Reoxygenation of Hypoxic Human Umbilical Vein Endothelial Cells Activates the Classic Complement Pathway
Background Ischemia-reperfusion injury leads to the activation and endothelial deposition of complement. We investigated whether exposure of human umbilical vein endothelial cells (HUVECs) to hypoxia and/or reoxygenation activates complement and decreases HUVEC-surface expression of the C3 regulatory proteins CD46 and CD55.
Methods and Results HUVECs were subjected to 0, 12, or 24 hours of hypoxia (O2=1%) and then reoxygenated for 3 hours (O2=21%) in the presence of 30% human serum. C3 deposition and HUVEC-surface expression of CD46 and CD55 were evaluated by ELISA and flow cytometry. C3 deposition on HUVECs subjected to 12 or 24 hours of hypoxia followed by 3 hours of reoxygenation was significantly greater than normoxic HUVECs. Inhibition of the classic but not the alternative complement pathway during reoxygenation attenuated C3 deposition. Western blot analysis of HUVEC lysates under reducing conditions demonstrated significantly increased iC3b deposition in hypoxic/reoxygenated HUVECs compared with normoxic HUVECs. FACS analysis confirmed iC3b deposition. HUVEC-surface expression of CD46 and CD55 increased after hypoxia and/or reoxygenation.
Conclusions We conclude that (1) hypoxia and reoxygenation of HUVECs significantly increases iC3b deposition on HUVECs, (2) C3 deposition after hypoxia and reoxygenation is largely mediated by the classic complement pathway, and (3) HUVEC-surface expression of CD46 and CD55 increases after hypoxia and reoxygenation. These data demonstrate that hypoxia and reoxygenation of human endothelial cells activates the classic complement pathway despite an increase in complement C3 regulatory proteins.
Ischemia-reperfusion injury leads to the activation and deposition of complement on the vascular endothelium.1 2 3 4 5 6 Experimental models of acute MI3 7 8 9 10 11 and human autopsy specimens taken from acute MI patients12 13 14 15 16 demonstrate that complement is selectively activated in areas of infarction only. Although inhibition of complement activation or depletion of complement proteins before reperfusion has been shown to reduce tissue injury,1 9 17 the mechanisms regulating complement activation at the endothelial cell surface during ischemia and reperfusion are unknown.
Complement is a cytotoxic host defense system composed of ≈20 intravascular plasma proteins subdivided into two cascade systems, the classic and alternative complement pathways.18 Complement activation and deposition in vivo are tightly regulated by both plasma and membrane-bound complement regulators. In particular, human endothelial cells express several membrane-bound complement regulatory proteins, including MCP (CD46), DAF (CD55), and protectin (CD59).19 DAF is a 70-kD glycoprotein that accelerates the decay of the classic and alternative C3 and C5 convertases. MCP is a 45- to 70-kD protein that binds C3b and C4b and possesses factor I–dependent cofactor activity for these two components. Whereas DAF and MCP inhibit complement activation at the level of C3, protectin is a 20-kD glycoprotein that interacts with both C8 and C9 during the assembly of C5b-9 at the membrane surface to inhibit formation of the membrane-inserted C9 homopolymer responsible for C5b-9 cytolytic activity. Despite endothelial cell-surface expression of these complement regulatory proteins, ischemia-reperfusion injury activates and deposits complement on ischemic tissues.
A recently published study using C3 and C4 knockout mice has shown an important role of the classic complement pathway in an in vivo model of hindlimb ischemia.20 The authors suggested that hypoxia and reoxygenation of endothelial cells are responsible for complement activation and the resulting injury in this model.20 Another study previously showed a decrease in membrane-bound complement regulatory proteins after ischemia/reperfusion injury.15 In the present study, we investigated whether hypoxia and/or reoxygenation increases endothelial C3 deposition and decreases HUVEC-surface expression of the C3 regulatory proteins DAF and MCP in a novel in vitro model. We demonstrate that hypoxia followed by reoxygenation activates the classic complement pathway and significantly increases iC3b deposition on HUVECs. Further, we demonstrate that cell-surface expression of MCP and DAF increases after hypoxia and reoxygenation. These data suggest that endothelial deposition of C3 in the setting of ischemia-reperfusion injury is augmented by reoxygenation and is not due to a loss of surface expression of the C3 complement regulatory proteins MCP and DAF.
