Repurposing of USFDA-approved drugs to identify leads for inhibition of acetylcholinesterase enzyme: a plausible utility as an anti-Alzheimer agent

Abstract

In the quest to identify new anti-Alzheimer agents, we employed drug repositioning or drug repositioning techniques on approved USFDA small molecules. Herein, we report the structure-based virtual screening (SBVS) of 1880 USFDA-approved drugs. The in silico-based identification was followed by calculating Prime MMGB-SA binding energy and molecular dynamics simulation studies. The cumulative analysis led to identifying domperidone as an identified hit. Domperidone was further corroborated in vitro using anticholinesterase-based assessment, keeping donepezil as a positive control. The analysis revealed that the identified lead (domperidone) could induce an inhibitory effect on AChE in a dose-dependent manner with an IC₅₀ of 3.67 μM as compared to donepezil, which exhibited an IC₅₀ of 1.37 μM. However, as domperidone is known to have poor BBB permeability, we rationally proposed new analogues utilizing the principles of bioisosterism. The bioisostere-clubbed analogues were found to have better BBB permeability, affinity, and stability within the catalytic domain of AChE via molecular docking and dynamics studies. The proposed bioisosteres may be synthesized in the future. They may plausibly be explored for their implication in the developmental progress of new anti-Alzheimer agent achieved via repurposing techniques in future.

In the quest to identify new anti-Alzheimer agents, we employed drug repositioning or drug repositioning techniques on approved USFDA small molecules.

1. Introduction

Acetylcholinesterase (AChE) (EC: 3.1.1.7) is a vital enzyme that plays a decisive role in the nervous system, allowing the breakdown of the neurotransmitter acetylcholine (ACh) into choline and acetate.^1,2 This break is indispensable for terminating the biological action of acetylcholine at the post-synapses.³ A reduction in the chemical level of acetylcholine characterizes Alzheimer's disease (AD).⁴ This chemical messenger, acetylcholine, is a vital factor correlated with cognitive decline and memory impairment.⁵ As per the reports published by Alzheimer's Association, AD is thought to affect 6.9 million Americans 65 years of age and older by 2024. Out of numerous biological druggable targets to treat AD, inhibiting AchE via small molecules to increase and replenish the levels of acetylcholine in the brain, which is further associated with improving cognitive function, is an essential therapeutic regimen. In the past year, a few small drug candidates, including donepezil, rivastigmine, and galantamine (Fig. 1), have been approved to manage the symptoms of AD via temporarily replenishing acetylcholine levels by the inhibition of AchE and acting as AChEIs. While the mentioned drugs have the potential to provide symptomatic relief, there are no reports associated with their involvement in the underlying progression of AD.^6–8 Besides this, one monoclonal antibody, aducanumab, was also approved in June 2021 that is known to modify the pathophysiology of Alzheimer's disease (AD) by targeting the specific Aβ protein. Being the first disease-modifying antibody, it has been the most alluring medicine for treating Alzheimer's disease.⁹ Its approval was also met with some controversy. Specifically, the EMA rejected this medication due to the contradictory findings of two similarly designed trials (ENGAGE and EMERGE) and the absence of proof that a decrease in brain Aβ correlates with improvement in clinical outcomes.¹⁰

Fig. 1. Chemical structure of key approved acetylcholinesterase inhibitors.

However, the currently approved regimen of approved AChEIs is also associated with numerous problems and side effects. The important ones include gastrointestinal issues (nausea and vomiting) along with the effect on the heart that further precipitates bradycardia and dizziness.^11,12 Owing to this, there is a considerable need to identify new chemical entities or repurpose or reposition the existing drugs for their use as AChEIs. Drug repurposing or repositioning techniques allow the exploration of chemical space from the enormous reservoir of FDA-approved drugs, thus identifying an existing drug for its new therapeutic applications.^13,14 In the case of AChEIs, this approach holds promise for identifying potent approved drugs to combat AD and related conditions where AChE is the culprit. This drug repurposing approach is essential since the approved drugs have already undergone clinical trial phases and possess data associated with their pharmacokinetic, pharmacodynamic and safety profiles. In addition, the accessibility of known dosage forms, including formulations and their routes of administration, allows a favorable transition of already established drugs into clinical trials for repurposed targets, thus ensuring faster market entry and acceptance by medical healthcare professionals, including patients.¹⁵

