Abstract
Undesired activation of the complement system is a major pathogenic factor contributing to various immune complex diseases and conditions such as hyperacute xenograft rejection. We aim for prevention of complement-mediated damage by specific inhibition of the classical complement pathway, thus not affecting the antimicrobial functions of the complement system via the alternative pathway and the lectin pathway. Therefore, 42 peptides previously selected from phage-displayed peptide libraries on basis of C1q binding were synthesized and examined for their ability to inhibit the function of C1q. From seven peptides that showed inhibition of C1q hemolytic activity but no inhibition of the alternative complement pathway, one peptide (2J) was selected and further studied. Peptide 2J inhibited the hemolytic activity of C1q from human, chimpanzee, rhesus monkey, rat, and mouse origin, all with a similar dose-response relationship (IC50 2–6 μM). Binding of C1q to peptide 2J involved the globular head domain of C1q. In line with this interaction, peptide 2J dose-dependently inhibited the binding of C1q to IgG and blocked activation of C4 and C3 and formation of C5b-9 induced via classical pathway activation, as assessed by ELISA. Furthermore, the peptide strongly inhibited the deposition of C4 and C3 on pig cells following their exposure to human xenoreactive Abs and complement. We conclude that peptide 2J is a promising reagent for the development of a therapeutic inhibitor of the earliest step of the classical complement pathway, i.e., the binding of C1q to its target.
Activation of the complement system plays a dual role in disease. On the one hand, complement activation has many protective functions in immunity, both as a first line defense mechanism against invading pathogens and as a potentiator of acquired immunity. On the other hand, complement activation is a major cause of tissue injury in many pathological conditions (1). Thus far, three different pathways have been identified via which the complement system can be activated, i.e., the classical pathway, the alternative pathway, and the lectin pathway. Activation of the classical complement pathway is involved in tissue damage resulting from deposition of autoantibodies and immune complexes, which may occur in autoimmune diseases such as systemic lupus erythematosus, myasthenia gravis, and Goodpasture’s syndrome (1, 2, 3). Furthermore, classical pathway activation is responsible for tissue injury in hyperacute xenograft rejection, triggered by the direct binding of preformed host Abs to the graft endothelium (4). Inappropriate complement activation is also an important mediator of ischemia/reperfusion injury occurring, for example, in stroke and myocardial infarction and after major surgery (5, 6, 7). Activation of the classical complement pathway in this type of tissue damage can occur via Ab-dependent as well as Ab-independent mechanisms, which in the latter case may involve the direct binding of C1q to damaged cells and in situ deposited acute phase proteins (6, 7).
The therapeutic application of complement inhibitors, to prevent undesired effects of complement activation, is currently under development. For example, C1 inhibitor, a physiological inhibitor of the serine proteases C1r and C1s of the classical pathway, has been preliminary used in man (8, 9). These studies have underscored the value of anticomplement strategies in the treatment of, e.g., acute myocardial infarction and sepsis. Next to inhibiting the C1 complex, C1 inhibitor also affects the lectin pathway of complement activation and the contact system (8, 10). Recent experiments indicate that high-dose i.v. Ig, which is frequently used as a broad anti-inflammatory treatment in patients, is also able to inhibit classical pathway activation in vivo (11). Other studies explored the use of soluble recombinant complement receptor 1 (CR1)3 as a therapeutic complement inhibitor. CR1 (CD35) is a membrane-bound receptor that binds the complement components C1q, C4, and C3. The soluble protein is able to prevent complement activation and complement-related damage in vivo (9). However, the disadvantage of this treatment is the lack of specificity, i.e., because soluble CR1 blocks C3, complement activation via the alternative pathway and the lectin pathway and the classical pathway are inhibited, all to the same extent. Recently, a peptide inhibitor of C3 has been developed, named compstatin (12, 13, 14, 15). Compstatin has promising properties for therapeutic complement inhibition both in vitro and in vivo.
Because the classical pathway in many cases is for a major part responsible for complement-related tissue damage, a specific and effective inhibitor of the classical pathway is desirable. Such an approach does not affect the alternative pathway and the lectin pathway, pathways known to play a key role in innate immunity against pathogens. Given that such an anticomplement treatment will potentially be useful in immune-compromised patients, the first line antimicrobial defense is of great importance to prevent life-threatening infections. Therefore, we aimed at the development of novel inhibitors of C1q, the recognition unit of the classical complement pathway.