HUVECs were obtained as previously described.21 22 Briefly, HUVECs were harvested by use of 0.1% collagenase (Worthington Biochemical Corp) and suspended in medium 199 containing 20% heat-inactivated BCS (Gibco Life Technologies Inc). The cells were initially seeded in either 75-cm2 flasks or 100-mm Petri dishes (Corning Costar) and incubated at 37°C in 95% air/5% CO2. When confluent, the endothelial cells were passaged with 0.5% trypsin-EDTA. Endothelial cell purity was assessed by a phase-microscopic “cobblestone appearance” and uptake of fluorescent acetylated LDL. All experiments were conducted on HUVECs during passages 1 through 3.
Cell-Surface ELISA Experiments
A C3-specific cell-surface ELISA was developed with a peroxidase-conjugated polyclonal goat anti–human C3 antibody (Cappel). HUVECs were grown to confluence on 0.1% gelatinized 96-well plastic plates (Corning Costar). The plates were then divided into the following groups: (1) normoxia (control), (2) hypoxia (12 hours), and (3) hypoxia (24 hours). Nontoxic hypoxic stress (O2=1%) was maintained in a humidified sealed chamber (Coy Laboratory Products, Inc) at 37°C gassed with 1% O2/5% CO2/balance N2 as we have shown previously.23 After the specified period of normoxia or hypoxia, the cell medium was aspirated and 100 μL of one the following was added to each well: (1) 30% HS, (2) HBSS, (3) 30% HS+100 mmol/L MgCl2/EGTA, (4) 100 mmol/L MgCl2/EGTA, (5) C2-depleted HS, (6) C2-depleted HS+245 μmol/L C2, or (7) factor B–depleted HS. The cells were then reoxygenated for 3 hours at 37°C in 95% air/5% CO2 or incubated an additional 3 hours in the hypoxia chamber. The cells were washed two times in an automated plate washer (Tri-Continental Scientific) and then fixed with 1% paraformaldehyde (Sigma Chemical Co) for 30 minutes. The cells were washed three times and incubated at 4°C for 1.5 hours with 50 μL of peroxidase-conjugated polyclonal goat anti–human C3 antibody (1:1000 dilution). After the cells were washed three times, the plates were developed with 50 μL of ABTS and read (Molecular Devices) at 405 nm. Background controls consisted of cells to which only the anti–human C3 antibody was added (ie, no HS). Background optical density was subtracted from all groups. All ELISA experiments were performed two or three times with six wells used per experimental group (n=12 to 18).
HUVECs were grown to confluence in 60-mm Petri dishes coated with gelatin. Cell-surface C3 deposition was measured by flow cytometry in normoxic HUVECs and HUVECs subjected to 24 hours of hypoxia followed by 3 hours of reoxygenation in the presence of 30% HS. After the cells were washed and fixed in 1% paraformaldehyde, they were incubated with FITC-conjugated goat anti–human C3 antibody (Cappel) for 30 minutes at 4°C.
In additional studies, a monoclonal antibody to a neoepitope on iC3b (Quidel) was used. An isotype control monoclonal antibody to porcine C5a was used in these studies. The monoclonal antibodies were identified with a FITC-labeled goat anti-mouse Ig F(ab′)2 (Jackson ImmunoResearch). C3 and iC3b deposition on HUVECs was measured with a FACSort flow cytometer (Becton Dickinson). All flow cytometry experiments were performed in duplicate.
Confluent HUVEC cultures grown in 96-well plates were incubated under normoxic or hypoxic conditions for 24 hours. The cell medium was then aspirated, and 30% HS was added to each plate. The cells were then allowed to reoxygenate for 3 hours at 37°C in 95% air/5% CO2 . The HUVECs were washed five times in an automated plate washer and solubilized with ice-cold lysis buffer (1% Nonidet-P40, 0.1% SDS, 3 mmol/L EDTA, 2 mmol/L PMSF, 3 μmol/L aprotinin, 29 μmol/L pepstatin, and 37 μmol/L leupeptin in PBS, pH 7.4). To demonstrate specific binding, HUVECs were suspended in 10 mmol/L EDTA/PBS buffer. Then the cells were incubated in a high-ionic-strength buffer (1 mol/L NaCl, 10 mmol/L Tris, pH 7.5) for 10 minutes. The cells were pelleted and suspended in a low-ionic-strength buffer (10 mmol/L Tris, 1 mmol/L EDTA, pH 7.4, with the same protease inhibitors) for 10 minutes on ice. The cells were then pelleted and suspended in the lysis buffer. The Western blot profile of these normoxic and hypoxic/reoxygenated samples did not differ from cells washed without high-stringency buffers (data not shown).
