Study On Pharmacokinetic Prediction Model Of Angiotensin-Converting Enzyme Inhibitor

Pharmacokinetic prediction model can provide decision support for drugresearchers, approvers and investors in the early stage of new drug research anddevelopment. Angiotensin-converting enzyme inhibitor is one of the six majorfirst-line antihypertensive agents. This paper is to build pharmacokinetic predictionmodel of angiotensin-converting enzyme inhibitor based on physicochemicalproperties, with the purpose of guiding its new drug development.In this paper, pharmacokinetic literature data and physicochemical properties of20angiotensin-converting enzyme inhibitors were widely collected. Then the methodof combining curve estimation with multiple linear regression was used to build theprediction models of six pharmacokinetic parameters which include apparent volumeof distribution (Vd/F), apparent total clearance (CL/F), area under concentration-timecurve extrapolated to infinity (AUC0-∞), peak concentration (Cmax), peak time (tmax)and half-life (t1/2). Scatterplot matrix and correlation analysis were used to judge thechanging trends of pharmacokinetic parameters by the changes of physicochemicalproperties and the collinearity between physicochemical properties respectively.Meanwhile, statistical plots were used for the hypothesis verification. Finally,Leave-One-Out was applied to validate the model prediction capability.The six pharmacokinetic prediction models of angiotensin-converting enzymeinhibitor based on physicochemical properties were successfully built in this paper.They are LogVd/F=0.206+0.098LogD2, LogCL/F=-0.149+0.156LogD+0.045LogD2, LogAUC0-∞=1.385-0.156LogD-0.049LogD2, LogCmax=1.120-0.088LogD2, Tmax=9.338-0.219PSA+0.001PSA2+0.016pKa22and Logt1/2=0.182-0.178LogP+0.036LogP2. The determination coefficients reached0.799,0.722,0.757,0.831,0.926and0.825. The correlation coefficients between predicted andexperimental values were0.894,0.850,0.870,0.911,0.962and0.908. Thecross-validation correlation coefficients reached0.765,0.636,0.673,0.776,0.744and0.691validated by Leave-One-Out. The results showed good goodness of fit andprediction performance.The six pharmacokinetic prediction models may be useful for pharmacokineticparameters prediction of new angiotensin-converting enzyme inhibitors in the initial stage of drug development based on LogD, LogP, PSA and pKa2. This can be taken asthe basis of screening candidate compound quickly and optimizing molecularstructure reasonably, so as to shorten the new drug development cycle, reduce costand risk and improve efficiency and success rate.