Cutting Edge: Human Anaphylatoxin C4a Is a Potent Agonist of the Guinea Pig But Not the Human C3a Receptor^1 2

The anaphylatoxins C3a, C4a, and C5a are small proteins of 74 to 77 residues that are generated in an inflammatory reaction by proteolytic cleavage from the complement components C3, C4, and C5, respectively. C3a and C5a are important chemotactic proinflammatory molecules, mediating smooth muscle contraction, increase in vascular permeability, and various cell activation and granule secretion reactions (for an overview, see Refs. 1 and 2).

Human C4a, first described in 1979, is regarded as the third anaphylatoxin (3) due to its structural similarity to C3a and C5a, its dependence on a carboxyl-terminal arginine residue for biologic activity, and its proinflammatory properties in guinea pigs. Although less active than C3a and C5a, in guinea pigs, human C4a induces smooth muscle contraction, increases vascular permeability (3), and induces granule secretion from platelets and O₂⁻ generation in macrophages (4, 5).

These biologic effects of C4a are subject to low dose desensitization (tachyphylaxis). Preincubation with substimulatory concentrations of C4a abrogates the functional response toward a subsequent 100% stimulus. In addition, cross-desensitization in guinea pig ileal contraction assays was observed between C4a and C3a, but not between C4a and C5a (3, 4). Based on these observations, C3a and C4a are thought to act on a common receptor. This view, however, has recently been questioned by Murakami and coworkers, who failed to detect desensitization of C3a-induced responses after pretreatment of guinea pig macrophages with human C4a (5). Furthermore, C4a did not inhibit binding of ¹²⁵I-labeled C3a to guinea pig macrophages (5). These findings would indicate distinct and separate receptors for C3a and C4a in guinea pigs.

In man, convincing evidence for anaphylatoxic effects of human C4a is still missing. C4a has been reported to inhibit C3a-induced chemotaxis of macrophages (6), although at extremely low concentrations (10⁻¹⁶ M!). Also, C4a and C4a analogue synthetic peptides lead to a dose-dependent wheal and flare generation when injected into human skin (3, 7), although no negative controls were included in these experiments. Evidence for C4a effects in species other than man or guinea pig has not been reported.

Recently, the C3aR from man (huC3aR)⁴ (8, 9, 10) and mouse (11, 12) have been cloned. Stably transfected cell lines expressing these receptors and human neutrophils have been tested for functional response towards serum-purified human C4a and found to be completely unresponsive (11, 13). This would suggest the existence of a separate C4a receptor in man and mouse, on the premise that C4a is an anaphylatoxin in these species as well. However, no positive control demonstrating biologic activity of the C4a preparation used in these experiments was presented.

Highly purified human C4a is difficult to prepare and, due to the high C4a concentrations required in most test systems, even trace contamination with other biologically active molecules, especially C3a and C5a, may jeopardize the experiments, as has been reported in one of the above-mentioned investigations for serum-purified C4a (13). This possible contamination may well be the cause for some of the discrepant results presented above. Using human rC4a and cells functionally expressing a cloned guinea pig (gp) or huC3aR, we provide evidence that C4a functions as an agonist of the gpC3aR but not the huC3aR.

Human C3a was obtained from Advanced Research Technologies (San Diego, CA), ¹²⁵I-labeled human C3a from NEN-DuPont (Boston, MA), and rC5a from Sigma (St. Louis, MO). The C3a carboxyl-terminal analogue synthetic peptide (WWGKKYRASKLGLAR, (W63,W64)C3a(63-77) (14)) was obtained from Bachem (King of Prussia, PA), and the C3a synthetic peptide P117 and control peptide P251 were prepared as described (9). Oligonucleotides were obtained from MWG Biotech (Ebersberg, Germany). N-terminal hexahistidine-tagged rC4a or rC4a with an N-terminal methionine residue was expressed in Escherichia coli and purified to homogeneity. The binding and functional activities of these two preparations of rC4a were equivalent.