HUVEC lysates (6 μg protein/lane) were resolved by SDS-PAGE (9%) under reduced conditions, transferred to nitrocellulose membranes (BioRad), and blocked with 10% nonfat dry milk in PBS buffer containing 0.1% Tween 20 and 0.1% BSA. The membranes were then incubated with peroxidase-conjugated goat IgG fraction to human complement C3 (Cappel) and/or rabbit anti–human C3d (Advance Research Technologies) for 1 hour at 20°C. A peroxidase-conjugated anti-rabbit IgG (1 hour at 20°C; Sigma) was used to detect the anti-C3d antibodies. The ECL system (Amersham International) was used to develop the Western blots. Purified C3, C3b, iC3b, and C3c standards were obtained from Advance Research Technologies. The Western blot experiments were performed at least three times. Densitometry from resulting bands was quantified from scanned images by use of NIH Image software.
Flow Cytometry and ELISA Analysis of MCP and DAF
Cell-surface expression of MCP and DAF was measured by flow cytometry and ELISA on normoxic HUVECs and HUVECs subjected to 24 hours of hypoxia followed by 3 hours of reoxygenation. After the cells were fixed with 1% paraformaldehyde, MCP and DAF were measured with polyclonal rabbit anti–human MCP and anti–human DAF antibodies (a gift from Dr B.P. Morgan, University of Wales College of Medicine, Cardiff, UK). A peroxidase-conjugated goat anti-rabbit secondary antibody (Cappel) was used for detection of the primary antibodies by ELISA. An FITC-conjugated donkey anti-rabbit F(ab′)2 (Jackson Immunoresearch) was used for detection of the primary antibodies by flow cytometry. Background controls for ELISA and flow cytometry consisted of cells to which a polyclonal rabbit anti-horse IgG (Jackson Immunoresearch) was added (species-matched inappropriate primary antibody). Background optical density was subtracted from all ELISA groups.
Immunoprecipitation of HUVEC MCP and DAF
To confirm the specificity of the anti–human MCP and DAF antibodies, confluent HUVEC cultures grown in 60-mm Petri dishes were washed with ice-cold HBSS and labeled with 1 mmol/L biotin (Immuno Pure Sulfo-NHS-Biotin) as previously described.24 Unbound biotin was quenched with 50 mmol/L NH4Cl in HBSS buffer. The HUVECs were then incubated with lysing buffer (150 mmol/L NaCl, 25 mmol/L Tris, 1 mmol/L MgCL2, 1% Triton X-100, 1% Nonidet P-40, 5 mmol/L EDTA, 5 μg/mL chymostatin, 2 μg/mL aprotinin, and 1.25 mmol/L PMSF, all from Sigma). Cell debris was removed by centrifugation (10 000g, 5 minutes). Cell lysates were precleared with 50 μg/mL preequilibrated protein G–sepharose (Pharmacia). Immunoprecipitation of MCP and DAF was performed by addition of polyclonal rabbit anti–human MCP and anti–human DAF antibodies. A polyclonal rabbit anti–horse IgG antibody (Jackson Immunoresearch) was also added as a species-matched inappropriate primary antibody control. Washed immunoprecipitates were boiled in nonreducing sample buffer (2.5% SDS, 0.38 mol/L Tris, pH 6.8, 20% glycerol, and 0.1% bromphenol blue), separated by SDS-PAGE (10% linear gel) under nonreducing conditions, and transferred to nitrocellulose by standard protocols. Biotinylated proteins were labeled with streptavidin-peroxidase (Pierce) and visualized by enhanced chemiluminescence.
All data are expressed as mean±SEM. Data analyses were performed with Sigma Stat (Jandel Scientific). Endothelial C3 deposition and HUVEC-surface expression of MCP and DAF in normoxic versus hypoxic HUVECs (ELISA) was analyzed by two-way ANOVA. All pairwise multiple comparisons were made with the Student-Newman-Keuls test. The reduced Western blots were scanned and then analyzed by a nonpaired Student’s t test. Optical density for all ELISA data are presented with background subtracted.
