Endothelial cell (EC) activation plays a key role in vascular inflammation, thrombosis, and angiogenesis. Allograft inflammatory factor-1 (AIF-1) is a cytoplasmic, calcium-binding, inflammation-responsive scaffold protein that has been implicated in the regulation of inflammation. The expression and function of AIF-1 in EC is uncharacterized, and the purpose of this study was to characterize AIF-1 expression and function in ECs. AIF-1 expression colocalized with CD31-positive ECs in neointima of inflamed human arteries but not normal arteries. AIF-1 is detected at low levels in unstimulated EC, but expression can be increased in response to serum and soluble factors. Stable transfection of AIF-1 small interfering RNA (siRNA) in ECs reduced AIF-1 protein expression by 73% and significantly reduced EC proliferation and migration (P < 0.05 and 0.001). Rescue of AIF-1 expression restored both proliferation and migration of siRNA-expressing ECs, and AIF-1 overexpression enhanced both of these activities, suggesting a strong association between AIF-1 expression and EC activation. Activation of mitogen-activated protein kinase p44/42 and PAK1 was significantly reduced in siRNA ECs challenged with inflammatory stimuli. Reduction of AIF-1 expression did not decrease EC tube-like structure or microvessel formation from aortic rings, but overexpression of AIF-1 did significantly increase the number and complexity of these structures. These data indicate that AIF-1 expression plays an important role in signal transduction and activation of ECs and may also participate in new vessel formation.
- small interfering RNA
- vascular inflammation
- allograft inflammatory factor-1
as the barrier between blood and tissues, endothelial cells (ECs) play a unique role in vascular biology in terms of both health and disease. Through synthesis and release of numerous small molecules, ECs maintain vascular tone; ECs maintain a nonthrombogenic surface to regulate adherence of platelets and leukocytes; ECs act as a permeability barrier to govern exchange of molecules from the blood stream into the wall of the blood vessel; and ECs can modify the oxidative state of lipoproteins (9, 28). When inflamed, ECs can synthesize and secrete multiple cytokines and growth factors, which act in a paracrine fashion to modify the activity of adjacent smooth muscle cells and also regulate leukocyte extravasation. EC activation, often resulting in migration and proliferation, is an essential component of multiple normal and pathophysiological processes including atherosclerosis, permeability, wound healing, angiogenesis, and regulation of vascular tone (1, 13, 17). Consequently, identification of proteins that regulate the EC response to inflammatory stimuli is key to our better understanding the multiple vascular diseases in which inflammation plays a role.
Allograft inflammatory factor-1 (AIF-1) is a 143 amino acid, cytoplasmic, evolutionarily conserved, calcium-binding protein. High levels of AIF-1 are constitutively expressed in inflammatory tissue and glial cells. Data in several groups, including ours, utilizing diverse systems advocate an important role for AIF-1 in inflammatory processes, primarily involving macrophages (14). These studies range from expression in infiltrating macrophages in rat cardiac allografts (33), in lesions of experimental autoimmune encephalomyelitis (30), and in rheumatoid arthritis (21). AIF-1 expression is evolutionarily conserved and is induced in the allograft response of such phylogenetically distant species as marine sponges (23). In humans, the AIF-1 gene maps to the major histocompatibility complex (MHC) class III region on chromosome 6p21.3, which is densely clustered with genes involved in the inflammatory response, including surface glycoproteins, complement cascade, tumor necrosis factor (TNF)-α and -β, and nuclear factor (NF)-κB (19). Single nucleotide polymorphisms (SNPs) in the AIF-1 gene have been strongly linked to systemic sclerosis, a fibrotic microvascular disease (12, 27). Structurally, AIF-1 has signatures of a cytoplasmic signaling protein, including several PDZ interaction domains, which are important in mediating interactions of multiprotein complexes (18). Consistent with this, it has been demonstrated that in some cell types, overexpression of AIF-1 leads to enhanced proliferation, migration, and activation of signal transduction pathways. Our previous work has reported the expression of AIF-1 in medial and neointimal vascular smooth muscle cells (VSMC) in human coronary arteries with coronary artery vasculopathy (CAV), a disease in which the endothelium of arteries in allografted human hearts are chronically inflamed. Furthermore, persistent expression of AIF-1 in cardiac allografts is associated with development of CAV (2). In VSMC, we have shown that overexpression of AIF-1 leads to increased migration and proliferation (3, 4). In addition to VSMC, we have shown that overexpression and attenuation of AIF-1 expression in monocytes and T-lymphocytes leads to changes in the activation state of those cells (32, 20).