In essence, drug repositioning exemplifies a planned and well-organized way for the identification of potent AChEIs, allowing the effective management of AD and other associated neurological disorders.^16–18 One primary drug repurposing method involves structure-based virtual screening (SBVS).¹⁹ SBVS is a computational way that allows rational assessment of the interactions between small molecules serving as the ligand and the targeted proteins. The screening of compound libraries allows the identification of potential lead candidates with high binding affinity towards the biological receptor. Therefore, the obtained leads are biologically corroborated to deduce their efficacy, which further paves the way to their entry into the clinical arena via cost-effective and time-efficient means.^20,21

Herein, we report the SBVS of 1880 FDA-approved drugs via the Glide module of the Schrodinger software. This was followed by calculating Prime MMGB-SA binding energy and molecular dynamics simulation studies. The cumulative analysis led to the identification of domperidone, as an identified hit. Domperidone was further corroborated in vitro using anticholinesterase-based assessment, keeping donepezil as a positive control. The analysis revealed the anticholinesterase potential of domperidone. However, as domperidone is known to have poor BBB permeability, we rationally proposed new analogues utilizing the principles of bioisosterism. The bioisostere-clubbed analogues were found to have better BBB permeability and may be utilized in developing new anti-Alzheimer agents achieved via the repurposing technique.

2. Materials & methods

2.1. Structure-based virtual screening

Maestro 13.3 of Schrodinger software is used to perform in silico studies. Structure-based drug designing is an extensively employed process in drug discovery. In the present work, the human acetylcholinesterase (hAChE) protein was downloaded from the RCSB protein data bank (https://www.rcsb.org/) having co-crystalized donepezil (PDB: 7E3H).²² For the protein preparation, the “protein preparation workflow” panel was used to optimize the hydrogen (H-bond), fill the missing chain, and minimize and delete the water molecules from the protein. In the next step, the grid generation was done by the “receptor grid generation” module. The FDA-approved ligands are downloaded from the ZINC database (https://zinc.docking.org/) or the FDA (https://www.fda.gov/).²³ The selected FDA-approved ligands (1880) were subjected to the “LigPrep” panel of the Glide module. The ligand preparation module generates the 3D structures of compounds with ionization around pH 7.0 ± 2.0 with OPLS2005 as the force field and generates the tautomers. In the last step, the ligand docking was executed using the ligand docking panel of the Glide module of Maestro Schrodinger software (https://www.schrodinger.com/).

2.2. Binding free energy calculations-MMGBSA

The top selected molecules from XP docking studies were subjected to MMGBSA calculations using Schrodinger software's Prime module. The solvation model VSGB 2.0 and force field, as well as OPLS2005, were used to perform the energy minimisation for receptor–ligand complexes. The results of MMGBSA were shown in the form of ΔG-bind, ΔG_H-bond, ΔG_lipo, ΔG_coul, and ΔG_vdW.²⁴

2.3. Molecular dynamics (MD) simulations

To mimic the biological system and make both ligands and proteins flexible, we performed molecular dynamics (MD) for 100 ns. The selected compound from docking and MM-GBSA studies (domperidone) was subjected to MD. MD utilises a system model panel, which is used as a predefined SPC (single point charge) solvent model for MD studies with orthorhombic boundary conditions that also employs a force field OPLS2005. Next, we did the energy minimization by using the energy minimization module, where the energy minimization of the protein–ligand (PL) complex is done. After achieving a stable ligand–protein complex, we performed MD for 100 ns using an NPT (isothermal–isobaric ensemble with constant temperature, pressure, and the number of particles) ensemble with atmospheric pressure (1.01 bar) and 310 K temperature while other parameters remain as default.²⁵

