Design, synthesis, and biological characterization of a peptide-mimetic antagonist for a tethered-ligand receptor

Abstract

Protease-activated receptors (PARs) represent a unique family of seven-transmembrane G protein-coupled receptors, which are enzymatically cleaved to expose a truncated extracellular N terminus that acts as a tethered activating ligand. PAR-1 is cleaved and activated by the serine protease alpha-thrombin, is expressed in various tissues (e.g., platelets and vascular cells), and is involved in cellular responses associated with hemostasis, proliferation, and tissue injury. We have discovered a series of potent peptide-mimetic antagonists of PAR-1, exemplified by RWJ-56110. Spatial relationships between important functional groups of the PAR-1 agonist peptide epitope SFLLRN were employed to design and synthesize candidate ligands with appropriate groups attached to a rigid molecular scaffold. Prototype RWJ-53052 was identified and optimized via solid-phase parallel synthesis of chemical libraries. RWJ-56110 emerged as a potent, selective PAR-1 antagonist, devoid of PAR-1 agonist and thrombin inhibitory activity. It binds to PAR-1, interferes with PAR-1 calcium mobilization and cellular function (platelet aggregation; cell proliferation), and has no effect on PAR-2, PAR-3, or PAR-4. By flow cytometry, RWJ-56110 was confirmed as a direct inhibitor of PAR-1 activation and internalization, without affecting N-terminal cleavage. At high concentrations of alpha-thrombin, RWJ-56110 fully blocked activation responses in human vascular cells, albeit not in human platelets; whereas, at high concentrations of SFLLRN-NH(2), RWJ-56110 blocked activation responses in both cell types. Thus, thrombin activates human platelets independently of PAR-1, i.e., through PAR-4, which we confirmed by PCR analysis. Selective PAR-1 antagonists, such as RWJ-56110, should serve as useful tools to study PARs and may have therapeutic potential for treating thrombosis and restenosis.

Figure 1

Structures of peptide-mimetic PAR-1 antagonists.

Figure 2

Synthesis of prototype RWJ-53052 (5). Reaction conditions (yield): a, BnNH₂, bis(2-oxo-3-oxazolidinyl)phosphinic chloride, CH₂Cl₂, 0°C (83%); b, 20% (vol/vol) piperidine in 1,4-dioxane (99%); c, Fmoc-4-methoxyphenylalanine, 1,3-diisopropylcarbodiimide, HOBt, CH₂Cl₂ (84%); d, 4-fluorobenzyl bromide, Cs₂CO₃, DMF; e, FeCl₃⋅6 H₂O, charcoal powder, Me₂NNH₂, MeOH, reflux (93% for 2 steps); f, 4-nitrophenylchloroformate, i-Pr₂NEt, CH₂Cl₂, −20°C—added 2 and warmed to 23°C (65%); g, 37% aqueous CH₂O, pyrrolidine, AcOH (63%); h, CF₃CO₂H/CH₂Cl₂ (97%). Abbreviations: Fmoc, 9-fluorenylmethoxycarbonyl; Pmc, 2,2,5,7,8-pentamethylchroman-6-sulfonyl; HOBt, 1-hydroxybenzotriazole; DMF, dimethylformamide.

Figure 3

Solid-phase parallel synthesis of analogues of RWJ-50352 (sphere = 2-chlorotrityl resin). Method A: a, BnNH₂, DCC, HOBt, MeCN; b, CF₃CO₂H, CH₂Cl₂; c, 2-Cl-trityl-Cl resin, i-Pr₂NEt, CH₂Cl₂-DMF; d, piperidine, DMF; e, Fmoc-3,4-diF-Phe-OH, HBTU, HOBt, i-Pr₂NEt, DMF; f, 6-amino-1-(2,6-dichlorobenzyl)indole (prepared as for 4, Fig. 2), 4-nitrophenylchloroformate, i-Pr₂NEt, CH₂Cl₂, −20°C; g, pyrrolidine, 37% aqueous CH₂O, HOAc, 1,4-dioxane. The overall purified yield from resin 7 was 76%. Method B: trityl resin-based synthesis. Abbreviations: DCC, 1,3-dicyclohexylcarbodiimide; HBTU, 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate; see Fig. 2.

Figure 4

(A) Three-point model showing spatial distribution of the key ammonium, phenyl, and guanidinium groups in the agonist peptide SFLLRN-NH₂ (doubly protonated form). Structural models were assembled computationally in extended and α-helical geometries by using the protein build command within the biopolymer mode of the sybyl software suite (Tripos Associates, St. Louis). The structures were energy minimized (Kollman United Atoms force field; conjugate gradient method) and subjected to molecular dynamics searches (sybyl; ref. 23). This search provided ranges of distances between the three groups, which were used as a starting point for peptide-mimetic design. (B) General formula for template-based SFLLRN-mimetic compounds, based on the distances in A.

Figure 5

Inhibitory effects of RWJ-56110 on increasing concentrations of PAR-1 agonists. Platelet aggregation induced by thrombin (A) or by SFLLRN-NH₂ (B). Platelet calcium mobilization induced by thrombin (C) or by SFLLRN-NH₂ (D). Calcium mobilization induced by thrombin in lung myofibroblasts from PAR-1-deficient mice, transfected with human PAR-1 (E), or in RASMC (F).

Figure 6

Expression of PAR-4. Southern blot analysis of PCRs on cDNA samples from RNA of HASMC, gel-filtered platelets, or CHRF-288-11 cells with RT. Control PCRs were done in parallel on each untreated RNA sample (−RT). The primers for the PCRs, the products of which were Southern blotted and probed with the appropriate nested primer probe, corresponded to PAR-4 and the positive control β-actin.

Figure 7

Inhibition of internalization of cleaved PAR-1 by RWJ-56110. CHRF-288-11 cells were pretreated with increasing concentrations of RWJ-56110 with or without α-thrombin. The surface expression of PAR-1 was monitored by flow cytometry with SPAN12 (solid bars) and ATAP2 (open bars).