An immune response is characterized by the activation of leukocytes and migration of hematopoietic cells into secondary lymphoid organs or to the site of inflammation. Under inflammatory conditions, the expression of specific chemokines (Yoshie et al., 2001), a class of small proteins with potent chemotactic activity, is up-regulated. For instance, the expression of the chemokines IL-8 (CXCL8), Gro-α (CXCL1), and ENA78 (CXCL5) is induced in inflammatory human diseases like rheumatoid arthritis, psoriasis, inflammatory bowel disease, acute respiratory distress syndrome, and chronic obstructive pulmonary disease (Gillitzer et al., 1996; Woods et al., 2000; Kurdowska et al., 2002; Banks et al., 2003; Beeh et al., 2003). IL-8, Gro-α, and ENA78 are potent activators of neutrophil chemotaxis (Patel et al., 2001; Feniger-Barish et al., 2003) and are ligands for the G-protein-coupled seven-transmembrane receptor CXCR2 (IL-8Rb). In addition, IL-8 is also a high-affinity ligand for CXCR1 (IL-8Ra). Additional ligands for CXCR2 are granulocyte chemotactic protein-2 (CXCL6) and neutrophil-activating peptide-2 (CXCL7) (Ahuja et al., 1996).

The role of CXCR1 and CXCR2 in the pathogenesis of inflammatory responses has been demonstrated in a number of animal models. For instance, neutralization of IL-8 by a monoclonal antibody resulted in the suppression of a delayed type hypersensitivity reaction, lung inflammation, or tissue injury in rabbits (Larsen et al., 1995; Matsumoto et al., 1997; Yamamoto et al., 1998). Similarly, a small, nonpeptide CXCR2 antagonist (SB225002; White et al., 1998) reduces significantly the number of infiltrating leukocytes into the inflamed joint and the expression of proinflammatory mediators in a rabbit model of arthritis (Podolin et al., 2002).

Human neutrophils express both receptors, CXCR1 and CXCR2, and chemotaxis of human neutrophils can be induced by IL-8, Gro-α, or ENA78 (Patel et al., 2001; Feniger-Barish et al., 2003). Interestingly, although in rodents neutrophil chemotaxis is mediated by a unique homolog of hCXCR2 (Bozic et al., 1994; Harada et al., 1994; Shibata et al., 2000), rabbit neutrophils express homologs of hCXCR1 and hCXCR2 (Beckmann et al., 1991; Thomas et al., 1991; Prado et al., 1994). Although homologs for CXCR1 and CXCR2 have been described for a variety on nonhuman species (Figs. 1 and 2), we wanted to model CXCR1- and CXCR2-mediated responses in cynomolgus monkey (Macaca fascicularis), a species where diseases with a neutrophil component like chronic obstructive pulmonary disease, acute lung injury, and rheumatoid arthritis can be studied (Rose et al., 1992; Mihara et al., 2001; Billah et al., 2002). Thus, we cloned the genes encoding CXCR1 and CXCR2 from cynomolgus monkey, analyzed the pharmacological characteristics of these receptors, and identified them as functional homologs of hCXCR1 and hCXCR2. In addition, comparison of the binding affinities of human chemokines with cynomolgus and human receptors with their specific amino acid sequences allowed the confirmation of formerly established sites critical for ligand binding.

