CC BY
3
CD ISSN 2310-5674 ppublishing.org
DOI: 10.29013/EJBLS-23-3-8-15
DESIGN OF ACETYLCHOLINESTERASE INHIBITORS FOR ALZHEIMER'S DISEASE
Jiawei Ren1
1 Lake Forest Academy, Illinois, USA
Cite: Jiawei Ren. (2023). Design of Acetylcholinesterase Inhibitors for Alzheimer's Disease. European Journal of Education and Applied Psychology 2023, No 3. https://doi. org/10.29013/EJBLS-23-3-8-15
Abstract
Introduction
Alzheimer's disease is a worldwide progressive neurodegenerative disease. In the past two decades, acetylcholinesterase (AChE) inhibitors have been the most popular medication in mitigating Alzheimer's disease symptoms. However, all common AChE inhibitors have side effects, and traditional inhibitor discovery is a very expensive and timely process. Replacing the traditional workflow with computational experiments, this research selected novel candidate AChE inhibitors for future drug development.
Methods
18 compounds were retrieved after virtually screening for pharmacophore structures affinitive to AChE. After compounds' interaction simulations with the AChE, their inhibition efficiencies were ranked based on Gibb's free energy of. Then the research further evaluated the compounds' oral bioavailability and ability to across the blood brain barrier by comparing their properties with the Lipinski's rule of 5.
Results
Overall, 17 out of 18 compounds passed the Lipinski's rule of 5, qualifying good absorption and permeation. While compound ZINC04713297 featured the best inhibition efficiency (-9.37kcal/mol), both ZINC92926669 and ZINC08756522 required the second lowest reaction energy (-9.20kcal/mol). ZINC92926669 stood out among the two in the absorption assessment for its stability and portability in the blood stream.
Conclusion
This research discovered 17 novel AChE inhibitors through the workflow combining virtual screening, interaction simulation and absorption assessment. It highlighted compound ZINC04713297 and ZIN92926669, which served as a starting point for development of novel acetylcholinesterase inhibitors for Alzheimer's disease.
Keywords: Alzheimer Disease, Acetylcholinesterase, Drug Discovery, Computational Molecular Biology
Introduction of dementia (Breijyeh, Z. and R. Karaman,
Alzheimer's disease (AD), a neurodegen- 2020). There are currently 55 million peo-erative disease, is the most common cause ple worldwide suffering from dementia. The
number is projected to double every 20 years, reaching 78million by 2030 (Yiannopoulou, K. G. and Papageorgiou S. G., 2020).The most popular current medicine for AD is acetylcholinesterase (AChE) inhibitor (Marucci, G., et al. 2021). By hindering the activity of AChE, this class of drug maintains a healthy level of the neurotransmitter acetylcholine at synapses (Sharma, K. 20190).
Most physicians would recommend AChE inhibitors donepezil, galantamine and rivastigmine as the first-line drug to cope with mild to moderate AD symptoms (Birks, J. 2006). However, all of them posted side effects such as nausea, vomiting and insomnia (in Liver Tox: 2012; Birks, J. S. and Harvey R. J., 2018; Hager, K., et al., 2014).
Considering the limitations of popular medicines, personalized drugs are brought to attention. Some current researches begin to design novel AChE inhibitor. Research in 2021 used virtual screening to test on 1220 galantamine derivatives (GAL-L-Ar) for anti-AChE activities (Stavrakov, G., et al., 2016). However, more drug options are still needed for doctors to match the drug with patients' physical characters. To meet the need for more candidate AChE inhibitors, this research virtually screened for compounds with high affinity to AChE as a start point for novel drug development.
Methods
Virtual screening
The virtual screening process aimed to find drugs that mimicked binding structures of AChE antibodies. These structures allowed drugs to have a high affinity to AChE.
PocketQuery
Fab410 is an AChE antibody, one of the largest peptide inhibitors targeting the peripheral site of AChE (Bourne, Y., et al., 2013). The entire antibody-protein complex has total six chains, and all antibodies constitute of a heavy chain and a light chain. This experiment chose to study two interactions: a. the interaction of AChE's chain A with antibody's light Chain E, b. the interaction between AChE's chain A with heavy Chain F (Charles A. Janeway, J., Paul Travers, Mark Walport, and Mark J. Shlomchik, 2001).
