Association Between Complement Factor H and Proteoglycans in Early Human Coronary Atherosclerotic Lesions
Implications for Local Regulation of Complement Activation
Objective— Complement activation has been suggested to play a role in atherogenesis. To study the regulation of complement activation in human coronary atherosclerotic lesions, we examined the spatial relationships between the major complement inhibitor, factor H, and the complement activation products C3d and C5b-9.
Methods and Results— In early lesions (American Heart Association types II and III), factor H was immunohistochemically found in the superficial proteoglycan-rich layer in association with numerous macrophages and C3d, whereas C5b-9 was found deeper in the intima, where factor H was virtually absent. In vitro experiments involving surface plasmon resonance and affinity chromatography analyses demonstrated that isolated human arterial proteoglycans bind factor H, and functional complement assays showed that glycosaminoglycans inhibit the complement activation induced by modified low density lipoprotein or by a foreign surface.
Conclusions— The present observations raise the possibility that proteoglycans, because of their ability to bind the major complement inhibitor factor H, may inhibit complement activation in the superficial layer of the arterial intima. In contrast, deeper in the intima, where factor H and proteoglycans are absent, complement may be activated and proceed to C5b-9. Thus, the superficial and the deep layers of the human coronary artery appear to differ in their ability to regulate complement activation.
Increasing evidence implies that one of the key factors generating and maintaining inflammation in the arterial intima is the complement system.1 Immunostaining experiments have shown the presence of large amounts of the terminal complement complex C5b-9 in association with smooth muscle cell α-actin and modified LDL.2,3 ⇓ Lipids isolated from human atherosclerotic lesions and enzymatically modified LDL have been shown to activate complement via the alternative pathway in vitro.4 Most interestingly, in rabbits deficient in complement component C6 of the terminal complement cascade, the development of experimental atherosclerosis has been shown to be retarded.5 Thus, although the present evidence suggests a proatherogenic role for the terminal complement pathway, the role in atherosclerosis of the early complement cascade is not known. Interestingly, in C3-deficient apoE-null mice, the formation of fatty streaks was increased, but the fatty streaks failed to mature into fibrotic lesions.6
Complement activation depends critically on the balance between complement activators and complement-regulatory proteins. Human atherosclerotic lesions have previously been shown to contain the fluid-phase inhibitors (C1 inhibitor, factor H, factor B,7 and C4 binding protein8), clusterin (apoJ),9 vitronectin/S-protein,10 and the cell membrane glycoproteins (complement receptors 1 and 3,7 decay-accelerating factor,11 and CD5912). Consistent with the presence of complement activation products in atherosclerotic lesions, it has recently been shown that mRNAs for complement components, but not for complement inhibitors, are upregulated in atherosclerotic plaques.13
In the present study, our aim was to find out how complement activation is regulated in the arterial intima. For this purpose, we collected human coronary arteries and studied the spatial relationships between the complement activation markers (C3d and C5b-9) and the major complement inhibitor (factor H) and C-reactive protein (CRP), a protein that can activate the early classical complement pathway and inhibit complement amplification by binding factor H.14 Moreover, in vitro binding studies and complement consumption assays were carried out to study the complement-regulating properties of human arterial proteoglycans.
An extended Methods section can be found in the online supplement (available at http://atvb.ahajournals.org).
Coronary artery specimens (n=36) were obtained from 29 patients; their characteristics are shown in online Table I (see http://atvb.ahajournals.org). Frozen and paraffin sections were stained by the indirect immunoperoxidase method by using 3-amino-9-ethylcarbazole or Vector NovaRED (Vector Laboratories) as substrate and were counterstained with hematoxylin. The primary antibodies used in the present study are listed in online Table II (see http://atvb.ahajournals.org).
Proteoglycans were isolated from the intima-media of human aortas.15 The proteoglycans were dissociated into monomers by incubation in buffer containing 4 mol/L guanidine HCl, and the proteoglycan monomers so formed were isolated by size-exclusion chromatography.16
For surface plasmon resonance measurements, human factor H and CRP were immobilized on carboxylated dextran CM5 chips (Biacore) by an amine-coupling procedure. Analyses of their binding with human aortic proteoglycans were performed in 1:3 veronal-buffered saline (50 mmol/L NaCl, 0.6 mmol/L sodium barbital, and 1.1 mmol/L barbituric acid; pH 7.5).