Effect of Hypoxia Versus Hypoxia and Reoxygenation
Cell-surface C3 deposition was measured by ELISA in normoxic HUVECs and HUVECs subjected to either hypoxia alone or hypoxia followed by reoxygenation. C3 deposition on HUVECs subjected to 12 or 24 hours of hypoxia followed by 3 hours of reoxygenation was significantly greater (P<.05) than normoxic HUVECs (with or without reoxygenation) or HUVECs subjected to only 12 or 24 hours of hypoxia (Fig 1⇓). C3 deposition on HUVECs subjected to only 12 or 24 hours of hypoxia did not differ significantly from normoxic HUVECs (with or without reoxygenation). C3 deposition after hypoxia and reoxygenation was also confirmed by flow cytometry (Fig 2⇓). MFI in HUVECs subjected to 24 hours of hypoxia was greater than that in normoxic HUVECs after 3 hours of incubation in 30% HS. C3 deposition (ie, MFI) was augmented in HUVECs subjected to 24 hours of hypoxia followed by reoxygenation (3 hours) in 30% HS. MFI in HUVECs not exposed to HS and subjected to normoxia, hypoxia, or hypoxia and reoxygenation did not differ.
Inhibition of the Classic Complement Pathway
To investigate whether C3 deposition involved activation of the classic complement pathway, human serum containing MgCl2/EGTA was added to inhibit the classic complement pathway. C3 deposition on HUVECs subjected to 24 hours of hypoxia followed by 3 hours of reoxygenation was significantly greater (P<.05) than normoxic (ie, control) HUVECs (0.22±0.03 versus 0.08±0.02 OD405, respectively). Treatment of HS with MgCl2/EGTA significantly inhibited C3 deposition (normoxic cells, 0.03±0.01 versus hypoxic/reoxygenated cells, 0.02±0.01; P<.05 compared with HUVECs not treated with MgCl2/EGTA). Control HUVECs receiving MgCl2/EGTA and no HS did not deposit C3.
Inhibition of the Classic and Alternative Complement Pathways With Complement-Depleted Sera
To further investigate the mechanism of C3 deposition, HUVECs were reoxygenated in the presence of C2-depleted HS or factor B–depleted HS to inhibit the classic and alternative complement pathway activity, respectively. C3 deposition on HUVECs subjected to 24 hours of hypoxia followed by 3 hours of reoxygenation in the presence of HS or factor B–depleted HS was significantly greater (P<.05) than normoxic HUVECs (Fig 3⇓). However, C3 deposition on HUVECs reoxygenated in the presence of C2-depleted HS did not differ significantly from normoxic HUVECs. When C2 was added back to the C2-depleted HS, C3 deposition on HUVECs after 24 hours of hypoxia and 3 hours of reoxygenation was significantly greater (P<.05) than normoxic HUVECs and did not differ significantly from HUVECs receiving HS. These data demonstrate that C3 deposition is specific and not a result of nonspecific C3 binding.
Western Blot Analysis
Western blotting was performed under reducing conditions using purified human C3 (lane 1) and C3b (lane 2) as standards and probed with the polyclonal anti–human C3 antibody (Fig 4A⇓). This antibody recognized the α-chain of C3 (α) and C3b (α1) but did not recognize an “α-chain” of iC3b (data not shown). This antibody recognized the β-chain of C3, C3b, and iC3b (data not shown). HUVECs not treated with 30% HS failed to show C3 deposition. Normoxic (lane 3) and hypoxic/reoxygenated (lane 4) HUVEC lysates revealed a C3 β-chain, but no α-chain was observed. Scanning revealed a significant increase (55±16%; P<.05; n=3) in β-chain deposition in the hypoxic/reoxygenated lysates compared with normoxic cells.
To further characterize the C3 species present on the HUVECs, we performed an additional Western blot that was probed with a polyclonal antibody against C3d. As shown in Fig 4B⇑ (left), this antibody recognized the C3 α-chain of purified iC3b (α2; lane 1). The anti-C3d antibody also recognized a band of molecular weight similar to that of the iC3b α2-chain in the normoxic (lane 2) and hypoxic/reoxygenated (lane 3) cell lysates. Similar to the β-chain, a significant increase in the α2 band density was observed in the hypoxic/reoxygenated HUVEC lysates compared with normoxic lysates. Another Western blot of normoxic or hypoxic/reoxygenated HUVEC lysates was probed with anti-C3d and anti-C3 antibodies. Fig 4B⇑ (right) demonstrates the position of the α2 band in relationship to the C3 β-chain in this Western blot. The α2 band ran at a slightly lower molecular weight than the C3 β-chain. This Western blot also demonstrates a significant increase in deposition of the α2 band and the β-chain in hypoxic/reoxygenated HUVECs (Fig 4B⇑, right, lane 2) compared with the normoxic HUVECs (Fig 4B⇑, right, lane 1).
