Sites of action for future therapy: an adenosine-dependent mechanism by which aspirin retains its antiinflammatory activity in cyclooxygenase-2 and NFkappaB knockout mice.

The antiinflammatory action of aspirin is generally attributed to inhibition of cyclooxygenases 1 and 2, but additional mechanisms are at work. These include inhibition of NFkappaB translocation to the nucleus and the capacity of aspirin to promote accumulation of adenosine, a potent antiinflammatory autocoid. We tested these hypotheses in the murine air pouch model of acute inflammation in wild type mice and in cyclooxygenase 2 or NFkappaB knockouts. The antiinflammatory effects of aspirin, sodium salicylate and indomethacin did not correlate with inhibition of cyclooxygenase in either group. Indeed, aspirin retained its antiinflammatory properties even in COX-2 knockouts. Similarly, aspirin was no less antiinflammatory in mice rendered deficient for NFkappaB (p105) than in wild type controls. In contrast, dexamethasone lost its antiinflammatory capacity in NFkappaB knockouts. Aspirin and sodium salicylate dramatically increased concentrations of adenosine in exudates, a property shared with methotrexate and sulfasalazine. Removal of adenosine by adenosine deaminase or specific antagonism of adenosine at A(2)receptors completely reversed the antiinflammatory effects of aspirin and sodium salicylate, but not those of dexamethasone. This adenosine-dependent, antiinflammatory effect of aspirin points to another target of drug development.

Salicylates and sulfasalazine, but not glucocorticoids, inhibit leukocyte accumulation by an adenosine-dependent mechanism that is independent of inhibition of prostaglandin synthesis and p105 of NFkappaB.

The antiinflammatory action of aspirin generally has been attributed to direct inhibition of cyclooxygenases (COX-1 and COX-2), but additional mechanisms are likely at work. These include aspirin's inhibition of NFkappaB translocation to the nucleus as well as the capacity of salicylates to uncouple oxidative phosphorylation (i.e., deplete ATP). At clinically relevant doses, salicylates cause cells to release micromolar concentrations of adenosine, which serves as an endogenous ligand for at least four different types of well-characterized receptors. Previously, we have shown that adenosine mediates the antiinflammatory effects of other potent and widely used antiinflammatory agents, methotrexate and sulfasalazine, both in vitro and in vivo. To determine in vivo whether clinically relevant levels of salicylate act via adenosine, via NFkappaB, or via the "inflammatory" cyclooxygenase COX-2, we studied acute inflammation in the generic murine air-pouch model by using wild-type mice and mice rendered deficient in either COX-2 or p105, the precursor of p50, one of the components of the multimeric transcription factor NFkappaB. Here, we show that the antiinflammatory effects of aspirin and sodium salicylate, but not glucocorticoids, are largely mediated by the antiinflammatory autacoid adenosine independently of inhibition of prostaglandin synthesis by COX-1 or COX-2 or of the presence of p105. Indeed, both inflammation and the antiinflammatory effects of aspirin and sodium salicylate were independent of the levels of prostaglandins at the inflammatory site. These experiments also provide in vivo confirmation that the antiinflammatory effects of glucocorticoids depend, in part, on the p105 component of NFkappaB.

Modes of action of aspirin-like drugs: salicylates inhibit erk activation and integrin-dependent neutrophil adhesion.

The anti-inflammatory effects of high-dose salicylates are well recognized, incompletely understood and unlikely due entirely to cyclooxygenase (COX) inhibition. We have previously reported a role for activation of the kinase Erk in CD11b/CD18 integrin-dependent adhesiveness of human neutrophils, a critical step in inflammation. We now report the effects of salicylates on neutrophil Erk and adhesion. Exposure of neutrophils to aspirin or sodium salicylate (poor COX inhibitor) inhibited Erk activity and adhesiveness of formylmethionyl-leucyl-phenylalanine- and arachidonic acid-stimulated neutrophils, consistent with anti-inflammation but not COX inhibition (IC50s = 1-8 mM). In contrast, indomethacin blocked neither Erk nor adhesion. Inhibition of Mek (proximal activator of Erk) also blocked stimulation of Erk and adhesion by formylmethionyl-leucyl-phenylalanineand arachidonic acid. Salicylate inhibition of Erk was independent of protein kinase A activation and generation of extracellular adenosine. These data are consistent with a role for Erk in stimulated neutrophil adhesion, and suggest that anti-inflammatory effects of salicylates may be mediated via inhibition of Erk signaling required for integrin-mediated responses.