Recently, Lauvrak et al. (16) reported the sequences of 42 peptides that were selected from phage display libraries on the basis of binding to human C1q. In the present study, we explored the effects of these peptides on complement activation. We selected one peptide that inhibits the classical pathway but not the alternative pathway and we present its mechanism of action. Because this peptide inhibits C1q from human, primate, and rodent origin, we propose that this peptide is a promising candidate for further development as a therapeutic C1q inhibitor.
Materials and Methods
Peptide synthesis
Peptide sequences were obtained from Lauvrak et al. (16), who selected 42 peptide sequences from five different phage-displayed peptide libraries (unconstrained and cysteine-constrained) on the basis of phage binding to human C1q. In this study, these 42 peptides were synthesized. Synthetic peptides were made by solid phase technology on TentagelS resin (Rapp, Tübingen, Germany) using N-Fmoc-t-butyl-protected amino acids, piperidine deprotection, and PyBOP/N-methylmorpholine activation. For the synthesis of peptide 2J-3, Fmoc-l-2-aminobutyric acid was used to introduce Xa. Peptide 2Jcy was obtained by first introducing an N-terminal bromoacetyl moiety (Xb), trifluoroacetic acid-induced cleavage and deprotection, ether precipitation of the cleaved product, and subsequent immediate cyclisation of the peptide in Na-phosphate pH 8 solution for 1 h at room temperature. Synthetic peptides were analyzed by analytical reversed phase HPLC and matrix-assisted laser desorption ionization time of flight mass spectrometry.
Peptides were dissolved in DMSO at a concentration of 10 mM and stored at −80°C. From this stock solution, appropriate dilutions were made in the different assay buffers as described below.
C1q isolation
C1q was isolated from human donor plasma as described previously (17). In brief, a protein precipitate was made from recalcified human plasma by addition of polyethylene glycol 6000 (3% w/v; Sigma-Aldrich, St. Louis, MO) and incubation for 1 h on ice. After centrifugation, the precipitate was dissolved in 2.5× Veronal-buffered saline (VBS; 1.8 mM Na-5,5-diethylbarbital, 0.2 mM 5,5-diethylbarbituric acid, 145 mM NaCl) containing 10 mM EDTA and loaded on an affinity column consisting of Sepharose-coupled human IgG that was previously incubated with rabbit IgG directed against human IgG. After washing, the column was eluted using 1 M NaCl in PBS/10 mM EDTA. Peak fractions containing C1q, as assessed by a C1q hemolytic assay, were pooled, stored at −80°C, and used for experiments.
C1q from rat serum and C1q from rhesus macaque serum (kindly provided by Dr. W. Bogers, Biomedical Primate Research Center, Rijswijk, The Netherlands) were isolated following the same procedure as described above. To isolate C1q from mouse serum, EDTA was added to the serum to a final concentration of 10 mM and the serum was directly loaded on the C1q affinity column as described above.
Hemolytic assays
For hemolytic assays of classical pathway complement activation, SRBC were sensitized using rabbit anti-SRBC Abs (Ab-coated erythrocytes (EA)). For a hemolytic activity of the classical component pathway (CH50) test, a total number of 1 × 107 EA was incubated in the presence of normal human serum diluted in dextrose gelatin Veronal buffer2+ (0.5× VBS, 0.05% gelatin, 167 mM glucose, 0.15 mM CaCl2, 0.5 mM MgCl2; final volume 200 μl). In a similar way, a C1q-dependent hemolysis test was performed using C1q-depleted human serum, diluted 1/75, and a limiting amount of purified C1q (obtained from human, mouse, rat, or rhesus monkey serum). In some experiments, a limiting amount of chimpanzee serum (kindly provided by Dr. W. Bogers) was used as a source of C1q. For analysis of alternative pathway activity, a hemolytic activity of the alternative complement pathway (AP50) test was performed, in which 1 × 107 rabbit erythrocytes were incubated with human serum diluted in dextrose gelatin Veronal buffer2+ containing 10 mM MgEGTA. In all these assays, peptide inhibitors were tested by premixing appropriate concentrations of peptide with the complement source before addition of the mixture to the erythrocytes, followed by a 60-min incubation at 37°C. After addition of 1.5 ml of PBS and centrifugation, hemolysis was assessed by measuring OD at 414 nm. The lytic activity of condition x was expressed in Z values:Z = −ln(1 − ((OD414 (x) − OD414 (0%))/(OD414 (100%) − OD414 (0%)))), in which the OD414 (0%) represents incubation of EA with buffer only and the OD414 (100%) was assessed after addition of H2O. The amount of complement added was chosen in such a way that the Z value in the absence of inhibitors was between 0.5 and 1.5. In general, serum was diluted 1/200 for a CH50 assay and 1/10 for an AP50 assay. The results are expressed as relative hemolytic activity and calculated as the ratio of the Z value in the presence of the inhibitor and the Z value in the absence of the inhibitor.
