Intravitreal clearance and volume of distribution of compounds in rabbits: In silico prediction and pharmacokinetic simulations for drug development.

The aims of this research were to (1) create a curated universal database of intravitreal volumes of distribution (Vss, ivt) and clearances (CL ivt) of small molecular weight compounds and macromolecules and (2) to develop quantitative structure property relationship (QSPR) and pharmacokinetic models for the estimation of vitreal drug concentrations based on the compound structure. Vss, ivt and CL ivt values were determined from the available literature on intravitreal drug administration using compartmental models and curve fitting. A simple QSPR model for CL ivt of small molecular weight compounds was obtained with two descriptors: Log D7.4 and hydrogen bond donor capacity. The model predicted the internal and external test sets reliably with a mean fold error of 1.50 and 1.33, respectively (Q(2)Y=0.62). For 80% of the compounds the Vss, ivt was 1.18-2.28 ml; too narrow range for QSPR model building. Integration of the estimated Vss, ivt and predicted CL ivt parameters into pharmacokinetic simulation models allows prediction of vitreous drug concentrations after intravitreal administration. The present work presents for the first time a database of CL ivt and Vss, ivt values and the dependence of the CL ivt values on the molecular structure. The study provides also useful in silico tools to investigate a priori the intravitreal pharmacokinetic profiles for intravitreally injected candidate compounds and drug delivery systems.