C5aR2 Activation Broadly Modulates the Signaling and Function of Primary Human Macrophages

The complement activation fragment C5a is a potent proinflammatory mediator and an increasingly recognized immune modulator. Human C5a is a 74-aa glycoprotein that binds to the seven-transmembrane domain receptors C5aR1 and C5aR2 (1). Both receptors are coexpressed on a broad range of cells, including myeloid cells (such as neutrophils, subpopulations of monocytes, and macrophages), certain lymphocytes, and many nonmyeloid cells (2). In a highly context- and cell type–dependent fashion, C5a powerfully modulates multiple aspects of immune cell functions (3). The dysregulation of the C5a-C5a receptor axis has been associated with a myriad of acute and chronic inflammatory diseases (4) and is, therefore, an active target for therapeutic drug development.

C5aR1 and C5aR2 display similar binding affinities toward C5a (5–8). From a molecular perspective, the primary C5a receptor, C5aR1, is a canonical G protein–coupled receptor (GPCR), coupling to Gα_i2 and Gα₁₆ (2, 4). C5a ligation to C5aR1 downregulates cAMP/protein kinase A signaling (9, 10), activates the Raf-MEK1-ERK1/2 cascade through multiple mediators (11–14), triggers intracellular calcium mobilization (15, 16), and recruits β-arrestins (17). The signaling mechanisms of C5aR2, however, remain to be clearly elucidated. In contrast to C5aR1, C5aR2 is commonly acknowledged to be incapable of G protein coupling because of its absence of the highly conserved DRY and NPXXY motifs that are indispensable for G protein recognition/coupling in Class A GPCRs (4, 18–21). Multiple studies have pointed to the roles of β-arrestins in C5aR2 signaling (21–26), whereas the secondary messengers downstream of β-arrestins remain inconclusive. Activated C5aR2, potentially through β-arrestins and heterodimerization with C5aR1, downregulates C5aR1-mediated ERK1/2 signaling (25–27).

Whereas the proinflammatory function of C5aR1 is largely undisputed, the potential roles played by C5aR2 remain enigmatic (21, 28, 29). One of the major factors that hindered progress in the C5aR2 field was the lack of selective pharmacological modulators (21, 28). In 2016, the first C5aR2-selective functional ligands were revealed by screening a panel of 61 linear peptides based on the C terminus of C5a (8). The lead compound, termed P32 (Ac-RHYPYWR-OH; Fig. 1), is a C5aR2 agonist that binds to C5aR2 within the high-micromolar range and displays selective partial C5aR2-agonistic activities by recruiting β-arrestin 2 (EC₅₀=5.2 μM) without triggering C3aR or C5aR1 activation (30). This compound has since served as a pharmacological tool to decipher the roles of C5aR2 in physiology and pathophysiology (8, 31–33). Functionally, P32 dampens C5a-mediated ERK1/2 phosphorylation and LPS-induced IL-6 release in human macrophages (8), and by acting through surface-expressed C5aR2, downregulates intracellular C5a-C5aR1 signaling and the subsequent canonical NLRP3 inflammasome activation and associated Th1 induction in CD4⁺ T lymphocytes (21, 31).

Despite studies demonstrating C5aR2 can modulate C5aR1 (and concomitant TLR4 and NLRP3) signaling in human cells, it is not yet known whether C5aR2 functionality is limited to C5aR1 activation or influences immune cells more broadly. The present study, therefore, used P32 to characterize, in depth, the innate immune roles of C5aR2 on primary human monocyte–derived macrophages (HMDMs), both on cell signaling and function. We found that although C5aR2 activation by itself was devoid of any detectable signaling activities, C5aR2 agonism dampened C5aR1-, C3aR-, and chemokine-like receptor 1 (CMKLR1)–mediated ERK signaling and altered calcium mobilization mediated by these receptors. Functionally, C5aR2 activation also predominantly downregulated cytokine responses triggered by various TLRs (TLR2, TLR3, TLR4, and TLR7), C-type lectin receptors (Dectin-1, Dectin-2, and Mincle), and the cytosolic DNA sensor stimulator of IFN genes (STING). This study, therefore, reveals that C5aR2 possesses pleiotropic immune-dampening functions in primary human macrophages.

