A lead compound which had sub-micromolar affinity for the rabbit NK1 receptor but negligible affinity for rat NK1 receptors, 3a, was discovered by directed screening. 2-Substitution in the ring of the benzylthiourea substituent in the initial lead was found to be important, and halogens (Cl, Br) in this position were found to improve affinity for the human receptor. The activity of a series of 2-halo-substituted benzylthioureas was then optimized by modification of the proline diphenylmethyl amide, guided by a simple conceptual model based on structural overlay between these early antagonists and NK1 selective peptides. In this way, aromatic amino acid amides were identified which had improved affinity with respect to the starting diphenylmethyl (DPM) amides. The first sub-nanomolar ligand for the human NK1 receptor which arose from this series, 4af, combined a 2- chlorobenzylthiourea unit with a 2-naphthylalanine amide. Contemporaneously it was discovered that the benzylthiourea unit could be simplified to a phenylthiourea providing that an appropriate 2-substituent was also incorporated. Combination of these two series gave 2-NO2 phenylthiourea analogues which led directly to the analogous urea, 5f (2- nitrophenylcarbamoyl-(S)-prolyl-(S)-3-(2-naphthyl)alanyl-N-benzy-N- methylamide, SDZ NKT 343), a highly potent ligand for the human NK1 receptor (K(i) = 0.16 nM). In addition to its high in vitro potency, 5f proved to be a potent orally active analgesic in guinea pig models of chronic inflammatory and neuropathic pain. The nature of the 2-aryl substituent was found to be critical for oral activity in this series. Clinical evaluation of 5f as a novel analgesic agent is currently underway.