The for Complement Anaphylatoxin C3a Is Expressed by Myeloid Cells and Nonmyeloid Cells in Inflamed Human Central : Analysis in Multiple Sclerosis and Bacterial Meningitis1

Philippe Gasque,2* Sim K. Singhrao,* Jim W. Neal,† Piao Wang,‡ Sakina Sayah,‡ Marc Fontaine,‡ and B. Paul Morgan*

The complement anaphylatoxins C5a and C3a are released at the inflammatory site, where they contribute to the recruitment and activation of leukocytes and the activation of resident cells. The distribution of the receptor for C5a (C5aR) has been well studied; however, the receptor for C3a (C3aR) has only recently been cloned, and its distribution is uncharacterized. Using a specific affinity-purified anti-C3aR peptide Ab and oligonucleotides for reverse transcriptase-PCR analysis, C3aR expression was char- acterized in vitro on myeloid and nonmyeloid cells and in vivo in the brain. C3aR was expressed by adult , cell lines, monocyte lines THP1 and U937, neutrophils, and monocytes, but not by K562 or Ramos. C3aR staining was confirmed by flow cytometry, confocal imaging, and electron microscopy analysis. A 65-kDa protein was immunoprecipitated by the anti-C3aR from astrocyte and monocyte cell lysates. Our results at the protein level were confirmed at the mRNA level. Using reverse transcriptase-PCR, Southern blot, and sequencing we found that C3aR mRNA was expressed by fetal astrocytes, astrocyte cell lines, and THP1, but not by K562 or Ramos. The astrocyte C3aR cDNA was identical with the reported C3aR cDNA. C3aR expression was not detected in normal brain sections. However, a strong C3aR staining was evident in areas of inflammation in multiple sclerosis and bacterial . In meningitis, C3aR was abundantly expressed by reactive astrocytes, , and infiltrating cells ( and neutrophils). In multiple sclerosis, infiltrating did not express C3aR, but a strong staining was detected on smooth muscle cells (pericytes) surrounding blood vessels. The Journal of , 1998, 160: 3543–3554.

he is an important component of the phages into the inflammatory site and activate the cells to phago- innate immune system, with the capacity to recognize and cytose invading pathogens. C3a and C5a also contribute to the T eliminate a large variety of pathogens without the recruit- activation of resident cells in the infected tissue, which will then ment of elements of the adaptive immune system (1, 2). Among express increased level of , chemokines, acute phase pro- the active products conferring this capacity, formation of the mem- teins (such as complement proteins), and adhesion molecules (3– brane attack complex and deposition of opsonins on the membrane 7). C3a and C5a are regulated physiologically by serum car- of the target cell are key elements for the efficient removal of boxypeptidase N, which inactivates by cleaving a single arginine pathogens. However, the anaphylatoxins released into the fluid residue from either molecule; some pathogens possess specific phase are critical, in that they attract and activate phagocytic cells peptidases that inactivate the anaphylatoxins (8). and may also activate resident cells at the inflammatory site. The Cells respond to anaphylatoxins via specific receptors. The re- complement anaphylatoxins C5a and C3a are powerful chemoat- ceptor for C5a (C5aR) was cloned in 1990 (9), and the generation tractants that recruit polymorphonuclear cells (PMN)3 and macro- of specific Abs using synthetic peptides enabled the tissue distri- bution of C5aR to be examined (10, 11). The C5aR was expressed on both myeloid cells (monocytes/macrophages, neutrophils, and *Department of Medical Biochemistry and † (Neuropathology Laboratory), eosinophils) and nonmyeloid cells (hepatocytes, epithelia, endo- University of Wales College of Medicine, Cardiff, Wales, United Kingdom; and thelia, mast cells, vascular smooth muscle, and glial cells) (3–7, ‡INSERM Unit 78, Institut Fe`de`ratif de Recherches Multidisciplinaires sur les Pep- tides, Chemin de la Brete´que, Bois-Guillaume, France 10–12, 13–16). Much less is known about the cellular distribution Received for publication August 21, 1997. Accepted for publication December of the C3aR. C3a binding experiments and assessment of C3a 9, 1997. functional effects indicated that a receptor for C3a was present on The costs of publication of this article were defrayed in part by the payment of page monocyte/ cells and cell lines, platelets, polymorpho- charges. This article must therefore be hereby marked advertisement in accordance nuclear leukocytes, mast cells, and adipocytes (3, 12, 17–23). The with 18 U.S.C. Section 1734 solely to indicate this fact. human C3aR cDNA has recently been cloned from the HL60 cell 1 This work was supported by the Medical Research Council, the Wellcome Trust, a Medical Research Council Career Development Award Fellowship (to P.G.), and a line (24), LPS-activated neutrophils (25), and the PMA-differen- Wellcome Trust Senior Research Fellowship (to B.P.M.). tiated U937 monocyte cell line (26). Northern blot analysis re- 2 Address correspondence and reprint requests to Dr. Philippe Gasque, Department of vealed that the 2.2-kb C3aR mRNA was expressed in all tissues, Medical Biochemistry, Brain Inflammation and Immunity Group, University of Wales College of Medicine, Cardiff, Wales CF44XN. E-mail address: WMBPG @cardiff.ac.uk 3 Abbreviations used in this paper: PMN, polymorphonuclear cells; C5aR, receptor ferase; GFAP, glial fibrillary acidic protein; MNC, mononuclear cells; PE, phyco- for C5a; C3aR, receptor for C3a; MS, multiple sclerosis; BM, bacterial meningitis; erythrin; DAB, diaminobenzidene; IF, immunofluorescence; DAF, decay-accelerating CNS, ; MAP, multiple array peptide; GST, -S-trans- factor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Copyright © 1998 by The American Association of Immunologists 0022-1767/98/$02.00 3544 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS particularly in lung, placenta, heart, spleen, and brain (25). How- ever, little is known about the expression of C3aR at the protein level and on nonmyeloid cells in these tissues. We and others are interested in the role of complement in the brain, and we have shown that a full complement system can be expressed by resident cells in the brain, particularly after stimulation (for review, see Ref. 27). The brain is an immuno- privileged organ isolated from the peripheral immune system, and we have proposed that complement expressed locally in response to infection or inflammation plays an important role as an an- tipathogen (27). Human and mouse glial cells respond to and are activated by anaphylatoxins (14, 28, 29), and we have recently shown that the C5aR was expressed by astrocytes, the most abun- dant glial cell type (14). Immunohistochemistry demonstrated that expression of C5aR was low in normal brain but abundant in in- flamed brain, suggesting that the anaphylatoxin receptor was in- volved in the inflammatory process (16). We here describe the expression of the C3aR by myeloid cells and brain cells in vitro and in brain tissue. C3aR expression was demonstrated in cells and tissue at the protein and mRNA levels, and immunohistochemistry was conducted on normal, multiple sclerosis (MS), and bacterial meningitis (BM) brain tissue sec- tions. In myeloid and glial cells the expressed C3aR was a protein of 65 kDa; expression of the receptor was highly elevated on my- FIGURE 1. Immunodetection by cytofluorometry of membrane C3aR eloid cells, astrocytes, microglia, and smooth muscle cells (peri- on the THP1 monocyte cell line and the Ramos B cell line cytes) in inflamed CNS tissue. The results further emphasize the using a specific affinity-purified Ab. THP1 and Ramos in suspension were incubated with different dilutions of the affinity-purified anti-C3aR from important role played by complement activation in CNS rabbit B (represented on the x-axis). The intensity of staining was analyzed inflammation. by FACS using PE-conjugated secondary Ab and the mean of fluorescence 2 (FL2) is given on the y-axis for each dilution of the Ab and for THP1 (open circle) and Ramos (closed circle). Each histogram shows data from Materials and Methods one experiment, but is representative of multiple analyses for each cell Peptide synthesis, production of Abs, and affinity purification type, THP1 (n ϭ 5) and Ramos (n ϭ 5). THP1 staining was completely abolished when cells were incubated with Ab and a 100-fold molar excess A multiple array, 30-amino acid, C3aR peptide PSGFPIEDHETSPLDNS of the competitive MAP-C3aR peptide (data not shown). The inset is a DAFLSTHLKLFPS (MAP-C3aR peptide) corresponding to amino acid hydrophilicity plot of the human C3aR protein (amino acids 160–340) to residues 270 to 300 of the C3aR loop (see Fig. 1) was synthesized on an indicate that the MAP-C3aR peptide (amino acids 270–300) was selected Applied Biosystems Synergy Synthesizer (Applied Biosystems, War- rington, U.K.). Anti-peptide antisera were raised in New Zealand White in the second extracellular loop between transmembrane domains 4 and 5. rabbits by repeated s.c. immunization (total of five) with peptide (100 ␮g/ animal/immunization) in CFA using standard protocols. The animals were test-bled at intervals, and the titer of the anti-peptide response was mea- rose 4B according to the manufacturer’s protocol (Pharmacia, sured. When the antipeptide response had plateaued, the animals were sac- Piscataway, NJ). rificed by exsanguination, and the antiserum was stored at Ϫ20°C. Total IgG was purified on glass-bound, protein A-Sepharose (Prosep A, Chemicals, cytokines, and Abs Bioprocessing, Princeton, NJ). Specific IgG was then affinity purified from Recombinant human IFN-␥, IL-1␤, and TNF-␣ were gifts from Hoff- the total IgG using the same C3aR peptide coupled to Sepharose. The C3aR mann-La Roche (Nutley, NJ). PMA, protein A-Sepharose, pepstatin, and peptide Sepharose was generated by coupling directly 5 mg of MAP-C3aR leupeptin were obtained from Sigma. Mouse and rabbit Abs against the peptide with 1.5 g (5–7 ml) of cyanogen bromide-activated Sepharose 4B N-terminal extracellular sequence of the human C5aR have been charac- according to the manufacturer’s protocol (Pharmacia Biotech, St. Albans, terized previously (11, 14). Abs against CD Ags (CD4, CD14, CD16, U.K.). IgG (200–250 mg) was incubated overnight (4°C with mixing) with CD18, CD19, CD35, and CD68) were purchased from Dako (High Wy- the C3aR peptide Sepharose. After washing repeatedly with PBS, the combe, U.K.). mAbs against CD44 (clones BRIC222 and BRIC235) and Sepharose was poured into a column, and Ab was eluted using 0.1 M CD59 (BRIC229) were obtained from IBGRL (Elstree, U.K.). Other Abs glycine-HCl buffer at pH 2.5. The yield of affinity-purified anti-peptide IgG used for immunocytochemistry of adherent cells on coverslips and for im- was 20 to 25 mg from 200 to 220 mg of total IgG (corresponding to 20–25 munohistochemistry of tissue sections were mouse anti-glial fibrillary ml of rabbit serum). acidic protein (anti-GFAP; clone GA5; 1/1000; Sigma), mouse anti- For FACS analysis, affinity-purified anti-C3aR (1.5 mg) was also bio- galactocerebroside (1/1000) (16), mouse anti--specific enolase tinylated with 1 mg of biotinamidocaproate N-hydroxysuccimide (Sigma, (clone BBS/NC/VI-H14; 1/2000; Dako), mouse anti-CD11b (clone Poole, U.K.) and dialyzed to eliminate free . All preparations of Abs 2LPM19c; 1/50; Dako), mouse anti-CD68 (1/50; clones KP1, PG-M1, and were stored at 4°C in PBS/0.1% NaN . The specificity and titer of the 3 EBM11; Dako), and mouse anti-HLA class II (1/50; clone LN3; Biotest, anti-peptide Ab were assessed in an ELISA using 96-well plates coated Solihull, U.K.). with the MAP-C3aR peptide or nonspecific peptides (C3a peptide, WWGKKYRASKLGLAR; MAP-C5aR peptide, MNSSFEINYDHYGT Cell preparation and culture MDPNIPADGIHLPKRQ) produced in our laboratory (14). In some exper- iments, the anti-C3aR was blocked by preincubation with various concen- Primary cultures of human fetal astrocytes were grown in our laboratory trations of the MAP-C3aR peptide before application to the 96-well plate. from tissue supplied by the Medical Research Council Fetal Tissue Bank Specificity was also assessed by flow cytometry on C3aRϩ (THP1) and (London, U.K.). All cultures were established from fetal brain (6- to 10- C3aRϪ (Ramos) cell lines and by Western blot analysis of cell lysates and wk-old fetus) as described previously (14). Primary cultures (passage 0, a C3aR loop-glutathione-S-transferase (GST) fusion protein. days 1–7) contained clusters of (demonstrated by immunostaining To detect C3aR extracted from cells, lysates were immunoprecipitated with an anti-70K neurofilament mAb) that survived 7 to 10 days, lying on using anti-C3aR Sepharose. For this purpose, 5 mg of affinity-purified anti- a monocellular layer. Some 60 to 80% of the cells in this monocellular C3aR was coupled to 1.5 g (5–7 ml) of cyanogen bromide-activated Sepha- layer expressed the astrocytic marker, GFAP; 5 to 7% were microglia, and The Journal of Immunology 3545

