Interleukin-8 (IL-8) is a dimeric, C-X-C chemokine, produced by a variety of cells and which elicits proinflammatory responses from the neutrophil. As a prelude to drug design, we have investigated the interactions between IL-8 and its receptor by preparing a number of single-site mutants of IL-8 and determining their activity in receptor-binding and functional assays. In order to define the binding surface as precisely as possible, we have used chemical shifts obtained from nuclear magnetic resonance spectroscopy to screen mutant proteins for structural changes which affect regions of the IL-8 surface remote from the site of mutation. In addition to a previously recognized sequence, Glu4-Leu5-Arg6 in the N-terminal peptide, we have identified a second epitope comprising a contiguous group of non-sequential, solvent-exposed, hydrophobic residues, Phe17, Phe2l, Ile22, and Leu43. These two receptor-binding regions are separated by over 20 A in the IL-8 structure and are important both for receptor binding and function. In addition, we have shown through the production of a covalently linked IL-8 dimer, that subunit dissociation is not necessary for biological activity.

Tumor necrosis factor-␣ (TNF-␣), a cytokine secreted by inflammatory cells, has been implicated in several inflammatory disease, is a potent, orally active inhibitor of the TNF-␣ convertase (TACE), an enzyme responsible for proteolytic cleavage of the membrane bound precursor, pro-TNF-␣. Ro 32-7315 inhibited a recombinant form of TACE (IC 50 ϭ 5.2 nM) with selectivity over related matrix metalloproteinases. In a cellular assay system, THP-1 cell line, and in human and rat whole blood, Ro 32-7315 significantly reduced lipopolysaccharide (LPS)-induced TNF-␣ release with IC 50 values of 350 Ϯ 14 nM (n ϭ 5), 2.4 Ϯ 0.5 M (n ϭ 5), and 110 Ϯ 18 nM (n ϭ 5), respectively. Oral administration of Ro 32-7315 to Wistar rats caused a dose-dependent inhibition of LPS-induced release of systemic TNF-␣ with an ED 50 of 25 mg/kg. Treatment (days 0 -14) of Allen and Hamburys hooded rats with Ro 32-7315 (2.5, 5, 10, and 20 mg/kg, i.p., twice daily) significantly reduced adjuvant-induced secondary paw swelling (42, 71, 83, and 93%, respectively) as compared with the vehicle group. In the Ro 32-7315-treated group, the reduced paw swelling was associated with improved lesion score and joint mobility. Furthermore, in a placebocontrolled, single-dose study, Ro 32-7315 given orally (450 mg) significantly suppressed ex vivo, LPS-induced TNF-␣ release in the whole-blood samples taken from healthy male and female volunteers (mean inhibition of 42% over a 4-h duration, n ϭ 6). These data collectively support the potential use of such a compound for the oral treatment of inflammatory disorders.

No abstract

who in her opening remarks exhorted the audience to come up with stimulating questions and not just try to indulge in industrial espionage (always a challenge for those of us in the pharmaceutical industry). The audience for the most part responded to the challenge very well and Professor Murphy was always ready with a question to start things off on the right track.The programme got underway with a high-tech presentation from Dr Neera Borkakoti (Roche Discovery Welwyn, Welwyn Garden City, UK) who gave a presentation entitled ''Zinc Endoproteases: A Structural Superfamily''. This superfamily can be divided into three broad classes, the aminopeptidases, the endopeptidases and the carboxypeptidases. 3-D structures exist for several members of each class together with inhibitors bound into the active site. The endopeptidase family, to which the matrix metalloproteinases belong, are a particularly rich source of 3-D structures with 11 proteins having been published. Dr Borkakoti pointed out that proteins with minimal primary sequence homology may have a very similar 3-D structure. This was illustrated (with the aid of stereo slides) by superimposing the structure of astacin on MMP-1. The two proteins have 15% homology at the amino acid level but a very similar 3-D structure. When the conformation of inhibitors bound in the active site is compared it is clear that hydroxamate inhibitors of MMP-1 and thermolysin bind to the zinc in the active site in a very similar conformation but the detail of the other interactions that the inhibitors make with the protein differ considerably because of the architecture of the enzymes. Dr Borkakoti went on to argue that homology models can provide useful information about the 3-D structure of a new protein before a structure is available. As an example she used the structure of the snake venom protein adamalysin, which has 50% homology to TNF converting enzyme (TACE) and displayed images of the homology derived structure of TACE with a modelled hydroxamate inhibitor bound. It will be extremely interesting to learn how accurate this type of model building is when the structure of TACE is solved.The second talk was given by Dr Dave Bradshaw also from Roche Discovery Welwyn. The subject of the talk was ''Cartilage Protective Agent: A novel therapy for Arthritis'' and described the discovery and development of a collagenase inhibitor for use in rheumatoid and osteoarthritis. Cartilage protective agent (Ro 32-3555) is an hydroxamate collagenase inhibitor (Ki for MMP-1 and MMP-13 of 3 nM) with lower potency versus gelatinase and stromelysin. The compound is orally active and prevents collagen but not proteoglycan degradation in a sponge/ cartilage implant model (ED50 10 mg.kg-1 b.i.d.). CPA also prevents cartilage degradation in models of inflammatory arthritis e.g. P.acnes induced monoarthritis in rats and rabbits (illustrated with an impressive video of MRI images taken through the knee joint) and in a spontaneous model of OA, the STR/ORT mouse. It has excellent pharmacokinetic properti...