Recombinant human C5a was purchased from Sino Biological (Beijing, China). Isolated human C3a, recombinant human chemerin, leukotriene B4 (LTB4), and fMLF was purchased from Merck (Perth, Australia). The panel of pattern recognition receptor (PRR) ligands, as detailed in Table I, were purchased from InvivoGen (San Diego, CA). BSA was purchased from Merck. For cell culture, trypsin-EDTA, Ham’s F-12 and IMDM, and penicillin-streptomycin were purchased from Thermo Fisher Scientific (Melbourne, Australia). Dulbecco PBS was purchased from Lonza (Melbourne, Australia).

The C5aR2 agonist P32 (Ac-RHYPYWR-OH) was assembled on 2-chlotrityl chloride resin (CSBio) using manual solid-phase peptide synthesis using Fmoc-based chemistry. The first Fmoc-protected amino acid (1 M equivalent relative to the resin) was dissolved in dichloromethane (DCM, 2 ml), and diisopropylethylamine (DIPEA, four equivalents) was added. After complete dissolution of the amino acid, the mixture was added to the resin and shaken for 2 h. The resin was washed with DCM/methanol/DIPEA (17:2:1, 3 × 10 ml), DCM (2 × 10 ml), and then flow washed with N,N-dimethylformamide (DMF) for 1 min. The Fmoc protecting group was removed by shaking the resin with 20% piperidine/DMF mixture (2 × 10 ml, each cycle for 1 min). After deprotection, the resin was flow washed with DMF for 1 min. The next amino acid (four equivalents) was activated with 0.5 M HBTU solution in DMF (four equivalents) and DIEA (four equivalents) and was added to the resin and the mixture shaken for 10 min. After completion of peptide chain assembly, the terminal Fmoc group was removed, and the N terminus was acetylated by incubation of the resin with a mixture of acetic anhydride (870 μl), DIPEA (470 μl), and DMF (13 ml) (2 × 15 min). The peptide was cleaved from the resin by shaking with 20 ml of cleavage mixture (95% trifluoroacetic acid/5% water) for 2 h. The trifluoroacetic acid was then evaporated, and the peptide was precipitated with ice-cold ether, filtered, dissolved in 50% buffer A/B (Buffer A, H₂O/0.05% trifluoroacetic acid; Buffer B, 90% CH₃CN/10%H₂O/0.045% trifluoroacetic acid), and lyophilized. Crude peptide was purified by reversed-phase high-performance liquid chromatography on a Phenomenex C18 column using a gradient of 0–80% B in 80 min, with the eluant monitored at 215 and 280 nm. Electrospray-mass spectroscopy confirmed the molecular mass of the fractions collected, and those displaying the correct molecular mass of linear peptide were pooled and lyophilized. Analytical reversed-phase high-performance liquid chromatography and electrospray-mass spectroscopy confirmed the purity and molecular mass of the synthesized peptide.

HMDMs were generated and cultured as previously described (27). Briefly, human buffy coat blood from anonymous healthy donors was obtained through the Australian Red Cross Blood Service (Brisbane, Australia). Human CD14⁺ monocytes were isolated from blood using Lymphoprep (STEMCELL, Melbourne, Australia) density centrifugation followed by CD14⁺ MACS magnetic bead separation (Miltenyi Biotec, Sydney, Australia). The isolated monocytes were differentiated for 7 d in IMDM supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 15 ng/ml recombinant human M-CSF (Lonza) on 10-mm square dishes (Thermo Fisher Scientific). Nonadherent cells were removed by washing with DPBS, and the adherent-differentiated HMDMs were harvested by gentle scraping. Chinese hamster ovary cells stably expressing human C5aR1 (CHO-C5aR1) were maintained in Ham’s F-12 medium containing 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 400 μg/ml G418 (InvivoGen, San Diego, CA) in T175 flasks as previously described (34).

