Design and synthesis of potent, selective, and orally bioavailable tetrasubstituted imidazole inhibitors of p38 mitogen-activated protein kinase.

Novel potent and selective diarylimidazole inhibitors of p38 MAP (mitogen-activated protein) kinase are described which have activity in both cell-based assays of tumor necrosis factor-alpha (TNF-alpha) release and an animal model of rheumatoid arthritis. The SAR leading to the development of selectivity against c-Raf and JNK2alpha1 kinases is presented, with key features being substitution of the 4-aryl ring with m-trifluoromethyl and substitution of the 5-heteroaryl ring with a 2-amino substituent. Cell-based activity was significantly enhanced by incorporation of a 4-piperidinyl moiety at the 2-position of the imidazole which also enhanced aqueous solubility. In general, oral bioavailability of this class of compounds was found to be poor unless the imidazole was methylated on nitrogen. This work led to identification of 48, a potent (p38 MAP kinase inhibition IC50 0.24 nM) and selective p38 MAP kinase inhibitor which inhibits lipopolysaccharide-stimulated release of TNF-alpha from human blood with an IC50 2.2 nM, shows good oral bioavailability in rat and rhesus monkey, and demonstrates significant improvement in measures of disease progression in a rat adjuvant-induced arthritis model.

Aggrecan degradation in human cartilage. Evidence for both matrix metalloproteinase and aggrecanase activity in normal, osteoarthritic, and rheumatoid joints.

To examine the activity of matrix metalloproteinases (MMPs) and aggrecanase in control and diseased human articular cartilage, metabolic fragments of aggrecan were detected with monospecific antipeptide antibodies. The distribution and quantity of MMP-generated aggrecan G1 fragments terminating in VDIPEN341 were compared with the distribution of aggrecanase-generated G1 fragments terminating in NITEGE373. Both types of G1 fragments were isolated from osteoarthritic cartilage. The sizes were consistent with a single enzymatic cleavage in the interglobular domain region, with no further proteolytic processing of these fragments. Both neoepitopes were also detected by immunohistochemistry in articular cartilage from patients undergoing joint replacement for osteoarthritis (OA), rheumatoid arthritis (RA), and in cartilage from adults with no known joint disease. In control specimens, the staining intensity for both G1 fragments increased with age, with little staining in cartilage from 22-wk-old fetal samples. There was also an increase with age in the extracted amount of MMP-generated neoepitope in relation to both aggrecan and collagen content, confirming the immunohistochemical results. After the age of 20-30 yr this relationship remained at a steady state. The staining for the MMP-generated epitope was most marked in control cartilage exhibiting histological signs of damage, whereas intense staining for the aggrecanase-generated fragment was often noted in adult cartilage lacking overt histological damage. Intense staining for both neoepitopes appeared in the more severely fibrillated, superficial region of the tissue. Intense immunostaining for both VDIPEN- and NITEGE- neoepitopes was also detected in joint cartilage from patients with OA or RA. Cartilage in these specimens was significantly more degraded and high levels of staining for both epitopes was always seen in areas with extensive cartilage damage. The levels of extracted VDIPEN neoepitope relative to collagen or aggrecan in both OA and RA samples were similar to those seen in age-matched control specimens. Immunostaining for both types of aggrecan fragments was seen surrounding the cells but also further removed in the interterritorial matrix. In some regions of the tissue, both neoepitopes were found while in others only one was detected. Thus, generation and/or turnover of these specific catabolic aggrecan fragments is not necessarily coordinated. Our results are consistent with the presence in both normal and arthritic joint cartilage of proteolytic activity against aggrecan based on both classical MMPs and "aggrecanase."

Quantification of a matrix metalloproteinase-generated aggrecan G1 fragment using monospecific anti-peptide serum.

Several members of the matrix metalloproteinase family have been reported to cleave aggrecan in the interglobular domain between Asn-341 and Phe-342. An antiserum was prepared against a peptide conjugate corresponding to the C-terminal sequence of the matrix metalloproteinase-generated aggrecan G1 fragment (Phe335-Val-Asp-Ile-Pro-Glu-Asn341). A quantitative radioimmunoassay, with a limit of detection of about 80 pM, was developed using this antiserum. This antiserum requires the free carboxyl group of the C-terminal asparagine for optimal recognition. If the C-terminal asparagine is excised from the sequence, replaced with closely related amino acids, or extended across the matrix metalloproteinase cleavage site, there is a 40-10,000-fold loss in detection. Using peptides cleaved from the N-terminus, it was determined that the antiserum requires the entire Phe-Val-Asp-Ile-Pro-Glu-Asn sequence for optimal recognition. The radioimmunoassay detects matrix metalloproteinase-generated G1 fragments with similar sensitivity to the Phe-Val-Asp-Ile-Pro-Glu-Asn peptide, but it does not recognize intact aggrecan. Immunoreactive aggrecan G1 fragments of molecular mass 50 kDa are generated by the matrix metalloproteinases stromelysin and gelatinase A. In contrast, under identical conditions, the closely related metalloproteinases, gelatinase B and collagenase, as well as cathepsin G, cathepsin B and human leucocyte elastase, did not generate a G1 fragment recognized by the antiserum. The anti-Phe-Val-Asp-Ile-Pro-Glu-Asn serum detects stromelysin-generated aggrecan G1 fragments from mouse, guinea pig, rabbit and human, indicating that the detection is not species-specific. This antiserum and radio-immunoassay should be useful for quantifying and characterizing matrix metalloproteinase-generated aggrecan G1 fragments in articular cartilage and synovial fluids from humans and various animal models of articular-cartilage destruction.

