Endothelial Cell Activation by Pore-Forming Structures
Pivotal Role for Interleukin-1α
Background—Interaction of complement with endothelial cells (ECs) underlies the development of inflammation and coagulation in disease. Assembly of the membrane attack complex (MAC) of complement on EC membrane, like stimulation with cytokines, upregulates tissue factor and cyclooxygenase-2 but does so via the intermediary action of IL-1α. We asked whether the MAC activates porcine aortic and microvascular ECs in a global manner by this mechanism and whether this mechanism is used by membrane pore-forming structures.
Methods and Results—Exposure of ECs to complement caused upregulation of mRNAs for E-selectin, intracellular adhesion molecule-1, vascular cell adhesion molecule-1, Iκ-Bα, interleukin (IL)-1α, IL-1β, IL-8, and plasminogen activator inhibitor-1 over a period of 6 hours. The expression of these genes was not a primary response to stimulation, however, because IL-1 receptor antagonist inhibited expression of these genes. Activation of ECs by complement depended on the autocrine action of IL-1α, because complement-mediated EC activation was inhibited by anti–IL-1α antibodies. Melittin and mastoparan, amphiphilic pore-forming peptides like the MAC, induced E-selectin through intermediary action of IL-1.
Conclusions—These findings suggest that transmembrane pore-forming proteins, as a class of molecules, activate ECs through the autocrine effects of IL-1α.
Vascular endothelial cells (ECs) exhibit 2 distinct physiological postures. One posture is the “quiescent” state manifested under normal physiological conditions. Quiescent ECs maintain blood and tissue homeostasis by elaborating cell-associated macromolecules, such as thrombomodulin1 and tissue factor pathway inhibitor,2 which prevent inadvertent clotting of blood. Quiescent ECs inhibit inflammation by presenting a nonadherent surface for leukocytes and by secreting prostacyclin. Quiescent ECs maintain blood flow by secreting a balance of vasodilators and vasoconstrictors.3 The second posture of ECs is the “activated” state exhibited as a response to inflammatory mediators. Activated ECs promote coagulation by expressing tissue factor4 and plasminogen activator inhibitor-1 (PAI-1).5 Activated ECs promote inflammation by expressing such molecules as E-selectin6 and interleukin (IL)-1.7 The activation of ECs is a primary response to stimulation by cytokines and is orchestrated, in part, by activation of nuclear factor (NF)-κB leading to upregulation of tissue factor, E-selectin, and IL-1 mRNAs.8
Interaction of complement anaphylatoxins, such as C5a, and terminal complement complexes with ECs triggers changes in ECs similar to those induced by cytokines. Anaphylatoxins and the membrane attack complex (MAC) may induce changes that occur within minutes and include formation of intercellular gaps,9 expression of P-selectin,10 and activation of proteinases that cleave and release heparan sulfate proteoglycan from EC surfaces.11 The MAC also induces changes that occur over a period of hours, which include upregulation of tissue factor, cyclooxygenase-2, and chemokines.12 13 14
Although the assembly of MAC in ECs causes procoagulant and inflammatory changes, like those induced by cytokines, the upregulation of tissue factor and cyclooxygenase-2 in ECs occurs much more slowly than corresponding changes induced by cytokines. For example, whereas tumor necrosis factor (TNF)-α leads to immediate upregulation of tissue factor,4 MAC upregulates tissue factor14 and Cox-213 over 6 to 8 hours, suggesting that the upregulation of these genes might not be a primary response in ECs. Indeed, MAC induces the IL-1α gene as a primary response that acts on ECs to promote expression of tissue factor and Cox-2.13 14
Whether this distinct pathway of EC activation is unique to MAC or reflects a response to the broader class of transmembrane pore-forming proteins is unknown. Here, we show that EC activation by MAC is a prototype for EC activation by transmembrane pore-forming peptides.
