C5a Initiates the Inflammatory Cascade in Immune Complex Peritonitis

Animals

Specific pathogen-free female BALB/c, FcγRIIB^−/−, FcR γ-chain^−/− (all on a BALB/c background), WBB6F-1 Kit^W/W-v, and WBB6-F₁-Kit^+/+ mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained under specific pathogen-free conditions. FcγRIII^−/− mice (on the C57BL/6 background) were purchased from The Jackson Laboratory and backcrossed to the BALB/c background (six generations). FcγRI^−/− (26) and C5aR^−/− (27) mice were generated as described and were backcrossed to the BALB/c background (six generations). All mice were used at 8–10 wk of age. Animal care was provided in accordance with National Institutes of Health guidelines. Animal studies were approved by either the Bezirksregierung Hannover (Hannover, Germany) or the Cincinnati Children’s Hospital Medical Center (Cincinnati, OH) institutional animal care and use committee.

Peritoneal reverse passive Arthus reaction

OVA (20 mg/kg body weight; Sigma-Aldrich, Munich, Germany) was injected i.v., followed by i.p. injection of IgG rich in Ab to OVA (800 μg/mouse; ICN, Costa Mesa, CA) exactly as previously described (20). Mice were killed at the indicated time points after injury, and the peritoneal cavity was lavaged with 2 ml of ice-cold PBS and 0.1% BSA. Cytospin slides (200 μl) were prepared from lavage fluid and stained using Diff-Quick (Baxter Merz & Dade, Dudingen, Switzerland). PMN numbers per microscopic field were counted (≥20 different fields). To block the C5aR and C5L2 in BALB/c mice, animals were treated with the C5aRA A8^Δ71–73 (28) (10⁻⁵ M, 200 μl i.v.); to exclusively block the C5aR (CD88), Fab of the neutralizing anti-C5aR mAb 20/70 (29) (2 × 10⁻⁶M, 200 μl i.v.) were administered 10 min (A8^Δ71–73) or 30 min (Fab 20/70) before the initiation of IC peritonitis and 2 h after IC challenge. To block the C3aR, the C3a receptor antagonist (C3aRA) N²-[(2,2-diphenylethoxy)-acetyl]-l-arginine (SB290157) (30) was given once (500 μg dissolved 200 μl of PBS and 0.5% DMSO i.p.) 1 h before initiation of the Arthus reaction. Neutrophil numbers in controls receiving 200 μl of PBS and 0.5% DMSO were as high as those in Ab controls. To block the functional activity of cytokine-induced neutrophil chemoattractant (KC) and/or MIP, 100 μg each of anti-KC and/or anti-MIP-2 polyclonal Abs or of control polyclonal Abs (R&D Systems, Wiesbaden, Germany) were injected i.p. 90 min before initiation of the Arthus reaction.

Reduction of peritoneal Mφ using clodronate

To selectively reduce the number of Mφ, liposomes containing dichloromethylene-bisphosphonate (clodronate; gift from Roche, Mannheim, Germany) were used according to published procedures (31)_. In brief, mice were treated twice i.p. with 100 μl of clodronate preparation or with 100 μl of PBS (72-h interval). The number of Mφ in peritoneal lavage fluid was determined 24, 48, 72, and 96 h after the last injection of clodronate or PBS by staining cytospins (55 × g, 10 min) of peritoneal lavage fluid with Turk’s solution. In addition, differential cell counts were determined by staining cytospins with Diff-Quick (Baxter Merz & Dade). We found >85% reduction of peritoneal Mφ 24–96 h after the last injection of clodronate, but no reduction of peritoneal Mφ in PBS-treated controls. The reduction of peritoneal Mφ was confirmed by flow cytometry using PE-conjugated F4/80 mAb or an isotype-matched control. Samples were analyzed on a FACSCalibur flow cytometer and were evaluated using the CellQuest software (BD Biosciences, San Diego, CA). Profound neutrophil accumulation was found in clodronate-treated mice, which declined to zero during the next 3 days. Thus, IC challenge was performed 4 days after the last clodronate treatment.

Chemokine and cytokine ELISAs

Immunoreactive KC and MIP-2 in peritoneal lavage fluid were quantified using commercially available ELISA kits (R&D Systems) according to the manufacturer’s protocol. The detection limit of the assays was 15.6 pg/ml.