2.4. In vitro inhibition

Molecular Probes Inc. Invitrogen's Amplex red acetylcholine/acetylcholinesterase assay kit (A12217) was used to measure the activity of acetylcholinesterase inhibition.^26,27 To summarize, in 96-well plates with flat bottoms (Tarsons), 100 ml of Tris-HCl buffer (0.05 M, pH 8.0) containing reference inhibitors and test drugs at 4 different concentrations were incubated for 15 minutes at 37 °C. To begin the reaction, 100 ml of a 400 mM Amplex Red reagent working solution containing 2 U mL⁻¹ horseradish peroxidase, 0.2 U mL⁻¹ choline oxidase, and 100 mM acetylcholine was added. At 37 °C, using a multi-detection microplate fluorescence reader (Synergy HI, Bio-Tek® Instruments) that measured fluorescence at 545 nm for excitation and 590 nm for emission, we were able to quantify the formation of resorufin and H₂O₂ after 30 minutes of dark incubation with Amplex red dye. Concurrently, the vehicle was used in place of the test drugs to conduct a positive control experiment. 20 mM H₂O₂ was employed as the second positive control. For the purpose of the negative control, 1× reaction buffer devoid of acetylcholinesterase was utilized. After subtracting the background activity, which was found from wells with all components except for AChE, which was replaced with a 1× buffer solution, the specific final fluorescence emission was calculated. We ran each experiment three times to ensure accuracy.

2.5. Scopolamine induced amnesia model

2.5.1. Animals

The study used adult male Swiss Albino mice weighing between 20 and 25 grammes. The specimens were housed in polyacrylic enclosures measuring 22.5 × 37.5 cm, maintained at ambient temperature (24–27 °C) and subjected to a 12 hour photoperiod. Food was refrained from being consumed for one hour prior to the behavioural study. Both the methodology and the necessary number of animals for the study were authorized by the Institutional Animal Ethical Committee under Protocol No. Dean/13-14/CAEC/342. The behavioural study was performed on Y maze (Make-Orchid Scientific India) and chemicals like scopolamine hydrobromide and donepezil for that were purchased from Sigma-Aldrich.

2.5.2. Experimental design

2.5.2.1. Administration of scopolamine

On the seventh day, one hour after the test or donepezil administration, scopolamine hydrobromide (3 mg kg⁻¹) was dissolved in distilled water and given intraperitoneally. The behavioural experiment was conducted 5 minutes following the administration of scopolamine.²⁸

2.5.2.2. Experimental protocol and drug administration

Animals were partitioned into seven cohorts, each consisting of six animals. Prior to administration, the test compounds domperidone and donepezil were freshly dissolved in distilled water. The mice were separated into seven experimental groups, each consisting of six animals: (i) vehicle (1 ml); the experimental groups included: (ii) scopolamine at a dosage of 3 mg kg⁻¹, (iii) scopolamine combined with donepezil at a dosage of 1 mg kg⁻¹, (iv) scopolamine combined with domperidone at 0.5 mg kg⁻¹, (v) scopolamine combined with domperidone at 1 mg kg⁻¹, (vi) scopolamine combined with domperidone at 2 mg kg⁻¹; and (vii) control conditions. Drug administration was conducted by intraperitoneal injection (i.p.) for all groups. In separate groups, donepezil and domperidone were given once daily for a duration of seven days. Except for the vehicle and control groups, all the animal groups were given scopolamine on the seventh day to cause amnesia.

2.6. Evaluation of memory function

2.6.1. Y-maze test

The examination was conducted to assess the participants' immediate working memory. The study's experimental design and dosages are already specified in the experimental protocol and drug administration corresponding sections. The effects of domperidone were assessed at doses of 0.5, 1, and 2 mg kg⁻¹. The test was conducted on the final day of the treatment, more precisely on the seventh day. Following dosing, a 15 minute training session was conducted where animals were positioned in a Y-maze with the novel arm left closed. Following four hours of training sessions, the main study was conducted five minutes after administering scopolamine hydrobromide intraperitoneally. During this session, the animal was positioned at the midpoint of the arm in order to investigate all three arms. The experiment lasted for 15 minutes and any time the mice entered each arm was captured on camera. Repeated arm entry was interpreted as an indication of memory impairment. Sequential arm selections (ABC, BCA, CAB, excluding BAB) and new arm entry were regarded as factors contributing to memory enhancement. The memory improvement score was determined by application of the following formula: % alternation = (number of alternations/(total arm entries) − 2) × 100.

The brief outline of the methodology employed is illustrated in Fig. 2.

Fig. 2. The methodology used in the present work.