Materials and Methods

Cloning and Expression of Monkey CXCR1 and CXCR2. Cynomolgus CXCR1 was cloned from genomic DNA using PCR primers derived from the human sequence for CXCR1 (GenBank accession no. L19591). For cloning and expression purposes, the 5′ primer (CAAGAATTCGCCGCCACCATGTCAAATATTACAGATCCACA) contained an EcoR1 restriction site (underlined) and a kozak sequence (italic), and the 3′ primer (CAACGCGGCCGCTCAGAGGTTGGAAGAGACATTGA) contained an NotI restriction site (underlined) and a stop codon (italic). The resulting PCR product was cloned into the pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA) and sequenced. To obtain the 5′ and 3′ prime ends of the cCXCR1 gene, primers (5′ primer GGCATAGGTGACAATCACAACATAC and 3′ primer CCAGGCTTATCATCCAAACAACGCCACTCC) were designed for the rapid amplification of cDNA ends (RACE) based on the existing cynomolgus sequence. RACE (BD Biosciences Clontech, Palo Alto, CA) was performed on RNA purified from cynomolgus blood using the PAXgene blood RNA kit by PreAnalytiX (Hombrechtikon, CH). Similarly, cCXCR2 was cloned from genomic cynomolgus DNA using PCR primers based on the hCXCR2 sequence (GenBank accession no. NM_001557): 5′ primer CACTGATGTCTACCTGCTGAACCT starting at position +230f and 3′ primer CCGGTGCTTCTGCCCCATGT starting at +729r (ATG = 1). The PCR product was cloned, sequenced, and based on the sequence of this CXCR2 fragment, and RACE primers were designed to obtain the 5′ and 3′ segments of the cynomolgus gene: 5′ RACE primer, TGTCCTCATAGCAGACTGGGCTAATATA; and 3′RACE primer, TCTTACTTTTCCGAAGGACTGTCTACCT. The RACE reaction was performed on RNA from cynomolgus blood. After sequencing of the RACE products, PCR primers were designed based on the sequence of the 5′ and 3′ end of cCXCR2 to amplify the entire CXCR2 cDNA from blood RNA. For cloning and expression purposes, the 5′ primer contained am EcoR1 restriction site and a kozak sequence (CAAGAATTCGCCGCCACCATGCAGAGTTTCAACTTTGAAGAT), and the 3′ primer contained an NotI restriction site (underlined) and a stop signal (italic) (CAACGCGGCCGCTCAGAGAGTAGTGGAAGTGTGCCC). Cynomolgus CXCR1 and CXCR2 were expressed in BaF3 cells; positive transfectants were selected by G418 resistance. BaF3 cells were maintained in RPMI containing 10% FCS, 1% penicillin-streptomycin and glutamine, and 1 ng/ml murine IL-3 (R&D Systems, Minneapolis, MN). All assays using these recombinant BaF3 cells were performed using pools of cells without further selection for high expressors.

Comparison of Receptor Sequences from Different Species. The amino acid sequences were aligned using the Vector NTI software program. The transmembrane regions of the receptors were predicted using TMpred on the ISREC server.

RNA Extraction and cDNA Synthesis. RNA was prepared from blood from different species using the PAXgene blood RNA kit (PreAnalytix) according to the manufacturer's protocol. RNA was isolated from other tissues using RNA STAT-60 (Tel-Test Inc., Friendswood, TX) according to the manufacturer's protocol. All samples were DNase treated using DNase 1 (Roche Diagnostics, Indianapolis, IN) prior to cDNA synthesis. cDNA was generated using the SuperScript First Strand Synthesis System for reverse transcription-PCR (Invitrogen). Both oligo(dT) (Invitrogen) and random hexomers (Promega, Madison, WI) were used to prime first strand synthesis.

Quantitative PCR Analysis. Quantitative PCR analysis was performed on an ABI 9700 sequence detection instrument (TaqMan) following the manufacturer's instructions. SYBR green TaqMan analysis was performed using primer sequences specific for mouse, rat, rabbit, cynomolgus, and human CXCR1/CXCR2 (Table 1). Ribosomal RNA primers (Applied Biosystems, Foster City, CA) were used to normalize cDNA levels to correct for sample to sample variation. TaqMan PCR conditions were as follows: 48°C for 30 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min. Data were analyzed using Sequence Detection Systems software version 1.7.

Flow Cytometry. Cynomolgus neutrophils were isolated from blood and stained with 20 μl of the phycoerythrin-conjugated anti-human CXCR1 antibody clone 42705 (R&D Systems) or the phycoerythrin-conjugated anti-human CXCR2 antibody clone 48311 (R&D Systems) in PBS, 1% BSA, and 0.1% sodium azide. Events were acquired on a BD Biosciences FACScan and analyzed using the CellQuest software.