DMEM, l-glutamine, penicillin, streptomycin, phenol, guanidinium thiocyanate, and Moloney murine leukemia virus reverse transcriptase (RT) were from Life Technologies. Actinomycin D, diethyl pyrocarbonate, gelatin, BSA, glutaraldehyde, limulus amoebocyte lysate, and alkaline phosphatase–conjugated antibodies were from Sigma. FCS was from HyClone. Human sera immunodepleted of C8, purified human C8, and anti-iC3b antibodies were from Quidel. Deoxynucleotide triphosphates, oligo d(T)16, and AmpliTaq DNA polymerase were from Perkin Elmer Cetus. Random hexanucleotide primers were from Boehringer Mannheim Corp. RNasin was from Promega Corp. Human recombinant IL-1α and TNF-α were from Genzyme. [α-32P]dCTP was from DuPont-NEN. IL-1 receptor antagonist (IL-1ra) and anti–human IL-1α–neutralizing antibodies were from R&D Systems. Melittin and mastoparan were from Calbiochem. sCR1 was from T-Cell Sciences.
Porcine aortic ECs were isolated, characterized, and cultured as described previously.9 Porcine aortic ECs expressing human DAF and CD59 were isolated from transgenic pigs obtained from Nextran Inc.15 Porcine cardiac microvascular ECs were isolated as described.16 EC monolayers (passages 3 to 6) were grown to confluence on gelatin-coated plates and were maintained in fresh DMEM with 10% FCS for 48 to 72 hours before each experiment. The purity of EC cultures was determined on the basis of uptake of acetylated LDL and the lack of staining for α-actin.
Anti-EC Antibodies and Complement
Samples of human, porcine, and rabbit blood were used as sources of anti-EC antibodies and complement as previously described.14 17 Complement activation was determined as described.9 When endotoxin contamination was of concern, the limulus amoebocyte lysate test18 was performed. Generally, serum, purified serum components, and various components used for stimulation of ECs did not show detectable amounts of endotoxin.
Expression of E-Selectin and Tissue Factor on ECs
Expression of E-selectin on monolayers of porcine aortic and microvascular ECs was measured by ELISA using monoclonal anti–human E-selectin antibodies as described.19 Expression of tissue factor on ECs was measured by assays for tissue factor activity as described.14
Analysis of mRNA by Polymerase Chain Reaction
Total RNA from ECs was analyzed as described.14 Sequences of primers used for polymerase chain reaction (PCR) are as follows: β-actin, ATGTTTGAGACCTTCAACAC, CACGTCACACTTCATGATGGA; E-selectin, AGCAAGCCAGAGCAACACAG, TTGCCCGAGTCCTTGTCTCT; VCAM-1, GAACGAACACTCTTACCT, TTCCATCCTCATAGCAAT; ICAM-1, GGGAGGAGGAGCTGTTCA, AGGTGGGAAGCTGTAGAA; IL-1β, GAGGCTGATGGCCCCAAAGAG, CCTCCAGCTGCAGGGTGGGCG; IL-6, CATTAAGTACATCCTCGGCAA, TAGTGTCCTAACGCTCA; PAI-1, GTGTTTCAGCAGGTGG, TGAGCCATCATGGGCAC; vWF, CGCCGCTGCACGACTTCT, GCTCCAGCTCGTCCACAC; IκBα, AGCGGCTACTGGATGAC, GCCTCCAAGCACGCAGTC; and IL-1α, GCAGCCACCGGGAAGATTCTG, GTACATACGGCCTGTCAACAC. The annealing temperature was 55°C for all reactions except IL-1β and vWF, for which the temperature was 63°C. All reactions were 30 cycles except for β-actin, which was 25 cycles. The identity of all PCR products was confirmed by sequencing (not shown).
