Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Polymorphonuclear leukocytes (PMNs) represent an important constituent of the innate immune system. Under normal circumstances, they circulate in the vasculature in a quiescent state. However, in response to inflammatory stimuli, they adhere, transmigrate across the vascular endothelium, and enter areas of tissue inflammation where they participate not only in the destruction and removal of pathogens but also amplify the process of inflammation (Kishimoto et al., 1999). Migration of PMNs and phagocytosis of bacteria are tightly controlled, and adhesion receptors present on the cell surface of PMNs are crucial for initiating these responses. Upon activation of PMNs in response to stimuli such as TNF, the adhesion receptors, members of the integrin superfamily, are activated, and additional receptors present in intracellular granules translocate to the cell surface (Carlos and Harlan, 1994). The interaction of integrins with ligands present on vascular endothelial cells, stromal cells, and bacteria promotes adhesion and its sequelae.

The major adhesion-promoting receptor present on the cell surface of activated PMNs is complement receptor type 3 (CR3), also called Mac-1, CD11b/CD18, or _M2 (Kishimoto et al., 1999). CR3 is a heterodimer belonging to the ₂ integrin subfamily. The two subunits, _M (M_r 165,000) and ₂ (M_r 95,000), are held together by noncovalent interactions. Studies have been done on the structure and functions of ₂ integrins and have been reviewed extensively (Harris et al., 2000; Shimaoka et al., 2002). Both  and  subunits, are transmembrane glycoproteins with long extracellular domains, short membrane spanning domains, and cytoplasmic tails. The extracellular portion of  and  subunits contain I- and I-like domains, respectively, near the N terminus that play an important role in ligand binding. The cytoplasmic tails contain sequences critical for inside-out signaling and cytoskeleton association (Shimaoka et al., 2002).

Ligand binding by CR3 is divalent cation-dependent. CR3 recognizes many different protein and nonprotein ligands, such as extracellular matrix proteins, blood coagulation proteins, intracellular adhesion molecules, the complement pathway product C3bi, and others (Zhang, 1999). Overlapping but nonidentical ligand binding pockets for distinct ligands have been identified by mutagenesis studies of _M and ₂ (Zhang and Plow, 1996).

All leukocytes are an integral component of the host defense system, but in some disease states their actions may instead be deleterious to the host. Leukocyte activation may cause tissue damage by inducing the release of reactive oxygen intermediates (ROIs) and other radicals, proteases, and arachidonic acid metabolites. PMNs from CR3 knockout mice demonstrate reduced spreading, phagocytosis, and oxygen radical generation and show an unanticipated defect in PMNs apoptosis (Coxon et al., 1996). CR3 deficiency abrogates sustained Fc receptor-dependent neutrophil adhesion and complement-dependent proteinuria in acute glomerulonephritis (Tang et al., 1997). These mice are less susceptible to cerebral ischemia/reperfusion injury and show decreased neointimal formation after balloon injury of the carotid artery (Soriano et al., 1999; Simon et al., 2000). Therefore, agents that block leukocyte recruitment into sites of inflammation have the potential to treat PMN-mediated inflammation in a variety of inflammatory diseases.

Antiadhesion therapies using antibodies, peptides, and peptidomimetic inhibitors against adhesion receptors or combination therapies directed against both adhesion receptors and their ligands are beneficial in a variety of animal models and disease processes (Jaeschke et al., 1993; Cornejo et al., 1997; Curley et al., 1999). A number of humanized antibodies against adhesion receptors are in phase I and II clinical trials, but the potential long-term risk of complications due to these antibodies will also have to be considered. A class of small organic molecules called leumidins with activity in a number of animal models of inflammation has also been described (Burch et al., 1992). They apparently block the process of inflammation by inhibiting the recruitment of leukocytes into tissue. Although the exact mechanism of action of leumedins is unknown, they do inhibit the up-regulation of CR3 expression on PMNs. We now report the discovery and characterization of two low-molecular-weight compounds 2-[4-(3,4-dihydro-2H-quinolin-1-yl)-buta-1,3-dienyl]-1-ethyl-naphtho[1,2-d]thiazol-1-ium; chloride (compound 1) and 1-ethyl-2-/3-/1-ethylbenzothiazolin-2-ylidiene/propenyl/-thiazolium; iodide (compound 2) that prevent CR3 binding to C3bi and also block the adhesion of human PMNs to fibrinogen and the adhesion-mediated production of ROIs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents and Antibodies. Calf intestinal alkaline phosphatase (AP), N-succinimidyl 3-(2-pyridithio) propionate, Tween 20, Super Block, Slide-A-Lyser dialysis cassettes, biotinyl-3-maleimidopropionamidyl-3-6-dioxaoctanediamine, 2-(4'-hydroxyazobenzene)-benzoic acid (avidin-HABA) reagent, and (N-hydroxysulfosuccinimide)-biotin (Sulfo-NHS) were all purchased from Pierce Chemical (Rockford, IL). Iodoacetamide, n-octyl -D-glucopyranoside, Sepharose CL-4B-200-activated thiol-Sepharose, cyanogen bromide-activated Sepharose, diethylenetriaminepentaacetic acid, diisopropyl fluorophosphate, and n-octyl -D-glucopyranoside, C5-deficient human serum and bovine serum albumin were from Sigma-Aldrich (St. Louis, MO). Dulbecco's phosphate-buffered saline (PBS) with and without Ca²⁺ and Mg²⁺, was from Invitrogen (Carlsbad, CA). Attophos was from Promega (Madison, WI). Enhancement solution was bought from PerkinElmer Wallac (Gaithersburg, MD). Ficoll was from Amersham Biosciences Inc. (Piscataway, NJ). All protease inhibitors were from Calbiochem (San Diego, CA). SDS gels, molecular weight markers, and silver staining kit were from Novex (San Diego, CA). Sheep red blood cells (SRBCs) were from BioWhittaker (Walkersville, MD).