I entered the PDB code of the Fab410-BfAChE complex, 4QWW, on PocketQuery (http: / / pocketquery.csb.pitt.edu/). Clicked "search" to run the search for binding clusters with high affinity to AChE. Pocke tQuery yielded Fab410's residue clusters at the interface of the interaction. The interactive clusters were ranked by scores, based on a support vector machine (SVM) classifier (Koes, D. R. and C. J., 2012).
Pharmacophore Screening
I uploaded three highest ranking clusters each from chain E and chain F onto ZINC-Pharmer. ZINCPharmer (http://zincpharm-er.csb.pitt.edu) found matching compound hits that shared similar structure with the cluster. For each residue clusters, three compounds that had the highest RMSD (similarity scores) were chosen for further study.
Docking Experiment
SwissDock (http: //www.swissdock.ch) calculated the Gibbs free energy of interaction between compounds and AChE. On the "Submit Docking" page, I selected "targets" by uploading the structure file for 3LII. Then ligands were entered by either using their identifiers from PDB or their structure files (Grosdidier, A., Zoete, V., and Michielin, O., 2011). Entering the job name and email, clicked "submit" to run the experiment. After approximately twelve hours, results were sent via email. I ranked the compounds' inhibition efficiencies based on their interactions' Gibbs free energy. A negative value of energy of interaction indicated that the reaction was spontaneous. Therefore, the lower the value was, the more likely a compound could inhibit AChE.
Absorption Assessment
In order to qualify as applicable drugs, the compounds had to meet specific features — Lipinski's Rule of 5 (Nogara, P. A., et al., 2015). This set of rules included restrictions on number of hydrogen bond acceptor, number of hydrogen bond donor, iLOGP, and molecular weight.
Swiss ADME
To prepare the input from Swiss ADME, SMILES format of the eighteen compounds were copied from ZINC database. By pasting
them on the Swiss ADME page and clicking "run", the program generated a range of compound data. The compound couldn't have more than ten H-bond acceptors, more than five H-bond donors, a iLOGP value greater than five, or the molecular weight bigger than five hundred gram per mole. Any compounds that violated the rules were eliminated.
Results Virtual Screening
PocketQuery
Fab410 was a non-competitive AChE antibody that had promise in further research
(Bourne, Y., Renault, L. and Marchot P., 2015). After submitting the PDB code for Fab410-BfAChE complex on Pocket Query, three clusters with the highest score each were chosen from chain E and chain F. Size represented the number of amino acid residues in the cluster, and the distance embodied the longest distance in Angstroms between the centroids of any two residues in the cluster. Overall, clusters from chain E earned a better score than clusters of chain F (Table 1).
Table 1. Chain, size, distance and score of the six clusters
Name
Chain
Size
Distance
Score
Clusterl Cluster2 Cluster3 Cluster4 Cluster5 Cluster6
E E E F F F
2 3 2 l 2 2
7.3511 7.3511 9.9599 0
11.1823 5.8989
0.981743 0.981391 0.979564 0.969102 0.95847 0.956513
Finding Matching Hits
The six clusters were sent to ZINC-Pharmer. The system searched for structurally similar compounds for each cluster. The compounds were given a value of RMSD (root mean square deviation) — a
larger value indicated a more significant deviation from the original cluster structure (Koes, D. R. and Camacho, C. J., 2011). The best three compound with the lowest RMSD were selected for each cluster (Fig-urel) (Table 2).
