The development of novel LTA4H modulators to selectively target LTB4 generation

Abstract

The pro-inflammatory mediator leukotriene B₄ (LTB₄) is implicated in the pathologies of an array of diseases and thus represents an attractive therapeutic target. The enzyme leukotriene A₄ hydrolase (LTA₄H) catalyses the distal step in LTB₄ synthesis and hence inhibitors of this enzyme have been actively pursued. Despite potent LTA₄H inhibitors entering clinical trials all have failed to show efficacy. We recently identified a secondary anti-inflammatory role for LTA₄H in degrading the neutrophil chemoattractant Pro-Gly-Pro (PGP) and rationalized that the failure of conventional LTA₄H inhibitors may be that they inadvertently prevented PGP degradation. We demonstrate that these inhibitors do indeed fail to discriminate between the dual activities of LTA₄H, and enable PGP accumulation in mice. Accordingly, we have developed novel compounds that potently inhibit LTB₄ generation whilst leaving PGP degradation unperturbed. These novel compounds could represent a safer and superior class of LTA₄H inhibitors for translation into the clinic.

Figure 1. LTA₄H inhibitors to have entered the clinic are non-selective and inhibit PGP degradation in vitro and in vivo.

Schematic representation of the binding site of LTA₄H showing the relative positions occupied by LTA₄ (a) and PGP (b). (c) Proposed solution for a new compound that sits in the narrow arm occupied by the hydrophobic tail of LTA₄, (Site A) blocking its hydrolysis, but retaining ability to cleave the tripeptide PGP (Site B). The capacity of DG-051 to inhibit LTA₄H epoxide hydrolase (LTB₄ generation; (d)) and aminopeptidase activities (PGP degradation; (e)). Overlay of the X-ray structures of ligand-LTA₄H structures of (f) SC57461 (blue; 3U9W.pdb) and (g) DG-051 (pink; 3FH7.pdb) with that of the stable PGP analogue OBP-Pro (white, 4MS6.pdb). The protein surface is coloured by amino acid hydrophobicity from most hydrophobic (orange) to hydrophilic (blue) through white regions. (h) JNJ-40929837 (green) was docked into the protein with Autodock Vina. Mice were orally administered SC57461A (i and j) or JNJ-40929837 (k and l) and serum aminopeptidase activity (i and k) or absolute PGP concentrations (j and l) were determined. IC₅₀ curves were generated from 3 technical replicates at each concentration of test compound and at least 2 biological replicates. Data for in vivo studies (i–l) are representative of 2 experiments with 5 mice per group. Results depicted as mean ± SEM. *P < 0.05, **P < 0.01 using Mann–Whitney statistical test.

Figure 2. Compounds with a resveratrol core are selective epoxide hydrolase inhibitors.

(a) LTA₄H contains two overlapping binding Sites A and B with previous X-ray crystallography studies revealing that two small molecules are capable of binding in each of the sites simultaneously (3FTX.pdb). Peptidase inhibitor Bestatin occupies Site B and 2H-Resveratrol sits in Site A. The protein surface is coloured by amino acid hydrophobicity from most hydrophobic (orange) to hydrophilic (blue) through white regions. LTA₄H epoxide hydrolase activity (b,c and f) and aminopeptidase activity (d and g) are presented following treatment with resveratrol (compound 1; b), compound 5 (c and d) and compound 9 (f and g). (e) Site A with docked pose of pinostilbene (compound 5; grey) overlaid on the X-ray structure of 2H-resveratrol bound to LTA₄H (blue; 3FTS.pdb). The protein surface is coloured by amino acid hydrophobicity from most hydrophobic (orange) to hydrophilic (blue) through white regions. The methyl ether exploits a hydrophobic patch in Site A, an interaction that is not favoured for Resveratrol and accounts for the increase in activity of compound 5 compared to Resveratrol. IC₅₀ curves were generated from 3 technical replicates at each concentration of test compound and at least 2 biological replicates. Results depicted as mean ± SEM.

Figure 3. Compounds 9 and 15 make a hydrogen bond with Asp-375 within Site A of the active site of LTA₄H.

(a) Green mesh surrounding Compound 9 (orange) shows the Fo-Fc difference map density (2σ level) having omitted the compound from the model. (b) Comparison of X-ray structure of 2H-resveratrol (blue; 3FTU.pdb) and compound 9 (bronze; our study) in Site A of LTA₄H. The protein surface is coloured by amino acid hydrophobicity from most hydrophobic (orange) to hydrophilic (blue) through white regions. (c) Relative positions of 2H-resveratrol (blue) and compound 9 (bronze) showing hydrogen bonding interactions. 2H-resveratrol H-bonds to the carbonyl of Trp-311 and 3 water molecules (pink) and its position is further reinforced by a π-stacking interaction with the indole of Trp-311. Compound 9 adopts a head-to-tail orientation such that its hydroxyl groups form H-bonds to Asp-375, key for hydrolase activity, and a single water molecule (orange). (d) Compound 9 (pink) does not impinge significantly on the peptidase site delineated by OBP-Pro (white; 4MS6.pdb) when the X-ray structures of Compound 9-LTA₄H (our study) and 4MS6.pdb are overlaid. Best scored poses of compound 12 (e; white) and compound 15 (f; orange) docked into site A of LTA₄H with Autodock Vina. The R2 OH of compound 12 is located in a hydrophobic region so replacing this group with a methyl substituent (compound 15) improves interaction with the protein. In addition, a hydrogen-bonding interaction between the phenol of compound 15 and Asp-375 is also observed.