Purpose: In vitro drug metabolism measurements often underpredict in vivo hepatic drug clearance, with the degree of underprediction varying greatly by drug. Currently, in vitro to in vivo extrapolation (IVIVE) is used to physiologically scale-up the measured in vitro rate of drug loss (kinc) by correcting for differences in enzyme concentration in the incubation vs. the average human liver (i.e. scale-up kinc to kss). When calculating clearance (CL) though, volume (V) must also be considered as CL = k • V. In chemistry, where reactions occur in fixed-volumes, the straightforward IVIVE method is applicable; however, in pharmacokinetics, where the volume of distribution (Vd) varies drug by drug, traditional scale-up methods may not yield accurate results. While Vd is often examined for its role in drug disposition, the parameter’s role in the IVIVE of hepatic clearance has not yet been considered.
Methods: The theoretical basis of IVIVE was re-examined and a new relationship was derived to include the Vd component in the scale-up. An in vitro system was also designed to evaluate the impact of a drug’s in vitro Vd on in vitro measurements of drug metabolism.
Results: Examination of the theoretical basis for IVIVE revealed that the liver is assumed to be a homogeneous, fixed-volume organ, rather than a heterogeneous organ composed of both hydrophilic and lipophilic regions into which drugs can potentially distribute differently than metabolic enzymes. To account for drug specific scale-up of volume (in addition to the physiologic scale-up of kinc), a new relationship for IVIVE was derived that includes the parameter Rss, or the ratio of the volume of distribution of drug in the whole liver at steady-state to the volume of distribution of drug in the hepatocyte water (where drug metabolizing enzymes are assumed to reside). Rss is expected to vary from drug to drug, and provides at least a partial explanation for the poor and variable IVIVE predictions previously observed. To investigate this, an in vitro system was designed to simulate a heterogeneous liver. By repurposing a Rapid-Equilibrium Dialysis device, microsomal incubations were performed adjacent to either a lipophilic or hydrophilic phase, with a liquid impermeable membrane separating the two. It was confirmed that only drug can cross the membrane. BDDCS class 1 drugs (primarily metabolized with negligible transporter effects) were tested and kinc was measured with either an adjacent lipophilic (1-octanol) or hydrophilic (buffer/microsomes) phase. Differences between lipophilic and aqueous rate constants were observed; their ratio (k_lipophilic / k_aqueous) ranged from 0.16 – 0.89, and the extent of difference moderately correlated with drug partitioning (logP, r2=0.29). However, the correlation with published IVIVE fold-difference values was not as strong (r2=0.13).
Conclusion: Based on the new theoretical relationship for IVIVE, underprediction is due in part to drug-dependent distribution properties that are evident in vivo but are not currently accounted for with in vitro methods. The in vitro system designed for these studies does not necessarily replicate the complexities of the liver, rather highlights the importance of lipophilic distribution and Vd in CL and IVIVE. Our laboratory is currently investigating approaches to account for this drug distribution within the liver such as using animal scale-up of individual drugs.
Leslie Benet– Professor, UCSF School of Pharmacy, San Francisco, California