Bulk Isolation of PMNs for Protein Purification. Large quantities of PMNs for protein purification were isolated at ambient temperature from leukocyte-enriched Polypacs. Red cells were removed by sedimentation for 1 h after the addition of an equal volume of freshly prepared 6% dextran (T-500) in 0.9% saline to the blood fraction. The leukocyte-enriched supernatant was centrifuged at 500g for 10 min. The cell pellets were resuspended in 50 ml of phosphate-buffered saline containing Ca²⁺ and Mg²⁺ (PBS⁺⁺), and 25 ml was layered onto 15 ml of Ficoll. The buffy coat overlaying the red cell layer was isolated, and contaminating red cells were removed by three rounds of hypotonic lysis with distilled water. PMNs were suspended in PBS⁺⁺, and the cell number was determined by counting in a hemocytometer. Cells were treated with 5 nM diisopropyl fluorophosphate for 5 min, pelleted, snap frozen in liquid nitrogen, and stored at 80°C.

Purification of CR3. CR3 was purified from isolated human PMNs by affinity chromatography using IB4 mAb as described previously, with modifications (Cai and Wright, 1995). PMNs (4-9 × 10¹⁰) were homogenized in 150 ml of freshly prepared lysis buffer (100 mM Tris, pH 8.0; 150 mM NaCl; 2 mM MgCl₂; 1% Triton X-100; 0.02% NaN₃; 1 µg/ml each of antipain, benzamidine, chymostatin, leupeptin, and pepstatin; and 2 mM phenylmethylsulfonyl fluoride) and lysed for 2 h at 4°C with stirring. The lysate was centrifuged at 35,000g for 30 min. The supernatant was transferred to a 250-ml conical tube containing 10 ml of Sepharose beads, complexed according to the manufacturer's instructions with IB4, and prewashed with equilibration buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl₂, and 0.1% Triton X-100). The contents were mixed end-over-end for 1.5 h at 4°C. The beads were removed by filtration and sequentially washed with at least 400 ml each of equilibration buffer and buffer composed of Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl₂, and 0.1% n-octyl -D-glucopyranoside. The beads were then packed tightly into a 10-ml column. CR3 was eluted with elution buffer (50 mM triethylamine, 150 mM NaCl, 2 mM MgCl₂, and 1% n-octyl -D-glucopyranoside) and collected in 3-ml fractions into tubes containing 450 µl of 1 M Tris, pH 8. The protein was dialyzed in a Slide-A-Lyser 10 K dialysis cassette (15-ml capacity; Pierce Chemical) against 2 liters of PBS⁺⁺ overnight in the cold room. The dialyzed protein was divided into 200-µl aliquots, snap frozen in liquid nitrogen, and stored at 80°C. The purity of CR3 was analyzed by SDS-polyacrylamide gel electrophoresis, using 4 to 12% Tris-glycine gels under reducing conditions that were stained with silver. Two strong bands corresponding to the  and  chains of CR3 (CD11b and CD18, respectively) were observed at 165 and 95 kDa (Fig. 1). The identity of the _M chain (165-kDa band) was confirmed by Western blotting with polyclonal anti-human CR3 Ab (data not shown). PMNs contain low amounts of CD11a/CD18, but our preparations did not show a CD11a band, which was further confirmed by Western blotting using monoclonal anti-human CD11a Ab (data not shown).

View larger version (100K): [in this window] [in a new window]   Fig. 1.   Purification of CR3 from human PMN lysates, as described under Materials and Methods. Proteins were analyzed on a 4 to 12% SDS-polyacrylamide gel electrophoresis gel under reducing conditions, and the bands were stained with a silver stain kit. Lane 1, Novex Mark 12 standards (200-55 kDa); lane 2, PMN (35,000g) supernatant; and lanes 3 and 4, purified CR3 (490 and 245 ng).

