Abstract
Tumor necrosis factor-α (TNF-α), a cytokine secreted by activated monocytes/macrophages and T lymphocytes, has been implicated in several disease states, including rheumatoid arthritis, inflammatory bowel disease, septic shock, and osteoporosis. Monocyte/macrophage production of TNF-α is dependent on the mitogen-activated protein kinase p38. RWJ 67657 (4-[4-(4-fluorophenyl)-1-(3-phenylpropyl)-5-(4-pyridinyl)-1H-imidazol-2-yl]-3-butyn-1-ol) inhibited the release of TNF-α by lipopolysaccharide (a monocyte stimulus)-treated human peripheral blood mononuclear cells with an IC50 of 3 nM, as well as the release of TNF-α from peripheral blood mononuclear cells treated with the superantigen staphylococcal enterotoxin B (a T cell stimulus), with an IC50 value of 13 nM. This compound was approximately 10-fold more potent than the literature standard p38 kinase inhibitor SB 203580 in all p38 dependent in vitro systems tested. RWJ 67657 inhibited the enzymatic activity of recombinant p38α and β, but not γ or δ, in vitro and had no significant activity against a variety of other enzymes. In contrast, SB 203580 significantly inhibited the tyrosine kinases p56 lck and c-src(IC50 = 5 μM). RWJ 67657 did not inhibit T cell production of interleukin-2 or interferon-γ and did not inhibit T cell proliferation in response to mitogens. RWJ 67657 inhibited TNF-α production in lipopolysaccharide-injected mice (87% inhibition at 50 mg/kg) and in rats (91% inhibition at 25 mg/kg) after oral administration. Based on these favorable biological properties, RWJ 67657 may have use as a treatment for inflammatory diseases.
Tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) are proinflammatory cytokines involved in the pathogenesis of rheumatoid arthritis, endotoxin-induced shock, inflammatory bowel disease, and osteoporosis, among others (Martin and Resch, 1988; Dinarello, 1991). The biosynthesis of these two cytokines is positively regulated at the transcriptional (Baldassare et al., 1999) and translational level by the stress-activated serine-threonine kinase p38 (Lee et al., 1994), making this kinase a potentially attractive target for treatment of the above diseases. The compounds SB 203580 (Badger et al., 1996) and SB 220025 (Jackson et al., 1998; Fig. 1) have been shown to be potent inhibitors of p38 in vitro and to produce significant anti-inflammatory effects in various animal models. In addition to blocking TNF-α and IL-1β production in various systems, inhibitors of p38 kinase activity also block IL-1α-stimulated nitric oxide production in chondrocytes (Badger et al., 1998) and IL-1-induced cyclooxygenase-2 (COX-2) and matrix metalloproteinases 1 and 3 mRNA in fibroblasts (Ridley et al., 1997) and decrease the stability of COX-2 mRNA in activated monocytes (Dean et al., 1999). These and other studies show that p38 inhibitors can inhibit a wide spectrum of anti-inflammatory activities.
Recent data suggest that p38 inhibitors may have therapeutic use beyond their well known anti-inflammatory properties. Several transgenic and knockout mouse models, as well as in vitro studies using SB 203580, strongly implicate p38 in the regulation of IL-12 and interferon-γ gene transcription (Rincon et al., 1998; Lu et al., 1999) and hence in the regulation of T helper cell 1 (Th1)-type immune responses. In contrast, in studies of human T cells, p38 inhibitors potently and selectively blocked IL-4 production, suggesting a predominant role in regulation of Th2-type immunity (Schafer et al., 1999a). Another study of human cells found evidence to support the involvement of p38 in regulation of both Th1 and Th2 responses, as well as IL-10 production (Koprak et al., 1999). p38 inhibitors can block replication of human immunodeficiency virus type 1 in primary human T cells (Cohen et al., 1997) and IL-1-induced human immunodeficiency virus type 1 replication in a macrophage cell line (Shapiro et al., 1998). Evidence also suggests that SB 203580 may decrease myocardial TNF-α production and cardiomyocyte apoptosis, while increasing myocardial function after ischemia and reperfusion (Cain et al., 1999; Ma et al., 1999). The potential uses of p38 inhibitors extend to any condition in which TNF-α or IL-1β plays a role in disease pathology.