Multiple cell types, including VSMC, monocytes, lymphocytes, and ECs participate in and contribute to the vascular response to injury, particularly atherosclerosis and restenosis. However, despite previous studies that have correlated increases in AIF-1 expression in arterial pathology, only one paper reports AIF-1 immunoreactivity in CD31 cells (12). Furthermore, characterization of AIF-1 expression in ECs has not been reported and a defined function for this protein in EC pathophysiology and vasculogenesis has not been described. On the basis of previous studies in other relevant cell types, we hypothesized that AIF-1 expression is central for EC activation. In this study, we report that AIF-1 expression colocalizes with ECs in neointima in human arteries with CAV. Abrogation of AIF-1 expression in ECs by constitutive expression of small interfering RNA (siRNA) suppresses cell proliferation and migration and also attenuates signal transduction cascades. Interestingly, the expression pattern of and participation and regulation of signal transduction cascades modified by AIF-1 in ECs are distinct from those in VSMC and macrophages. In EC angiogenesis assays, modulation of AIF-1 expression can regulate tube-like structure formation and aortic sprouting. Together, these data suggest that AIF-1 expression regulates EC activation in specific ways and plays a role in vascular inflammation, and potentially, in the formation of new blood vessels.
MATERIALS AND METHODS
Tissues were obtained and processed as described (2). Coronary arteries were removed from nonfailing hearts from organ donors that were deemed not appropriate for transplantation and from hearts that needed to be removed from patients with severe CAV at the time of retransplantation (2). The length of time for heart removal due to CAV ranged from 3 to 7 yr. Five different normal and seven different CAV arteries, from males and females, were tested with identical results; representative sections are shown. Tissue used in this study is from a tissue bank of sections obtained from standard pathology tissue collection. Use of these tissue blocks were approved by the Institutional Review Board of Temple University Hospital. Human CD31 (EC marker) (Neo Markers, San Diego, CA) were used at a concentration of 2 μg/ml. AIF-1 antibody, which has been previously described, was used at 1.0 μg/ml (18). Sections were then incubated with biotinylated secondary antibody (1:200) followed by avidin-biotin-peroxidase complex in a Vectastain Elite kit (both from Vector Labs, Burlingame, CA). The reaction product was visualized with DAB (Vector Labs) used as the chromogenic substrate, which produces a reddish-brown stain. The sections were counterstained with hematoxylin. For immunofluorescence, primary antibody incubation was followed by a 30-min incubation with secondary antibody conjugated to AlexaFluor 568 (red) and AlexaFluor 488 (green) (Molecular Probes, Eugene, OR). For junctional adhesion molecule (JAM) rearrangement, ECs that overexpressed AIF-1 or siRNA were grown on coverslips coated with collagen. Some were unstimulated and others stimulated with bFGF for 30 min to induce JAM rearrangement, fixed with 4% paraformaldehyde, and immunostained with anti-JAM-A antibody as described (26).
siRNA expression plasmids.
AIF-1 siRNA constructs as previously described were synthesized by GenScript (Scotch Plains, NJ) (32). The 19 base pair regions of human AIF-1 mRNA were targeted (5′ to 3′ AGAGAGGCTGGATGAGATC) and chemically synthesized as part of a small, double-stranded 70-bp DNA insert containing the target sequence in the sense orientation, followed by a short loop region, the target in the antisense orientation, and six thymidines added to the 3′ end, which serves as a polymerase III transcription termination site. This insert was flanked by BamHI and HindIII restriction sites and cloned into the siRNA expression vector pRNA-U6.1/shuttle (pShuttle), which contains a RNA polymerase III promoter, that initiates the transcription of a short hairpin RNA rapidly processed by cellular machinery into 19–22 nt double-stranded RNA (siRNA). The pShuttle plasmid contains the selectable marker Neomycin to facilitate selection of stably transfected cells.
Cells and culture.