Functional characterization of rC4a was performed in a guinea pig ATP release assay, as described (15). Guinea pigs from strains C2BB/R⁺ (C3aR positive) and C2BB/R⁻ (C3aR-negative) from our own breeding colonies were used as platelet donors (16). Desensitization was measured by preincubation of the platelets with the deactivating stimulus (a concentration leading to <10% ATP release, determined empirically at the beginning of each experiment) for 5 min at 37°C and subsequent addition of a 100% stimulus of either C3a (c = 10 nM), rC4a (c = 4 μM), or rC5a (c = 250 nM).

A partial DNA sequence of the gpC3aR was obtained by PCR amplification of genomic DNA using oligonucleotide primers derived from regions conserved between the human and mouse C3aR sequences (8, 9, 10, 11, 12). This fragment was used to screen a guinea pig genomic DNA library in λFIXII (Stratagene, La Jolla, CA). A λ clone containing a genomic insert, which encoded the gpC3aR, was identified. The open reading frame of this clone plus an extra 48 bp of genomic DNA sequence at the 3′ end was subcloned into pcDNA3 (Invitrogen, San Diego, CA) and designated pSL94.

Competitive binding assays were performed essentially as described (9). A microtiter plate-based calcium mobilization assay, utilizing a fluorometric imaging plate reader (FLIPR; Molecular Devices, Sunnyvale, CA), was used for the functional characterization of HEK-293 cells transiently expressing either the gpC3aR or huC3aR and Gα-16 (17). Briefly, cells were transfected using Lipofectamine Plus reagent (Life Technologies, Gaithersburg, MD), and the following day they were plated in poly-d-lysine-coated 96-well black/clear plates (Becton Dickinson, Bedford, MA). After 18 to 24 h, the medium was aspirated from each well, and the cells were loaded with Fluo-3-AM (Molecular Probes, Eugene, OR). At initiation of the assay, fluorescence is read once every second for 60 s and then every 3 s for the following 60 s. Agonist was added at 10 s, and the maximal fluorescence count above background after addition was used to define maximal activity for that concentration of agonist. FLIPR software normalizes fluorescence readings to give equivalent readings for all wells at zero time.

To avoid the problem of contamination in serum-purified human C4a preparations, we expressed rC4a. As shown in Figure 1,A, human rC4a was approximately three orders of magnitude less potent in a guinea pig ATP release assay than C3a, but induced ATP release at high nanomolar concentrations (ED₅₀ ≈ 400 nM), comparable with the ED₅₀ of 600 nM determined previously for serum-purified C4a in the serotonin release assay (4). Furthermore, in this assay, complete cross-tachyphylaxis between human C3a and rC4a, but not C5a, was observed (Fig. 1 B). In addition, platelets of the C3aR-negative guinea pig strain C2BB/R⁻, an inbred strain of our own breeding colonies with a functional C3aR deficiency (16), did not respond to either C3a or rC4a (data not shown). These data demonstrate the functional activity of the rC4a and confirm previous data on guinea pig cells showing cross-tachyphylaxis between C3a and C4a (3, 4). The simplest explanation of these data would be that C3a and C4a act through the same receptor. However, postreceptor mechanisms may account for the observed cross-tachphylaxis, as in the case of the FMLP-R and C5aR (18). We, therefore, set out to clone a C4a-binding C3aR from the guinea pig.

The coding regions of many G protein-coupled receptors, including the human C3aR (19), are not interrupted by introns. Therefore, we cloned the gpC3aR directly from a genomic DNA library. Using primer combinations conserved in the C3aR sequences of man (8, 9, 10) and mouse (11, 12), a genomic DNA fragment was amplified with high homology to the huC3aR. Using this fragment as a probe to screen a guinea pig library, a genomic λ clone was isolated with an open reading frame of 1428 bp, which encoded a protein of 475 residues with a calculated molecular mass of 53,570 Da and four potential N-glycosylation sites (Fig. 2). The sequence encodes a G protein-coupled receptor with seven hydrophobic transmembrane domains, a large second extracellular loop of 165 amino acid residues, and a high homology to the C3aR sequences of man, mouse, and rat (8, 9, 10, 11, 12, 20). However, only half of the residues are conserved in all four known C3aR sequences (240/475 = 50.5%), and only 37 of the 165 residues in the second extracellular loop (22.4%) in the gpC3aR are found at the same position in the other C3aR sequences. The peptide sequence of this gpC3aR was disclosed in a recently published book (2); however, the nucleotide sequence has not been published nor does it appear in GenBank, and no functional or binding data have been presented. It is unlikely that this receptor is a pseudogene, since by RT-PCR, expression of this transcript was detected in guinea pig heart, lung, and spleen RNA (data not shown).