Generation of Anti–Porcine IL-1α Antibodies
Porcine IL-1α cDNA (M86730) encoding mature IL-1α was cloned in plasmid PHIL-S1 (Invitrogen). The resulting recombinant plasmid that carries 195 amino acids of putative mature IL-1α was used to transform the yeast Pichia and the recombinant IL-1α were overexpressed as suggested by the manufacturer. Recombinant porcine IL-1α was used to immunize rabbits. The rabbit polyclonal anti-IL-1α antibodies inhibited the activity of IL-1α.
Immunoprecipitation of IL-1α
ECs in 24-well plates were pulse-labeled with 100 μCi/mL of 35S-methionine (specific activity >1000 Ci/mmol) for 2 hours. The immunoprecipitation was performed as described.20 For each reaction, 50 μg rabbit anti–porcine IL-1α antibodies was used. Samples were resolved on 18% polyacrylamide gels, dried, and exposed to Kodak X-Omat AR film for 21 days.
Complement-Mediated Changes in ECs
To determine whether complement might cause global changes in ECs, characteristic of activation, ECs were exposed to a 15% concentration of a human serum containing 0.8 μg/mL anti–porcine EC antibodies and complement. Under these conditions, complement activation occurs on ECs in the absence of cell lysis.9 The treated porcine aortic ECs had increased levels of mRNA for E-selectin, intracellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM)-1, Iκ-Bα, IL-1α, IL-8, and PAI-1 (Figure 1⇓). The response to complement did not represent a nonspecific increase in expression of EC genes, because there was no change in the expression of von Willebrand factor (vWF) or β-actin mRNA.
Role of Complement in EC Activation
We showed that complement activation was actually required for activation of ECs by the following experiments. First, upregulation of genes for E-selectin, IL-1α, and tissue factor did not occur when soluble CR1 (sCR1) was added to the human serum applied to porcine aortic ECs (Figure 2A⇓). Second, porcine aortic ECs transgenic for human decay-accelerating factor (DAF) and CD59,21 which inhibit complement at the levels of C3 and C8/C9, respectively, did not show evidence of activation after stimulation with 15% human complement (Figure 2B⇓), although the transgenic cells were readily activated by IL-1α (Figure 2B⇓).
Effect of Complement on Microvascular ECs
Because much of the tissue damage brought about by activation of complement occurs in small blood vessels, we asked whether complement activates microvascular ECs as it does aortic ECs. Stimulation of porcine cardiac microvascular ECs by human anti-EC antibodies and complement caused expression of E-selectin and tissue factor over a period of 16 hours (Figure 3⇓). sCR1 also prevented activation of cardiac microvascular ECs by human serum, indicating that complement activates microvascular ECs as it does aortic ECs.
Kinetics of EC Activation
We next questioned whether the kinetics of EC activation in response to complement differs from the kinetics of the response to cytokines. In 5 experiments, stimulation of porcine aortic and microvascular ECs by IL-1α or TNF-α led to expression of E-selectin protein within 2 hours (Figure 4⇓). Conversely, stimulation of aortic ECs with complement led to expression of E-selectin only at ≈8 hours (Figure 4A⇓) and stimulation of microvascular ECs by complement led to expression of E-selectin only at ≈6 hours after stimulation (Figure 4B⇓). The more rapid expression of E-selectin on microvascular ECs was seen in repeated experiments. The delay in expression of E-selectin by ECs after exposure to complement was not limited by the rate of complement activation, because iC3b was deposited on ECs as early as 30 minutes after exposure to human serum (Figure 5⇓). Figure 5A⇓ also shows that whereas complement activation peaked at 1 to 2 hours, induction of E-selectin mRNA by complement did not occur until 8 hours had elapsed.