ELISA
In general, ELISA were performed using Maxisorb plates (Nunc, Roskilde, Denmark) that were coated with different proteins or peptides diluted in coating buffer (100 mM Na2CO3/NaHCO3, pH 9.6), either for 2 h at 37°C or overnight at room temperature. Nonspecific binding sites were blocked using incubation with 3% BSA in PBS for 1 h at 37°C. All subsequent steps were performed in PBS containing 0.05% Tween 20 and 1% BSA unless otherwise indicated, and each step was followed by washing for three times using PBS/0.05% Tween 20. Enzyme activity of HRP was assessed by addition of ABTS (Sigma-Aldrich) and H2O2. The OD at 415 nm was measured using a microplate biokinetics reader (EL312e; Bio-Tek Instruments, Winooski, VT).
To assess binding of C1q to peptides, peptides were coated on the plate at various concentrations, followed by a blocking step and addition of purified human C1q (0.1 μg/ml). Binding of C1q was detected using F(ab′)2 from rabbit IgG anti-human C1q, conjugated to digoxygenin-3-O-methylcarbonyl-ε-aminocaproic acid-N-hydroxysuccinimide ester (dig; Boehringer Mannheim, Mannheim, Germany) according to instructions provided by the manufacturer, followed by HRP-conjugated F(ab′)2 from goat IgG anti-dig (Boehringer Mannheim). To assess the effect of peptides on C1q binding to IgG, plates were coated with rabbit IgG (5 μg/ml) and purified human C1q (0.4 μg/ml) was added in the presence or absence of peptides in various concentrations. C1q binding was assessed using a biotinylated mouse mAb directed against human C1q (mAb 2214; kindly provided by Dr. C. E. Hack, Sanquin Blood Supply Foundation, Amsterdam, The Netherlands) (18), followed by HRP-conjugated streptavidin (Boehringer Mannheim). In some experiments, C1q was allowed to bind to coated rabbit IgG or peptide 2J in the presence or absence of purified mAb 2204 directed against the globular head portion of C1q or mAb 2214 directed against the collagenous portion of C1q (both mouse IgG1; kindly provided by Dr. C. E. Hack) (18), followed by detection of C1q binding using dig-conjugated rabbit IgG F(ab′)2 anti-human C1q, as indicated above.
ELISA-based assays to detect classical pathway complement activation were performed in plates coated with purified human IgM that were incubated with normal human serum (generally diluted 1/200) in the presence or absence of peptides diluted in gelatin Veronal buffer2+ (VBS, 0.1% gelatin, 0.5 mM MgCl2, 2 mM CaCl2, 0.05% Tween 20). Complement activation was assessed using mouse mAb to detect deposition of activated C4 (C4-4a anti-human C4d (19), from Dr. C. E. Hack, conjugated to dig), activated C3 (RFK22 anti-human C3 (20), conjugated to dig) and the membrane attack complex C5b-9 (unconjugated AE11 anti-C5b-9; kindly provided by Dr. T. E. Mollnes, Nordland Central Hospital, Bodø, Norway). Ab binding was detected using either HRP-conjugated F(ab′)2 anti-dig (Boehringer Mannheim) or HRP-conjugated goat anti-mouse Igs (DAKO, Glostrup, Denmark).
