Development of a generalized, quantitative physicochemical model of CYP3A4 inhibition for use in early drug discovery

To examine the structure-activity relationships for the inhibition of the activity of recombinant human CYP3A4 and to establish a generalized, quantitative physicochemical model for use in early drug discovery.

Inhibition of the activity of recombinant human CYP3A4 (erythromycin N-demethylase) by 30 diverse chemicals was studied using enhanced throughput methodology.

There was a general, strong correlation between the IC50 value determined against erythromycin N-demethylase activity and lipophilicity (LogD7.4) (r2 = 0.68, p <0.0001). This relationship was strengthened further by subdividing the structures studied into two distinct subpopulations of chemistry within the dataset. These could be identified by the absence (r2 = 0.80, p <0.0001) or presence (r2 = 0.69, p <0.0001) of a sterically uninhindered N-containing heterocycle, more specifically a pyridine, imidazole, of triazole function. The presence of these structural motifs increased the potency of CYP3A4 inhibition by approximately 10-fold for a given lipophilicity (LogD7.4.value). More detailed analyses of AstraZeneca compounds demonstrated that the inhibitory potency of the pyridine structure can be attenuated through direct steric effects or electronic substitution resulting in a modulation of the pKa of the pyridine nitrogen, thereby influencing its ability to interact with the CYP heme.

A generalized, quantitative model is proposed for the inhibition of the major drug metabolizing enzyme, CYP3A4. This model indicates the importance of lipophilicity and rationalizes increased potency arising through additional interactions with the heme iron. These general relationships were shown to be applicable to a selection of compounds of interest to several early research projects.