Abstract

The potential of substrates and modifiers of CYP3A4 to show differential effects, attributed to the existence of multiple binding sites, confounds the straightforward prediction of in vivo drug-drug interactions from in vitro data. A set of in vitro interaction studies was performed in human lymphoblast-expressed CYP3A4 involving representatives of two CYP3A4 subclasses, midazolam (MDZ) and testosterone (TST); a distinct subgroup, nifedipine (NIF); and its structural analog, felodipine (FEL). Mechanistic insight into the interaction of each pair of substrates was provided by employing a range of multisite kinetic models; most were subtypes of a generic two-site model, but a three-site model was required for TST interactions. The complexity of the inhibition profiles and the selection of the kinetic model with appropriate interaction factors were dependent upon the kinetics of substrates involved (hyperbolic, substrate inhibition, or sigmoidal for MDZ/FEL, NIF, and TST, respectively). In no case was a simple reciprocity seen between pairs of substrates. The interaction profiles observed between TST, MDZ, NIF, and FEL involved several atypical inhibition features (partial, cooperative, concentration-dependent loss of characteristic homotropic behavior) and pathway-differential effects reflecting an 80-fold difference in K_i values and a δ factor (defining the alteration in the binding affinity in the presence of a modifier) ranging from 0.04 to 2.3. The conclusions from the multisite kinetic analysis performed support the hypothesis of distinct binding domains for each substrate subgroup. Furthermore, the analysis of intersubstrate interactions strongly indicates the existence of a mutual binding domain common to each of the three CYP3A4 substrate subclasses.