Primary human vascular ECs (HUVECs) were obtained as cryopreserved secondary culture from Cascade (Portland, OR) and subcultured in 200 medium with low serum growth supplement. Primary bovine aortic endothelial cells (BAEC), also purchased from Cascade Biologics, were cultured in MCDB 131 medium with 10% fetal bovine serum (FBS) and 100 IU/ml penicillin and 50 mg/ml streptomycin. Cells in passages 3–5 were used in the described studies. Preconfluent ECs were serum starved in 0.3% FBS for 24 h and then exposed to 10% FBS, 15% T cell-conditioned media (Fisher Biotech), 5 μg/ml oxidized LDL, 20 ng/ml VEGF, 10 ng/ml bFGF for another 72 h, at which times samples were processed for protein isolation. Some samples remained untreated and were used as controls. All cytokines were purchased from Sigma (St. Louis, MO).
Stable transfections and adenovirus infection.
Cells were transfected with control vector (pRNA-U6.1/shuttle) alone or with pRNA-U6.1/shuttle-AIF-1-siRNA by electroporation (Amaxa, Gaithersburg, MD) according to manufacturers instructions. Antibiotic G418 (300 ng/ml) was used to select for transfected cells. Stably transduced bovine ECs were pooled from each transduction to avoid the effects of clonal variation. For rescue, siRNA stably transfected cells were incubated with 20 multiplicities of infection AIF-1 adenovirus (AdAIF-1) or GFP adenovirus (AdGFP) overnight and used 48 h later.
Equal numbers of stable transfectants were seeded into 24-well plates at a density of 7,500 cells/ml as described previously (15). Medium was changed on the third day, and after 2 and 5 days, cells were counted by using a standard hemocytometer.
Migration assay and scratch wounding.
Boyden chamber migration assays were performed as we described (3). 6.5-mm diameter Transwell Boyden chamber plates (Fisher Biotech, Pittsburgh, PA) with 8-μm polycarbonate membrane pore size were seeded with stably transfected or adenoviral rescued cells (4×104 cells per membrane) in medium containing 0.5% FCS. Twenty nanograms of VEGF (Sigma) were placed in the lower chamber. Cells were incubated for 3 h at 37°C, at which time cells were fixed and stained in Dif-Quick Cell Stain (American Hospital Supply, Baltimore, MD). The upper layer was scraped free of cells. Cells that had migrated to the lower surface of the membrane were quantitated by counting four high-powered fields per membrane. Experiments were performed in triplicate from three independently transfected groups of bovine ECs. Scratch wounding was performed as we described (3). Briefly, equal numbers of stable ECs and those infected with AIF-1 adenovirus 48 h earlier were grown in growth media on glass slides to confluence, at which time monolayers were scraped with a cell scraper to create a 3-mm track devoid of cells in the center of the chamber. The wound tracks were immediately washed to remove any detached cells, and fresh medium was added. At different times after wounding, cells were fixed and stained with hematoxylin.
Tube-like structure and aortic ring assay.
As previously described (25), low-growth factor Matrigel (BD Biosciences) was thawed at 4°C for 48 h. Matrigel (200 μl) was added to each well of a 24-well tray and allowed to polymerize at 37°C for 30 min. ECs (200 μl; 2 × 105 cells/ml) were added on the top of matrigel suspended in MCDB-131 medium with 10% FCS and kept incubated at 37°C for 16 h. Images were taken on an inverted microscope using ×10 objective. Three images were taken per well from random fields. Each condition was performed in triplicate. The number of tube-like structures were counted manually per image, and an average was calculated for each condition. Mouse aortic ring assay was carried out as described (7, 35) using C57BL/6 mice (3–4 mo old). Briefly, thoracic aortas were excised from mice, and periadventitial fibroadipose tissues were removed. Aortas were then cut into 1-mm rings, placed into six-well trays containing MCDB medium 131 and 10 μl of adenovirus at 1.75 × 109 pfu/ml for 1 h. AIF-1, GFP, and siRNA adenovirus hve been previously described (32). Rings were then transferred to 48-well tissue culture plates coated with Matrigel (BD Biosciences) and overlaid with an additional 100 μl of Matrigel and allowed to gel for 30 min at room temperature. The plates were incubated at 37°C with MCDB medium 131 medium containing 2% autologous mouse serum and 50 ng/ml VEGF (R&D, Minneapolis, MN). Aortic rings were examined daily and digital images were taken at day 6 for quantitative analysis of the area of vessel outgrowth by the SPOT Advanced program (Media Cybernetics, Sterling Heights, MI). Microvessel outgrowth was calculated by circling the extent of microvessel outgrowth at 6 days and subtracting the area of the aortic ring (36). All animal procedures were approved by the Institutional Animal Care and Use Committee of Temple University.
