28049 Madrid, Spain. E-mail: cmartineza{at}cnb.uam.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: vasculogenesis • apoptosis • caspase inhibitors
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Remodeling and pruning processes are common to angiogenesis and vasculogenesis during embryonic development, both poorly understood events that include the growth of new vessels and regression of others (2)
. In adult mammals, angiogenesis takes place only during wound healing and in the female reproductive cycle; it nonetheless has a crucial role in pathological situations, such as malignant tumor growth, for which tumor cells require a vascular network to supply oxygen and nutrients as well as to remove waste products (3)
. These cells also exploit the vessel network for cell intravasation and subsequent metastasis (4)
.
Apoptosis is required by vertebrates for tissue repair and remodeling throughout life to cope with alterations produced by infection and injury (5)
. As angiogenesis occurs in de novo tissue formation and during wound healing, we hypothesize that apoptosis may be involved in angiogenesis during vascular network formation; disturbance of these apoptosis-based rearrangement mechanisms could thus result in defective angiogenesis.
Several recent reports provide indirect evidence supporting this hypothesis (6
, 7)
. They show that tumors induce regression of the preexisting co-opted host vasculature via apoptosis, leading to massive tumor cell loss. Tumors then restart growth by inducing angiogenesis (6
, 7)
; apoptosis is thus implicated in the early stages of tumorigenesis. Thrombospondin-1 (TSP-1) inhibits angiogenesis in vivo and in vitro through CD36 binding (8)
, triggering a signaling cascade through the Src family kinase p59fyn, caspase-3-like proteases and the stress-activated p38 mitogen-activated protein kinases. The biological result of this complex activation pathway is apoptosis (8)
. TSP-1 may thus participate in vascular remodeling, triggering local apoptosis. In addition, the serine/threonine protein kinase Akt/PKB may have an alternative role in regulating VEGF- and attachment-mediated survival signals in endothelial cells (9)
. These data contrast with earlier reports showing that activation of the apoptotic pathway blocks angiogenesis (10
, 11)
.
We used the standardized short-term Matrigel in vitro angiogenesis assay (12)
; our results indicate a role for apoptosis in correct endothelial cell rearrangement during angiogenesis in this assay with human umbilical vein endothelial cells (HUVEC) and human lung microvascular endothelial cells (HMVEC-L), where apoptotic cells were found within capillary-like structures. The mechanism underlying activation of the highly regulated apoptotic process appears to require procaspase-3 processing and caspase-3 activity. Caspase activation is initiated, followed by cytochrome c release from mitochondria; this is a pathway involved in stress-activated apoptosis. We further demonstrate that caspase inhibitory peptides or overexpression of anti-apoptotic genes blocked angiogenesis in vitro, giving rise to cell cumuli. These results were confirmed in other in vitro angiogenesis models based on 2- and 3-D collagen I gels, with similar results. Finally, we show that caspase inhibitory peptides blocked both functional and structural angiogenesis in vivo at an early vascular cell alignment step in two different mouse strains. Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) showed that in vivo endothelial apoptosis occurred before formation of new capillaries. Taken together, these data indicate that apoptosis has a major role in correct vascular network formation during angiogenesis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and cultured on plastic coated with 0.1% gelatin (Sigma, St. Louis, MO) in M199 medium (Life Technologies, Paisley, Scotland) with 20% fetal calf serum (FCS; BioWhittaker, Walkersville, MD), 10 mM HEPES, 2.5 µg/ml fungizone (both from Life Technologies), 100 µg/ml heparin, and 100 µg/ml endothelial cell growth supplement (ECGS) (both from Sigma). HMVEC-L were cultured in EGM-MV (cells and medium were from Clonetics, Walkersville, MD). The Jurkat T cell leukemia cell line (ATCC, Manassas, VA) was cultured in RPMI 1640 (BioWhittaker) with 10% FCS.
In vitro angiogenesis assays
Ninety-six-well plates were coated with 50 µl undiluted Matrigel (Becton Dickinson, San Jose, CA), which was allowed to gel (1 h, 37°C). HUVEC (3x104) or HMVEC-L (2x104) were seeded in M199 medium containing 5% FCS or in EGM-MV, respectively.