Integrin-dependent homotypic adhesion of neutrophils. Arachidonic acid activates Raf-1/Mek/Erk via a 5-lipoxygenase- dependent pathway.

AA stimulates integrin-dependent neutrophil adhesion, a critical early step in acute inflammation. However, neither the signaling pathway(s) of AA-stimulated adhesion, nor whether AA acts directly or through the generation of active metabolites, has been elucidated. Previously, we have observed a tight association between neutrophil Erk activation and homotypic adhesion in response to chemoattractants acting through G protein-linked receptors. We now report a similar association between homotypic adhesion and Erk activation in response to AA. Erk activation was cyclooxygenase independent and required AA metabolism to 5(S)- hydroperoxyeicosatetraenoic acid (5-HpETE) via 5-lipoxygenase, but not the further lipoxygenase-dependent metabolism of 5-HpETE to leukotrienes. AA stimulation of Erk was accompanied by Raf-1 activation and was sensitive to inhibitors of Raf-1 and Mek. Whereas activation of Erk by AA was pertussis toxin sensitive, [3H]-AA binding to neutrophils was not saturable, suggesting that an AA metabolite activates a G protein. Consistent with this hypothesis, Erk activation by 5(S)-hydroxyeicosatetraenoic acid (5-HETE; lipoxygenase-independent metabolite of 5-HpETE) was also pertussis toxin sensitive. These data suggest that a 5-lipoxygenase metabolite of AA, e.g., 5-HETE, is released from AA-treated cells to engage a plasma membrane-associated, pertussis toxin-sensitive, G protein-linked receptor, leading to activation of Erk and adhesion via the Raf-1/Mek signal transduction pathway.

Reiter's syndrome among Asian shipboard immigrants: the case of The Golden Venture.

OBJECTIVE: To assess the incidence of Reiter's syndrome aboard The Golden Venture, a ship carrying illegal immigrants from China to the United States. METHODS: After identification of an index case, we conducted telephone interviews with medical staff at immigrant detention centers in Pennsylvania, New York, and Virginia. When a potential case was identified at one facility, we performed a site inspection, reviewing the medical records of all detainees and performing histories and physicals on all those with joint and/or ocular complaints. RESULTS: We identified two patients, both HLA B27 positive, with Reiter's syndrome. The observed incidence (0.87%) approximated the predicted incidence but may have underestimated the actual incidence. We review the history of shipboard Reiter's syndrome, and discuss the pathogenic roles of HLA B27 and particular infectious agents. CONCLUSION: Continued transportation of illegal immigrants from China and other parts of the world is likely to result in occasional clusters of Reiter's syndrome. Physicians treating immigrant populations should remain aware of the possibility of reactive arthritis.

Colchicine alters the quantitative and qualitative display of selectins on endothelial cells and neutrophils.

Since colchicine-sensitive microtubules regulate the expression and topography of surface glycoproteins on a variety of cells, we sought evidence that colchicine interferes with neutrophil-endothelial interactions by altering the number and/or distribution of selectins on endothelial cells and neutrophils. Extremely low, prophylactic, concentrations of colchicine (IC50 = 3 nM) eliminated the E-selectin-mediated increment in endothelial adhesiveness for neutrophils in response to IL-1 (P < 0.001) or TNF alpha (P < 0.001) by changing the distribution, but not the number, of E-selectin molecules on the surface of the endothelial cells. Colchicine inhibited stimulated endothelial adhesiveness via its effects on microtubules since vinblastine, an agent which perturbs microtubule function by other mechanisms, diminished adhesiveness whereas the photoinactivated colchicine derivative gamma-lumicolchicine was inactive. Colchicine had no effect on cell viability. At higher, therapeutic, concentrations colchicine (IC50 = 300 nM, P < 0.001) also diminished the expression of L-selectin on the surface of neutrophils (but not lymphocytes) without affecting expression of the beta 2-integrin CD11b/CD18. In confirmation, L-selectin expression was strikingly reduced (relative to CD11b/CD18 expression) on neutrophils from two individuals who had ingested therapeutic doses of colchicine. These results suggest that colchicine may exert its prophylactic effects on cytokine-provoked inflammation by diminishing the qualitative expression of E-selectin on endothelium, and its therapeutic effects by diminishing the quantitative expression of L-selectin on neutrophils.

