Chemometric modeling of PET imaging agents for diagnosis of Parkinson’s disease: a QSAR approach

Recently, adenosine A2A receptor antagonists have been identified as an interesting drug target for the treatment of Parkinson’s disease (PD). Radiolabelled molecular imaging technologies such as positron emission tomography (PET) have emerged in the research field of medicinal chemistry as a diagnostic tool for PD. In the current study, we have performed quantitative structure–activity relationship (QSAR) analysis of 35 xanthine ligand PET tracers as A2AR (adenosine receptors) antagonists in order to determine their structural features required to have binding affinity and selectivity towards A2AR. The division of the dataset into training and test sets was done using a random method, while the feature selection for the binding affinity was done using Genetic Algorithm (GA). The best model with five descriptors was obtained using the spline option in the GA run. QSAR models with four descriptors were also developed for A2AR selectivity, where significant descriptors were selected from the large pool of descriptors using stepwise regression method followed by Best Subset Selection (BSS) method. Furthermore, to improve the quality of the external predictions, we used the “Intelligent Consensus Predictor” tool (http://teqip.jdvu.ac.in/QSAR_Tools/DTCLab/). Both the models showed robustness in terms of statistical parameters. Molecular docking studies have been carried out to understand the molecular interactions between the ligand and receptor, and the results are then correlated with the structural features obtained from the QSAR models. Furthermore, the information derived from the newly found descriptors gives an insight for the development of new candidate PET tracers for the use in PD.