Time-lapse videomicroscopy analysis was performed using an Olympus IX70 inverted microscope equipped with a COHU high performance CCD black-and-white videocamera coupled to a Sony SVT-S3050P time-lapse videocassette recorder and video monitor. In this case, HUVEC (7.5x105 cells) were seeded on 10 cm2 sterile Flaskette glass slides (Nalge Nunc, Naperville, IL) coated with 500 µl Matrigel. Capillary-like structure formation was filmed for 24 h under phase contrast using a 10x objective. Images were acquired every 60 s and sequential frames were captured.
A model of angiogenesis based on rat tail collagen type I (Becton Dickinson) was also used. HUVEC (3x104 per well) were resuspended in 50 µl of a 0.5 mg/ml solution of collagen in DMEM (BioWhittaker) with 1% HEPES, seeded on a 96-well plate, and incubated (30 min, 37°C). M199 medium (160 µl) with 1% ECGS and 1% heparin was then added. VEGF (50 ng/ml) was used to induce differentiation. Inhibitor peptides or equivalent volumes of solvent solution dimethyl sulfoxide (DMSO) were added together with the cell suspension and medium. Two-dimensional collagen type I gel angiogenesis experiments were evaluated over a 7-day period.
Cell viability was assessed by the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) method. When angiogenesis was complete, a solution of MTT in phosphate-buffered saline (PBS) was added to each well to a final concentration of 0.5 µg/µl, incubated (4 h, 37°C), and plates were photographed. When indicated, crystals were solubilized with 100 µl DMSO. Absorbance was measured at 570 nm on a Dynatech MR 5000 multiwell plate reader (Dynatech, Burlington, MA).
Three-dimensional collagen gels were prepared by mixing 4 x 106 HUVEC/ml with PBS, 10X DMEM, 1% HEPES, 1% ECGS, 1% heparin, and 20 ng/ml bFGF to a final collagen type I concentration of 628 µg/ml. Cell suspensions (10–20 µl) were added to wells of a 48-well culture dish, then inverted and warmed (37°C, 30 min) to initiate collagen polymerization. M199 medium containing z-VAD (500 µM) or peptide solvent was added as appropriate.
Cell proliferation
HUVEC (104 cells) were seeded on 0.1% gelatin-covered 96-well plates. After 24 h, 3H-thymidine was added to medium for 5 h. Cells were washed three times with PBS, then lysed. Released radioactivity was measured on a 1205 Betaplate (Wallac Oy, Turku, Finland).
Analysis of endothelial cell migration
Polyvinylpyrrolidone-free polycarbonate filter Transwell inserts (6.5 mm diameter, 8 µm pores; Costar, Cambridge, MA) were incubated overnight with 0.1% gelatin and dried. Inserts were placed in a 24-well plate (Falcon, Franklin Lakes, NJ) containing 600 µl M199 medium and 0.2% BSA alone or in the presence of chemoattractant (10 nM bFGF alone or with 0.5 mM z-VAD). HUVEC (105) were seeded on the upper filter face in 200 µl medium. After incubation (24 h, 37°C), cells on the upper face were removed with cotton swabs. Filters were then quantitated by modified Wright-Giemsa staining (Grifols, Barcelona, Spain).