15 to 40% were fibroblasts. Microglia were removed by shaking (350 rpm, suspension. The remainder of the FACS protocol was identical with that 37°C, for 1.5 h followed by a second shaking at 250 rpm and 37°C over- described above. night). Adherent cells after these two sequential shakings were subcultured in DMEM containing 10% FCS (Life Technologies, Paisley, U.K.), 1% Immunocytochemistry L-glutamine (Life Technologies), and 1% penicillin/streptomycin (Life For CD44 and C3aR staining, adult astrocytes, fetal astrocytes, and astro- Technologies). Experiments were conducted from passages 2 to 5, when cyte cell lines were cultured on sterile glass coverslips for 5 to 8 days, and Ͼ ϩ Ͻ cultures contained 95% GFAP ve cells and 5% neurons, fibroblasts, after washing with PBS, cells were fixed with 1% formaldehyde for 20 and microglia. min. Cells were then washed intensively in PBS/0.2 M glycine to block Normal adult temporal lobe tissue was obtained fresh from of aldehyde groups. For GFAP immunostaining, cells were fixed and perme- patients undergoing therapeutic resection for intractable . The tis- abilized with a mixture of 95% ethanol/5% acetic acid for 5 min at Ϫ20°C. sue was used for culture of adult astrocytes as previously described (16, Abs were used at optimal dilution (1 ␮g/ml in PBS/1% BSA) and incu- 30). The use of human tissue was in accordance with procedures and reg- bated with fixed cells overnight at 4°C in a humid chamber. Affinity- ulations established by the local ethical committee of the University of purified rabbit anti-C3aR was tested over the concentration range 7.5 to Wales College of Medicine. 0.81 ␮g/ml in the presence or the absence of competing peptide. After The two human astrocyte cell lines, T98G and CB193, were cultured washing, coverslips were incubated for 30 min at 37°C with either FITC- and characterized as described previously (14). Neutrophils (PMN) and Ј labeled secondary Ab (F(ab )2 rabbit anti-mouse IgG (1/100; Dako) or goat mononuclear cells (MNC) were isolated from heparinized blood obtained anti-rabbit IgG (Sera-Lab, Sussex, U.K.)) or peroxidase-labeled secondary by venepuncture of healthy volunteers. Briefly, leukocytes were separated Ab (rabbit anti-mouse IgG (1/100; Bio-Rad, Richmond, CA) or goat anti- from erythrocytes by dextran sedimentation using 0.6% (w/v) dextran rabbit IgG (1/100; Bio-Rad)). After intensive washing, coverslips for flu- (Fisons, Loughborough, U.K.). The leukocyte-rich upper layer was then orescence microscopy were mounted in Citifluor (Citifluor, London, U.K.) fractionated by layering on Histopaque (Sigma) followed by centrifugation and sealed. For peroxidase immunostaining, coverslips were incubated in a at 220 ϫ g for 25 min at room temperature. MNC were collected from the solution of DAB/H2O2 (Sigma) before hematoxylin counterstaining and interface and washed in PBS/BSA. Residual erythrocytes in the PMN-rich mounting. Fluorescence was imaged by confocal laser scanning micros- cell pellet were removed by hypotonic lysis, and PMN were washed in copy on a Leica TCS microscope (Leica, Heidelberg, Germany). Twelve PBS-BSA. Cytocentrifuge preparations of cells were stained with Wright’s optical sections were collected per field at 0.3- to 0.5-␮m intervals from the stain, and Ͼ90% of cells were PMN. FACS analysis of PMN preparations ϩϩϩ ϩϩϩ ϩϩ bottom to the top of the cell. Sections were then assembled as extended revealed that 90 to 95% of cells were CD11b , CD18 , CD16 , focus views, or individual sections were viewed as a gallery. DAB immu- ϩ Ϫ ϩ Ͼ CD14 , CD19 , CD35 . Lymphocytes and monocytes constituted 90% nostaining (brown positive staining) was photographed on a Leica DMLB of the MNC preparation, and no contaminating PMN were identified. In the microscope with brightfield at two magnifications (ϫ500 and ϫ1250). MNC preparation, 10 to 15% of cells analyzed by FACS were strongly ϩ Double staining of the CB193 astrocyte cell line for CD44 and C3aR CR3 . Human monocyte-derived cell lines, THP1 and U937, and neuro- was performed in an identical manner, except that both Abs were applied blastoma cell lines, IMR32 and SKNSH, were obtained from American to the same coverslip and were detected using a rhodamine-conjugated goat Type Culture Collection (Rockville, MD). The human endothelial cell line anti-mouse IgG (Sigma) and a FITC-conjugated goat anti-rabbit IgG ECV 304, the human erythroleukemia cell line K562, the human B lym- (Seralab), respectively. Fluorescence was imaged by confocal microscopy phocyte cell line Ramos, and the human T lymphocyte cell line Molt 4 using specific cut-off filters for FITC and rhodamine. were obtained from European Cell Culture Collection (ECACC) (Salis- bury, U.K.). A subclone of the B lymphocyte cell line Raji (Rajiϩ3) ex- Sources of tissues, processing, and immunohistochemistry pressing high levels of CD59 (31) was routinely grown in our laboratory and used for RT-PCR analysis. The human hepatoma cell line HepG2 was Brain tissue was obtained locally at or from specialist tissue col- obtained from Dr. M. Daveau (INSERM Unit 78, Rouen, France). The lections. Tissue was collected from individuals with a variety of brain human NK cell line (YT) was a gift from Dr. G. Griffiths (London, U.K.) disorders (demyelination and CNS infection) and from normal surgical and was cultured in the presence of IL-2. controls. Tissue from cases of MS (three acute plaques and one chronic Cell phenotype was confirmed using immunocytochemistry or FACS plaques) were obtained from Dr. Jia Newcombe (MS Society Laboratory, analysis, as described previously (14). Astrocytes and the two astrocyte cell London, U.K.) and locally (J. W. Neal, Neuropathology Laboratory, lines presented the phenotype (GFAPϩ, CD11bϪ, HLAIIϪ). For some ex- Cardiff, U.K.). Tissues from four cases of BM were obtained locally, all of periments, cells were stimulated with cytokines (200 IU/ml of IFN-␥ or which were characterized by a marked infiltration of neutrophils in the IL-1␤ and 1000 IU of TNF-␣) or with PMA (10 ng/ml) for 24 h. To induce meninges. Normal control brain tissues was obtained at autopsy or brain differentiation of the two monocyte cell lines, THP1 and U937, cells were surgery from individuals with no evidence of neurodegenerative disease, treated with 10 ng/ml of PMA over a period of 3 days, replacing medium ischemia, or . Autopsy samples were obtained at a similar postmor- with fresh PMA-containing medium every day. Under these conditions, tem interval (maximum of 30 h) as the disease samples. Brains were cut PMA was not toxic to the cells, which by day 2 had adhered strongly to the coronally, and individual blocks from areas of the brain containing mac- plastic. roscopic evidence of pathology were dissected. Tissue was either snap- Ϫ Monocyte cell lines THP1 and U937 were characterized as the mono- frozen and kept at 40°C or fixed in 10% formalin before processing for cyte/macrophage (the latter after PMA (10 ng/ml) treatment for 3 days) cell cryosections or paraffin wax embedding and sectioning, respectively. phenotype with, respectively, for monocyte/macrophage: GFAPϪ/Ϫ, Rehydrated paraffin sections were counterstained with hematoxylin/eo- CD11bϩ/ϩϩϩ, CD68ϩ/ϩϩϩ, CD35ϩ/ϩϩ, HLAIIϩ/ϩϩ, C5aRϩ/ϩϩϩ, as pre- sin to display morphology, and was used to identify viously described (14). demyelinating plaque areas in MS tissue (16). Rehydrated paraffin wax sections and cryosections (8 ␮m) from normal and diseased brains were immunostained with different dilutions of the affinity-purified anti-C3aR ␮ Flow cytometry and testing specificity using competitive (7.5 to 0.81 g/ml) and anti-cell markers diluted in PBS/BSA using an indirect immuno-horseradish peroxidase/DAB method as described previ- MAP-C3aR peptide ously (32). Peroxidase-labeled swine anti-rabbit IgG and rabbit anti-mouse Astrocytes or cell lines were harvested from culture by incubation in FACS IgG (Dako; 1/100 dilution) were used as secondary Abs. For double immunofluorescence (IF), tissue sections were incubated buffer (PBS containing 2% BSA and 0.1% NaN3) supplemented with 10 mM EDTA. Cells were washed and resuspended at 106 cells/ml in the same simultaneously with the rabbit anti-C3aR and mouse anti-brain cell mark- buffer without EDTA, incubated with the appropriate primary Ab (0.5–5 ers (clone GA5 for GFAP and clone LN3 for HLA class II) followed by ␮g/ml unless stated otherwise) for 30 min on ice, washed three times in FITC-labeled goat anti-rabbit (Sera-Lab) and rhodamine-labeled goat anti- cold FACS buffer, incubated with the appropriate phycoerythrin (PE)-la- mouse (Sigma) Abs. Fluorescence was imaged on a Leica DMLB epiflu- beled secondary Ab for 30 min on ice, and washed an additional three times orescence microscope using specific filters. before analysis on a Becton Dickinson FACScan (San Jose, CA). PE-con- Cell lysates, Western blotting, and immunoprecipitation jugated rabbit anti-mouse IgG (Dako) and PE-conjugated goat anti-rabbit IgG (Sigma) were used at a 1/100 dilution. When cells were incubated with Western blotting was performed on cells (astrocytes and cell lines) solu- the biotin-labeled affinity-purified anti-C3aR, the staining was visualized bilized in PBS containing 2% Nonidet P-40 together with enzyme inhibi- after incubation with FITC-labeled avidin (Sigma) (1/100) or with R-PE- tors as previously described (14). To enrich for C3aR before Western blot- labeled streptavidin (Jackson ImmunoResearch, Stratech, Luton, U.K.). To ting, cell lysates were immunoprecipitated. Freshly made cell lysate (1 ml; confirm the specificity of the anti-C3aR, dilutions of Abs (0.4–3.25 ␮g/ml) 2 ϫ 107 cells/ml) was incubated with mixing overnight at 4°C with 50 ␮l were preincubated with a fixed concentration of competing MAP-C3aR of affinity-purified rabbit anti-C3aR Sepharose. After five washings with peptide (5 ␮g/ml) at 4°C for 30 min to 1 h before applying them to the cell 0.5% Nonidet P-40 in PBS, C3aR retained on the Ab-Sepharose was eluted 3546 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS

by mixing in an Eppendorf microfuge with 50 ␮l of 0.1 M glycine, pH 2.5, (EMBL J03564) were CCC ATT GCT GTT GGT ACC GT (181) and ACA for 30 min at room temperature. Sepharose was removed by centrifugation CAG GTT GGT AGT CGT (503). The sequences of human C5aR primers at 10,000 ϫ g for 10 min, and the eluted protein was diluted in an equal (EMBL no. M62505/JO5327) were GAC CAG AAC ATG AAC TCC TT volume of SDS-PAGE sample buffer. The incorporation of the glycine (16) and TGT CGC CTA CAC TGC CTG (1083). The sequences of human elution step prevented the release of rabbit anti-C3aR, simplifying inter- C4 primers (EMBL no. K02403) were CGG GTC TTT GCW CTG GAT pretation of the Western blots. THP1 and Ramos cell lysates were tested as CA (515) and CTT CAC CTC RAA GTT GGG AA (773). Samples of positive and negative controls for C3aR expression to validate the immu- RT-PCR products were loaded onto a 1% agarose gel in 1ϫ TBE buffer, noprecipitation protocol. Samples were fractionated on 15% or 7.5% SDS- separated by electrophoresis at 50 to 100 mA, and transferred onto nylon PAGE, electroblotted onto nitrocellulose and stained overnight at 4°C with membranes (Hybond Nϩ, Amersham, Aylesbury, U.K.) for Southern blot- either the affinity-purified anti-C3aR (0.75 ␮g/ml) or rabbit anti-decay- ting. Ladders (123 bp and 1 kb; Life Technologies) were used as DNA size accelerating factor (anti-DAF; CD55) also produced in our laboratory (0.25 markers. All C3aR cDNAs obtained after RT-PCR of the astrocyte cell line ␮g/ml), essentially as previously described (14). After washing and incu- CB193 were purified (Quiaquick PCR purification column, Qiagen, Dork- bation with peroxidase-labeled goat anti-rabbit IgG (Bio-Rad; 1/4000), ing, U.K.) and sequenced using the ABI PRISM dye terminator cycle se- blots were developed using the enhanced chemiluminescence system quencing kit from Perkin-Elmer (Warrington, U.K.) and the 373A auto- (ECL, Pierce, Chester, U.K.). Prestained broad range protein markers from matic sequencer from Applied Biosystems (Foster City, CA). Sequences New England Biolabs (Beverly, MA) were used as m.w. standards. Human were compared with the published human C3aR using the DNAstar soft- THP1, unstimulated or differentiated with PMA (10 ng/ml, 3 days), were ware package (DNA Star, London, U.K.). used as the C3aR-positive control. To eliminate the possibility of cross-reactivity of affinity-purified anti- cDNA probes, nucleotide sequencing, and Southern blot C3aR Ab with C5aR, PMA-differentiated THP1 cell lysate (1 ml; 2 ϫ 107 cells) was immunoprecipitated with W17 mouse monoclonal anti-C5aR The human C3aR cDNA probe was obtained from PMA-differentiated preadsorbed on protein A-Sepharose (Sigma; 20 ␮g of Ab for 6.25 mg of U937 RNA by RT-PCR using two C3aR oligonucleotides (754 and 1391). PAS). After elution in Laemmli buffer, samples were tested by Western The RT-PCR product was purified by the PCR DNA purification system blot using the affinity-purified anti-C3aR. (Hybaid) and subcloned in the PGEM-T plasmid (Promega, Southampton, U.K.). Positive bacterial clones were identified by blue/white screening and were selected. Inserts were PCR amplified using primers SP6 and T7 (de- RNA extraction and RT-PCR for C3aR mRNA veloped in-house) derived from vector sequence flanking the insert site and Total RNA from unstimulated and stimulated cultures of human astrocytes then sequenced as described above. The probe isolated by this procedure and cell lines was prepared using the Ultraspec RNA isolation system was 100% homologous with the reported C3aR cDNA (24) between po- (Biotecx, Houston, TX) according to the manufacturer’s instructions. RNA sitions 754 and 1391 bp. integrity was confirmed on agarose gels, and concentrations were deter- For Southern blotting the C3aR probe was labeled by the random oligo ␣ 32 mined from absorbance at 260 nm. Before reverse transcription, total RNA procedure with [ - P]dCTP (3000 Ci/mmol; Redivue, Amersham). Blots (50 ␮g) was treated for 20 min at 37°C with 10 U of RQ1 RNase-free were prehybridized (2 h) and hybridized (16–20 h) in a hybridization oven ϫ ϫ DNase (Promega, Madison, WI) in 100 ␮l of buffer (40 mM Tris/HCl (pH at high stringency (50% deionized formamide, 5 SSPE, 1% SDS, 5 Denhardt’s reagent, 5% dextran sulfate, and 100 ␮g/ml denatured and frag- 8), 10 mM NaCl, 6 mM MgCl2, and 10 mM CaCl2) and 200 U of RNAsin (Promega) to remove all trace of DNA. The mixture was then phenol and mented herring sperm DNA). Posthybridization washings were performed ϫ ϫ chloroform extracted. The aqueous phase was ethanol precipitated, and the in (2 SSPE/0.1% SDS, twice at room temperature; 2 SSPE/0.1% SDS ϫ ϫ pellet after centrifugation was resuspended in diethylpyrocarbonate-treated for1hat68°C; 1 SSPE/0.1% SDS for1hat68°C; 0.1 SSPE/0.1% water. SDSfor1hat68°C). Blots were then exposed to Kodak 4 film (Eastman The reverse transcription was conducted at 37°C for 120 min in 30 ␮l Kodak, Rochester, NY) for 5 to 10 min at room temperature. (final volume) with 2 ␮g of total RNA (DNA-free), 60 U of RNAsin (Pro- mega), 1 mM dNTPs (Bioline, London, U.K.), 250 pmol of random hex- GST-C3aR loop fusion protein amer primers (pdN6 from Pharmacia), and 400 U of Moloney murine leu- To further confirm the specificity of the anti-C3aR peptide we generated a kemia -RT (Life Technologies, Paisley, U.K.) in the reaction buffer fusion protein by cloning the cDNA encoding the second C3aR extracel- (10 mM Tris/HCl, 15 mM KCl, 0.6 mM MgCl2, and 5 mM DTT). The lular loop (between TM4 and TM5 domains, amino acids 161–332) in a absence of contaminants was routinely checked by RT-PCR assays of neg- GST fusion protein system (Pharmacia). The C3aR loop cDNA was pro- ative control samples, in which the RNA samples were replaced with sterile duced by RT-PCR from PMA-differentiated THP1 RNA using two specific water, or Moloney murine leukemia virus-RT was not added. oligonucleotides (F-C3aR-s (562, CGG GAT CCC GGG AAA TCT TCA ␮ PCR was conducted with 3 l of reverse transcribed RNA mixture in a CTA CAG AC) and R-C3aR-s (1077, CGG GAT CC TCA G GGT GTT ␮ 50- l final reaction volume with 100 pmol of each C3aR primer (generated GGC ACT TGA TCG TC) and Vent DNA polymerase (New England in-house on a Beckman oligo synthesizer, Beckman, Palo Alto, CA) in Biolabs, Hertfordshire, U.K.) to minimize PCR errors. The sequence TCA ϫ ␮ 10 buffer (Promega) containing 1.5 mM MgCl2, 200 M dNTP, and 1.25 (double underlined) was introduced in the downstream primer to create a U of Taq DNA polymerase (Promega). The PCR protocol used was: de- stop codon TGA. A BamHI site (underlined) was introduced in both prim- naturation step at 94°C for 4 min; five cycles of 94°C for 30 s, annealing ers and was used to clone the C3aR cDNA into the pGex-2T expression 60°C for 1 min, and extension at 72°C for 2 min; 25 cycles of 94°C for vector. After transfection in Escherichia coli (strain BL21, Pharmacia), 10 s, 60°C for 30 s, and 72°C for 45 s; and elongation at 72°C for 15 min. plasmid from individual bacterial colonies was tested to confirm the cDNA PCR was performed in a Hybaid (Teddington, U.K.) Omnigene insert orientation and the fidelity of the C3aR cDNA sequence. The ex- thermocycler. pression of the fusion protein was induced by the addition of 1 mM iso- Seven oligonucleotides were selected from the published C3aR cDNA propyl-␤-thiogalactopyranoside (Promega) in 0.5 l bacterial culture (OD sequence (23–25). Their sequences from 5Ј to 3Ј and their positions in 600 nm ϭ 0.8) for 2 to 3 h. Bacteria were pelleted at 5000 ϫ g for 30 min, and parentheses are given according to the sequence published by Roglic et al. lysates were prepared using 1% Nonidet P-40/PBS. Crude lysate was an- (EMBL access no. U28488) (24): GCT CAT CCC CTC CAT CAT TG alyzed by Western blot, and the expressed fusion protein was affinity pu- (, 369), GGG TGG TGG CTT TTG TGA TG (sense, 521), GTC CCC rified on glutathione-Sepharose 4B according to the manufacturer’s in- ACT GTC TTC CAA CC (sense, 754), GGT TGG AAG ACA GTG GGG structions. The predicted size of the fusion protein was 50 kDa (30 kDa for AC (antisense, 754), CAG CAG GAA ACC CAC TA (antisense, 1122), GST and 19.4 kDa for the C3aR loop). AGG GCA TAA AGG AAG GGA TT (antisense, 1391), and TGA ATG GAC TGC CTT GCT TT (antisense, 1433). All samples were subjected to RT-PCR for ␤-actin and glyceraldehyde-3-phosphate dehydrogenase Results (GAPDH) as a positive control and as an internal standard not affected by Production of a specific anti-C3aR Ab any cytokine or PMA treatment. The sequences of human ␤-actin (EMBL M10278) primers were: actin-1, CTA CAA TGA GCT GCG TGT GG A 30-amino acid sequence from the second extracellular loop (311); and actin-2, CTC ATT GCC AAT GGT GAT GA (800). The se- (amino acids 270–300; inset in Fig. 1) was chosen for Ab produc- quences of GAPDH (EMBL M33197) primers were: GAPDH-Up, GAA tion based upon predictions of hydrophilicity/antigenicity. This CGG GAA GCT TGT CAT CA; and GAPDH-Do, TGA CCT TGC CCA peptide had no significant homology with any other cloned protein CAG CCT TG. RT-PCR analysis for complement receptor type 2 (CR2), C5aR, and C4 reported in the BLAST databank and had no homology with the mRNAs was also conducted on all RNA samples to confirm the specificity human C5aR. The C3aR peptide was synthesized as a multiple of the RT-PCR protocol (20, 33). The sequences of human CR2 primers array on a lysine core and was used for immunization of rabbits. The Journal of Immunology 3547