iC3b deposition after hypoxia and reoxygenation was then confirmed by flow cytometry using a monoclonal antibody to a neoepitope on iC3b (Quidel). MFI was increased on normoxic cells exposed to 30% HS compared with normoxic cells not exposed to 30% HS (430±12 versus 275±7 arbitrary fluorescence units, respectively). MFI on HUVECs subjected to 24 hours of hypoxia was greater than that of normoxic HUVECs after 3 hours of incubation in 30% HS (584±30 versus 430±12 arbitrary fluorescence units, respectively). iC3b deposition (ie, MFI) was augmented in HUVECs subjected to 24 hours of hypoxia followed by reoxygenation (3 hours) in 30% HS compared with hypoxia alone (806±15 versus 584±30 arbitrary fluorescence units, respectively). MFI in HUVECs not exposed to HS and subjected to normoxia or hypoxia and reoxygenation did not differ (275±7 versus 254±5 arbitrary fluorescence units, respectively). These flow cytometric data using a monoclonal antibody to iC3b are similar to the flow cytometric results obtained with the polyclonal antibody to C3 (Fig 2⇑). However, these data demonstrate that the C3 species attached to the HUVECs is iC3b.
HUVEC-Surface Expression of MCP and DAF
Cell-surface expression of MCP and DAF after hypoxia and reoxygenation was investigated by ELISA (Fig 5⇓) and flow cytometry (Fig 6⇓) to investigate whether the increased iC3b deposition was a result of a decrease in surface expression of these C3 complement regulators. ELISA demonstrated that cell-surface expression of DAF and MCP in hypoxic/reoxygenated HUVECs was significantly increased compared with normoxic HUVECs (P<.05). Flow cytometry demonstrated an increase in DAF and MCP surface expression after hypoxia that was further augmented with reoxygenation.
Immunoprecipitation of MCP and DAF
To confirm the specificity of the polyclonal anti-human DAF and MCP antibodies used in these studies, both MCP and DAF were immunoprecipitated from normoxic HUVEC lysates (Fig 7⇓). Western blots of the immunoprecipitates revealed 55-kD and 70-kD bands consistent with the known molecular weights of MCP and DAF, respectively. Furthermore, no bands of similar molecular weight were observed when HUVEC lysates were immunoprecipitated with a species-matched inappropriate primary antibody.
We have shown that human endothelial cells after hypoxia induced complement activation demonstrated by increased iC3b deposition. C3 deposition on hypoxic HUVECs was augmented by reoxygenation. Complement activation in this model is largely dependent on activation of the classic complement pathway. We also demonstrate that HUVEC-surface expression of the C3 regulators MCP and DAF increases after hypoxia and reoxygenation. We conclude that hypoxia and reoxygenation of human endothelial cells activates the classic complement pathway despite an increase in surface expression of complement regulatory proteins.
Complement Activation During Ischemia/Reperfusion
Since complement concentrations are highest in the plasma, one might expect that complement activation first takes place intravascularly on the endothelium and then proceeds extravascularly during ischemia/ reperfusion. Weisman and colleagues9 have shown that C5b-9 is deposited only on the coronary endothelium during the early phase of reperfusion. Buerke and colleagues17 recently demonstrated that C1q is deposited on the coronary endothelium after myocardial ischemia and reperfusion in cats. A recent in vivo study by Weiser and colleagues20 suggested that ischemia and reperfusion of endothelial cells may activate complement by expression of neoantigens at the endothelial cell surface. Thus, previous studies have demonstrated that complement is deposited on the vascular endothelium. Our study demonstrates not only that complement is deposited on endothelial cells but also that hypoxic/reoxygenated endothelial cells become complement activators.
Complement is known to be activated during ischemia and reperfusion.1 2 3 4 5 6 However, the mechanisms regulating complement activation during ischemia/reperfusion are not understood. It is possible that the complement regulatory proteins MCP, DAF, and protectin are shed from the endothelial surface or become inactivated during reperfusion. This is particularly important because DAF and protectin are glycosylphosphatidyl-inositol–linked proteins, and phospholipases are known to be activated during ischemia and reperfusion.25 We demonstrate that hypoxia and reoxygenation significantly increase HUVEC-surface expression of DAF and MCP in the present study. Furthermore, we recently identified the presence of another complement regulatory protein on HUVECs, complement receptor one (CR1; CD35).26 HUVEC-surface expression of CR1 is also increased after hypoxia.26 Thus, complement activation and C3 deposition occur in this model despite increased surface expression of these C3 complement regulatory proteins. It should be noted that although MCP, DAF, and protectin are known to be present on human endothelial cells,25 27 only protectin is present on cardiomyocytes.16 Therefore, myocytes may be even more vulnerable to complement-mediated injury than the endothelium.