CXC chemokine release from elicited peritoneal Mφ

BALB/c mice were injected i.p. with thioglycolate (3%, 2 ml). After 72 h, peritoneal cells were harvested from lavage fluid as previously described (20). Peritoneal cells were transferred into six-well plates and cultured overnight in RPMI 1640 medium to allow Mφ to adhere to the plastic surface. After washing twice with PBS and 0.1% BSA, cells were incubated with three different concentrations of recombinant human C5a (rhC5a; 10⁻⁷, 5 × 10⁻⁷, and 10⁻⁶ M; Sigma-Aldrich, St. Louis, MO) or IC (10, 50, and 100 μg/ml OVA/anti-OVA polyclonal rabbit Ab). Soluble IC were formed by incubation of OVA with rabbit anti-OVA in a molar ratio of 1:4 (32). The precipitate was centrifuged and then resuspended in medium at the indicated concentration. No detectable LPS was found in the C5a and IC preparations as determined by a Limulus amebocyte lysate assay (BioWhittaker, Walkersville, MD; <3 pg/ml).

Real-time RT-PCR for KC and MIP-2 gene transcription

RNA from thioglycolate-elicited peritoneal Mφ (1 × 10⁶) was prepared using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA) according to manufacturer’s recommendations. cDNA was generated from 2 μg of DNase-treated total RNA, using Superscript II reverse transcriptase (Invitrogen Life Technologies) following the manufacturer’s protocol. Gene expression levels were determined using iQ-SYBR Green Supermix (Bio-Rad, Hercules, CA) containing 5 μl of cDNA (diluted 1/30) and 500 nM primer in a total volume of 50 μl. Samples were amplified by 40 cycles at 58°C on an iCycler Real-Time PCR System (Bio-Rad). For data analysis, expression levels were normalized to β-actin expression. The following oligonucleotides were used: mouse KC (72-bp fragment): forward primer, 5′-TTCTCTGTGCAGCGCTGCTG-3′; reverse primer, 5′ CGCAGCTCATTGGCGATAGG-3′; and mouse MIP-2 (87-bp fragment): forward primer, 5′-TCAGTGCTGCACAGTTCACTG-3′; reverse primer, 5′-CATTGACAGCGCAGTTCACTG-3′.

Expression of C5aR and FcγR on peritoneal resident cells

To determine the impact of C5aR signaling on FcγR expression, naive BALB/c mice were injected with rhC5a (5 × 10⁻⁸ M i.p.). Binding studies with peritoneal Mφ and hC5aR transfected RBL-2H3 cells revealed indistinguishable binding affinities of rhC5a to murine CD88 and human CD88 (2.7 ± 0.2 × 10⁻⁹ M (28) vs 2.9. ± 0.1 × 10⁻⁹ M; data not shown). Two hours later, mice were killed, and the peritoneal cavity was lavaged with 2 ml of ice-cold PBS and 0.1% BSA. Receptor expression levels on peritoneal resident cells were determined by flow cytometry. Cells obtained from peritoneal lavage (1 × 10⁶) were fixed with paraformaldehyde and stained with the following Abs or isotype controls (Immunotech, Westbrook, ME): anti-F4/80 (PE-conjugated; Serotec, Raleigh, NC), anti-FcγRIIB (Ly17.2, clone K9.361 (33) Alexa 488-conjugated; provided by F. D. Finkelman, Cincinnati, OH); anti-FcγRIIB/III (clone 2.4G2. Alexa 488-conjugated; BD Pharmingen, San Diego, CA); and biotinylated anti-FcγRI (clone X54-5/7.1) (34). Binding of anti-FcγRI mAb was detected using streptavidin-FITC (BD Pharmingen).

To determine C5aR expression on resident peritoneal Mφ or resident peritoneal MCs, the following Abs or isotype controls (Immunotech) were used: anti-CD117 (c-Kit, clone 2B8, PE-conjugated; Pharmingen); anti-F4/80 (PE-conjugated; Serotec), and anti-C5aR (clone 20/70 (29), Alexa 488-conjugated). Samples were analyzed on a FACSCalibur flow cytometer (BD Biosciences), and data were evaluated using CellQuest Pro software.

Statistical analysis

Statistical analysis was performed using the SigmaStat version 2.0 statistical package (Jandel, Erkrath, Germany). All data are given as the mean ± SEM. First we tested for a normal distributed population using the Kolmogoroff-Smirnov test. To analyze differences between two normally distributed groups, an unpaired t test was used. Comparison of the means of more than two groups was performed using one-way ANOVA. When the mean values of the groups showed a significant difference, pairwise comparison was performed using Tukey’s test.