Figure1. Structures of the compounds
Table 2. Chain, cluster and RMSD value of compounds
Chain_Cluster_Compound_RMSD
ZINC89638859 0.465
Cluster 1 ZINC92926669 0.480
ZINC89639139 0.480
ZINC88456369 0.337
Chain E Cluster 2 ZINC35334680 0.398
ZINC16942442 0.453
ZINC13377881 0.234
Cluster 3 ZINC00056472 0.247
ZINC03847223 0.456
ZINC93096123 0.268
Cluster 4 ZINC00274695 0.272
ZINC17031267 0.272
ZINC83262622 0.315
Chain F Cluster 5 ZINC04713297 0.359
ZINC04904284 0.402
ZINC08756522 0.242
Cluster 6 ZINC64438930 0.289
ZINC01633524 0.338
Docking Experiment
The docking algorithm worked by first generating 5000 to 15000 binding models. The binding energy were later evaluated, and those that had the most favorable energies were selected to the file shown as results (Grosdi-dier, A., Zoete, V., and Michielin, O., 2011).
Every compound's lowest Gibbs free energy of reaction were sorted into the table (Table 3). The lowest AG ranged from -9.37kcal/ mol of ZINC04713297 to -7.36 kcal/mol of ZINC64438930. The compounds were ranked based on the estimated AG indicating the spontaneity of its interaction with AChE.
Table 3. Estimated AG and rank of the interactions between compounds and AChE
Cluster Compound Estimated AG (kcal/mol) Rank
ZINC89638859 - 7.69 14
Cluster 1 ZINC92926669 - 9.20 2
ZINC89639139 - 8.13 7
ZINC88456369 - 7.65 16
Cluster 2 ZINC35334680 - 9.19 4
ZINC16942442 - 8.90 5
ZINC13377881 - 7.64 17
Cluster 3 ZINC00056472 - 7.72 13
ZINC03847223 - 8.68 6
ZINC93096123 - 7.75 12
Cluster 4 ZINC00274695 - 7.88 9
ZINC17031267 - 7.80 11
Cluster Compound Estimated AG (kcal/mol) Rank
ZINC83262622 - 7.89 8
Cluster 5 ZINC04713297 - 9.37 1
ZINC04904284 - 7.82 10
ZINC08756522 - 9.20 2
Cluster 6 ZINC64438930 - 7.36 18
ZINC01633524 - 7.67 15
Absorption Assessment Swiss ADME
According to the Lipinski's Rule of 5, a compound would have a poor absorption or permeation if it had more than ten H-bond acceptors, more than five H-bond donors, a iLOGP value greater than five, or the molecular weight bigger than five hundred
gram per mole (Lipinski, C.A., et al., 2001). Based on the data acquired from SwissAD-ME (Table 4), all of the compounds passed Lipinski's Rule of 5, except ZINC13377881 of cluster 3. Therefore, 17 out of 18 compounds qualified the assessment with good absorption and ability to cross blood brain barrier.
Table 4. Number of H-bond acceptors, donors, Log Po/w, and molecular weight of
compounds
Cluster
Cluster 1
Cluster 2
Cluster
3
Cluster
4
Cluster
5
Cluster
6
Compound Num. H-bond Num. H-bond Log Po/w Molecular
acceptors donors (iLOGP) weight
ZINC89638859 3 3 3.40 367.46 g/mol
ZINC92926669 4 3 3.03 353.40 g/mol
ZINC89639139 3 3 2.34 327.40 g/mol
ZINC88456369 5 2 2.87 330.38 g/mol
ZINC35334680 5 1 2.25 389.20 g/mol
ZINC16942442 7 4 1.01 298.32 g/mol
ZINC13377881 6 6 2.13 348.39 g/mol
ZINC00056472 4 4 2.44 302.36 g/mol
ZINC03847223 3 4 2.06 260.31 g/mol
ZINC93096123 4 3 3.09 303.33 g/mol
ZINC00274695 7 2 2.32 316.28 g/mol
ZINC17031267 8 2 1.91 306.21 g/mol
ZINC83262622 4 5 2.16 453.23 g/mol
ZINC04713297 4 5 2.21 406.23 g/mol
ZINC04904284 4 5 2.08 361.78 g/mol
ZINC08756522 7 3 3.03 498.13 g/mol
ZINC64438930 2 5 3.02 420.89 g/mol
ZINC01633524 4 3 1.51 332.74 g/mol
Discussion Virtual screening and docking experiment
AChE antibody was screened to locate structures highly affinitive to AChE. Then, a
searching tool discovered compounds similar to those structures. Results from ZINC-Pharmer showed that compounds mimicking cluster 4 had the lowest RMSD. Although this group of compounds were most structur-
ally similar to the original cluster, they only ranked at the 9th, 11th, and 12th in the docking experiment. ZINC04713297 (-9.37kcal/ mol) with a higher RMSD score showed the best inhibition efficiency, followed by ZINC92926669 (-9.20kcal/mol) and ZINC08756522 (-9.20kcal/mol). Therefore, duplicating the original cluster structure didn't guarantee the best inhibition. A slight deviation might lead to a better effect.