Activation of ECs in Response to Complement Is IL-1α–Mediated
We tested whether upregulation of E-selectin by complement was mediated by IL-1α, as reported for tissue factor and cyclooxygenase-2.13 14 Four lines of evidence supported this mechanism. First, as shown in Figure 1A⇑, mRNA for IL-1α appeared as early as 1 hour before expression of genes for E-selectin. Second, human recombinant IL-1 receptor antagonist (IL-1ra), which is cross-reactive among species,22 23 blocked complement-mediated upregulation of genes for E-selectin, ICAM-1, VCAM-1, Iκ-Bα, IL-1β, IL-6, and PAI-1 in porcine aortic ECs (Figure 6A⇓). The human recombinant IL-1ra also prevented complement-induced expression of E-selectin protein and tissue factor activity in cardiac microvascular ECs (Figure 6⇓, B and C). Third, the response to complement was mediated by IL-1α rather than by IL-1β, because IL-1β mRNA was detected only 16 hours after stimulation with complement. Fourth, IL-1α was the active component in conditioned medium obtained from ECs treated with complement, because the ability of conditioned medium to activate quiescent ECs was inhibited by addition of anti–porcine IL-1α antibodies (Figure 7C⇓). The inhibition of complement-mediated EC activation by the IL-1ra was not caused by a generalized suppression of EC responses, because IL-1ra did not inhibit EC activation by endotoxin (not shown).
The synthesis of IL-1α protein in ECs stimulated by complement was examined by immunoprecipitation using polyclonal anti–IL-1α antibodies generated against recombinant porcine IL-1α. In 6 experiments, cell-associated IL-1α polypeptide was detected as early as 3 hours after stimulation of porcine ECs with human complement. EC-associated IL-1α was detected as a precursor and processed products of 30 to 35 kDa in cells. The anti–pig IL-1α antibodies, like other anti–IL-1 antibodies,20 immunoprecipitated IL-1α from cell lysates. Although IL-1α could be detected in the supernatant by functional assay based on specific inhibition (Figure 7C⇑), the processed protein was not detected. Although the extent of processing varied between experiments,24 cell-associated IL-1α was absent in resting ECs and complement-treated porcine ECs transgenic for human DAF and CD59 (Figure 7B⇑).
We next questioned whether ECs store IL-1α, which would be released by complement, resulting in EC activation. The following suggested the absence of IL-1α in quiescent ECs. First, immunoprecipitation experiments (Figure 7B⇑) revealed the lack of IL-1α in resting ECs. Second, conditioned medium obtained from ECs exposed to complement for 30 minutes lacked IL-1α activity (Figure 7C⇑, inset). Third, expression of IL-1α mRNA at 1 hour required active RNA synthesis in ECs, because actinomycin D inhibited its expression in this system (Figure 7A⇑).
Role of Pore-Forming Proteins in Activation of ECs by Complement
The MAC was necessary for the activation of ECs, because a complement source lacking C8 did not induce E-selectin, whereas this C8-depleted serum reconstituted with purified C8 did so (Figure 8⇓). To test whether in activating ECs, the MAC was representative of membrane pore-forming proteins, we asked whether melittin, an amphipathic pore-forming peptide from bee venom similar to C9,25 activates ECs like MAC. Treatment of aortic ECs with melittin caused upregulation of IL-1α mRNA at a rate similar to that observed with complement and induced expression of E-selectin at 8 hours, with a peak expression at 12 hours (Figure 9⇓). Induction of E-selectin in ECs by melittin, as with complement, was dependent on IL-1, because IL-1ra inhibited E-selectin expression by ECs exposed to melittin (Figure 9⇓, B and C). Melittin also activated microvascular ECs via the IL-1α pathway, because the induction of E-selectin was inhibited by IL-1ra (Figure 10⇓). Mastoparan, another pore-forming peptide, also induced E-selectin in an IL-1–dependent manner (not shown). These findings suggest that transmembrane pore-forming structures activate ECs through the autocrine effects of IL-1α and that a common mechanism governs activation of aortic and microvascular ECs by the MAC and the MAC-like transmembrane pore-forming structures.