C5a receptor signaling is the critical initial event for neutrophil recruitment in IC peritonitis

We used the well-established reverse passive Arthus reaction peritonitis model to assess the importance of C5aR, C5L2, and C3aR signaling to initiation of the inflammatory response in IC disease. As a readout, we determined the recruitment of neutrophils into the peritoneal cavity, the hallmark of the Arthus reaction. The naive peritoneum consists mainly of Mφ, lymphocytes, and a small population of MCs. In the course of IC peritonitis, neutrophils accumulate in the peritoneal cavity and comprise 40–50% of the cells in the peritoneum 6 h after IC challenge (data not shown). To assess the impact of C5aR ligation on neutrophil attraction, we blocked the C5aR using Fab of the neutralizing anti-C5aR mAb 20/70 (29) and used C5aR^−/− mice (27). As shown in Fig. 1⇓a, neutrophil recruitment was abrogated. We have recently shown that C5aRA A8^Δ71–73 blocks the interaction of C5a with CD88 and with the second C5aR, C5L2 (28). Administration of C5aRA A8^Δ71–73 blocked neutrophil migration completely. To assess whether C3a contributes to neutrophil recruitment in IC peritonitis, e.g., through the activation of peritoneal resident cells, we blocked C3aR by a specific C3aRA (30). No reduction of PMN numbers occurred, arguing against an important role for this anaphylatoxin in IC peritonitis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Effect of C5aR blockade and FcγR deletion on PMN recruitment. a, Blocking the C5aR abrogates PMN recruitment in immune complex peritonitis. IC peritonitis was induced as described in Materials and Methods. To block the C5aR, mice were treated with the C5aRA A8^Δ71–73 or anti-C5aR Fab 20/70. To block the C3aR, mice were treated with the C3aRA SB290157. Mice were killed after 6 h, and neutrophil numbers per microscopic field were counted (≥20 different fields). In all subsequent figures, data are the mean ± SEM, and symbols denote significant differences between experimental groups (single symbol, p < 0.05; double symbols, p < 0.001; by ANOVA; n = 5–15/group). Significant differences were found between the following groups: ∗∗, BALB/c mice treated with A8^Δ71–73, anti-C5aR Fab 20/70 or C5aR^−/− mice vs untreated BALB/c mice. b, FcγRI and FcγRIII play codominant roles in neutrophil recruitment. Neutrophil recruitment was determined in BALB/c, FcγRI^−/−, FcγRIII^−/−, and FcR γ-chain^−/− mice as described in a. Significant differences: ++, BALB/c vs FcR γ-chain^−/− mice; ∗, BALB/c vs FcγRI^−/− mice (ANOVA: n = 5–6/group). c, Blocking of C5aR signaling reduces neutrophil recruitment in FcγRIIB^−/− mice. IC peritonitis was induced in BALB/c and FcγRIIB^−/− mice in the presence or the absence of C5aR blockade by anti-C5aR Fab 20/70. Significant differences: ∗∗, BALB/c vs FcγRIIB^−/− mice; ++, BALB/c vs BALB/c treated with anti-C5aR Fab; ††, FcγRIIB^−/− vs FcγRIIB^−/− mice treated with anti-C5aR Fab (by ANOVA; n = 5–6/group).

To determine the contribution of activating FcγR to neutrophil recruitment, we used mice with engineered deficiencies of 1) FcγRI, 2) FcγRIII, or 3) the FcR γ-chain, which leads to a combined functional deficiency of FcγRI and FcγRIII. These mice were backcrossed onto the BALB/c background. In FcγRI^−/− mice, neutrophil accumulation in the peritoneum was slightly (although significantly) reduced by 25% compared with BALB/c controls, whereas the reduction of neutrophils in FcγRIII^−/− mice was only 15% (and did not reach statistical significance; Fig. 1⇑b). However, neutrophil trafficking was abolished in FcR γ-chain^−/− mice, confirming our previous results (obtained with FcR γ-chain^−/− mice on the C57BL/6 background) (20). The fact that neither the absence of FcγRIII nor the absence of FcγRI (Fig. 1⇑b) reduced neutrophil recruitment substantially (in contrast to FcR γ-chain^−/− mice) strongly suggests a codominant role for activating FcγRI and FcγRIII in neutrophil recruitment.

To assess a possible interrelationship between C5aR and FcγR signaling, we investigated whether C5a affects FcγR-induced neutrophil recruitment. Deficiency of FcγRIIB resulted in a strikingly enhanced accumulation of neutrophils during IC peritonitis (Fig. 1⇑c). Neutrophil numbers were twice as high as in BALB/c controls. Notably, blocking of C5aR signaling prevented the increase in neutrophil migration into the peritoneum of FcγRIIB^−/− mice. These data strongly suggest that C5a acts upstream of FcγR signaling and that C5aR signaling is a prerequisite for FcγR-mediated inflammatory responses.