Ligand-induced phosphokinase signaling was assessed using the respective AlphaLISA Surefire Ultra phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴), p-AKT1/2/3 (Ser⁴⁷³), p-AKT1/2/3 (Thr³⁰⁸) and p-p38 MAPK (Thr¹⁸⁰/Tyr¹⁸²) kits following the manufacturer’s protocols (PerkinElmer, Melbourne, Australia). Briefly, HMDMs or CHO-C5aR1 cells (50,000 per well) were seeded in tissue culture–treated 96-well plates (Corning, Corning, NY) for 24 h and serum-starved overnight. All ligand dilutions were prepared in serum-free IMDM containing 0.1% BSA. For CHO-C5aR1 cells, serum-free F-12 medium was used instead. For stimulation, cells were incubated with respective ligands for 10 min or as otherwise stated at room temperature and then immediately lysed using AlphaLISA lysis buffer on a microplate shaker (450 rpm, 10 min). For the quantification of individual phosphokinases, cell lysate (5 μl per well) was transferred to a 384-well ProxiPlate (PerkinElmer) and combined with the respective donor and acceptor reaction mixes (2.5 μl per well, respectively), followed by a 2-h incubation at room temperature in the dark. On a Tecan Spark 20M (Männedorf, Switzerland), upon laser irradiation of donor beads at 680 nm, the chemiluminescence of acceptor beads at 615 nm was recorded.

Ligand-induced intracellular calcium mobilization was assessed using Fluo-4 NW Calcium Assay Kit (Thermo Fisher Scientific) following the manufacturer’s instructions. Briefly, HMDMs were seeded (50,000 per well) in black clear-bottom 96-well plates (Corning, Corning, NY) for 24 h before the assay. Cells were first stained with the Fluo-4 dye in assay buffer for 30 min (37°C, 5% CO₂). Respective ligands were prepared in assay buffer containing 0.1% BSA. Cells were then pretreated with assay buffer/BSA only (control) or 100 μM P32 for an additional 20 min before commencing recording. On a Flexstation 3 platform, the fluorescence (Ex/Em: 494/516 nm) was continually monitored for a total of 100 s with respective agonists added at 16 s. Data were recorded as the magnitude of signal deviation from the baseline.

The immunomodulatory effect of C5a was assessed in HMDMs as previously described (8). All treatment ligands were prepared in serum-free IMDM containing 0.1% BSA. HMDMs (100,000 per well), seeded in 96-well tissue culture plates (Corning), were treated with respective PRR ligands at final concentrations detailed in Table I in the absence or presence P32 (100 μM) for 24 h (37°C, 5% CO₂). Cell culture supernatant was collected and stored at −20°C till use. The TNF-α, IL-6, and IL-10 levels in the supernatant were quantified using respective ELISA kits (BD Biosciences, Sydney, Australia) as per the manufacturer’s protocols.

All experiments were conducted in triplicates using HMDMs derived from different donors (donor n-values for each experiment are provided in figure legends). Data were analyzed using GraphPad software (Prism 8.4.2) and expressed as mean ± SEM unless otherwise stated. For the signaling assays, data from each individual repeat was normalized accordingly before being combined. For all dose-response studies, logarithmic concentration-response curves were plotted using combined data and analyzed to determine the respective potency values. For the cytokine release assays, the original measurements from each donor were combined where indicated. Statistical analysis was performed through two-way ANOVA with Sidak posttest for the signaling data, and two-tailed paired t tests for the cytokine data, unless otherwise described.

We first examined the potential response of C5aR2 activation on HMDM signaling using a sensitive AlphaLISA Surefire Ultra assay (25). One major outcome of C5aR1 activation is ERK1/2 phosphorylation and signaling, which can be driven by both a G protein–dependent pathway and through β arrestins, although the latter has not been confirmed for C5aR1 (4, 35). Acute C5aR2 activation was previously observed to be devoid of any detectable ERK1/2 signaling capabilities in HMDMs (8). We extended from this by conducting a time-course experiment to assess any potential delayed ERK1/2 activity caused by C5aR2 agonism, using P32 (Fig. 1), for up to 50 min (Fig. 2A). Selected doses of C5a (1 and 100 nM), representing the doses below and above the EC₅₀ of C5aR2 activation (53.7 nM), respectively, were included for comparison (8). C5a triggered a rapid and profound ERK1/2 phosphorylation in HMDMs, with the activity level being comparable for 1 and 100 nM of C5a at 5 min poststimulation, but markedly higher for 1 nM relative to 100 nM at 10 min. Selective C5aR2 agonism using P32 (100 μM), however, did not cause any significant ERK1/2 phosphorylation over 50 min. We also examined other key signaling assays downstream of C5a, including phosphorylation of p38 MAPK and Akt. Similar to ERK, however, despite strong C5a-mediated signaling, no detectable C5aR2-mediated activity was observed for Akt or p38 MAPK phosphorylation (Supplemental Fig. 1).