Membrane pore-forming peptides have been implicated in the pathogenesis of several disorders. For example, MAC is thought to play a role in the rejection of cardiac xenografts.26 Melittin induces edema and influx of polymorphonuclear leukocytes27 and causes cardiovascular dysfunction, manifested by widespread thrombosis and leukocyte infiltration.28 Our studies suggest that membrane pore-forming peptides mediate coagulation and inflammation by activating ECs through a distinct pathway, involving primarily upregulation of IL-1α and autocrine action of IL-1α on endothelium, leading to localized inflammation and coagulation.
The difference between the mechanism of EC activation by pore-forming structures and the mechanism of EC activation by cytokines may stem from the way these mediators interact with ECs. IL-1 and TNF-α bind to their corresponding receptors, which transduce signals, leading to activation of NF-κB and subsequent induction of genes possessing NF-κB binding sites. Pore-forming proteins, conversely, do not interact with specific receptors but rather randomly insert in cell membranes. Insertion of the MAC in EC membranes may activate the extracellular signal-regulated kinase (ERK) pathway, as it does in B cells,29 leading to the induction of IL-1α.
One prominent effect of the MAC on many cell types is an increase in cytoplasmic Ca2+. Alterations in intracellular concentration of Ca2+ may in turn play a role in the induction of IL-1α by the MAC and other pore-forming peptides. Increased concentration of intracellular Ca2+ is also thought to be important in surface expression of P-selectin from ECs exposed to the MAC.30 Consistent with the importance of Ca2+ in this system, we have found that a decrease in extracellular concentration of Ca2+ inhibits EC expression of E-selectin in response to complement (not shown).
Porcine aortic or microvascular ECs appear to respond differently than human umbilical vein ECs (HUVECs) to pore-forming peptides. Our preliminary experiments have shown that pore-forming peptides also activate human aortic ECs. HUVECs were recently found to respond rapidly to cytolytically inactive terminal complement complexes, upregulating E-selectin at a rate resembling the response to endotoxin.31 How the cytolytically inactive terminal complement complexes activated the HUVECs is not known, however, and it is possible that EC activation in this system is also IL-1α–dependent. The apparent difference between responses of HUVECs and aortic ECs to terminal complement complexes may also stem from different biological properties of these cells. For example, HUVECs, being fetal cells, express platelet-derived growth factors that maintain a partially activated state.32
Clearly, the activation of ECs by MAC can account for widespread inflammatory reactions seen in complement-mediated disorders. Our results suggest that activation of ECs in this way is likely to be the result of the synthesis and release of pluripotent IL-1α. IL-1α might impair regional blood flow by causing expression of vasoconstrictors such as thromboxane A2 and endothelin-1, which not only localize the injury but also amplify the effect by increasing the local concentration of IL-1α, causing expression of tissue factor, PAI-1, chemokines, and adhesion molecules. This impairment in regional blood flow could also result in a decrease in local pH, which in turn would optimize the activity of heparanase that is inactive at physiological pH but is active under acidic conditions.33 Activation of heparanase, in turn, might cause degradation and removal of heparan sulfate proteoglycans from ECs, with consequent loss of anticoagulant, anti-inflammatory, and barrier function, adding to the changes associated with activation of ECs.
This study was supported by grants from the National Institutes of Health (HL-52297 and HL-46810).
- Received July 21, 1999.
- Revision received November 19, 1999.
- Accepted December 2, 1999.
- Copyright © 2000 by American Heart Association
Esmon CT. The roles of protein C and thrombomodulin in the regulation of blood coagulation. J Biol Chem. 1989;264:4743–4746.
Luscher TF, Boulanger CM, Dohi Y, Yang Z. Endothelium-derived contracting factors. Hypertension. 1992;19:117–130.
Bevilacqua MP, Pober JS, Majeau GR, Cotran RS, Gimbrone MA Jr. Interleukin 1 (IL-1) induces biosynthesis and cell surface expression of procoagulant activity in human vascular EC. J Exp Med. 1984;160:618–623.