C5aR signaling on peritoneal Mφ regulates the expression of FcγR

In a search for mechanisms that account for this crucial role of C5a, we hypothesized that C5a up-regulates the expression of activating FcγR on peritoneal resident cells. To determine whether this is the case, we determined the impact of C5a challenge on the expression of activating FcγRs (FcγRI and FcγRIII) on peritoneal Mφ. Furthermore, we tested whether C5a down-regulates inhibitory FcγRIIB. In fact, the expression of activating FcγRs was up-regulated, whereas that of FcγRIIB was down-regulated (Fig. 2⇓) 2 h after i.p injection of C5a into BALB/c mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

C5aR signaling regulates the expression of activating and inhibitory FcγR. BALB/c mice were injected i.p. with rhC5a (5 × 10⁻⁸ M) or PBS (as a control). Cells were gated on F 4/80 and stained for FcγRI (mAb X54-5/7.1), FcγRIIB (mAb Ly 17.2), and FcγRIII (mAb 2.4G2) or the respective isotype controls and analyzed on a FACSCalibur (BD Biosciences). To prevent binding of FcγRIIB/III-specific mAb 2.4G2 to FcγRIIB, experiments were performed in the presence of a 10-fold molar excess of FcγRIIB-blocking mAb Ly 17.2. Data are representative of a total of three independent experiments.

Peritoneal Mφ and MCs contribute to PMN recruitment

Our finding that FcγRI contributes to PMN recruitment led us to hypothesize that C5aR and FcγR signaling in peritoneal resident cells is important for triggering neutrophil recruitment in IC peritonitis, because murine neutrophils do not express FcγRI (34). We assessed the contribution of peritoneal Mφ by depleting these cells using the well-established method of clodronate-induced apoptosis (31). To assess the importance of MCs, we used the MC-deficient mouse strain Kit^W/W-v along with MC-competent controls (Kit^+/+). In Kit^+/+ mice, neutrophil numbers increased substantially between 2 and 6 h after IC challenge. At 6 h, neutrophils comprised 40–50% of the cells in the peritoneum. Thereafter, neutrophil numbers declined over the next 24 h (Fig. 3⇓a). Either genetic deficiency of MCs (Kit^W/W-v), or pharmacological depletion of Mφ significantly inhibited neutrophil recruitment 4 and 6 h after IC challenge (Fig. 3⇓, a and b). However, the kinetics of inhibition were different. In MC-deficient Kit^W/W-v mice, neutrophil influx was virtually abrogated during the first 4 h, but increased with a slope similar to that in Kit^+/+ mice between 4 and 6 h. After depletion of Mφ, neutrophils migrated almost normally during the first 4 h; however, their trafficking was essentially blocked between 4 and 6 h. These data suggest that MCs contribute mainly to early neutrophil recruitment, whereas Mφ are needed to mediate neutrophil trafficking between 4 and 6 h after IC challenge. To ensure that the crucial effect of C5aR blocking is not restricted to the BALB/c background, we blocked C5aR signaling in Kit^+/+ and Kit^W/W-v mice. In both cases, we found neutrophil migration to be abolished (Fig. 3⇓b).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

a, Resident peritoneal cells contribute to PMN migration into the peritoneal cavity. Kinetics of neutrophil recruitment in MC-sufficient Kit^+/+, MC-deficient Kit^W/W-v, and Mφ-depleted Kit^+/+ mice. Mice were killed at the indicated time points, and neutrophil numbers were determined as described in Materials and Methods. Mφ depletion was achieved by two i.p. injections of clodronate, inducing apoptosis in peritoneal Mφ as described in Materials and Methods. Significant differences: ∗∗, Kit^+/+.vs Kit^W/W-v mice; ∗, Kit^+/+ vs Mφ-depleted Kit^+/+mice. b, C5aR blockade abrogates neutrophil recruitment in Kit mice. Kit^+/+ mice, Mφ-depleted Kit^+/+ mice, and Kit^W/W-v mice were treated with neutralizing anti-C5aR Fab 20/70. Neutrophil numbers were counted 6 h after IC challenge in peritoneal lavage fluid. Significant differences: ∗, Kit^+/+ vs Mφ-depleted Kit^+/+ (Kit^+/+ plus clodronate); ++, Kit^+/+ vs Kit^W/W-v; ††, Kit^+/+vs Kit^+/+ plus anti-C5aR; ‡‡, Mφ-depleted Kit^+/+ vs Mφ-depleted Kit^+/+ plus anti-C5aR; °°, Kit^W/W-v vs Kit^W/W-v plus anti-C5aR (by ANOVA; n = 5–6/group). c, C5aR is expressed by peritoneal Mφ and MCs. Resident cells harvested from peritoneal lavage fluid were incubated with anti-c-Kit mAb (MCs) or F4/80 mAb (Mφ) in combination with anti-C5aR Fab 20/70 and analyzed by flow cytometry.