C5aR2 activity has been shown to dampen ERK signaling in primary human polymorphonuclear leukocytes and macrophages, induced by a fixed dose of C5a (8, 25, 26). To extend from this, we next examined how pre-exposure of HMDMs to P32 may affect the cell’s dose-dependent response to C5a. Elevating concentrations of C5a induced a rapid upregulation of ERK1/2 phosphorylation, peaking at 0.8 nM of C5a (EC₅₀ = 0.2 nM) (Fig. 2B). Further increases in C5a doses (up to 100 nM) instead caused a significant reduction in ERK1/2 phosphorylation to ∼27% of the peak level, as observed in the time-course experiment (Fig. 2A). This downregulatory effect on C5a-mediated signaling is consistent with previous findings and reported to be associated with C5aR2 activity (25, 34). Indeed, when testing C5a-mediated ERK1/2 phosphorylation in CHO-C5aR1 cells which do not express C5aR2, no similar downregulation was observed (Fig. 2C), supporting the involvement of C5aR2. Furthermore, in line with a prior study (8), we confirmed that C5aR2 activation in HMDMs reduced the magnitude of the peak response of C5a, down to 77% of the peak response (Fig. 2B). We also identified that C5aR2 activation significantly shifted the dose-response curve of C5a, reducing its EC₅₀ potency from 0.20 to 0.27 nM. The dose required for a maximal effect of C5a also increased 5-fold from 0.8 to 4 nM.

In addition to C5a-induced ERK signaling, which mainly reflects an integrated cell signaling outcome (36), we also examined C5a-mediated intracellular calcium mobilization, a possible G_α16-mediated signaling event in human macrophages (Fig. 2D–I) (15). C5a triggered profound intracellular calcium mobilization in HMDMs, characterized by an immediate elevation in intracellular calcium concentration within 5 s followed by a sustained “second phase” of calcium influx from 20 to 30 s post stimulation for C5a concentrations 1 nM and above (Fig. 2D–I). This is consistent with previous observations and can be attributed to store-operated calcium entry from extracellular spaces in human macrophages (15, 37, 38). In contrast to the dose-dependent downregulation observed for phospho-ERK1/2 activity, a saturated dose of C5a did not cause any reduction in calcium response. This indicates that the dampened response for ERK signaling is pathway specific rather than a consequence of global C5aR1 downregulation and desensitization. Furthermore, by contrast to the dampening effect observed for ERK signaling, activation of C5aR2 using P32 significantly enhanced intracellular calcium mobilization triggered by 0.1 nM (Fig. 2E, 2I) and 10 nM (Fig. 2H, 2I) C5a, whereas the baseline intracellular calcium level was unaffected (Fig. 2F, 2I). Overall, these results demonstrate that although C5aR2 is unable to directly signal through classical G protein–mediated pathways (i.e., MAPK, Akt phosphorylation, or calcium mobilization), its activation alters C5aR1-mediatd signaling events in macrophages. This leads us to next question whether similar modulatory activities hold true for other closely-related GPCRs.

Complement C3a receptors are the closest related receptor to C5a receptors, and have previously been linked to C5aR2 in mice (39). We therefore next examined how C5aR2 activation with P32 might impact C3a-mediated signaling in human macrophages. C3a potently induced ERK1/2 phosphorylation in HMDMs with an EC₅₀ of 0.22 nM (Fig. 3A). Interestingly, C3aR also demonstrated a significant downregulation of phopho-ERK1/2 activity when stimulated by higher concentrations of C3a, showing a comparable trend to that of C5aR1. Pre-exposure of HMDMs to the C5aR2 agonist significantly dampened the cell’s response to C3a, characterized by a marked 35-fold reduction in potency from 0.2 to 7.1 nM, and a delay in peak concentration from 4 to 100 nM of C3a. The profound suppression by C5aR2 agonism on C3aR-mediated ERK signaling is in sharp contrast with the relatively mild effect observed for C5a. Similar P32-mediated downregulation was also observed for C3a-induced calcium signaling in HMDMs, albeit to a lesser extent (Fig. 3B–F). A significant reduction in overall calcium influx was observed for 1 nM of C3a following P32 pretreatment (Fig. 3C, 3F), which diminished when higher C3a concentrations were applied.

