Thein vitro–in vivocorrelations have been focused on the use of hepatocytes
and liver microsomal preparations to predict those parameters describing
metabolic stability, namely, intrinsic clearance (CLint) and hepatic clearance
(CLh) (Houston and Galetin, 2003), supported by the successful in vivo
predictions reported for the clinical drugs such as verapamil, diazepam,
lidocaine, and phenacetin (Bargetzi et al., 1989; Iwatsubo et al., 1997; Kroemer
et al., 1992; Seddon et al., 1989; Smith and Timbrell 1974). Basically, two
stepwise processes are involved in the prediction of PK properties using thesein
vitrodata: scaling-up from CLintdeterminedin vitro(CL
0
int), and estimates of
the PK parameters using one of the previously-mentioned mathematical
models.
CL
0
intcan be estimated via either the depletion of substrate or the formation
of metabolites as described. Substrate disappearance, based on presumed
exponential decay is often the choice in early discovery. Mechanisms of
metabolite formation, founded primarily on the understanding of enzyme
kinetic behavior, can be complicated, and is typically the preference in later
discovery and development. In fact,CL
0
intis appropriately determined using
either approach, given the consistency of their results (Obach and Reed-Hagen,
2002).
13.5.2 Procedure ofIn vitro–In vivoCorrelation
13.5.2.1 Scaling-Up
Weighing As previously described,CL
0
intis equivalent toVmax/Km. However,
compounds frequently undergo metabolism via more than one metabolic
pathway, possibly mediated by several enzymes. The relationship between
CLintandVmax/Kmhas accordingly been modified:
CL
0
int¼SCL0
int;i¼SðEi=ETÞVmax;i=Km;i: ð^13 :^12 ÞThe values ofVmax,iandKm,iare assigned to one of the relevant enzymesi
involved in one of the primary biotransformation pathways (Tang et al., 2001).
Eiis the quantity (or level of protein expression) of the enzyme i, whileETis the
total quantity (or the total protein expression level) of all relevant enzymes.
Therefore,Ei/ET= 1 when only one enzyme is significantly contributing to the
overall metabolism. As discussed previously, the kinetic parameters (Kmand
Vmax) for individual contributing enzymes can be determined using either the
classic biochemical plots or, preferably, nonlinear regression analyses.
Alternatively, CL
0
int can be estimated by measuring the rates at a single,
relatively low, substrate concentration, for example, 1mM. The substrate
concentration relevant to the anticipated therapeutic levels will presumably be
much lower than theKm.
PREDICTION OF HUMAN HEPATIC CLEARANCE 435