The crystal structure of SB 203580 bound to inactive p38α has been solved (Tong et al., 1997) and shows that the compound binds in the ATP-binding pocket, consistent with biochemical data suggesting that the compounds compete with ATP for binding to the enzyme (Griswold and Young, 1996). SB 203580 is known to inhibit the p38α and β isoforms but not the γ and δ isoforms (Kumar et al., 1997). This selectivity is due largely to the presence of threonine at position 106 in p38α and β but also involves His107 and Leu108 (Gum et al., 1998). Additional data suggest that this class of compounds binds equally well to the activated (phosphorylated on Thr180 and Tyr182) and nonactivated (monophosphorylated or nonphosphorylated) p38 enzyme, whereas ATP binds only to the activated form (Frantz et al., 1998). This property has led to the proposal that these compounds have a significant advantage in cells because the high local concentration of ATP (millimolar) will not compete with compound for binding to inactive enzyme. This phenomenon and the known anti-inflammatory and antiangiogenic properties of various p38 inhibitors in vivo (Jackson et al., 1998) suggest that inhibitors of p38 kinase may have potential as human therapeutics.
We report here the biological evaluation in vitro and in vivo of the novel p38 kinase inhibitor RWJ 67657 (Fig.2). RWJ 67657 inhibited the release of TNF-α by lipopolysaccharide (LPS)-stimulated human peripheral blood mononuclear cells (PBMCs) with an IC50 value of 3 nM and inhibited the release of TNF-α from PBMCs stimulated with the superantigen staphylococcal enterotoxin B (SEB; a T cell stimulus), with an IC50 value of 13 nM. This compound was approximately 10-fold more potent than SB 203580 in in vitro systems. RWJ 67657 inhibited the enzymatic activity of recombinant p38α and β, but not γ or δ, in vitro, with no significant activity against a variety of other enzymes. In contrast, SB 203580 exhibited significant activity against the tyrosine kinases p56 lckand c-src (IC50 = 5 μM). RWJ 67657 did not inhibit T cell production of the cytokines IL-2 or interferon-γ and did not inhibit T cell proliferation in response to mitogens. RWJ 67657 inhibited TNF-α production in LPS-injected mice (87% inhibition at 50 mg/kg) and in rats (91% inhibition at 25 mg/kg) after oral administration. Based on these favorable biological properties, RWJ 67657 may have use as a potential treatment for inflammatory diseases.
Experimental Procedures
Materials.
RWJ 67657 (4-[4-(4-fluorophenyl)-1-(3-phenylpropyl)-5-(4-pyridinyl)-1H-imidazol-2-yl]-3-butyn-1-ol) was synthesized at The R. W. Johnson Pharmaceutical Research Institute according to methods described in world patent application WO 9847892. SB 203580 [4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridinyl)imidazole] was purchased from Calbiochem (La Jolla, CA).
p38 Cellular Assay.
Whole blood from healthy human donors or beagle dogs was obtained by venipuncture. Rat blood was obtained via cardiac puncture using anesthetized male Lewis rats (200–300 g; Charles River Laboratories, Wilmington, MA). All blood was collected into heparinized vacuum tubes and centrifuged at 600gfor 10 min at room temperature. Buffy coats were harvested, diluted to 15 ml with PBS, underlaid with 15 ml of Accu-Paque (Accurate Chemical and Scientific Corp., Westbury, NY), and centrifuged at 1100g for 20 min at room temperature. PBMCs were harvested, washed twice with Hanks’ balanced salt solution (HBSS; Life Technologies, Gaithersburg, MD), and resuspended in low endotoxin RPMI 1640 culture medium (Sigma Chemical Co., St. Louis, MO) containing 1% FCS and 1× penicillin-streptomycin-glutamine (Life Technologies) at 1.67 × 106 cells/ml. Cells (180 μl) were added to duplicate wells of a flat-bottomed 96-well plate and allowed to settle for 1 h at 37°C. Test compounds (10 μl) or vehicle [2% dimethyl sulfoxide (DMSO)] was added to each well, and the plate was incubated for 1 h at 37°C. Finally, 10 μl/well LPS (200 ng/ml; Sigma Chemical Co.) was added, for a final concentration of 10 ng/ml. Plates were incubated overnight at 37°C and 5% CO2. Supernatants were harvested, diluted 1:5, and assayed for TNF-α by enzyme-linked immunosorbent assay (ELISA) as described by the manufacturers of the kits (Genzyme, Cambridge, MA, for human TNF-α; Biosource Intl., Camarillo, CA, for rat TNF-α). An ELISA for dog TNF-α was developed in our laboratory, using a mouse-anti-human TNF-α primary antibody and a rabbit-anti-human TNF-α secondary antibody, both of which cross-react with dog TNF-α (Research Diagnostics, Inc., Flanders, NJ). For T cell stimulation, SEB (Sigma Chemical Co.) was added to PBMCs at a final concentration of 100 ng/ml; other conditions were the same as for LPS treatment. The human IL-2 and IL-1β ELISA kits were from Genzyme. The IFN-γ ELISA kit was from Endogen (Woburn, MA).