Table I. Immunodetection of C3aR by FACS analysis a

Mean of Fluorescence 2 (PE staining)

Non. Rb Mab Imm. Rb anti-C3aR anti-C5aR Mab-anti-CD44

THP1, monocyte (n ϭ 7) 3 Ϯ 1.5 30 Ϯ 430Ϯ 7 1750 Ϯ 180 THP1/PMA 3 days 5 Ϯ 257Ϯ 8 737 Ϯ 55 1560 Ϯ 133 (n ϭ 5) Ramos B cells (n ϭ 7) 1 Ϯ 0.5 1 Ϯ .2 3 Ϯ .2 ND Neutrophils (n ϭ 3) 1.5 Ϯ 145Ϯ 4 377 Ϯ 26 607 Ϯ 24 Monocyteb (n ϭ 3) 3 Ϯ 129Ϯ 7 246 Ϯ 65 2776 Ϯ 709 Lymphocyteb (n ϭ 3) 1 Ϯ .2 4 Ϯ 2 2.2 Ϯ 1 921 Ϯ 75 CB193, astrocyte (n ϭ 4) 2 Ϯ 121Ϯ 776Ϯ 13 3830 Ϯ 214 T98G, astrocyte (n ϭ 2) 1 Ϯ 0.5 10 Ϯ 3 124 Ϯ 13 ND

a Cells were stained according to the protocol described in Materials and Methods using nonimmune rabbit antiserum (Non. Imm. Rb, IgG cut), affinity-purified rabbit anti-C3aR, W17 mouse monoclonal anti-C5aR (Mab anti-C5aR; Professor O. Go¨tze), and, as a positive control, mouse anti-CD44, clone BRIC222. ND, not determined. b Gated population from FACS analysis of the mononuclear cell preparation.

(Table I). To characterize in more detail the regulation of C3aR expression by monocyte cell lines, U937 were cultured for 48 h in the presence of three different concentrations of recombinant cy- FIGURE 2. Titration of the affinity-purified anti-C3aR Ab in the pres- tokines (IFN-␥, IL-1␤, and TNF-␣) or phorbol ester (PMA). C3aR ence of the competitive MAP-C3aR peptide. Serial dilutions of the affinity- expression on the cell membrane was analyzed by FACS, and the purified anti-C3aR (x-axis, nanograms per milliliter of Ab) were tested by results are presented in Table II. PMA increased the expression of ELISA on 96-well plates coated with either MAP-C3aR peptide (circle) or C3aR, but only by a factor of 1.5 at 48 h poststimulation; the most unrelated C3a peptide (square) in the absence (open circle or open square) dramatic effect was obtained after stimulation of U937 with IFN-␥. or the presence (filled circle or filled square) of a fixed concentration of the Even at a very low concentration (10 IU/ml) IFN-␥ increased ␮ competitive MAP-C3aR peptide (5 g/ml). Plates were then incubated C3aR expression by 2-fold, and at 1000 IU/ml, C3aR expression with peroxidase-conjugated secondary Ab, and specific binding of the Ab was elevated by 6-fold. TNF-␣ and IL-1␤ had no effect on C3aR was assessed by developing the plate with OPD substrate (Dako) and mea- ␥ surement of the OD at 490 nm. The y-axis represents the mean of OD expression by U937 (Table II). IFN- also up-regulated C3aR ex- readings of three wells for each dilution of Ab. The three curves are the pression by THP1. Neutrophils and monocytes both expressed results of three independent experiments. C3aR, whereas the various lymphocyte subsets were all negative (Table I). C3aR was not detected on the membranes of the undif- ferentiated Rajiϩ3 line, the K562 erythroleukemia cell line, The antiserum was purified by affinity chromatography on peptide- the Molt 4 T lymphocyte cell line, or the YT NK cell line (data not Sepharose. The yield of affinity-purified anti-C3aR Abs was ap- presented). proximately 10% of the applied IgG. Affinity-purified anti-C3aR IgG was used throughout the study. The specificity of the anti- Astrocyte cell lines and fetal and adult astrocytes express C3aR was tested first by staining the THP1 monocyte line and the constitutively membrane C3aR Ramos B cell line followed by FACS analysis (Fig. 1). THP1 was Two astrocyte cell lines were tested by FACS for C3aR expres- strongly stained even when cells were incubated with as little as sion. The well-differentiated astrocytoma cell line CB193 ex- 0.81 ␮g/ml of the Ab. Ramos was consistently negative for C3aR. pressed twice as much C3aR as the undifferentiated glioblastoma Preincubation of anti-C3aR with a 100-fold molar excess of the cell line T98G (Table I). Both lines express the astrocyte-specific C3aR peptide completely abolished staining of THP1 (not shown). marker GFAP and have been extensively used as a model of the The titer and specificity of the anti-C3aR were further tested by human astrocyte (33). Indirect IF and confocal microscopy con- ELISA (Fig. 2). The anti-C3aR gave a strong positive signal even firmed that C3aR was expressed abundantly on the membrane of at high dilutions of C3aR peptide, and preincubation with peptide the CB193 astrocyte cell line (Fig. 3, b and c). No staining was prevented binding of Ab; no binding of anti-C3aR occurred on detected in the nucleus or the cell cytoplasm. C3aR staining was plates coated with C3a or C5aR peptide (latter not shown). In all patchy, a pattern identical with that described for expression of subsequent experiments anti-C3aR was used at 0.4 ␮g/ml unless C5aR on astrocyte lines (14). The C3aR staining intensity and stated otherwise. pattern were reproducible (n Ͼ 10), specific, and blocked by the competitive peptide. To exclude the possibility that the patchy Expression of C3aR by neutrophils, monocytes, C3aR staining was an artifact of the immunochemistry procedure, PMA-differentiated THP1, and other cells we simultaneously stained CB193 for membrane-associated CD44 We have previously described the expression of C5aR by leuko- and C3aR. Figure 3, d and e, clearly shows that CD44 membrane cytes (primary cells and cell lines); the same approach was used to staining was distributed homogeneously, whereas C3aR staining characterize the expression of C3aR using FACS analysis. THP1, on the same cells was patchy and clustered. undifferentiated or differentiated with PMA (3 days), were stained Expression of C3aR was also examined on primary astrocytes for C3aR; PMA-differentiated cells had a mean fluorescence dou- cultured from fetal (n ϭ 5) and adult (n ϭ 2) brain using a classic ble that of undifferentiated cells (Table I). We have previously immunoperoxidase/DAB staining technique. Fetal astrocytes were shown that PMA differentiation increased the expression of C5aR faintly but consistently stained (data not shown), whereas adult on THP1 by a factor of 10- to 20-fold, and these results were here astrocytes were strongly stained by anti-C3aR (Fig. 4). Adult as- confirmed in parallel with the measurements of C3aR expression trocytes are adherent cells with long, ramified processes and stain 3548 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS

Table II. Effects of IFN-␥, IL1-␤, TNF-␣, and PMA on C3aR expression by monocyte cell line U937 a

FACS Analysis (mean of fluorescence)

IFN-␥ (48 h, IU/ml) IL-1␤ (48 h, IU/ml) TNF-␣ (48 h, IU/ml) PMA (48 h, ng/ml)

Non Stim. 10 100 1000 10 100 1000 10 100 1000 0.2 2 20

CD 44 (B222) 638.8 586.2 514.9 502 632.7 699.9 654.1 754.5 891.6 919.3 733.9 1387.3 1268.1 CR3 (OKM1) 59.4 64.8 78.5 93.5 63.4 66.2 65.5 68.2 63.1 51.4 77.3 537.8 601.8 C5aR (P12/1) 105.9 221.8 447.1 490.7 115.5 120 115.4 132.3 159.9 148.1 106.8 524.8 528.7 C3aR(Rb ␣-pept)b 75.77 162.2 367.9 431.2 78.1 80.1 79.2 88.9 101.6 98.4 79.4 109.5 111.5

a U937 monocyte cell line (105 cell/ml) was cultured for 48 h in a 24-well plate, in medium alone (Non Stim.) or with medium supplemented with three different concentrations of IFN-␥, IL-1␤, TNF-␣, or phorbol ester (PMA). Cells were then analyzed by FACS for CD44 (mouse clone BRIC 222), CR3 (mouse clone OKM1), and C5aR (mouse clone P12/1). A significant difference of 50% in the level of expression between control cells and stimulated cells is presented in bold. Data from one typical experiment and the results were confirmed in two additional separate experiments. b Biotin-conjugated rabbit anti-C3aR (3.75 ug/ml) was used in this experiment with RPE-conjugated streptavidin (1/200, Jackson ImmunoResearch). strongly for GFAP (Fig. 4a). The C3aR staining on adult astro- lysates (Fig. 5B), no positive results were obtained for C3aR from cytes was again patchy and confirmed our results with the CB193 cell lysates. To improve sensitivity, cell lysates were first immu- astrocyte cell line (Fig. 4d and Fig. 3, b, c, and e). Neurons and noprecipitated on Sepharose-bound anti-C3aR. Western blotting of in fetal brain cultures did not express C3aR at a immunoprecipitates revealed a protein with an apparent molecular detectable level (not shown), whereas microglia in fetal and adult mass of 65 kDa in lysates from unstimulated CB193, undifferen- brain cultures were stained for C3aR (Fig. 4b). Adult and fetal tiated THP1, and PMA-differentiated THP1 (Fig. 5C). The inten- brain cultures and CB193 were stimulated for 24 h with various sity of the 65-kDa band was greater in the PMA-stimulated THP1 cytokines (IFN-␥, IL-1␤, and TNF-␣) or with PMA. No change in lane compared with that of undifferentiated THP1. No protein was C3aR expression was detected on any cell type in the cultures (data detected when Ramos or K562 cell lysates were immunoprecipi- not shown). tated under the same conditions. When THP1/PMA cell lysate was immunoprecipitated with either Sepharose-bound polyclonal anti- C3aR on monocyte lines and astrocytes is a 65-kDa protein C5aR or monoclonal anti-C5aR (W17/1) adsorbed on protein A- Although the GST-C3aR fusion protein was readily detected by Sepharose to enrich the preparation for C5aR, no protein was de- immunoblotting (Fig. 5A) and DAF was easily detected in cell tected upon Western blotting with anti-C3aR (not shown). This

FIGURE 3. Immunolocalization of C3aR on the human astrocyte cell line CB193 by indirect IF and confocal microscopy analysis. IF was conducted on CB193 astrocytes cultured and fixed on glass coverslips and incubated with Abs against two membrane receptors, CD44 and C3aR. Cells were incubated with either a) mouse anti-CD44 clone BRIC222 (TCS; 1/200) followed by an FITC-conjugated rabbit anti-mouse Ig Ab, or b and c) affinity-purified rabbit anti-C3aR (3.75 ␮g/ml) followed by an FITC-conjugated goat anti-rabbit Ig Ab. In d and e, cells were double immunostained for CD44 and C3aR. Cells were simultaneously incubated with BRIC222 and anti-C3aR, followed by rhodamine-conjugated goat anti-mouse Ig (Sigma) and FITC-conjugated donkey anti-rabbit Ig (SeraLab). IF was analyzed by confocal microcopy. Extended focus views demonstrated that fluorescence was localized on the membrane for both Abs. CD44 staining was homogeneously distributed on the membrane (a and d); C3aR staining (b, c, and e), however, was patchy and localized to restricted areas of the cell membrane. Optical sections of the image did not show any CD44 or C3aR staining in the nucleus or the cytoplasm of CB193 (data not shown). The arrowheads mark the locations of cells in the same field either on the rhodamine channel (CD44 staining; d) or on the FITC channel (C3aR staining; e). Magnifications: a, c, d, and e, ϫ500; b, ϫ1250. The Journal of Immunology 3549

FIGURE 4. Immunolocalization of C3aR on hu- man adult astrocyte and astrocyte cell line CB193 cul- tures using peroxidase/DAB staining. Adherent cells in culture on glass coverslips (a and b, adult brain cells; c and d, CB193) were either ethanol/acetic acid permeabilized (a) or formalin fixed (b–d) and incu- bated with polyclonal Abs against GFAP (a)or against C3aR (b and d; affinity purified; 1.62 ␮g/ml). Cells were counterstained with hematoxylin (nucleus staining). a, Adult astrocytes were easily identified by their long ramified processes and strong staining for GFAP. Two large flattened cells (GFAP negative) were also present in the field (arrows) and are prob- ably fibroblast derived cells (magnification, ϫ500). b, Astrocyte (large arrow) and microglia/macrophage (small arrows) present in the adult brain culture were strongly stained for C3aR. The staining was again patchy (see inset; magnification, ϫ1250) and re- stricted to some areas of the cell membrane. C3aR staining was also strongly detected on CB193 cell membrane (d; inset is high magnification, ϫ1250). No nuclear or cytoplasmic staining of adult astrocytes or CB193 was detected using the anti-C3aR (c) CB193 negative control staining with nonimmune rabbit antiserum (IgG cut).

simple experiment eliminated the possibility that the anti-C3aR expression by U937 was slightly (IFN-␥ effect) or highly (PMA might cross-react with the C5aR. effect) up-regulated, in agreement with previous reports (34, 35). C4 mRNA expression by CB193 was up-regulated by IFN-␥ and C3aR mRNA expressed by astrocytes is 100% identical with the TNF-␣, but was not affected by the other treatments as reported leukocyte C3aR mRNA previously (36). RT-PCR analysis was used to characterize the expression of C3aR Expression of C3aR mRNA in astrocyte cell lines and a range of in isolated cells from . The cDNAs so obtained were other cell lines of myeloid and nonmyeloid lineage was compared cloned, sequenced, and compared with the published sequence. by RT-PCR (Fig. 7). C3aR mRNA was expressed by the THP1 Seven oligonucleotides were designed and used for RT-PCR anal- monocyte cell line and the ECV 304 endothelial cell line, but not ysis. The human C3aR intron/exon organization has not yet been by B and T lymphocyte cell lines, the YT NK cell line, or the reported, so we were not able to choose primers spanning an intron HepG2 hepatoma cell line (Fig. 7A). Southern blot analysis using sequence. Thus, all RNA samples were treated with Rq1 DNase to a specific C3aR cDNA probe cloned from U937 confirmed the eliminate contamination with genomic DNA, and it was confirmed identity of the cDNA and revealed that C3aR mRNA was present that no PCR fragment was obtained when the RT step was omitted, at a very low level in Molt4 T lymphocyte, YT NK cell, and when RNA was substituted with water, or when no primer was HepG2 cell lines (Fig. 7B). Multiple controls for false positive and added to the PCR mix. The different combinations of C3aR prim- false negative results were performed. Figure 7, C and D, shows ers were used in RT-PCR with total RNA extracted from control RT-PCR analysis for CR2 cDNA and GAPDH cDNA on the same U937 (Fig. 6A) and PMA-differentiated U937 (Fig. 6B). The same samples (same RT) as those used for C3aR RT-PCR. GAPDH strategy was used for RT-PCR with total RNA from the CB193 cDNA was present in all RNA samples and at the same level; CR2 astrocyte cell line and fetal astrocytes (data not shown). C3aR cDNA was detected only in Raji, Molt4, and CB193 astrocyte cell mRNA was detected in U937, particularly after PMA differentia- lines as previously described (37). The expression of CR2 mRNA tion, and also in CB193 and fetal astrocytes. We were not able to by the ECV endothelial cell line demonstrated here is a new find- obtain adult astrocytes at sufficient purity in culture for analysis of ing and merits further study. the expression of C3aR mRNA. The astrocytoma (CB193) C3aR cDNAs were sequenced and were 100% identical between posi- tions 369 and 1433 bp with the C3aR cDNA cloned from C3aR is present in inflamed brain tissues and is expressed by HL60 (24). myeloid and nonmyeloid cells Stimulation of the CB193 astrocytoma line with IFN-␥, TNF-␣, By RT-PCR analysis the expression of C3aR mRNA was detected or PMA for 24 h followed by RT-PCR analysis had no effect on in the temporal and frontal lobes, the , and the caudate C3aR mRNA expression, whereas IL-1 at 200 IU/ml decreased nucleus of normal control human brain (data not shown). To lo- expression of C3aR mRNA (Fig. 6D). calize the cellular distribution of the C3aR, anti-C3aR Ab was The activities of all cytokines and PMA treatment were checked applied to sections of normal and diseased brain. The Ab was by examining the effects on expression of C5aR by U937 (Fig. 6C) highly specific for C3aR and was positive on frozen or formalin- and of C4 by CB193 (data not shown). After 24 h, C5aR mRNA fixed tissue. In all sections from normal brain, C3aR staining was 3550 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS

FIGURE 5. Characterization of C3aR protein in control cell lysates (THP1, THP1/PMA, and K562) and astrocyte cell lysates by Western blotting and immunoprecipitation. Details of Western blot analysis (A and B) and immunoprecipitation (C) are given in Materials and Methods. A and B, Cell lysates (10 ␮l) and 100 ng of the purified GST-C3aR loop fusion protein were separated by SDS-PAGE in nonreducing conditions (15% gel) and Western blotted using the affinity-purified rabbit anti-C3aR (A) or the rabbit anti-human DAF (B), the latter to test for the quality of the cell lysates. C, SDS-PAGE and Western blot analysis of cell lysates (10 ␮l) or cell lysate immunoprecipitated (IP) with anti-C3aR Sepharose to enrich the preparation for C3aR. One milliliter of GST-C3aR loop bacteria cell lysate or lysate from CB193, THP1, and PMA-differentiated THP1 were immunoprecipitated. Eluate (50 ␮l) in glycine buffer was diluted in Laemmli buffer (nonreducing condition) before loading onto a 7.5% acrylamide gel. Western blot was conducted using the affinity-purified anti-C3aR and a peroxidase-conjugated goat anti-rabbit Ab. Using this protocol we were able to detect a band at 50 kDa present in the GST-C3aR bacteria lysate and a bands at 62–65 kDa in the cell lysate of THP1, THP1/PMA, and astrocyte cell line CB193. The band above the 175-kDa marker is the Ab released from contaminating Sepharose beads after the glycine elution. This band is detected even when PBS is used instead of cell lysate in the immunoprecipitation protocol. No band at 62 to 65 kDa was detected in Ramos and K562 cell lysate even after immunoprecipitation (data not presented). To exclude the possibility that the 62- to 65-kDa band was a degradation product of the Ab used for the immunoprecipitation, the membrane was blotted with biotinylated anti-C3aR followed by a peroxidase-conjugated avidin (Sigma) for ECL detection. Only the band at 62–65 kDa, not that at 175 kDa, was visualized in this case (data not presented). almost undetectable except for a very faint, but reproducible, stain- phocytes (data not shown). In the lumen of the vessels, erythro- ing on some astrocytes and a stronger staining on peripheral mac- cytes were always C3aR negative (see Fig. 8b, inset). Infiltrating rophages. However, sections of MS (Fig. 8, a–d) and acute BM PMN in BM were also stained for C3aR (Fig. 8, f and h). At high (Fig. 8, e–h) brain gave a strong and specific C3aR staining. Re- magnification it was noted that C3aR staining on neutrophils was active astrocytes (Fig. 8, b and e), ameboid microglia (Fig. 8c), variable in distribution, i.e., patchy on some cells and homoge- and ramified microglia (Fig. 8d) were consistently and strongly neously distributed on others (Fig. 8, f and h). Occasional perivas- stained for C3aR in both conditions. In MS, but not in BM, cells cular macrophages present in BM sections were also stained for in the brain blood vessel wall were also strongly stained for C3aR C3aR (Fig. 8g). Double IF staining of MS tissue sections (Fig. 9) (Fig. 8b), and these cells were identified as smooth muscle cells confirmed that all cells positive for C3aR were also either (pericytes). In acute MS, C3aR was expressed at high level on GFAP positive (astroglia) or HLA class II positive infiltrating perivascular macrophages, but not on infiltrating lym- (macrophage/microglia). The Journal of Immunology 3551

FIGURE 6. Expression of C3aR mRNA by the con- trol cell line (U937) and the astrocyte cell line CB193: effects of cytokines and phorbol ester (PMA). Total RNA (2 ␮g) from undifferentiated (A) and PMA-differ- entiated (B) U937 were analyzed by RT-PCR using dif- ferent combinations of C3aR primers. Lane 1, 369–1122; lane 2, 369–1391; lane 3, no primer; lane 4, 369–1433; lane 5, 521–754; lane 6, 521–1122; lane 7, 521–1391; lane 8, 521–1433; lane 9, 754–1122; lane 10, 754–1391; lane 11, 754–1433. An aliquot (20 ␮l) of each RT-PCR reaction was analyzed on a 1.2% aga- rose gel, and DNA was visualized using ethidium bro- mide. Total RNA (2 ␮g) from nonstimulated cells (Ns), 24-h stimulated cells (200 IU/ml IFN-␥, 200 IU/ml IL- 1␤, 1000 IU/ml TNF-␣) and 24-h phorbol ester-treated cells (PMA; 10 ng/ml) were analyzed by RT-PCR for C5aR (C; U937 monocyte cell line), C3aR (D; CB193 astrocyte cell line), and the housekeeping GAPDH (E; CB193). Primers (C5aR, 16–1083; C3aR, 369–1122) were selected for RT-PCR of C5aR and C3aR.

Discussion 1996 with the cloning of cDNA encoding the C3aR (24–26). C3aR is predicted to be a member of the seven-transmembrane-spanning Despite a substantial body of functional evidence indicating the receptor family, but differs from previously characterized members existence of a specific receptor on some cell types for the C3a anaphylatoxin, the receptor was only conclusively identified in of the family in that the second extracellular loop is very large (172 amino acids in total). Data concerning the protein encoded by the cDNA and its tissue distribution are sparse because of the lack of reagents for immunodetection. We here describe the generation of highly specific Abs reactive with the predicted large extracellular loop of the C3aR and the use of these reagents and molecular probes to characterize the expression and molecular mass of the C3aR protein in leukocytes, glia, and brain tissue. The presence of a C3aR on monocyte cell lines was first dem- onstrated by showing a specific and saturable binding of C3a to U937 cells (18, 34). Differentiation of U937 using PMA only slightly increased C3a binding, whereas differentiation with dibu- tyryl cAMP increased binding 7-fold. We here show, using the specific anti-C3aR Ab, that monocyte lines THP1 and U937 ex- press C3aR and that PMA differentiation caused an up-regulation of C3aR levels. Up-regulation of C3aR was detectable only after prolonged incubation with PMA; in contrast, CR3 was up-regu- lated within1hofPMAexposure. The rapid up-regulation of CR3 is a consequence of mobilization of an intracellular pool, and the slow response of C3aR suggests than no PMA-responsive pool is present. Up-regulation of C5aR by PMA followed kinetics similar to those of C3aR, but the increase was much greater, up to 10-fold. IFN-␥ in our hands was the most powerful regulator of C3aR ex- pression by two different monocyte cell lines. Even at 10 IU/ml, IFN-␥ caused a 2-fold increase in C3aR expression on U937. The effect of IFN-␥ was time and dose dependent. Our data confirm previous reports that showed that [125I]C3a binding to U937 was elevated when cells were cultured for 3 days with IFN-␥ (34). FIGURE 7. Expression of C3aR by myeloid- and nonmyeloid-derived Neutrophils isolated from blood expressed C3aR and at higher cell lines and comparison with the expression of another C3 receptor, CR2 levels than did undifferentiated THP1 or blood-derived monocytes. (CD21). Total RNA (2 ␮g) was prepared from various human cell lines: Lymphocytes were consistently negative for C3aR. We have yet to THP1 monocytes, THP1/PMA, Rajiϩ3 B lymphocytes, K562 erythroleu- examine whether C3aR can be up-regulated on monocytes and kemia cells, Molt4 T lymphocytes, YT NK cells, ECV304 endothelial neutrophils or expressed de novo by lymphocytes when cells are cells, HepG2 hepatocytes, and CB193 from three different cell culture pas- primed with cytokines or chemokines. sages. A, RT-PCR analysis for C3aR using primers (C3aR, 369–1122); B, We have previously reported the expression of C5aR on glial Southern blot of the same gel using a 32P-labeled C3aR cDNA probe; C, RT-PCR analysis for CR2 mRNA using specific primers (CR2, 181–503); cells in vitro and the remarkable correlation between expression by D, RT-PCR analysis for GAPDH. glia in vivo and brain inflammation (14–16). We have proposed 3552 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS

FIGURE 8. Expression of C3aR by infiltrating my- eloid and nonmyeloid cells in normal and diseased brain tissues. Highly purified and specific rabbit anti- C3aR (1.62 ␮g/ml) was used to immunostain rehy- drated paraffin (a, b, c, e, f, g, and h) and frozen tissue (d) sections from normal (a) and diseased human brains (MS and BM) using a classic peroxidase/DAB development protocol. Sections were counterstained with hematoxylin to display morphology. a, Normal brain: no staining was detected in the white or the gray (inset) matter when anti-C3aR was applied to rehy- drated paraffin sections of normal surgical brain. Stain- ing on frozen tissue sections revealed moderate, but reproducible, staining on ramified microglia, rare perivascular macrophages, and a few astrocytes (data not shown). b, Acute MS: C3aR-positive staining on the majority of reactive astrocytes as well as on cells in the blood vessel wall. Note that erythrocytes were all negative for C3aR. Higher magnification (inset) re- vealed that the C3aR staining in blood vessel wall was on pericytes (smooth muscle cells). c, Acute MS: a cluster of ameboid microglia rich in vacuoles was strongly stained for C3aR. d, Chronic MS: ramified microglia were C3aR positive, and here the staining was confirmed on frozen tissue sections. e and h, BM: The expression of C3aR was dramatically elevated in BM brains. Astrocytes (e and inset), microglia (not shown) and infiltrating cells, macrophages ( g), and neutrophils ( f and h) were all strongly stained for C3aR. Not all neutrophils in the meninges expressed C3aR at a high level. h, C3aR distribution on the mem- brane of an infiltrating neutrophil is clustered. No C3aR staining was detected in the blood vessel wall in BM. All photographs were taken at ϫ500, except for b, d, e, and f insets (ϫ1250).

