Combined Molecular and Machine Learning Models to Probe the Activation Potential of ACE2 in Pulmonary Arterial Hypertension (PAH)

Abstract

Pulmonary Arterial Hypertension (PAH) is a severe cardiovascular disorder, characterized by high blood pressure. Untreated PAH can lead to heart failure and disrupt lung functions. Many recent studies have suggested that an altered renin-angiotensin-aldosterone (RAAS) can be a causative factor in PAH pathogenesis. Therefore, an increased level of Angiotensin II (Ang II) has been associated with the development of PAH. Previously, various ACE inhibitors (ACEI) have been proposed as potential drug candidates to mitigate the detrimental effects of Ang II. ACE2, a recently discovered homolog of ACE, opposes the effect of Ang II by converting Ang II into Ang 1-7. Briefly, targeting the activation of ACE2 by ACE inhibitors may act as a counterbalance to the effect of Ang II. In this study, we employed molecular docking guided machine learning models to predict the binding potential of ACE inhibitors to activate ACE2 for PAH treatment. Predictive machine learning models were implemented on docked complexes of ACE inhibitors for the prediction of the activation potential of ACE2. Support Vector Machine (SVM) and Artificial Neural Network (ANN) models accurately classified ACE inhibitors and ACE2 activators with overall accuracies of 99.57% and 90.69%, respectively. Ligands with ChEMBL ids, CHEMBL273140 and CHEMBL10521, demonstrated the most effective dual functionality as both ACE inhibitor and ACE2 activator. Our findings aided in understanding the binding attributes of ACE2 activators at the molecular level, which can assist in developing novel pharmaceutical agents for the treatment of PAH.