Proliferation Assay.
T cells were purified from human PBMCs by negative selection as previously described (Schafer et al., 1999b). CD8+ T cells were magnetically immunodepleted using anti-CD8 Dynabeads M-450 (Dynal, Lake Success, NY) according to the manufacturer’s instructions. Unbound cells were washed and determined to be >98% CD3+CD4+CD28+ using a FACSort flow cytometer (Becton Dickinson, Mountain View, CA). T cells were plated in flat-bottomed 96-well tissue culture plates in complete medium (RPMI containing 10% FCS, 50 U/ml penicillin G, 50 μg/ml streptomycin, and 2 mM glutamine) at 2 × 105cells/well. SB 203580 or RWJ 67657 was serially diluted in complete medium, at a constant final DMSO concentration, and added to cells for a 1-h pretreatment at 37°C and 5% CO2 (final volume, 0.2 ml/well). OKT3 (anti-CD3ε, 1 μg/ml), IgG2aκ (1 μg/ml), CD28.2 (anti-CD28, 10 μg/ml), or IgG1κ (10 μg/ml) was added, maintaining a total IgG concentration in all samples at 11 μg/ml, followed by F(ab′)2 goat anti-mouse IgG (30 μg/ml). Cells were cultured at 37°C and 5% CO2 for 3 days. Proliferation was measured by culturing cells with [3H]thymidine (1 μCi/well; Amersham, Arlington Heights, IL) for 18 h, from day 3 to day 4. The 3H-labeled cells were harvested onto filter mats and counted in a 1205 Betaplate liquid scintillation counter (Wallac, Gaithersburg, MD).
Purification of Adherent Monocytes.
Human peripheral blood leukocyte preparations (Gulf Coast Regional Blood Center, Houston, TX) were underlaid with Accu-Paque (Accurate Chemical and Scientific) and centrifuged at 1100g for 20 min at room temperature. PBMCs were harvested from the interface, washed twice with HBSS (Life Technologies), resuspended in RPMI 1640 (Sigma Chemical Co.)/10% FBS (Hyclone, Logan, UT)/1× penicillin-streptomycin-glutamine (Life Technologies) at 1 × 107 cells/ml, and incubated in 6-well plastic tissue culture plates at 4 ml/well for 1 h at 37°C and 5% CO2 to allow monocytes to adhere. Nonadherent cells were removed by washing wells twice with RPMI 1640. Adherent monocytes constituted approximately 10% of unseparated PBMCs.
Immune Complex Kinase Assays.