For affinity chromatography, human aortic proteoglycans were coupled to a N-hydroxysucciminide-activated HiTrap column (Amersham Pharmacia Biotech). Human factor H (20 μg) was injected into the column that had been equilibrated with a buffer containing 10 mmol/L HEPES, pH 7.4, and either 2 mmol/L CaCl2 or 100 μmol/L EDTA. The material bound to the column was then eluted with a gradient of NaCl (0→250 mmol/L) in the buffer.
For assays of complement consumption, LDL was modified enzymatically, as described previously,17 to yield E-LDL. Complement activation in normal human serum (NHS) was studied by incubating NHS (50 μL) in the absence or presence of Sepharose, heparin-Sepharose, proteoglycan-Sepharose, or CRP (50 μg/mL) in a total volume of 100 μL PBS. In other experiments, NHS (50 μL in a total volume of 100 μL PBS) was incubated with E-LDL (50 μg/mL) in the absence or presence of heparin (100 μg/mL). The samples were incubated for 20 minutes at 37°C in a shaker, and generation of the complement activation product C3a-desarg was quantified by ELISA (Quidel Co).
For analysis of factor H–like proteins in human atherosclerotic lesions, 2 samples from grossly normal human coronaries and samples from normal and atherosclerotic aortas were homogenized and analyzed by Western blotting. The blots were probed with goat α-human factor H (Calbiochem, diluted 1:5000), rabbit α-human short consensus repeat (SCR) 19-20 (diluted 1:1000, kindly provided by Dr Jens Hellwage, Hans Knoell Institute, Jena, Germany), or mouse α-factor H antibody 196X18 at 5 μg/mL.
The distribution patterns of complement activation products, the complement inhibitor factor H, and CRP were compared with those of proteoglycans and apoB by immunohistochemistry. As a control, online Figure I (data supplement; please see http://atvb.ahajournals.org) shows a grossly normal human coronary artery (American Heart Association type I lesion) with the adaptive intimal thickening typical of coronary arteries. Neither the superficial nor the deep layer of the intima showed staining for C3d or C5b-9, indicating the absence of immunohistochemically detectable complement activation. The narrow proteoglycan-rich layer stained positively for factor H and partially for CRP.
Figure 1 and online Figure II (data supplement; please see http://atvb.ahajournals.org) show early to intermediate atherosclerotic lesions with intimal thickening and fatty streak formation. Typically, the macrophages and apoB were located in the superficial thick proteoglycan-rich layer, which was visualized by immunostaining for versican. Interestingly, these areas were also positive for C3d, whereas C5b-9 was found as distinct deposits deeper in the intima (in the musculoelastic layer). Thus, in the superficial intima, complement activation had been limited to the C3 level, whereas in the deep intima, it had proceeded to the C5b-9 level. Importantly, factor H, the major complement inhibitor, was present only in the superficial intima, in which it showed strict colocalization with versican. The observed colocalization of proteoglycans and factor H was obvious in all the early atherosclerotic lesions studied in which a distinct superficial proteoglycan-rich layer and deep musculoelastic layer could be clearly distinguished (n=10). In addition, a reciprocal staining pattern for factor H and C5b-9 was found in most of the samples studied. Three examples of such staining, with the use of immunofluorescence double staining, are shown in Figure 2. When the overlap between C5b-9 and factor H was estimated in samples in which distinct staining for factor H and C5b-9 was found (n=24), it was found that a reciprocal staining pattern (≤20% overlap) was present in 62.5% of the lesions, partial overlap (21% to 80% overlap) in 25% of the lesions, and colocalization (>80% overlap) in 12.5% of the lesions. Finally, CRP, which has the ability to initiate the classical complement pathway but also to terminate complement activation at the C3 level by binding factor H,14 was mainly found in the superficial intima colocalized with proteoglycans and factor H.