Purification of C3bi and Preparation of C3bi-Alkaline Phosphatase. C3bi was purified from fresh human serum and coupled to AP exactly as described previously (Cai and Wright, 1995). Briefly, serum was treated with 20 mM iodoacetamide for 1 h at 37°C, to block all free sulfhydryl groups. After extensive dialysis, the serum was incubated with 8 g of activated thiol-Sepharose for 2 h at 37°C, to promote activation of complement and deposition of C3bi on the Sepharose via reactivity with the internal thioester. After thorough washing with 20 mM Tris, pH 7.4, C3bi was eluted with 10 mM L-cysteine in 20 mM Tris, pH 7.4. C3bi was further purified on a Mono Q column with a gradient of NaCl. C3bi was conjugated to AP through a disulfide bridge with N-succinimidyl 3-(2-pyridithio) propionate. Conjugated and free C3bi were separated on a Mono Q column and eluted with a NaCl gradient. The column fractions were further analyzed by electrophoresis on 8 to 25% polyacrylamide native gels. The presence of a major band whose mobility was slower than purified C3bi or AP, indicated the conversion of C3bi to C3bi-AP. The R_F value of this conjugated band also matched with the purified C3bi-AP, provided by Dr T.-Q. Cai (Merck Research Labs) (data not shown). The C3bi-AP fractions were pooled, and 100-µl aliquots were snap frozen in liquid nitrogen and stored at 80°C.

Biotinylation of C3bi. Human purified C3bi was biotinylated on the free sulfhydryl group of C3bi after elution from thiol-Sepharose and further purification on a Mono Q column. Then 100 µl of 10 mM biotinyl-3-maleimidopropionamidyl-3-6-dioxaoctanediamine-activated biotin was added to 2.5 mg of purified C3bi in 1 ml of 50 mM Tris-HCl, pH 7.5, and incubated overnight at ambient temperature with shaking. C3bi-biotin was purified on a Mono Q Sepharose column, as described above, and C3bi-biotin fractions were pooled. The amount of biotin in C3bi-biotin was estimated using avidin-HABA, and stoichiometric coupling between C3bi and biotin was observed. C3bi-biotin was further biotinylated by adding 0.5 mg of Sulfo-NHS to 2.5 mg of C3bi biotin in 1 ml and incubating for 30 min at 4°C. The mixture was dialyzed against 20 mM Tris-HCl, pH 7.5, 5 mM EDTA in the cold to remove excess Sulfo-NHS-biotin. The amount of biotin in C3bi-biotin was estimated, using avidin-HABA, to be 3 to 10 molecules of biotin per C3bi, depending on the preparation.

Preparation of Biotinylated EC3bi. Biotinylated EC3bi were prepared as described by Li and Arnaout (1999). SRBCs were sensitized with 10 µl of anti-SRBC Ab 19S for 30 min at 37°C and then transferred to ice for another 30 min. Sensitized SRBCs were washed twice in DHVB⁺⁺ buffer (0.15 M glucose, 0.15 mM CaCl₂, 0.5 mM MgCl₂, 0.5% human serum albumin, 1 mM sodium barbital, and 30 mM NaCl, pH 7.4), resuspended at 10⁹ cells/ml in DHVB⁺⁺ buffer, and stored on ice at 4°C. To biotinylate the cells, 1 mg of Sulfo-NHS-biotin was added to 2 ml of sensitized SRBCs and incubated at 4°C for 30 min. Biotinylated SRBCs were washed twice in DHVB⁺⁺ buffer, resuspended at 10⁹ cells/ml in DHVB⁺⁺ buffer, and stored on ice. To make biotin-EC3bi, 100 µl of C5-deficient human serum was added to 1 ml of biotinylated SRBCs at 37°C for 1 h. Opsonization was terminated by adding 1 ml of HVB/EDTA buffer (1 × Veronal buffer, 1.88 g/l N-5,5-diethyl barbital, 2.88 g/l N-5,5-diethyl barbituric acid, pH 7.4, 0.1% human serum albumin, and 1 mM EDTA) and incubating the mixture on ice for 10 min. Biotinylated EC3bi were washed twice in DHVB⁺⁺ buffer, resuspended at 10⁸ cells/ml of DHVB⁺⁺ buffer, and stored on ice at 4°C. Cells were used within 2 weeks after preparation.