Given the marked effect of C5aR2 activation on C3a-mediated signaling, we next asked whether C5aR2 activation could influence cellular signaling at a more global level rather than being specific to complement receptors. To investigate this, we chose several chemokine receptors, formyl peptide receptor (FPR) 1, CMKLR1 (or ChemR23), and LTB4 receptor 1 (LTB4R), based on their close structural relationship with C5aR1 (40). For ERK signaling (Fig. 4A), fMLF-activating FPR1 induced ERK1/2 phosphorylation in HMDMs with an EC₅₀ of 2.9 nM and displayed a marked dose-dependent downregulation at higher fMLF concentrations in a similar manner as C5aR1 and C3aR. Chemerin, through its major receptor CMKLR1, also induced ERK signaling with an estimated EC₅₀ of 16 nM (Fig. 4B). Pretreating cells with P32 significantly dampened CMKLR1-mediated phospho-ERK1/2 signaling, as shown by the significant reduction in the peak level reached (to 60% of the control level) within the ligand range concerned (Fig. 4B). The FPR1-mediated response, however, remained largely unaffected (Fig. 4A). LTB4R did not trigger significant ERK signaling in our preliminary trials (data not shown), and thus was not pursued for this signaling pathway.

For the calcium signaling pathway, all the three chemokine receptors (FPR1, LTB4R, and CMKLR1) induced intracellular calcium mobilization in HMDMs (Fig. 4C–E, Supplemental Fig. 2). Pretreating cells with P32 significantly elevated the LTB4R-mediated calcium response triggered by 10 and 100 nM LTB4 (Fig. 4E, Supplemental Fig. 2C). Subtle C5aR2 agonist–mediated modulations was also seen with the FPR1- and CMKLR1-mediated calcium traces (Fig. 4C, 4D). P32 markedly enhanced the peak calcium response induced by 10 nM fMLF by 14% (Supplemental Fig. 2A) and reduced that by 50 nM chemerin by 28% (Supplemental Fig. 2B). However, the overall calcium response mediated by these ligands, as measured by the area under the curve, was not significantly altered (Fig. 4C, 4D).

Next, to explore the role of C5aR2 on macrophage functions, we examined how C5aR2 agonism could modulate cytokine responses upon PRR activation. Several PRRs were chosen based on their expression and robust response following stimulation in human macrophages (41, 42) and activated using respective agonists (Table I). We then assayed the production of the acute response cytokines IL-6, TNF-α, and IL-10 (43, 44).

For the TLRs (TLR1/2, TLR3, TLR4, and TLR7), stimulation using their respective agonists trigged significant release of the cytokines IL-6, TNF-α, and IL-10 (Fig. 5). The magnitude of the responses varied, however, ranging from below 100 pg/ml for TLR3 (Fig. 5B, 5G, 5L) to more than 10,000 pg/ml for TLR4-induced IL-6 (Fig. 5C). Expectedly, large variations also existed among HMDMs derived from different human donors, in line with population heterogeneity (45). The spontaneous secretion of the three cytokines by HMDMs was negligible under nonstimulated conditions (Fig. 5, refer to dotted lines). Activation of C5aR2 robustly dampened IL-6 production triggered by TLR3, TLR4, and TLR7 in all five donors (Fig. 5B–D), with the changes for TLR4 and TLR7 reaching statistical significance, consistent with previous findings (8). For the proinflammatory cytokine TNF-α, C5aR2 agonism dampened TLR3- and TLR4-mediated responses from four out of five donors assessed (Fig. 5G, 5H) but increased TNF-α in one donor, and so the average level of change did not reach statistical significance. TLR1/2- and TLR7-mediated TNF-α release was largely unaltered by P32 treatment (Fig. 5F, 5I, 5J). A dampening response following C5aR2 activation was also observed for TLR3- and TLR7-mediated release of the anti-inflammatory IL-10 (Fig. 5L, 5N).