In the advanced atherosclerotic lesions studied (n=14), staining for both proteoglycans and factor H was more diffuse and showed only partial colocalization, regardless of the presence of a distinct proteoglycan-rich layer (found in 4 samples). In the advanced lesions, in accord with the previous results, extensive deposition of C5b-9 was also occasionally detected in areas positive for factor H19 and CRP.20
Because the immunohistochemical findings demonstrated that factor H is present predominantly in the proteoglycan-rich layer of the intima, we experimentally tested for the ability of human arterial proteoglycans to bind factor H. Using the surface plasmon resonance technique, we were able to observe binding of human arterial proteoglycans to immobilized factor H (Figure 3A). Moreover, because CRP colocalized with the proteoglycans in the arterial intima, we also tested for its ability to bind to the proteoglycans isolated from human aortas. However, the proteoglycans failed to bind to immobilized CRP in vitro (online Figure III; please see http://atvb.ahajournals.org).
Binding of factor H to surface-immobilized proteoglycans was analyzed further by affinity chromatography (Figure 3B). In this system, almost all of the factor H injected into the proteoglycan affinity column was eluted as a single peak at ≈50 mmol/L NaCl. This binding was found to be independent of the presence of Ca2+.
Because the results obtained by immunohistochemistry suggested a role for proteoglycans in complement inhibition, we studied whether glycosaminoglycans would inhibit complement activation in vitro. For this purpose, complement activation in NHS was studied by using enzymatically modified LDL and a foreign surface (Sepharose) as complement activator. We found that heparin glycosaminoglycan, when added to the solution with enzymatically modified LDL, strongly inhibited the production of C3a (Figure 4A). Moreover, heparin, when coupled to Sepharose, was able to fully inhibit the Sepharose-induced complement activation (Figure 4B). CRP, when added to the incubation mixture, caused a nonsignificant increase in Sepharose-induced complement activation and had no effect in the presence of heparin-Sepharose. Most important, proteoglycans isolated from human aortas were able to inhibit the complement activation induced by Sepharose beads (Figure 4C). Taken together, it appears that glycosaminoglycans can effectively inhibit complement activation in vitro.
To ascertain which proteins of the factor H family are deposited in human arterial tissue, the factor H–like proteins in homogenates of human coronary arteries and of the aorta were detected by immunoblotting (Figure 5).21 With a polyclonal antibody against factor H (Figure 5A), which recognizes all the proteins in the factor H family, we detected a band at 150 kDa corresponding to factor H, a band at 42/43 kDa corresponding to both factor H–like protein-1 (FHL-1, ≈42 kDa) and the α-form of factor H–related protein-1 (FHR-1, 43 kDa), and a band at 37 kDa corresponding to the β form of FHR-1. We next probed the blot with an antibody against the SCR 1 of factor H and FHL-1 that does not bind to the FHRs. As shown in Figure 5B, relative to human serum, all the arterial preparations were enriched in FHL-1. Finally, we probed the blot with an antibody against SCR 19-20, which recognizes factor H and the different FHRs but not FHL-1. As shown in Figure 5C, the amount of FHR-1 in the vascular samples appeared not to be increased but reduced. Moreover, the relative intensities of the FHR-1 bands corresponded to those of albumin, suggesting that at least a fraction of the FHR-1 in the intima was plasma-derived.
In the present study, we found that in early atherosclerotic lesions, the superficial layer of the arterial intima containing the versican proteoglycan also contained the major complement inhibitor, factor H. In most lesions, this area was also positive for CRP and C3d. However, in the deeper parts of the intima and also to some extent in the media, there was a relative absence of factor H and prominent deposition of the terminal complement complex C5b-9. This mutually exclusive staining pattern of factor H and C5b-9 suggests that factor H tends to prevent the assembly of the terminal complement complex in the arterial intima. In addition, the immunoblotting experiments revealed that various arterial preparations were relatively enriched not only in factor H but also in FHL-1, a protein that is encoded by an alternatively spliced mRNA derived from the factor H gene and that shares the complement-regulatory properties of factor H. Macrophages have previously been shown to synthesize factor H and FHL-1.22 Thus, these proteins, in addition to infiltration from the plasma, may also have originated from local synthesis in the macrophage-rich areas of the coronary intima.