Binding of C3bi-AP to Immobilized CR3. To measure the binding of C3bi-AP to purified CR3, 96-well Immulon 4HBX flat bottom microtiter plates (Dynex Technologies Inc., Chantilly, VA) were coated overnight with 50 µl/well of 25 nM CR3 diluted in PBS⁺⁺. The wells were washed four times with 0.1% Tween 20 in PBS⁺⁺, and nonspecific sites on the plastic were blocked for 1 h at ambient temperature with 100 µl/well of Superblock prepared in PBS⁺⁺ according to the manufacturer's protocol. After an additional four washes with 0.1% Tween 20 in PBS, 25 µl of 25% Superblock was added to the wells, with or without the test reagent, and the plates were shaken on a titer plate shaker for 3 min. Then 25 µl of 10 nM C3bi-AP ligand was added per well, and the plates were shaken an additional 3 min and placed at 37°C for 60 min. The concentration of C3bi-AP was chosen from titration experiments indicating that this concentration was within the linear range for C3bi-AP binding to CR3. Unbound C3bi-AP was removed by four washes with 0.1% Tween 20 in PBS⁺⁺. Then 50 µl of Attophos substrate for AP activity was added to the wells and incubated for 45 min at 37°C. The reaction was stopped by adding 3 µl of 5 M inorganic phosphate. Fluorescence was measured by CytoFluor (Series 4000; Applied Biosystems, Foster City, CA) at 450/50-nm excitation and 580/25-nm emission. CR3 immobilized in wells that did not contain C3bi-AP was used for the determination of background fluorescence. Percentage of inhibition of C3bi-AP binding to immobilized CR3 was calculated as 100  [(fluorescence with test reagent  background fluorescence)/(fluorescence with no test reagent  background fluorescence) × 100].

Binding of Biotinylated C3bi to Immobilized CR3. An additional assay was developed to measure binding of C3bi to immobilized CR3 via a dissociation-enhanced lanthanide fluorescent immunoassay (DELFIA) method, with time resolved fluorescence. Maxisorb 96-well plates (Nalge Nunc International, Rochester, NY) were coated overnight at 4°C with 50 µl/well of 20 nM CR3 diluted in PBS⁺⁺ buffer. Nonspecific sites were blocked with 200 µl/well of 2 mg/ml casein in PBS⁺⁺ buffer for 1 h at ambient temperature. Wells were then washed three times with 0.01% Tween 20 and 0.01% sodium azide in PBS⁺⁺. Then 25 µl of test reagent in 0.5 mg/ml casein in PBS⁺⁺ was added per well, along with 25 µl of 20 nM biotinylated human-C3bi diluted in 0.5 mg/ml casein in PBS⁺⁺. Plates were shaken briefly, incubated at 37°C for 1 h, and washed with 0.01% Tween 20 and 0.01% sodium azide in PBS⁺⁺. Then 100 µl of streptavidin-labeled Europium in 1% bovine serum albumin, 0.05% Tween 20, 7 µg/ml diethylenetriaminepentaacetic acid in PBS⁺⁺ was added to each well, and the plate was incubated for 45 min in the dark at room temperature. The plates were washed as described above, and 100 µl/well of Enhancement solution was added. After 15 min, the fluorescence was read on a Victor 2 plate reader (PerkinElmer Wallac, Boston, MA) at 340-nm excitation and 615-nm emission. Data are presented as arbitrary fluorescence units.

Binding of Biotinylated EC3bi to Immobilized CR3. To quantitate inhibition of multimeric interactions between C3bi and CR3, binding of biotinylated EC3bi to immobilized CR3 was measured using a DELFIA method. Nunc Maxisorb 96-well plates were coated overnight at 4°C with 50 µl of 25 nM purified CR3 in PBS⁺⁺. The plates were blocked and washed as described above. To each well, 25 µl of test reagent in DHVB⁺⁺ buffer was added along with 1.5 × 10⁵ biotinylated EC3bi in 25 µl of DHVB⁺⁺ buffer. The plates were incubated at 37°C for 30 min and then inverted to dislodge unbound EC3bi. Then 50 µl of 1% glutaraldehyde solution in PBS⁺⁺ was added to each well, for 1.5 to 2 h at 37°C. After washing with 0.01% Tween 20 and 0.01% sodium azide in PBS⁺⁺, 50 µl of 1% human serum albumin/PBS⁺⁺ was added for 1.5 to 2 h at 37°C to quench excess glutaraldehyde. After further washing with 0.01% Tween 20 and 0.01% sodium azide in PBS⁺⁺, Enhancement solution was added and fluorescence was read as for the biotinylated C3bi assay described above.