To detect activation of intracellular signaling proteins, cells were rinsed with PBS, starved in 0.3% MCDB for 24 h, and stimulated with 10% FBS for the indicated times. Cell extracts were prepared as described (4). Extract proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and blocked. AIF-1 rabbit polyclonal antibody (1:3,000 dilution) has been previously described (5). Anti-phospho p44/42 (Santa Cruz, Santa Cruz, CA), anti-phospho-PAK, and antibody to the total protein of each of these kinases were from Cell Signaling Technology (Beverly, MA), and a 1:2,000 dilution of secondary antibody were used. Equal loading of protein extracts on gels was verified by Ponceu S staining of the membrane and normalization to either a housekeeping gene (GAPDH) or in the case of phosphoprotein, total protein. Blots were then stripped and reprobed with the housekeeping proteins anti-actin (1:1,000 dilution, Neomarkers) or anti-GAPDH (1:5,000 dilution, Biolegend, San Diego, CA). Reactive proteins were visualized using enhanced chemiluminescence (GE Healthcare, Piscataway, NJ).
RNA isolation and quantitative RT-PCR.
RNA was isolated from EC and reverse transcribed into cDNA as we have described (2). JAM-A, GAPDH, and β-actin mRNA were targeted by using primer pairs from Integrated DNA Technologies (Coralville, IA) and amplified by using an Eppendorf MCEP RealPlex 4× thermocycler. Product was quantitated by Eppendorf software.
Experiments were repeated at least three times. Results from proliferation and migration are presented as means ± SD and are compared by one-way ANOVA. Results form cell signaling transduction are presented as means ± SE and are compared by two-way ANOVA and Bonferroni post-tests. A value of P < 0.05 was considered significant.
AIF-1 expression is inducible in endothelium and cultured ECs. Inducible AIF-1 expression in smooth muscle cells, macrophages, and T-lymphocytes in injured and atherosclerotic human coronary arteries has been reported. One paper reports AIF-1 immunoreactivity in CD31 cells, but characterization of expression and function in EC has not been described (12). Representative immunohistochemical staining of tissue sections from normal and human coronary arteries from patients with cardiac allograft vasculopathy (CAV) were used because endothelium in these arteries are chronically inflamed. Costaining with antibody to AIF-1 and EC-specific CD31 antibody demonstrates AIF-1 immunoreactivity in endothelium in arteries with CAV but not normal arteries (Fig. 1). As expected, VSMC in these sections also stain positive for AIF-1. Increased AIF-1 immunoreactivity in endothelium in inflamed arteries was confirmed in multiple sections from several different patients. The lack of AIF-1 in uninjured vessels and its inducible expression in endothelium in inflamed, but not normal, arteries suggested that expression of this protein was cytokine inducible. We examined the induction of AIF-1 protein in both human (HUVECs) and bovine (BAEC) ECs stimulated with serum and a variety of soluble factors. In these experiments, cells were quiesced by incubation in 1% FCS for 48 h and exposed to 10% FBS, 15% T-lymphocyte-conditioned media (TCM), 20 μg/ml oxidized low-density lipoprotein (LDL), 20 ng/ml VEGF, and 20 ng/ml bFGF for 48 h, and AIF-1 was detected by Western blot analysis. Figure 2 indicates that a basal level of AIF-1 is detectible in unstimulated ECs of both species. However, AIF-1 expression is differentially induced by soluble factors in cultured ECs. Both 10% FBS and TCM can significantly increase (P < 0.05 and < 0.001) AIF-1 expression above basal levels. These experiments are the first to demonstrate differential expression of AIF-1 in ECs by soluble factors. This expression is similar to VSMC in that AIF-1 is inducible by 10% FCS and TCM but different in that growth factors do not induce increased expression. This inducible expression in cultured ECs agree with injury-responsive expression in vivo and strongly suggest that in ECs, AIF-1 expression is an injury-responsive protein induced by soluble factors.
Expression of AIF-1 regulates EC proliferation.