Because proteoglycans colocalized with factor H in the arterial intima, we hypothesized that they might be involved in controlling excessive complement activation. Two independent types of analysis showed that factor H binds to isolated arterial proteoglycans in vitro. This binding is likely to involve ≥1 of the heparin-binding sites present in SCRs 7,13, and 20 of factor H.23–26 ⇓ ⇓ ⇓ Importantly, we could also show that human arterial proteoglycans, like heparin, can inhibit complement activation in vitro. The inhibitory effect of heparin appears to depend on its ability to potentiate the effect of factor H in C3b inactivation.27,28 ⇓ Taken together, our observations made in vivo and in vitro suggest that factor H binds to the intimal proteoglycans and may restrict complement activation locally. In the glomerular basement membranes of pigs genetically deficient in factor H29 and in humans either genetically deficient in factor H30 or with dysfunction of factor H due to autoantibodies,31 the in vivo findings of extensive complement deposition in the glomerular membrane of the kidney demonstrate the importance of factor H in protecting the extracellular matrix components against attack by complement. The present analogous findings in the human coronary artery extend the repertoire of protected tissues to the proteoglycan-rich layer of the arterial intima. Indeed, this intimal layer can be considered functionally analogous to the glomerular membrane in terms of its high content of glycosaminoglycans and in its capacity to retain molecules with an affinity for heparin.
The present study did not elucidate the nature of the agents that activate the complement system in the arterial intima. However, the findings of (1) C3d in the absence of C5b-9 in the proteoglycan-rich layer of the arterial intima and (2) the deposition of C4 also in this layer (not shown) led us to infer that the classical pathway of complement is activated in this layer, although activation via mannose-binding lectin cannot be excluded.32 One of the mediators of the classical complement pathway activation is CRP. We, among others, have found CRP in the arterial intima,20,33–37 ⇓ ⇓ ⇓ ⇓ ⇓ and interestingly, intimal cells in addition to hepatocytes have been shown to encode a high level of mRNA for CRP.37,38 ⇓ Previous studies have shown that human atherosclerotic lesions also contain deposits of IgM and IgG,39,40 ⇓ both of which have been shown to activate the classical pathway of complement in vitro.41 CRP and immunoglobulins have been shown to bind to modified lipoproteins,36,42 ⇓ and CRP also binds to apoptotic cells,43 chromatin,44 and fibronectin,45 all of which are present in atherosclerotic lesions. The presence, in the deep musculoelastic layer of the arterial intima, of C5b-9 and extracellular lipids, including cholesterol crystals, is compatible with the view that the extracellular lipids have activated the alternative pathway of complement. This view is supported by previous reports demonstrating that intimal extracellular liposomes,46 enzymatically modified LDL,4 and cholesterol crystals47 can all activate the alternative pathway of complement in vitro.
As demonstrated by a study involving experimental atherosclerosis in C6-deficient rabbits,5 excessive activation of the terminal complement pathway, like excessive activation of other pathways of the inflammatory system, is proatherogenic. Although activation of the early complement cascade may lead to full activation of complement with ensuing inflammation, its role in opsonization, the evolutionarily old function of complement,48 may be protective. Indeed, this is supported by the recent findings showing that the formation of fatty streak lesions is increased in apoE-null mice deficient in C3.6 According to this view, CRP and factor H may work together in clearing debris from the intima via opsonization by iC3b and promote uptake by macrophages via the CR3 (CD11b/CD18) receptor.49 The present findings suggest that proteoglycans exert local control of complement activation in the human coronary intima by allowing activation of the early complement cascade involved in opsonization but limiting complement activation to the C3 level. In light of the above-mentioned considerations, this regulatory function by proteoglycans in the early coronary lesions is likely to be protective; ie, it would tend to prevent the progression of atherosclerosis from early to more advanced lesions. Indeed, recent experiments with C3-null mice have suggested that the maturation of early lesions into fibrotic lesions depends on an intact complement system.6
The Wihuri Research Institute is maintained by the Jenny and Antti Wihuri Foundation. This study was supported by grants QLG1-1999-01007 from the European Commission to the consortium “Macrophage Function and Stability of the Atherosclerotic Plaque (MAFAPS)” as part of the Fifth Framework Program of the European Union (P.T.K. and M.O.P.), the Academy of Finland (P.T.K. and S.M.), the Federation of Finnish Insurance Companies (P.T.K.), the Finnish Heart Foundation (S.M.), and the Sigrid Juselius Foundation (P.T.K. and S.M.). We thank Suvi Mäkinen, Laura Vatanen, and Päivi Ihamuotila for expert technical assistance and Dr Katariina Öörni for help with preparation and characterization of the aortic proteoglycans.
