Functional Expression and Characterization of Macaque C-C Chemokine Receptor 3 (CCR3) and Generation of Potent Antagonistic Anti-macaque CCR3 Monoclonal Antibodies*

in a of chronic in rhinitis. The (cid:1) -chemokine receptor C-C chemokine receptor a mechanism for the selective recruitment of eosinophils into and become an attractive biological target for therapeutic intervention. identical homologues, CCR3 in the pre-B L1-2 cell line macaque eotaxin with high affinity ( K d (cid:2) 0.1 n M a robust eotaxin-induced Ca 2 (cid:3) flux and chemotaxis. Characterization of (cid:1) -chemokines on native macaque CCR3 on eosinophils performed by means of eotaxin-induced shape change in whole blood using a novel signaling regions of the human CCR3 gene. The 5 (cid:5) -primer the sequence (cid:6) 23 (cid:6) 3 bp upstream of the ATG initiation codon and a Hin dIII site (5 (cid:5) -GGC- TTA-AGC-TTC-TAT-CAC-AGG-GAG-AAG-TG-3 (cid:5) 3 (cid:5) -primer contained the sequence 11–29 bp downstream of the TAG termination codon and a Not I site (5 (cid:5) -CTT-CAT-CTC-CTT-GCG-GCC-GCT-CCT-CTT-TAG-GCA-ATT-TTC-3 (cid:5) 30 94 60 s, 55 s, and 72 °C a PerkinElmer Life Sciences model 9600 DNA thermal cycler. PCR products from cynomolgus and rhesus DNA subcloned into expression vector pBJ-Neo and pcDEF3 J. Langer, University of Medicine and Dentistry of New Jersey), respectively. Rhesus Macaque Whole Blood Gated Autofluorescence Forward Scat- ter (GAFS) Assay— GAFS assays were performed as described (44) 3 with modifications for whole blood. Briefly, 90 (cid:3) l of fresh rhesus ma- caque blood was mixed with varying concentrations of chemokines (in 10 (cid:3) l) in 1.2-ml polypropylene tubes, incubated at 37 °C for 10 min, and then transferred onto ice. To preserve the shape change, the cells were fixed in 250 (cid:3) l of ice-cold optimized fixative solution containing 2.5% Cytofix (BD Biosciences, San Jose, CA), 22.5% H 2 O, and 75% FACSflow (BD Biosciences) and left on ice for 2 min. The mixture was then transferred into 1 ml of ice-cold lysis buffer (155 m M NH 4 Cl, 10 m M KHCO 3 ) in a Falcon tube (catalog no. 2052) and left on ice for another 20 min to achieve uniform red cell lysis. For blocking experiments, aliquots of blood were preincubated with mAbs in a total volume of 90 (cid:3) l for 20 min at room temperature before 10 (cid:3) l of macaque eotaxin (7.5 n M final concentration) was added. The samples were run on a FACScan flow cytometer (BD Biosciences), and data from 30,000 cells from each sample were collected. Eosinophils were gated out based on their high autofluorescence, and mean forward scatter was calculated by Consort 40/VAX software (Becton Dickinson, Immunocytometry Systems). Typ- ically, more than 4,500 eosinophils in each sample were analyzed. The mean forward scatter of the eosinophils in each sample was determined, and a dose-response for each chemokine or mAb was Two used the macaque

Bronchial asthma is a multifactorial disease characterized clinically by reversible bronchoconstriction leading to shortness of breath. In the pathophysiology of the disease, a chronic inflammatory condition persists in the airways of most patients, which involves a complex interplay between blood leukocytes, airway epithelial cells, and bronchial smooth muscle cells. One of the most striking aspects of asthma is the selective accumulation and activation of distinct subtypes of leukocytes into the airways, particularly eosinophils, and it is these cells that are postulated to play a key role in the pathophysiology of the disease (1,2).
Inflammatory mediators, such as chemoattractants, generated at the involved sites, promote the migration of eosinophils from the vasculature into the tissue. Unlike eicosanoids and complement cleavage fragments, which display activities on a wide variety of cells, candidate molecules for the selective recruitment of eosinophils into the airways are a class of proteins called chemotactic cytokines or chemokines. Chemokines are a growing superfamily of Ͼ50 small molecular mass proteins (ϳ8 -10 kDa) and are characterized by their actions on distinct subtypes of leukocytes (3,4). These proteins can be classified into two major subfamilies based on the arrangement of the first two conserved cysteines in the protein. In the ␣ or C-X-C family, these two cysteines are separated by any amino acid, whereas in the ␤ or C-C family, these two cysteines are adjacent to one another. Some members of this latter family have been discovered to possess strong eosinophil migratory and activating properties.
Chemokines exert their effects by binding to members of the G-protein-coupled receptor superfamily of receptors, which contain seven transmembrane domains. The discovery of a potent eosinophil-specific ␤-chemokine, eotaxin, isolated from the bronchoalveolar lavage (BAL) 1 fluid from ovalbumin-challenged guinea pigs (5), led to the identification of CCR3 (6 -8), the third ␤-chemokine receptor characterized in an expanding family that numbers close to 20 to date (3). Although its expression was first thought to be limited to eosinophils, CCR3 is now known to be more widely expressed on cells involved in allergic inflammation, such as basophils (9), mast cells (10), airway epithelial cells (11), and potentially TH 2 T-lymphocytes (12). Chemokines that are selective for CCR3 include eotaxin, eotaxin-2, and eotaxin-3, whereas the potent eosinophil-activating ␤-chemokines RANTES, MCP-3, and MCP-4 bind to other chemokine receptors in addition to CCR3 (13).
The relatively high expression of CCR3 on eosinophils, along with expression of specific CCR3-activating ␤-chemokines in the involved tissue affords a potent mechanism for the selective recruitment, activation, and retention of this cell type into tissue sites during allergic inflammation. It has been shown with a blocking monoclonal antibody directed against human CCR3 that the actions of multiple ␤-chemokines on eosinophils are mediated through CCR3 (27), indicating that CCR3 is the major ␤-chemokine receptor expressed on these cells. For these reasons, CCR3 has recently emerged as a significant antiinflammatory pharmacological target, and CCR3 antagonists are currently being developed for the treatment of asthma and other allergic disorders (28,29).