that expression of C5aR in the inflamed brain is important in per- cytes. To further confirm that the receptor expressed by astrocytes petuation of the inflammatory response. The primary aim in gen- was identical with that cloned from leukocyte cell lines we used a erating reagents for detection of C3aR was to ascertain whether it PCR strategy to clone and sequence astrocyte C3aR cDNA. The follows a similar pattern of expression in brain cells and tissue. In majority of the astrocyte C3aR cDNA sequence (residues 369– the original descriptions of cloning, the 2.2-kb C3aR mRNA was 1433) was obtained in this manner and was 100% identical with detected by Northern analysis in human brain tissue (24–26); in the published sequence obtained from HL60 cells (24). the present report we confirm this finding. However, no data were Although all cells in astrocyte cultures expressed C3aR, the dis- available on the cell types expressing C3aR in brain. Staining with tribution on individual cells was not homogeneous; C3aR was the specific anti-C3aR Ab demonstrated that astrocyte cell lines present in dense patches on the membrane, particularly near the and primary fetal and adult astrocytes expressed C3aR in vitro, ends of cellular processes (Fig. 3). The distribution is very remi- with the more differentiated of the cell lines and adult astrocytes niscent of that reported for C5aR on astrocytes (14); double-stain- expressing more than the undifferentiated cell line and fetal astro- ing studies will be necessary to demonstrate whether these two The Journal of Immunology 3553

FIGURE 9. Double IF staining of MS tissue sections for C3aR. Frozen MS brain tissue sections were double stained for C3aR (FITC, rabbit anti-C3aR peptide) and either for astrocyte cell marker (GFAP, mouse Ab) or microglia cell marker (HLA class II, mouse LN3 clone). Specific binding of the primary Ab was detected using FITC-conjugated goat anti-rabbit Ab (SeraLab) and rhodamine-conjugated goat anti-mouse Ab (Sigma). a and b, MS sections stained for C3aR and GFAP. c and d, MS sections stained with two irrelevant Abs. e and f, MS sections stained for C3aR and HLA class II. All GFAP-positive cells expressed C3aR. HLA class II-positive cells were mainly identified as microglia and infiltrating macrophages, and they all expressed C3aR. No C3aR-positive cells were detected that did not also express markers for the astrocyte or the microglia/macrophage population.