Monocytes were cultured overnight in RPMI 1640/1% FBS/1× penicillin-streptomycin-glutamine, stimulated with LPS at 10 ng/ml for 15 min at 37°C, and lysed in 400 μl of Nonidet P-40 (NP-40) lysis buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 1% NP-40, and 1 mM Na3VO4) containing 1× EDTA-free complete protease inhibitor cocktail (Boehringer Mannheim, Indianapolis, IN) for 1 h on ice. Lysates were centrifuged at 16,000g for 10 min and precleared for 30 min with 50 μl of slurry of GammaBind G Sepharose (Pharmacia Biotech, Uppsala, Sweden) prebound with rabbit IgG (Jackson ImmunoResearch, West Grove, PA) at 5 mg/ml slurry. Lysates were precleared a second time with 50 μl of GammaBind G Sepharose slurry and immunoprecipitated for 1 h with 50 μl of GammaBind G Sepharose slurry prebound with either 2 μg of anti-p38 polyclonal rabbit antibody, specific for the 20 C-terminal residues of human p38α (Santa Cruz Biotechnology, Santa Cruz, CA), or 3 μg of anti-MAP kinase-activated protein kinase-2 (MAPKAPK-2) polyclonal sheep antibody, specific for residues 310 to 325 of MAPKAPK-2 (Upstate Biotechnology, Lake Placid, NY). Immunoprecipitates were washed twice with NP-40 lysis buffer and twice with kinase reaction buffer (25 mM HEPES, pH 7.5, 10 mM MgCl2, 10 mM MnCl2, 20 mM β-glycerophosphate, and 1× EDTA-free complete protease inhibitor cocktail). Kinase reactions were performed in kinase reaction buffer containing 90 μM ATP and 5 μCi of [γ-32P]ATP (3000 Ci/mmol; Amersham, Arlington Heights, IL) with 0.5 μg of glutathione-S-transferase (GST)-MAPKAPK-2 (Upstate Biotechnology) as p38 substrate or 1 μg of heat shock protein 27 (StressGen, Victoria, British Columbia, Canada) as MAPKAPK-2 substrate. After 20 min at 30°C, reactions were stopped by the addition of 2× SDS sample buffer (Novex, San Diego, CA) containing 10% 2-mercaptoethanol and boiling for 5 min. Samples were electrophoresed in 10% or 16% Tris-glycine polyacrylamide gels and transferred to polyvinylidene difluoride membranes (Novex). Membranes were exposed to a Storage Phosphor Screen and analyzed on a Storm 840 PhosphorImager System using ImageQuant software (Molecular Dynamics, Sunnyvale, CA). MAPKAPK-2 activity was measured from monocytes pretreated with RWJ 67657 or SB 203580 in 0.1% DMSO for 1 h before LPS stimulation. p38 activity was measured from monocytes stimulated without pretreatment but with the addition of RWJ 67657 or SB 203580 in 0.1% DMSO to immunoprecipitated p38 enzyme for 5 min before the addition of ATP.
p38 and Extracellular Signal Receptor-Activated Kinase 2 (ERK-2) Assays.
Recombinant p38α, β, γ, and δ GST-fusion proteins were purchased from Upstate Biotechnology. Activated MAP kinase ERK-2 was purchased from Stratagene (La Jolla, CA). Either 30 ng of p38 or 10 ng of ERK-2 was incubated in kinase reaction buffer (25 mM HEPES, pH 7.5, 10 mM MgCl2, 10 mM MnCl2) containing 50 μM ATP, with 30 μg of myelin basic protein as substrate (Life Technologies) and 1 μCi of [γ-33P]ATP (3000 Ci/mmol; Amersham Life Science), with or without test compounds or vehicle (DMSO, 2% final concentration), in a total volume of 60 μl, in a round-bottomed polypropylene 96-well plate. After 30 min at 30°C, reactions were stopped and proteins were precipitated by the addition of 60 μl/well of 50% trichloroacetic acid, and the precipitates were transferred to a 96-well Durapore membrane filterplate (Millipore, Bedford, MA). Wells were filtered using a Millipore vacuum manifold, washed 5× with 200 μl/well of 10% trichloroacetic acid/10 mM sodium phosphate, and briefly air dried. Microscint-20 scintillant (30 μl/well; Packard, Meriden, CT) was added, and the plate was sealed with plastic film (Packard) and counted in a Packard TopCount microplate scintillation counter.
c-Src and Lck Assays.
