Monitoring C5aR2 Expression Using a Floxed tdTomato-C5aR2 Knock-In Mouse

Canonical and noncanonical cleavage of C5 results in the generation of C5a, which plays important roles in host defense, immune surveillance, and tissue homeostasis (1, 2). However, C5a also drives detrimental immune responses in autoimmune (3), inflammatory, and neurodegenerative (4) diseases. Strong and sustained production of C5a can lead to downregulation of immune responses in leukocytes, but it may have opposing effects in other cell types (5). These ambivalent biological functions of C5a have been explained by differential expression of C5a receptor (C5aR)1 and C5aR2. The expression and effector functions of C5aR1 are well characterized (6–8). It belongs to the large family of G protein–coupled receptors (GPCRs). Activation of C5aR1 by its cognate ligand C5a leads to the recruitment of G_αi subunits, resulting in the recruitment of β-arrestin, which is critical for receptor internalization, desensitization, recycling, and, in some cases, G protein–independent signaling (9–11).

The functional properties and expression patterns of C5aR2 are still enigmatic (5, 12). Like C5aR1, C5aR2 is a seven-transmembrane receptor that belongs to the GPCR family; however, as a result of alterations in the DRY and NPXXY motifs, as well as other properties, it does not couple to G proteins (13) but still binds to β-arrestin. Because C5aR2 lacks coupling to G proteins, it was initially considered a decoy receptor that inhibits the function of C5aR1 by capturing excessive C5a (13). In line with this view, Scola et al. (14) showed that cells that solely express C5aR2 bind C5a independently of C5aR1. This binding leads to ligand degradation, a mechanism by which C5aR2 might block the proinflammatory properties of C5a mediated by activation of C5aR1. Studies by Bamberg et al. (15) supported the view that C5aR2 interacts with β-arrestin, thereby regulating other signaling pathways. Binding of C5a to both receptors triggers their phosphorylation, thereby facilitating their association with β-arrestin. C5aR1/β-arrestin association leads to ERK1/2 phosphorylation, whereas the formation of C5aR2/β-arrestin complexes results in inhibition of ERK1/2 phosphorylation, suggesting that the ratio of C5aR1/C5aR2 expression determines C5a-mediated cellular responses. Also, C5aR2 acts as a negative regulator of C5aR1 in a mouse model of acute lung injury (16). However, the narrowed view of C5aR2 as a sole decoy receptor has been challenged by several findings demonstrating that binding of C5a to C5aR2 induces pro- and anti-inflammatory functions (17–19). In neutrophils, C5aR2 deficiency leads to an altered cytokine profile when these cells are stimulated with C5a. In this setting, C5aR2 controlled the C5a-induced upregulation of IL-6, TNF-α, and CR3 (14, 17). Further, C5aR2-deficient mice showed improved survival in the cecal ligation puncture model of sepsis, which is in line with the observation that C5aR2 deficiency results in reduced inflammatory cell infiltration and HMGB-1 production from macrophages and neutrophils (20). Recently, Croker et al. (5) described the specific C5aR2 agonist P32. Stimulation of C5aR2 with this molecule modulated IL-6 release from human macrophages. Further, they observed that stimulation of C5aR2 inhibited the C5a-mediated effector functions of macrophages, as well as neutrophil mobilization without blocking it completely (5).

To better understand the various functions of C5aR2 as a decoy receptor, a regulator of C5aR1 functions, and as a second C5aR that activates cells independently of C5aR1, a detailed understanding of C5aR2 expression and regulation is necessary. In this study, we developed a floxed tandem dye Tomato (tdTomato)-C5aR2 knock-in mouse, which allows reliable tracking of C5aR2 expression in immune and tissue cells. The two loxP sites were introduced to create cell- or tissue-specific C5aR2 knockout mice by mating the floxed tdTomato-C5aR2 knock-in mouse with corresponding Cre deleter strains. Using this reporter mouse, we provide a detailed map of C5aR2 expression in innate and adaptive immune cells residing in the bone marrow (BM), blood, spleen, lung, lamina propria (LP) of the small intestine, visceral adipose tissue (VAT), and brain. Further, we assessed the expression of C5aR2 in the lung and the airways under inflammatory conditions upon IL-33 challenge. Our findings demonstrate differential expression of C5aR2 in myeloid cells and lack of C5aR2 expression in T cells. Surprisingly, we observed C5aR2 expression in B cells and subsets of NK cells. Importantly, C5aR2 activation on NK cells controlled IL-12/IL-18–mediated IFN-γ production. Further, we found IL-33–mediated regulation of C5aR2 expression in pulmonary eosinophils and monocyte-derived dendritic cells (moDCs).

Monoclonal allophycocyanin-Cy7–labeled Ab against CD11b (M1/70), PerCP-Cyanine (Cy)5.5–labeled Ab against CD8a (53-6.7), allophycocyanin or V450-labeled Ab against Ly6G (1A8), V450-labeled Ab against CD45R (RA3-6B2), FITC-labeled Ab against CD90.2 (30-H12), allophycocyanin-H7–labeled Ab against CD19 (1D3), and Brilliant Violet (BV)421-labeled Ab against Siglec-F (E50-2440) and CD43 (1G10) were purchased from BD Biosciences. PerCP-Cy5.5–labeled Ab against Ly6C (RB6-8C5); allophycocyanin-labeled Ab against CD115 (AFS98), CD11c (N418), NK1.1 (PK136), CD4 (GK1.5), and CD11b (M1/70); PE-Cy7–labeled Ab against CD25 (PC61.5) and CD4 (RM4-5); eFluor (eF)450-labeled Ab against CD19 (1D3), CD3e (145-2C11), and CD49b (DX5); Alexa Fluor (AF)488-labeled Ab against CD3 (17A2); allophycocyanin-eF780–labeled Ab against MHC class II (MHCII; M5/144.15.2); BV510-labeled Ab against CD11b (M1/70); AF711-labeled Ab against CD64 (X54-5/7.1); BV421-labeled Ab against F4/80 (BM8); and PerCP-Cy5.5–labeled Ab against CD103 (2E7) were purchased from BioLegend. PerCP-Cy5.5–labeled Ab against CD3 (OKT3), CD5 (53-73), CD27 (LG.7F9), NK1.1 (PK136), TCR-β (H57-597), and CD11b (MI/70); eF780-labeled Ab against CD11c (N418), B220 (RA3-6B2), and CD49b (DX5); eF450-labeled Ab against CD25 (PC61.5) and CD317 (ebio129c); PE-Cy5–labeled Ab against CD127 (A7R34); and PE-Cy7–labeled Ab against IgM (11/41) were purchased from eBioscience (Affymetrix). For surface and intracellular staining, unlabeled rat C5aR2-specific Ab (clone 468705; IgG2_B; R&D Systems) was used, the binding of which was detected using a secondary AF647-labeled F(ab)₂ anti-rat Ab fragment (Abcam). For intracellular staining of phosphorylated ERK1/2, the AF647-labeled Ab against ERK1/2 (pT202/pY204; BD) was used. Further, we used rabbit unlabeled polyclonal Ab against C5aR2 (catalog number HP8015; Hycult Biotech), the binding of which was detected using a secondary AF647-labeled F(ab)₂ anti-rabbit Ab fragment (Abcam).

