Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The initial adhesive interaction of leukocytes with the endothelium is a loose "rolling" adhesion. This "rolling" of the leukocytes along the endothelial surface is mediated by the selectins (Lawrence and Springer, 1991; Abbassi et al., 1993; von Andrian et al., 1991) but has also recently been found to be mediated by 4 integrins (Berlin et al., 1995) and CD44 (DeGrendele et al., 1996). The selectins are a three-member family of cell surface adhesion molecules. L-selectin is expressed on leukocytes, whereas E- and P-selectin are expressed on activated endothelium, and P-selectin is present on the surface of activated platelets (reviewed in Lasky, 1995). These receptors have a similar extracellular domain structure that includes an N-terminal lectin-like domain, an epidermal growth factor-related domain and several complement regulatory protein repeat elements. The lectin domain is analogous to other C-type Ca⁺⁺-dependent lectins, and binding studies have established the structure of the carbohydrate ligands. The selectins are known to recognize sialylated, fucosylated lactosaminoglycans such as sLe^x. When linked to protein, the sLe^x structure is clearly contained in the structure of the carbohydrate ligand for the selectins. However, isolated sLe^x oligosaccharides are low-potency inhibitors of the selectins. The inhibition of binding to the selectins requires near millimolar concentrations of sLe^x analogs for E-selectin and greater than millimolar concentrations for P- and L-selectin (Brandley et al., 1993).

The role of selectins in the development of disease has been well established in a number of animal models. Blocking antibodies, sLe^x oligosaccharides and soluble selectin proteins have been used to prevent disease development. P-selectin-deficient mice show delayed PMN infiltration into thioglycollate-induced peritonitis (Mayadas et al., 1993). Studies in E-selectin-deficient mice show that both E- and P-selectin must be blocked to prevent development of either thioglycollate-induced peritonitis or a cutaneous DTH response (Labow et al., 1994). In a P-selectin-dependent model of cobra venom factor-induced lung inflammation, sLe^x oligosaccharides protected from tissue damage (Mulligan et al., 1993). Selectins clearly play a role in the development of inflammatory disease, and the lectin portion of selectins suggests a role for a carbohydrate-based inhibitor, but sLe^x-based compounds may not be potent enough inhibitors to be drug candidates. Such antagonists are, however, being developed for clinical use (Mousa, 1996).

One ligand for the selectins is sulfatide [Gal(3-SO₄)1-1Cer]. This compound is able to inhibit binding of ligand expressing cells to P- (Aruffo et al., 1991), E- (Nair et al., 1994) and L- (Watson et al., 1990) selectin. The concentration of sulfatide necessary to inhibit P-selectin binding (IC₅₀ = 10-20 µM) distinguishes its potency from that of the sLe^x oligosaccharides. The ability of sulfatide to block the selectins is consistent with its anti-inflammatory activity in vivo. Sulfatide can block the development of the RPA reaction in rats (Nair et al., 1994), as well as cobra venom factor- and IgG-induced rat lung injury (Mulligan et al., 1995) and CCI₄-induced liver inflammation in rats (Kajihara et al., 1995). These findings led to a synthetic effort to identify compounds that were structurally related to sulfatide but had enhanced inhibitory activity. In this study, we present the characterization of BMS-190394, a sulfatide analog with dramatically enhanced in vivo anti-inflammatory properties.

To characterize the efficacy of the compounds tested, we used the dermal RPA reaction and the DTH reaction. The RPA reaction is an example of an induced inflammatory response caused by the deposition of immune complexes, where tissue damage occurs as a result of complement activation, infiltration of PMN and increased vascular permeability (Cochrane and Aikin, 1966). The DTH reaction is a T-cell-dependent reaction that induces marked PMN influx. Infiltrating PMN cause damage to tissue during many inflammatory reactions, including those associated with asthma, allergic rhinitis, rheumatoid arthritis, inflammatory bowel disease and psoriasis (Lewis et al., 1990). At sites of insult-induced tissue damage and injury, the endothelium up-regulates selectin expression, thus facilitating leukocyte adherence. Both the RPA and DTH reactions result in increased PMN accumulation in the tissue because of the transendothelial migration of adherent leukocytes. We have thus tested whether a compound found to have in vitro selectin-inhibiting activity can block the development of an inflammatory reaction that is marked by PMN infiltration.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Male Sprague-Dawley rats (250-300 g b.wt.) with or without jugular vein cannulas are obtained from Hilltop Lab Animal (Scottdale, PA). The rats are housed individually in stainless steel cages and in accordance with NIH guidelines. All animals were given free access to food and water and housed in rooms with 12-h-cycle lighting.