Next, we examined the potential effects of C5aR2 activation on non-Toll PRR–induced macrophage responses (Fig. 6). Activation of the C-type lectin receptors Dectin-1, Dectin-2, and Mincle, and the cytosolic DNA sensor STING, all triggered significant secretion of IL-6, TNF-α, and IL-10 from human macrophages; however, the magnitudes of the responses were generally lower in comparison with the TLR responses. By contrast to the variable responses observed following C5aR2 activation among the TLRs, the C5aR2-dependent modulatory effects on the non-Toll PRRs appeared to be more consistent. C5aR2 activation via P32 demonstrated significant and marked downregulation of Mincle- and STING-mediated responses for all cytokines (Fig. 6C–E, 6H–J, 6M–O), with the most significant reduction observed for Mincle-mediated IL-6 and TNF-α, reaching 13 and 21% of control levels, respectively. Similar inhibitory trends were also observed for Dectin-2–mediated TNF-α and IL-10, reaching 37 and 52% of the control levels, respectively (Fig. 6G, 6J, 6L, 6O). In contrast to Dectin-2, however, C5aR2 activation had minimal and variable effects on Dectin-1–mediated responses (Fig. 6A, 6E, 6F, 6J, 6K, 6O).

After 20 y since its initial discovery, the alternative C5a receptor C5aR2 still remains an enigmatic player in innate immunity, owing in part to the initial lack of selective and potent C5aR2 modulators which could be used as pharmacological tools for research. Significant advance was made with the identification of the first C5aR2-selective agonist termed P32 (Ac-RHYPYWR-OH), which displayed selective C5aR2-agonistic activities by recruiting β-arrestin 2. Although several studies have demonstrated a modulatory role of C5aR2 toward human immune cell functions involving C5a signaling (25–27, 31, 32, 46, 47), it is not yet known whether C5aR2 functionality can influence immune cells more broadly. The present study thus characterized the roles of C5aR2 on the signaling and functions of primary human macrophages by using P32 as a selective probe. Mature macrophages express a myriad of phagocytic receptors, complement receptors, and PRRs and play critical roles in fighting infections, maintaining tissue homeostasis, and regulating inflammatory responses (48), and thus were an ideal and relevant cell to study. We found that in addition to the well-documented modulatory activities of C5aR2 on C5a-induced signaling, C5aR2 activation also altered intracellular signaling elicited by several other complement and chemokine receptors. Functionally, C5aR2 activation displayed an inhibitory effect toward cytokine responses triggered by a multitude of PRRs, most notably on Mincle and cGas-STING pathways.

Macrophage C5aR2 activation via P32 demonstrated a generally suppressive effect toward ligand-induced intracellular ERK1/2 phosphorylation. The observed C5aR2-mediated dampening effect on C5aR1 is consistent with previously published data (8). Mechanistically, activated C5aR2 signals through β-arrestins (25, 26), and likely several other downstream, yet-to-be identified mediators, to promote C5aR1-C5aR2 heterodimer formation (27) and subsequently modulate C5a-induced ERK1/2 phosphorylation. It is important to note, however, despite P32’s activity on C5aR2, this ligand is unable to upregulate or alter the basal level of heterodimer formation between C5aR1 and C5aR2 (8), excluding this as the mechanism responsible for the ERK-modulatory action of P32. Alternatively, C5aR2 has been suggested to be involved in the endocytosis of C5aR1, possibly by regulating AP2 recruitment, such that the deficiency of C5aR2 prohibited ligand-induced internalization of C5aR1 and subsequent endosomal ERK1/2 signaling (49). It is therefore likely that P32, by activating C5aR2 and its associated signaling, is capable of altering the trafficking and ERK signaling of C5aR1 in a C5aR1-C5aR2 heterodimerization–independent manner.