Recombinant 6xHis-tagged Lck was produced in Escherichia coli in our laboratory and purified from soluble cell extracts by Talon metal affinity chromatography (Clontech Laboratories, Palo Alto, CA). Recombinant c-Src was from Upstate Biotechnology. p34 cdc-2 substrate peptide was synthesized in-house and conjugated to biotin at the N terminus using a double 6-aminocaproic acid linkage. Enzyme was diluted into reaction buffer (62.5 mM HEPES, pH 7.5, and 12.5 mM MgCl2) with 0.6% ovalbumin. Substrate was diluted into reaction buffer to 10 μg/ml, with 125 μM γ-ATP and 2.5 mM sodium orthovanadate. The kinase reaction was initiated by mixing (in order) in a U-bottomed polypropylene test plate: 10 μl of test compound diluted in 10% DMSO, 20 μl of enzyme, and 20 μl of substrate ATP. After 30 min at room temperature, reactions were stopped by the addition of 15 μl of 100 mM EDTA. Aliquots of 50 μl were transferred to an ELISA plate previously coated with NeutrAvidin (Pierce Chemical Co., Rockford, IL). Binding of the biotinylated substrate occurred over 15 min at room temperature, after which wells were washed three times with 200 μl of PBS-0.05% Tween 20. Optimally diluted anti-phosphotyrosine monoclonal antibody (50 μl/well of RC20 conjugated to horseradish peroxidase; Transduction Laboratories, Lexington, KY) was added and incubated 30 min at room temperature, followed by three washes with PBS-Tween 20. TMB solution (100 μl/well of 3,3′,5,5"-tetramethyl benzidine with H2O2; Sigma Chemical Co.) was added for 30 min at room temperature in the dark. Reactions were stopped with 50 μl/well of 1 M H2SO4 and absorbance was read at 450 nm, using a Molecular Devices microplate reader (Menlo Park, CA). Uninhibited enzyme activity was determined by mixing enzyme and substrate in the presence of 10% DMSO. Nonspecific background was determined in the presence of substrate and DMSO alone. Specific inhibition by test compounds was calculated as:
PKA Enzyme Assay.
The catalytic subunit of PKA (Pierce Chemical Co.) was diluted to 5 U/ml in 25 mM Tris, pH 7.6, 150 mM NaCl, and 50% glycerol. Kemptide substrate peptide (Pierce Chemical Co.) was diluted to 1 mg/1.7 ml in water and mixed with equal parts reaction buffer (50 mM MgCl2, 0.01% Triton X-100, 100 mM Tris, pH 7.4), 250 mM nonradioactive ATP, and 5 μCi/ml [γ-33P]ATP (3000 Ci/mmol; Amersham Life Science). In a 96-well U-bottomed polypropylene plate, the following were mixed, in order, and incubated for 30 min at room temperature: 10 μl of inhibitor or vehicle (10% DMSO), 20 μl of diluted enzyme, and 20 μl of substrate/ATP. Reactions were stopped with 70 μl of 75 mM H3PO4, and 100-μl aliquots were transferred to a 96-well phosphocellulose filter plate (Millipore). Wells were washed four times with 75 mM H3PO4 using a Millipore vacuum manifold. Microscint-20 (30 μl/well; Packard) was added. The plate was sealed with plastic film and counted in a Packard TopCount microplate scintillation counter. Kinase activity was calculated by subtracting cpm obtained in the absence of enzyme from cpm obtained in the presence of enzyme. The ability of a test compound to inhibit this activity was determined as:
Thromboxane Assay for COX-1 Activity.
Fresh plasma from healthy donors was used as a source of platelets. Plasma was diluted 1:20 in 0.9% NaCl containing 14.4 U/ml heparin, and 180 μl/well was added to round-bottomed 96-well plates. After 5 min at 37°C, 10 μl of test compounds or vehicle (2% DMSO) was added (final 0.1% DMSO). After 15 min at 37°C, 10 μl/well of the calcium ionophore A23187 (Calbiochem, San Diego, CA) was added (A23187 was initially dissolved at 7 mg/ml in DMSO and diluted 1:25 in HBSS just before use). After incubation for 15 min at 37°C, the reactions were stopped by placing the plate on ice for 5 min. Samples were diluted 1:10 with HBSS and assayed immediately for thromboxane B2 using a fluorescent enzyme immunoassay kit according to the manufacturer’s instructions (PerSeptive Biosystems, Framingham, MA).
Inhibition of TNF-α Release in Mice.
Groups of five 6- to 8-week-old female BALB/cJ mice were fasted for 4 to 6 h. The mice were then injected orally with 200 μl of the chosen compounds (∼10 ml/kg), diluted in 0.01 to 0.1 N HCl, as required for solubilization. A negative control group of mice was injected with HCl alone. At 30 min after oral dosing, each mouse was injected i.p. with 400 μl ofE. coli LPS (Sigma Chemical Co.) at a concentration of 50 μg/ml in saline; because BALB/cJ female mice at this age weigh approximately 20 g, the dose of LPS was about 1 mg/kg. At 1 h later, after sacrifice by CO2 inhalation, blood was obtained by cardiac puncture. Serum was prepared, aliquoted, and frozen at −80°C. Freshly thawed samples were typically tested within 1 to 2 days by a commercially available ELISA specific for mouse TNF-α (Endogen, Woburn, MA). Using this protocol, 1 to 4 ng/ml TNF-α was found in sera from negative control animals. Mice injected i.p. with saline alone, instead of LPS, did not produce detectable levels of TNF-α. Serum TNF-α levels for compound-treated animals were compared with those for vehicle-treated animals by use of the Student’s t test.