Test compound. BMS-190394 was prepared by Bristol-Myers Squibb (Candiac, Quebec, Canada) as described (Patent No. 95118068.6-2110) and was used as the sodium salt form.

Materials. Rabbit polyclonal IgG rich in antibovine albumin was purchased from Organon Teknika (West Chester, PA). ¹²⁵I-BSA (sp. activity 1-5 mCi/mg) was purchased from Dupont New England Nuclear (Boston, MA). Calcein acetoxymethylester (calcein-AM) was obtained from Molecular Probes (Eugene, OR). Unless otherwise indicated, reagents were from Sigma Chemical Co. (St. Louis, MO).

Cell culture. HL-60 cells obtained from American Type Culture Collection were cultured in RPMI 1640 medium (GIBCO, Grand Island, NY) supplemented with 20% fetal calf serum (Sigma) and 50 µg/ml gentamicin (GIBCO).

Preparation of recombinant globulins. Recombinant fusion proteins composed of extracellular portions of the selectins fused to the human immunoglobulin heavy chain CH3, CH2 and hinge regions were prepared as described previously (Aruffo et al., 1991, Walz et al., 1990, Aruffo et al., 1992). These soluble selectin-immunoglobulin fusion proteins contain the signal sequence, lectin domain and EGF repeat, along with six, two and two of the complement regulatory-like modules for E-, P- and L-selectin, respectively. A plasmid coding for the cDNA coding for each fusion protein was transfected into COS cells. The fusion proteins were purified from the tissue culture supernatant using protein A as described (Aruffo et al., 1990).

HL-60 cell binding assay for P- and E-selectin Rg. The assay for cell binding to immobilized selectin Rg was performed as described (Nair et al., 1994). Briefly, the wells of a 96-well dish (Corning) were coated overnight with anti-human Fc antibody diluted into 50 mM Tris pH 9.1 buffer, blocked with 1% nonfat dry milk in DPBS and allowed to bind selectin Rg. HL-60 cells were labeled with 10 µM calcein for 30 min at 3 × 10⁷ cells/ml at room temperature. When inhibitors were tested, the Rg-coated wells were preincubated at room temperature for 15 min with the inhibitor, and 200,000 cells were added to yield the final indicated inhibitor concentration in 160 µl of DPBS. In each analysis, a vehicle control was utilized to assess the role of DMSO; this control had no significant effect on binding or membrane integrity at the maximal final concentration of 0.5 volume percent. The blocked Rg-bound wells were rinsed twice, and labeled cells were added for 30 min at room temperature. Unbound cells were removed by aspiration and three washes of the wells. Fluorescence in each well was determined using a Millipore Cytofluor fluorescent plate reader.

LS180 cell binding assay for L- and P-selectin Rg. The wells of a 96-well dish were prepared as described above for the HL-60 binding assay using P- and L-selectin Rg. LS180 colon tumor cells, obtained from the ATCC, can bind L- and P-selectin (Nelson et al., 1993). The LS180 cells were used in a manner similar to that described but were labeled with calcein-AM as described for the HL-60 cells. After a 30-min binding period, the unbound cells were washed off as described for the HL-60 cells, and the inhibition of binding was determined by the number of fluorescent cells bound per well.

Preparation of platelets. Blood from normal human donors was anticoagulated with citrate, layered over 1-Step platelets (Accurate Chemical, Westbury, NY) and subjected to centrifugation at 350 × g for 20 min at room temperature. The platelet band was collected, diluted in 2 volumes of Tyrode's salts solution with 5 mM HEPES, 10 mM EDTA and 0.2% BSA (THEB) and centrifuged at 600 × g for 10 min. There were no leukocytes present when the platelet preparation was examined microscopically or when the platelets were analyzed for forward and side scatter parameters via flow cytometry. The platelet pellet was resuspended in THEB and incubated at room temperature for 1 hr. Calcein-AM was added to the platelets at a final concentration of 10 µM and incubated for 10 min at 37°C to label the platelets. Without washing, the platelets were counted on a Coulter counter model ZM, and the concentration was adjusted to 1 × 10⁷/ml. The platelets were activated with 2 U/ml of human thrombin for 10 min in Tyrode's salts containing 2 mM CaCl₂, 5 mM HEPES and 0.2% BSA (THB) at 37°C and immediately fixed with 1% buffered formalin for 1 to 2 hr at room temperature. A small aliquot of labeled platelets was removed before activation and designated as nonactivated.

