Computational Screening of 1,348 FDA-Approved Drugs as Potential BACE1 Inhibitors for Alzheimer's Disease Drug Repurposing

Technical Summary: Computational Screening of FDA-Approved Drugs for BACE1 Inhibition

Problem Statement
Alzheimer's disease (AD) affects over 55 million individuals globally, yet no approved therapy modifies the underlying disease course. The amyloid cascade hypothesis posits that the accumulation of amyloid-beta (Aβ) peptides drives neurodegeneration. Beta-secretase 1 (BACE1) is the enzyme responsible for the rate-limiting cleavage of amyloid precursor protein (APP) to initiate Aβ production. While BACE1 is a validated pharmacological target, all major clinical trials of BACE1 inhibitors (e.g., verubecestat, atabecestat) have failed to demonstrate cognitive benefits, often due to safety concerns or insufficient efficacy. Consequently, there is an urgent need to identify novel inhibitors with improved selectivity and structural profiles. Drug repurposing—identifying new uses for FDA-approved compounds—offers a strategy to accelerate discovery by leveraging existing safety data, though systematic screening of the entire approved pharmacopoeia against BACE1 has not been previously conducted.

Methodology
This study employed a high-throughput, unbiased computational screening approach to evaluate 1,348 FDA-approved small molecules against the BACE1 target.

• Target Structure: The three-dimensional structure of human BACE1 (UniProt P56817) was retrieved from the AlphaFold2 Protein Structure Database. Only regions with a predicted local distance difference test (pLDDT) score >70 were utilized to ensure structural reliability.
• Compound Library: The complete library of 1,348 FDA-approved drugs was sourced from the ZINC15 database in 2D SDF format. Structures were converted to 3D PDBQT format using Open Babel with MMFF94 force field optimization.
• Docking Protocol: A custom Python-automated pipeline executed blind molecular docking using AutoDock Vina 1.2.5. A 40×40×40 Å search volume centered on the geometric centroid of the BACE1 structure encompassed the entire protein surface to identify optimal binding sites without bias toward the known catalytic cleft.
• Analysis: Binding affinity scores (kcal/mol) were extracted and ranked. A significance threshold of −7.0 kcal/mol was applied to identify meaningful interactions, with scores below −10.0 kcal/mol classified as "very strong" binders. Results were benchmarked against published docking scores for clinical-stage BACE1 inhibitors.

Key Results
The screening yielded 1,325 successful docking results. Of these, 809 compounds (60%) met the significance threshold (≤ −7.0 kcal/mol), and 19 compounds (1.4%) achieved "very strong" binding scores (≤ −10.0 kcal/mol).

• Top Hits: The three highest-scoring compounds were Hepatitis C (HCV) antiviral agents:
  • Paritaprevir: −12.48 kcal/mol (NS3/4A protease inhibitor)
  • Ledipasvir: −10.88 kcal/mol (NS5A inhibitor)
  • Simeprevir: −10.78 kcal/mol (NS3/4A protease inhibitor)
    These scores exceeded the published docking scores of known clinical-stage BACE1 inhibitors (e.g., verubecestat at −10 to −11 kcal/mol).
• Cross-Validation: Nilotinib (a BCR-ABL kinase inhibitor for leukemia), which has independent preclinical and clinical evidence for reducing tau and amyloid pathology in AD, ranked highly at −10.59 kcal/mol.
• Pilot Confirmation: Sertraline (an SSRI), identified in a preliminary six-compound pilot screen at −7.50 kcal/mol, was confirmed in the full library screen.
• Structural Pattern: The clustering of HCV antivirals at the top of the ranking suggests a structural basis for BACE1 affinity, likely due to the shared requirement for large, macrocyclic scaffolds to occupy deep protease binding clefts (HCV NS3/4A and BACE1 are both proteases).

Significance and Claims
The authors claim this is the first comprehensive, unbiased computational screen of the entire FDA-approved drug library against the AlphaFold2-predicted BACE1 structure. The study's primary significance lies in identifying Hepatitis C antivirals as a novel, structurally compelling class of BACE1 repurposing candidates, a class never previously associated with BACE1 inhibition.

The paper posits that the high docking scores of Paritaprevir, Ledipasvir, and Simeprevir suggest a structural complementarity between HCV protease inhibitor scaffolds and the BACE1 active site. Furthermore, the inclusion of Nilotinib among the top hits serves as methodological cross-validation, indicating the screen can identify drugs with known biological relevance to AD pathology. The authors conclude that while these findings are computational predictions requiring experimental validation (e.g., in vitro FRET assays, BBB penetration studies), they establish a strong rationale for prioritizing HCV antivirals and Nilotinib for immediate experimental investigation as potential disease-modifying therapies for Alzheimer's disease.