When several CYPs are involved in the metabolism of a compound, the use of cDNA-expressed systems may not always properly estimate the inhibitory effects of a given drug. Contradictory results on the correlation level of inhibitory potential using different fluoroprobes have been shown in the literature.65,76
In any case it remains a very useful tool for preliminary screening stages, completed later on selected compounds with conventional probes in models showing a full metabolic competence.99 When working on small series of compounds, human liver microsomes are the preferred test system to determine the inhibitory potential,46,65 because CYP kinetic measurements are not confounded with other metabolic processes or cellular uptake, as can be the case in hepatocytes.
The major issue in using human liver microsomes is the selection of highly specific CYP probes or index reaction (Table 5). The kinetic parameters should be assessed and the conditions of incubation should be validated by testing and comparing the inhibitory potential of known inhibitors with literature values.
In the past, the index reactions were monitored using several analytical techniques namely, LC coupled with UVV fluorescence, radiochemical, or MS/MS detections. Today a single LC/MS/MS technique can follow most of the CYP substrates and their respective metabolites within the same run, supplanting the standard use of radiolabeled substrates.42 A method is in use in our laboratory enabling us to follow 11 metabolic reactions, each specific for one human CYP42 in a single incubate.
IC50 or Ki generated in vitro can be corrected for nonspecific microsomal binding (not including plasma binding) in order to avoid an underestimation of the drug interaction risk or even false negatives. Tran et a/.77 have shown that the inhibition of several CYP3A4 inhibitors on diazepam metabolism decreased when microsomal proteins increased. This highlights the interest of correcting inhibition constants for microsomal binding.
These IC50 determinations are sufficient in early stages to extrapolate a preliminary risk of drug interaction using a pragmatic safety approach. They do not, however, allow one to distinguish the type of inhibition that is normally required to select the right model for drug interaction predictions (Table 2).
It can be assumed, as a simplification, that the vast majority of inhibitors involved in drug interactions are competitive inhibitors. In this case when the substrate concentration is equal to the Km, the Ki is theoretically equal to IC50/2.
Was this article helpful?