Platelet:HL-60 cell adhesion assay. Log-phase HL-60 cells were washed, resuspended in THB and fixed with 1% buffered formalin. Both platelets and HL-60 were washed in a 5-fold excess volume of HBSS (GIBCO), resuspended in THB and counted. Cell concentrations were adjusted to 2 × 10⁷/ml for platelets and 4 × 10⁶/ml for HL-60, which was determined to be an optimal ratio for adhesion. Antibodies or compounds were incubated with 50 µl of platelets for 30 min at room temperature before the addition of 50 µl of HL-60. This 5:1 ratio of platelets to HL-60 was incubated for 30 min at room temperature before the addition of 0.2 ml of THB to increase the volume so the samples could be analyzed on a FACScan cytometer (Becton Dickinson, San Jose, CA). Nonactivated platelets, and activated platelets with 10 mM EDTA were included as controls. Data were collected within a region set for the forward scatter channel corresponding to HL-60 size events. The % HL-60 containing bound platelets was calculated from FL-1 histograms where two separate peaks represented HL-60 alone and HL-60 with bound platelets.

Inhibitor of cell binding. BMS-190394 was prepared by dissolution to a final concentration of 2 mg/ml in deionized water by sonication in a bath sonicator for 5 minutes, heating to 90°C for 10 min, sonicating for an additional 10 min, again heating to 90°C and slow cooling (in the heating block).

Pharmacokinetics. BMS-190394 was dosed i.v. at 1.0 mg/kg and i.p. at 5.0 mg/kg in 0.5% Tween 80 in PBS adjusted to pH 8.0. Blood samples were taken at selected times until 24 hr after dosing. Plasma was isolated immediately, and 50 µl was mixed with 4 µl 2 N HCl and 200 µl acetonitrile, vortexed and centrifuged. The supernatant was analyzed by reverse-phase HPLC.

RPA reaction in rat skin. The rats were anesthetized with ketamine/rompun (100 mg/12 mg/300 g b.w.t.) given i.p. The anesthetized rats were injected intradermally with anti-BSA (0.6 mg/site) over the previously clipped dorsal skin and dosed via the jugular vein with BSA (10 mg) containing 1 µCi of ¹²⁵I-BSA. BMS-190394 was dosed i.v. and i.p. in a vehicle of 0.5% Tween 80 in phosphate buffer (0.05 M), selected on the basis of the solubility of the compound and reported clinical experience with the vehicle.

DTH reaction in rat. The DTH reaction is induced by topical application of oxazolone. On day 0, an inductive dose of 4% oxazolone was applied to the shaved belly of the rat. Four days later, a second inductive dose of 2% oxazolone was applied to the belly. Six days after the first induction, a challenge dose of 1% oxazolone was applied to the ears. BMS-190394 or vehicle was administered i.p. immediately after the challenge dose of oxazolone. Twenty-four hours later, the rats were sacrificed, and punch biopsies of the ears were removed, weighed and processed for MPO activity as described above.

Statistical analysis. Significant differences were determined by paired Student's t test. Statistical significance was defined as P < 0.05. All values are expressed as mean ± S.D.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

BMS-190394 inhibits HL-60 cell binding to P- and E-selectin Rg. BMS-190394 (fig. 1) is a sulfatide analog designed to inhibit in vitro selectin binding as well as in vivo anti-inflammatory activity. As an in vitro model of the selectin-ligand interaction, we have employed the promyelocytic leukemia cell line HL-60 that expresses the ligands for P- and E-selectins. These cells bind to surfaces coated with P-selectin Rg or E-selectin Rg. The inhibition curve (fig. 2) shows that BMS-190394 inhibits the binding of HL-60 cells to each of these receptors in a dose-dependent manner. Galactosylceramide, a nonsulfated version of sulfatide, did not inhibit the binding of the HL-60 cells to P- or E-selectin, as previously shown (Nair et al., 1994). The dose-inhibition curves for E- and P-selectin show nearly equipotent inhibition of each selectin. These curves yield IC₅₀ values of 9.5 and 11 µM for P-selectin and E-selectin, respectively. The P- and E-selectin IC₅₀ values ranged from 9 to 20 µM in the course of several experiments, but there was no significant difference between the inhibition of P-selectin and that of E-selectin in any experiment. In comparison, sulfatide inhibited the binding with an IC₅₀ value of 16 µM (Nair et al., 1994).