To ensure that C5a is able to activate resident peritoneal cells, we assessed C5aR expression on peritoneal Mφ and peritoneal MCs. Although some reports have suggested the presence of C5aR expression on murine MCs (35), the expression of C5aR on resident peritoneal MCs has not yet been determined. As shown in Fig. 3⇑C, peritoneal-derived Mφ and MCs clearly express the C5aR. Of the Mφ population (F4/80⁺ cells), 98–99% of cells are C5aR⁺, whereas 80–85% of the MC population (c-Kit⁺; 2% of total cells) express C5aR. Thus, resident peritoneal cells express activating FcγR (36, 37) as well as C5aR.

CXC chemokines KC and MIP-2 are important secondary mediators of neutrophil recruitment

To begin to investigate the mechanism by which resident cells contribute to peritoneal recruitment of neutrophils, we determined the release of chemoattractant chemokines into the peritoneum during IC peritonitis. We focused on the CXC chemokines KC and MIP-2. Neither of the chemokines was detected during the first 10 min after IC challenge (Fig. 4⇓), suggesting that these mediators are not released from preformed stores. Measurable amounts of KC and MIP-2 appeared after 1 h. Between 1 and 2.5 h, KC and MIP-2 concentrations continued to rise, thereafter declining to reach baseline levels after 6 h. In the absence of MCs, KC and MIP-2 levels started to increase later (at 2 h), reached only 30% of the maximum in MC-competent animals, and had returned to baseline levels after 4 h (Fig. 4⇓, a and b). These data suggest that MCs contribute significantly to the early release of KC and MIP-2. Mφ depletion did not affect KC release (Fig. 4⇓a). However, it diminished the release of MIP-2 similar to that seen in the absence of MCs (Fig. 4⇓b), indicating that Mφ contribute significantly to MIP-2, but not to KC, release.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Mφ and MCs contribute to KC (a) and MIP-2 (b) release during IC peritonitis. Kinetic of KC and MIP-2 release during IC peritonitis. Chemokine concentrations were determined by ELISA in the peritoneal lavage fluid of Kit^+/+ mice, Kit^W/W-v mice, and Mφ-depleted Kit^+/+ mice. Samples were taken at the indicated time points. Significant differences: ∗ or ∗∗, Kit^+/+ vs Kit^W/W-v mice; + or ++, Kit^+/+ vs Mφ-depleted Kit^+/+ mice (by ANOVA; n = 5–6/group). c, KC and Mip-2 contribute to PMN recruitment. Kit^+/+ mice were treated with neutralizing anti-KC and anti-MIP-2 polyclonal Ab, either alone or in combination or with a polyclonal control Ab i.p. 90 min before the initiation of IC peritonitis. Neutrophil numbers were measured 6 h after IC challenge. Significant differences: ∗, Kit^+/+ vs Kit^+/+ mice treated with neutralizing anti-KC and anti-MIP-2 Ab (by ANOVA; n = 5–6/group).

To determine whether these chemokines contribute to neutrophil recruitment, we neutralized these mediators with specific polyclonal Ab. Inhibition of KC and MIP-2 reduced neutrophil accumulation in the peritoneal cavity by 50% (Fig. 4⇑c). However, blockade of KC or MIP-2 alone did not affect neutrophil recruitment. These data suggest that Mφ and MCs contribute significantly to neutrophil recruitment during IC peritonitis by mechanisms that involve CXC chemokines.

C5aR signaling and FcγR signaling in resident cells is critical to induce CXC chemokines

Our data suggest that MCs and Mφ are an important source of KC and MIP-2 in IC peritonitis (Fig. 4⇑). As we found C5aR expression on both cell types, we wondered whether C5aR signaling in these cells is an important trigger for CXC chemokine production and release. C5aR blockade decreased KC and MIP-2 concentrations by 66% 2.5 h after IC challenge (Fig. 5⇓a). Of note, KC and MIP-2 levels did not increase at later time points, ruling out the possibility that C5aR blockade results in a delay of CXC chemokine release. Although these data provide no direct evidence, they strongly suggest that C5aR ligation on resident cells contributes to CXC chemokine release. Further, the data indicate that signaling through C5aR on MCs triggers the release of KC, as depletion of MCs, but not of Mφ, impairs the release of KC (Fig. 5⇓a, left panel). Finally, the release of MIP-2 is most likely to depend on C5aR ligation on MCs and Mφ, because both cell populations contribute to MIP-2 secretion (Fig. 5⇓a, right panel).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Impact of C5aR blockade (a) or genetic ablation (b) of activating FcγRs on KC and MIP-2 release. a, Kit^+/+ mice were treated with neutralizing anti-C5aR Fab 20/70. KC and MIP-2 concentrations were determined 2.5 h after initiation of IC peritonitis. Significant differences: KC: ∗∗, Kit^+/+ vs Kit^W/W-v or Kit^+/+ plus anti-C5aR; ++, Mφ-depleted Kit^+/+ vs Kit^W/W-v, or Kit^+/+ and anti-C5aR; MIP-2: ∗∗, Kit^+/+ vs all other groups (by ANOVA; n = 5–6/group). b, KC and MIP-2 concentrations were determined in the peritoneal lavage fluid of BALB/c, FcγRI^−/−, FcγRIII^−/−, and FcR γ-chain^−/− mice 2.5 h after IC challenge. Significant differences: ∗∗, BALB/c vs FcR γ-chain^−/− mice (by ANOVA; n = 5–6/group).