Inhibition of TNF-α Release in Rats.
Groups of five fasted male Lewis rats (200–300 g; Charles River Laboratories) were dosed orally with compound (5 ml/kg). Thirty minutes later, each rat was injected i.p. with 2 ml of saline containing LPS such that the final dose was 1 mg/kg. At 1 h later, the rats were sacrificed and blood was obtained by cardiac puncture. Levels of TNF-α were determined by ELISA (Biosource Intl.). At this dose of LPS, 10 to 30 ng/ml TNF-α was found in the sera from control rats dosed orally with HCl only. Student’s t test was used to compare serum TNF-α levels between compound-treated and vehicle-treated rats.
Results
Inhibition of p38 Kinase and Cytokine Release.
RWJ 67657 inhibited the production of TNF-α and IL-1β by LPS-stimulated human PBMCs, with IC50 values of 3 and 11 nM, respectively, approximately 10-fold less than the corresponding IC50values for the literature standard p38 kinase inhibitor SB 203580 (Table 1). To confirm that RWJ 67657 inhibited p38 kinase activity, the compound was added to an immune complex kinase assay using p38α immunoprecipitated from purified LPS-activated human monocytes. The IC50 value was 30 nM (Fig. 3A), indicating that RWJ 67657 was approximately 10-fold more potent in cells than in the in vitro kinase assay. This difference was significantly greater for SB 203580, which was 20- to 50-fold less potent in the in vitro kinase assay compared with the LPS/PBMC assay. Our IC50 values for SB 203580 in the in vitro kinase assay (600–1500 nM) agree with other studies using myelin basic protein as a substrate (Gum et al., 1998) but not with studies using GST-activating transcription factor-2 as a substrate, where the values are considerably lower (30 nM; Kumar et al., 1997; Lisnock et al., 1998).
The serine-threonine kinase MAPKAPK-2 is an in vivo substrate for p38 (Cuenda et al., 1995), which becomes activated via phosphorylation by p38. Therefore, as a means of demonstrating inhibition of p38 kinase activity within the cells at the time of stimulation, we assessed the activity of MAPKAPK-2 immunoprecipitated from monocytes pretreated with RWJ 67657 or SB 203580 before LPS activation (Fig. 3B). MAPKAPK-2 kinase activity was not inhibited by either compound when added directly to immunoprecipitated enzyme (data not shown). The cellular IC50 values for both RWJ 67657 (2 nM) and SB 203580 (20 nM) determined in the MAPKAPK-2 assay were significantly lower than those obtained when adding compounds to immunoprecipitated p38α in vitro (see Discussion).
Specificity Assays.
SB 203580 is known to inhibit the p38α and β isoforms but not the γ and δ isoforms (Kumar et al., 1997). Table 2 shows that RWJ 67657 behaved similarly; both compounds were more potent inhibitors of p38α than p38β. The IC50 values for both RWJ 67657 and SB 203580 in these in vitro assays were significantly higher than those obtained when adding compounds to immunoprecipitated p38α, due at least in part to the fact that the recombinant preparations contained a significant proportion of inactive enzyme that acts as a sink for compound binding (see Discussion).
The specificity of RWJ 67657 was further assessed using a variety of recombinant enzyme and cellular assays. There was little or no cross-reactivity in many other systems, some of which are shown in Tables 1 and 3. Importantly, RWJ 67657 did not inhibit other MAP kinase family members tested, such as ERK-2, other serine-threonine kinases such as PKA, or any tyrosine kinase tested, including those in the c-src family. SB 203580, however, had an IC50 value of ∼5 μM against c-src and the related tyrosine kinase lck, which should be taken into consideration when interpreting studies usinglck+ cells in which this compound is effective only at micromolar concentrations (Crawley et al., 1997). Importantly, RWJ 67657 had less activity than SB 203580 in a platelet assay measuring the generation of thromboxane B2as an indicator of COX-1 activity. This suggests that there could be less gastrointestinal toxicity associated with the RWJ compound in vivo. SB 203580 has been shown to inhibit COX-1 directly (IC50 = 2 μM) but to also inhibit thromboxane synthase activity (Borsch-Haubold et al., 1998), which may also lead to decreased thromboxane B2 production. This possibility has not been ruled out for RWJ 67657.