Analysis of endothelial cell apoptosis
TUNEL analysis was performed at different times after HUVEC seeding on Matrigel using POD In situ Cell Death Detection (Roche, Mannheim, Germany), according to the manufacturer’s protocol. Caspase inhibitors z-VAD, Ac-LEHD-CMK, or Ac-DEVD-CHO were from Bachem (Bubendorf, Switzerland) and z-IETD-FMK was from Calbiochem (La Jolla, CA). For caspase activity measurement, HUVEC (7.5x105 cells/plate) were seeded on 8 cm2 Petri dishes (Falcon). Uncoated controls and Matrigel-coated (500 µl) dishes were detached with 250 µl dispase (Life Technologies) (1 h, 37°C). Cells were centrifuged (250 g, 5 min); pellets were washed twice in PBS and frozen at -80°C. After thawing, pellets were resuspended in extraction buffer with Nonidet P-40. Cytosolic extracts were used to determine enzymatic activity in aliquots containing 1 mg/ml protein. The fluorescent caspase-3 substrate acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin (Ac-DEVD-AMC; Bachem) was used. Cleaved substrate fluorescence was determined by C18 reverse phase HPLC using fluorescence detection (338 nm excitation, 455 nm emission). Control experiments were performed to confirm that substrate release was linear with time and protein concentration under these conditions. z-VAD was used at 0.3–0.6 mM. Caspase-3 processing during in vitro angiogenesis was studied by Western blot in cell lysates obtained from HUVEC after 8 h culture. Cells were lysed with a chilled Nonidet P-40 buffer solution. After centrifugation (20,000 g, 15 min), supernatants were resolved in SDS-PAGE; proteins were transferred to nitrocellulose membrane and incubated 1 h with anti-caspase-3 antibody (1:1000; PharMingen, San Diego, CA). Blots were washed extensively and developed with peroxidase-conjugated goat anti-rabbit antibody (Dako, Glostrup, Denmark) using the ECL system (Amersham Pharmacia, Buckinghamshire, UK).
An anti-Fas antibody was used that blocked the ability of a Fas ligand (FasL) agonist antibody to trigger apoptotic cell death in Jurkat T cells. Neutralizing anti-human tumor necrosis factor 
(anti-hTNF-) and -ß (anti-hTNF-ß) antibodies (both from Sigma) were used to block the biological activity of TNF- and TNF-ß in the angiogenesis assay and for ligand detection in cytometric analysis.
Assessment of cytochrome c release
In vitro cultured HUVEC were analyzed 8 h after seeding on Matrigel. Double immunofluorescence was performed on 2% paraformaldehyde-fixed cells by incubation with a mouse anti-cytochrome c antibody (PharMingen) and a human anti-mitochondrial antiserum that recognizes the E2 polypeptide of the mammalian mitochondrial pyruvate dehydrogenase complex. Bound antibodies were detected with Cy2-conjugated goat anti-mouse (green; cytochrome c staining) and Cy3-anti-human IgG (red; mitochondrial staining) (both from Jackson ImmunoResearch, West Grove, PA). Nuclei were blue counterstained with TOPRO-3 (1:1000; Molecular Probes, Eugene, OR). Optical sections were obtained using an Ar-Kr laser and a TCS-NT Leica confocal imaging system.
Flow cytometry analysis
HUVEC cultured on plastic or differentiated on Matrigel were collected, plated in V-bottom 96-well plates (2x105 cells/well), then incubated with mouse anti-Fas (Immunotech, Marseille, France), FasL (PharMingen), or anti-endothelial cell (Chemicon) antibody and washed twice; FITC-labeled rabbit anti-mouse antibody (Dako) was added and the plate was incubated. Plates were washed twice and cell-bound fluorescence was determined in a profile EPICS-Elite Cytometer (Coulter, Miami, FL). M199 medium with 1% FCS was used in the wash steps and for antibody solutions; antibodies (5 µg/ml) were incubated for 30 min at 4°C. Mouse anti-human TNFR-I and goat anti-human TNFR-II antibodies, both phycoerythrin-conjugated (R&D Systems, Minneapolis, MN), were used for receptor expression.
DNA constructs, transient transfection, and retroviral transfection procedures
Using an optimized calcium phosphate precipitation technique (14)
, HUVEC were transiently transfected with 2.5 µg/ml plasmid DNA (pEF-BOS-CX-bcl2, pEF-BOS-CX-p35). pEF-BOS-CX was used as a void vector control. The retroviral pCL-bcl2-Neo vector containing human Bcl-2 cDNA was generated by cloning DNA into the EcoRI site of the pCLXSN retroviral plasmid (15)
. Retroviral production was by transient transfection of 293T cells, as described (16)
. For viral transduction, 105 cells were incubated (overnight, 37°C) with 1 ml of retroviral supernatant or virus-free medium in 5 µg/ml protamine sulfate (Sigma); infection was repeated 24 h later under the same conditions. Infected cells were selected in 0.1 mg/ml G418 (Life Technologies).
