(640c) Identification of Apolipoprotein E4 Inhibitors for Alzheimer’s Disease Therapy through Large-Scale Virtual Screening

Alzheimer’s disease (AD), a neurodegenerative disorder, is characterized by its ability to cause memory loss and damage other cognitive functions. In 2015, it was estimated that almost 50 million people were affected by this disease, and this number is expected to increase in coming years as the increasing average life expectancy increases ¹. There are also concerns over the social and economic impact of the disease since the loss of function that affected individuals experience can lead to their dependence on outside care. Since current treatments mainly focus on managing symptoms without directly targeting AD pathology, there is a need to investigate new approaches and treatments for inhibiting potential protein targets involved in AD². The two main pathological hallmarks of AD are the amyloid beta (Aβ) and tau proteins which accumulate in the form of plaques and neurofibrillary tangles, respectively². Aggregation of beta amyloid plaques (Aβ) and neurofibrillary tangles in brain are responsible for development of AD. Nonetheless, many attempts targeting Aβ and tau proteins have been proven mostly unsuccessful³. Currently, evidences mounts that targeting Apolipoprotein E ε4 may combat Alzheimer’s Pathogenesis⁴. Apolipoprotein E has three different major alleles that are found in humans, leading to different protein isoforms, apoE2, apoE3, and apoE4. APOE ε4 is the strongest genetic risk factor for sporadic Alzheimer’s disease⁵. ApoE4 has been shown to increase Aβ accumulation by directly binding to Aβ and stabilizing it as well as interfering with Aβ and tau protein clearance by competing with receptors⁵. In this study, Apolipoprotein E4 (ApoE4) was chosen as the target based on its role in the main pathological hallmarks of AD. The structure and function of ApoE4 protein were investigated by previous studies⁶. The crystal structures can be obtained from Protein Data Bank with code 6NCN and 6NCO. We aimed to discover small-molecule inhibitors binding to ApoE4 protein to provide potential therapeutical options for Alzheimer’s disease.

Since there are millions of chemical compounds available in compound databases, such as FDA approved drugs that were listed in the DrugBank database, and experimental compounds that were registered and are available for purchase through ChemBridge, it is nontrivial to conduct experimental screening of each chemical compound against ApoE4. Therefore, a computational approach was then developed to identify small-molecule inhibitors of ApoE4 protein. Ligand-protein docking was conducted large-scale virtually screening. Among the existing ligand-protein docking, such as AutoDock, FlexX, ICM and GOLD, ICM was reported with the best performance⁷. It was able to predict structures for 93% complexes within the acceptable accuracy. Therefore, ICM was used as the docking platform to screen potential small molecule inhibitors for the target protein ApoE4 from large compounds databases. The crystal ligand was redocked in ApoE4 protein and got binding energy score -18.01 kcal/mol. In the protein-ligand docking software, lower scores indicate stronger binding affinity between the compounds and ApoE4. Compounds with binding scores below -20.0 were of special interest in this work, as they were predicted to have better binding affinity than all the crystal ligands, regardless of pocket. We virtually screened 1.5 million compounds from FDA-approved drug library and Chembridge database. By attempting to reposition FDA-approved drugs (approximately 2,500 drugs), resources are saved since identified compounds should already have data related to clinical trials and toxicity and be classified as safe for human use. The ChemBridge database was further used to evaluate approximately 1.5 million compounds binding to ApoE4 protein. The ChemBridge was a much more comprehensive database of drug-like compounds; however, since these compounds are experimental, they are not necessarily FDA-approved or extensively studied. Subsequently, chemical clustering program select putative inhibitors with diverse common substructures. The rationale for this is that compounds share similar structure may have similar properties. Therefore, the number of ApoE4 inhibitor candidates can be further reduced after compounds are clustered according to their structure similarity. An unweighted pair group method with arithmetic mean (UPGMA) clustering method was conducted to cluster compounds based on substructures. UPGMA, which achieved the best result compared to Butina-based clustering, is an agglomerative, hierarchal approach which is used for drug similarity studies⁸. We further selected inhibitors and extracted common structures from each group. The final candidates were evaluated on the following drug-like properties: partition coefficient (LogP), solubility (LogS), permeability through blood brain barrier (BBB score), predict CACO-2 permeability LogP (CACO2 score), toxicity (LD50).