It is worth noting that a similar downregulation of ERK signaling at higher ligand concentrations was also experienced by the related receptors C3aR and FPR1. Although this trend could be associated with the regulatory roles played by the alternative receptors FPR2 and FPR3 for the fMLF-triggered response (50), no alternate inhibitory receptor for C3a has been identified. A GRK- and β-arrestin–mediated receptor internalization and downregulation could, therefore, be at play (51). Contrary to our initial expectations, however, C5aR2-mediated downregulation of ERK signaling does not only apply to C5aR1, but also, to an even greater extent, toward the related receptors C3aR and CMKLR1. Chen et al. (39) has previously reported the critical involvement of C5aR2 in optimizing C3a-mediated signaling and function in murine polymorphonuclear leukocytes and macrophages; as such, the C5aR2 agonist–induced downregulation of ERK signaling observed in this study is not entirely surprising. Given that neither C3aR nor CMKLR1 has been found to dimerize with C5aR2, the ERK-dampening effect observed in this study adds to the evidence that C5aR2 exerts its modulatory actions independently of C5aR1. One possible explanation is that C5aR2 activation via P32 may alter the basal physiology of cells by modifying the available pool of secondary messengers and trafficking molecules, which in turn influences intracellular signal transduction. Indeed, this indirect mechanism of regulation has been demonstrated for the CXCL12 receptors CXCR4 and CXCR7, which both recruit and signal through β-arrestin 2 (52). The presence of CXCR7, possessing a higher basal association with β-arrestin 2, competes with CXCR4 for the available pool of β-arrestin 2, and thereby controls the integrated response to CXCL12 in cells. Other than CXCR7, several atypical chemokine receptors, such as the decoy D6 receptor and CCRL1, also demonstrate strong preference for coupling to β-arrestins over G proteins (53, 54). The potential analogy between these receptors and C5aR2 may serve as an avenue to further uncover the signaling mechanisms of C5aR2 (53).

In contrast to the clear downregulatory trend observed with ERK signaling, C5aR2-dependent modulation of ligand-induced calcium mobilization is more subtle. A slight C5aR2 agonist–induced upregulation of calcium influx was observed with selected concentrations of C5a. This could be attributed to altered C5aR1 internalization and downregulation in the presence of the agonist, which unavoidably competes with the endogenous ligand C5a for C5aR2 binding. However, this fails to account for the dampened calcium mobilization triggered by low concentrations of C3a and LTB4 following P32 pretreatment. This suggests that C5aR2 activation in macrophages may alter basal cell response thresholds in a broad and sophisticated manner, such as by modulating basal ion homeostasis, which could be overcome by higher ligand concentrations.

Functional cross-talk between C5a-C5aR1 and the pattern recognition system has been depicted in several prior studies, with its importance in immune regulation becoming increasingly emphasized (46, 47, 55–59). We therefore examined how selective activation of C5aR2 using P32 may modulate various PRR-meditated cell responses. C5aR2 activation was previously suggested to be critically involved in LPS-induced in vivo production of HMGB1 (60), upregulation of LPS-induced release of G-CSF (27, 47), and modulation of LPS-induced IL-6 and TNF-α release (8, 39). In the current study, C5aR2 activation exerted a largely suppressive effect toward the cytokine responses mediated by several PRRs. For instance, P32 treatment markedly reduced IL-6 generation in five out of eight PRRs examined, and similarly for IL-10. Significant reductions were also recorded for TNF-α responses triggered by three out of four non-Toll PRRs. The most prominent C5aR2-mediated suppression, however, was observed on STING- and Mincle-mediated cytokine output, in which a significant dampening effect was observed for all cytokines examined. Thus, we may conclude that C5aR2 activation has a generalized and largely suppressive effect on multiple pattern recognition–triggered cytokine responses by human macrophages.

In a pathophysiological context such as during an acute infection, in which high concentration of endogenous C5a is present, this C5aR2-mediated suppression of the signaling and functions of other chemokine receptors may assist macrophages in prioritizing C5aR1-driven functions. This modulatory role of C5aR2 may indeed be quintessential for an efficient and coordinated cell response to be achieved upon infection or inflammation. The observed downregulatory effects of C5aR2 activation may also help explain the multiple research findings indicating anti-inflammatory and protective roles of C5aR2 in pathological contexts, for instance in experimental sepsis (61, 62), experimental lung inflammatory injury (63), bone fracture, and bullous pemphigoid (21, 64, 65). Mechanistically, these anti-inflammatory actions were predominantly attributed to C5aR2 acting as either a decoy receptor to sequester excessive C5a and/or as a negative regulator for C5a-C5aR1–induced signaling (18, 25–27). The results from our study, however, suggest that rather than serving as a sole negative regulator of C5a-C5aR1 functionality, C5aR2 activation possesses independent and largely downregulatory effects on multiple pattern recognition–induced signaling responses.