Cloning, Expression, and Purification of Human ICAM-1/Ig. Domains 1 through 5 of human ICAM-1 (GenBank no. X06990) were amplified by PCR using Raji QuickClone cDNA (BD Biosciences Clontech, Palo Alto, CA) as a template and forward (5'-taaagacctcagcctcgctatg-3') and reverse (5'-ggagagcacattcacggtcac-3') primers. At the 5' end of the forward primer, a HindIII site was placed, and at the 5' end of the reverse primer, BamHI and splice donator sites were placed. PCR was performed with Expand (Roche Diagnostics, Indianapolis, IN) in a PerkinElmer (PE 480) thermocycler by using 40 cycles with the following parameters: 90 s at 94°C, 90 s at 53°C, and 90 s at 72°C. PCR products from six separate reactions were pooled and purified using the QIA quick PCR purification kit (QIAGEN, Valencia, CA). The PCR products were digested with HindIII and BamHI, purified as described above, and ligated into a HindIII- and BamHI-digested pCDNA3.1/human Ig fusion protein construct. Clones were checked for inserts and sequenced. Sequence results revealed all clones had the same point mutation three base pairs upstream of the reverse primer. A second round of PCR was performed with the forward primer mentioned above and a new reverse primer (5'-ggagagcacattcacggtcacctcg-3') using 100 ng of DNA from the first round of PCR as the template. PCR was performed with Expand in a PE 480 thermocycler by using 10 cycles with the following parameters: 30 s at 94°C, 60 s at 58°C, and 3 min at 72°C. PCR products from six separate reactions were pooled, purified, digested, and ligated into pCDNA3.1/human Ig fusion protein construct as described above. Clones were sequenced and the correct sequence for the resulting human ICAM-1 fragment fused to human Ig was verified. The plasmid was transfected into Chinese hamster ovary cells using LipofectAMINE (Invitrogen) and grown under 0.5 mg/ml G418 (Invitrogen) selection. Culture supernatants from single cell clones were assayed by Ig enzyme-linked immunosorbent assay, and a high expressing clone was adapted to Chinese hamster ovary-serum-free media (Invitrogen) for large-scale expression. Secreted ICAM-1/Ig was purified from crude culture supernatants using protein A/G affinity chromatography and dialyzed against PBS without Ca²⁺ and Mg²⁺ (PBS).

Preparation of Biotinylated Human ICAM-1/Ig Fusion Protein and binding to Immobilized CR3. Purified human ICAM-1/Ig (2 mg/ml) was dialyzed overnight in 50 mM sodium carbonate/bicarbonate buffer, pH 8.5. Then 0.135 ml of Sulfo-NHS-biotin (1 mg/ml) was added to 1 ml of purified human ICAM-1/Ig and incubated for 6 h at ambient temperature. Unreacted biotin was removed using Bio-Gel P30 columns (Bio-Rad, Hercules, CA). The amount of biotin in ICAM-1/Ig-biotin was estimated using avidin-HABA, and there were approximately three molecules of biotin per ICAM-1/Ig in most preparations.

Cell Preparation for Functional Assays. PMNs were prepared from human venous blood freshly drawn into heparinized syringes and separated on PMNs isolation medium (Cardinal Associates, Inc., Santa Fe, NM) exactly as described previously (Detmers et al., 1991). Contaminating erythrocytes were removed by hypotonic lysis. PMNs were suspended either in Dulbecco's PBS with 0.5 mg/ml human serum albumin, 0.3 U/ml aprotinin, and 3 mM glucose (HAP buffer), for adhesion assays; or Krebs-Ringer phosphate buffer with 5.5 mM glucose (KRPG buffer), for assays of superoxide production.

PMNs Adhesion Assay. Adhesion of human PMNs to fibrinogen-coated Terasaki plates (Robbins Scientific, Sunnyvale, CA) was performed exactly as described previously (Detmers et al., 1998). Briefly, PMNs were labeled with the fluorescent dye carboxyfluorescein diacetate succiminidyl ester (Molecular Probes, Eugene, OR), and 10⁴ cells were added to each fibrinogen-coated Terasaki well. After addition of test reagents, the plates were incubated for 10 min at 37°C before addition of agonists (15 ng/ml TNF, 30 ng/ml PMA) and continued incubation at 37°C. Adhesion was quantitated by measuring the fluorescence in each well in a CytoFluor (Applied Biosystems, Foster city, CA) before and after washing. Percentage of adhesion was calculated as (fluorescence after washing/fluorescence before washing) × 100. Samples for each condition were run in triplicate, and the data are presented as the mean ± S.E.M. Experiments representative of at least two repetitions are shown.