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Structure of BMS-190394. The O-Bz represents the O-benzoyl group.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Inhibition of HL-60 cells binding to immobilized P- (closed symbols) and E- (open symbols) selectin Rg. After preincubation of the P- and E-selectin Rg-coated 96-well plate with BMS-190394 (circles) or galactosylceramide (triangles), HL-60 cells were added and incubated for 30 min. Percent inhibition of binding of the HL-60 cells relative to untreated wells is shown for each of the indicated concentrations of the added compound. In the absence of the compound, the P- and E-selectin wells bound 110,000 and 97,000 cells respectively, whereas wells containing no selectin Rg bound 3400 cells. The data points represent mean ± S.D. of values from an experiment that is representative of three experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Inhibition of LS180 cell binding to L-selectin Rg (closed symbols) and P-selectin Rg (open symbols) by BMS-190394. Rg binding was performed as described in the legend to figure 2, using LS180 cells. In the absence of compound, L-selectin Rg bound 11,000 cells, and P-selectin Rg bound 5200 cells, whereas wells containing no Rg bound 2000 cells.

BMS-190394 inhibits cell surface P-selectin. After activation with thrombin, P-selectin is expressed on the surface of platelets. We exploited this characteristic in an assay that can address the cross-species activity of P-selectin inhibitors. This assay utilizes flow cytometry to monitor the binding between fluorescent-labeled platelets and HL-60 cells. As shown in figure 4, BMS-190394 inhibits human platelet binding to HL-60 cells in a dose-dependent manner with an IC₅₀ of 28 ± 10 µM for the combined data from four experiments, whereas galactosylceramide had essentially no effect. The compound inhibited rat platelet binding to HL-60 cells with an IC₅₀ of 38 µM, which suggests potency similar to that in human platelets. The blocking activity of BMS-190394 in this assay indicates that this compound inhibits naturally expressed P-selectin, not just the truncated fusion protein presented in figure 2. Because the platelet P-selectin and the endothelial P-selectin are products of the same gene, the primary structure of the binding domains are identical, and the binding characteristics are likely to be similar.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Inhibition of HL-60 cell binding to platelets by BMS-190394 or galactosylceramide. Platelets were isolated from humans and rats and labeled by uptake of a fluorescent reagent. After preincubation with the indicated concentration of BMS-190394 (circles) or galactosylceramide (triangles), the platelets were allowed to adhere to HL-60 cells, as described in the text. The numbers of HL-60 cells containing bound human (closed symbols) and rat (open symbols) platelets were 2861 and 3349, respectively. These values were compared with the numbers of HL-60 cells with adherent platelets in the presence of EDTA (440 and 425, respectively) and used to calculate the percent inhibition of platelet binding. These data are typical of at least three separate determinations.

Phamacokinetics of BMS-190394. The time course of disappearance of BMS-190394 from the circulation of the rat was determined by an RP-HPLC procedure. Figure 5 shows the blood concentration of BMS-190394 after i.v. administration. After i.v. dosing with 1 mg/kg, the elimination half-life was found to be approximately 7 hr. When rats were dosed i.p. with 5 mg/kg of BMS-190394, significant absorption to the systemic circulation was observed. After i.p. dosing, the plasma level of BMS-190394 was found to reach 17 µg/ml in 4 hr, and detectable levels were still found at 24 hr. BMS-190394 was not detectable in the plasma after p.o. dosing.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Pharmacokinetic profile of BMS-190394 after dosing in the rat. The plasma concentrations were determined as described in the text, after i.v. dosing at 1.0 mg/kg (triangles) or i.p. dosing at 5.0 mg/kg (circles).