RBC lysis (RBCL) buffer was prepared using 155 mM NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA (all from Sigma-Aldrich). BSA and a Lamina Propria Dissociation Kit were purchased from Miltenyi Biotec. Recombinant murine GM-CSF and Flt3 were from PeproTech. Liberase TL was from Roche Diagnostics, and DNase I for cell isolation was from Sigma-Aldrich. The KAPA Express Extract Kit was from Peqlab. FBS, RPMI 1640, and HBSS were from PAA. PBS, l-glutamine, penicillin, and streptomycin were from Life Technologies. DNase I for total RNA isolation was from Fermentas. The RevertAid First Strand cDNA Synthesis Kit was from Thermo Scientific. Primers for real-time RT-PCR were from Eurofins Genomics, and all other reagents for RT-PCR were from Bio-Rad. GelRed was purchased from BIOTREND. The Cytofix/Cytoperm kit was purchased from BD Biosciences; C5a was from Hycult Biotech; IL-33 was from eBioscience; DAPI was from Life Technologies; and mounting Medium Fluoromount-G was purchased from SouthernBiotech.

C57BL/6J wild-type (WT) mice were purchased from Janvier Labs. The tdTomato-C5ar2^fl/fl and tdTomato-C5ar2^fl/+ knock-in mice were generated as described below. All animals were used at 8–12 wk of age and were handled in accordance with the appropriate institutional and national guidelines. Animals were used for organ removal according to protocols approved by local authorities of the Animal Care and Use Committee (Ministerium für Energiewende, Landwirtschaft, Umwelt, Natur, und Digitalisierung, Kiel, Germany). Intratracheal (i.t.) injections of IL-33 were also approved by this committee (V242-81505/2016 [19-2/2017], V242-30278/2016 [50-4/16]). All experiments were performed by certified personnel.

The floxed tdTomato-C5aR2 knock-in mice were generated by gene targeting. A diagram of the targeting strategy is shown in Fig. 1. The targeting vector, the embryonic stem (ES) cells, and the mice were generated by Ozgene (Bentley, Australia). Our strategy was to insert tdTomato immediately following the splice acceptor of exon 2 in-frame with the coding sequence of C5ar2. The self-cleaving peptide derived from porcine teschovirus-1 (P2A) and the cDNA of C5ar2 were also present downstream of tdTomato to permit expression of the endogenous peptide from the knock-in locus. This region was floxed with loxP sites to permit a knockout of the locus by Cre recombinase.

The targeting construct was generated by the sequential cloning of four fragments. The final recipient plasmid encoded a neomycin selection cassette. Primers used to amplify the fragments also encoded the restriction enzyme sites required for the assembly of the fragments generated. All fragments were housed in the Surf2 vector backbone (Ozgene). The first fragment was a 4.4-kb fragment encoding the 3′ homology arm. This fragment was amplified from BACsRP23-388G21 and RP23-93F22 using primers 1617_461 (5′-CTAAGGGGTGAATGTTCAGTGAGAAA-3′) and 1617_561 (5′-TAAGCATTGGTAAGACGTCGCAAGAGGGCTTTGGATTTCCTAGGC-3′). The second fragment was a 1.2-kb fragment encoding the P2A and cDNA of C5ar2. This fragment was amplified from BACsRP23-388G21 and RP23-93F22 using primers 1617_44 (5′-TAAGCATTGGTAAGACGTCGACACGTGTCATAACTTCGTATAGCATACATTATACGAAGTTATCTACACCGGCATCTCAGACACCATTTCGTG-3′) and 1617_54 (5′-CTAAACGCGTCTGGAGCTACTAACTTCAGCCTGCTGAAGCAGGCTGGAGACGTGGAGGAGAACCCTGGACCTATGAAATATGAGGGGCTGGGAGACGGGTGTCTGGGGGCCAGAATGATGAACCACACCACCAG-3′); primer 1617_44 encoded a loxP site, and primer 1617_54 encoded the P2A sequence and the coding sequence of exon 1.

The third fragment was a 1.6-kb fragment encoding the coding sequence of tdTomato. This fragment was amplified in one of Ozgene’s proprietary vectors with primers 1617_45 (5′-TAAGCATTGGTAAGGCGCGCCCTTGTACAGCTCGTCCATGCCGTA-3′) and 1617_55 (5′-GCCGCAATGCTGGTACCCTGGAGTGCAGGCACCCTGCGCCACAGTGTTGGTGTCTCTAGCATGGAGCCGCTGGACCCGCCTTGCCCGGGGACAGCAAAGCTGGAGTTTTCCGTTGACTTTCCCTCCTGTCCCCTGCAGGTGTCTGATGGTGAGCAAGGGCGAGGAGGTCATCAAAGA-3′); primer 1617_55 also incorporated the splice acceptor of C5ar2 exon 2. The fourth fragment encoded the 3.6-kb 5′ homology arm. This fragment was amplified from BACsRP23-388G21 and RP23-93F22 using primers 1617_43 (5′-CTAAAGATCTCTGTGCCGGTTATTGCCAACTAAG-3′) and 1617_53 (5′-TAAGCATTGGTAAGGGAGTCAGAATTAGCATTCCAGA-3′).