Furthermore, we investigated the contribution of activating FcγR to KC and MIP-2 release. KC and MIP-2 levels in FcR γ-chain^−/− mice were reduced by 60% compared with those in BALB/c mice (Fig. 5⇑b). However, KC and MIP-2 levels in FcγRI^−/− or FcγRIII^−/− mice were unchanged. These data suggest a significant role for FcγRI and FcγRIIII in CXC chemokine release. To ascertain that the chemokine release in Kit^+/+ mice (Fig. 4⇑) and BALB/c mice is comparable, we assessed the kinetics and the concentrations of KC and MIP-2 in BALB/c mice. We found that they were almost identical with those determined in Kit^+/+ mice (data not shown).

Cross-talk between C5aR and FcγR signaling amplifies effector functions in peritoneal Mφ

To assess whether the induction of CXC chemokine release through C5a in vivo results from direct C5aR signaling on peritoneal resident cells, thioglycolate-elicited peritoneal Mφ were stimulated with different concentrations of C5a. We found dose- and time-dependent production of MIP-2 (Fig. 6⇓, a and b) and KC (similar to MIP-2; data not shown). To assess whether C5a regulates the production of the chemokines at the transcriptional level, we performed real-time RT-PCR. We found some mRNA induction as early as 15 min after C5a stimulation, which strongly increased over time. Two hours (MIP-2) or 1 h (KC) after C5a challenge, mRNA transcripts were 35.4 ± 4.2-fold (MIP-2) or 30.3 ± 3.8-fold (KC) higher compared with medium controls (Fig. 6⇓c). These data provide evidence that C5a induces de novo synthesis of KC and MIP-2 from peritoneal Mφ and regulates the production of these chemokines at the mRNA level.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

C5aR and FcγR signaling synergize in the induction of MIP-2 release from peritoneal Mφ. a, Mφ were incubated with increasing amounts of C5a or IC. MIP-2 concentrations were determined in supernatants by commercially available ELISA after 2-h incubation. b, Kinetics of spontaneous, rhC5a-induced (10⁻⁶ M), or IC-induced (100 μg/ml) MIP-2 release. MIP-2 concentrations in supernatants were determined 10, 30, 60, and 120 min after rhC5a or IC challenge. c, Kinetics of C5a-induced (5 × 10⁻⁸ M) mRNA expression of KC (left panel) and MIP-2 (right panel) as determined by real-time RT-PCR.

Furthermore, we assessed the impact of OVA/anti-OVA ICs on chemokine production by elicited peritoneal Mφ. Similar to C5a, we found a dose- and time-dependent induction of both chemokines (Fig. 6⇑, a and b). Finally, we stimulated peritoneal Mφ with a combination of C5a and IC. As shown in Fig. 6⇑a, MIP-2 and KC release increased substantially. In fact, it exceeded by far the sum of the amounts induced by C5aR or FcγR ligation alone.

Critical importance of complement and both activating FcγRs in neutrophil recruitment

All available data suggest that the main contribution of complement to IC inflammation is through C5aR (CD88) signaling (4). However, recent findings that a second C5aR, C5L2, provides high affinity binding sites for C5a and its degradation product, C5adesArg (15), raises the possibility that C5L2 signaling accounts for some of the C5a effects. In particular, blocking or deletion of C5aR (CD88) does not rule out that the reduced inflammatory response results from inhibitory signaling through C5L2 (similar to the inhibitory effect of FcγRIIB on activating FcγRs). Our findings that ablation of C5aR signaling as well as ablation of C5aR- and C5L2-signaling abrogate neutrophil recruitment argue against such an inhibitory role of C5L2.

C3a is not a chemoattractant for neutrophils. However, it is a powerful chemoattractant and activator of MCs and Mφ, suggesting a pathogenetic role for C3aR signaling in IC disease. In fact, C3a was shown to attract neutrophils through such indirect pathways (40). However, blocking of the C3aR pathway had no impact on neutrophil migration, suggesting that C3aR signaling plays no or at best a redundant role in the pathogenesis of IC peritonitis.