Activity of BMS-190394 in the RPA reaction. The RPA reaction is initiated by the formation of immune complexes between the circulating BSA and the anti-BSA antibody instilled in the skin. This reaction is marked by leukocyte infiltration and increased VP. After enough time for the inflammation to develop, skin punches taken from the site of the immune complex deposition were used for analysis of two endpoints to determine the magnitude of the response. The efficacy of the drug is reflected by reduction in accumulation of ¹²⁵I-BSA from the circulation in the skin punch and the MPO levels. The activity is reported as the ED₅₀, the concentration necessary to block the change in VP or MPO by 50%. The pharmacokinetic data showed that the half-life of the drug was 7 hr, longer than the time for development of the peak RPA response, when the drug was delivered i.v. and that the drug did not undergo significant metabolism. The target of the drug is within the vasculature, on the endothelium and on the platelets, so the drug was tested via i.v. delivery. BMS-190394 elicited a dose-dependent inhibition of VP and MPO, as shown in figure 6. The ED₅₀ values after i.v. administration were 0.2 mg/kg and 0.04 mg/kg, respectively, for VP and MPO.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Dose-dependent inhibition of the RPA reaction by BMS-190394 after i.v. () or i.p. () administration. The compound was administered by the indicated route at the time of dermal injection. Control animals received an injection of the vehicle by the indicated route. The parameters were measured 4 hr after initiation of the reaction. The values represent the mean inhibition relative to the vehicle of eight animals per treatment condition. Inhibition of the vascular permeability (top panel) was relative to the control values of 145 µl of plasma (no RPA) and 677 µl (RPA reaction) for the i.p. treatment and 602 µl (no RPA) and 1122 µl (RPA reaction) for the i.v. treatment. Inhibition of the cell infiltration (as reflected by MPO activity, bottom panel) were performed as described in the text and are calculated from the control values of 3 mO.D. units/min/biopsy (no RPA) vs. 2825 mO.D./min/biopsy (RPA reaction) via the i.p. route and 5.8 mO.D./min/biopsy (no RPA) vs. 1621 mO.D./min/biopsy (RPA reaction) for the i.v. treatment route. Inhibition relative to vehicle-treated animals that was statistically significant (P < .05) is indicated (*).

DTH reaction. The DTH response is a T-cell-dependent model of allergic contact dermatitis. After sensitization, this response is induced by a challenge dose of oxazolone to the ear. This model is highly responsive to corticosteroid treatment but not so responsive to nonsteroidal therapy (Cavey et al., 1990; Chapman et al., 1986). The DTH reaction results in a marked neutrophil infiltration into the site of challenge on the ear 24 hr after the challenge. BMS-190394 was administered i.p. immediately after the challenge dose of oxazolone. Figure 7 shows that BMS-190394 blocks the oxazolone-induced DTH response in rats in a dose-dependent manner. The ED₅₀ value for blockade of PMN infiltration was approximately 2.8 mg/kg. Dexamethasone was used as a positive control in the DTH reaction; when administered i.v. at 3.3 mg/kg, it inhibited the development of the DTH reaction. The increase in PMN influx was inhibited 99.5% by dexamethasone. In data not shown, histology sections taken from the inflammatory site were examined for the presence of leukocytes. Histological analysis confirmed that treatment with BMS-190394 or dexamethasone blocked the infiltration of all cell types, not just PMN.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Effect of BMS-190394 on the oxazolone-induced DTH reaction. BMS-190394 was injected i.p. at the time of administration of the challenge dose of oxazolone on the ears of sensitized rats. Control animals were injected with the vehicle at the time of the challenge. The endpoint was measured as described in the text 24 hr after the challenge dose of oxazolone was applied. The values are the mean of eight animals at each dosing amount. The percent inhibition was determined by comparing the noninflamed animals' MPO values of 2.0 ± 0.2 mO.D./min/biopsy and the vehicle-treated DTH reaction animals' MPO value of 637 ± 285, as described in "Materials and Methods." Each of the data points represents a significant difference (P < 0.05) between the test group and the non-treated group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Several types of selectin inhibitors block the development of inflammation in vivo. Examples include the use of antibodies against E-selectin to block the development of skin and lung inflammatory reactions (Mulligan et al., 1991), and sLe^x, which has been shown to block the development of lung inflammatory reactions (Mulligan et al., 1993). In addition to the in vivo studies on sLe^x (Lasky, 1995) and sulfatide (Nair et al., 1994; Mulligan et al., 1995; Kajihara et al., 1995), some lower-weight compounds such as glycyrrhizin (Rao et al., 1994) and inositol polyphosphates (Cecconi et al., 1994) have been reported to inhibit selectin-dependent processes. sLe^x, glycyrrhizin and the inositol phosphates have, in general, required near millimolar concentrations to achieve inhibition in vitro.