- Received September 5, 2002.
- Accepted December 4, 2002.
- ↵Torzewski M, Torzewski J, Bowyer DE, Waltenberger J, Fitzsimmons C, Hombach V, Gabbert HE. Immunohistochemical colocalization of the terminal complex of human complement and smooth muscle cell alpha-actin in early atherosclerotic lesions. Arterioscler Thromb Vasc Biol. 1997; 17: 2448–2452.
- ↵Torzewski M, Klouche M, Hock J, Messner M, Dorweiler B, Torzewski J, Gabbert HE, Bhakdi S. Immunohistochemical demonstration of enzymatically modified human LDL and its colocalization with the terminal complement complex in the early atherosclerotic lesion. Arterioscler Thromb Vasc Biol. 1998; 18: 369–378.
- ↵Bhakdi S, Dorweiler B, Kirchmann R, Torzewski J, Weise E, Tranum J, Walev I, Wieland E. On the pathogenesis of atherosclerosis: enzymatic transformation of human low density lipoprotein to an atherogenic moiety. J Exp Med. 1995; 182: 1959–1971.
- ↵Schmiedt W, Kinscherf R, Deigner HP, Kamencic H, Nauen O, Kilo J, Oelert H, Metz J, Bhakdi S. Complement C6 deficiency protects against diet-induced atherosclerosis in rabbits. Arterioscler Thromb Vasc Biol. 1998; 18: 1790–1795.
- ↵Buono C, Come CE, Witztum JL, Maguire GF, Connelly PW, Carroll M, Lichtman AH. Influence of C3 deficiency on atherosclerosis. Circulation. 2002; 105: 3025–3031.
- ↵Seifert PS, Hansson GK. Complement receptors and regulatory proteins in human atherosclerotic lesions. Arteriosclerosis. 1989; 9: 802–811.
- ↵Mackness B, Hunt R, Durrington PN, Mackness MI. Increased immunolocalization of paraoxonase, clusterin, and apolipoprotein A-I in the human artery wall with the progression of atherosclerosis. Arterioscler Thromb Vasc Biol. 1997; 17: 1233–1238.
- ↵Seifert PS, Hansson GK. Decay-accelerating factor is expressed on vascular smooth muscle cells in human atherosclerotic lesions. J Clin Invest. 1989; 84: 597–604.
- ↵Yasojima K, Schwab C, McGeer EG, McGeer PL. Complement components, but not complement inhibitors, are upregulated in atherosclerotic plaques. Arterioscler Thromb Vasc Biol. 2001; 21: 1214–1219.
- ↵Jarva H, Jokiranta TS, Hellwage J, Zipfel PF, Meri S. Regulation of complement activation by C-reactive protein: targeting the complement inhibitory activity of factor H by an interaction with short consensus repeat domains 7 and 8-11. J Immunol. 1999; 163: 3957–3962.
- ↵Öörni K, Pentikäinen MO, Annila A, Kovanen PT. Oxidation of low density lipoprotein particles decreases their ability to bind to human aortic proteoglycans: dependence on oxidative modification of the lysine residues. J Biol Chem. 1997; 272: 21303–21311.
- ↵Camejo G, Fager G, Rosengren B, Hurt-Camejo E, Bondjers G. Binding of low density lipoproteins by proteoglycans synthesized by proliferating and quiescent human arterial smooth muscle cells. J Biol Chem. 1993; 268: 14131–14137.
- ↵Chao FF, Blanchette-Mackie EJ, Tertov VV, Skarlatos SI, Chen YJ, Kruth HS. Hydrolysis of cholesteryl ester in low density lipoprotein converts this lipoprotein to a liposome. J Biol Chem. 1992; 267: 4992–4998.
