engleski

QSAR Studies in a Series of Protein Tyrosine Kinase Inhibitors

Recent years have witnessed a remarkable progress in the medicinal chemistry design of selective protein tyrosine kinase inhibitors (PTKIs) because these drugs block grow signals in the cell and consequently stop the cell growth and cell dividing. Over 20 kinase inhibitor drugs are available on the market and numerous PTK inhibitors are currently in discovery process and preclinical phases. However the number of inhibitors that has been approved for the market still remains low. The present paper, aimed at obtaining a deeper insight into the structure-properties relationships of this class of molecules, reports the results of QSAR studies between computed molecular descriptors and predicted ADMET properties of a series of PTKIs (n = 28) that are in clinical use or in preclinical phases of drug development. Structural features of PTKIs, derivatives of quinoline, quinazoline, pyrido- and pyrimido-pyrimidine, were explored using molecular descriptors (MDs), topological indices (TIs), drug-likeness scores (DLs) and ADMET parameters. MDs and DLs of investigated PTKIs were calculated using Molinspiration engines v2013.09 and v2011.06. TIs were computed using Chemicalize.org/Properties Viewer and ADMET properties were predicted by MedChem StudioTM and ADMET PredictorTM 7.0 (Simulations Plus, Inc., USA). All analyses were performed by OriginPro 8.0 (Origin Laboratories, USA). The study results revealed that the highest scores for kinase inhibitor likeness (KI DLs 0.90-1.27) were computed for pyrimido[5.4-d]pyrimidin-4-amine and pyrido[3.4-d]pyrimidin-4, 6-diamines. For these compounds DLs with GPCR ligand (GPCR-l DLs 0.21-0.45), ion channel modulator (ICM DLs 0.22-0.33) and enzyme inhibitor (EI DLs 0.21-0.36) were also computed. Lower values of KI DLs (0.36-0.74) were computed for quinazoline derivatives. The results of correlation studies revealed significant correlation coefficients (R = 0.8869–0.9873) between MDs (Mr, V, TPSA) and topological indices (TIs), i.e. Wiener number (W), Randić connectivity index (X1) and Szeged index (Sz). No significant correlations were found between MDs, TIs or ADMET parameters and biological activity (IC50) in a series of investigated anilinoquinazolines. ADMET Predictor analyses of PTKIs with multiple DLs revealed that these molecules are CYP 2D6 and CYP 3A4 substrates, with ADMET Risk 3 or 4, CYP Risk 1 and TOX Risk 3. The multi-drug-likeness scores computed for PTKIs with highest KI DLs indicate the potential multi-target activity of these drugs with possible anti-target outcomes, so these findings should be taken into consideration for the evaluation of their potential toxicity. References: 1. Mishra H et al. 2009, A comparative study on the molecular descriptors for predicting drug-likeness of small molecules, Bioinf. 3, 384-388. 2. Han C et al. 2011, An integrated drug-likeness study for bicyclic privileged structures: from physicochemical properties to in vitro ADME properties, Mol. Divers. 15, 857-876.

Protein kinase inhibitors; molecular descriptors; drug-likeness; QSAR

BM048 - Potpora Sveučilišta u Zagrebu, biomedicinsko područje, projekt 'Istraživanje značajki inhibitora protein tirozin kinaze i histon deacetilaze u antitumorskoj terapiji' (voditelj M. Jadrijević-Mladar Takač)