It was important to characterize the functional significance of AIF-1 expression, and we hypothesized that inducible expression is an important event in EC activation. We tested this hypothesis by inhibiting AIF-1 expression in ECs using stable, antibiotic selectable vector-based siRNA delivery. For these, and subsequent experiments, BAEC were used because they were more amenable to stable transfection and antibiotic selection than HUVECs. Establishment of stable cell populations enabled consistent and reproducible amounts of AIF-1-specific siRNA to reduce experimental variability as we have described (32). Vector-based delivery, rather than transfection of oligonucleotides, was used because it has been found to be more effective than synthetic siRNA for inhibition of gene expression, likely due to increased stability (8). An AIF-1 siRNA construct cloned into the pShuttle vector and empty vector (control) stable transfectants were isolated by antibiotic selection. Populations of resistant cells were pooled to avoid the effects of clonal selection. Extracts were made from these cells, and AIF-1 expression was determined by Western blot analysis. These cells retained their EC phenotype as determined by maintained PECAM expression (not shown). Figure 3A shows that stable expression of the AIF-1 siRNA construct reduced AIF-1 protein by an average of 73%, whereas vector alone (pShuttle) did not.
Increased proliferative capacity of ECs is a hallmark of EC activation during wound healing, development, and angiogenesis. To elucidate whether AIF-1 expression was required for EC proliferation, we tested whether silencing AIF-1 expression suppresses EC growth. Equal numbers of AIF-1 siRNA stable transfectants were seeded into replicate 24-well plates and counted on the third and sixth day. Figure 3B shows ECs that express AIF-1 siRNA grow 56% more slowly than do control vector-only cells (44.1 × 103 ± 8.2 and 99.4 × 103 ± 18.0 for siRNA and control cells for 6 days, respectively; P < 0.05). To validate that growth inhibition was due to abrogation of AIF-1 expression, we infected AIF-1 siRNA cells with AdAIF-1 to rescue AIF-1 expression. These experiments show that restoration of AIF-1 expression rescued the ability of these cells to proliferate by 239% (44.1 × 103 ± 8.2 and 149.2 ×103 ± 9.0 for siRNA and siRNA infected with AdAIF-1, respectively; P < 0.001) (Fig. 3B). Moreover, cells infected with AdAIF-1 grew 51% more rapidly than did pShuttle control cells (99.0 × 103 ± 18 for control and 149.2 ×103 ± 9.0 for rescued cells, respectively; P < 0.05). siRNA cells infected with green fluorescent protein adenovirus (AdGFP) were used as an adenoviral control and showed no significant change in growth capacity (Fig. 3B). Together, these data indicate that AIF-1 expression is tightly associated with EC proliferation.
Expression of AIF-1 regulates EC migration.
We hypothesized that AIF-1 expression would also participate in regulation of EC migration. To determine whether AIF-1 plays a role in EC migration, pShuttle control cells, siRNA stable transfectants, and siRNA ECs that had been rescued with AdAIF-1 were seeded into Boyden chambers, and migration was induced by addition of VEGF, a strong EC chemoattractant. Figure 4A shows that EC migration is regulated by AIF-1 expression, as AIF-1 siRNA expressing cells migrate significantly more slowly than control cells in both VEGF and unstimulated ECs (44.0 ± 0.1 vs. 26.3 ± 0.9 for unstimulated, and 107 ± 7.1 vs. 31.9 ± 5.1 in VEGF stimulated, P < 0.05 and 0.001, respectively). Importantly, there was no significant difference between unstimulated and VEGF-stimulated EC siRNA expressing cells, underscoring the importance of basal levels of AIF-1 expression for EC migration. siRNA expressing ECs rescued with AdAIF-1 migrated more rapidly than did control cells in the absence of chemoattractant (73.0 ± 0.9 Vs 44.0 ± 0.0, P < 0.001, for AdAIF-1 rescued and control EC, respectively). To further demonstrate a function for AIF-1 expression in EC motility, directional migration of EC monolayers following mechanical wounding was performed. Equal numbers of confluent, stably transfected pShuttle, siRNA, or siRNA rescued with AdAIF-1 were scraped to create a 3-mm wide wound track devoid of cells, and the time needed to fill the wound was compared. The data presented in Fig. 4B demonstrate that abrogation of AIF-1 expression reduces EC migration, and restoration of AIF-1 expression not only restored, but enhanced, the ability of these cells to migrate into the wound. Collectively, these data suggest AIF-1 involvement in EC migration.