2.7. Hit optimization

Hit optimisation is one of the methods in the initial drug discovery development process. Various methods are used for lead optimisation, such as structure activity–relationship and computational methods. However, we used the bioisosteric replacement method in the present study to improve the pharmacokinetic profile. In Schrodinger, the Maestro software includes the bioisosteric module.

3. Results and discussion

3.1. Structure-based virtual screening

We begin the expedition by selecting the hAChE from the protein data bank (PDB) with PDB ID: 7E3H. The selected protein 7E3H active site with co-crystallized donepezil is shown in Fig. 3. The protein selected for study has a resolution of 2.45 Å. After protein preparation, the grid generation was done. Initially, molecular docking was validated by redocking the co-crystallized ligand (donepezil) onto the primary site where the co-crystallized ligand was bound. The analysis revealed by RMSD (root mean square deviation) is indicated in Fig. 4.

Fig. 3. Protein belonging to AChE PDB ID: 7E3H (A); representation of the binding site, highlighting essential residual amino acids in the active domain (B).

Fig. 4. A. RMSD calculations of the co-crystallized ligand (green colour) and the redocked pose (purple colour), B. 3D pose of co-crystallized ligand, C. 2D interaction of co-crystallized ligand.

The standard precision (SP) mode of the Glide module was used for the molecular docking study. Docking scores are calculated based on ligand binding affinity (conformation) with the protein (receptor and macromolecule). Based on the docking score, the top 100 ligands were selected, and the docking score ranged from −16.81 to −15.91 kcal mol⁻¹. Further, these molecules were subjected to extra precision (XP) docking, and the top 10 molecules were selected based on vital interaction and docking score in the range of 17.30 to 15.99 kcal mol⁻¹. The ligands' docking score with pivotal interaction is presented in Table 1.

Selection of top ligands as deduced via SBVS and MMGBSA analysis (scores in kcal mol⁻¹).

S. no	Compound name	Dock (glide) score	MM-GBSA SCORE
1	Protokylol	−17.30	−71.65
2	Nebivolol	−17.30	−82.17
3	Domperidone	−16.62	−101.33
4	Omadacycline	−16.43	−88.68
5	Dobutamine	−16.28	−65.34
6	Ipratropium	−16.22	−65.89
7	Trazodone	−16.08	−80.13
8	Oxitropium	−15.64	−87.61
9	Haloperidol	−15.35	−73.57
10	Reproterol	−15.99	−70.11
11	Co-crystalized ligand	−17.56	−78.36

The top two compounds, protokylol (−17.308 kcal mol⁻¹) and nebivolol (−17.301 kcal mol⁻¹), showed docking scores comparable to co-crystallized donepezil (17.567 kcal mol⁻¹). The co-crystallized ligand donepezil (Fig. 5A) also has essential interaction with amino acid residues TRP86, PHE295, and TRP286. Similarly, compared to co-crystallized ligands, protokylol (Fig. 5B) interacts with amino acid residues TYR341, TYR337, TRP86 and PHE295, while nebivolol (Fig. 5C) was found to interact with TYR341, TYR86, TRP286 and Tyr124. The next hit, domperidone (Fig. 5D), also interacts with amino acid residues TRP86, TYR133, TYR337, TYR286, SER293 and PHE295. Next, oxitropium showed interactions with amino acid residues TRP86, TYR337, PHE295, ARG296, and TYR124 (Fig. 5E).

Fig. 5. A) Co-crystalized ligand. (B) Protokylol. (C) Nebivolol. (D) Domperidone. (E) Oxitropium.

3.2. MM-GBSA calculations

Next, docking scores provide a rapid and effective way to predict binding affinities. However, this solely relies on the geometric and energetic considerations of the receptor and ligand. Considering this, MMGBSA (molecular mechanics generalized born surface area) was used to calculate the free energy for binding. This computational method offers a more laborious and physically accurate approach that accounts for the solvation effects and entropic contributions in addition to the protein and ligands, allowing the free binding energy to be calculated. Considering the utility, we performed the MMGBSA calculation of the obtained top ligand via SBVS. The analysis revealed that domperidone was found to have the highest binding free energy −101.33 kcal mol⁻¹. However, both docking and MMGBSA methods have similarities in that molecular docking allows a preferred binding pose by providing a snapshot of the most energetically favourable interaction.