In vivo angiogenesis
The procedure used has been described (17)
. Briefly, 500 µL Matrigel containing 100 ng/ml VEGF (R&D Systems) and 64 U/ml heparin (Sigma), alone, with the indicated inhibitors or equivalent volume of solvent (negative control), was injected subcutaneously (s.c.) into the abdominal tissue of 6- to 8-wk-old female BALB/c or C57BL/6 mice along the peritoneal midline. BALB/c and C57BL/6 genetic backgrounds were used in different assays and gave similar results. For quantitation, mice were killed after 4 days, Matrigel pieces were excised and homogenized in Drabkin’s reagent (Sigma) with a glass douncer (17)
, centrifuged (1000 g, 10 min), and supernatants filtered through 45 µm pore filters (Millipore, Bedford, MA). Hemoglobin content was measured at 540 nm.
To analyze tube structures formed in vivo, excised Matrigel pieces were snap-frozen in liquid nitrogen, embedded in OCT compound and stored at -70°C. Cryosections (8 µm) were fixed in cold acetone. Immunofluorescence was analyzed on endothelial cells by incubation with a rabbit anti-von Willebrand factor antibody (Dako). Bound antibodies were detected with Cy3-conjugated goat anti-rabbit IgG (red) (Jackson ImmunoResearch). Nuclei were blue counterstained with 4', 6-diamidino-2-phenylindole (DAPI).
For TUNEL analysis, Matrigel was excised from mice 24 or 48 h after injection; cryosections were prepared and fixed in 4% formaldehyde and the MebStain Apoptosis Kit II (Immunotech) was used, with modifications. Fixed samples were washed twice with PBS and incubated with 0.2% Triton X-100 (15 min, room temperature). Samples were incubated with a 10% BSA-buffered solution, after which manufacturer’s indications were followed. Streptavidin-Cy3 (Jackson ImmunoResearch) was incubated overnight with a FITC-conjugated anti-CD31 antibody (PharMingen). Nuclei were counterstained with DAPI.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
19
20)
. In the presence of angiogenic factors, seeded endothelial cells initially dispersed in culture (Fig. 1
A) are induced to arrange (Fig. 1B
), align and form capillary-like structures (Fig. 1C, D
). A gradual decrease in cell number during tube-like structure formation can readily be observed during in vitro culture (Fig. 1A, B
). Moreover, single cell counts before structure breakdown at the end of the assay showed that only 50% of seeded cells remained alive in cultures 24 h after seeding (not shown). TUNEL analysis of cultures showed positive staining, indicating that apoptosis occurs during capillary-like structure maturation both outside and in the lumen of the tube-like structures (Fig. 2
A, B). Time course analysis of HUVEC during angiogenesis showed an increase with time in TUNEL-positive cells compared with controls on plastic (Fig. 2C
). These findings suggest that apoptosis may be responsible for the decrease in cell number. Cell lysates were evaluated for enzymatic activity of prominent apoptosis-associated caspases. In the Matrigel assay, caspase-3-like activity peaks 8 h after initiation of culture, declining slightly thereafter (Fig. 2D
). Other caspase activities such as caspase-1, -2, and -9 were not found (not shown). Biochemical characterization of procaspase-3 activation was confirmed with anti-caspase-3 antibodies in Western blot (Fig. 2E
); the 17 and 19 kDa active caspase-3 fragments were observed in Western blot during the angiogenic process (Fig. 2E
).
|
|
These preliminary observations prompted us to analyze the mechanism involved in apoptosis activation during HUVEC angiogenesis. Two major mechanisms participate in the initiation of apoptotic pathways. One depends on activation of initiator caspases that require specific cofactors, such as triggering of caspase-8, which is recruited to the CD95 death-inducing signaling complex via an adapter, FADD (Fas-associated protein with death domain) (21)
. This pathway is dependent on cell surface Fas receptor expression and FasL binding. This mechanism was excluded in these experimental conditions, as an antagonist anti-Fas antibody did not block the apoptotic process (see below). Alternatively, under stress conditions, mitochondrial cytochrome c is released into the cytosol (22
, 23)
, where it activates the cofactor Apaf-1 (apoptotic protease-activating factor 1) with subsequent proteolytic processing of another initiator caspase, caspase-9 (24
, 25)
. Confocal microscopy colocalization studies revealed cytochrome c release in mitochondria; at 7–8 h of in vitro angiogenesis progression, cytochrome c was found in the cytosol (Fig. 2F
, see inset). We conclude from these data that in the first steps of blood vessel formation, an apoptotic pathway is activated in some cells, leading to cytochrome c release from mitochondria and endothelial cell death. The question then arises whether these apoptotic cells and related apoptotic mechanisms participate in these angiogenesis models.
