Home Services Projects Software Resources Publications About Contact
  Case Study

Cheminformatics-Aided Discovery of Novel BACE Inhibitors for Alzheimer's Disease

How NovaMechanics identified potent multi-target drug candidates for Alzheimer's disease using structure-based and ligand-based computational methods, validated through in vitro testing and molecular dynamics simulations.

Co-funded by the European Regional Development Fund and the Republic of Cyprus through the Research and Innovation Foundation

The Challenge

Alzheimer's Disease: A Multi-Target Problem

Alzheimer's disease is the most common neurodegenerative disorder worldwide, affecting millions of patients with no disease-modifying treatment available. BACE1 (beta-secretase 1) is a key enzyme involved in the production of amyloid-beta peptides — a hallmark of Alzheimer's pathology. Additionally, tau protein aggregation plays a central role in neurodegeneration. Multi-target drug candidates that can inhibit both BACE1 and tau aggregation represent a promising therapeutic strategy, as they address multiple disease mechanisms simultaneously.

55M+
People living with dementia worldwide
No Cure
No disease-modifying treatment currently available
Dual Target
BACE1 + Tau aggregation for maximum therapeutic impact

Our Approach

A comprehensive computational pipeline from virtual screening to multi-target validation

Data Assembly

Curated 9,433 chemical entities from ChEMBL with BACE1 inhibitory activity data. After duplicate elimination and activity optimization, a final dataset of 6,599 unique compounds was compiled, with 68% classified as active (IC50 ≤ 1000 nM) and 32% inactive.

Ligand-Based Modeling

Developed predictive machine learning models using J48, kNN, and Random Forest algorithms. A consensus model achieved 85.9% accuracy, 91.4% sensitivity, and 85.1% precision for predicting BACE1 inhibitory activity.

Structure-Based Virtual Screening

Performed high-throughput virtual screening of 14,400 small molecules using molecular docking against the BACE1 crystal structure (PDB: 6BFE). Selected 35 commercially available compounds with satisfactory binding scores for biological testing.

Molecular Dynamics Simulations

Ran 50ns molecular dynamics simulations using Enalos Asclepios KNIME nodes to validate docking results and compute MM-GBSA binding free energies for top compounds. Confirmed stable protein-ligand interactions.

In Vitro BACE1 Inhibition Testing

Tested 35 compounds for BACE1 inhibitory activity. Three potent inhibitors identified: Compound 4 (IC50 = 160 nM — exceptional potency), Compound 7 (IC50 = 8.43 µM), and Compound 13 (IC50 = 28 µM). Lead optimization using chemical similarity searches identified additional analogue C5II.

Multi-Target Validation

Best compounds evaluated for tau aggregation inhibition, BBB permeability (PAMPA assay), antioxidant activity (DPPH assay), and cytotoxicity (MTT assay on human cells). Compound C5II showed potent tau aggregation inhibition (~60%, comparable to curcumin). Compound 4 showed excellent antioxidant activity. All compounds proved non-toxic.

Key Results

160 nM
IC50 of Compound 4
Exceptional BACE1 inhibitory potency
85.9%
Consensus Model Accuracy
Predicting BACE1 inhibition from molecular descriptors
Dual Activity
C5II Compound
Inhibits both BACE1 and tau aggregation (~60%)
Zero Toxicity
All Compounds
Non-toxic in human cell MTT assay at 24h
BBB Permeable
Compounds 13 & C5II
Show blood-brain barrier permeability via PAMPA
Antioxidant
Compound 4
Activity comparable to curcumin in DPPH assay

Powered by NovaMechanics Software

Enalos Asclepios KNIME Nodes

Used for molecular dynamics simulations and structure-based virtual screening workflows. Enabled automated, zero-code docking pipelines and binding free energy calculations via MM-GBSA.

Learn more

Isalos Analytics Platform / Enalos Cloud

Used for building consensus ML models (J48, kNN, Random Forest) for BACE1 inhibition prediction and hosting the BACE predictive web application with comprehensive model validation.

Learn more

This project identified multi-target drug candidates that qualify as disease-modifying ligands for Alzheimer's disease. The discovered compounds — particularly the exceptional BACE1 inhibitor C4 (160 nM) and the dual BACE1/tau inhibitor C5II — represent promising starting points for further optimization. These results demonstrate NovaMechanics' ability to deliver validated hit compounds through an integrated computational and experimental pipeline.