In most models of IC disease, signaling through FcγRIII is critical for the inflammatory phenotype (1, 7, 23). Studies with FcγRI^−/− mice revealed a critical role for FcγRI in inflammation and hypersensitivity (26, 41). A role for FcγRI in IC peritonitis was suggested by comparison of two mouse strains: 1) mice with a combined functional deficiency of FcγRI and FcγRIII (lacking the common γ-chain), and 2) FcγRIII^−/− mice (20). Data from this study suggested a dominant role for FcγRI and only a minor role for FcγRIII in IC peritonitis. However, because the γ-chain is not exclusively used to transmit signals downstream of activating FcγRs, but for many other receptors as well (e.g., activating isoforms of the paired Ig receptor (42), and certain subsets of Ig-like transcripts (43)), these data do not necessarily reflect the role of FcγRI. Further, codominant effects of both activating FcγRs may not be adequately addressed by this approach. Comparing the impacts of deletion of FcγRI, FcγRIII, and FcR γ-chain on neutrophil recruitment, we found that deletion of either activating FcγR had only a minor effect on neutrophil attraction; however, deletion of the γ-chain abrogated neutrophil migration. In contrast to other IC disease models (1, 7), these data suggest a codominant role for both activating FcγRs in the inflammatory response, although we cannot rule out at this point that the activation of other receptors, the signaling of which is γ-chain dependent, may account for this effect as well.

Complement activation is a crucial event upstream of FcγR activation

Much of the data available from experimental IC disease support a model in which C5aR signaling and FcγR signaling form a network that orchestrates the inflammatory phenotype (1, 7, 19, 20, 21, 22, 23, 24, 38, 39). In contrast, some data suggest an exclusive role for FcγR (5, 6, 8, 44), which has led to the imputation of distinct roles for complement and FcγR in autoimmunity and infection (25). Our recent data (24) along with that from the current study may help resolve this apparent discrepancy. In fact, we found a direct regulatory effect of C5aR signaling on the balance of activating (FcγRIII) and inhibitory (FcγRIIB) FcγR expression in IC alveolitis (24). C5a up-regulates FcγRIII and down-regulates FcγRIIB on alveolar Mφ, thereby reducing the threshold for FcγR activation. In this study, we show that the same regulatory pathway is present in peritoneal Mφ, and that C5aR signaling up-regulates the expression of FcγRI as well. Thus, C5a modulates the entire FcγR network, suggesting a general mechanism in IC disease. Although other factors have been described that regulate the balance between activating FcγRs (such as TNF-α and IFN-γ) and inhibitory FcγRs (IL-4), they play no role in the initial events leading to IC-induced inflammation. As we have shown previously (20), TNF-α blockade has no impact on PMN recruitment. Further, we found no IFN-γ or IL-4 in peritoneal lavage samples taken during the first 6 h after IC challenge (our unpublished observations). In further support of a critical role for C5a in regulating FcγR expression, we found that C5aR blockade in FcγRII^−/− mice (which have a markedly increased inflammatory phenotype) reduced neutrophil recruitment to the level in wild-type mice. Together these data suggest a model in which C5a acts upstream of FcγR activation by setting the threshold for FcγR activation in IC disease (Fig. 7⇓). This model integrates the findings that the inflammatory phenotype in anti-glucose-6-phosphate isomerase arthritis (1), autoimmune vitiligo (38), autoimmune hemolytic anemia (39), and the various models of the Arthus reaction (4) depend on both C5aR and FcγR signaling. In a broader sense, data that link the development of many autoimmune disease to impaired FcγR regulation and function (for review, see Ref. 45), such as Goodpasture’s syndrome, autoimmune arthritis, and systemic lupus erythematosus, may be integrated by including complement activation into the network of FcγR regulation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Model of the C5aR/FcγR network that controls the inflammatory response in IC disease. IC activate the complement system, resulting in the generation of C5a. Ligation of the C5aR results in activation of signaling pathways that up-regulate activating FcγRs and down-regulate inhibitory FcγRIIB, thus shifting the balance toward a proinflammatory phenotype. Aggregation of activating FcγRs together with C5aR signaling leads to the release of secondary chemotactic mediators of the CXC chemokine family. CXC chemokines in concert with C5a eventually recruit neutrophils into the peritoneum and orchestrate the inflammatory response.