To gain a better understanding of the role of CCR3 and its ligands in this important animal model of asthma, we have cloned and functionally expressed the macaque CCR3 in the murine pre-B L1-2 cell line. We have also cloned the macaque eotaxin and employed chemically synthesized protein to functionally characterize the macaque CCR3. Moreover, we have generated murine monoclonal antibodies directed against the macaque CCR3 using two different immunizing antigens and have demonstrated that these mAbs are potent functional antagonists of this receptor through a series of receptor binding and signaling assays in vitro.

EXPERIMENTAL PROCEDURES
Southern Hybridization-Genomic DNA was isolated from venous blood of cynomolgus 2 and rhesus macaques (Macaca fascicularis and Macaca mulatta, respectively; Merck Research Laboratories, in-house colony) using the QIAamp DNA blood kit (Qiagen, Valencia, CA). Also, cynomolgus macaque DNA was purchased from Therion (Troy, NY), and rhesus macaque DNA was purchased from CLONTECH (Palo Alto, CA). Southern blot hybridization was performed using standard procedures (41). Briefly, genomic DNA (20 g) was digested with a set of restriction enzymes for 7 h, followed by phenol/chloroform extraction, ethanol precipitation, and then separation on a 0.7% agarose/Tris acetate-EDTA gel. The gel was then saturated in 1.5 M NaCl and 0.5 M NaOH (to denature the DNA) for 45 min followed by neutralization in 1.5 M NaCl and 1 M Tris-Cl (pH 8.0) for 45 min. DNA was transferred onto a Hybond-N ϩ membrane (Amersham Biosciences) using 20ϫ SSC as transfer buffer. Prehybridization was carried out at 42°C for 1 h in 6ϫ SSC, 5ϫ Denhardt's solution, 0.5% SDS, 50% formamide, and 100 g/ml salmon sperm DNA, followed by hybridization for 16 h in the identical solution containing 2 ϫ 10 6 cpm/ml 32 P-labeled 1.1-kb human CCR3 DNA fragment (Ready-to-Go DNA labeling kit; Amersham Biosciences) comprising the open reading frame (7). The membrane was washed twice in 2ϫ SSC, 0.1% SDS at room temperature for 15 min; twice in 0.1ϫ SSC, 0.1% SDS at room temperature for 15 min; once at 55°C for 15 min; and once at 65°C for 15 min. The membrane was then exposed onto Eastman Kodak Co. X-OMAT film and developed.
Cloning of Cynomolgus and Rhesus Macaque CCR3-PCR was performed using cynomolgus and rhesus macaque genomic DNA as template with primers designed from the 5Ј-and 3Ј-untranslated regions of the human CCR3 gene. The 5Ј-primer contained the sequence Ϫ23 to Ϫ3 bp upstream of the ATG initiation codon and a HindIII site (5Ј-GGC-TTA-AGC-TTC-TAT-CAC-AGG-GAG-AAG-TG-3Ј). The 3Ј-primer contained the sequence 11-29 bp downstream of the TAG termination codon and a NotI site (5Ј-CTT-CAT-CTC-CTT-GCG-GCC-GCT-CCT-CTT-TAG-GCA-ATT-TTC-3Ј). PCR was performed for 30 cycles: 94°C for 60 s, 55°C for 60 s, and 72°C for 2 min in a PerkinElmer Life Sciences model 9600 DNA thermal cycler. The resultant PCR products from the cynomolgus and rhesus DNA were subcloned into expression vector pBJ-Neo (7) and pcDEF3 (42) (a generous gift from Dr. J. Langer, University of Medicine and Dentistry of New Jersey), respectively.
Cloning of Rhesus Macaque Eotaxin-Poly(A) ϩ mRNA was isolated by oligo(dT)-cellulose chromatography from rhesus macaque small intestine and used to generate first strand cDNA using a tagged oligo(dT) primer (3Ј-RACE kit; Invitrogen). The cDNA was then used as template for PCR performed with the following two primers: 5Ј-primer (5Ј-AAC-CAC-CAC-CTC-TCA-CGC-C-3Ј) and 3Ј-primer (5Ј-CAG-GCT-CTG-GTT-TGG-TTT-CAA-3Ј). PCR was performed for 30 cycles: 94°C for  30 s, 55°C for 30 s, and 72°C for 1 min. The resultant PCR product was subcloned into pCR2.1-TOPO (Invitrogen). A consensus sequence for rhesus macaque eotaxin was identified by DNA sequencing of six independent clones. The rhesus macaque eotaxin protein was chemically synthesized (Gryphon Sciences, South San Francisco, CA) based on the predicted amino acid sequence.
Expression of Macaque CCR3 in the Murine Pre-B Cell Line, L1-2, and the Human Eosinophilic Cell Line, AML14.3D10 -The murine pre-B L1-2 cell line was a generous gift from Dr. I. Weissman (Stanford University). Transfection of the L1-2 and AML14.3D10 cell lines were carried out as described (7,8) with slight modifications. Briefly, 5 ϫ 10 6 cells were washed in Hanks' balanced saline solution, mixed with 20 g of pcDEF3-rhesus CCR3 or pcDEF3-human CCR3 (for L1-2 transfection), and pBJ/Neo-cynomolgus CCR3 (for AML14.3D10 transfection) in a 0.4-cm electroporation cuvette, electroporated at 250 V and 960 microfarads, and then cultured in complete medium (RPMI 1640 with 10% fetal bovine serum). After 48 h, the cells were placed in medium containing 0.8 mg/ml G418 and plated onto 96-well plates at 18,000 cells/ well. Clones were transferred into six-well plates, and positive clones were selected by their ability to bind 125 I-human eotaxin (Amersham Biosciences).