Assessment of complement deposition on PK15 cells
Complement activation induced by xenoreactivity of human serum to pig cells was assessed essentially as described previously (20). PK15 cells (pig kidney epithelial cell line, obtained from the American Type Culture Collection, Manassas, VA) were cultured in DMEM (Life Technologies, Breda, The Netherlands) supplemented with 10% FCS, penicillin (100 IU/ml) and streptomycin (100 μg/ml). Cells were detached by trypsinization, followed by incubation in normal human serum (5%) diluted in culture medium, for 30 min at 37°C, in the presence or absence of peptides at 100 μM. Subsequently, cells were washed with cold washing buffer (PBS, 1% BSA, 0.01% NaN3) followed by incubation with mAb directed against C4 (C4-4) and C3 (RFK22). Binding of mouse mAb was detected using PE-conjugated goat anti-mouse Ig Abs (DAKO). Cells were analyzed by flow cytometry using a FACScan (BD Biosciences, Erembodegem, Belgium) in the presence of propidium iodide (1 μg/ml; Molecular Probes, Leiden, The Netherlands) for exclusion of dead cells. Deposition of C4 and C3 was expressed as mean or median fluorescence intensity of cells that were gated based on scatter parameters and negative staining for propidium iodide. Trypsinization of PK15 cells does not modify the surface expression of the major xenoantigen Galα1-3Gal or their response to exposure to human serum (20).
Results
Selection of peptides on the basis of C1q inhibition
To select useful C1q-inhibiting peptides from the 42 peptide sequences previously reported by Lauvrak et al. (16), we synthesized these peptides and tested them in four different assay systems (Table I⇓). First, all peptides were tested in a sensitive C1q-dependent hemolytic assay. Of 42 peptides, 30 peptides showed inhibition of the hemolytic activity of C1q. Next, all peptides were tested for their ability to inhibit the classical complement pathway in a CH50 test. Of the 30 selected peptides, 20 peptides showed inhibition of the CH50 test. Furthermore, all peptides were examined for whether they affected the alternative complement pathway in an AP50 test. From the 20 selected peptides, eight were positively selected because they inhibited the classical pathway but not the alternative pathway. Finally, peptides were coated in an ELISA plate and C1q binding was assessed. Seven of the eight selected peptides showed clear binding of C1q to the immobilized peptide. These seven peptides have the desired properties for a specific C1q inhibitor; i.e., they inhibit the classical complement pathway at the level of C1q but not the alternative pathway. Based on its strong inhibitory capacity, peptide 2J was selected for further analysis.
Functional characterization of 42 peptides regarding their effect on the classical and the alternative complement pathways and their binding to C1q
In Fig. 1⇓, representative test results are shown for peptide 2J, in comparison with peptide 4B and peptide 2L that were negatively selected during the test procedure. Peptide 2J showed a clear dose-dependent inhibition of C1q-dependent hemolysis (Fig. 1⇓A) and of total classical pathway activity of normal human serum in the CH50 test (Fig. 1⇓B), but had no effect on the activity of the alternative pathway in the AP50 test (Fig. 1⇓C). In contrast, peptide 4B inhibited both the classical pathway and the alternative pathway, and peptide 2L had no effect on any test for complement activation (Fig. 1⇓, A–C). Peptides 2J and 4B, but not peptide 2L, showed a clear dose-dependent interaction with C1q when coated on ELISA plates (Fig. 1⇓D). As an additional test for the specificity of peptide 2J, we established that the peptide does not bind to mannan-binding lectin, which initiates the lectin pathway of complement activation (data not shown).
Selection of specific peptide inhibitors of the classical complement pathway. Various concentrations of the peptides 2J, 4B, and 2L, as indicated, were tested in different complement activation tests (A–C) and in ELISA (D). A, Peptides were preincubated with purified human C1q and tested in a C1q-dependent hemolytic assay. B, Peptides were preincubated with normal human serum and tested in a CH50 test for classical pathway activity. C, Peptides were preincubated with normal human serum and tested in an AP50 test for alternative pathway activity. D, Peptides were coated on ELISA plates followed by incubation with purified human C1q (0.1 μg/ml) and detection of C1q binding. A–C, Data represent the mean ± SD of duplicate analysis. Results are representative of three independent experiments.