In contrast, MMGBSA is associated with the calculation of the binding free energy between an associated ligand and proteins, considering molecular mechanics interactions and the solvation effects. Nevertheless, molecular docking and MMGBSA do not provide the scope to study the explicit time-dependent simulations between the targeted proteins and ligands. The molecular dynamics technique is usually employed to address this, identify the real-time simulations, and decipher the stability. Thus, based on the docking core and MMGBSA, we selected domperidone for MD simulation studies because it possesses all the essential interactions.

3.3. Molecular dynamics study

From molecular docking and MMGBSA studies, the domperidone complex was selected for MD simulation for 100 ns. The protein's RMSD analysis showed that it went through a slight fluctuation. It remained stable thereafter throughout the 100 ns simulation time within the 1.2–4.0 Å range. From the ligand-RMSD graph (Fig. 6A), it was observed that there are fluctuations around the initial 10 ns. After that, the ligand remains stable in the active site of the protein's cavity. The protein root means square fluctuation (RMSF) plot indicated (Fig. 6B) that fluctuations were observed around amino acid (AA) residues in the range of 90–105 (2.4 Å) and 250–300 residues (3.2 Å). The protein–ligand (PL) contact (Fig. 6C) histogram showed that ASP 74 (98%), PHE 295 (91%), and SER 293 (46%) formed an H-bonding interaction with the ligand. The imidazole ring of the ligand formed π–π stacking with the TRP 86 (35%) residue and π–cation interaction with TYR 341 (55%). All these amino acids aid the stability of the complex over 100 ns, as shown in Fig. 6D.

Fig. 6. Molecular dynamics studies. (A) RMSD graph of domperidone with protein. (B) RMSF plot of protein. (C) Histogram interaction of protein residues with domperidone. (D) 2D interaction diagram of domperidone (E). Timeline depiction of amino acid residues with time.

3.4. In vitro analysis: corroborating the in silico findings

The in vitro analysis of the lead hit domperidone was done on AChE to understand the corroboration of the in silico findings. The assay was done in the presence of donepezil, which was used as the positive control and reported to exhibit potent AChE inhibition. The analysis conducted at three varying concentrations (0.2 μM, 2 μM, and 20 μM) revealed that the identified lead (domperidone) was able to induce an inhibitory effect on AChE in a dose-dependent manner (Fig. 7) with an IC₅₀ of 3.67 μM as compared to donepezil which exhibited an IC₅₀ of 1.37 μM. Though donepezil was found to be better at low doses and got saturated at higher doses, domperidone exhibited a dose-dependent inhibition of acetylcholinesterase at the chosen dose.

Fig. 7. Dose response curve of donepezil and domperidone portraying AChE inhibition at three varying concentrations (0.2 μM, 2 μM, and 20 μM).

Assessment of memory function using Y maze

Consecutive arm choices and novel arm entries estimated memory function. The improvement in the score was correlated with better memory function. Donepezil, the marketed AChE inhibitor, was reported to significantly improve the percentage alternation score at the dose of 3 mg kg⁻¹. At 1 mg kg⁻¹ dose, domperidone significantly improved memory function compared to scopolamine. However, the improvement was significantly less than that of the control group.

Further, 2 mg kg⁻¹ was found to be the most potent dose. No significant difference in the alternation score was observed between the control and donepezil. The same set of experiments was used to assess the novel arm entries of the animals (Fig. 8A). An increased number of novel arm entries is a sign of better memory function. The novel arm entries were found to be significantly improved at the doses of 1 and 2 mg kg⁻¹ as compared with scopolamine (Fig. 8B).

Fig. 8. Assessment of memory function using the Y maze test: (A) score of the percentage alternation and (B) novel arm entries.

However, domperidone exhibited weaker AChE inhibition than donepezil. The literature reveals the combination of domperidone with AChE inhibitors such as rivastigmine. The combination has been explored to reduce the gastrointestinal side effects in patients with AD.²⁹ However, the issue with domperidone is that it has partial access to the CNS as it cannot cross the BBB and might not show potent AChE inhibition in an actual biological scenario. To compensate for this, we have thought to rationalise the lead optimization technique to score better ligands with better efficacy and the ability to cross the BBB efficiently.