Dexamethasone attenuation of cytokine-mediated articular cartilage degradation in experimental lapine Haemophilus arthritis.

The role of cytokines in the regulation of articular inflammation and cartilage degradation was evaluated in the rabbit model of Haemophilus influenzae type b arthritis. At 6 and 12 h after intraarticular infection, treatment with IB4 monoclonal antibody to the CD18 leukocyte receptor alone or in combination with dexamethasone resulted in significant reduction of synovial fluid (SF) neutrophil concentration. Treatment with dexamethasone alone was associated with lower SF concentrations of interleukin-1 (IL-1), tumor necrosis factor-alpha, and stromelysin than in other groups. At 24 h after infection, increased cartilage degradation was detected in untreated controls and in animals treated with IB4 alone or in combination with dexamethasone compared with those treated with dexamethasone alone. Multiple regression analyses indicated SF concentration of IL-1 and stromelysin as the significant predictors of cartilage degradation. These data suggest that IL-1 mediates cartilage degradation by regulation of metalloproteinases, such as stromelysin, during acute experimental bacterial arthritis.

Production, purification and characterization of canine prostromelysin.

One of the best studied animal models of osteoarthritis is a dog model in which the anterior cruciate ligament of the hind limb stifle joint is transected (Pond-Nuki model). To determine whether stromelysin might play a role in this model, it was necessary to purify the enzyme for production of suitable probes. In the present study, dog synovial fibroblasts were stimulated to express a metalloproteinase that was demonstrated to be canine prostromelysin by Northern blot, protein purification and amino-terminal sequence analyses. Unlike rabbit synoviocytes, passaged dog synoviocytes did not express stromelysin mRNA in response to recombinant human IL-1, but expressed stromelysin mRNA only upon treatment with dog monocyte-conditioned medium (dMCM). The aminophenylmercuric acetate (APMA)-activatable metalloproteinase present in the culture supernatants of stimulated dog synoviocytes was purified using a combination of ion-exchange and dye matrix affinity chromatography. The purified canine metalloproteinase co-migrated on reducing SDS-PAGE with recombinant human prostromelysin-1 as a doublet with apparent molecular masses of 54 and 56 kDa. Similar to APMA-activated human prostromelysin-1, the APMA-activated canine metalloproteinase was inhibited by 1,10-phenanthroline or recombinant human tissue inhibitor of metalloproteinase (TIMP). The amino-terminal sequences of the canine pro- and APMA-activated enzymes were compared with those of human, rabbit and rat stromelysin. The striking homologies among the sequences demonstrated that the purified canine metalloproteinase was indeed canine prostromelysin. A rabbit anti-canine prostromelysin polyclonal antiserum was generated and used to localize the enzyme within cultured dog synoviocytes and articular cartilage stimulated with dMCM. The reagents developed in this study should be useful for examining the expression and distribution of prostromelysin in canine models of osteoarthritis.

Specificity of glycosphingolipid recognition by Entamoeba histolytica trophozoites.

The ability of purified glycosphingolipids to enhance liposome-stimulated Entamoeba histolytica actin polymerization was assessed as a means of defining the specificity of mammalian cell membrane lipid glycan recognition by this parasite. Synthetic liposomes containing a variety of individual glycosphingolipids bearing neutral, straight-chain oligomeric glycans with galactose or N-acetylgalactosamine termini stimulated rapid (90-s) polymerization of amoeba actin. Glycans with terminal N-acetylglucosamine residues were not stimulatory at all or were only weakly stimulatory. Glycans with glucose, N-acetylglucosamine, galactose, and N-acetylgalactosamine as the penultimate residue were recognized. Attachment of N-acetylneuraminate to the terminal residue of a stimulatory glycosphingolipid eliminated activity; attachment of fucose to the penultimate sugar reduced activity. Glycans with a terminal beta 1-4 or 1-3 glycosidic bond were most effective; glycans with terminal alpha 1-4 or 1-3 glycosides were less effective. The activity of glycans with both beta- and alpha-linked terminal glycosides was inhibited by lactose, suggesting recognition of both configurations by a single amoeba protein. The ability of liposomes to stimulate actin polymerization reflected the extent of liposome phagocytosis.

Use of non-cellular models to study the interaction of E. histolytica with mammalian cells.

We have utilized liposomes constructed with individual mammalian cell membrane glycosphingolipids and latex beads conjugated with glycoproteins as models to investigate the molecular specificity and mechanism of interaction of E. histolytica with target cells. Synthetic liposomes constructed with a variety of glycosphingolipids bearing neutral, straight chain glycans with galactose or N-acetylgalactosamine termini stimulated rapid (90 sec), contact dependent polymerization of E. histolytica actin. Glycans with terminal N-acetylglucosamine residues were not, or only weakly, stimulatory. Attachment of N-acetylneuraminate to the terminal residue of stimulatory glycosphingolipids eliminated activity. Attachment of fucose to the penultimate sugar reduced glycan recognition. Latex beads conjugated with asialofetuin (galactose beta 1-4 glycan termini) adhered to amoebae more effectively than fetuin conjugated beads (N-acetylneuraminate termini) or agalactosyl asialofetuin conjugated beads (N-acetylglucosamine termini). However, none of the glycoprotein conjugated beads stimulated contact mediated polymerization of E. histolytica actin. The carbohydrate specificity of E. histolytica interaction with non-cellular particles resembles that observed with whole target cells our results demonstrate that carbohydrate recognition specificity extends to lipid as well as protein bound glycoconjugates. Further, these studies suggest that the biochemical consequences of binding to glycosphingolipids differ from those resulting from interaction with glycoprotein. This may be relevant to the mechanism of mammalian cell attack by this pathogen.