Downstream of C5aR and FcγR signaling: resident peritoneal cells are important

The pathways downstream of C5aR and FcγR signaling are poorly understood, as are the qualitative and quantitative contributions of resident tissue cells, such as MCs and Mφ. Depletion of MCs results in significant reduction of the inflammatory response in Arthus reaction models (46, 47, 48, 49). In the absence of MCs, neutrophil recruitment is delayed, and the maximum response is only 50% of that in controls (Ref. 50 and this study). The pathways that activate MCs are incompletely understood. Experiments in which MCs from FcR γ-chain-deficient mice were adoptively transferred to wild-type mice indicated a crucial role for FcγRIII in a cutaneous Arthus model (37). Depletion of complement by cobra venom factor or genetic ablation of C5 prevents degranulation of peritoneal MCs, suggesting a mechanistic role for C5 (35). However, C5aR expression has only been described on distinct subtypes of MCs, including skin MCs (51), a subfraction of cardiac MCs (52), synovial MCs in rheumatoid arthritis (53), and neoplastic MCs in patients with MC neoplasms (and the human MC line HMC-1) (52). By contrast, no expression of C5aR was found on lung MCs or MCs in other visceral organs. In this study, we demonstrate C5aR expression on resident peritoneal MCs. These data along with our finding that specific blockade of the C5aR reduces KC release in Kit^+/+ mice to the same extent as MC depletion (in Kit^W/W-v) and our data showing that KC release is critically dependent on the presence of MCs suggest an important role for C5aR signaling in MC activation in vivo (see below).

The remaining neutrophil migration seen in the absence of MCs points toward another cell type contributing to neutrophil migration. The codominant roles of FcγRI and FcγRIII in IC peritonitis imply a significant contribution of tissue Mφ in neutrophil attraction. In support of this view, we found a significant reduction of neutrophil recruitment after Mφ depletion. Although previous studies (19, 20, 21) suggested a significant role for tissue Mφ, we now provide evidence that this is indeed the case. Of note, Mφ depletion had a negative impact on neutrophil elicitation that became evident only >4 h after IC challenge, thus highlighting the important role of Mφ in late (>4 h) neutrophil recruitment.

Downstream of C5aR and FcγR signaling: CXC chemokines are critical

Ligation of C5aR and activating FcγRs results in pleiotropic proinflammatory effector functions, including the release of cytokines and chemokines (3, 9). Of particular interest are the CXC chemokines that harbor the conserved glutamic acid-leucine-arginine tripeptide motif, such as KC and MIP-2. These chemokines bind to CXCR2, which belongs to the group of chemoattractant, G protein-coupled receptors present on neutrophils (54), MCs (55), and Mφ (56). Substantial release of these chemokines has been demonstrated in mouse (23, 24) and rat (32) models of IC alveolitis, something that was markedly reduced in the absence of C5aR signaling (24, 32). In line with these data, we found huge local accumulation of KC and MIP-2 in IC peritonitis that was reduced by C5aR blockade. These in vivo data were mirrored by in vitro data showing that C5a is a strong, FcγR-independent inductor of KC and MIP-2 release from peritoneal Mφ (in contrast to alveolar Mφ (23)). Furthermore, we found that both activating FcγRs (FcγRI and FcγRIII) promote the synthesis of MIP-2 and KC in vivo and in vitro, matching our data for neutrophil recruitment. Again, these data are in contrast to IC alveolitis, in which FcγRIII was found to be crucial for MIP-2 release. Only when both chemokines were blocked in vivo did neutrophil migration decrease substantially, suggesting that a certain amount of CXCR2 triggering is sufficient to mediate efficient neutrophil migration. However, because neutrophil recruitment was affected by, at best, 50%, other CXCR2 ligands (e.g., LPS-inducible CXX chemokine) may contribute to neutrophil migration as well. Consequently, the contribution of CXCR2 signaling to neutrophil trafficking may be underestimated. Importantly, combined activation of C5aR and FcγR synergized in promoting MIP-2 and KC release from peritoneal Mφ in vitro. This synergism may result from C5a-mediated reduction of the threshold for FcγR activation and/or a positive feedback between C5aR signaling and signaling pathways of activating FcγRs. Together our data demonstrate that CXCR2 signaling is a critical effector mechanism downstream of C5aR/FcγR signaling. Furthermore, our data provide a mechanistic link between C5aR/FcγR signaling, Mφ activation, CXC chemokine release, and neutrophil recruitment.

In summary, we propose a model in which C5a acts upstream of FcγRs, initiating the inflammatory response by acting as a chemoattractant for neutrophils and by activating local MCs and Mφ to release CXC chemokines. In this model, C5a functions as the dominant regulator of FcγR expression decreasing the threshold for FcγR activation (Fig. 7⇑). This regulation is the prerequisite for adequate inflammatory function of the FcγR system, because it shifts FcR signaling from the inhibitory, ITIM-dominated phenotype to the activating, ITAM-dominated phenotype. Understanding the mechanisms underlying the complex network of C5aR and FcγR regulation is likely to lead to novel, urgently needed therapeutics for autoimmune diseases.