The four fragments were assembled sequentially as follows. Fragment 1 was excised by digestion with AatII and ligated into fragment 2, which had been digested with the same enzyme. Combined fragments 1 and 2 were excised from the resulting construct by digestion with MluI and ligated into fragment 3, which had been digested with the same enzyme. Combined fragments 1, 2, and 3 were excised from the resulting construct by digestion with NotI and ligated into fragment 4, which had been digested with the same enzyme, resulting in the targeting vector 1617_pTV. This vector was linearized by digestion with SacII prior to electroporation into C57BL/6 Bruce4 ES cells.

For genotyping, we used biopsies from the ear. DNA extraction was performed using the KAPA Express Extract Kit (Peqlab), following the manufacturer’s instructions. To amplify the different DNA fragments, we used following primers: GK 357: 5′-CGTGGAGTTCCTAGAGGCTCA-3′; GK 353: 5′-AATACTCGCTGGTGGTGTGG-3′ (Eurofins Scientific). PCRs were run using the following conditions: 95°C for 3 min, followed by 35 cycles at 95°C for 15 s, 65°C for 15 s, and 72°C for 60 s, followed by 72°C for 120 s. The samples were transferred to a 1.0% sodium borate agarose gel, which was stained for amplification products with GelRed (BIOTREND).

Mice were killed by cervical dislocation under anesthesia, and the heart was perfused with 10 ml of cold PBS before organ removal. For BM cell isolation, femurs and tibias were removed, placed in PBS on ice, and subsequently flushed with PBS. RBCs were removed by incubating the cells in RBCL buffer for 3 min. The reaction was stopped by adding a large volume of PBS. BM-derived DCs (BMDCs) were generated by 9 d-differentiation with GM-CSF (20 ng/ml) as previously described (9, 21). BMDCs were defined as CD11c⁺CD11b⁺MCHII^hiCD115⁻ cells. To obtain BM-derived plasmacytoid DCs (BMpDCs), BM cells were differentiated for 9 d in the presence of Flt3 ligand (200 ng/ml). BM-derived macrophages (BMMs) were obtained after 6 d of differentiation in the presence of M-CSF, which is secreted by L929 cells and is used in the form of L929-conditioned medium (22), or as CD11c⁺CD11b⁺CD115⁺MHC^lo BMMs in a 9 d-differentiation with GM-CSF (20 ng/ml) (21). For pulmonary cell preparation, lungs were harvested and digested using 0.25 mg/ml Liberase TL and 0.5 mg/ml DNase I in pure RPMI medium for 45 min at 37°C, and single-cell suspensions were prepared as described (7). For VAT cell preparations, perigonadal fat tissue was harvested, taking care not to remove the gonads, and minced into small pieces. Tissue was digested using 0.25 mg/ml Liberase TL and 0.5 mg/ml DNase I in pure RPMI 1640 medium for 45 min at 37°C. The cell suspension was filtered through a cell strainer (40 μm Nylon; BD) and washed with 10 ml of RPMI 1640 medium complemented with 10% FCS containing 0.5 mg/ml DNase I. Cells were centrifuged at 350 × g for 10 min at 4°C, resuspended in PBS, transferred into a fresh 15-ml tube, and washed again. Bronchoalveolar lavage fluid (BAL) samples were obtained as described (9). Lavage of the peritoneal cavity was performed with 10 ml of ice-cold PBS. Collected cells were washed once with PBS. RBCs were removed by incubating the cells in RBCL buffer for 3 min and then washed with PBS (7). Blood was collected by cardiac puncture and immediately transferred into tubes containing 10 mM EDTA to prevent coagulation. Diluted blood samples were incubated several times in RBCL buffer to remove RBCs and then washed with PBS (7). Isolation of cells from the spleen and mesenteric lymph nodes (mLNs) was performed by mechanical disruption using a cell strainer (40 μm Nylon; BD) and the plunger of a 5-ml syringe (BD). The cell strainer was flushed three times with 5 ml of PBS. Cells were then incubated in RBCL buffer for 3 min and finally washed with PBS. LP cell suspension was obtained using a Lamina Propria Dissociation Kit (Miltenyi Biotec), following the manufacturer’s recommendations. Briefly, the small intestine was removed from the mice and quickly cleared of feces, residual fat, and Peyer’s patches. Then, they were cut longitudinally and incubated twice in HBSS, once with the addition of EDTA (5 mM) and once without. The tissue was digested using the manufacturer’s enzyme mixture in HBSS with Ca²⁺ and Mg²⁺ for 30 min at 37°C and homogenized using a gentleMACS Dissociator. After the last washing step, cell number was determined using a Neubauer chamber, and cells were resuspended in PBS containing 1% BSA. Viability of the collected cells was determined by trypan blue exclusion or a LIVE/DEAD Cell Viability Assay (BD) using flow cytometry (LSR II; BD). Brain tissue was collected, sliced into pieces, and digested with Liberase TL (0.1 mg/ml) and DNase I (0.1 mg/ml) for 30 min at 37°C. Brain homogenate was passed through a cell strainer (40 μm Nylon; BD), and the cell suspension was washed and fractionated on 30–75% discontinuous Percoll gradient (GE Healthcare) for 30 min at 1350 × g without brake. The interphase-containing cells were collected and washed with Dulbecco’s PBS, and the cellular composition was determined by flow cytometry.

Phenotypic characterization of cells was performed using a BD Aria III cell sorter (BD) and a MoFlo Legacy cell sorter (Beckman Coulter). The discriminative markers and gating strategies used to identify different cell types in the different organs and compartments have been described previously (23) and are described briefly in the figure legends and text. Flow cytometric data were analyzed using FlowJo 10 (TreeStar, Ashland, OR).