Ligand-induced Ca 2ϩ Mobilization-Changes in intracellular calcium level were measured on a fluorescence imaging plate reader (FLIPR; excitation 488-nm argon laser line/emission 530-nm bandpass interference filter; Molecular Devices, Inc., Sunnyvale, CA) following the manufacturer's instructions with modifications. L1-2/macaque CCR3 or L1-2/human CCR3 cells were washed in wash buffer (Hanks' balanced saline solution containing 20 mM Hepes, pH 7.2, 0.1% bovine serum albumin) and labeled with fluo-3 (Molecular Probes, Inc., Eugene, OR) at 1.5 ϫ 10 6 cells/ml in labeling buffer at 37°C for 1 h. Labeling buffer was prepared as follows. 50 g of fluo-3 was dissolved in 44 l of 10% pluronic F-127 in Me 2 SO and then added to 11 ml of wash buffer. Following labeling with fluo-3, cells were washed twice in wash buffer. About 1.5 ϫ 10 5 cells in 135 l of wash buffer were added to each well of a 96-well plate (black, clear bottom; Corning Costar, Cambridge, MA) and then centrifuged for 5 min (no brake). The plate was placed on FLIPR, 67.5-l ligands (one-half of the cell volume) at various concentrations were added, and fluorescence changes were recorded. For blocking experiments, mAbs were preincubated with cells for 20 min before 3 nM macaque eotaxin was added. Results were calculated using KaleidaGraph.
Ligand-induced Chemotaxis-L1-2/macaque CCR3 cells were labeled with 5 M calcein (Molecular Probes), washed, and resuspended in chemotaxis buffer (RPMI 1640 plus 0.5% bovine serum albumin) at 5 ϫ 10 6 cells/ml. To the bottom chamber of a 96-well Neuroprobe ChemoTx plate (5-m pore size; NeuroProbe, Inc., Gaithersburg, MD), increasing concentrations of chemokine were added in a volume of 29 l of chemotaxis buffer. To the top chamber, 30 l of cells (1.5 ϫ 10 5 total cells) were added, and chemotaxis was allowed to proceed at 37°C for 1 h. Unmigrated cells were removed from membrane with a Kimwipe. The plate was then analyzed in a CytoFluor II fluorometer (excitation, 485 nm; emission, 530 nm; PerSeptive Biosystems, Framingham, MA). For blocking experiments, 5 nM macaque eotaxin was added to the bottom chamber. To the top chamber, 15 l of cells (1.5 ϫ 10 5 total) was mixed together with 15 l of mAb at various concentrations.
Rhesus Macaque Whole Blood Gated Autofluorescence Forward Scatter (GAFS) Assay-GAFS assays were performed as described (44) 3 with modifications for whole blood. Briefly, 90 l of fresh rhesus macaque blood was mixed with varying concentrations of chemokines (in 10 l) in 1.2-ml polypropylene tubes, incubated at 37°C for 10 min, and then transferred onto ice. To preserve the shape change, the cells were fixed in 250 l of ice-cold optimized fixative solution containing 2.5% Cytofix (BD Biosciences, San Jose, CA), 22.5% H 2 O, and 75% FACSflow (BD Biosciences) and left on ice for 2 min. The mixture was then transferred into 1 ml of ice-cold lysis buffer (155 mM NH 4 Cl, 10 mM KHCO 3 ) in a Falcon tube (catalog no. 2052) and left on ice for another 20 min to achieve uniform red cell lysis. For blocking experiments, aliquots of blood were preincubated with mAbs in a total volume of 90 l for 20 min at room temperature before 10 l of macaque eotaxin (7.5 nM final concentration) was added. The samples were run on a FACScan flow cytometer (BD Biosciences), and data from 30,000 cells from each sample were collected. Eosinophils were gated out based on their high autofluorescence, and mean forward scatter was calculated by Consort 40/VAX software (Becton Dickinson, Immunocytometry Systems). Typically, more than 4,500 eosinophils in each sample were analyzed. The mean forward scatter of the eosinophils in each sample was determined, and a dose-response curve for each chemokine or mAb was generated accordingly.
Generation of Mouse Anti-macaque CCR3 mAb-Two immunizing antigens were used to generate monoclonal antibodies against the macaque CCR3. The first approach used viable whole L1-2 cells stably expressing the full-length macaque CCR3 sequence as immunogen, essentially as described (27) with some modifications. About 1.5 ϫ 10 7 L1-2/rhesus macaque CCR3 cells were injected into C57b/6 mice. Intraperitoneal injections were performed four times at 2-week intervals (Cell Essentials, Boston, MA). Postimmune mouse sera were tested by flow cytometry and inhibition of 125 I-human eotaxin binding on AML14.3D10-cynomolgus macaque CCR3 transfectants versus untransfected AML14.3D10 cells. The most potent serum blocked 125 Ihuman eotaxin binding by 50% at 1:1000 serum dilution and blocked this binding completely at 1:125 dilution (data not shown). The mice whose sera were most potent in flow cytometry and inhibition of 125 Ihuman eotaxin binding were chosen for fusion. The final injection of whole L1-2/rhesus macaque CCR3 cells was performed intravenously. Three days postinjection, mouse splenocytes were fused with SP2/0 myeloma cells. Approximately 1600 hybridoma clones were obtained from two fusions (Cell Essentials). To eliminate any clones that reacted to native L1-2 cell surface antigens, screening of hybridoma supernatants was performed by flow cytometry on AML14.3D10/cynomolgus macaque CCR3 cells. In the second approach, a peptide (30-mer) was synthesized (Gryphon Sciences), corresponding to the predicted NH 2 terminus of the macaque CCR3 amino acid sequence (TTS-LDT-VET-FGP-TSY-DDD-MGL-LCE-KAD-VGA-norleucinal-amide). The peptide was conjugated to thyroglobulin and injected into BALB/c mice at 3-week intervals. Mouse sera were tested by flow cytometry on L1-2/ macaque CCR3 cells versus L1-2 untransfected cells and inhibition of 125 I-human eotaxin binding to L1-2/macaque CCR3 cells. Postimmune sera possessed similar potency as those using whole L1-2/macaque CCR3 cells as immunogen. As above, the mouse that produced the most potent serum was used for fusion. A total of about 600 hybridoma clones were obtained by this method from a single fusion. Positive hybridoma clones were selected by enzyme-linked immunosorbent assay (see below) and tested by flow cytometry on L1-2/macaque CCR3 cells. Positive clones identified from both approaches tested positive on L1-2/macaque CCR3 and negative on the parental L1-2 cell line by flow cytometry. Hybridoma clones producing anti-macaque CCR3 mAbs were subcloned to ensure purity and then injected into mice for ascites production. The IgG fraction from the ascites fluid was purified by protein A chromatography and dialyzed against PBS. Antibody concentration was determined by the Bradford protein assay kit (Bio-Rad). The isotype of the anti-macaque CCR3 mAbs was determined by enzyme-linked immunosorbent assay.