AIF-1 inhibition reduces activation of signal transduction kinases.
The EC response to inflammatory and angiogenic stimuli leading to migration and proliferation is mediated at least in part by the MAPK p44/42 pathway (29). AIF-1 contains motifs consistent with a scaffold signaling protein, and we hypothesized that diminution of AIF-1 expression would disrupt signal transduction pathways. For these experiments, stable transfectants were incubated in serum-reduced medium for 48 h and then stimulated with 10% FCS for different times. Activation of several cytoplasmic kinases in extracts from these cells were quantitated by Western blot analysis using phospho-specific antibodies. Figure 5 shows that activation of MAPK p44/42 was significantly reduced in siRNA-expressed ECs at 5 and 15 min poststimulation (P < 0.05). PAK1 activity is also involved in EC migration (22). Because AIF-1 can induce motility, we hypothesized that abrogation of AIF-1 would also adversely affect PAK1 activation. Figure 6 demonstrates that abrogation of AIF-1 expression can significantly reduce serum-stimulated PAK1 activation (P < 0.05). We have previously shown in murine macrophages that knock down of AIF-1 inhibits p38 activation (32), and chronic overexpression of AIF-1 activates p38 in cultured VSMC and arteries from AIF-1 transgenic mice (31). In contrast, no inhibition of p38 activation in AIF-1 siRNA or activation of p38 in AIF-1 overexpressing EC was noted (data not shown). Together, these data indicate a close association between AIF-1 expression and activation of p44/42 and PAK1 in EC and suggests cell-type specificity of AIF-1 effects.
AIF-1 expression participates in tube formation and endothelial sprouting.
Migration of ECc in response to angiogenic factors is an obligate event in angiogenesis. Because PAK1 is an important regulator of EC migration, and AIF-1 knockdown reduced PAK1 activation, we hypothesized that AIF-1 expression might play a role in endothelial migration and organization leading to tube formation. For these experiments, 1) ECc stably transfected with either control vector, or siRNA, or were seeded onto growth factor-reduced matrigel, in the presence of 10% FCS; 2)the number of tube-like structures were counted manually per multiple representative images; and 3) an average was calculated for each condition 16 h later. Some samples of siRNA ECs were infected with AdAIF-1 or AdGFP to restore AIF-1 expression (Fig. 7). This experiment gave a somewhat unexpected result in that while inhibition of AIF-1 expression did not adversely effect formation of tube-like structures, exogenous expression of AIF-1 by adenoviral delivery can significantly enhance formation of these structures (23.1 ± 1.3 vs. 11.1 ± 3.1, 14.7 ± 1.5, and 11.2 ± 0.8 for AdAIF-1 rescue, empty pShuttle vector, siRNA, and AdGFP rescue, respectively, P < 0.01).
Angiogenesis occurs when ECs sprout from preexisting vessels to form new structures. Because AIF-1 expression effected tube-like structure formation, we tested to determine whether AIF-1 could also regulate EC sprouting ex vivo from aortic rings (36). For these experiments, mouse thoracic aorta was sectioned into 1-mm rings, incubated with AdGFP, AdsiRNA, or AdAIF-1 for 1 h, and then cultured in triplicate in growth factor-reduced matrigel. Sprouting from rings were analyzed daily, and on the sixth day the rings were photographed, and outgrowth area was quantitated. Figure 8 corroborates the EC tube formation experiments in that while inhibition of AIF-1 expression did not reduce sprouting, overexpression of AIF-1 did significantly increase the area of sprouting (7,917 ± 386, 3,018 ± 442, and 3,601 ± 479 μm2 for AdAIF-1, AdGFP, and AdsiRNA, at 6 days, respectively, P < 0.01). Qualitatively, the barbarization of EC networks emanating from aortic rings were also more complex in the AdAIF-1-infected rings than either AdGFP or AdsiRNA. Together, these data suggest that while AIF-1 deficiency does not impair angiogenic capacity of ECs, increased amounts of AIF-1 can enhance the angiogenic potential of ECs.