Cellular specificity data are shown in Table 3. Neither RWJ 67657 nor SB 203580 inhibited IL-2 or IFN-γ release from T cells or T cell proliferation induced by antibody-mediated CD3 plus CD28 cross-linking. In addition, neither compound inhibited a variety of other cellular responses, including erythropoietin-driven proliferation and neutrophil chemotaxis in response to either IL-8 or f-Met-Leu-Phe (data not shown). Because erythropoietin and f-Met-Leu-Phe have been shown to induce p38 activation (Nagata et al., 1997; Zu et al., 1998), our data suggest that p38 may not be essential for these cellular responses.
Species Differences.
Interestingly, both RWJ 67657 and SB 203580 exhibited species differences in p38-dependent cellular assays (Table 4). It is of note that TNF-α production by rat PBMCs was ∼90-fold less sensitive to RWJ 67657 than was the human PBMC response, whereas the beagle dog PBMC response was only 2- to 3-fold less sensitive. The reason for this species difference in sensitivity to p38 inhibitors is unknown but may be related to differences in the amino acid sequences of human versus rat p38α (Lee et al., 1994; Nemeth et al., 1998).
Animal Experiments.
RWJ 67657 inhibited TNF-α release in LPS-treated mice (ED50 = 25 mg/kg) and rats (ED50 = 10 mg/kg), respectively (Tables5 and 6). These data are similar to those reported for SB 203580 (Badger et al., 1996). The relatively high doses required for 50% inhibition in vivo, despite the potent activity in the cellular and enzyme assays, suggest that the compound may be poorly absorbed and/or rapidly metabolized in rodents. In fact, the oral bioavailability of RWJ 67657 was determined to be 3.7 ± 1.9% in male rats, which were used for the LPS experiments, and 32.3 ± 8.6% in female rats, in separate experiments. This also may explain the observation that RWJ 67657 was less potent than SB 203580 in the in vivo mouse assay (Table 5), even though it was more potent than SB 203580 in mouse cellular and enzyme assays, similar to human cells (data not shown). Other possibilities include greater binding of RWJ 67657 to mouse serum proteins, or greater metabolism of the compound in mice, relative to SB 203580. In addition, as seen in Table 4 for rat monocytes, rodent cells may generally be less sensitive to these inhibitors than human cells.
Discussion
We report the biological evaluation of RWJ 67657, a novel inhibitor of p38 kinase. In vitro, this compound potently inhibited TNF-α production from LPS- or SEB-activated human PBMCs and selectively inhibited the enzymatic activity of the α and β isoforms of p38. As shown in Fig. 2, RWJ 67657 was a significantly more potent inhibitor of native activated p38α in vitro (IC50 = 30 nM) than the literature standard SB 203580 (IC50 = 0.6–1.5 μM). This difference in IC50 values was less dramatic when MAPKAPK-2 activity was measured as an indicator of p38 inhibition within the cells (2 versus 20 nM). This superior potency was also seen at the cellular level, where RWJ 67657 inhibited LPS-induced production of TNF-α and IL-1β with IC50 values of 3 and 11 nM, respectively, whereas the corresponding IC50values for SB 203580 were 30 and 94 nM (Table 1). The bulk of the data suggest that RWJ 67657 is ∼10-fold more potent than SB 203580 for inhibition of p38α kinase activity and inhibition of the p38-dependent cellular responses measured here.