Enzyme-linked Immunosorbent Assay-Bovine serum albumin-conjugated macaque CCR3 NH 2 -terminal peptide (4 g/ml in PBS) was loaded onto 96-well plates (100 l/well) and incubated at 4°C for 18 h. The plates were then washed three times with PBS. Blocking solution (3% fish gel in PBS, 300 l/well) was added, and the plates were incubated at room temperature for 2 h. The plates were then washed, and hybridoma supernatant was added to each well and incubated for 1 h. The supernatant was removed, and the plates were washed in PBS containing 0.05% (v/v) Nonidet P-40. Horseradish peroxidase enzymeconjugated secondary antibody was added to each well, and the plates were incubated for 30 min. The plates were washed again in PBS containing 0.05% Nonidet P-40. ABTS-100 substrate solution was then added and incubated for 30 min for color development. Clones that gave a signal/noise ratio of greater than 3 were picked for further analysis.
Partial Purification of Rhesus Macaque Leukocyte Subtypes-Rhesus macaque eosinophils were prepared by the following method. In a 50-ml conical tube, 7 ml of NIM2a and 7 ml of NIM2b solutions (Cardinal Associates, Santa Fe, NM) were prewarmed to room temperature and carefully layered sequentially. Ten ml of rhesus macaque whole blood was slowly layered on top and centrifuged at 1000 ϫ g in a swinging bucket rotor at room temperature for 40 min (brake off). The top portion containing plasma and NIM solution was removed, and the granulocyte cell layer was then transferred to another tube. The granulocyte layer was washed once in wash buffer (Hanks' balanced saline solution containing 0.1% bovine serum albumin) and incubated in lysis buffer (155 mM NH 4 Cl, 10 mM KHCO 3 ) for 15 min to lyse the red blood cells. The remaining cells were then washed twice in wash buffer. The final granulocyte cell preparation contained ϳ25% eosinophils by differential staining with the remaining cells mostly consisting of neutrophils. For the preparation of rhesus macaque PBMCs, the following method was employed. In a 50-ml conical tube, 6.8 ml of Lympholyte-Mammal solution (Cedarlane Laboratories, Hornby, Ontario, Canada) was added. Subsequently, 9 ml of diluted blood (1:1 with PBS) was added and centrifuged at room temperature for 20 min. The PBMCs were carefully removed from the interface with a Pasteur pipette, diluted with PBS, and centrifuged as above. The PBMCs were then washed twice in PBS. The final rhesus macaque PBMC preparation contained ϳ52% T-lymphocytes by anti-CD3 staining followed by flow cytometry.

Cloning of CCR3 from Macaque
To determine the complexity of the CCR3 genes from different non-human primate species, a Southern blot analysis was performed on both cynomolgus and rhesus macaque genomic DNAs. Southern blots hybridized with the 32 P-labeled human CCR3 cDNA probe surprisingly revealed that the hybridization banding pattern was identical for both of these macaque species when the genomic DNAs were digested with BamHI, BglII, EcoRI, and HindIII (Fig. 1). Although this banding pattern was different from that of human genomic DNA (8), 4 a simple hybridization pattern was observed, indicating that CCR3 is encoded by a single copy gene in both the cynomolgus and rhesus macaque.
Given that human CCR3 lacked intervening sequences within the coding region (45), it was postulated that the coding region of the rhesus and cynomolgus macaque CCR3 genes would be intronless as well. Using PCR based on primers designed from the 5Ј-and 3Ј-untranslated regions of the human CCR3 gene, we independently cloned the rhesus and cynomolgus macaque CCR3 genes from genomic DNA, and the sequences we obtained were genetically distinct from those pre-viously reported (46 -48). The sequences of the cynomolgus and rhesus CCR3 genes contained open reading frames of 1065 nucleotides in length, which encoded G-protein-coupled receptor proteins of 355 amino acids in length. Genomic DNA from six cynomolgus monkeys were used separately as templates in PCR to amplify the CCR3 gene. For each individual, the PCR products from at least four separate reactions were pooled for cloning, and six clones were sequenced to obtain a consensus sequence. Analysis of the cynomolgus macaque CCR3 sequences revealed a nucleotide polymorphism of either G or A at the first position of codon 250, encoding either a valine or isoleucine at that position, respectively (one animal was heterozygous, containing both A and G at that position, whereas four animals contained only an A and one animal contained only a G). To clone the rhesus macaque CCR3, genomic DNA from three rhesus monkeys were used as templates in PCR. As with cloning the cynomolgus CCR3, four PCR products from each of the rhesus template genomic DNAs were pooled and subcloned, and six clones were sequenced for each. Consensus sequences of clones obtained from these three animals contained only an A at the first position of codon 250, encoding only isoleucine. All other nucleotide polymorphisms in either the cynomolgus or rhesus macaque CCR3 sequences were silent (see GenBank TM entries). Throughout this report, we will refer to the sequence containing isoleucine at position 250 as that of macaque CCR3. Despite the fact that multiple clones were sequenced from at least two sources each of rhesus macaque and cynomolgus macaque DNAs, we were unable to confirm the amino acid changes previously reported for rhesus macaque CCR3 (Lys 176 (46) (48)). These differences probably reflect polymorphisms in the various rhesus and cynomolgus macaque CCR3 genes present in these colonies.