Cells were collected from the BM, as described above, and stained with DAPI or allophycocyanin-labeled Ab recognizing Ly6G (clone 1A8). The cells were incubated at 4°C for 15 min, transferred to a slide, and mounted with Fluoromount-G. Images were obtained using an Olympus FV 1000 confocal microscope with a 40× oil immersion 0.95 numerical aperture objective. FluoView 2.1c software (Olympus) was used as acquisition software. Image processing was conducted using Imaris software (Bitplane).

BM cells from tdTomato-C5ar2^fl/fl reporter and C5ar2^−/− mice were left unstimulated or were stimulated for 8 min with C5a (20 nM), with or without treatment with the C5aR2 agonist P32 (100 μM) 30 min prior to C5a addition. The C5aR2 agonist P32 was described recently (5). BM cells were then fixed with 1.8% formaldehyde for 15 min before 500 μl of ice-cold methanol was used to permeabilize the cell membranes. Next, methanol incubation was stopped after 20 min by washing with 500 μl of cold PBS. After resuspending the cell pellet in 100 μl of PBS, cells were stained for 15 min with an AF647-labeled Ab against ERK1/2 (pT202/pY204). Unbound Ab was removed by PBS washing. After pellet resuspension in 300 μl of PBS, samples were analyzed for ERK1/2 phosphorylation by flow cytometry.

The spleen was harvested and dissociated by mechanical disaggregation. After RBC lysis (as described above), cells were counted, and MACS-purified T cells were used for in vitro activation as described (7). To stimulate CD4⁺ T cells, a 96-well cell culture plate (Greiner Bio-One) was coated with unconjugated anti-CD3 (17A2) and anti-CD28 (37.51; both from eBioscience) in 150 μl of PBS at a final concentration of 2 and 10 μg/ml, respectively. The plate was kept at 37°C, 5% CO₂ for ≥2 h before discarding the coating solution and adding the cell suspension. T cells were cultured for 5 d in complete RPMI 1640 culture medium supplemented with 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cells were subsequently analyzed by flow cytometry for tdTomato-C5aR2 expression.

IL-33 (500 ng) was administered i.t. to anesthetized mice on four consecutive days (24). Twenty-four hours after the final IL-33 challenge, BAL and lung samples were collected as described above and used to determine immune cell composition by flow cytometry.

NK cells were purified from the spleen using a cell sorter. NK cells were identified as NKp46⁺CD3⁻ cells. After sorting, 1 × 10⁵ NK cells were cultured in RPMI 1640 (with 10% FCS, 1% penicillin/streptomycin, and 1% l-glutamine) and stimulated with 10 ng/ml IL-12 and 10 ng/ml IL-18 for 24 h. In some experiments, NK cells were also stimulated with 100 μM of the C5aR2-specific agonist P32. As a readout, IFN-γ production was measured by ELISA.

RNA was isolated using TRIzol Reagent, according to the manufacturer’s instructions. After DNase I treatment of the RNA, a reverse-transcription reaction was performed using a RevertAid First Strand cDNA Synthesis Kit. Real-time PCR was done using iQ SYBR Green Supermix on a CFX96 Real-Time System (Bio-Rad) with the following primers: β-actin 5′-GCACCACACCTTCTACAATGAG-3′ and 5′-AAATAGCACAGCCTGGATAGCAAC-3′; C5ar2 5′-CTGGGCCTCTTGCTGACTGTGC-3′ and 5′-GCCCCAGGAAGCCAAAGAGGA-3′. The temperature profile of the RT-PCR was 95°C for 3 min, followed by 40 cycles at 95°C for 5 s, 58°C (β-actin) or 54°C (C5ar2) for 5 s, and 72°C for 30 s, followed by 30 cycles for 5 s with a 1°C temperature increase starting at 65°C to confirm the expected PCR products by melting curve analysis. Real-time RT-PCR data were analyzed using CFX Manager Software 3.1 (Bio-Rad). Amplification products were also separated using a 2% sodium borate agarose gel and were detected by GelRed staining (BIOTREND).

Statistical analysis was performed using GraphPad Prism version 6 (GraphPad). Normal distribution of data was tested using the Kolmogorov–Smirnov and D’Agostino–Pearson tests. Statistical differences between groups were assessed using the nonpaired Student t test (Figs. 7D, 7E, 8) or the paired Student test (Fig. 7F). A p value <0.05 was considered significant (*p < 0.05, **p < 0.01, ***p < 0.001).

We generated a floxed tdTomato-C5aR2 knock-in mouse that allows cell- and tissue-specific tracking and deletion of C5aR2. For this purpose, we cloned a tdTomato cassette directly upstream of the 5′ end of exon 2 of the C5ar2 gene. To allow conditional deletion of the C5ar2 gene, we placed two loxP sites upstream of the tdTomato coding sequence and downstream of exon 2 of the C5ar2 gene (Fig. 1A). We obtained several ES cells that incorporated tdTomato and the two loxP sites into the C5ar2 gene locus. One of such cells was successfully used to produce heterozygous (tdTomato-C5ar2^fl/+) and homozygous (tdTomato-C5ar2^fl/fl) floxed tdTomato-C5aR2 knock-in mice (Fig. 1B).

The initial analysis of the tdTomato signal in different organs using flow cytometry showed that tdTomato⁺ cells were present in all organs analyzed, with the highest frequency in brain, BM, and BAL. The frequency of tdTomato⁺ cells was lower in spleen, lung, blood, VAT, and the LP of the small intestine (Fig. 1C). Of note, in lung and the LP of the small intestine, the tdTomato signal was not associated with epithelial cells (data not shown).