Cell Membrane Preparation. Ba/F3-monkey CXCR1 and monkey CXCR2 cell membranes were prepared as previously described (Hipkin et al., 1997). Cells were pelleted by centrifugation, incubated in homogenization buffer (10 mM Tris-HCl, 5 mM EDTA, and 3 mM EGTA, pH 7.6) and 1 mM phenylmethylsulfonyl fluoride for 30 min on ice. The cells were then lysed with a Dounce homogenizer using a stirrer-type RZR3 polytron homogenizer (Caframo, Wiarton, ON, Canada) with 12 strokes at 900 RPM. The intact cells and nuclei were removed by centrifugation at 500g for 5 min. The cell membranes in the supernatant were then pelleted by centrifugation at 100,000g for 30 min. The membranes were resuspended in glygly buffer (20 mM glycylglycine, 1 mM MgCl₂, and 250 mM sucrose, pH 7.2), aliquoted, quick frozen, and stored at -80°C. Protein concentration in membrane preparations was determined using the method of Bradford (Bradford, 1976).

Guanosine 5′-O-(3-[³⁵S]thio)triphosphate ([³⁵S]GTPγS) Binding Assay. The exchange of [³⁵S]GTPγS (triethylammonium salt; specific activity = 1250 Ci/mmol; PerkinElmer Life and Analytical Sciences, Boston, MA) was measured using a scintillation proximity assay (SPA) as previously described (Gonsiorek et al., 2003). For each assay point, 2 μg of membrane was preincubated for 15 min at room temperature with 200 μg of wheat germ agglutinin (WGA)-coated SPA beads (Amersham Biosciences Inc., Piscataway, NJ) in SPA binding buffer (50 mM TRIS-HCL, 1 mM CaCl₂, 5 mM MgCl₂, 50 mM NaCl, 0.002% NaN₃, 0.1% BSA, and 10 μg/ml saponin, pH 7.6). The beads and membranes were transferred to a 96-well Isoplate (Amersham Biosciences AB, Uppsala, Sweden) and preincubated for 60 min with 10 μM GDP in the presence of increasing concentrations of recombinant hIL-8, Gro-α, or ENA-70 (R&D Systems). The GTPγS exchange reaction was initiated by the addition of 0.1 nM [³⁵S]GTPγS and was carried out for 60 min at room temperature. Membrane-bound [³⁵S]GTPγS was measured using a 1450 Microbeta Trilux counter (Amersham Biosciences AB).

Radioligand Binding Assay.¹²⁵I-IL-8 (specific activity = 2200 Ci/mmol) were obtained from PerkinElmer Life and Analytical Sciences. Radioligand competition and saturation binding assays were done using SPA technology (Cox et al., 2001). Membranes (2 μg per assay point) in SPA binding buffer were preincubated for 30 min at room temperature with 200 μg WGA-SPA, transferred to a 96-well Isoplate, and further incubated at room temperature with the radioligand and the indicated concentrations of chemokines for 6 h. Ligand affinities (K_i) from competition binding experiments were calculated from binding IC₅₀ using the Cheng-Prusoff equation (Cheng and Prusoff, 1973).

Chemotaxis Assay. The 96-well 5-micron pore size disposable ChemoTx number 101-5 was purchased from Neuro Probe (Gaithersburg, MD). Lower chambers were filled with 30 μl of various concentrations of IL-8 (R&D Systems) in PBS containing 0.2% BSA. Wells were then covered with a filter. Onto each filter spot, 5 × 10⁴ cells in 25-μl volume were loaded and incubated at 37°C for 90 min. The migrated cells in the lower chamber were transferred into a Microfluor 2 White plate (ThermoLabsystems, Franklin, MA) by spinning through a Funnel plate (Neuro Probe). 25 μl of PBS was added to each well and then filled with 50 μl of CelTiter-Glo luminescent solution (Promega). After 10 min at room temperature, the plate was read on a Luminoskan (ThermoLabsystems) using Acsent Software.

Results

Sequence Analysis of the Cynomolgus Monkey CXCR1 and CXCR2 Receptors. To clone the CXCR1 and CXCR2 receptor from cynomolgus monkey, primers based on the 5′ and 3′ sequences of the respective human receptor were used to amplify gene fragments from monkey blood or spleen. Gene fragments of the expected size were sequenced. Sequence analysis confirmed that these gene fragments indeed resembled primate CXCR1 and CXCR2 sequences. Subsequent RACE reactions then allowed us to obtain the entire cynomolgus genes, including the 5′ and 3′ ends.

