6.6.1. Intestinal Absorption and Bioavailability
Running an HTPK simulation to calculate intestinal absorption and/or bioavailabilty is very fast. This makes it a valuable tool for compound optimization and prioritization in a discovery setting where values are needed for large numbers of compounds and wet chemistry data is limited or absent altogether. %Fa and %Fb can be calculated from in silico or in vitro input parameters or some combination thereof.
HTPK simulations run in ADMET Predictor assume perfusion rate-limited kinetics for the liver and kidney, which are the archetypal clearance routes. The default renal clearance is calculated from the glomerular filtration rate and the fraction unbound in plasma (fup*GFR) while the hepatic clearance is calculated using the microsomal or hepatocyte clearance value supplied as an input to the program.
Hepatic plasma clearance is calculated from the fraction of drug unbound in plasma (fup), ratio of blood to plasma (Rbp), blood flow to the organ (QH) and the unbound intrinsic clearance (CLint,u):
Please see the Intrinsic Clearance section for a description of how the intrinsic clearance is corrected for non-specific binding: i.e., how the bound intrinsic clearance is converted to an unbound intrinsic clearance. As stated earlier, the mechanistic corrections described in the Intrinsic Clearance section occur only when the user has selected <Mechanistic> as the %unbound option in the dialog. Note that the equation describing the liver microsome correction uses the fumic_hum model (for human), but in HTPK this can be changed using the RefFumic parameter in the relevant HIA file (SimHIA.hia, SimRIA.hia, SimMIA.hia, SimDIA.hia or SimMKYIA.hia).
Metabolism in the gut and liver can be nonlinear processes. Therefore, the fraction of drug extracted due to first pass (and hence the bioavailability) can vary with the administered dose. If the fraction of first pass extraction (FPE) and the fraction absorbed (%Fa) are known, then the fraction bioavailable (%Fb) can be calculated as
HTPK simulations use a somewhat simplified version of the ACAT model used in GastroPlus, so small differences in estimated %Fa and %Fb between the two programs are expected. The observed agreement between bioavailability estimates obtained from GastroPlus and those generated by the HTPK Simulation Module is therefore quite remarkable: the root mean square deviation (RMSD) between predictions was only 4.5% for a group of 118 passively permeable compounds largely drawn from Zhao et al. (2001).