In vitro angiogenesis is impaired by blocking apoptosis through caspase inhibitors and by expression of anti-apoptotic genes in endothelial cells
To confirm the role of apoptosis during angiogenesis, we included caspase inhibitors in the angiogenesis assays and analyzed their effect on tube-like structure formation. Both HUVEC (Fig. 3
A–F) and HMVEC (Fig. 3G-K
) differentiate on Matrigel, forming tube-like structures (Fig. 3A, G
). These structures were unaffected when only the peptide solvent (1% v/v DMSO) was added (Fig. 3B, H
). Addition of the broad range caspase inhibitor z-VAD (0.5 mM) partially blocked the elongation step during tube-like structure formation on Matrigel, however, resulting in distorted cord-like structures composed of rounded cells and cell cumuli (Fig. 3C, I
). Other caspase inhibitors produced different effects: IETD (80 µM; a specific caspase-8 inhibitor) and DEVD (0.5 mM; a caspase-3 inhibitor) induced malformation in tube-like structures (Fig. 3D, F, J
) similar to that induced by z-VAD, but LEHD (0.15 mM; a specific inhibitor of caspase-9) had no effect (Fig. 3E, K
).
|
This alteration in the angiogenesis pattern concurred with the inhibition of caspase-3 activity during the first 8 h of angiogenesis and was confirmed by the decrease in pro-caspase-3 activation in fluorogenic and Western blot assays (Fig. 2D, E
). Cell viability during the assay was analyzed by the MTT method. HUVEC included in tube-like structures were alive, but cells outside the structures were dead and dispersed (Fig. 3M
). z-VAD addition to the in vitro Matrigel angiogenesis assay simultaneously inhibited tube-like structure formation and cell death (Fig. 3N
). In this case, 95% of cells remaining in the culture were alive, although no tube-like structures were found. In contrast, endothelial cell proliferation (MTT, 3H-thymidine) and migration (Transwell filter), steps also involved in angiogenesis, were unaffected by caspase inhibitors used at the same concentrations (not shown). Inhibition thus occurs only at the rearrangement step.
Collagen type I and Matrigel endothelial differentiation models were compared to study the effect of z-VAD in angiogenesis, using the MTT assay to confirm viability. In both models, z-VAD inhibited formation of vascular-like structures. VEGF induced reorganization of HUVEC into tube-like structures (Fig. 4
A) even in the presence of DMSO (Fig. 4B
). This was not the case when 0.5 mM z-VAD was added, after which the cells appeared completely disorganized (Fig. 4C
), like control cells (Fig. 4D
). Similar results were obtained when HUVEC were induced with bFGF to form tube-like structures on 3-D collagen gels. Controls, including the peptide solvent, showed normal vascular structures (Fig. 4E, F
), which were inhibited by 0.5 mM z-VAD (Fig. 4G
).
|
To verify the role of apoptosis in endothelial remodeling during angiogenesis, HUVEC were transfected with the anti-apoptotic human Bcl-2 and baculovirus p35 genes using an optimized calcium phosphate technique (14)
that renders 50% transfection efficiency. HUVEC were also retrovirally infected with void, human Bcl-2, and baculovirus p35 recombinant retroviral vectors to confirm results. Overexpression of the corresponding proteins was confirmed by Western blot (not shown). Compared with wild-type HUVEC control and results obtained with z-VAD-treated HUVEC, Bcl-2, or p35 overexpression prevented correct organization of cells and completion of structures to a similar degree (Fig. 5
B, C, respectively). HUVEC transfected with the void vector formed normal tube-like structures (Fig. 5A
).
|
Exogenous addition of pharmacological inhibitors of apoptosis or overexpression of anti-apoptotic genes in culture thus promote survival of rounded cells, resulting in malformed cord-like structures. We conclude that apoptosis is a necessary early event for correct vessel formation in the in vitro Matrigel assay, since apoptosis inhibition causes incomplete, probably nonfunctional, angiogenesis.