Absorption assessment
Last part of the research was to test the absorption and permeation of the candidate AChE inhibitors using SwissADME. Compounds' properties had to pass the Lipins-ki's Rule of 5 by having no more than five H-bond donors, no more than ten H-bond acceptors, Log Po/w value smaller than five, and molecular weight less than 500g/mol. Number of H-bond donor and acceptors correlate to reactivity of the compounds. If a compound had too many reacting groups, it could interact with other molecules in the blood before reaching the brain. Log Po/w and molecular weight couldn't be too high either. A high Log Po/w value would lead to poor solubility in water and fast metabolism rate. A heavy molecular weight would render a compound harder to be transported through the blood stream. While Lipins-ki's rule of 5 wasn't a hard cutoff line for absorption and permeation, it was a good standard to compare with. For the results, ZINC13377881 had six H-bond donors, making it potentially overly reactive. The rest seventeen compounds passed the Lip-inski's Rule of 5, which made them qualified with good absorption and ability to cross the blood brain barrier.
Best candidate compounds
The docking experiment highlighted ZINC04713297, ZINC92926669 and ZINC08756522 as the three best compounds at inhibiting AChE. Since the virtual screening didn't aim for a specific pharmacophore structure, there wasn't an established mechanism that contributed to the high inhibition efficiencies of all three compounds. However, possible interactions included hydrophobic interactions, hydrogen bonding, and ionic bonds. In addition, the carbon oxygen double
bonds in all three compounds resembled the acetyl group, which might facilitate the binding to the anionic site of acetylcholinesterase.
While both ZINC92926669 and ZINC08756522 had the same Gibb's free energy, ZINC08756522 weighed 498.13 g/mol, which almost violated the Lipinski's rule of 5. This left our focus on ZINC04713297 and ZINC92926669. Compared to ZINC04713297 which had the lowest Gibb's free energy of -9.37kcal/mol, ZINC92926669 featured a better absorption with lower molecular weight (353.40 g/mol) and only 3 H-bond donors. It's not reasonable to decide on the best compound without the context. ZINC92926669 had a more comprehensive performance with good absorption and inhibition. But once the compound reaches the target site, ZINC04713297 would inhibit AChE more spontaneously. Therefore, ZINC04713297 and ZINC92926669 were the two best compounds discovered from this research
Workflow evaluation
This study established the workflow combining binding site identification, virtual screening, docking experiment, and ADME assessment to identify compounds that inhibit the acetylcholinesterase. The research introduced candidate compounds as a starting point for production of effective drug in Alzheimer's Disease therapy. All the experiment were on computer, which didn't involve real chemicals. Therefore, there were no exposures to dangerous compounds. In addition, all the compound data and interactions simulations were online. So, the research required neither purchasing nor shipping of any molecules, which saved both money and time. However, due to the lack of real experiments, the actual inhibition efficiency, properties and toxicity of the candidate drugs required future studies. The next step of this research would be testing compounds through in vivo study, particularly reaction with AChE for inhibition efficiency and trial on mice for toxicity study.
Conclusion
Through this virtual screening, seventeen compounds were chosen as potential therapeutics for Alzheimer's Diseases. Overall, ZINC04713297 and ZINC92926669 from
heavy chain F of AChE antibody stood out for their outstanding inhibition efficiencies. While ZINC04713297 exhibits a more spontaneous inhibition, ZINC92926669 featured better absorption and permeation. This research recognized the necessity of testing the compounds properties in person. Before going into future drug development, candidate compounds had to be reacted with AChE for actual inhibition efficiency. It's also crucial conduct in vivo studies to research on the effect of drugs on living organism and their toxicities.
