Discovery, modeling, and human pharmacokinetics of N-(2-acetyl-4,6-dimethylphenyl)-3-(3,4-dimethylisoxazol-5-ylsulfamoyl)thiophene-2-carboxamide (TBC3711), a second generation, ETA selective, and orally bioavailable endothelin antagonist.

Sitaxsentan (1) (Wu et al. J. Med. Chem. 1997, 40, 1690) is our first endothelin antagonist being evaluated in clinical trials. It has demonstrated biological effects in an acute hemodynamic study in CHF (Givertz et al. Circulation 2000, 101, 2922), an open-label 20-patient pulmonary hypertension trial (Barst et al. Chest 2002, 121, 1860-1868), and a 31-patient trial in essential hypertension (Calhoun et al. AHA Scientific Sessions 2000). In a phase 2b/3 pulmonary arterial hypertension trial, once a day treatment of 100 mg of sitaxsentan statistically significantly improved 6-min walk distance and NYHA class at 12 weeks (Barst et al. Am. J. Respir. Crit. Care Med. 2004, 169, 441). We have since reported on our efforts in generating follow-up compounds (Wu et al. J. Med. Chem. 1999, 42, 4485) and recently communicated that an ortho acyl group on the anilino ring enhanced oral absorption in this category of compounds (Wu et al. J. Med. Chem. 2001, 44, 1211). Here we report an expansion of this work by substituting a variety of electron-withdrawing groups at the ortho position and evaluating their effects on oral bioavailability as well as structure-activity relationships. As a result, TBC3711 (7z) was identified as our second endothelin antagonist to enter the clinic due to its good oral bioavailability (approximately 100%) in rats, high potency (ET(A) IC(50) = 0.08 nM), and optimal ET(A)/ET(B) selectivity (441 000-fold). Compound 7z has completed phase-I clinical development and was well tolerated with desirable pharmacokinetics in humans (t(1/2) = 6-7 h, oral availability > 80%).

A 3D structure model of integrin alpha 4 beta 1 complex: I. Construction of a homology model of beta 1 and ligand binding analysis.

It is well established that integrin alpha 4 beta 1 binds to the vascular cell adhesion molecule (VCAM) and fibronectin and plays an important role in signal transduction. Blocking the binding of VCAM to alpha 4 beta 1 is thought to be a way of controlling a number of disease processes. To better understand how various inhibitors might block the interaction of VCAM and fibronectin with alpha 4 beta 1, we began constructing a structure model for the integrin alpha 4 beta 1 complex. As the first step, we have built a homology model of the beta 1 subunit based on the I domain of the integrin CD11B subunit. The model, including a bound Mg(2+) ion, was optimized through a specially designed relaxation scheme involving restrained minimization and dynamics steps. The native ligand VCAM and two highly active small molecules (TBC772 and TBC3486) shown to inhibit binding of CS-1 and VCAM to alpha 4 beta 1 were docked into the active site of the refined model. Results from the binding analysis fit well with a pharmacophore model that was independently derived from active analog studies. A critical examination of residues in the binding site and analysis of docked ligands that are both potent and selective led to the proposal of a mechanism for beta 1/beta 7 ligand binding selectivity.