Comparison of the amino acid sequences of primate CXCR1 receptors revealed that the cCXCR1 receptor is most closely related to CXCR1 of rhesus monkey (Macaca mulatto; 98% homology) and has about 94% homology to human, chimpanzee, or gorilla CXCR1 (Fig. 1). In contrast to the amino acid sequences of other primate CXCR1s, both macaque CXCR1 receptors have an additional tyrosine inserted at position 14 in their N termini. More surprisingly, however, we observed differences in the cynomolgus sequence at amino acid positions that otherwise were highly conserved in primates and nonprimate sequences, i.e., a tyrosine to histidine change at the end of the transmembrane domain 1 at position 74, a serine to alanine change at position 190, and an arginine to histidine change at the C-terminal border of TM5 (Fig. 1). Since amino acids at these positions are highly conserved among species, the sequence for cCXCR1 was confirmed by sequencing this gene from at least five individual monkeys.

In contrast to CXCR1, cCXCR2 was only 93% homologous to rhesus CXCR2 and had even less homology to human, chimpanzee, or gorilla CXCR1 (∼92%). The cCXCR2 receptor sequence showed the highest variance within the first 12 N-terminal amino acids when compared with rhesus and primate CXCR2 (Fig. 2). Additional differences in the amino acid sequence were noted in regions that otherwise were highly conserved among primate and nonprimate sequences (rabbit, rat, and mouse), i.e., a tyrosine to histidine exchange at position 75 similar to CXCR1, a valine to isoleucine change at position 82, a serine to alanine change at position 107, a glutamic acid to glutamine change at position 287, and a stretch of four nonconserved amino acids at positions 189 to 192 (Fig. 2).

Thus, in comparison with known primate IL-8 receptors, the CXCR1 and CXCR2 receptors of cynomolgus monkey have a significant number of differences in their amino acid sequence. Most noteworthy are a considerable number of changes in domains that are highly conserved in primate and nonprimate species.

Expression Analysis of CXCR1 and CXCR2. We compared the pattern of cCXCR1 and cCXCR2 expression in selected monkey tissues. Highest expression of both receptors was found in blood (Fig. 3, A and B). The expression level of both receptors was about 100-fold lower in spleen and more than 1000-fold lower in brain and spinal cord when compared with blood. The predominant expression of CXCR1 and CXCR2 in cynomolgus blood is similar to the expression pattern seen in human, where high expression of both receptors was found in blood and spleen (Fig. 3, C and D). In a nonprimate species like rabbit, CXCR1 and CXCR2 are again predominantly expressed in blood (Fig. 3, E and F), whereas in contrast, rats express only CXCR2 in blood and spleen. Rat CXCR1, in contrast to the other species investigated, is expressed at only very low levels in the analyzed tissues and blood (Fig. 3F).

Thus, cCXCR1 and cCXCR2 are expressed, like in human, primarily in blood, indicating that the novel cynomolgus genes are likely homologs of the human genes. Like human neutrophils, cynomolgus neutrophils express CXCR1 and CXCR2 as was demonstrated by flow cytometry using antibodies to human CXCR1 or CXCR2 (Fig. 3I).

Pharmacological Characterization of Cynomolgus CXCR1 and CXCR2. Because the cloned cynomolgus receptors CXCR1 and CXCR2 showed differences in otherwise conserved amino acid residues, we tested whether cCXCR1 and cCXCR2 had ligand binding characteristics similar to their human homologs. To this end, we expressed recombinant human and monkey receptors in the pro-B cell line BaF3. To determine the affinity of the receptors to IL-8 and Gro-α, membranes of recombinant BA/F3 cells were incubated in the presence of ¹²⁵I-IL-8 and increasing concentrations of unlabeled human chemokines. The amount of radio-labeled ligand bound to the receptor was determined by scintillation proximity assay technology.