We have demonstrated that reoxygenation augments complement deposition in this model. Interestingly, studies of acute MI lesions demonstrate increased myocardial deposition of complement in reperfused MI lesions compared with nonreperfused MI lesions.10 Oxygen-derived free radicals have been shown to activate the terminal complement cascade by converting C5 to a functionally active C5b-like metabolite.28 In addition, oxygen-derived free radicals have been shown to mediate endothelial cell damage by complement-stimulated granulocytes.29 30 However, direct activation of C3 or the classic complement pathway by oxygen-derived free radicals has not been demonstrated. Because the increase in C3 deposition on HUVECs is augmented by reoxygenation, we are currently investigating whether the formation of oxygen-derived free radicals during reoxygenation activates complement in this model.
The importance of complement activation after reoxygenation of hypoxic HUVECs is highlighted by experimental studies in which inhibition of complement activation, complement depletion, or functional inhibition of C5a during reperfusion significantly decreases infarct size.8 9 31 32 However, the initiating step and site of complement activation in vivo during ischemia and reperfusion is not known. Myocardial deposition of the classic complement pathway components C1q, C4, and C3 has been demonstrated, whereas myocardial deposition of properdin (a specific marker for alternative complement pathway activation) was not observed.8 16 17 Similarly, Rossen and colleagues33 demonstrated that mitochondrial membrane and cardiolipin-containing fragments released from ischemic tissues can bind C1q and activate the classic complement pathway. Thus, complement can be activated by myocardial fragments and deposited within the myocardium. The mechanism of complement activation in our model is presently under investigation.
Characterization of C3 Deposition
Our data demonstrate that C3 is deposited on endothelium after hypoxia and reoxygenation. C3, the most abundant complement component, is composed of an α-chain (with an internal disulfide bond) linked to a β-chain by a disulfide bond. In the presence of the alternative or classic C3 convertase, C3 is cleaved into the anaphylatoxin C3a and C3b. The C3 activation step exposes a thioester bond in the C3d region of the C3 α-chain. The thioester bond is very unstable and allows C3b to attach covalently to cell membranes. This is an important step, because C3b deposition to a cell surface allows initiation of the terminal complement complex, C5b-9, and can serve to amplify the alternative complement pathway. In the presence of the fluid-phase factor I and one of several cofactors (MCP, CR1, CR2, factor H, or C4bp), C3b is degraded to iC3b. iC3b is further degraded to the fluid-phase complement component C3c and the membrane-bound C3dg molecule.
Under reducing conditions, iC3b liberates three separate peptides: a β-chain (75.5 kD), an α2-chain (63 kD), and another 39.5-kD peptide from the α-chain. We did not observe a “C3” α-chain in the reduced cell lysates or from purified iC3b antigen on Western blots using the polyclonal anti–human C3 antibody (Fig 4A⇑). These data suggest that the C3 species present on the HUVECs is not a result of nonspecific C3 binding or C3b but rather another C3 species. We did observe a significant increase in the β-chain density in hypoxic HUVEC lysates compared with normoxic lysates. To identify the C3 species present on the HUVECs, we used a polyclonal anti-C3d antibody. The C3d antibody identified a band, α2, in normoxic and hypoxic/reoxygenated lysates under reducing conditions with a molecular weight equivalent to the reduced iC3b α-chain (Fig 4B⇑, left). The α2 band density was significantly greater in the hypoxic/reoxygenated lysates than with normoxia. Dual probing of a Western blot with the polyclonal anti-C3 and anti-C3d antibodies allowed us to identify the iC3b α-chain (α2) and the β-chain on the same Western blot (Fig 4B⇑, right). We further documented that the C3 species on the HUVECs was iC3b by additional flow cytometry experiments using a monoclonal antibody to a neoepitope on iC3b. These flow cytometry data confirmed our Western blot data demonstrating significantly enhanced iC3b under hypoxic conditions compared with normoxia. Reoxygenation of the hypoxic HUVECs augmented iC3b deposition. These data demonstrate that (1) the polyclonal antibody to human C3 does not recognize the iC3b α-chain under reduced conditions and (2) iC3b but not C3 or C3b deposition is significantly increased on hypoxic/reoxygenated HUVECs.