The mode of action of aspirin-like drugs: effect on inducible nitric oxide synthase.

Nitric oxide synthesized by inducible nitric oxide synthase (iNOS) has been implicated as a mediator of inflammation in rheumatic and autoimmune diseases. We report that exposure of lipopolysaccharide-stimulated murine macrophages to therapeutic concentrations of aspirin (IC50 = 3 mM) and hydrocortisone (IC50 = 5 microM) inhibited the expression of iNOS and production of nitrite. In contrast, sodium salicylate (1-3 mM), indomethacin (5-20 microM), and acetaminophen (60-120 microM) had no significant effect on the production of nitrite at pharmacological concentrations. At suprapharmacological concentrations, sodium salicylate (IC50 = 20 mM) significantly inhibited nitrite production. Immunoblot analysis of iNOS expression in the presence of aspirin showed inhibition of iNOS expression (IC50 = 3 mM). Sodium salicylate variably inhibited iNOS expression (0-35%), whereas indomethacin had no effect. Furthermore, there was no significant effect of these nonsteroidal anti-inflammatory drugs on iNOS mRNA expression at pharmacological concentrations. The effect of aspirin was not due to inhibition of cyclooxygenase 2 because both aspirin and indomethacin inhibited prostaglandin E2 synthesis by > 75%. Aspirin and N-acetylimidazole (an effective acetylating agent), but not sodium salicylate or indomethacin, also directly interfered with the catalytic activity of iNOS in cell-free extracts. These studies indicate that the inhibition of iNOS expression and function represents another mechanism of action for aspirin, if not for all aspirin-like drugs. The effects are exerted at the level of translational/posttranslational modification and directly on the catalytic activity of iNOS.

Activation-dependent carboxyl methylation of neutrophil G-protein gamma subunit.

The gamma subunits of heterotrimeric guanine nucleotide-binding regulatory (G) proteins (G gamma) are post-translationally processed at their C termini by prenylation, proteolysis, and carboxyl methylation. Whereas prenylation of G gamma is required for membrane association of G proteins, the role of carboxyl methylation is unknown. Here we show that human neutrophils express G gamma 2 but not G gamma 3 or G gamma 7 and that carboxyl methylation of G gamma 2 is associated with signal transduction. In a reconstituted cell-free system, neutrophil G gamma 2 was labeled by the methyl donor S-[methyl-3H]adenosyl-L-methionine. Carboxyl methylation was confirmed by alkaline hydrolysis and quantitation of volatile [3H]methanol. Neutrophil G gamma 2 methylation was stimulated by activation of G protein with guanosine 5'-[beta, gamma-thio]triphosphate. We estimate that after 1 hr of G-protein activation at least 6% of the total pool of G gamma 2 was carboxyl-methylated. The inflammatory agonist fMet-Leu-Phe stimulated guanosine 5'-[beta,gamma-thio]triphosphate-dependent carboxyl methylation. Methylation of G gamma 2 was inhibited by the carboxyl methyltransferase inhibitor N-acetyl-S-trans,trans-farnesylcysteine at concentrations that affected signal transduction in neutrophils. These results demonstrate that activation of neutrophil Gi is associated with alpha-carboxyl methyl esterification of G gamma 2 and suggest that carboxyl methylation of G gamma may play a role in signal transduction.

Targets for antiinflammatory drugs.

Aspirin-like nonsteroidal antiinflammatory agents, corticosteroids, and methotrexate are the mainstays of therapy in rheumatoid arthritis and other inflammatory diseases. It is now clear that these agents share at least one characteristic: All of these agents diminish the adhesive interactions required for the accumulation of white blood cells at an inflamed site. We discuss the biochemical and functional mechanisms by which these drugs alter the inflammatory response, and the potential for development of new antiinflammatory agents based upon these pharmacological actions.

Nonsteroidal antiinflammatory agents inhibit stimulated neutrophil adhesion to endothelium: adenosine dependent and independent mechanisms.