In their first description of C5aR2, Ohno et al. (25) stated that C5ar2 mRNA is strongly expressed in granulocytes under steady-state conditions. Since then, several groups determined the potential functions of C5aR2 as a regulator of C5aR1-mediated effects (15) or independent of C5aR1 functions (19, 20). Thus, we determined tdTomato expression in neutrophils from BM, blood, and the peritoneal cavity. In line with previous reports, we found strong expression of tdTomato-C5aR2 in all tested neutrophil populations. When we compared tdTomato-C5aR2 expression in neutrophils from tdTomato-C5ar2^fl/+ and tdTomato-C5ar2^fl/fl mice, we found that the mean fluorescence intensity (MFI) was higher in tdTomato-C5ar2^fl/fl mice (Fig. 2A). Next, we tested whether the tdTomato signal matches C5ar2 mRNA expression. For this purpose, we analyzed sorted neutrophils from blood and BM from WT mice and assessed C5ar2 expression by RT-PCR. In both neutrophil subsets, we found a strong signal for C5ar2 mRNA (Fig. 2B). Further, we assessed C5aR2 protein expression using Ab-based surface staining. Using two commercially available Abs (clone 468705 [R&D Systems] and a polyclonal Ab [Hycult Biotech]), we found a strong and homogeneous signal in BM neutrophils from WT (data not shown) and tdTomato-C5ar2^fl/fl mice (Fig. 2C), suggesting C5aR2 expression in such cells. Surprisingly, we also observed a strong signal when we stained BM neutrophils from C5ar2^−/− mice, demonstrating that these Abs recognize structures independently of C5aR2 (Fig. 2C). Finally, we stained BM neutrophils from WT and tdTomato-C5ar2^fl/fl mice with Ly6G and DAPI and transferred them onto glass slides to analyze the tdTomato signal by fluorescence microscopy. We observed a strong tdTomato signal in Ly6G⁺ neutrophils from tdTomato-C5ar2^fl/fl mice but not from WT mice, and the signal was primarily located in the cytoplasm of these cells (Fig. 2D).

To delineate whether the tdTomato construct has any impact on the function of C5aR2, we determined C5a-induced ERK1/2 phosphorylation in BM neutrophils in the presence of the C5aR2-specific agonist P32. Bamberg et al. (15) showed that C5aR2 acts as a negative regulator of ERK1/2 signal transduction. These findings could be corroborated using the C5aR2-selective ligand P32 (5). Stimulation of C5aR2 with P32 in tdTomato-C5ar2^fl/fl mice abrogated ERK1/2 phosphorylation after C5a stimulation. This effect was C5aR2 specific, because C5a-mediated ERK1/2 phosphorylation in C5aR2-deficient neutrophils was not reduced by the presence of the C5aR2 agonist (Fig. 2E). These findings demonstrate that C5aR2 in tdTomato-C5ar2^fl/fl mice is fully functional.

Previous studies reported important functional roles for C5aR2 in macrophages. For example, C5a-induced G-CSF production is markedly reduced in LPS-activated macrophages from C5ar2^−/− mice compared with WT mice (26). Also, C5a-induced HMGB-1 production in macrophages depends on the presence of C5aR2 (20). We determined the expression of C5aR2 in macrophages from M-CSF– and GM-CSF–stimulated BM cell cultures, peritoneal cavity, LP of the small intestine, airways, lung tissue, VAT, and brain. All tested macrophage populations expressed tdTomato-C5aR2, although the expression levels were different (Fig. 3A, 3B, 3D–H). Peritoneal macrophages strongly and homogeneously expressed tdTomato-C5aR2 (Fig. 3D). In contrast, tdTomato-C5aR2 expression was much lower in M-CSF– or GM-CSF–derived BMMs: ∼40% of M-CSF BMMs (Fig. 3A) and 20% of GM-CSF BMMs (Fig. 3B) stained positive for tdTomato-C5aR2. Macrophages from the LP of the small intestine (Fig. 3E), lung and airways (Fig. 3F), and VAT (Fig. 3G), as well as macrophages or microglia cells from the brain (Fig. 3H), showed a heterogeneous expression pattern of tdTomato-C5aR2, with a tdTomato⁺ and a tdTomato⁻ population. Around 80–85% of LP (Fig. 3E) and airway (Fig. 3F) macrophages, as well as microglia cells (Fig. 3H), were tdTomato-C5aR2⁺. Approximately 60–65% of the brain macrophages (Fig. 3H) and 55–60% of pulmonary macrophages (Fig. 3F) were tdTomato-C5aR2⁺, whereas only 20–25% of the VAT macrophages stained positive for tdTomato-C5aR2 (Fig. 3G). These findings demonstrate tissue-specific differences in macrophage tdTomato-C5aR2 expression.

We also determined surface and intracellular C5aR2 expression in GM-CSF BMMs and surface expression in peritoneal macrophages using the two anti-C5aR2 Abs (as described above). In GM-CSF BMMs from tdTomato-C5ar2^fl/fl mice, we found strong C5aR2 surface and intracellular staining, which was like the signal obtained using cells from C5ar2^−/− mice (Fig. 3C, data not shown). In line with this observation, peritoneal macrophages from tdTomato-C5ar2^fl/fl and C5ar2^−/− mice stained positive using the Abs against C5aR2. These data confirm our findings with neutrophils (Fig. 2C) and provide additional evidence that the two commercially available Abs used in this study recognize C5aR2, as well as other structures on neutrophils and macrophages and, thus, are not specific for C5aR2.

Little is known about the biological role or function of C5aR2 in eosinophils, which play important roles in allergic diseases such as allergic asthma. Previously, we found that C5aR2 drives eosinophilic airway inflammation in different models of experimental allergic asthma (27). In this study, we assessed the expression of tdTomato-C5aR2 in eosinophils from blood, lung, LP of the small intestine, BM, and VAT. BM and VAT are considered reservoirs of eosinophils (28, 29). In all tested eosinophil populations, we found a clear tdTomato signal (Fig. 4). In eosinophils from BM (Fig. 4A), blood (Fig. 4B), lung (Fig. 4C), and LP of the small intestine (Fig. 4D), we observed two distinct populations. In BM, lung, and in the LP of the small intestine (65–80%) as well as in the blood (90–95%), most eosinophils expressed tdTomato-C5aR2; however, a fraction of the eosinophils were clearly tdTomato-C5aR2⁻ (Fig. 4A–D). In VAT (Fig. 4D), dtTomato-C5aR2 expression was homogeneous but weaker than in all other compartments.