As shown in Fig. 4, cCXCR1 binds human IL-8 (B_max = 1.96 pmol/mg; data not shown) with an affinity similar to that of hCXCR1 (K_d = 170 ± 84 and 103 ± 37 pM, respectively). Similar to hCXCR1, cCXCR1 has very low binding activity to the CXCR2-specific ligand Gro-α (250 ± 84 nM and 1.4 ± 1.1 μM, respectively). Comparison of the chemokine binding affinities to CXCR2 of both species revealed that hCXCR2 binds hIL-8 with an affinity of 210 ± 14 pM, whereas IL-8 binds to cCXCR2 receptor with higher affinity at 46 ± 28 pM (B_max = 0.75 pmol/mg; data not shown). However, although hCXCR2 binds Gro-α with an affinity of 540 ± 140 pM, Gro-α binding to cCXCR2 was 12-fold lower at about 3.7 ± 2.2 nM. Binding of radiolabeled ligand to membranes of parental BA/F3 cells could not be detected (data not shown). Thus, the cynomolgus receptors CXCR1 and CXCR2 have a ligand binding profile similar to that of their human homologs; both receptors bind IL-8 with comparable affinity, whereas Gro-α binding occurs preferentially to CXCR2. However, Gro-α binds with significantly lower affinity to cCXCR2 than to hCXCR2.

To test whether chemokine binding by cynomolgus receptors translated in receptor activation, ligand-dependent [³⁵S]GTPγS exchange was tested on membranes prepared from BaF3 cells expressing cynomolgus CXCR1 or CXCR2. Cynomolgus CXCR1-dependent GTPγS exchange was halfmaximal stimulated by IL-8 concentrations between 84 and 240 pM, whereas 1000-fold higher concentrations of Gro-α were required (EC₅₀ = 260 ± 160 nM; Fig. 5A). Similarly, a half-maximal response of cCXCR2 to IL-8 was reached at concentrations between 1.8 and 10 pM, whereas Gro-α-induced CXCR2 activation occurred at only 1000-fold higher concentrations (EC₅₀ = 27 ± 14 nM; Fig. 5B). In contrast, although ENA-70 was less potent than IL-8 on CXCR1-mediated GTPγS exchange, it was more potent than IL-8 and Gro-α in activating the CXCR2 receptor.

Cynomolgus Receptors CXCR1 and CXCR2 Confer IL-8-Dependent Chemotactic Activity to Recombinant BaF3 Cells. Chemokine binding and activation of hCXCR1 and hCXCR2 results in chemotaxis of cells (Feniger-Barish et al., 2003). To test whether the newly identified cynomolgus receptors CXCR1 and CXCR2 can induce chemotaxis, we tested the migration of recombinant BaF3 cells in response to increasing concentrations of hIL-8. As shown in Fig. 6, recombinant BaF3 cells expressing either cCXCR1 or cCXCR2 migrate toward hIL-8 in a dose-dependent manner (EC₅₀ = 190 and 930 pM, respectively). Parental BaF3 cells did not chemotax to hIL-8, nor did recombinant cCXCR1 or cCXCR2 BaF3 chemotax to hMCP-1, a ligand for CCR2 (data not shown). This indicated that the newly identified cynomolgus receptors CXCR1 and CXCR2 are functional homologs of hCXCR1 and hCXCR2 in their abilities to bind chemokines, activate signal transduction, and confer chemotactic activity to cells.

Ligand binding sites on the hCXCR1 and hCXCR2 receptors have been mapped by analysis of ligand binding to receptor chimeras, to receptors with point mutations, or by inhibition of ligand binding in the presence of an antagonist like SB225002 (Catusse et al., 2003). Since the cynomolgus receptors CXCR1 and CXCR2 show differences in the amino acid sequences at sites that are usually highly conserved among species, we determined the effect of SB225002 on the ligand binding activities of hCXCR1 and hCXCR2 to that of cCXCR1 and cCXCR2. SB225002 only poorly inhibited IL-8 binding to CXCR1 of either species (IC₅₀ in low micromolar range; Fig. 7), confirming it as a CXCR2 selective antagonist (White et al., 1998). However, the binding affinity of SB225002 was about 10-fold more potent in inhibiting IL-8 binding to hCXCR2 (IC₅₀ = 4.8 ± 1.6 nM) than to cCXCR2 (IC₅₀ = 34 ± 11 nM; Fig. 7). These data were consistent with the IC₅₀ of 9.9 nM reported by Catusse et al. for the effect of SB225002 on IL-8 binding to hCXCR2 (Catusse et al., 2003).