EC migration and proliferation have an important role in numerous physiological responses such as angiogenesis, the vascular response to interventional procedures such as angioplasty and stent placement, wound healing, and embryogenesis (1, 9, 13, 17, 28). Previous work by our laboratory has demonstrated the expression AIF-1 protein in human arteries with CAV and in neointimal cells in animal models of various types of arterial injuries (2, 5). AIF-1 has also been described to play a role in macrophage, T-lymphocyte, and VSMC activation (4, 12, 32, 34). ECs, macrophages, T-lymphocytes, and VSMC all participate in atherogenesis and the vascular response to injury, but the functional characterization of AIF-1 expression in ECs has not been determined. In this study, immunohistochemistry of human coronary arteries with CAV indicated AIF-1 expression in ECs. Because CAV is initiated and propagated by the expression and presence of multiple cytokines, we performed experiments to investigate differential expression of AIF-1 by cytokines in ECs. Noticeable differences in AIF-1 inducible expression among ECs, VSMC, and macrophages can be observed. In both ECs and VSMC, complex stimuli such as FBS and T-lymphocyte-conditioned media elicited the most robust expression of AIF-1. In contrast, this pattern of expression in ECs differs from VSMC, in which VSMC growth factors platelet-derived growth factor (PDGF) and bFGF, can induce significant AIF-1 expression, but EC growth factors, such as VEGF and bFGF, are not capable of inducing AIF-1 to a significant degree. Furthermore, low, but basal, levels of AIF-1 can be detected in cultured BAEC, whereas little to no AIF-1 can be detected in VSMC (5). In macrophages, high-level AIF-1 expression is constitutive, but significantly more robust expression can be induced by oxidized LDL (32). This report is the first to demonstrate AIF-1 expression in ECs of inflamed arteries and can be induced by soluble stimuli in cultured ECs. It also suggests tissue-specific differences in basal and inducible AIF-1 expression in VSMC, macrophages, and ECs.
Inducible AIF-1 expression in ECs suggested an important role for this protein in EC activation. Using bovine ECs as a model system, our approach was to silence AIF-1 expression in these cells by constitutive siRNA expression, and analysis of important EC physiological indexes such as proliferation, migration, and tube-like formation. For these experiments, vector-based delivery, rather than transfection of oligonucleotides, was used because it has been found to be more effective than synthetic siRNA for inhibition of gene expression, likely due to increased stability (8). Use of an antibiotic selectable vector also allowed establishment of stable cell lines expressing consistent and reproducible amounts of siRNA to reduce experimental variability. Abrogation of AIF-1 expression significantly inhibited EC proliferation and migration. It was interesting in that VEGF did not induce increased AIF-1 expression, but basal levels of AIF-1 were necessary for VEGF-induced EC migration. Moreover, when AIF-1 expression in the siRNA expressing ECs was rescued with AdAIF-1, these capacities of ECs were not only restored, but enhanced significantly, greater than control cell levels, suggesting a strong association between AIF-1 expression and EC activation.
We previously demonstrated that AIF-1 participates in signaling cascades important in macrophage and smooth muscle cell activation (31, 32). Together, these characteristics of AIF-1 suggested that abrogation of AIF-1 protein expression might interrupt signaling cascades important in EC activation. It is known that activation of mitogen-activated protein (MAP) kinases plays a critical role in regulation of EC activation, including proliferation and migration in response to vessel injury (29, 6). p44/42 in particular is strongly activated by various growth factors and agents that stimulate cell growth and thus an important effector for cell proliferation and migration (15). Because of its prominence in EC activation events, we focused on p44/42 kinase and found that reduction of AIF-1 expression significantly inhibited the FBS-stimulated activation of p44/42, suggesting that decreased growth capacity in ECs with AIF-1 abrogation may be attributed, at least in part, to the reduction in the signaling pathways involved in p44/42 activation. We also examined activation of PAK1 because PAK1 is a direct effector of Rac1 (16), and we have shown AIF-1 can activate Rac1 in human VSMC (3). In this study, abrogation of AIF-1 expression can significantly reduce serum-stimulated PAK1 activation, but interestingly, no significant inhibition of several other kinases was noted, suggesting a specificity of pathways mediated by AIF-1 in ECs in response to inflammatory stimuli. In tracheal smooth muscle cells, PAK1 activation is proximal to p38 activation (11). In macrophages, AIF-1 attenuation reduced activation of p38, AKT, and p90RSK kinases (32). In murine VSMC, chronic overexpression of AIF-1 increases p38 MAPK activation (31). However, in this study in ECs, abrogation of AIF-1 did not reduce nor did overexpression increase p38 MAPK activation. No inhibition of any of this kinase in ECs again illustrates an interesting distinction between AIF-1 expression and function in macrophages, VSMC, and EC. AIF-1 has no catalytic or protein kinase domains, yet, composite data from several cell types suggests that in the absence of AIF-1, signaling cascades are interrupted, resulting in suppressed activation of kinases distal from AIF-1. This implies that AIF-1 may function as an injury responsive scaffold protein in ECs.