Adhesion-Dependent Oxidative Burst Assay. The production of H₂O₂ by adherent PMNs was measured using the previously published assay of horseradish peroxidase (HRP) catalyzed oxidation of scopoletin (Nathan et al., 1989). Polystyrene Primaria 96-well tissue culture plates (BD Biosciences, Lincoln Park, NJ) were coated with fetal bovine serum for 30 min at 37°C. Each well was washed vigorously three times by forceful squirting of 3 ml of 0.9% NaCl, to prevent interference with scopoletin fluorescence by soluble bovine serum albumin. To each well was added 50 µl of scopoletin (39 µM in KRPG), 10 µl of HRP (0.5 U/ml), 15 µl of NaN₃ (1 mM), 10 µl of test reagent or vehicle, and 3 × 10⁴ PMNs in KRPG. The reaction was initiated by adding agonist (TNF or PMA) at the indicated concentrations, giving a final assay volume of 125 µl. The scopoletin fluorescence was measured every 10 min (360-nm excitation, 460-nm emission) in a CytoFluor fluorescence plate reader equipped with a temperature control device to maintain the plates at 37°C throughout the assay. Samples were run in quadruplicate, and the data are presented either as the mean ± S.E.M. of three experiments or as a representative experiment. The nanomoles of H₂O₂ produced per 30,000 PMNs were calculated as described previously (Vaidya et al., 1999). Unless otherwise stated, percentage of inhibition of H₂O₂ production was calculated as 100  ((nmol of H₂O₂ from test reagent- and agonist-treated PMNs  nmol of H₂O₂ from buffer-treated PMNs)/(nmol of H₂O₂ from agonist-treated PMNs  nmol of H₂O₂ from buffer-treated PMNs) × 100), using values taken at 60 min.

Protein Determination Assay. Protein concentration was determined by the Bio-Rad protein assay (Bio-Rad), with bovine serum albumin as a standard.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of Two Compounds that Inhibit CR3 Binding to C3bi but Not to ICAM-1. An assay measuring binding of C3bi-AP to immobilized CR3 was used to screen for small molecule antagonists of CR3. The receptor purified by affinity chromatography contained both  and  chains (Fig. 1) and bound avidly to C3bi-AP. CR3 interacts with a variety of protein and nonprotein ligands, but monomeric C3bi was chosen as a ligand because it binds CR3 with high affinity (0.127 µM) (Cai and Wright, 1995). In the development of the assay, variation in the concentration of CR3 used to coat the wells from 6.25 to 25 nM yielded a linear signal with 10 nM of C3bi-AP. Similarly, varying C3bi-AP concentrations from 10 to 100 nM resulted in a linear signal with 25 nM of CR3. Based on these preliminary studies, 25 nM CR3 and 10 nM C3bi-AP were chosen as the standard concentrations. The assay was also linear with time up to 60 min, so a 60-min time point was chosen for convenience. As would be expected for the interaction of an integrin with its ligand, C3bi-AP binding activity to CR3 was absolutely dependent on divalent metal ions (Ca²⁺ and Mg²⁺) and was abolished by 10 mM EDTA. Binding of C3bi-AP to CR3 was completely inhibited by mAb 44a (10 µg/ml), an anti-CR3 antibody known to block ligand binding, further demonstrating the specificity of the interaction. As a result of the screen, two compounds were identified, compound 1 and compound 2 (Fig. 2), that inhibited ligand binding with IC₅₀ values of 0.14 and 0.33 µM (Table 1), respectively, in eight-point titrations. For comparison, mAb 44a blocked binding of C3bi-AP to CR3 with an IC₅₀ value of 0.002 µM. These compounds were shown to be pure by liquid chromatography-mass spectrometry, demonstrating a single peak at the expected molecular weights (433 for compound 1 and 492 for compound 2).

View larger version (6K): [in this window] [in a new window]   Fig. 2.   Chemical structures of compound 1 and compound 2.

Compounds 1 and 2 Inhibit Multimeric Binding between CR3 and EC3bi. These compounds were also tested for inhibition of the multimeric interaction between biotinylated EC3bi and immobilized CR3. Binding of EC3bi to CR3 was inhibited 70 and 85% by anti-CR3 mAb 44a at 10 and 50 µg/ml, respectively, whereas anti-human LFA-1 (clone 38) at a concentration up to 50 µg/ml did not show any inhibition, demonstrating that any potential contamination by LFA-1 in the CR3 preparation did not contribute to the interaction. Compounds 1 and 2 inhibited EC3bi binding to CR3 with IC₅₀ values of 6.7 and >10 µM (Table 1), respectively, 30- to 48-fold higher than the IC₅₀ values for inhibition of monomeric C3bi-AP binding to CR3. Both compounds were less potent in blocking EC3bi binding to CR3 than divalent mAb 44a, which had an IC₅₀ value of 0.04 µM in this assay. Because EC3bi binding to CR3 is a multimeric event (Li and Arnaout, 1999), it is likely to be more difficult to inhibit than the binding of monomeric C3bi-AP to CR3. Therefore, higher concentrations of compound were required to block EC3bi binding to CR3, and multivalent reagents (mAb) showed an advantage.

