Abstract:Parkinson's disease is characterized by a multifactorial nature that is linked to different pathways. Among them, the abnormal deposition and accumulation of α-synuclein fibrils is considered a neuropathological hallmark of Parkinson's disease. Several synthetic and natural compounds have been tested for their potency to inhibit the aggregation of α-synuclein. However, the molecular mechanisms responsible for the potency of these drugs to further rationalize their development and optimization are yet to be determined. To enhance our understanding of the structural requirements necessary for modulating the aggregation of α-synuclein fibrils, we retrieved a large dataset of α-synuclein inhibitors with their reported potency from the ChEMBL database to explore their chemical space and to generate QSAR models for predicting new bioactive compounds. The best performing QSAR model was applied to the LOTUS natural products database to screen for potential α-synuclein inhibitors followed by a pharmacophore design using the representative compounds sampled from each cluster in the ChEMBL dataset. Five natural products were retained after molecular docking studies displaying a binding affinity of - 6.0 kcal/mol or lower. ADMET analysis revealed satisfactory properties and predicted that all the compounds can cross the blood-brain barrier and reach their target. Finally, molecular dynamics simulations demonstrated the superior stability of LTS0078917 compared to the clinical candidate, Anle138b. We found that LTS0078917 shows promise in stabilizing the α-synuclein monomer by specifically binding to its hairpin-like coil within the N-terminal region. Our dynamic analysis of the inhibitor-monomer complex revealed a tendency towards a more compact conformation, potentially reducing the likelihood of adopting an elongated structure that favors the formation and aggregation of pathological oligomers. These findings offer valuable insights for the development of novel α-synuclein inhibitors derived from natural sources.

Using Generative Modeling to Endow with Potency Initially Inert Compounds with Good Bioavailability and Low Toxicity

Exploration and Exploitation Approaches Based on Generative Machine Learning to Identify Potent Small Molecule Inhibitors of α-Synuclein Secondary Nucleation

Probing the molecular mechanisms of α-synuclein inhibitors unveils promising natural candidates through machine-learning QSAR, pharmacophore modeling, and molecular dynamics simulations

Optimizing Drug Design by Merging Generative AI With Active Learning Frameworks

Improving drug discovery with a hybrid deep generative model using reinforcement learning trained on a Bayesian docking approximation

De novo generation of multi-target compounds using deep generative chemistry

In-Pocket 3D Graphs Enhance Ligand-Target Compatibility in Generative Small-Molecule Creation

Enhancing molecular design efficiency: Uniting language models and generative networks with genetic algorithms

Guided Multi-objective Generative AI to Enhance Structure-based Drug Design

Molecular generation strategy and optimization based on A2C reinforcement learning in de novo drug design

Multi-objective generative AI for designing novel brain-targeting small molecules

De novo generation of dual-target ligands using adversarial training and reinforcement learning

SynthFormer: Equivariant Pharmacophore-based Generation of Molecules for Ligand-Based Drug Design

Generative Active Learning for the Search of Small-molecule Protein Binders

aSynPEP-DB: a database of biogenic peptides for inhibiting α-synuclein aggregation

DrugGen: Advancing Drug Discovery with Large Language Models and Reinforcement Learning Feedback

Protein mimetic 2D FAST rescues alpha synuclein aggregation mediated early and post disease Parkinson’s phenotypes

Unlocking Novel Therapies: Cyclic Peptide Design for Amyloidogenic Targets through Synergies of Experiments, Simulations, and Machine Learning

Machine learning-aided generative molecular design

Leveraging molecular structure and bioactivity with chemical language models for de novo drug design

Prediction of Inhibitors Against Alpha-Synuclein Fibrils Formed in Parkinson’s Disease