Several DC subsets in lymphoid and nonlymphoid compartments can be defined based on genetic and phenotypic markers (30). These include conventional DCs (cDCs), inflammatory moDCs, and pDCs. In mice, cDCs can be further classified into CD11b⁺-like and CD8⁺-like subsets. Following this differentiation scheme, CD11b⁺ cDCs are composed of splenic CD11b⁺CD8⁻ DCs and pulmonary CD11b⁺ cDCs. The CD8-like cDC subset comprises splenic CD8⁺CD11b⁻ and pulmonary CD103⁺CD11b⁻ cDCs (31). At steady-state, only a very small fraction of CD11b⁺ DCs in the lung are moDCs, which increase under inflammatory conditions (32). Similarly, CD11b⁺ and CD103⁺ cDC subpopulations have been described in the LP of the small intestine (33). In contrast, DCs in the VAT are poorly defined. We identified a F4/80⁻Siglec-F⁻MHCII⁺CD11c⁺CD11b⁺ population in the VAT that we consider CD11b⁺ cDCs. However, we did not find any CD8-like cDC equivalent (data not shown). Finally, we observed a mixed BMDC/BMM population, as recently described (21), comprising 20–30% CD11c⁺CD11b⁺CD115⁻MHCII⁺ BMDCs and 70–80% CD11c⁺CD11b⁺CD115⁺ BMMs when BM cells were differentiated in GM-CSF–supplemented medium.

The expression of tdTomato-C5aR2 within the DC compartment was heterogeneous. In the lung, a minor subpopulation of CD11b⁺ cDCs (15–20%) and a large subset of moDCs (60–65%) stained positive for tdTomato-C5aR2, whereas CD103⁺ cDCs did not express tdTomato-C5aR2 (Fig. 5A). Similarly, at best, a very minor fraction of GM-CSF–derived BMDCs expressed tdTomato-C5aR2 (Fig. 5B). Also, both splenic DC subsets stained negative for the tdTomato signal (Fig. 5C). In the LP of the small intestine, CD11b⁺ and CD103⁺ cDCs expressed tdTomato-C5aR2, although the expression levels were low (Fig. 5D). The VAT DC subset did not express tdTomato (Fig. 5E). Similarly, pDCs from spleen, the LP of the small intestine, lung, and VAT, as well as BMpDCs, stained negative for tdTomato (Fig. 5F).

Currently available data suggest that C5aR2 is expressed in human and murine T lymphocytes. C5aR2 expression was shown in human naive and central memory T cells (34). In mice, it was recently shown that C5aR2 might be involved in T cell functions (14, 34); however, only one study showed C5ar2 expression in murine T cells, at least at the mRNA level (6). To gain insight into the expression of C5aR2 in CD4⁺ and CD8⁺ T cells, we determined tdTomato-C5aR2 expression in naive CD4⁺ and CD8⁺ T cells from spleen, blood, and mLNs, as well as CD3⁺ T cells from the lung. Further, we assessed tdTomato-C5aR2 expression in splenic CD4⁺ T cells in response to stimulation with CD3/CD28. We found no tdTomato-C5aR2 expression in naive CD4⁺ or CD8⁺ T cells from different tissues (Fig. 6A–D). Also, CD3/CD28 stimulation did not result in the induction of tdTomato-C5aR2 (Fig. 6E). This observation was confirmed at the mRNA level using naive CD4⁺ T cells purified from spleen, in which we found no amplification of C5ar2 transcripts (Fig. 6F).

For human B cells, C5ar2 mRNA expression has been shown on plasma cells and memory B cells (15). Detailed information about C5aR2 expression in mouse B cells is lacking. When we assessed tdTomato-C5aR2 expression in CD19⁺ B cells from spleen, blood, mLN, and lung, we found a clear and homogeneous tdTomato signal in B cells from all tissues (Fig. 6A–C) with the exception of the lung, in which the signal was weak (Fig. 6D). C5ar2 expression was confirmed at the mRNA level in splenic B cells, in which we consistently amplified a faint band (Fig. 6F).

Previously, human NK cells were found to express C5ar2 mRNA (15) and C5aR2 protein intracellularly (15, 35). Furthermore, mouse NK cells have been reported to express low levels of C5ar2 mRNA, which was upregulated upon Escherichia coli infection (36). In this study, we found a clear tdTomato-C5aR2 signal in NK cells under steady-state conditions in spleen and blood (Fig. 7A, 7B), whereas NK cells from the spleen stained tdTomato-C5aR2 (Fig. 6D). Interestingly, only 20–25% of splenic NK cells and 30–35% of blood NK cells stained positive for tdTomato-C5aR2. C5ar2 expression was confirmed at the mRNA level using purified splenic CD49⁺CD3⁻ NK cells (Fig. 7C). To assess the functional role of C5aR2 expression in NK cells, we first determined the maturation profile of NK cells from BM, spleen, and lung. As shown in Fig. 7D, we found the expected low frequency of mature CD11b⁺CD27⁻ NK cells in the BM and an increased frequency of such cells in spleen and lung. C5aR2 deficiency had no impact on the distribution of mature NK cells (Fig. 7D). We also analyzed other maturation stages of NK cells (CD11b⁻CD27⁻, CD11b⁻CD27⁺, CD11b⁺CD27⁺) and found no differences between WT and C5ar2^−/− mice (data not shown). One important function of NK cells is to produce IFN-γ in response to IL-12 and IL-18 stimulation (37, 38). When we treated purified NK cells from WT and C5ar2^−/− mice with IL-12 and IL-18, we found a significantly enhanced production of IFN-γ in C5aR2-deficient NK cells (Fig. 7E), suggesting that C5aR2-activation suppresses IL-12/IL-18–mediated IFN-γ production. Thus, we stimulated purified splenic NK cells from WT mice with IL-12/IL-18 in the presence or absence of the specific C5aR2 agonist P32. Indeed, such treatment significantly suppressed IL-12/IL-18–driven production of IFN-γ (Fig. 7F).