HEK-293 cells transiently transfected with this receptor specifically bound human C3a. As shown in Figure 3, competitive displacement studies revealed the presence of a high affinity receptor for C3a with an apparent K_d of ∼2 nM. As previously shown with the huC3aR, this binding was specific for C3a, because ¹²⁵I-labeled C3a could be displaced in a dose-dependent manner by the synthetic C3a analogue peptide P117, but not by the unrelated peptide P251 (9). In addition, supramicromolar concentrations of rC4a (IC₅₀ > 1 μM) were able to competitively displace radiolabeled C3a from both the guinea pig and huC3aR. Although human C4a is able to bind weakly to both the gpC3aR and huC3aR, the affinity of this interaction is three orders of magnitude lower than the affinity of C3a binding to the same receptors.

To compare the functional activity of the gpC3aR and huC3aR, we cotransfected the cDNA for each receptor, together with a cDNA clone encoding Gα-16, into HEK-293 cells and tested for intracellular calcium mobilization in response to rC4a, rC5a, C3a, or a C3a analogue peptide. Previously, we reported that cotransfection of the huC3aR sequence with Gα-16 was necessary to render transfected HEK-293 cells responsive to C3a (9). As shown in Figure 4 A, cells expressing the gpC3aR responded in a dose-dependent manner not only to C3a (ED₅₀ = 0.18 ± 0.02 nM) and the C3a peptide (ED₅₀ = 0.15 ± 0.01 nM), but also to rC4a (ED₅₀ = 8.7 ± 0.52 nM). Cells transfected with the cDNA for the gpC3aR or Gα-16 alone did not respond to any of the peptides tested (data not shown).

In marked contrast to the results obtained with rC4a on cells expressing the gpC3aR, cells transiently expressing the huC3aR did not respond to rC4a concentrations as high as 1 μM, but did respond to C3a (ED₅₀ = 0.36 ± 0.07 nM) and the C3a peptide (ED₅₀ = 3.1 ± 0.3 nM; Fig 4 B). Previously, comparable results were obtained with stable cell lines expressing the cloned huC3aR (13). The gpC3aR and huC3aR responded differently to human rC4a as well as to the C3a synthetic peptide, which was virtually equipotent with C3a vs the gpC3aR but approximately an order of magnitude less potent than C3a vs the huC3aR. This may not be an unexpected result, as the biological assay originally used to characterize this “superagonist” C3a analogue peptide were all performed on guinea pig cells (14). Minimal activity was noted with rC5a on cells expressing either the gpC3aR or huC3aR, but only at supramicromolar concentrations (ED₅₀ > 10 μM).

These data confirm that the receptor naturally expressed on guinea pig platelets is promiscuous with respect to human C3a and rC4a, and they extend these observations to demonstrate that both anaphylatoxins are also potent agonists of the cloned gpC3aR. In contrast, rC4a is not active on the huC3aR or mouse C3aR (this study and 13 . The C3aR is a single-copy gene in man (19) and guinea pig (data not shown). Thus, human C4a cannot have a biologic effect in man that is mediated via the C3aR, whereas such a pathway apparently exists in guinea pigs. However, it remains to be determined whether gpC4a will also stimulate the gpC3aR.

Our data are difficult to reconcile with the findings of Murakami and coworkers who describe separate receptors for C3a and C4a on guinea pig macrophages (5), although these differences may result from the different reagents and assays used. While there are no additional reports describing a C4aR that is distinct from the C3aR, our data do not exclude such a possibility. Through an interaction with the C3aR, human C4a may function as an anaphylatoxin in guinea pigs, but not in mouse or man.

We thank D. Bitter-Suermann for his continuous support and helpful advice. The help of Katia Rech, Claudia Rheinheimer, C. Schiebl, and Kerstin Teschner is gratefully acknowledged.