proteins are colocalized. Several other membrane proteins stained this cell line is reported to express the C5aR (5, 13); it may thus using an identical protocol gave a homogeneous distribution pat- represent the unusual occurrence of expression of one anaphyla- tern on astrocytes, eliminating the possibility that this unusual dis- receptor without the other. To confirm specificity and in- tribution pattern of C3aR and C5aR is an artifact of the staining crease sensitivity, a combination of RT-PCR and Southern blotting procedure. We suggest that the enrichment of the receptor at spe- was used. With this protocol, C3aR message was detected at very cific areas on the cell membrane is of functional significance, per- low levels in the various lymphocyte cell lines and in HepG2 haps enhancing the capacity of the cells to respond to the anaphy- (Fig. 7). latoxins. It will be interesting to examine whether the distribution Given the above data on glial cell expression of C3aR and the of the anaphylatoxin receptors alters when cells are stimulated presence of C3aR message in brain tissue (24–26), it was expected with C3a or C5a and when cells are migrating along an anaphy- that C3aR would be detected by immunohistochemistry in brain. latoxin gradient. However, normal brain tissue was almost completely negative for Microglia derived from fetal and adult brain were also strongly C3aR expression, although the Ab readily detected C3aR in other stained for C3aR (Fig. 4b). This finding was anticipated because normal tissues (lung, liver, and adrenal gland) subjected to a sim- microglia are derived from the monocyte/macrophage lineage (38) ilar processing protocol (data not included). In marked contrast, and strongly express C5aR (16). The demonstration of C3aR ex- inflamed brain tissue (MS or meningitis) was strongly positive for pression by human microglia also supports the recent report that C3aR (Figs. 8 and 9). This finding is, again, very similar to that C3a induces calcium fluxes in cultured murine microglia (29). Oli- obtained for C5aR, which was barely detectable in normal brain godendrocytes and neurons in culture were always negative but was highly expressed in inflammation (16). The distribution of for C3aR. C3aR expression differs between the two conditions chosen for The molecular mass of expressed C3aR has not previously been study. In both, astrocytes and microglia in the areas of pathology determined. The cDNA sequence predicts a mature protein of 482 and infiltrating phagocytes (macrophages in MS, neutrophils in amino acids and a molecular mass of 54 kDa; however, the se- meningitis) express C3aR. In MS, a strong perivascular staining quence contains two putative N-glycosylation sites. Immunopre- was observed, which was not present in meningitis tissue, that cipitation of C3aR from leukocyte lines or from astrocyte lines appeared to be associated with pericytes. The functional signifi- followed by Western blotting demonstrated that the protein from both sources had a molecular mass of 65 kDa, implying that the cance of expression of C3aR by pericytes in the vessel wall is C3aR is heavily glycosylated in both cell types (Fig. 5C). Al- uncertain, but it does suggest that the anaphylatoxins might influ- though all immunoprecipitates were from whole cells, the pre- ence vessel tone or permeability in MS brain. Elevated expression dicted 54-kDa unglycosylated C3aR precursor was never detected. of anaphylatoxin receptors in CNS tissue appears to be a hallmark We are currently confirming the above data by growing cells in of inflammatory processes and is probably associated with intra- tunicamycin to inhibit glycosylation before immunoprecipitation. thecal complement activation and local generation of the anaphy- Experiments are also underway to define the molecular mass of latoxins (27, 39). The expressed receptors may then contribute to C3aR from eosinophils, neutrophils, monocytes, and primary as- an autocrine pathway, activating glia and recruiting myeloid and trocytes. The principal difficulty in these experiments is to obtain nonmyeloid cells into the brain tissue (38). This scenario is sup- sufficient numbers of cells in sufficient purity for use in the im- ported by the in vitro observation that C3a and C5a stimulate and munoprecipitation protocol. also induce of astrocytes and microglia (14, 28, 29, Expression of C3aR was confirmed for the various cell lines and 40). Monocyte/macrophage expression of cytokines, chemokines, primary cells at the mRNA level by RT-PCR. All cells positive by and other immune molecules is regulated by C3a and C5a (3, 12, IF were also positive by RT-PCR (Fig. 7). The endothelial cell line 19, 41); we are now testing whether C3a and/or C5a have similar ECV304, untested in IF, was also positive, whereas the hepatoma effects on astrocytes and microglia in culture. cell line HepG2, also untested in IF, was negative. The absence of To address the roles of the anphylatoxins and their receptors in detectable message for C3aR in HepG2 cells is interesting in that CNS inflammation it will be necessary to use models in which the 3554 C3aR EXPRESSION BY MYELOID AND NONMYELOID CELLS IN INFLAMED CNS receptors and their ligands can be manipulated. We have estab- trophils: the human C3a receptor characterised by functional responses and 125I- lished rodent models of demyelination and ischemia (both condi- C3a binding. Biochemistry 31:11274. 19. Takabayashi, T., E. Vannier, B. D. Clark, N. H. Margolis, C. A. Dinarello, tions characterized by a severe brain inflammation with comple- J. F. Burke, and J. A. Gelfand. 1996. A new biologic role for C3a and C3adesarg: ment activation and genesis of anaphylatoxins), and we plan to regulation of TNF-␣ and IL-1␤ synthesis. J. Immunol. 156:3455. block specifically the effects of both anaphylatoxins using specific 20. Daffern, P. J., P. H. Pfeifer, J. A. Ember, and T. E. Hugli. 1995. C3a is a che- motaxin for human eosinophils but not for neutrophils. I. C3a stimulation of antagonist peptides and neutralizing Abs for ligands and receptors. neutrophils is secondary to eosinophil activation. J. Exp. Med. 181:211. Mouse and rat C5aR have been cloned, and recently, mouse C3aR 21. Legler, D. F., M. Loetscher, S. A. Jones, C. A. Dahinden, M. Arock, and has been cloned from a brain cDNA library (42–44). The reagents B. Moser. 1996. Expression of high- and low-affinity receptors for C3a on the human mast cell line, HMC-1. Eur. J. Immunol. 26:753. necessary for receptor blockade in rodents are currently being 22. Nilsson, G., M. Johnell, C. H. Hammer, H. L. Tiffany, K. Nilsson, D. D. Metcalfe, generated. A. Siegbahn, and P. Murphy. 1996. C3a and C5a are chemotaxins for human mast cells and act through distinct receptors via a pertussis toxin-sensitive signal trans- duction pathway. J. Immunol. 157:1693. Acknowledgments 23. Baldo, A., A. D. Sniderman, S. St-Luce, R. K. Avramoglu, M. Maslowska, B. Hoang, J. C. Monge, A. Bell, S. Mulay, and K. Cianflone. 1993. The adipsin- We thank Dr. B. Delpech (Centre H. Becquerel, Rouen, France) for the gift acylation stimulating protein system and regulation of intracellular triglyceride of the astrocyte cell line CB193, Dr. G. Griffiths (University College, Lon- synthesis. J. Clin. Invest. 92:1543. don, London, U.K.) for the gift of the NK cell line, Dr. C. Harris (Uni- 24. Roglic, A., E. R. Prossnitz, S. L. Cavanagh, Z. Pan, A. Zou, R. D. Ye. 1996. versity of Wales College of Medicine, Cardiff, Wales) for the Rajiϩ3 cDNA cloning of a novel G protein-coupled extracellular loop structure. Biochim. Biophys. Acta 1305:39. clone, Dr. P. Chan for his excellent technical assistance, and Prof. Otto 25. Ames, R. S., Y. Li, H. M. Sarau, P. Nuthulaganti, J. J. Foley, C. Ellis, Z. Zheng, Go¨tze (University of Gottingen, Gottingen, Germany) for the generous gift K. Su, A. J. Jurewicz, R. P. Hertzberg, D. J. Bergsma, and C. Kumar. 1996. of the anti-C5aR mAbs. Molecular cloning and characterization of the human anaphylatoxin C3a receptor. J. Biol. Chem. 271:20231. 26. Crass, T., U. Faffetseder, U. Martin, M. Grove, A. Klos, J. Ko¨hl, and W. Bautsch. References 1996. Expression cloning of the human C3a anaphylatoxin receptor (C3aR) from differentiated U-937 cells. Eur. J. Immunol. 26:1944. 1. Morgan, B. P. 1990. Complement: Clinical Aspects and Relevance to Disease. 27. Morgan, B. P., and P. Gasque. 1996. Expression of complement in the brain: role Academic Press, London. in health and disease. Immunol. Today 17:461. 2. Frank, M. M., and L. F. Fries. 1991. The role of complement in inflammation and phagocytosis. Immunol. Today 12:322. 28. Armstrong, R. C., L. Harvath, and M. E. Dubois-Dalcq. 1990. Type 1 astrocytes 3. Wetsel, R. A. 1995. Structure, function and cellular expression of complement and -type 2 astrocyte glial progenitors migrate toward distinct anaphylatoxin receptors. Curr. Opin. Immunol. 7:48. molecules. J. Neurosci. Res. 27:400. 4. Morgan, E. L. 1993. Complement C5a and immunoregulation. In Complement 29. Moller, T., C. Nolte, R. Burger, A. Verkhratsky, and H. Kettenmann. 1997. 2ϩ Today. Complement Profiles Series, Vol. 1. J. M. Cruse and R. E. Lewis, eds. Mechanisms of C5a and C3a complement fragment-induced [Ca ]i signalling in Karger, Basel, p. 56. mouse microglia. J. Neurosci. 17:615. 5. Buchner, R. R., T. E. Hugli, J. A. Ember, and E. L. Morgan. 1995. Expression of 30. Armstrong, R. C., H. H. Dorn, C. V. Kufta, E. Friedman, and functional receptors for human C5a anaphylatoxin (CD88) on the human hepa- M. E. Dubois-Dalcq. 1992. Pre-oligodendrocytes from adult human CNS. J. Neu- tocellular carcinoma cell line HepG2. J. Immunol. 155:308. rosci. 12:1538. 6. Foreman, K. E., A. A. Vaporciyan, B. K. Bonish, M. L. Jones, K. J. Johnson, 31. Harris, C. L., K. S. Kan, G. T. Stevenson, and B. P. Morgan. 1997. Tumour cell M. M. Glowsky, S. M. Eddy, and P. A. Ward. 1994. C5a-induced expression of killing using chemically engineered constructs specific for tumour cells P-selectin in endothelial cells. J. Clin. Invest. 94:1147. and the complement inhibitor CD59. Immunology 107:364. 7. McCoy, R., D. L. Haviland, E. P. Molmenti, T. Ziambaras, R. A. Wetsel, and 32. Singhrao, S. K., J. W. Neal, P. Gasque, B. P. Morgan, and G. R. Newman. 1996. D. H. Perlmutter. 1995. N-formylpeptide and complement C5a receptors are ex- Role of complement in the aetiology of Pick’s disease? J. Neuropathol. Exp. pressed in liver cells and mediate hepatic acute phase gene regulation. J. Exp. Neurol. 55:578. Med. 182:207. 33. Gasque, P., P. Chan, C. Mauger, M. T. Schouft, S. Singhrao, M. P. Dierich, 8. Reed, S. L., J. A. Ember, D. S. Herdman, R. G. DiScipio, T. E. Hugli, and I. Gigli. B. P. Morgan, and M. Fontaine. 1996. Identification and characterization of com- 1995. The extracellular neutral cysteine proteinase of Entamoeba histolytica de- plement C3 receptors on human astrocytes. J. Immunol. 156:2247. grades anaphylatoxins C3a and C5a. J. Immunol. 155:266. 34. Burg, M., U. Martin, D. Bock, C. Rheinheimer, J. Kohl, W. Bautsch, and A. Klos. 9. Gerard, N. P., and Gerard, C. 1991. The chemotactic receptor for human C5a 1996. Differential regulation of the C3a and C5a receptors (CD88) by IFN-␥ and anaphylatoxin. Nature 349:614. PMA in U937 cells and related myeloblastic cell lines. J. Immunol. 157:5574. 10. Morgan, E. L., J. A. Ember, S. D. Sanderson, W. Scholz, R. Buchner, R. D. Ye, 35. Burg, M., U. Martin, C. Rheinheimer, J. Kohl, W. Bautsch, E. C. Bottger, and and T. E. Hugli. 1993. Anti-C5aR : characterization of neutralizing A. Klos. 1995. IFN-␥ upregulates the human C5a receptor (CD88) in myeloblas- antibodies specific for a peptide, C5aR (9–29), derived from the predicted amino- tic U937 cells and related cell lines. J. Immunol. 155:4419. terminal sequence of the human C5a receptor. J. Immunol. 151:377. 36. Gasque, P., A. Ischenko, J. Legoedec, C. Mauger, M. T. Schouft, and 11. Oppermann, M., U. Raedt, T. Hebell, B. Schmidt, B. Zimmermann, and O. Go¨tze. M. Fontaine. 1993. Expression of the complement classical pathway by human 1993. Probing the human receptor for C5a anaphylatoxin with site-directed an- glioma in culture. J. Biol. Chem. 268:25068. tibodies: identification of a potential ligand binding site on the NH2-terminal 37. Gasque, P., P. Chan, C. Mauger, M. T. Schouft, S. Singhrao, M. P. Dierich, domain. J. Immunol. 151:3785. B. P. Morgan, and M. Fontaine. 1996. Identification and characterization of com- 12. Ko¨hl, J., and B. Bitter-Suermann. 1993. Anaphylatoxins. In Complement in plement C3 receptors on human astrocytes. J. Immunol. 156:2247. Health and Disease. K. Whaley, M. Loos, and J. M. Weiler, eds. Kluwer Aca- 38. Shrikant, P., and E. N. Benveniste. 1996. The central nervous system as an im- demic Publishers, Hingham, MA, pp. 299–324. munocompetent organ: role of glial cells in antigen presentation. J. Immunol. 13. Haviland, D. L., R. L. McCoy, W. T. Whitehead, H. Akama, E. P. Molmenti, 157:1819. A. Brown, J. C. Haviland, W. C. Parks, D. H. Perlmutter, and R. A. Wetsel. 1995. Demonstration of C5aR on non-myeloid cells of the liver and lung. J. Immunol. 39. Ernst, J. D., K. T. Hartiala, I. M. Goldstein, and M. A. Sande. 1984. Complement 154:1861. (C5)-derived chemotactic activity accounts for accumulation of polymorphonu- 14. Gasque, P., P. Chan, M. Fontaine, A. Ischenko, M. Lamacz, O. Go¨tze, and clear leukocytes in cerebrospinal fluid of rabbits with pneumococcal meningitis. B. P. Morgan. 1995. Identification and characterization of the complement C5a Infect. Immun. 46:81. anaphylatoxin receptor on human astrocytes. J. Immunol. 155:4882. 40. Yao, J., L. Harvath, D. L. Gilbert, and C. A. Colton. 1990. Chemotaxis by a CNS 15. Fureder, W., H. Agis, M. Willheim, H. C. Bankl, U. Maier, K. Kishi, macrophage, the microglia. J. Neurosci. Res. 27:30. M. R. Muller, K. Czerwenka, T. Radazkiewicz, J. H. Butterfield, 41. Hugli, T. E. 1983. Structure and function of anaphylatoxins. Springer Semin. G. W. Klappacher, W. R. Sperr, M. Oppermann, K. Lechner, and P. Valent. 1995. Immunopharmacol. 6:173. Differential expression of complement receptors on human basophils and mast 42. Gerard, C., L. Bao, O. Orozco, M. Pearson, D. Kunz, and N. P. Gerard. 1992. cells: evidence for mast cell heterogeneity and CD88/C5aR expression on skin Structural diversity in the extracellular faces of peptidergic G-protein-coupled mast cells. J. Immunol. 155:3152. receptors: molecular cloning of the mouse C5a anaphylatoxin receptor. J. Immu- 16. Gasque, P., S. K. Singhrao, J. W. Neal, O. Gotze, and B. P. Morgan. 1997. nol. 149:2600. Expression of the receptor for complement C5a (CD88) is upregulated on reactive 43. Sayah, S., C. Patte, P. Gasque, P. Chan, A. Ischenko, H. Vaudry, and astrocytes, microglia, and endothelial cells in the inflamed human central nervous M. Fontaine. 1997. Characterization of rat C5a anaphylatoxin receptor (C5aR): system. Am. J. Pathol. 150:31. partial sequencing of rat C5aR cDNA and study of its expression in rat brain. 17. Fukuoka, Y., and T. E. Hugli. 1988. Demonstration of a specific C3a receptor on Mol. Brain Res. 48:215 guinea pig platelets. J. Immunol. 140:3496. 44. Tornetta, M. A. J. J. Foley, H. M. Sarau, and R. S. Ames. 1997. The mouse 18. Klos, A., S. Bank, C. Gietz, W. Bautsch, J. Ko¨hl, M. Burg, and T. Kretzschmar. anaphylatoxin C3a receptor: molecular cloning, genomic organization and func- 1992. C3a receptor on dibutyryl-cAMP-differentiated U937 cells and human neu- tional expression. J. Immunol. 158:5277.