The predicted amino acid sequence of macaque CCR3 is shown as a serpentine diagram (Fig. 2A). The macaque CCR3 sequence is 92 and 94% identical to the human CCR3 sequence at the amino acid and nucleotide level, respectively. Specifically, there are 28 amino acid differences between the human and macaque CCR3 proteins. Interestingly, of the 17 noncon-servative changes between human and macaque CCR3, 13 of these residues are located in the extracellular domains of the molecule and evenly distributed along the NH 2 terminus and the first, second, and third extracellular loops ( Fig. 2A), regions important for chemokine binding and activation of the receptor (49). Fig. 2B represents an amino acid lineup of the macaque CCR3 to all of the known CCR3 sequences to date, which include human, African green monkey (96% identical at the amino acid level), sheep (72% identity), guinea pig (68% identity), rat (67% identity), and mouse (69% identity for both SV/129 and Balb/c strains). Some important characteristics of the macaque CCR3 sequence are worth noting. Like human CCR3, macaque CCR3 appears to lack any triplet amino acid consensus sites for N-linked glycosylation (N-X-S/T; where X is any amino acid), motifs that are present in nearly all ␣and ␤-chemokine receptors to date. The four cysteine residues postulated to form two disulfide bonds, one in each of the four extracellular domains, are located at positions 24, 106, 183, and 273. The macaque CCR3 also contains the amino acid motif, DRYLAIVHA, distal to TMIII in intracellular loop 2 (i2), which is strikingly conserved in all species of CCR3 sequences to date (Fig. 2B) and is critical for G-protein coupling and signal transduction upon ligand binding to G-protein-coupled receptors (50). The cytoplasmic tail of macaque CCR3 also contains eight serine/threonine residues at positions 333, 339, 340, 341, 343, 345, 346, and 352. Serine-and threonine-containing cytoplasmic tails are a common feature of chemokine receptors that may become phosphorylated by G-protein coupled receptor kinases, resulting in desensitization (51).

Cloning of Eotaxin from Macaque
Since it has been previously shown that eotaxin is highly expressed in the gastrointestinal tract in humans (52), the rhesus macaque small intestine was chosen as a source of mRNA to clone the eotaxin cDNA by reverse transcriptase-PCR from this species. The two-loop structure of the predicted amino acid sequence of rhesus macaque eotaxin is shown (Fig. 3A) (hereafter referred to as macaque eotaxin throughout this report). The macaque eotaxin is 88 and 94% identical to the human eotaxin sequence at the amino acid and nucleotide level, respectively. As shown, there are only seven amino acid differences between the mature human and macaque eotaxins (six nonconservative, one conservative), with six of the differences occurring within the NH 2 terminus-proximal 30 amino acids of the protein. Fig. 3B represents an amino acid lineup of the macaque eotaxin to all of the known eotaxin sequences to date, which include human, horse (66% identical at the amino acid level), bovine (59% identity), guinea pig (58% identity), rat (59% identity), and mouse (56% identity).

Expression of Macaque CCR3 in Murine Pre-B L1-2 Cells
Both the human and macaque CCR3 were subcloned into expression vector pcDEF3 (42), which uses the human elongation factor 1a promoter and has been proven to drive high level expression of proteins in heterologous cell lines. Given that the murine pre-B L1-2 cell line was successfully used for the heterologous expression of human CCR3 (8), we also chose to use this cell line for functional expression of macaque CCR3. Positive clones were picked by their ability to bind 125 I-human eotaxin, and one, designated clone 19, was chosen for further analysis.

Competition Binding of Various ␤-Chemokines on L1-2/Macaque CCR3 Cells
Competition binding studies were performed with 125 I-macaque eotaxin on clone 19 in order to characterize the pharma-  Table I, unlabeled macaque eotaxin competed with a K d of 0.1 nM. Scatchard analysis demonstrated that macaque eotaxin bound with a single affinity and that clone 19 expressed 3 ϫ 10 5 receptors/cell (Fig. 4A, inset). These results were similar to those obtained when human CCR3 was expressed in AML14.3D10 cells when 125 I-human eotaxin was used as the radiolabeled ligand (7). The rank order of potency of the various ␤-chemokines on macaque CCR3 upon competition with 125 I-macaque eotaxin is macaque eotaxin Ͼ murine eotaxin ϳ human eotaxin Ͼ human MCP-4 Ͼ human eotaxin-2 Ͼ human eotaxin-3 Ͼ human MCP-3 Ͼ human RANTES (Fig. 4B and Table I).
The pharmacological properties of macaque CCR3 were compared with that of human CCR3 by competition binding of the various ␤-chemokines against 125 I-human eotaxin on recombinant L1-2 cells expressing these receptors (Table I). Most of the ligands have comparable binding affinity on the human and macaque CCR3, with the exception of human eotaxin-3 and RANTES. Human eotaxin-3 has high affinity on human CCR3 (K d ϭ 0.9 nM) yet binds to macaque CCR3 with Ͼ5-fold lower affinity (K d ϭ 5.2 nM). Likewise, human RANTES binds human CCR3 with 6-fold higher affinity (K d ϭ 1.5 nM) than macaque CCR3 (K d ϭ 9.2 nM).

Functional Coupling of Macaque CCR3 in L1-2 Cells
Ligand-induced Ca 2ϩ Mobilization-To investigate the ability of ␤-chemokines to activate signal transduction in L1-2/ macaque CCR3 cells, agonist-induced calcium mobilization was measured (Table II). The amplitude at which intracellular calcium reaches its highest level in L1-2 cells occurs within 10 s of ligand addition. The rank order of potency obtained by binding affinity via competition of ␤-chemokines on macaque CCR3 with 125 I-macaque eotaxin was comparable with the agonist potency obtained by calcium mobilization (see Tables I and II); however, the functional EC 50 values were ϳ10-fold lower than the binding K d values for most of the chemokine ligands. Interestingly, macaque eotaxin was shown to be equipotent as an agonist on the human and macaque CCR3. Furthermore, human RANTES was unable to trigger a calcium flux even at a concentration of 100 nM.