Induction of Fas–FasL interaction or blockade of TNF- and TNF-ß action does not modify in vitro angiogenesis of endothelial cells
Fas–FasL interaction was excluded as a component of this apoptotic pathway, since a blocking anti-Fas antibody did not interfere with the in vitro angiogenesis process and tube-like HUVEC organization was observed (Fig. 5D
); this inducer of apoptosis effect was corroborated in a parallel assay using Jurkat cells (Fig. 5E, F
). Fas and FasL expression was also analyzed by flow cytometry; neither Fas nor FasL expression was observed during angiogenesis on Matrigel or proliferation on plastic (not shown). Similarly, incubation with neutralizing anti-TNF- or -TNF-ß antibodies modified neither the normal angiogenic pattern nor TNF-, TNF-ß, TNFR-I, or TNFR-II expression levels in HUVEC (not shown).
Caspase inhibitors block in vivo angiogenesis
Having established that apoptosis is important during early in vitro angiogenesis, we analyzed its significance in vivo by testing the effect of caspase inhibitors in a murine angiogenesis model based on s.c. injection of Matrigel. In this assay, angiogenesis is dependent on the angiogenic factor VEGF; in vivo Matrigel angiogenesis is then indirectly determined by hemoglobin content measurement (17)
. z-VAD or Ac-DEVD-CHO inhibited VEGF-promoted angiogenesis (80% and 65%, respectively; Fig. 6
A); inhibition was not observed when a random control peptide was added (not shown). BALB/c and C57BL/6 genetic backgrounds were used in different assays and gave equivalent results.
|
Reduced hemoglobin content correlated with reduced in vivo formation of capillary structures (Fig. 6B
). We used anti-von Willebrand factor antibody for immunofluorescent labeling on Matrigel cryosections excised from in vivo assays for induction with VEGF (Fig. 6B
, panel a), induction in the presence of peptide solvent (panel b), and induction in the presence of z-VAD (panel c). For immunofluorescence analysis, endothelial cells in the sections were developed in red (upper images) and nuclei DAPI counterstained in blue (lower images). Both controls showed normally formed capillary structures. In vivo angiogenesis in the presence of 0.5 mM z-VAD formed incomplete, distorted capillaries (panel c).
In vivo endothelial cell apoptosis was further analyzed histologically in Matrigel pieces.
TUNEL-positive nuclei are observed in cryosections before the endothelial cells form capillary structures; nuclei are not apoptotic when these structures become evident (Fig. 6C
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 27)
. Here we show not only that apoptosis occurs during angiogenic processes in in vitro models, but that it is necessary for correct in vivo remodeling of endothelial cells. Specific inhibition of caspase activation during angiogenesis results in distorted tube-like structures in vitro as well as in vivo, as indicated by immunohistology of cryosections.
Physiological cell death is clearly an essential component of animal development, critical in establishing and maintaining functional tissue architecture (28)
. Tissues are constructed following three general principles: cell proliferation, differentiation, and destruction. The most dramatic cases are vertebrate nervous and immune systems in which, after extensive proliferation and differentiation, apoptosis has an important role in cell selection; unnecessary neurons and lymphocytes bearing inoperative or self-reactive receptors are deleted (5)
.
We hypothesized that apoptosis may also participate early in the formation of new vascular tissue. Angiogenesis would thus be a multistep process comprising activation of endothelial cells, proliferation, migration, apoptosis of superfluous cells, and differentiation of remaining cells into a new, functional blood vessel network. We show that blockage of caspase activity through the action of z-VAD, a broad-spectrum caspase inhibitor, produces alterations in the normal angiogenic pattern. TUNEL assay pointed to apoptosis rather than necrosis of endothelial cells as the process implicated in in vitro angiogenesis, where activation of caspase-3 occurs with maximum activity 8 h after cell seeding.