Complement activation leads to the production of several important biologically active components, namely, C5a, iC3b, and C5b-9. Endothelial deposition of iC3b has been shown to be a potent stimulus for increased adhesion of neutrophils.34 35 36 Similarly, C5b-9 and C5a induce endothelial expression of the neutrophil adhesion molecule P-selectin.37 38 Our study demonstrates that iC3b is the C3 complement component deposited in this model. Thus, activation of complement may initiate the recruitment and endothelial attachment of neutrophils in ischemic/reperfused tissues. Future studies observing the increased adherence of neutrophils in this model are warranted.
Limitations to the Model/Study
We recognize that this study is an in vitro model using hypoxia/reoxygenation. Hypoxia/reoxygenation is only one variable among many present during ischemia/reperfusion and may not accurately represent what happens in vivo. The period of hypoxia (12 to 24 hours) used in this study decreases the Po2 of HUVEC culture media from ≈160 to 16 to 24 mm Hg (unpublished observations). Although HUVECs in vivo normally experience an environmental Po2 of ≈39 mm Hg, HUVECs are grown in vitro at a Po2 of 160 mm Hg. Thus, the decrease in cell medium Po2 from 160 mm Hg to ≈20 mm Hg is a true hypoxic condition for HUVECs grown in vitro. Second, this study uses HUVECs, which may not represent other vascular endothelial cells (ie, microvascular endothelium). However, all studies were performed onHUVECs during passages 1 through 3, and these cells did express complement regulatory molecules. Future studies using arterial endothelial cells are planned. Third, although DAF and MCP are physically present on HUVECs and are increased by hypoxia and reoxygenation, the mere presence of these complement regulators does not necessarily denote functional activity. However, functional data for these complement regulators would be difficult to obtain because of the overlapping functional properties and the lack of sufficient quantities of functionally inhibitory Fab fragments.
In summary, we conclude that (1) hypoxia leads to complement activation and deposition of C3 on HUVECs, (2) reoxygenation of hypoxic HUVECs augments C3 deposition, (3) bound C3 was characterized as iC3b by Western analysis and flow cytometry, (4) C3 deposition after hypoxia and reoxygenation is largely mediated by the classic complement pathway, and (5) HUVEC-surface expression of DAF and MCP increases after hypoxia and reoxygenation. These data demonstrate that hypoxia and/or reoxygenation of human endothelial cells activates the classic complement pathway despite increased expression of C3 regulatory proteins.
Selected Abbreviations and Acronyms
|DAF||=||decay acceleration factor|
|HUVEC||=||human umbilical vein endothelial cell|
|MCP||=||membrane cofactor protein|
|MFI||=||mean fluorescence intensity|
These studies were partially funded by a grant from the Milton Fund (Dr Stahl), HL-52886 (Dr Stahl), HL-56086 (Dr Stahl), and DK-50189 (Dr Colgan). We wish to thank Margaret M. Morrissey for assistance with HUVEC culture.
Drs Collard, Väkevä, and Büküsoglu share first authorship.
- Received October 28, 1996.
- Revision received December 17, 1996.
- Accepted January 9, 1997.
- Copyright © 1997 by American Heart Association
Kilgore KS, Friedrichs GS, Homeister JW, Lucchesi BR. The complement system in myocardial ischaemia/reperfusion injury. Cardiovasc Res. 1994;28:437-444.
Lucchesi BR. Complement activation, neutrophils, and oxygen radicals in reperfusion injury. Stroke. 1993;24:41-47.
Entman ML, Smith CW. Postreperfusion inflammation: a model for reaction to injury in cardiovascular disease. Cardiovasc Res. 1994;28:1301-1311.
Maroko PR, Carpenter CB, Chariello M. Reduction by cobra venom factor of myocardial necrosis after coronary artery occlusion. J Clin Invest. 1978;61:661-670.
Pinckard RN, O’Roarke RA, Crawford MH. Complement localization and mediation of ischemic injury in baboon myocardium. J Clin Invest. 1980;66:1050-1056.