All nonsteroidal antiinflammatory drugs (NSAIDs) inhibit neutrophil aggregation (homotypic cell-cell adhesion) and do so without affecting expression of CD11b/CD18. Since the first step in acute inflammation is a critical interaction between neutrophils and the vascular endothelium (heterotypic cell-cell adhesion), we determined whether NSAIDs diminish the adherence of neutrophils to the endothelium. At antiinflammatory concentrations (0.5-5 mM) sodium salicylate, an NSAID that does not inhibit prostaglandin synthesis, inhibited stimulated but not unstimulated neutrophil adherence to endothelial cells (IC50 < 1 mM, P < 0.00001). Salicylates have previously been shown to inhibit oxidative phosphorylation and, predictably, sodium salicylate inhibited oxidative phosphorylation, as evidenced by depletion of ATP stores (875 +/- 75 pmol/10(6) PMN, [2.92 +/- 0.25 mM]) in stimulated (FMLP, 0.1 microM) but not resting neutrophils treated with antiinflammatory doses of sodium salicylate (EC50 = 1 mM, P < 0.00001). Indomethacin and piroxicam (10 and 30 microM) only minimally decreased ATP concentrations in stimulated and resting neutrophils. ATP is metabolized to adenosine, and we have previously demonstrated that both endogenously released (180-200 nM) and exogenous adenosine (IC50 = 250 nM) inhibit stimulated neutrophil adherence to endothelial cells. To determine whether the increased metabolism of ATP and the resultant increase in adenosine release were responsible for inhibition of neutrophil adhesion to endothelium, we determined whether addition of adenosine deaminase (ADA, 0.125 IU/ml), an enzyme that converts extracellular adenosine to its inactive metabolite, inosine, affected inhibition of neutrophil adhesion to endothelium by stimulated neutrophils. ADA significantly reversed inhibition of neutrophil adherence to endothelium by sodium salicylate (0.5-5 mM, P < 0.00001). This suggests that sodium salicylate inhibits neutrophil adherence by increasing adenosine release. Whereas indomethacin and piroxicam (10-50 microM) also inhibited stimulated neutrophil adherence to endothelial cells, ADA did not affect their inhibition of adherence. These studies demonstrate a heretofore unexpected antiinflammatory mechanism for salicylates: salicylates increase ATP hydrolysis and thereby enhance release of adenosine. Moreover, these data are consistent with the hypothesis that NSAIDs differ from one another with respect to their mechanisms of action.

Inhibition of neutrophil function by aspirin-like drugs (NSAIDS): requirement for assembly of heterotrimeric G proteins in bilayer phospholipid.

Non-steroidal anti-inflammatory drugs (NSAIDs) inhibit neutrophil functions via mechanisms that are independent of their effects on prostaglandin biosynthesis. We examined the effects of sodium salicylate and piroxicam on GTP/GDP exchange by a regulatory G protein (G alpha i). Plasma membrane and cytosol of human neutrophils were prepared by nitrogen cavitation and discontinuous sucrose density centrifugation. Salicylate (3 mM) and piroxicam (50 microM) reduced [35S]GTP gamma S binding to purified plasma membranes [65 +/- 3.7 and 75 +/- 5.3% (P < 0.003) of control, respectively]. Membrane-associated G alpha/beta gamma was solubilized by treatment of plasma membranes with sodium cholate. NSAIDs did not inhibit binding of GTP to solubilized G alpha/beta gamma derived from detergent-treated plasma membranes. Lipid reconstitution was achieved by detergent dialysis followed by the addition of bilayer liposomes (phosphatidylcholine). Salicylate and piroxicam inhibited GTP gamma S binding to G alpha/beta gamma derived from solubilized plasma membranes reconstituted in phosphatidylcholine vesicles (bilayer structures) but had no effect when phosphatidylethanolamine (hexagonal phase II structure) was used for reconstitution. Salicylate and piroxicam had no effect on GTP binding to cytosolic fractions in which soluble G alpha i exists as a free subunit, suggesting that the effect required either assembly of G alpha i/beta gamma heterotrimer or the presence of a lipid bilayer. Although the addition of purified bovine beta gamma subunits to dialyzed cytosol increased both the total GIP binding capacity and the pertussis toxin-dependent ADP-ribosylation of G alpha i, consistent with assembly of a G protein heterotrimer, NSAIDs had no effect on GTP binding. In contrast, NSAIDs inhibited GTP binding to heterotrimeric G alpha cytosol/beta gamma bovine when the complex was inserted into bilayer liposomes. The data indicate that salicylate and piroxicam disrupt neutrophil function via their capacity to interfere with GTP/GDP exchange at an alpha subunit of a regulatory G protein, an effect which requires assembly of the active heterotrimer G alpha i/beta gamma in a phospholipid bilayer.