Caspase-3 is an executioner, not an initiator caspase; thus it must be activated by another caspase such caspase-2, -8, -10 (in the case of TNF family death receptors) or caspase-9 (in the case of intracellular stress and cytochrome c release from mitochondria). We found both putative pathways, 1) caspase-8 activation, as inhibition of caspase-8 by IETD peptide produces equivalent alterations to angiogenesis in presence of z-VAD and 2) cytochrome c release from mitochondria, as shown by confocal microscopy. Incubation of endothelial cells with LEHD, a specific inhibitor of caspase-9, nonetheless showed no difference compared to control angiogenesis. Cytochrome c release thus may be a consequence, not the primary effector, of the apoptosis observed.
These results suggested an apoptotic pathway whose initial step would be activation of a death receptor but not internal cellular stress. Several cases have been reported of cytochrome c release after caspase-8 activation in which the cytosolic Bcl-2 family protein BID (BH3 interacting domain death agonist) is cleaved by caspase-8, then translocates to mitochondria, where it triggers cytochrome c release (29)
. Considering these precedents, we studied three possible initiator molecules: Fas ligand, TNF-, and TNF-ß. In all three cases, the result was negative; all apoptotic pathways should nevertheless be taken into consideration. The putative role of other TNF family members such TRAIL or other unknown effectors requires further investigation. Recent data confirming a mediator role for p35 in a TRAIL-dependent apoptosis pathway (30)
, as well as the inhibitory effect on angiogenesis observed in p35-transfected HUVEC, support this suggestion. Several authors nonetheless propose that cell shape has a very important role in the signal transduction events that drive cells to one fate or another. This would explain why one cell proliferates while neighboring cells only micrometers away turn on entirely different gene programs that lead to differentiation or apoptosis (31)
.
We used the in vitro and in vivo Matrigel assays, which generate highly reproducible results (12)
. Similar results were obtained in in vitro VEGF- and bFGF-induced angiogenesis on 2- and 3-D collagen type I gels, respectively. The correspondence of three different in vitro models of tube-like structure formation points to apoptosis as a common step in the angiogenic process.
Our results with two different in vitro human endothelial cell systems, HUVEC and HMVEC-L, in three different in vitro models extend this observation and show that apoptosis may be a general mechanism in angiogenesis, probably to eliminate superfluous cells not included in the vascular network. In the in vivo chorioallantoic membrane (CAM) model, monoclonal anti-vß3 integrin antibodies trigger apoptosis and prevent angiogenesis, also preventing tumor cell progression (10)
. Similarly, retrovirally mediated overexpression of caspase-resistant Bcl-2 protein delays HUVEC apoptosis in vitro for > 7 days and to 31 days in vivo (11)
. Two closely related observations are germane. First, we studied angiogenesis at a very early stage, when endothelial cell proliferation has not yet begun or is just commencing; we thus analyze initial phases of vessel formation. Second, this early phase is observed in the absence of vß3 expression, as the anti-vß3 antibody LM609 had no effect in the in vitro Matrigel angiogenesis assay, nor did the vß3-specific disintegrin kistrin block VEGF-induced angiogenesis in vitro (not shown). In the CAM model, the inhibition mediated by integrin targeting probably occurs at a much later stage, when vessels are already formed, acting on their survival. We thus identified a mechanism that operates before vessel formation, whereas previous work analyzed mechanisms involved in vessel maintenance. We cannot rule out that other integrins such as vß5 may be implicated at this early stage of vessel formation, as they have been described in VEGF-mediated angiogenesis (32)
; other apoptosis-related phenomena may be involved in very early stages of angiogenesis.
Vertebrate cell death machinery is redundant; thus, it is not surprising that most mutations that compromise this machinery have little effect on the overall ability of vertebrate tissue to develop. Many apoptotic gene knockout models result in lethality during embryogenesis or the neonatal period, although death usually originates from local failure in a specific tissue rather than from a general embryonic apoptosis blockade. Due to promiscuity, different tissue types activate apoptosis mechanisms by alternative pathways. In our angiogenesis models, caspase-3, caspase-8, and cytochrome c release appear to have a crucial role. In summary, these observations may provide a better understanding of the mechanisms involved in angiogenesis, which will help to develop new therapeutic approaches to cancer and other angiogenesis-based disorders.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 15, 2001.