Acknowledgement
I want to thank Dr. Gabr for mentoring me throughout the research. And I really appreciate the developers of all the websites that I used, including but not limited to Pock-etQuery, ZINCPharmer, SwissDock, SwissADME. Finally, as far as I am aware, except for the reference, the research paper does not include any content that has been published. If the statement above doesn't hold truthful, I am willing to take all responsibilities.
References
Breijyeh, Z. and Karaman R. Comprehensive Review on Alzheimer's Disease: Causes and Treatment. Molecules, 2020.- 25(24).
Yiannopoulou, K.G. and Papageorgiou S. G. Current and Future Treatments in Alzheimer Disease: An Update. J Cent Nerv Syst Dis, 2020.- 12.- p. 1179573520907397.
Marucci, G., et al. Efficacy of acetylcholinesterase inhibitors in Alzheimer's disease. Neuropharmacology, 2021.- 190.- p. 108352.
Sharma, K. Cholinesterase inhibitors as Alzheimer's therapeutics (Review). Mol Med Rep, 2019.- 20(2).- P. 1479-1487.
Birks, J. Cholinesterase inhibitors for Alzheimer's disease. Cochrane Database Syst Rev, 2006.- (1).- P. Cd005593.
in Liver Tox: Clinical and Research Information on Drug-Induced Liver Injury. 2012, National Institute of Diabetes and Digestive and Kidney Diseases: Bethesda (MD).
Birks, J.S. and Harvey R. J. Donepezil for dementia due to Alzheimer's disease. Cochrane Database Syst Rev, 2018.- 6(6).- P. Cd001190.
Hager, K., et al. Effects of galantamine in a 2-year, randomized, placebo-controlled study in Alzheimer's disease. Neuropsychiatr Dis Treat, 2014.- 10.- P. 391-401.
Stavrakov, G., et al. Docking-based Design of Galantamine Derivatives with Dual-site Binding to Acetylcholinesterase. Mol Inform, 2016.- 35(6-7).- P. 278-85.
Bourne, Y., et al., Molecular characterization of monoclonal antibodies that inhibit acetylcholinesterase by targeting the peripheral site and backdoor region. PLoS One, 2013.-8(10).- P. e77226.
Charles A. Janeway, J., Paul Travers, Mark Walport, and Mark J. Shlomchik., Immunobiolo-gy: The Immune System in Health and Disease. 5 ed. 2001, New York: Garland Science.
Koes, D.R. and C. J. Camacho, Small-molecule inhibitor starting points learned from protein-protein interaction inhibitor structure. Bioinformatics, 2012.- 28(6).- P. 784-91.
Grosdidier, A., Zoete, V., and Michielin, O. SwissDock, a protein-small molecule docking web service based on EADock DSS. Nucleic Acids Res, 2011.- 39(Web Server issue).-P. W270-7.
Nogara, P.A., et al., Virtual screening of acetylcholinesterase inhibitors using the Lipinskis rule of five and ZINC databank. Biomed Res Int, 2015.- P. 870389.
Bourne, Y., Renault, L. and Marchot P. Crystal structure of snake venom acetylcholinesterase in complex with inhibitory antibody fragment Fab410 bound at the peripheral site: evidence for open and closed states of a back door channel. J Biol Chem, 2015.- 290(3).-P. 1522-35.
Koes, D.R. and Camacho, C. J. Pharmer: efficient and exact pharmacophore search. J Chem Inf Model, 2011.- 51(6).- P. 1307-14.
Lipinski, C.A., et al., Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev, 2001.— 46(1-3).- P. 3-26.
submitted 22.08.2023;
accepted for publication 20.09.2023;
published 8.10.2023
© Hasa A. Barbullushi A., Kolici E., Xhixha A., Nake, A. Contact: ren.jiawei@outlook.com