From the DrugBank screening, there were 15 compounds that had a docking score below -20.0. In addition, several compounds were shown to bind to more than one pocket. Docking results ranged from -24.19 to -20.03 kcal/mol. From the ChemBridge screening, there were 1391 putative strong binders of ApoE4 were identified by docking-based virtual screening with docking score ranged from -27.35 to -20.02 kcal/mol. A key group of compounds that were identified as potentially effective were non-steroidal anti-inflammatory drugs (NSAIDs). Anti-inflammatory effects could be beneficial since neuroinflammation is a contributing factor to AD and its pathology¹. Numerous epidemiological studies in a range of ethnogeography populations have identified that NSAID use is associated with a lower risk of developing Alzheimer disease, and this association was suggested to be causal by numerous intervention-based studies in animal models⁹. Therefore, our findings suggested an interesting hypothesis that NSAID drugs might act as ApoE4 inhibitors against Alzheimer’s disease. The 1391 compounds identified from Chembridge were clustered based on UPGMA clustering method. In total, ten common substructures were identified. Since the compounds clustered in the same group share similar structure, core compounds from each group were extracted as the representative inhibitor. This results in a final hit list that contains 577 compounds. The putative inhibitors were found containing key scaffold/functional groups including benzene sulfonamide, diaromatic rings, pyrazole, thiophene, quinoline.

In conclusion, total 592 potential ApoE4 inhibitors were newly identified through computational approaches, which would provide drug candidates for further experiment assays. Therefore, in vitro testing is needed to confirm binding predictions; however, this could be challenging due to the stability of ApoE4. If in vitro testing were to confirm effective compounds, another potential direction could be to use data about binding and PROTAC technology to design a linker that would allow entire protein degradation rather than just inhibition.

References

1. Loera-Valencia, R. et al. Current and emerging avenues for Alzheimer’s disease drug targets. Journal of Internal Medicine (2019) doi:10.1111/joim.12959.
2. Cummings, J. New approaches to symptomatic treatments for Alzheimer’s disease. Molecular Neurodegeneration (2021) doi:10.1186/s13024-021-00424-9.
3. Congdon, E. E. & Sigurdsson, E. M. Tau-targeting therapies for Alzheimer disease. Nature Reviews Neurology (2018) doi:10.1038/s41582-018-0013-z.
4. Uddin, M. S. et al. APOE and Alzheimer’s Disease: Evidence Mounts that Targeting APOE4 may Combat Alzheimer’s Pathogenesis. Molecular Neurobiology (2019) doi:10.1007/s12035-018-1237-z.
5. Serrano-Pozo, A., Das, S. & Hyman, B. T. APOE and Alzheimer’s disease: advances in genetics, pathophysiology, and therapeutic approaches. The Lancet Neurology (2021) doi:10.1016/S1474-4422(20)30412-9.
6. Petros, A. M. et al. Fragment-Based Discovery of an Apolipoprotein E4 (apoE4) Stabilizer. J. Med. Chem. (2019) doi:10.1021/acs.jmedchem.9b00178.
7. Bursulaya, B. D., Totrov, M., Abagyan, R. & Brooks, C. L. Comparative study of several algorithms for flexible ligand docking. J. Comput. Aided. Mol. Des. (2003) doi:10.1023/B:JCAM.0000017496.76572.6f.
8. Kruger, F., Fechner, N. & Stiefl, N. Automated Identification of Chemical Series: Classifying like a Medicinal Chemist. J. Chem. Inf. Model. (2020) doi:10.1021/acs.jcim.0c00204.
9. Rivers-Auty, J., Mather, A. E., Peters, R., Lawrence, C. B. & Brough, D. Anti-inflammatories in Alzheimer’s disease—potential therapy or spurious correlate? Brain Commun. (2020) doi:10.1093/braincomms/fcaa109.