Weisman HF, Bartow T, Leppo MK, Marsh HCJ, Carson GR, Concino MF, Boyle MP, Roux KH, Weisfeldt ML, Fearon DT. Soluble human complement receptor type 1: in vivo inhibitor of complement suppressing post-ischemic myocardial inflammation and necrosis. Science. 1990;249:146-151.
Mathey D, Schofer J, Schafer H, Hamdoch T, Joachim H, Ritgen A, Hugo F, Bhakdi S. Early accumulation of the terminal complement-complex in the ischemic myocardium after reperfusion. Eur Heart J. 1994;15:418-423.
Schafer H, Mathey D, Hugo F, Bhakdi S. Deposition of the terminal C5b-9 complement complex in infarcted areas of human myocardium. J Immunol. 1986;137:1945-1949.
Buerke M, Murohara T, Lefer AM. Cardioprotective effects of a C1 esterase inhibitor in myocardial ischemia and reperfusion. Circulation. 1995;91:393-402.
Weiser MR, Williams JP, Moore FD Jr, Kobzik L, Ma MH, Hechtman HB, Carroll MC. Reperfusion injury of ischemic skeletal muscle is mediated by natural antibody and complement. J Exp Med. 1996;183:2343-2348.
Gimbrone MA Jr, Shefton EJ, Cruise SA. Isolation and primary culture of endothelial cells from human umbilical vessels. TCA Manual. 1978;4:813-818.
Maciag T, Hover GA, Stemerman MB, Weinstein J. Serial propagation of human endothelial cells in vitro. J Cell Biol. 1981;91:420-428.
Zünd G, Nelson DP, Neufeld EJ, Dzus AL, Bischoff J, Mayer JE, Colgan SP. Hypoxia enhances stimulus-dependent induction of E-selectin on aortic endothelial cells. Proc Natl Acad Sci U S A. 1996;93:7075-7080.
LeBivic A, Francisco XR, Rodriquez-Boulan E. Vectorial targeting of apical and basolateral plasma membrane proteins in a human adenocarcinoma epithelial cell line. Proc Natl Acad Sci U S A. 1989;86:9313-9317.
Collard CD, Bukusoglu C, Colgan SP, Morrissey MA, Morgan BP, Frendl G, Stahl GL. Surface expression of complement receptor one (CR1) on human umbilical vein endothelial cells (HUVECs): modulation by endothelial hypoxia. Circulation. 1995;92(suppl I):I-38. Abstract.
Cole J, Housley GA, Dykman TR, MacDermott RP, Atkinson JP. Identification of an additional class of C3-binding membrane proteins of human peripheral blood leukocytes and cell lines. Proc Natl Acad Sci U S A. 1986;82:859-864.
Sacks T, Moldow CF, Craddock PR, Bowers TK, Jacob HS. Oxygen radicals mediate endothelial cell damage by complement-stimulated granulocytes: an in vitro model of immune vascular damage. J Clin Invest. 1978;61:1161-1167.
Amsterdam EA, Stahl GL, Pan H-L, Rendig SV, Fletcher MP, Longhurst JC. Limitation of reperfusion injury by a monoclonal antibody to C5a during myocardial infarction in pigs. Am J Physiol.. 1995;268:H448-H457.
Mulligan MS, Yeh CG, Rudolph AR, Ward PA. Protective effects of soluble CR1 in complement- and neutrophil-mediated tissue injury. J Immunol. 1992;148:1479-1485.
Rossen RD, Michael LH, Hawkins HK, Youker K, Dreyer WJ, Baughn RE, Entman ML. Cardiolipin-protein complexes and initiation of complement activation after coronary artery occlusion. Circ Res. 1994;75:546-555.
Vercellotti GM, Platt JL, Bach FH, Dalmasso AP. Neutrophil adhesion to xenogeneic endothelium via iC3b. J Immunol. 1991;146:730-734.
Tsuji S, Kaji K, Nagasawa S. Activation of the alternative pathway of human complement by apoptotic human umbilical vein endothelial cells. J Biochem (Tokyo). 1994;116:794-800.
Kilgore KS, Miller BF, Imlay MM, Warren JS. P-selectin and PAF-dependent neutrophil adhesion to endothelial cells is induced by the membrane attack complex (MAC). FASEB J. 1995;9:A35. Abstract.
Foreman KE, Vaporciyan AA, Bonish BK, Jones ML, Johnson KJ, Glovsky MM, Eddy SM, Ward PA. C5a-induced expression of P-selectin in endothelial cells. J Clin Invest. 1994;94:1147-1155.