Discussion

We identified the novel genes encoding for cynomolgus monkey CXCR1 and CXCR2 and successfully expressed them in recombinant cells. Like for hCXCR1, cCXCR1 binds preferentially IL-8 with high affinity, whereas cCXCR2 binds IL-8 and Gro-α, albeit the binding affinity of human Gro-α to cCXCR2 occurred with a 10-fold lower affinity than to the human receptor. Ligand binding to cCXCR1 or cCXCR2 resulted in receptor activation as determined by GTPγS exchange with an EC₅₀ that correlated with their binding activity, and IL-8 induced a chemotactic response of recombinant cells expressing either cynomolgus receptor. This identifies the cynomolgus receptors as functional homologs of hCXCR1 and hCXCR2.

Alignment of primate and nonprimate receptor sequences reveals a number of unique amino acid differences in the cynomolgus receptors when compared with homologous genes of other primates (Figs. 1 and 2). Outside the N-terminal domain, the majority of amino acid changes are located in transmembrane regions; thus, they are not likely to affect ligand binding. Indeed, this is supported by the fact that binding of IL-8 to both cynomolgus receptors occurs with an affinity similar to that of IL-8 binding to the human receptors.

Extensive analysis of the hCXCR1 receptor has identified two domains as important ligand binding sites: the N-terminal domain (LaRosa et al., 1992) and the third extracellular loop (EL3) (Hebert et al., 1993) suggesting that, in its three-dimensional conformation, these domains are in close contact. Cynomolgus CXCR1 has 98% homology to hCXCR1, and the highest number of differences in the amino acid sequence occur in the N-terminal domain, the fourth transmembrane domain, and the third extracellular loop. Despite these differences in proposed ligand binding sites, cynomolgus and human CXCR1 bind human IL-8 with similar affinity. This suggests that the conformation of the N-terminal domain is flexible to the extent that the addition of an amino acid residue like tyrosine at position 14 in cCXCR1 does not influence IL-8 binding. The role of EL3 in ligand binding to CXCR1 has been demonstrated by binding studies on CXCR1 molecules with mutated amino acid residues. Substitution of amino acids in EL3 at positions 271, 274, 275, or 278 to 280 by alanine abolished IL-8 binding (Hebert et al., 1993; Leong et al., 1994). Although in cCXCR1 the amino acid residues at positions 275 and 278 to 280 are conserved, glutamine is changed to histidine or lysine at position 271 and 274, respectively. These amino acid changes do not affect the affinity of hIL-8 to cCXCR1, suggesting that substitution of polar amino acids by positively charged residues in EL3 does not alter IL-8 binding sites or the conformation of the receptor.