Revision received January 31, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
vß3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels. Cell 79,1147-1164[CrossRef][Medline]
v integrins. Science 270,1500-1502This article has been cited by other articles:
|
|
 |
|
  G. M. Birdsey, N. H. Dryden, V. Amsellem, F. Gebhardt, K. Sahnan, D. O. Haskard, E. Dejana, J. C. Mason, and A. M. Randi Transcription factor Erg regulates angiogenesis and endothelial apoptosis through VE-cadherin Blood, April 1, 2008; 111(7): 3498 - 3506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Ferrari, G. Pintucci, G. Seghezzi, K. Hyman, A. C. Galloway, and P. Mignatti VEGF, a prosurvival factor, acts in concert with TGF-beta1 to induce endothelial cell apoptosis PNAS, November 14, 2006; 103(46): 17260 - 17265. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. de Resende, S. L. Amaral, D. H. Munzenmaier, and A. S. Greene Role of endothelial cell apoptosis in regulation of skeletal muscle angiogenesis during high and low salt intake Physiol Genomics, April 13, 2006; 25(2): 325 - 335. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Wagner, I. Petrache, B. Schofield, and W. Mitzner Pulmonary ischemia induces lung remodeling and angiogenesis J Appl Physiol, February 1, 2006; 100(2): 587 - 593. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ben-Yosef, A. Miller, S. Shapiro, and N. Lahat Hypoxia of endothelial cells leads to MMP-2-dependent survival and death Am J Physiol Cell Physiol, November 1, 2005; 289(5): C1321 - C1331. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Yap, D. Veliceasa, U. Emmenegger, R. S. Kerbel, L. M. McKay, J. Henkin, and O. V. Volpert Metronomic Low-Dose Chemotherapy Boosts CD95-Dependent Antiangiogenic Effect of the Thrombospondin Peptide ABT-510: A Complementation Antiangiogenic Strategy Clin. Cancer Res., September 15, 2005; 11(18): 6678 - 6685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Mezentsev, R. M. H. Merks, E. O'Riordan, J. Chen, N. Mendelev, M. S. Goligorsky, and S. V. Brodsky Endothelial microparticles affect angiogenesis in vitro: role of oxidative stress Am J Physiol Heart Circ Physiol, September 1, 2005; 289(3): H1106 - H1114. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. Enis, B. R. Shepherd, Y. Wang, A. Qasim, C. M. Shanahan, P. L. Weissberg, M. Kashgarian, J. S. Pober, and J. S. Schechner Induction, differentiation, and remodeling of blood vessels after transplantation of Bcl-2-transduced endothelial cells PNAS, January 11, 2005; 102(2): 425 - 430. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Kirby Anthrax Lethal Toxin Induces Human Endothelial Cell Apoptosis Infect. Immun., January 1, 2004; 72(1): 430 - 439. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Cascone, E. Giraudo, F. Caccavari, L. Napione, E. Bertotti, J. G. Collard, G. Serini, and F. Bussolino Temporal and Spatial Modulation of Rho GTPases during in Vitro Formation of Capillary Vascular Network: ADHERENS JUNCTIONS AND MYOSIN LIGHT CHAIN AS TARGETS OF Rac1 AND RhoA J. Biol. Chem., December 12, 2003; 278(50): 50702 - 50713. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||
HUVEC during angiogenesis on Matrigel; 
HUVEC during angiogenesis on Matrigel with 0.4 mM z-VAD; 
HUVEC on plastic; x HUVEC on plastic with 0.4 mM z-VAD). Values represent the mean of three assays. E) Caspase-3 is processed after 8 h of in vitro angiogenesis. Cell lysates were obtained from HUVEC after 8 h culture on Matrigel, resolved by SDS-PAGE, and analyzed by Western blot (Materials and Methods). Lane a, HUVEC on plastic; lane b, HUVEC after 8 h angiogenesis on Matrigel; lane c as for lane b, with 0.4 mM z-VAD. F) Confocal microscopy of apoptotic activity requiring cytochrome c release from mitochondria. In vitro cultured HUVEC were analyzed by confocal microscopy 8 h after seeding (Materials and Methods). Cytochrome c (green) and mitochondrial staining (red); nuclei are counterstained with TOPRO-3 (blue). Bar: 50 µm. Inset shows a representative HUVEC in which cytochrome c is clearly localized in the cytosol. Bar: 10 µm.