Binding of Compounds 1 and 2 to CR3 Is Not Easily Reversed and Is Enhanced by Light. The presence of unsaturated double bonds in the carbon chain of the compounds (Fig. 2) suggested the possibility that a covalent adduct was formed with CR3. To test this possibility, the ability of compounds 1 and 2 to block C3bi-AP binding was tested after removal of the compounds from the test wells. Compounds 1 or 2 were preincubated with immobilized CR3 for 2 h at 37°C at 1 µM, 3- to 7-fold higher than their IC₅₀ values for inhibition of C3bi-AP binding to CR3. The unbound compounds were then removed with several washes in PBS⁺⁺. In parallel, 1 µM compound 1 or 2 was added to CR3-coated wells just before the addition of C3bi-AP. Whether compounds 1 and 2 were preincubated with CR3 and removed or added just before the addition of C3bi-AP, there was similar inhibition of C3bi-AP binding (Fig. 3) (86% inhibition with compound 1 and 70% with compound 2). To confirm these observations, the effect of these compounds at 1 µM was tested using a higher concentration of C3bi-AP, but the inhibition of C3bi-AP binding to CR3 was not reversed even by ligand concentration up to 0.1 µM, suggesting that these compounds are not easily displaced from CR3.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Inhibition of CR3 activity by compound 1 or 2 is not readily reversed. Compounds were incubated at 1 µM for 2 h with immobilized CR3. The wells were left unwashed or washed extensively with PBS++. Control wells with no compounds were run in parallel in the assay for C3bi-AP binding to CR3, as described under Materials and Methods.

Compounds 1 and 2 Inhibit Cellular Functions that Depend upon CR3. Compounds 1 and 2 were tested for their ability to block CR3-dependent adhesion by human PMNs. Fibrinogen, a ligand for CR3 (Wright et al., 1988), was chosen as the substrate for CR3-mediated adhesion, and adhesion was measured using a previously validated assay (Detmers et al., 1998). PMNs were preincubated for 10 min at 37°C with increasing concentrations of the compounds before addition of TNF (15 ng/ml) or PMA (30 ng/ml) for 15 to 20 min at 37°C and quantitation of adhesion, as described under Materials and Methods. Compound 1 blocked PMNs adhesion to fibrinogen in response to TNF and PMA with IC₅₀ values of 3.7 ± 0.008 and 8.6 ± 0.004 µM, respectively, in eight-point titrations. Compound 2 also blocked CR3-dependent adhesion with IC₅₀ values of 2.5 and >10 µM with TNF and PMA, respectively (Table 1). Tests for cytotoxicity indicated that the compounds were not toxic to human PMNs up to 10 µM for 60 min (data not shown). These studies indicate that CR3-mediated multimeric binding to fibrinogen can be blocked by both compounds 1 and 2, albeit with IC₅₀ values higher than those observed for blockade of binding monomeric C3bi-AP. For comparison, mAb 44a blocked TNF- or PMA-stimulated PMNs adhesion with IC₅₀ values of 0.02 µM and 0.04, respectively (Table 1).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Inhibition of the adhesion-dependent oxidative burst by compounds 1 and 2. Human PMNs were incubated in serum-coated wells for 10 min at 37°C with 1 µM compound 1 or 2 before addition of TNF (100 ng/ml) or KRPG and continued incubation at 37°C. H2O2 produced was measured at 10-min intervals, as described under Materials and Methods, with a decrease in arbitrary fluorescent units indicating production of H2O2. Each compound was tested twice, and the values are the mean of the four wells from one representative experiment. , blank (no TNF); , control (+TNF); , TNF + compound 1; , TNF + compound 2.

Structure-Activity Relationships for Compounds Related to Compound 2. The two benzothiazole rings of compound 2 are connected through a diene linkage, and analogs with increased chain length between the rings (compounds 3 and 4) were evaluated at a single concentration (1 or 10 µM) in assays for CR3 binding to C3bi-AP, adhesion of PMNs to fibrinogen, and ROIs by adherent PMNs (Table 3). Changing from a diene to a triene linkage, in compound 3, had little effect on the efficacy of blocking C3bi-AP binding or PMNs adhesion. Compound 3, however, had reduced ability to inhibit the adhesion-dependent oxidative burst, compared with compound 2. Further increasing the chain length to a tetraene linkage, in compound 4, resulted in no blockade of C3bi-AP binding to CR3 up to 10 µM, reduced ability to inhibit PMNs adhesion, but no further reduction in inhibiting the adhesion-dependent oxidative burst. Thus, there was a decrease in inhibition with the increase in the carbon chain length between the ring structures.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

CR3-mediated cell adhesion plays a critical role in PMNs functions such as host defense and cell trafficking. The severity of illness in leukocyte adhesion deficiency patients, who fail to express CR3 and other 2 integrins, emphasizes the importance of these receptors in maintaining immunological homeostasis. Inappropriate activation of PMNs, however, can have detrimental consequences (Weiss, 1989). At sites of tissue inflammation, PMNs interact with stromal cells and extracellular matrix, adhere, become activated, and produce a multitude of cytokines (Meda et al., 1994), superoxide radicals, and proteolytic enzymes (Albelda et al., 1994) that cause injury to the surrounding tissue. PMNs have been implicated in causing ischemia/reperfusion injury, as well as the development of various autoimmune diseases (Tang et al., 1997; Soriano et al., 1999).

