DECIPHERING THE MECHANISMS OF ALZHEIMER'S AND PARKINSON'S DISEASES USING NETWORK BIOLOGY AND A FUNCTIONAL GENOMICS APPROACH

Abstract

Neurodegenerative disorders are known to exhibit genetic overlap and shared pathophysiology. This study aims to find the shared genetic architecture of Alzheimer's disease (AD) and Parkinson's disease (PD), two major age-related progressive neurodegenerative disorders. The gene expression profiles of GSE67333 (containing samples from AD patients) and GSE114517 (containing samples from PD patients) were retrieved from the Gene Expression Omnibus (GEO) functional genomics database managed by the National Center for Biotechnology Information (NCBI). The web application GREIN (GEO RNA-seq Experiments Interactive Navigator) was used to identify differentially expressed genes (DEGs). 617 DEGs (239 upregulated and 379 downregulated) were identified from the GSE67333 dataset. Likewise, 723 DEGs (378 upregulated and 344 downregulated) were identified from the GSE114517 dataset. The protein-protein interaction (PPI) networks of the differentially expressed genes (DEGs) were constructed, and the top 50 hub genes were identified from the network of the respective dataset. Of the 4 common hub genes between the two datasets, CXCR4 was selected due to its gene expression signature profile and the same direction of differential expression between the two datasets. Mavorixafor was chosen as the reference drug due to its known inhibitory activity against CXCR4 and its ability to cross the blood-brain barrier. Molecular docking and molecular dynamics simulation of 51 molecules having structural similarity with Mavorixafor were performed to find two novel molecules, ZINC49067615 and ZINC103242147. Natural compounds are gaining prominence in the therapy of neurodegenerative disorders due to their biocompatibility and potential neuroprotective properties, including their ability to modulate CXCR4 expression. Recent advancements in artificial intelligence (AI) and machine learning (ML) algorithms have opened new avenues for drug discovery research across various therapeutic areas, including neurodegenerative disorders. We produced an ML model using cheminformatics-guided machine learning algorithms using data of compounds with known CXCR4 activity, retrieved from the Binding Database, to analyse diverse physicochemical attributes of natural compounds obtained from the COCONUT Database and predict their inhibitory activity against CXCR4.