Bevilacqua MP, Schleef RR, Gimbrone MA Jr, Loskutoff DJ. Regulation of the fibrinolytic system of cultured human vascular endothelium by interleukin 1. J Clin Invest. 1986;78:587–591.
Bevilacqua MP, Pober JS, Mendrick DL, Cotran RS, Gimbrone MA Jr. Identification of an inducible endothelial-leukocyte adhesion molecule. Proc Natl Acad Sci U S A. 1987;84:9238–9242.
Warner SJC, Auger KP, Libby P. Interleukin 1 induces interleukin 1, II: recombinant human interleukin 1 induces interleukin 1 production by adult human vascular EC. J Immunol. 1987;139:1911–1917.
Collins T, Read MA, Neish AS, Whitley MZ, Thanos D, Maniatis T. Transcriptional regulation of endothelial cell adhesion molecules: NF-κB and cytokine-inducible enhancers. FASEB J. 1995;9:899–909.
Saadi S, Platt JL. Transient perturbation of endothelial integrity induced by antibodies and complement. J Exp Med. 1995;181:21–31.
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.
Selvan RS, Kapadia HB, Platt JL. Complement-induced expression of chemokine genes in endothelium: regulation by IL-1-dependent and -independent mechanisms. J Immunol. 1998;161:4388–4395.
Saadi S, Holzknecht RA, Patte CP, Stern DM, Platt JL. Complement-mediated regulation of tissue factor activity in endothelium. J Exp Med. 1995;182:1807–1814.
Nishida M, Carley WW, Gerritsen ME, Ellingsen O, Kelly RA, Smith TW. Isolation and characterization of human and rat cardiac microvascular EC. Am J Physiol. 1993;264:H639–H652.
Malyguine AM, Saadi S, Holzknecht RA, Patte CR, Sud N, Platt JL, Dawson JR. Induction of procoagulant function in porcine EC by human NK cells. J Immunol. 1997;159:4659–4664.
Hogquist KA, Unanue ER, Chaplin DD. Release of IL-1 from mononuclear phagocytes. J Immunol. 1991;147:2181–2186.
Russell DA, Tucker KK, Chinookoswong N, Thompson RC, Kohno T. Combined inhibition of interleukin-1 and tumor necrosis factor in rodent endotoxemia: improved survival and organ function. J Infect Dis. 1995;171:1528–1538.
Carinci V, Guida S, Fontana MR, Palla E, Rossini M, Melli M. Processing of interleukin-1 in cells of monocytic lineage is differentiation-dependent. Eur J Biochem. 1992;203:293–301.
Niculescu F, Rus H, van Biesen T, Shin ML. Activation of Ras and mitogen-activated protein kinase pathway by terminal complement complexes is G protein dependent. J Immunol. 1997;158:4405–4412.
Hattori R, Hamilton KK, McEver RP, Sims PJ. Complement proteins C5b-9 induce secretion of high molecular weight multimers of endothelial von Willebrand factor and translocation of granule membrane protein GMP-140 to the cell surface. J Biol Chem. 1989;264:9053–9060.
Tedesco F, Pausa M, Nardon E, Introna M, Mantovani A, Dobrina A. Cytolytically inactive terminal complement complex activates EC to express adhesion molecules and tissue factor procoagulant activity. J Exp Med. 1997;185:1619–1627.
Ihrcke NS, Parker W, Reissner KJ, Platt JL. Regulation of platelet heparanase during inflammation: role of pH and proteinases. J Cell Physiol. 1998;175:255–267.The mechanism by which pore-forming structures disturb the functions of endothelium was investigated. Amphiphilic pore-forming peptides, such as the membrane attack complex, melittin, and mastoparan, activated aortic and microvascular endothelial cells, not directly but through the autocrine action of IL-1α. Thus, pore-forming structures as a class act differently from other agonists on endothelial cells.