Structural variants of peptide 2J
Peptide 2J is a 15-meric peptide containing cysteine residues at positions 1 and 14 (Table I⇑). Depending on the redox status of the environment, such a sequence may form a cyclic structure upon oxidation. We examined whether a cyclic structure is required for the C1q-inhibiting properties of peptide 2J. Therefore, five variants of peptide 2J were synthesized. Four variants were obligatory linear due to a deletion of either one cysteine (2J-1,2) or both cysteines (2J-4), or due to the replacement of both cysteine residues by 2-aminobutyric acid, an isosteric homolog of cysteine (Xa) that cannot form S-S bridges (2J-3) (Table II⇓). These variants were tested for C1q inhibition in a C1q-dependent hemolytic assay. Peptide 2J inhibited C1q hemolytic activity with an IC50 of ∼2 μM, but the linear variants did not show inhibition of C1q when applied at a concentration of up to 100 μM (Fig. 2⇓A). Furthermore, these linear variants did not bind C1q in ELISA, suggesting that the circular structure is required for the interaction of peptide 2J with C1q. Alternatively, a stable cyclic variant of peptide 2J (2Jcy) was constructed by replacing the N-terminal cysteine with bromoacetic acid, which was used to cyclise the peptide (N terminus attached to the cysteine side chain). Peptide 2Jcy inhibited the C1q hemolytic activity in a dose-dependent way, but the IC50 was approximately eight times higher than that of the original peptide 2J (Fig. 2⇓B). Apparently, the two cysteines in the original peptide 2J are required for optimal activity.
The C1q-inhibiting activity of structural variants of peptide 2J. Various concentrations of peptide 2J, its linear variants (2J1–4, A) and a stable cyclic variant (2JCy, B) were preincubated with human C1q and tested in a C1q-dependent hemolytic assay. Peptide 2L was included as a negative control. Data are presented as the mean ± SD from duplicate analysis. Similar results were obtained in at least two independent experiments.
Structural variants of peptide 2Ja
Peptide 2J inhibits the classical complement pathway by blocking the binding of C1q to its ligand
Inhibition of the different steps of the classical complement pathway by peptide 2J was further studied by ELISA. Incubation of IgM-coated ELISA plates with human serum resulted in deposition of activated C4 and C3 as well as the terminal complement complex C5b-9 on the plate. Incubation of human serum in the presence of peptide 2J inhibited activation of C4 and C3 as well as formation of C5b-9 in a dose-dependent way, whereas the linear peptide 2J-3 had no effect (Fig. 3⇓). Furthermore, peptide 2J showed a dose-dependent inhibition of the binding of purified human C1q to coated IgG (Fig. 4⇓A), suggesting an interaction of peptide 2J with the globular head domain of C1q. To further characterize the interaction between C1q and peptide 2J, an experiment was performed using mAb directed against various domains of C1q. A mAb directed against the head portion of C1q blocked the binding of C1q to peptide 2J and to IgG, whereas a mAb against the collagenous part of C1q did not affect the binding of C1q to either IgG or peptide 2J (Fig. 4⇓B).
Peptide 2J inhibits C4 and C3 activation and formation of C5b-9. ELISA plates were coated with purified human IgM and incubated with human serum in the presence or absence of different concentrations of peptide 2J and its linear variant peptide 2J-3. Binding of C4 (A), C3 (B), and C5b-9 (C) was assessed using specific mAb as described in Materials and Methods. Results represent one of three experiments.
Peptide 2J inhibits the binding of the globular head of C1q to IgG. A, Rabbit IgG was coated to ELISA wells and a preincubated mixture of peptide 2J or peptide 2J3 with human C1q was allowed to bind to the plate. Bound C1q was subsequently detected with a mAb directed against C1q. B, Wells were coated with rabbit IgG or with peptide 2J (10 μM) followed by incubation with human C1q (0.4 μg/ml) in the presence or absence of mAb directed against the globular head of C1q (mAb 2204) or the collagenous tail of C1q (mAb 2214) (0.6 μg/ml). C1q binding was assessed using a polyclonal Ab directed against C1q. Data represent mean ± SD of duplicate OD values obtained after subtraction of the OD obtained in the absence of C1q. Similar results were obtained in three independent experiments.
Peptide 2J inhibits human, primate, and rodent C1q
Peptide 2J was selected from a phage display library on the basis of binding to human C1q. To examine the potency of peptide 2J for inhibition of C1q from other species, hemolytic assays were performed in which human C1q-deficient serum was reconstituted with limiting amounts of C1q from human, chimpanzee, rhesus monkey, rat, and mouse origin. Peptide 2J dose-dependently inhibited the hemolytic activity of C1q from all the species tested with a comparable strength (Fig. 5⇓), suggesting that this peptide binds to a region of C1q that has a high degree of similarity among the various species.
Peptide 2J inhibits C1q from human, primate, and rodent origin. Peptide 2J (closed symbols) or peptide 2L (open symbols) were preincubated with 1/2000 diluted chimpanzee serum, or with purified C1q obtained from human, rat, mouse, or rhesus macaque serum. Samples were tested for C1q hemolytic activity using C1q-deficient human serum.