It should be noted that the potency of both compounds was much less when recombinant p38 kinases were assayed, compared with cellular responses. There are at least two reasons for this. First, these compounds bind equally well to active (phosphorylated on Thr180 and Tyr182) and inactive (monophosphorylated or nonphosphorylated) enzyme (Frantz et al., 1998). The recombinant enzymes used in our study contained a substantial proportion of inactive enzyme that can bind inhibitors, thereby acting as sinks for compound and artifactually increasing the concentrations of compound required for significant inhibition of activity. Second, because compounds of this class block LPS-induced phosphorylation of p38 in monocytic cells, it has been proposed that the p38-inhibitor complex may be a poor substrate for the upstream kinases that activate p38 (Frantz et al., 1998). Therefore, in cells, compounds of this class may inhibit p38-dependent responses by two mechanisms, inhibition of p38 kinase activity via inhibition of ATP binding to the enzyme, and inhibition of p38 activation by either upstream kinases, autophosphorylation, or both, which is reflected in the increased potency in the cellular assays.
This may also explain the greater potency seen when cells were treated with inhibitors and IC50 values determined by measuring the activity of the p38 substrate MAPKAPK-2 in vitro, versus the lesser potency observed when compounds were added only in vitro to p38 immunoprecipitated from activated cells. In the latter case, inhibitors were not present in the cells, and therefore p38 activation by upstream kinases was unaffected. Furthermore, the higher IC50 value for inhibition when compound was added to the kinase reaction rather than directly to cells could also reflect differences in the ability of the compounds to bind to free (intracellular) versus antibody-bound kinase.
RWJ 67657 exhibited negligible inhibition of a number of other enzymes, including the tyrosine kinases p56 lck and c-src, the unrelated serine-threonine kinase PKA, or the related MAP kinase ERK-2, even at micromolar concentrations. The IC50 value for RWJ 67657 in the COX-1 assay was >10 μM, well above the nanomolar concentrations required to inhibit TNF-α release by 50%. This suggests that there may be a decreased likelihood of gastrointestinal toxicity associated with RWJ 67657 at therapeutic concentrations in vivo.
RWJ 67657 exhibited species differences in our p38-dependent cellular assays (Table 4). It is of note that LPS-stimulated rat PBMCs were ∼90-fold less sensitive to RWJ 67657 than were human PBMCs, whereas beagle dog PBMCs were only 2- to 3-fold less sensitive. This species difference in sensitivity to p38 inhibitors may be related to species differences in the amino acid sequences of p38α. The rat p38α protein sequence (Nemeth et al., 1998) differs by only five amino acids from the human sequence (Lee et al., 1994). However, one of the differences is a leucine-to-histidine change at residue 48 in the rat protein, which introduces a basic amino acid with a bulkier R group close to residue 51, which was shown by X-ray crystallography of p38/inhibitor cocrystals to be a point of contact with thepara-fluorophenyl-containing pyridynlimidazoles, such as SB 203580 (Wilson et al., 1997). Other possibilities include decreased intracellular exposure to the compounds, due to differences in membrane permeability, intracellular metabolism of the compounds, transport of the compounds out of the cell, or binding to nontarget proteins. If this sensitivity difference holds in vivo, a therapeutic dose of such compounds in humans might be lower than that observed in rodent models, assuming similar bioavailability and pharmacokinetics.
Finally, RWJ 67657 showed a dose-dependent reduction of LPS-induced serum TNF-α levels in both mice and rats after oral administration. Near-total inhibition of TNF-α production was observed at 50 mg/kg in mice and 25 mg/kg in rats. Because of this favorable pharmacological profile, RWJ 67657 is being further evaluated for preclinical development.
Footnotes
-
Send reprint requests to: Dr. John Siekierka, OMP Admin. 2360, Drug Discovery, The R. W. Johnson Pharmaceutical Research Institute, 1000 Route 202 South, Raritan, NJ 08869. E-mail:jsiekier{at}prius.jnj.com
- Abbreviations:
- TNF-α
- tumor necrosis factor-α
- IL-1β
- interleukin-1β
- IFN-γ
- interferon-γ
- HBSS
- Hanks’ balanced salt solution
- ELISA
- enzyme-linked immunosorbent assay
- COX-1 and -2
- cyclooxygenase-1 and -2
- DMSO
- dimethyl sulfoxide
- IL-2
- interleukin-2
- LPS
- lipopolysaccharide
- SEB
- staphylococcal enterotoxin B
- NP-40
- Nonidet P-40
- PBMC
- peripheral blood mononuclear cell
- MAPKAPK-2
- mitogen-activated protein kinase-activated protein kinase-2
- GST
- glutathione-S-transferase
- Received April 7, 1999.
- Accepted July 22, 1999.
- The American Society for Pharmacology and Experimental Therapeutics