IL-33 is abundantly expressed in lung epithelial cells and regulates the development of Th2 immunity in experimental allergic asthma (39). In this study, we administered IL-33 i.t. and determined the impact of this treatment on the frequency of tdTomato-C5aR2⁺ immune cells in the lung and airways. Further, we assessed the change in tdTomato-C5aR2 expression in such cells. Administration of IL-33 into WT and tdTomato-C5ar2^fl/fl mice resulted in similar accumulation of inflammatory cells in the airways and the lung tissue (data not shown). However, IL-33 treatment reduced the frequency and the expression levels of tdTomato-C5aR2 in lung eosinophils and moDCs but not in macrophages or CD11b⁺ cDCs (Fig. 8A, 8C, 8E). Further, the frequency of tdTomato-C5aR2⁺ airway macrophages or their tdTomato-C5aR2 expression was not altered (Fig. 8B, 8D, 8F).

C5a exerts critical roles in the regulation of innate and adaptive immune responses through activation of its cognate receptors C5aR1 and C5aR2. C5aR1 signaling pathways intersect with TLR (40, 41) and FcγR signaling (3). Also, they modulate and control T cell proliferation and differentiation (34, 42). The role of C5aR2 in C5a-driven effector and regulatory functions is less well understood (12, 43). Initially, C5aR2 was considered solely as a decoy receptor counterbalancing the proinflammatory effector functions mediated by C5aR1 (13); however, more recent data show that C5aR2 exerts pro- and anti-inflammatory properties independently of C5aR1 (17–19). For example, C5aR2 selectively modulates LPS-driven IL-6 production from human macrophages (5). Previous studies suggested that C5aR2 is broadly expressed on several myeloid and lymphoid cells together with C5aR1 (43); however, C5aR2 expression levels are lower than those of C5aR1 (15, 19, 25). Until now, a detailed picture of C5aR2 expression in murine cells and tissues was lacking. Taking advantage of the floxed tdTomato-C5aR2 reporter knock-in mouse, we provide a detailed map of C5aR2-expressing immune cells in different tissues under steady-state conditions and upon pulmonary inflammation. We also evaluated tdTomato-C5aR2 expression in lung and intestinal epithelial cells. In contrast to C5aR1 and C5aR2 expression in human intestinal epithelial cells and a colonic cell line (44), we found no evidence for tdTomato-C5aR2 in intestinal or pulmonary epithelial cells.

First, we assessed the distribution of C5aR2 expression in neutrophils from different organs. In a previous article, neutrophils were identified as C5aR2⁺ cells; however, the same article stated that cells from the BM were negative for C5aR2 (25). This view was challenged by later studies showing C5ar2 mRNA expression, as well as protein expression, in BM neutrophils (19). In line with these studies, we found that neutrophils from BM, blood, and the peritoneal cavity were strongly positive for tdTomato-C5aR2. tdTomato expression mirrored C5ar2 mRNA expression, demonstrating that the reporter mouse strain is a reliable tool to determine C5aR2 expression. Moreover, C5aR2 expressed by neutrophils in tdTomato-C5ar2^fl/fl mice is fully functional, as evidenced by the fact that the recently described C5aR2 agonist P32 (5) suppressed C5a-mediated ERK1/2 phosphorylation downstream of C5aR1 signaling in neutrophils from the reporter mice. Also, we demonstrate that the C5aR2 reporter mouse allows tracking of C5aR2 expression by immunofluorescence microscopy. Importantly, we tested two commercially available Abs that are reported to specifically recognize C5aR2. To our surprise, these Abs showed strong signals with neutrophils or macrophages from WT mice, as well as from C5aR2-deficient mice, clearly demonstrating that they recognize structures other than C5aR2. This finding further supports the view that tracking of C5aR2 expression independently of Ab staining is required to provide a reliable picture about C5aR2 expression in tissues.

In the next step, we assessed the distribution of tdTomato-C5aR2⁺ cells in blood, spleen, brain, VAT, lung, LP of the small intestine, BAL, and BM. We found tdTomato-C5aR2⁺ cells in all tested organs, with the highest abundance in brain and BM.

Previous studies found important roles for anaphylatoxin receptors in the development of neurodegenerative diseases in mice and humans (10, 45). In line with these observations, we already observed strong expression of tdTomato-C5aR2 in naive mice. In a recent study, Fonseca et al. (46) elegantly showed the upregulation of C5aR1 and C5aR2 in human Alzheimer’s disease patients using anaphylatoxin receptor Abs. To ensure the specificity of the staining and identify C5aR1- and C5aR2-expressing cells in that study, the investigators combined several Ab clones, confirming the difficulties that we observed using C5aR2 Abs in mice. This finding is in line with a recent study in which the investigators encountered problems with the reliable use of C5aR2 Abs in the spinal cord, because all Abs showed staining in C5ar2^−/− mice (47). Using the tdTomato-C5aR2 reporter mouse, we observed C5aR2 receptor expression in resident microglia cells and macrophages. The strong tdTomato signal that we observed in both populations is in line with studies indicating that C5aR2 is expressed in microglia cells (45, 48, 49). Further, another study showed that astrocytes and neurons also express C5aR2 (46). A detailed and comprehensive picture of C5aR2 expression in the multiple cell types in the brain and neuronal tissue is beyond the scope of the current study and will be addressed in a future analysis.

In allergic asthma models, it was shown that the C5aR2 is required for full manifestation of the allergic phenotype, including the development of airway hyperresponsiveness and airway eosinophilia (27). In this article, we show that tdTomato-C5aR2 is strongly expressed in eosinophils from the BM, blood, lung, and LP of the small intestine and, to a lesser extent, the VAT. Interestingly, our findings suggest that the expression levels of C5aR2 in eosinophils are somewhat lower than in neutrophils, particularly in BM. In line with our observations in mice, human eosinophils were reported to express C5aR2, at least at the mRNA level (15). Pulmonary challenge with IL-33, a strong proinflammatory cytokine involved in the early response of the lung mucosal surface to airway insults had a strong impact on C5aR2 expression in eosinophils and moDCs. In fact, C5aR2 expression was decreased, as well as the frequency of C5aR2⁺ eosinophils and moDCs. We have previously shown that pulmonary eosinophils and moDCs also express C5aR1 (7) and that eosinophils upregulate C5aR1 during allergic asthma (50). Given that C5aR2 can control and counteract C5aR1-driven cell activation (18), these findings suggest that eosinophils and moDCs may become more sensitive to C5a-mediated cell activation in a pulmonary environment with high IL-33 levels.