Peptide 2J inhibits xenoreactivity of human serum to pig cells
Hyperacute rejection of xenografts is initiated by activation of the classical complement pathway following the binding of preformed xenoreactive Abs from the host to the graft endothelium. Accordingly, when PK15 cells (pig kidney epithelial cells) are exposed to human serum, binding of human anti-pig Abs results in complement activation and surface deposition of C3, as detected by flow cytometry (Fig. 6⇓). C3 deposition was clearly detectable on the total cell population, as indicated by a shift of the fluorescence. Furthermore, a subpopulation of the cells showed particular strong fluorescence (C3bright, fluorescence intensity > 200, 15.5%; Fig. 6⇓, upper panel). Incubation of PK15 cells with human serum in the presence of peptide 2J, but not the linear variant peptide 2J-3, strongly inhibited the deposition of C3 on PK15 cells, as compared with the control (Fig. 6⇓). This inhibitory effect of peptide 2J was apparent from the mean and the median fluorescence, from a shift of the major fluorescent peak as well as from a decrease of the C3bright population (from 15.5 to 3.8%). Similar results were obtained for surface deposition of C4 (data not shown).
Peptide 2J inhibits xenoreactivity of human serum to pig cells. PK15 cells were incubated with normal human serum (1/20) in the presence or absence of peptide 2J or peptide 2J-3 (100 μM). Complement deposition on the cell surface was subsequently assessed by flow cytometry using a mAb directed against human C3. A, Representative profiles of staining for C3. Histograms with dashed lines represent staining for C3 after incubation with heat-inactivated serum. Histograms with solid lines represent incubation with normal serum in the presence or absence of peptides, as indicated. B and C, Quantification of data, expressed as mean fluorescence intensity (B) or median fluorescence intensity (C). The mean ± SD of duplicate analysis is shown. The dashed line indicates the background staining obtained in the presence of heat-inactivated serum.
Discussion
The present study describes a peptide that binds to C1q and is able to inhibit the classical pathway of complement activation by interfering in the interaction between C1q and IgG.
This peptide was selected out of a set of 42 phage-displayed peptides previously reported to bind C1q (16). It is striking that 30 of these 42 peptides showed inhibition of C1q activity in our assays, which is >70%. One-third of these peptides did not inhibit the classical pathway when tested in a CH50 assay, which is most likely a matter of sensitivity. Two peptides showed inhibition of the CH50 assay but not of the C1q-dependent hemolytic assay. These peptides did not show C1q binding in ELISA, and they may interfere in the complement activation test at a different level, or in a nonspecific way. Furthermore, a group of peptides appeared to inhibit not only the classical pathway but also the alternative pathway, as demonstrated for peptide 4B. Peptide 4B, as well as other peptides from this group (4D, 4E, 2C, 2D, 2E, 6B), showed a clear binding to C1q in ELISA. It is possible that the complement-inhibiting potential of these peptides is due to (additional) interactions at another step in the complement activation cascade that is currently unidentified. Because the alternative pathway assay was performed in a calcium-free buffer, the inhibitory potential of 4B and related peptides in this assay is probably independent of its interaction with C1.
Some of the peptides examined did not show any interaction with C1q in our assays, neither functionally nor in ELISA, such as peptide 2L. Lauvrak et al. (16) reported a similar binding strength to C1q for peptides 2J and 2L, whereas we could demonstrate strong C1q binding for the former peptide only. This apparent contradiction may be related to the fact that Lauvrak et al. (16) assessed binding of phage-displayed peptides to coated C1q, whereas our binding assays used coated peptides and fluid phase C1q. The conformation of peptides in soluble form may differ from that of peptides displayed on a phage.
Based on the strength of the inhibitory capacity of the peptide for C1q and based on the specificity, peptide 2J was selected as the most promising candidate for a specific inhibitor of the classical complement pathway. This peptide inhibits the recognition phase of the classical pathway, namely the binding of C1q, resulting in inhibition of subsequent activation of C4 and C3, formation of the membrane attack complex, and erythrocyte lysis. Peptide 2J is a 15-meric peptide containing two cysteines at positions 1 and 14. Our results show that linearized variants of the peptide are not active. In contrast, a stable cyclic variant showed C1q inhibitory activity and C1q binding. These results strongly suggest that peptide 2J is acting as a circular structure. Alternatively, it could be possible that peptide 2J forms active dimers or multimers via intermolecular S-S bridges. However, because the nonactive linear peptides 2J-1 and 2J-2 are able to form similar structures, whereas the active cyclic peptide 2Jcy is unable to form dimers, we consider this possibility as unlikely. Peptide 2Jcy is less active than the original peptide 2J, indicating that the two cysteines are important for the activity of peptide 2J, either by affecting the basic structure or size of the molecule or by a direct interaction with C1q. Interestingly, 31 of the 42 previously identified C1q-binding peptides contain at least two cysteines, and most C1q-binding peptides were identified in cysteine-constrained libraries (16).