Significant P32-mediated downregulation of cytokine response was observed for TLR3, TLR4, TLR7, Mincle, and STING. These TLRs share signaling components, such as TGF-β–activated kinase 1, at the convergence of MyD88-dependent and -independent pathways (66). This in turn upregulates the transcription factors AP-1 through p38 MAPK and JNK, and NF-κB through the IκB kinase (IKK) complex (66, 67). A certain degree of commonality is also shared by Mincle, which (through spleen tyrosine kinase) modulates multiple MAPKs and IKKs, and STING, which modulates IKKs through TANK-binding kinase 1 (68). It is therefore plausible that P32 and C5aR2 activity modulates PRR-driven cell responses through IKK regulation.

In addition, it has been previously proposed that complement-TLR cross-talk occurs at the level of MAPKs, including ERK1/2, p38 MAPK, and JNK, which subsequently modulate transcription factors such as NF-κB and AP-1 (55). For instance, C5a was found to negatively regulate the TLR4- and CD40-induced synthesis of IL-12, IL-23, and IL-27 by macrophages, likely by inhibiting the TLR4-induced upregulation of the transcriptional factors IFN regulatory factor 1 and IFN regulatory factor 8 in an ERK1/2 and PI3K-dependent manner (55, 57). Our study, however, could not detect any MAPK activation in human macrophages following C5aR2 activation. To the best of our knowledge, apart from the human mast cell line LAD2 in which potential C5aR2-dependent ERK and PI3K signaling was observed (69), no other C5aR2-mediated kinase signaling has been identified. C5aR2 does, however, recruit β-arrestins (25, 26), and considering the important role of β-arrestins as signal transduction scaffolds for multiple MAPKs (70), it remains an interesting area of future research to confirm whether C5aR2 directly modulates the expression of various transcription factors involved in cytokine production.

One limitation of our study was the reliance on using the agonist P32 to selectively activate C5aR2. First reported in 2016, P32, as one of the few available selective C5aR2 activators, has been employed as a useful pharmacological tool in several studies, and has demonstrated significant modulatory activities both in vitro and in vivo (8, 31, 32). There are, however, several considerations associated with using P32 as a proxy for C5aR2 activation. Primarily, being a partial agonist, P32 does not achieve full receptor activation compared with C5a (8), which may have limited us observing the full effects of C5aR2 activation in the current study. Furthermore, unlike C5a, P32 is unable to induce C5aR1-C5aR2 heterodimerization, an important step mediating the downregulation of C5aR1 activities (27, 71). P32 also displays weak binding to C5aR1, albeit devoid of any functional consequence (8, 30), and this weak interaction may have potentially contributed to the modulatory actions observed for experiments using C5a. As such, future studies exploiting C5aR2 knockdown or knockout human or mouse primary immune cells could help support the findings presented in the current study.

In conclusion, this study highlights novel pleiotropic functions for C5aR2 in human primary macrophages. C5aR2 activation significantly modulated the signaling activities of both C5aR1 and other complement and chemokine receptors such as C3aR and CMKLR1. Functionally, selective C5aR2 activation downregulated cytokine production triggered by various TLRs (TLR2, TLR3, TLR4, and TLR7), C-type lectin receptors (Dectin-1, Dectin-2, and Mincle), and the cytosolic DNA sensor STING. Our findings contribute to the better understanding of the immunopharmacology underlying this enigmatic complement receptor.

We thank Australian Red Cross Lifeblood and human donors for providing the cells used in these studies.

T.M.W. and R.J.C. are inventors on a patent (United States patent application, Publication No. US2017/0129921 A1, 2017 May 11) pertaining to the C5aR2 agonist used in this study. The other author has no financial conflicts of interest.