Ligand-induced Chemotaxis-Since heterologous chemokine receptors are functionally coupled to the chemotaxis pathway in murine pre-B L1-2 cells, chemotaxis assays were carried out with various ␤-chemokines on L1-2/macaque CCR3 cells. The rank order of potency of the various ␤-chemokine ligands was as follows: macaque eotaxin ϳ human eotaxin Ͼ human MCP-4 Ͼ human eotaxin-2 ϳ human MCP-3 Ͼ human eotaxin-3 Ͼ human RANTES (Fig. 5, b and d; Table III). Similar to human CCR3, no response was observed with human MIP-1␣. This rank order of potency was generally consistent with radiolabeled binding and calcium flux data, except for human MCP-3, which induces chemotaxis with a higher relative potency than it does in agonist-induced calcium flux. Unlike the other chemotactic ␤-chemokines that are active on L1-2/macaque CCR3, human eotaxin-2 is only a partial agonist, exhibited by a chemotactic index of less than 50% of the amplitude observed with the other eotaxins.

Activation of Native Macaque CCR3: Ligand-induced Shape Change of Rhesus Macaque Eosinophils
Activation of native macaque CCR3 was assessed by ␤-chemokine-induced shape change of rhesus macaque eosinophils in whole blood by GAFS assay (44). Eosinophils, like all leukocytes, undergo cytoskeletal rearrangement (via actin polymerization) upon chemokine receptor activation, resulting in a change of cellular morphology. These shape changes can be readily monitored via increases in forward scatter by flow cytometry. Moreover, eosinophils are highly autofluorescent and therefore can be gated out in whole blood using flow cytometry without separation of the different leukocytes. For the GAFS assay, blood from a rhesus macaque that contained a high circulating level of eosinophils (Ͼ15%) was used. The rank order of potency for the various ␤-chemokine ligands in the rhesus eosinophil GAFS assay was as follows: macaque eotaxin Ͼ human eotaxin ϳ human MCP-4 Ͼ human MCP-3 Ͼ human eotaxin-2 Ͼ human eotaxin-3 Ͼ human RANTES (Fig.  5, a and c; Table III). This rank order was very similar, but not identical, to that from the chemotaxis assay. Most notably, macaque eotaxin was more potent than human eotaxin, and human eotaxin-2 did not appear to be acting as a partial agonist in the GAFS assay. Unlike human eosinophils, which express CCR1 (7,44), the rhesus macaque eosinophils did not respond to human MIP-1␣ in the GAFS assay, most likely reflecting the absence of CCR1 on these cells or species differences for activity of human MIP-1␣ between human and macaque CCR1.

Identification of Monoclonal Antibodies against Macaque CCR3
Two approaches were taken to produce mouse mAbs against macaque CCR3 (see "Experimental Procedures"). The first approach used the L1-2/macaque CCR3 cell line as the immunizing antigen, and one clone, designated mAb 5B9 (IgG2a/), was identified by flow cytometry (Fig. 6a). The second approach used a 30-amino acid NH 2 -terminal macaque CCR3 synthetic peptide as the immunizing antigen, and six positive clones were identified by enzyme-linked immunosorbent assay. Only two of these clones, designated mAb 51 (IgG1/) and mAb 52 (IgG2b/), were positive by flow cytometry on L1-2/macaque CCR3 cells (Fig. 6, b and c). All three of these anti-macaque CCR3 mAbs were negative on the parental L1-2 cell line by flow cytometry (Fig. 6, a-c), indicating specificity for macaque CCR3. Only mAb 52 exhibited some specificity for human CCR3 by flow cytometry, whereas none of these three antimacaque CCR3 mAbs were found to be positive for murine CCR3 (data not shown).

Anti-macaque CCR3 mAbs Recognize Native CCR3 on Rhesus Macaque Eosinophils
The anti-macaque CCR3 mAbs were assessed by flow cytometry on primary rhesus eosinophils to determine binding to native CCR3. Partial purification of the rhesus eosinophils yielded a final mixture that contained 24% eosinophils, 1% lymphocytes, and 75% neutrophils by differential staining. Of the three anti-macaque CCR3 mAbs, only two of these (mAb 5B9 and mAb 51) bound to a subset of the rhesus leukocyte cell suspension by flow cytometry (ϳ22%), the exact proportion by differential staining representative of eosinophils (Fig. 6, e and f), whereas mAb 52 did not react with this subset at all (Fig.  6g). An anti-integrin mAb (anti-CD11b) was used as a positive control, which stained the entire population of cells in the rhesus leukocyte mixture (Fig. 6h). In contrast to data observed with human T-lymphocytes (12), rhesus macaque T-lymphocytes were not shown to express CCR3 via double staining the rhesus macaque PBMC mixture with anti-CD3 and anti-macaque CCR3 (5B9) mAbs, followed by flow cytometry (data not shown).

Anti-macaque CCR3 mAbs Block Eotaxin Binding to L1-2/Macaque CCR3 Cells
The anti-macaque CCR3 mAbs were next tested for inhibition of eotaxin binding. As shown in Fig. 7a, mAbs 5B9 and 51 exhibited highly potent inhibition of 125 I-human eotaxin bind-   CCR3 cells (b and d). Increase in eosinophil shape change in rhesus macaque whole blood was determined by GAFS as described under "Experimental Procedures." Chemotactic index was determined by the fluorescence reading ob- ing to L1-2/macaque CCR3 cells with IC 50 values of 0.027 and 0.051 g/ml, respectively. Also shown in Fig. 7a, mAb 52 or mouse IgG control did not demonstrate any inhibition of eotaxin binding at concentrations up to 100 g/ml. None of the three anti-macaque CCR3 mAbs inhibited 125 I-human eotaxin binding to L1-2/human CCR3 (data not shown).