There is strong evidence in the literature, based on studies conducted in animal models, that anti-CR3 antibodies decrease ischemia/reperfusion injury (Jaeschke et al., 1993), decrease the area of myocardial infarction (Curtis et al., 1993), decrease liver cell injuries (Jaeschke et al., 1991), and diminish neointimal thickening and restenosis after balloon injury of carotid arteries (Rogers et al., 1998). These antibodies are also effective in the treatment of endotoxic challenge and hemorrhagic shock (Burch et al., 1993). Although the use of antibody therapy seems to be very promising, adverse effects due to nonselective blockade of various other leukocyte functions may lead to severe complications (Ramamoorthy et al., 1997). The identification of small molecules that selectively block the binding CR3 to its ligands may prove to be a better therapeutic agent than antibodies.

Blockade of the binding sites of integrins with peptides based on their ligands or small molecules derived from the structure of such peptides has proven effective in inhibiting the activities of ₁ and ₃ integrins. Several peptide and nonpeptide inhibitors (integrilin, lamifiban, xemilobifan) of _IIb3 based on the arginine-glycine-aspartic acid motif sequence of the ligand fibronectin have been identified. Integrilin was found to be effective in minimizing platelet aggregation and preventing coronary thrombosis in humans (Curley et al., 1999). Peptidomimetic inhibitors of ₄ integrins have been found to be effective in vivo in a mouse model of delayed type hypersensitivity and in a sheep model of airway hyperresponsiveness (Zimmerman, 1999). However, peptides derived either from the ligands of CR3 or from the complementarity-determining region of anti-CR3 antibodies were not very efficacious in blocking ligand binding in vitro (Feng et al., 1998). The failure of the peptides to block the interaction between C3bi and CR3 may be due to their improper conformation in solution or to the size of the ligand binding sites, which may be too extensive to block with a small peptide.

A 41-kDa glycoprotein isolated from canine hookworms known as neutrophil inhibitory factor (NIF) has been shown to interact with CR3 and inhibit a number of PMNs functions. NIF binds to the I-domain of CR3 with high affinity and blocks the binding of several ligands to CR3, including C3bi, ICAM-1, and fibrinogen as well as adhesion of PMNs to protein-coated surfaces (Zhang and Plow, 1997). NIF has been shown to be effective in attenuating the deleterious effects of excessive PMNs activation in animal models (Barnard et al., 1995; Bauer et al., 1995).

A number of reports also describe a class of small organic molecules known as leumidins that exhibit significant inhibitory activities in several animal models of PMNs-mediated inflammation. Leumedins seem to elicit pharmacological action by inhibiting cell adhesion processes in PMNs. They have been shown to be beneficial in several animal models such as contact dermatitis, the Arthus reaction, Crohn's disease, colitis, and sepsis (Burch et al., 1992). Although the precise mechanism of inhibition by leumedins is unknown, they do inhibit up-regulation of CD11b/CD18 expression (Bator et al., 1992). At micromolar concentrations, leumedins have also been shown to induce the loss of shedding of L-selectin and block primary capture of neutrophils under flow conditions (Endemann et al., 1997). During activation of PMNs, CR3 translocates from secondary granules to the plasma membranes where it is clustered (Detmers et al., 1987) and present in a high-affinity state for ligand binding and promoting adhesion of PMNs. Thus, preventing this translocation process seems to be the main mechanism by which leumedins prevent inflammation caused by PMNs.

In the present investigation, we have identified two low-molecular-weight compounds that block the binding of CR3 to its ligand C3bi. In contrast to leumedins, compounds 1 and 2 directly inhibit the binding of CR3 with C3bi by binding to CR3 with a reaction likely to involve photoactivation and formation of a covalent adduct. Although these are not desirable properties for a potential therapeutic agent, these compounds nonetheless demonstrate the possibility of blocking CR3 binding with a small molecule, and elucidation of their site of interaction on CR3 may lead to new avenues for achieving this goal with other structures.

In summary, we have identified two compounds that are capable of inhibiting binding of CR3 to C3bi. Limited structure-activity relationship studies on these compounds indicate the importance of the benzothiazole ring, the ethyl side chains, and the length of the carbon chain linking the ring structure. Further modification or substitution at these sites has the potential to prepare better analogs that may show higher affinity for CR3, not only in in vitro assays but also in animal models of inflammation. These studies may help in the design of future anti-inflammatory therapeutic agents that can act as specific therapeutic agents targeting PMNs-mediated inflammation.