It was shown that C5aR2 expression in human monocytes was upregulated in patients with Mycobacterium tuberculosis infection (51), suggesting a role for this receptor in monocyte/macrophage-mediated control of this pathogen. In line with an important role for C5aR2 in macrophage function, it was shown that atherosclerosis and vascular remodeling is attenuated in C5aR2-deficient mice. In vitro assays further revealed that proinflammatory and proatherosclerotic mediators from macrophages were reduced significantly when C5aR2 was depleted (52). Further, C5aR2 activation in macrophages has been suggested to regulate the development of obesity, because C5aR2 deficiency increased macrophage numbers in white adipose tissues, thereby contributing to obesity-associated pathologies (53). We found tdTomato-C5aR2 signals in all macrophage populations under steady-state conditions, including those from VAT, although the expression was lower compared with that of macrophages from other compartments. In future work, tdTomato-C5aR2 reporter mice might prove useful to track C5aR2 expression in VAT macrophages under obese conditions.

Investigating macrophages from BM cell culture, we found that only one third of these cells expressed tdTomato-C5aR2 with a relatively low fluorescence intensity. Recently, it was hypothesized that, in this setting, C5aR2 might play a more important role than C5aR1 in regulating IL-1β responses (54). Clearly, specific deletion of C5aR2 using the floxed tdTomato-C5aR2 reporter mouse will help to clarify the exact role of C5aR2 in NLRP3 inflammasome activation in macrophages.

In the initial article describing C5aR2, Ohno et al. (25) found that C5aR2 is expressed in immature, but not in mature, DCs. This view was challenged by Gerard and colleagues (15), who showed that C5aR2 expression is absent in immature DCs but was upregulated upon TLR4 or TLR9 stimulation in these cells. Using the tdTomato-C5aR2 reporter mouse, we demonstrate that the C5aR2 expression pattern strongly depends on the DC subset investigated. Although moDCs and CD11b⁺CD103⁻ DCs from lung express low levels of tdTomato-C5aR2 under steady-state conditions, the receptor is barely detectable in other DC subpopulations. Also, GM-CSF–differentiated DCs from BM were largely negative for C5aR2, which is in line with the observations by Ohno et al. (25). This weak expression of C5aR2 in DCs might explain the minor influence of C5aR2 activation on the process of splenic DC–driven naive T cell differentiation into Th1 or Th17 effector or regulatory T cells (55).

Taking into account the recently discovered important role for C5aR2 in complement-driven inflammasome activation of human CD4⁺ T cells (34) and considering the ongoing controversy regarding the expression of C5aR1 in murine T cells (6–8, 42), we tested T cell subsets from different sources for tdTomato-C5aR2 expression. Several articles have highlighted the role of C5aR1 for the regulation of T cell survival, proliferation, and differentiation into Th1 (40), Th2 (56), or Th17 (55, 57, 58) effector cells. Importantly, C5aR2 was also recently shown to regulate Th1 lineage commitment through its impact on NLRP3 activation in human CD4⁺ T cells (34). Similar to what we had previously observed for C5aR1 using GFP-C5aR1 reporter mice (7), we found that T cells completely lacked tdTomato-C5aR2 and C5ar2 mRNA expression, even after CD3/CD28 stimulation. Our findings suggest that the differences that we observed previously for C5aR1 expression between human and mouse T cells also apply to the expression pattern of C5aR2.

With regard to C5aR2 expression in murine B cells, very limited information is available. In the human system, C5aR2 expression was shown in plasma and memory B cells (15). In this article, we demonstrate that naive mouse B cells from spleen, blood, and mLNs homogeneously express C5aR2. These data were corroborated by C5ar2 mRNA expression in splenic B cells. These results are in sharp contrast to our findings in GFP-C5aR1 reporter mice (7); B cells from different lymphoid organs did not express C5aR1. The expression of C5aR2 in murine B cells independently from C5aR1 suggests an exclusive functional role for C5a on mouse B cells through C5aR2.

Finally, we determined tdTomato-C5aR2 expression in NK cells from different tissues. Interestingly, we found that only subpopulations of NK cells expressed the receptor in blood and spleen. In a previous report, low levels of C5ar2 mRNA expression were reported for NK cells from spleen, which was upregulated upon exposure to E. coli (15, 35, 36). Also, C5aR2 mRNA expression has been shown in human NK cells (15). More recently, this finding has been confirmed by another group, who also demonstrated intracellular expression of C5aR2 (35). The expression of C5aR2 in NK cells contrasts with previous findings obtained with GFP-C5aR1 reporter mice (7). In that study, we found no C5aR1 expression in NK cells from different tissues, suggesting that, similar to what we have observed in B cells, C5aR2 is the C5aR receptor that is exclusively expressed in mouse NK cells. Assessing the functional consequence of the expression of C5aR2, we found that NK cells, which lack C5aR2, are much more potent IFN-γ producers than their WT counterparts in response to IL-12/IL-18 stimulation. Further, specific ligation of C5aR2 with the C5aR2 agonist P32 suppressed IFN-γ production in WT NK cells. These findings demonstrate that C5aR2 activation controls IL-12/IL-18–induced IFN-γ production from NK cells, which plays an important role during early infection with intracellular pathogens; these data suggest that the complement system regulates early immune responses during such infections through NK cell control.

In summary, we provide a detailed map of C5aR2 expression in different cells from the innate and adaptive immune system. Our data clearly show C5aR2 expression in neutrophils, eosinophils, macrophages, moDCs, CD11b⁺CD103⁻ cDCs, B cells, and NK cells from different tissues. In contrast, we demonstrate that most of the DC subsets, as well as naive and activated T cells, do not express C5aR2. Also, we show a novel function for C5aR2 in NK cells and a novel role for IL-33 as regulator of C5aR2 expression in pulmonary eosinophils and moDCs. The floxed tdTomato-C5aR2 reporter mouse will prove useful to track C5aR2 expression and to conditionally delete this receptor in experimental disease settings to further delineate the enigmatic role of this anaphylatoxin receptor in health and disease.

The authors have no financial conflicts of interest.