Anti-macaque CCR3 mAbs Display Potent Functional Antagonism
The anti-macaque CCR3 mAbs were analyzed in a series of functional assays to determine antagonist activity. As shown in Fig. 7b, pretreatment of L1-2/macaque CCR3 cells with mAbs 5B9 and 51 inhibited macaque eotaxin-induced calcium flux with IC 50 values of 0.5 and 1.6 g/ml, respectively. Additionally, none of the anti-macaque CCR3 mAbs blocked human eotaxin-induced calcium flux in L1-2/human CCR3 cells (data not shown). Moreover, the mAbs were also evaluated in both the rhesus macaque whole blood GAFS assay and in chemotaxis on L1-2/macaque CCR3. As shown in Fig. 7c, mAbs 5B9 and 51 antagonized macaque eotaxin-induced shape change of rhesus eosinophils with IC 50 values of 0.088 g/ml and 2.1 g/ml, respectively. Finally, mAbs 5B9 and 51 blocked macaque eotaxin-induced chemotaxis of the L1-2/macaque CCR3 cell line with IC 50 values of 0.05 and 0.3 g/ml, respectively (Fig. 7d). These two anti-macaque CCR3 antagonist mAbs also inhibited human MCP-4-induced chemotaxis of L1-2/macaque CCR3 cells with a potency comparable with that of macaque eotaxin (data not shown). Consistent with the binding studies, mAb 52 or control mouse IgG did not exhibit any inhibition in either the calcium flux assay (at concentrations up to 1 mg/ml) or the GAFS and chemotaxis assays (at concentrations up to 100 g/ml for both) (data not shown).

DISCUSSION
In this paper, we describe the molecular cloning and functional expression of the macaque CCR3 ␤-chemokine receptor. Previous studies investigating macaque CCR3 have principally focused on AIDS research exploring the mechanisms of human immunodeficiency virus type 1 and 2 co-receptor usage (46) and AIDS-associated neuropathogenesis in the macaque (53)(54)(55). In addition, two sequence notes have previously been reported (47,48). This is the first report describing full pharmacological characterization of macaque CCR3 with a variety of ␤-chemokine ligands in a series of radiolabeled binding and signal transduction assays. The macaque CCR3 sequences described in this report are genetically distinct from those previously reported, most likely reflecting genetic polymorphisms in the macaque CCR3 gene. We have also cloned the macaque ␤-chemokine eotaxin and were able to employ chemically synthesized protein in the biochemical and functional assays described in this report.
In addition, we describe the generation of monoclonal antibodies directed against the macaque CCR3 and have demonstrated that these mAbs are functional antagonists in a series of in vitro assays. The in vitro potency of the anti-macaque CCR3 mAbs reported here compares favorably with other antagonistic murine anti-CCR3 mAbs raised against CCR3 from other species, which include human CCR3 (mAb 7B11) (27) and guinea pig CCR3 (mAb 2A8) (56). In addition, in vivo studies demonstrated that the anti-guinea pig CCR3 mAb 2A8 was able to block eotaxin-induced eosinophil recruitment into the guinea pig skin. Furthermore, a rat mAb was generated against murine CCR3 (57). Although this antibody was nonneutralizing, it was very effective at depleting eosinophils in the blood, lung, and BAL fluid in a Nippostrongylus brasiliensis-infected model of eosinophil accumulation in mice. This is the first report describing the development of blocking mAbs specifically raised against a macaque chemokine receptor.
The anti-macaque CCR3 mAbs were generated by two different immunizing antigens, intact macaque CCR3-transfected cells, and a macaque CCR3 NH 2 -terminal synthetic peptide of 30 amino acids in length. The method utilizing heterologous stably transfected cell lines as immunogens has proven most successful at producing antagonist mAbs against chemokine receptors (27, 56, 58 -65). We were able to produce two antagonist mAbs against macaque CCR3 (5B9 and 51), one from each of the two methods; however, the more potent mAb was the one employing intact macaque CCR3-transfected L1-2 cells as the immunogen (mAb 5B9). This mAb (5B9) was 2-fold more potent in blocking radiolabeled eotaxin binding and, in functional assays, 3-, 24-, and 6-fold more potent in eotaxin-induced calcium flux, shape change, and chemotaxis, respectively, than mAb 51 (see Fig. 7).
Antagonistic monoclonal antibodies raised against chemokine receptors have been invaluable tools for deciphering ligand binding domains on these receptors. These mAbs essentially fall into two main categories, those that bind to the NH 2 -terminal region and those that recognize the second extracellular loop, although some mAbs generated against human CCR5 have been mapped to multidomain conformational epitopes (63). Besides the anti-macaque CCR3 mAb 51 described in this report, neutralizing mAbs that have been shown to bind to the NH 2 terminus of chemokine receptors include those that were raised against human CXCR1 and CXCR2 (58,59,61), human CXCR3 (66), human CCR4 (67), and human CCR5 (63). Neutralizing mAbs that recognize the second extracellular loop of chemokine receptors include human CCR2 (68), CCR5 (62,63), and CXCR4 (69). Although beyond the scope of this report, it would be interesting to compare the neutralizing epitopes on CCR3 of the anti-macaque CCR3 mAb 5B9 as described here with that of the anti-human CCR3 mAb 7B11 (27), given the close homology between these two sequences and that both of these mAbs were generated with identical strategies using intact L1-2 transfectants as immunogens in mice.
Whereas we were able to generate a total of three mAbs that reacted with recombinant macaque CCR3 expressed on either L1-2 or AML14.3D10 cells, only two of these were able to recognize native CCR3 on primary rhesus macaque eosinophils by flow cytometry (mAbs 5B9 and 51; see Fig. 6). The inability of mAb 52 to recognize macaque CCR3 on the rhesus eosinophils has several plausible explanations. The epitope recognized by this mAb is masked by physical association with some other molecular entity on the cell surface (i.e. endogenous surface proteins, sugar moieties, etc.). Alternatively, this antigenic determinant on macaque CCR3 expressed on recombinant cell lines is in a conformation that is distinct from that on native rhesus eosinophils, resulting in the lack of reactivity with mAb 52. In support of this concept, the principle human immunodeficiency virus type 1 co-receptors CCR5 and CXCR4 have been shown to exist in multiple conformational states on the cell surface using a panel of specific neutralizing mAbs (63,70,71). Interestingly, mAb 52 was the only one of the three anti-macaque CCR3 mAbs that recognized recombinant human CCR3.