3.5. Hit optimization

The hit optimization of domperidone as a potential treatment for Alzheimer's disease offers several significant benefits. Firstly, it can improve the drug's ability to penetrate the blood–brain barrier (BBB), ensuring that a higher concentration of the active compound reaches the brain. This is crucial for enhancing its therapeutic effect on acetylcholinesterase (AChE), which is a key target in Alzheimer's treatment. Secondly, hit optimization can increase the potency and selectivity of domperidone against AChE. Higher potency means that lower doses are needed, which can reduce side effects and improve patient compliance. Thus, bioisosteric replacement was done to improve the pharmacokinetic profile of the domperidone hit. A total of 89 bioisosteres were generated among the top five bioisosteres with high docking scores in the range of −17.78 to −16.11 kcal mol⁻¹, and then reference −17.56 kcal mol⁻¹ was selected (Table 2). Bioisosters_1 (Fig. 9) interacts with amino acid residues TYR341, TYR337, TRP286, TRP86 and PHE295 with a docking score of 17.78 kcal mol⁻¹ as shown in Fig. 10A. Similarly, bioisosters_2 interacts with amino acid residues TYR341, TYR337, TRP286 and PHE295 with a docking score of 17.18 kcal mol⁻¹ (Fig. 10B).

Selection of top 5 ligands as deduced via docking score and MMGBSA analysis.

Sr. no.	Compound name	Dock score	MM-GBSA SCORE
1.	Bioisosters_1	−17.78	−98.51
2.	Bioisosters_2	−17.18	−101.3
3.	Bioisosters_3	−16.90	−91.15
4.	Bioisosters_4	−16.24	−94.27
5.	Bioisosters_5	−16.11	−94.94
6.	Co-crystalized ligand	−17.56	−78.36

Fig. 9. Structure of proposed top-scoring bioisosteres for plausible synthesis and exploring anti-AChE inhibition.

Fig. 10. (A) Representation of 3D and 2D diagrams of bioisosters_1. (B) Representation of 3D and 2D diagrams of bioisosters_2.

Further, the top five compounds were subjected to MM-GBSA calculation to understand the free binding energy calculation. All the bioisosteres have ΔG binding energy higher than that of co-crystallized donepezil as shown in Table 2. Bioisosters_2 (Fig. 9) showed the highest binding energy of −101.3 kcal mol⁻¹.

Next, the top five selected compounds were subjected to ADME studies to check whether there was an improvement in the pharmacokinetic properties of the bioisosteres. All the bioisosteres follow the rule of five, and QPlogPo/w and QPlogS were within the acceptable range (Table 3). The main properties of QPlogBB are improved, ensuring that the active compound crosses the BBB and reaches the brain. Therefore, we can deduce from the docking score as well as MM-GBSA and ADME studies that the designed bioisosteres have improved binding affinity and pharmacokinetic properties.

ADME properties of bioisosteres.

Compound id	Mol Wt.	QPlogPo/w	QPlogS	QPlog HERG	QPlogBB	% Human oral absorption	Ro5
Bioisosters_1	427.988	4.455	−5.269	−6.314	0.215	100	0
Bioisosters_2	428.976	3.639	−4.532	−5.92	−0.057	90.699	0
Bioisosters_3	429.963	2.665	−3.254	−5.322	−0.176	81.507	0
Bioisosters_4	429.963	2.72	−3.75	−5.711	−0.316	80.654	0
Bioisosters_5	429.963	2.884	−3.721	−5.712	−0.157	84.166	0
Domperidone	425.917	3.899	−5.379	−6.884	−0.642	85.493	0
Co-crystalized ligand	379.498	4.409	−4.572	−6.592	0.115	100	4

After analyzing the pose of donepezil and bioisosters_1 in the active site, it was found that both occupy the similar biding pocket and have similar binding poses (Fig. 11A) in the active site. By analyzing the poses of co-crystallized donepezil and bioisosters_1 and their pharmacophoric domains, we found that they consist of a hydrogen bond acceptor and donor linker, and both terminals contain a ring system (Fig. 11B). Thus, we can conclude that the bioisosteres efficiently possessed an essential feature required for AChE inhibition.

