Browsing by Subject "QSAR"

QSPR (Quantitive structure property relationship) describes relationship between descriptors and biological activity. Therefore, QSPR models are useful tools in drug discovery. The literature review summarizes existing corneal, intestinal and blood brain barrier permeability models. The most common descriptors are hydrophobicity, polar surface and H-bonding capability. Also, the size of molecule may have influence on permeability even though the results are sometimes contradictory. Descriptors might have limiting values such as those presented in Lipinski's ‖rule of five‖. Drug candidate should not have 'rule of 5' values outside of the useful range, otherwise the per oral absorption of the compound may be compromised. In the literature review the transporter activity in cornea, intestine and blood brain barrier is described. Currently, many QSPR-models have been developed to predict interactions of drug candidates with transporters. The purpose of experimental part was to build in silico -model of corneal passive permeability for early ocular drug discovery. QSPR-model was built using permeability data and molecular descriptors of 54 molecules. Corneal permeability coefficients in rabbits were obtained from the literature. Octanol-water partitition coefficient at pH 7,4 (logD) and the total number of hydrogen bonds were the descriptors in the final model. The final equation was log(permeability coefficient) = -3,96791 - 0,177842*Htotal + 0,311963*logD(pH7,4). For this model R2 was 0,77 ja Q2 was 0,75. The model was evaluated using an external data set of 15 compounds and by pharmacokinetic modeling. Predicted permeability coefficients were used to simulate the aqueous humour concentrations of sevent compounds at steady-state. In addition corneal absorption coefficient (Kc) was simulated for 13 compounds and these values were compared to predicted permeability. The predicted permeability coefficients correlated well with experimental permeability coefficients. In addition aqueous humour concentrations can be simulated in steady state using predicted (QSPR) permeability coefficients. The final QSPR-model may be used in ocular drug discovery and development.