Several other peptides have been described that bind C1q. An 11-meric peptide derived from the constant region of IgG1 inhibits the lysis of pig erythrocytes by human serum in the millimolar range (21). Furthermore, a dimeric peptide WY, derived from the CH2 domain of human IgG, inhibits lysis of EA, also in the millimolar range (22). Multimeric forms of these peptides had a strongly increased potency. The reported activity of the monomeric peptides is much less than that of peptide 2J, which is active in the micromolar range, although variation in the sensitivity of the different assay systems may explain part of these differences. Recently, a number of C1q-binding peptides were identified by an iterative panning and blocking strategy (23). Two different motifs were identified in C1q-binding peptides, i.e., the NPF motif ([N/S]PFxL) and the SHY motif, of which the former motif was also present in some peptides identified by Lauvrak et al. (Ref. 16 ; peptides 1B, 2B, and 2C, Table I⇑). Peptides containing the NPF motif were shown to induce complement activation and consumption in the fluid phase (24).
Inhibition of C1q activity can also be accomplished using purified or recombinantly expressed natural C1q-binding proteins, such as the endothelial C1q receptor (17), calreticulin (25), the globular C1q receptor (26), and a C1q-binding protein derived from Escherichia coli (27). Furthermore, C1q-binding peptides derived from the globular C1q receptor (26) and from calreticulin (25) have been shown to inhibit C1q hemolytic activity. A related strategy for the development of C1q inhibitors is the synthesis of protein parts of the C1q molecule that can act as a competitive inhibitor of ligand binding. For this purpose, a 12-meric peptide was developed from the B chain of the globular head region of C1q (ghB) that inhibits complement-mediated lysis of EA with an IC50 of 130 μM (22). Furthermore, the ghB expressed as a fusion protein with maltose binding protein inhibits the lysis of EA (28, 29), and slightly better results were obtained with a trimerized ghB protein (29). These recombinant proteins have the potency to be developed into a therapeutic C1q inhibitor. In comparison with peptides, the production and administration of such recombinant proteins is more complicated and expensive. Therefore, the potential applicability of peptides for therapeutic complement inhibition is something worthwhile to be considered.
Recently a peptide was identified that can inhibit the complement system at the level of C3 (12). Compstatin is a 13-meric cyclic C3-binding peptide that inhibits complement activation in vitro and in vivo. Treatment with compstatin significantly prolonged the survival of pig kidneys perfused with human blood (15). Furthermore, compstatin blocked complement activation in baboons induced by administration of heparin-protamine complexes (14). Recent data indicate that the cyclic structure of compstatin is important to prevent its breakdown by plasma proteases (30). The cyclic structure of 2J is required for its interaction with C1q but may in a similar way also be an advantage for its stability in plasma.
A major advantage of the use of C1q-specific peptides for complement inhibition in vivo is the specificity for classical pathway inhibition only. Activation of the complement system via the alternative pathway and via the lectin pathway is a major mechanism in innate immunity and takes place upon direct contact with microorganisms, without the need of Ab production. Therefore, these pathways are more than the classical pathway directly involved in prevention of infections. In view of the potential treatment of severely ill and immune-compromised patients with complement inhibitors, it is of great importance to allow activation of these pathways.
Treatment with peptide 2J will inhibit the earliest step of undesired classical pathway activation, i.e., the binding of the head domain of C1q to its target. This is of relevance in view of the proinflammatory effects of early products of the complement activation cascade, such as C4a (1). Furthermore, C1q binding may trigger effects mediated via C1q receptors, which are present on, e.g., phagocytes and endothelial cells (31). For the potential treatment of hyperacute rejection of xenografts, early inhibition of complement activation is advantageous, because endothelial cell activation induced by complement is most likely involved in various phases of graft rejection. Therefore, complement inhibition at the level of C1q, using fluid phase inhibitors, could be used in combination with transgenic expression of molecules, such as CD46, CD55, and CD59, which affect the complement system at a later stage.