Fig. 11. (A) Overlaying representation of donepezil and bioisosters_1 at the active site. (B) Pharmacophore feature alignment for donepezil and bioisosters_1.

Further, to understand the stability of bioisosters_1 and protein, MD simulation for 100 ns was carried out. The protein's RMSD analysis showed a slight fluctuation around 20 ns. After that, it remained stable throughout the simulation within the 1.50–4.0 Å range. The ligand-RMSD graph (Fig. 12A) shows that it remains stable in the active site of the protein's cavity. The protein root means square fluctuation (RMSF) plot indicated (Fig. 12B) that fluctuations were observed around amino acid (AA) residues in the range of 250–300 residues (4.0 Å). The protein–ligand (PL) contact (Fig. 12C) histogram showed that ASP 74 (99%), PHE 295 (92%), and ARG 296 (61%) formed H-bonding interaction, and TRP86, TYR337, and HIS447 formed π–π stacking interactions with the ligand. All these amino acids aid the stability of the complex over 100 ns, as shown in Fig. 12D.

Fig. 12. Molecular dynamics studies. (A) RMSD graph of bioisosters_1 with protein, (B) RMSF plot of protein. (C) Histogram interaction of protein residues with bioisosters_1. (D) 2D interaction diagram of bioisosters_1. (E) Timeline depiction of amino acid residues with time.

4. Density functional theory analysis of the hit domperidone and most potent bioisostere (bioisosters_1 or BOM)

Density functional theory (DFT) plays a crucial role in advancing the study of optical materials by offering detailed insights into their electronic and optical properties at both atomic and molecular scales. DFT allows for accurate predictions of a material's electronic structure, such as its band structure and band gap, which are key to understanding its intrinsic properties.¹ Additionally, DFT can be applied to predict non-linear optical behaviours, providing valuable information on how materials respond to intense light. The frontier molecular orbitals, namely the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), are vital in molecular interactions, where the HOMO acts as an electron donor and the LUMO as an electron acceptor.² The energy levels of these orbitals give insights into a molecule's stability and reactivity. Koopmans' theorem provides an effective approximation of these orbital energies by using the molecule's ionization potential and electron affinity.

Global reactivity descriptors are theoretical concepts in quantum chemistry and molecular modelling that offer a comprehensive view of a molecule's reactivity. These descriptors provide insights into the overall reactivity of a system, rather than focusing on specific local regions, making them particularly valuable for understanding and predicting chemical behaviour in the context of electronic structure calculations. Commonly studied global reactivity descriptors include electronegativity (χ), chemical potential (μ), chemical hardness (η), and softness (s).³ To better understand the comparative stability of BOM and DOM, the energies of their frontier molecular orbitals were examined, and various global reactivity descriptors were calculated (as shown in Table 4). DFT-based studies were conducted in different environments, including the gas phase, aqueous medium, and organic phase (DMSO). A higher chemical hardness indicates a larger energy gap between the HOMO and LUMO, which is associated with a more stable system. Among the compounds studied, DOM demonstrated greater stability, with a chemical hardness of 2.754 compared to 2.335 for BOM. In terms of medium effects, BOM exhibited the highest chemical hardness in the gas phase, followed by DMSO and then water. Similarly, DOM showed the highest chemical hardness in DMSO, followed by the gas phase and then water. The enhanced stabilization of DOM compared to BOM in both organic and aqueous environments was attributed to the presence of hydrogen bonding sites in DOM.

Outcome of the molecular descriptor analysis, BOM represents bioisosters_1 and domperidone is represented by DOM.

Cd	Vehicle	HOMO	LUMO	Energy gap	Chemical potential (μ)	Chemical hardness (η) = (I − A)/2	Chemical softness	Electrophilicity index
BOM	DMSO	−6.131	−1.462	4.669	−3.796	2.335	0.214	3.087
	Water	−6.116	−1.477	4.639	−3.796	2.319	0.216	3.107
	Gas phase	−6.199	−1.496	4.703	−3.847	2.352	0.213	3.147
DOM	DMSO	−6.055	−0.546	5.508	−3.300	2.754	0.182	1.977
	Water	−6.055	−0.935	5.120	−3.495	2.560	0.195	2.385
	Gas phase	−6.021	−0.682	5.340	−3.351	2.670	0.187	2.104