The in vitro and in vivo activity of sulfatide suggested that structural analogs could be devised with the goal of improved selectin-blocking and anti-inflammatory activity. The outcome of this sulfatide analog synthetic program was the synthesis of BMS-190394. BMS-190394 differs from sulfatide in the nature and position of the anionic group attached to the galactose; a single sulfate group is present in sulfatide and two carboxylate groups at 4- and 6- in BMS-190394. In addition, the two remaining hydroxyl groups of galactose and the hydroxyl of the aglycone have been acylated with benzoyl substituents.

In this study, we used in vitro assays to show that BMS-190394 is an inhibitor of P-, E- and L-selectin-dependent adhesion. Inhibition of the binding to both P- and E-selectin is significant because both are expressed on the endothelium in response to proinflammatory stimuli and can initiate the process of leukocyte adherence to endothelial surfaces. The leukocyte cell surface receptor L-selectin mediates binding to the endothelium, and BMS-190394 is capable of blocking this interaction. The binding of cells to platelets provides a selectin adhesion assay with an added level of complexity compared with those that use isolated selectin proteins. This assay relies on the platelet surface expression of P-selectin, a protein in its native state that has been fully processed and glycosylated. A further advantage of the platelet binding assay is that the binding is between a cell and platelets in suspension, not to an immobilized protein, and this allows for a more dynamic interaction, and facilitating the potential interaction of multiple selectin:ligand pairs. The platelet:HL-60 cell assay has provided a means of comparing the binding activity of P-selectin from nonhuman test species. BMS-190394 inhibits binding of both rat and human platelet P-selectin.

We extended these in vitro findings to address the ability of BMS-190394 to block inflammatory reactions in vivo. BMS-190394 inhibited the rat RPA reaction when administered through either the i.v. or the i.p. route. BMS-190394 had potent activity against both endpoints tested: VP and PMN infiltration. Comparison of the ED₅₀ values for i.v. doses suggests that BMS-190394 is as potent as dexamethasone, a commonly used anti-inflammatory therapeutic. In addition, in i.p. comparisons, BMS-190394 is far more potent than sulfatide, because sulfatide was unable to inhibit the VP significantly and had an ED₅₀ of 50 mg/kg for MPO (Nair et al., 1994), compared with ED₅₀ values of 0.6 and 0.1 mg/kg for BMS-190394 VP and MPO inhibition, respectively. BMS-190394 showed no activity when administered p.o., a result consistent with its lack of oral bioavailability. Thus a compound that inhibits P-, E- and L-selectin is capable of blocking the development of an inflammatory reaction in the rat. This is consistent with a mechanism of action in which the compound acts by blocking the interaction of leukocytes with endothelium in a selectin-dependent manner. Blocking the selectins prevents the initial step in PMN infiltration, thereby preventing subsequent damage to the tissue. Our data are consistent with a primary role for the selectins in dermal immune complex-induced responses in the rat. PMN play a strong role in the development of the models used, and a reduction in the number of circulating PMN could elicit a reduction in the magnitude of an inflammatory response. However, BMS-190394 did not affect the number of circulating PMN.

The DTH response is a T-cell-dependent inflammatory reaction. Administration of BMS-190394 blocked the infiltration of PMN induced by the DTH reaction, as would be expected of an agent that interferes with the initial interaction of the leukocyte with the endothelium. A role for E-selectin in a monkey DTH response has been demonstrated through the use of blocking antibodies (Silber et al., 1994). In E-selectin-deficient mice, elimination of both E- and P-selectin binding was necessary to block the development of a cutaneous DTH response (Labow et al., 1994). In the present study, we have demonstrated the effectiveness of a low-molecular-weight inhibitor of all three selectins in blocking the rat DTH response.

The RPA and DTH models require 4 and 24 hr, respectively, to develop. In spite of the different time courses of development of the two inflammatory reactions, BMS-190394 was capable of blocking the inflammatory response in each. This is consistent with the pharmacokinetic data showing that BMS-190394 has a relatively long half-life and resides primarily in the circulation.

The two models presented in this study, the RPA and the DTH, are both characterized by the infiltration of leukocytes as one of the primary steps in the development of the inflammatory response. In each of these models, the selectin antagonist BMS-190394 was able to block the development of the infiltration of PMN and the concomitant tissue edema. This finding is consistent with other models in which antibody blockade has shown that selectins play a primary role in the development of the inflammatory response (reviewed in Lasky, 1995). The results of this study support the role of selectins in mounting responses to immune complex-induced inflammation in rodents. This study also shows that a low-molecular-weight selectin inhibitor drug candidate can act by a specifically targeted mechanism to block inflammation.