PAK1 is involved in cell cytoskeletal dynamics, locomotion, migration, and angiogenic tube formation. Dominant-negative PAK1 and a peptide corresponding to an adapter protein binding site on PAK1 also inhibited angiogenesis (22). It has also been shown that p44/42 MAPK kinase inhibitors can block endothelial tube formation (24, 10). Together, these studies provided impetus to explore the role of AIF-1 in angiogenesis assays. Angiogenesis is a complex, and as yet, incompletely understood process that involves matrix degradation and orientation of ECs into new vessels. In contrast to the proliferation and migration studies, AIF-1 diminution did not reduce formation of tube-like structures on growth factor-reduced matrigel nor did it reduce microvascular outgrowths in the aortic ring assay. Yet, overexpression of AIF-1 could enhance both tube formation and vessel outgrowth from isolated rings. The most direct explanation is that both tube formation and aortic sprouting are complex processes that require more than solely migration. This suggests that while AIF-1 is not absolutely necessary for vasculogenesis, but its presence makes these processes more efficient.
Expression of JAMs, particularly JAM-A, is increased and redistributed on the surface of ECs in inflammatory conditions (34). Redistribution of JAM-A in particular, facilitates signal transduction initiated by angiogenic signals to promote EC migration and angiogenesis (34, 26). Considering previous studies involving AIF-1 expression and vascular inflammation, and present studies linking AIF-1 expression and signal transduction, migration, and angiogenesis, we investigated whether modified AIF-1 expression could modulate JAM-A expression or redistribution. Interestingly, we found no change in JAM-A mRNA or protein expression, or JAM-A cellular distribution, in ECs in which AIF-1 was either overexpressed or knocked down when compared with control ECs. Even when bFGF was added to ECs, the change in JAM-A distribution was similar in control and in ECs in which AIF-1 expression was modulated. This suggest no relationship between AIF-1 activity and JAM-A function in ECs. Because JAMs function in cell-cell interactions, this also implies that AIF-1 may not play a role in extracellular interactions as well.
The functional consequences of AIF-1 expression in ECs has previously not been explored, and thus there are several novel points in this study. First, AIF-1 is detected in ECs within the intima of inflamed human arteries, and its expression can be induced in cultured ECs by inflammatory and angiogenic factors. Second, knock down of AIF-1 protein by stable transfection of siRNA reduces the several indexes of EC pathophysiology, including proliferation and migration. These functions could be rescued by exogenous expression of AIF-1. Third, signal transduction cascades could be reduced by AIF-1 abrogation. Fourth, though angiogenesis assays were not negatively effected by reduction of AIF-1, angiogenic potential of ECs was enhanced by AIF-1 overexpression. An additional important point is that the cytokine induction of AIF-1 in ECs is similar to, but not identical to, that of VSMC and macrophage. Together, these data suggest an important function of AIF-1 in inflammation-driven EC activation makes some distinction between AIF-1 function in ECs and other cell types and affords further insight into the mechanisms by which AIF-1 participates in the pathogenesis of vascular diseases.
NOTE ADDED IN PROOF
The lane orders in the final-published version of Figs. 2A and 3A and the total p44/42 lanes in Fig. 5B are different from the Articles in PresS version of this manuscript. We regret these errors. Also, the legend of Fig. 4B now states that this figure is a composite of nine different images.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-63810 and American Heart Association Grant 0455562U to M. V. Autieri. Y. Tian is the recipient of American Heart Association Predoctoral Fellowship 0515320U.
The authors acknowledge the expert technical assistance of Yuko Nakashima.
↵* Y. Tian and S. Jain contributed equally to this report.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society