Cynomolgus CXCR2 binds human IL-8 with an affinity comparable with its human homolog, but the binding affinity of Gro-α is approximately 10-fold lower. Alignment of human and cynomolgus receptors identify major differences in the amino acid sequence of the N-terminal domain and the second and third extracellular loops. These regions were implicated as important chemokine binding sites by analysis of ligand binding to human CXCR1-CXCR2 chimeras (Ahuja et al., 1996) and prediction from molecular modeling of CXCR2 (Luo et al., 1997). Within the N-terminal domain, point mutational analysis identified glutamic acid at position seven (E7Q) and aspartic acid at position nine (D9F) of hCXCR2 as sites for interaction with IL-8 and Gro-α (Katancik et al., 2000). Comparison with the cynomolgus N-terminal domain showed differences in the amino acid sequence at these positions (E7Q and D9F) and other changes (K16E, G17N, and T28D). In addition, a stretch of amino acids in EL2 of cCXCR2 differs from the human sequence (LTYI compared with SSNV starting at position 188). Although previous studies have identified these regions as critical IL-8 binding sites of hCXCR2 (Luo et al., 1997; Catusse et al., 2003), binding of IL-8 to cCXCR2 occurs with an affinity similar to hCXCR2. This suggests that despite differences in the amino acid sequence, the three-dimensional conformation and net charge of IL-8 binding sites on cCXCR2 is similar to hCXCR2. However, although these amino acid differences between cynomolgus and human CXCR2 do not influence IL-8 binding, they appear to affect Gro-α binding. An important Gro-α binding site is located in the third extracellular loop of hCXCR2. When within this loop the sequence HID at position 291 was changed to NXG in the human receptor, binding of Gro-α occurred with a 10-fold lower affinity (Catusse et al., 2003). Interestingly, the cynomolgus amino acid sequence differs from the human sequence at this site (human HIDR to cynomolgus NIDQ), which could explain the 10-fold lower binding affinity of human Gro-α to the monkey receptor. Interestingly, the human ligand ENA-70, a truncated version of the hCXCR2 ligand ENA-78, is highly efficacious on cCXCR2 (Fig. 5), suggesting that binding restrictions caused by species-specific differences in the receptor structure do not apply to all ligands.

The influence of HID291 on chemokine binding is further supported by the fact that this sequence has been identified as one of the binding sites for the CXCR2-selective antagonist SB225002 (Catusse et al., 2003). SB225002 inhibits binding of Gro-α and IL-8 to hCXCR2 but has no effect on binding of IL-8 to hCXCR1 (Catusse et al., 2003). Compared with hCXCR2, 10-fold higher concentrations of SB225002 are required to inhibit IL-8 binding to cCXCR2 than to hCXCR2, suggesting that in cCXCR2 at least one binding site for SB225002 is altered. In summary, correlation of the amino acid sequence of human and nonhuman receptors with their ligand binding characteristics substantiates the proposed position of ligand binding sites on these receptors. Additional studies with receptor antagonists support these results, suggesting that comparison of sequences and function of receptors from different species can identify potential ligand binding sites and may guide rational drug design.

The products of the newly identified cynomolgus genes were shown by sequence analysis and functional characterization to be true homologs of hCXCR1 and hCXCR2. These receptors are, like their human and rabbit homologs, highly expressed in blood at similar levels. Similarly, rat CXCR2 is highly expressed in blood, but only low levels of CXCR1 expression could be detected in selected rat tissues and blood. This low expression of rat CXCR1 might occur in macrophages, since expression of CXCR1 on rat neutrophils has not been detected (Dunstan et al., 1996). Interestingly, a gene with 86% homology to rat CXCR1 but only 76% homology to hCXCR1 has recently been identified in mice (GenBank accession no. BC051677), but the function of the proposed rodent CXCR1 receptors remains to be identified. Thus, evaluation of the function of CXCR1 and CXCR2 on neutrophil migration in vivo and their effect on the pathogenesis of inflammatory diseases can only be evaluated in species like rabbit or monkey that show a pattern of CXCR1 and CXCR2 expression similar to that in humans, whereas neutrophil function in inflammatory models in rodents is most likely dependent on the CXCR2 receptor only.

So far, the cynomolgus ligands for CXCR1 and CXCR2 have not been identified, but the sequences for rhesus macaque IL-8 and Gro-α have been reported (Minnerly et al., 1995; Basu et al., 2002). The amino acid sequence of rhesus IL-8 and Gro-α are 94% and 90% identical to the sequence of their human homologs. Based on the low binding affinity of human Gro-α to cCXCR2 and the observed differences in the cCXCR2 amino acid sequence compared with the human and rhesus receptors, we would predict that the sequence for at least cynomolgus Gro-α is significantly different from that of other primates.

In conclusion, this study demonstrates that the pharmacology of homologous gene products can differ substantially even between closely related species. This is emphasized by our observation that the CXCR2 antagonist SB225002 is at least 10-fold less effective in inhibiting ligand binding to cCXCR2 than to hCXCR2. This emphasizes the importance of information on the pharmacology of receptors targeted for the interpretation of animal models and testing the effects of antagonists in vivo.