- ↵Jokiranta TS, Zipfel PF, Hakulinen J, Kuhn S, Pangburn MK, Tamerius JD, Meri S. Analysis of the recognition mechanism of the alternative pathway of complement by monoclonal anti-factor H antibodies: evidence for multiple interactions between H and surface bound C3b. FEBS Lett. 1996; 393: 297–302.
- ↵Seifert PS, Hansson GK. Complement receptors and regulatory proteins in human atherosclerotic lesions. Arteriosclerosis. 1989; 9: 802–811.
- ↵Torzewski J, Torzewski M, Bowyer DE, Frohlich M, Koenig W, Waltenberger J, Fitzsimmons C, Hombach V. C-reactive protein frequently colocalizes with the terminal complement complex in the intima of early atherosclerotic lesions of human coronary arteries. Arterioscler Thromb Vasc Biol. 1998; 18: 1386–1392.
- ↵Meri S, Pangburn MK. Discrimination between activators and nonactivators of the alternative pathway of complement: regulation via a sialic acid/polyanion binding site on factor H. Proc Natl Acad Sci U S A. 1990; 87: 3982–3986.
- ↵Pangburn MK, Atkinson MA, Meri S. Localization of the heparin-binding site on complement factor H. J Biol Chem. 1991; 266: 16847–16853.
- ↵Blackmore TK, Sadlon TA, Ward HM, Lublin DM, Gordon DL. Identification of a heparin binding domain in the seventh short consensus repeat of complement factor H. J Immunol. 1996; 157: 5422–5427.
- ↵Blackmore TK, Hellwage J, Sadlon TA, Higgs N, Zipfel PF, Ward HM, Gordon DL. Identification of the second heparin-binding domain in human complement factor H. J Immunol. 1998; 160: 3342–3348.
- ↵Kazatchkine MD, Fearon DT, Silbert JE, Austen KF. Surface-associated heparin inhibits zymosan-induced activation of the human alternative complement pathway by augmenting the regulatory action of the control proteins on particle-bound C3b. J Exp Med. 1979; 150: 1202–1215.
- ↵Jokiranta TS, Solomon A, Pangburn MK, Zipfel PF, Meri S. Nephritogenic lambda light chain dimer: a unique human miniautoantibody against complement factor H. J Immunol. 1999; 163: 4590–4596.
- ↵Bhakdi S, Torzewski M, Klouche M, Hemmes M. Complement and atherogenesis: binding of CRP to degraded, nonoxidized LDL enhances complement activation. Arterioscler Thromb Vasc Biol. 1999; 19: 2348–2354.
- ↵Pentikäinen MO, Lindstedt KA, Shiota N, Vehmaan-Kreula P, Laine P, Kovanen PT. C-reactive protein is expressed in human arterial intima. Atherosclerosis. 2000; 151: 199. Abstract.
- ↵Ylä-Herttuala S, Palinski W, Butler SW, Picard S, Steinberg D. Rabbit and human atherosclerotic lesions contain IgG that recognizes epitopes of oxidized LDL. Arterioscler Thromb. 1994; 14: 32–40.
- ↵Gershov D, Kim S, Brot N, Elkon KB. C-Reactive protein binds to apoptotic cells, protects the cells from assembly of the terminal complement components, and sustains an antiinflammatory innate immune response: implications for systemic autoimmunity. J Exp Med. 2000; 192: 1353–1364.
- ↵Robey FA, Jones KD, Tanaka T, Liu TY. Binding of C-reactive protein to chromatin and nucleosome core particles: a possible physiological role of C-reactive protein. J Biol Chem. 1984; 259: 7311–7316.
- ↵Salonen EM, Vartio T, Hedman K, Vaheri A. Binding of fibronectin by the acute phase reactant C-reactive protein. J Biol Chem. 1984; 259: 1496–1501.
- ↵Seifert PS, Hugo F, Tranum-Jensen J, Zahringer U, Muhly M, Bhakdi S. Isolation and characterization of a complement-activating lipid extracted from human atherosclerotic lesions. J Exp Med. 1990; 172: 547–557.
- ↵Mevorach D, Mascarenhas JO, Gershov D, Elkon KB. Complement-dependent clearance of apoptotic cells by human macrophages. J Exp Med. 1998; 188: 2313–2320.