The potential efficacy of peptide 2J in xenotransplantation is illustrated in this study by showing that the peptide strongly inhibits complement deposition on pig cells exposed to human serum. Following their exposure to human serum, C3 deposition on the pig cells shows marked heterogeneity, which is possibly related to heterogeneous expression of xenoantigens on these cells as well as variability in sensitivity to complement activation. Addition of peptide 2J significantly decreased complement activation, resulting in C3 binding of low intensity in a more homogenous pattern. This may indicate that activation of human complement on the pig cell surface is not completely dependent on the classical pathway. Alternatively, it could be a matter of peptide efficacy and/or stability. Assessment of the resistance of peptide 2J against degradation by plasma proteases, which is important for potential in vivo applications, requires experiments of longer duration (30).
Information from C1q-deficient patients and mice indicates that a lack of C1q is directly involved in the induction of systemic lupus-like autoimmunity (32). Recent discoveries point to a role of C1q in the clearance of apoptotic and injured cells, thus leading to persistence of damaged self material in the case of C1q deficiency. C1q binds to apoptotic cells via its head domain (Ref. 33)4. Therefore, it is important to consider the effects of potential C1q inhibitors on the complement-mediated clearance of self material. However, treatment of patients with complement inhibitors will in most cases take place on a short term to overcome a severe injury, the initial consequences of an acute disease, or an exacerbation of a chronic disease. We consider that the induction of autoimmunity could be a long-term consequence of C1q-targeted complement inhibition, requiring proper testing in animal models before patient studies should be undertaken. In this respect, it is a major advantage that peptide 2J inhibits C1q from rodent, primate, and human origin, which will facilitate experimental and preclinical testing of its efficacy and safety.
Although activation of the classical complement pathway is causally involved in a number of situations of harmful complement activation, other pathways are likely to be involved as well. The alternative pathway amplifies the complement activation cascade induced via any pathway at the level of C3. In addition, recent experiments in C1q knockout mice suggest that at least part of the glomerular complement deposition that accompanies the renal disease in these animals is induced via the alternative pathway (32). Furthermore, the lectin pathway can be involved in complement activation after endothelial oxidative stress (34). A peptide inhibitor of mannan-binding lectin that has been developed recently was able to attenuate endothelial complement binding induced by hypoxic treatment in vitro (35). Therefore, further definition of the contribution of the various complement pathways in specific pathological situations is of major importance for the development of an effective, specific, and safe treatment.
In conclusion, peptide 2J has promising properties for therapeutic complement inhibition because it specifically inhibits the classical complement pathway at the earliest possible level. Further studies will be undertaken to develop this peptide into an effective drug for in vivo use.
Acknowledgments
We thank Isabelle van der Borch tot Verwolde and Willemien Benckhuijsen for excellent technical assistance. Dr. W. Bogers, Dr. C. E. Hack, and Dr. T. E. Mollnes are acknowledged for providing valuable reagents and advice.
Footnotes
-
↵1 This work was supported by grants from the Dutch Organization for Scientific Research (901-12-094), the European Community (Biotech; Bio4-CTD97-2242), and the Dutch Kidney Foundation (PC95).
-
↵2 Address correspondence and reprint requests to Dr. Anja Roos, Department of Nephrology, Leiden University Medical Center, D3P, Postbox 9600, 2300 RC Leiden, The Netherlands. E-mail address: A.Roos{at}LUMC.NL
-
↵3 Abbreviations used in this paper: CR1, complement receptor-1; dig, digoxygenin-3-O-methylcarbonyl-ε-aminocaproic acid-N-hydroxysuccinimide ester; EA, Ab-coated erythrocyte; VBS, Veronal-buffered saline; CH50, hemolytic activity of the classical component pathway; AP50, hemolytic activity of the alternative complement pathway; ghB, B chain of the globular head region of C1q.
-
↵4 A. Nauta, M. Daha, O. Tijsma, R. Nieuwland, C. Hack, and A. Roos. C1q binding to apoptotic cells and cell blebs induces complement activation. Submitted for publication.
- Received July 20, 2001.
- Accepted October 19, 2001.
- Copyright © 2001 by The American Association of Immunologists
References
In this issue