There have been numerous reports describing the functionality of the ␤-chemokine RANTES as a very potent activator of human eosinophils. These biological effects include calcium mobilization, chemotaxis, integrin up-regulation, degranulation, and transendothelial migration in vitro (72)(73)(74)(75)(76)(77). In addition, intradermal injection of human RANTES into allergic human subjects was shown to cause a rich eosinophil recruitment into the skin in vivo (78). In another human study, intranasal administration of RANTES into allergic rhinitis patients resulted in an inflammatory infiltrate into the nasal mucosa that was rich in eosinophils (79). Furthermore, we and others have demonstrated that this ␤-chemokine binds to CCR3 on human eosinophils and stimulates signal transduction through CCR3 with high potency (7,8,27). In contrast, human RANTES competed with low potency against radiolabeled macaque eotaxin (K d ϭ 80 nM, Table I; Fig. 4) and was ineffective at stimulating a calcium flux in macaque CCR3transfected L1-2 cells (Table II). RANTES was able only at very high concentrations to trigger a shape change in rhesus macaque eosinophils and chemotaxis in L1-2/macaque CCR3 cells (Table III; Fig. 5). Additional evidence for the lack of human RANTES to activate macaque CCR3 was attained from studies in which human RANTES was injected intradermally into cynomolgus macaques in vivo. In these studies, a complete absence of eosinophil recruitment was observed, whereas the other human CCR3-active ␤-chemokines eotaxin, MCP-4, and MCP-3 stimulated a robust eosinophil accumulation into the skin of this species (80). Taken together, these results would indicate that human RANTES is not an efficacious ligand for macaque CCR3 due to species differences. It is unknown whether native macaque RANTES is a functional agonist for macaque CCR3.
We have been able to assess functionality of native macaque CCR3 on rhesus eosinophils through a novel signaling assay known as GAFS (44). This assay proved to be straightforward, efficient, and extremely rapid. The GAFS assay is based on the fact that leukocytes undergo a rapid shape change through actin polymerization upon activation by chemokine receptor agonists, and this signal transduction pathway is necessary for directional migration. This increase in shape change can readily be measured by a shift in forward scatter using flow cytometry. The GAFS assay was initially developed on partially purified human leukocytes (PBMCs versus granulocytes). We were able to adapt this assay to rhesus leukocytes, and the ability of the flow cytometer to gate out eosinophils in rhesus whole blood based on their high autofluorescence was extremely effective. Agonist properties using GAFS of an assortment of ␤-chemokines on rhesus eosinophils in whole blood were very analogous to those in chemotaxis assays using the transfected macaque CCR3 in the murine pre-B L1-2 cell line. Furthermore, the GAFS assay readily determined antagonistic properties of the anti-macaque CCR3 mAbs on eosinophils in rhesus whole blood, an assay condition not suitable for agonistinduced calcium flux and chemotaxis.
The most well established model of asthma in the nonhuman primate is the cynomolgus macaque (M. fascicularis) model (30,31). These animals demonstrate a naturally occurring and reproducible airway sensitivity to Ascaris suum extract via inhalation. In the model, animals are exposed to single or multiple challenges of A. suum antigen, resulting in a prominent BAL eosinophilia. Indeed, eotaxin has been observed in the BAL rapidly within 6 h of after allergen challenge and remained high for 24 h in this cynomolgus macaque asthma model (17). Moreover, these non-human primates exhibit an early and late phase bronchoconstriction as well as a prominent airway hyperresponsiveness (AHR). The degree of AHR has been shown to be directly correlated with the level of eosinophil-derived major basic protein recovered from the BAL fluid, implying that eosinophil activation products may mediate the development and maintenance of AHR (30). Further support of eosinophil involvement in this primate asthma model has resulted from studies in which direct instillation of purified major basic protein into the airways caused the AHR (81) observed in these animals.
The use of the cynomolgus macaque model has proven crucial in the preclinical evaluation of therapeutic agents for the treatment of asthma in humans. Inhibition of airway eosinophilia and AHR in this non-human primate model was observed with several classes of compounds with disparate mechanisms of action. They include Rolipram, a specific inhibitor of phosphodiesterase IV (32), the corticosteroid dexamethasone (31,33), the leukotriene D 4 receptor antagonist, ICI 198,615 (34), and mAbs directed against intercellular adhesion molecule-1 (35) and interleukin-5 (TRFK-5) (36). Other therapeutic agents that have shown efficacy in AHR include the long acting ␤ 2 -agonist, Salmeterol (33), and the leukotriene B 4 antagonist, CP-105,696 (37). Furthermore, agents that have shown efficacy in late phase bronchoconstriction include the platelet-activating factor receptor antagonist WEB 2170 (38), the 5-lipoxygenase inhibitor, BI-L-239 (39), and a monoclonal antibody directed against endothelial leukocyte adhesion molecule-1 (40), although it is unknown what effect these compounds have on eosinophil recruitment into the airways. In a related non-human primate asthma model, the leukotriene D 4 receptor antagonist Montelukast sodium (Singulair), also known as MK-0476 (82), inhibited early and late phase bronchoconstriction in an A. suum challenge model in squirrel monkeys; however, the effect on BAL eosinophils in this model are unknown.
In summary, we have cloned and pharmacologically characterized the recombinant macaque CCR3 expressed in the murine pre-B cell line L1-2. We have also generated blocking monoclonal antibodies against macaque CCR3 using two different immunogens in mice. The mAbs were very potent in the inhibition of radiolabeled eotaxin binding to macaque CCR3, in addition to being potent CCR3 receptor antagonists in a variety of in vitro functional assays. These signaling assays include eotaxin-induced calcium mobilization and chemotaxis on recombinant macaque CCR3 and shape change on rhesus macaque eosinophils in whole blood. It is hoped that the blocking anti-macaque CCR3 mAbs reported here will be used in the macaque in proof-of-principle experiments to validate the CCR3 hypothesis preclinically as a potential therapeutic for asthma and other allergic diseases in humans.