Characterization of a plasma membrane-associated prenylcysteine-directed alpha carboxyl methyltransferase in human neutrophils.

Signal transduction in human neutrophils requires prenylcysteine-directed carboxyl methylation of ras-related low molecular weight GTP-binding proteins. We now report the subcellular localization and characterization of a neutrophil prenylcysteine alpha carboxyl methyltransferase. The highest carboxyl methyltransferase activity copurified with biotinylated neutrophil surface membranes, supporting a plasma membrane localization of the enzyme. Neutrophil nuclear fractions contained little or no methyltransferase activity. Methyltransferase activity was detergent-sensitive but could be reconstituted by removal of detergent in the presence of phosphatidyl choline and an anionic phospholipid. N-Acetyl-S-trans,trans-farnesyl-L-cysteine (AFC) and N-acetyl-S-all-trans-geranylgeranyl-L-cysteine (AGGC) were effective substrates for neutrophil prenylcysteine-directed methyltransferase; Vmax values for AFC and AGGC (16.4 and 22.1 pmol of methylated/mg protein/min, respectively) are among the highest yet reported. Although both GTP gamma S and the chemoattractant fMet-Leu-Phe stimulated methylation of ras-related proteins, neither affected methylation of AFC. These data suggest that neutrophil plasma membranes contain a phospholipid-dependent, prenylcysteine-directed carboxyl methyltransferase of relatively high specific activity that modifies ras-related protein substrates in the GTP-bound, activated state.

Chemoattraction of neutrophils by substance P and transforming growth factor-beta 1 is inadequately explained by current models of lipid remodeling.

"Classical" chemoattractants, such as FMLP, C5a, or leukotriene B4, not only elicit directed motility but also activate neutrophils (degranulation, release of active oxygen species). Signal transduction after ligation of receptors for these classical chemoattractants is mediated by pertussis toxin (PT)-sensitive, heterotrimeric G proteins and the early production of lipid messengers via phospholipases. In contrast, we have previously shown that substance P (SP) and transforming growth factor-beta 1 (TGF-beta 1) are "pure" chemoattractants in that they elicit chemotaxis without activating neutrophils. Paradoxically, pure chemoattractants also activate G proteins (plasmalemmal GTPase activity) without eliciting increments in cytosolic calcium ([Ca]i) and thus inositol trisphosphate. We therefore determined lipid remodeling and signal transduction in response to pure chemoattractants. Increments in plasmalemmal GTPase activated by SP (0.1 microM) and TGF-beta 1 (40 fM), like that after FMLP, were PT-sensitive (SP = 6.6 +/- 2 pm/mg/min vs SP + PT = 1.1 +/- 0.9 over basal activity; TGF-beta 1 = 4.3 +/- 1.6 vs TGF-beta 1 + PT = 2.3 +/- 0.9). In parallel, treatment of PMN with PT (1 microgram/ml, 30 min) inhibited chemotaxis (under agarose) after FMLP (2175 +/- 176 (SEM) microns vs 726 +/- 267) and SP (411 +/- 99 microns vs 103 +/- 62 microns) and TGF-beta 1 (40 fM, 375 +/- 53 microns vs 83 +/- 47). However, G proteins coupled to receptors for SP and TGF-beta 1, unlike FMLP, did not appear to be linked to phospholipases in that neither increments in diacylglycerol were detected after receptor ligation (FMLP = 152 +/- 22% resting levels; SP = 101 +/- 5%; TGF-beta 1 = 105 +/- 4%) nor was alkylacylglycerol increased by exposure to SP or TGF-beta 1 (SP = 92 +/- 4%; TGF-beta 1 = 101 +/- 8%; FMLP = 226 +/- 40%). Moreover, polymorphonuclear leukocytes failed to generate phosphatidates (PA) of either species after SP (DA-PA = 79 +/- 9% resting at 60 s; EA-PA = 103 +/- 4%) or TGF-beta 1 (DA-PA = 101 +/- 5%; EA-PA = 98 +/- 9%) in contrast to FMLP (DA-PA = 155 +/- 22%; EA-PA = 149 +/- 16%). The data clearly contravene the current dogma that all chemoattractants use inositol trisphosphate and diglycerides as intracellular signals and suggest the presence of a unique subset of PT-sensitive G proteins, not coupled to "classical" phospholipases, transduce chemoattraction.