The HOMO and LUMO are primarily associated with the electron-donating and accepting ability of a molecule. In the context of ligand–protein interactions, the localization of these molecular orbitals on specific functional groups of the ligand can significantly impact how the ligand donates electrons to the protein's active site. When the HOMO is localized on atoms or groups that are in close proximity to the protein's electron-accepting sites, it facilitates stronger interactions, such as hydrogen bonding, pi-stacking, or electrostatic interactions. Similarly, the localisation of the LUMO on functional groups that align well with these nucleophilic sites can result in more effective electron transfer processes, which are critical for the formation of stable covalent or non-covalent bonds between the ligand and the protein. This enhanced interaction can lead to increased binding affinity, making the ligand more effective at stabilizing the protein–ligand complex. In order to gain a detailed idea of the localisation, the frontier molecular orbitals were mapped over the surface of both the molecules. It has been revealed that the HOMO and LUMO of DOM are well separated and the molecule behaves in a donor–acceptor manner. As far as the HOMO–LUMO of BOM is concerned, the HOMO of the molecule was mainly localised over the piperidine unit and the LUMO was dispersed over the indanone unit. The frontier molecular orbitals of BOM and DOM are represented in Fig. 13.

Fig. 13. Frontier molecular orbitals of BOM and DOM.

Fukui function

Local reactivity descriptors such as the Fukui function also help to determine the selectivity and reactivity of a molecule. It represents the tendency of electron density to distort at a given position upon accepting or donating electrons. It helps in assessment of reactive sites liable to electrophilic, nucleophilic or radical attack. The atomic or condensed Fukui function on the ith atom site for nucleophilic (f⁺_i), electrophilic (f⁻_i) and free radical (f⁰_i) attack can be represented as:f⁺_i = qⁱ (N + 1) − qⁱ (N)f⁻_i = qⁱ (N) − qⁱ (N − 1)f⁰_i = (qⁱ (N + 1) − qⁱ(N − 1))/2where qⁱ is the atomic charge at the ith atomic site in the cation (N − 1), anion (N + 1) or neutral (N) molecule. The natural charge was determined by NBO analysis for evaluation of the charge present on each atom. Morrel et al. offered a new dual descriptor f(r) that distinguishes the site for electrophilic and nucleophilic attack existing inside a molecule and is well-defined by the equation:Δf(r) = f⁺_i − f⁻_iThe atomic site with Δf(r) > 0 is susceptible to electrophilic attack, whereas a site prone to nucleophilic attack must have Δf(r) < 0. The Fukui functions evaluated for BOM and DOM are summarized in Table 5 listed below, and Fig. 14 displays optimised structures of BOM and DOM.

Fukui functions for BOM and DOM.

BOM		DOM
Atom	Δf	Atom	Δf
C1	0.18	N1	0.61
C2	0.06	C2	0.16
N3	0.53	C3	0.39
C4	0.06	C4	0.08
C5	0.19	C5	0.08
C6	0.10	C6	−0.14
C7	0.16	C7	−0.13
C8	0.04	N8	0.55
C9	0.01	C9	0.01
C10	0.02	C10	0.40
C11	0.02	C11	0.17
C12	0.11	C12	0.27
C13	−0.01	C13	0.08
Cl14	0.04	Cl14	0.04
C15	0.26	N15	0.62
C16	−0.44	C16	−0.84
O17	0.22	O17	0.71
C18	0.06	C18	0.16
C19	0.19	C19	0.40
C20	0.18	C20	0.15
C21	0.15	N21	0.54
C22	0.07	C22	−0.12
C23	0.11	C23	0.12
C24	0.21	C24	0.23
S25	−0.20	C25	0.27
C26	0.17	C26	0.11
C27	0.25	C27	−0.11
C28	−0.44	N28	0.64
O29	0.24	C29	−0.84
		O30	0.72

Fig. 14. Optimised structures of A. BOM and B. DOM.

5. Conclusion

AChE is a vital enzyme that plays a decisive role in various neurodegenerative diseases such as Alzheimer's disease, etc. In the quest to identify new anti-Alzheimer agents, we employed drug repositioning or drug repositioning techniques on approved USFDA small molecules. The SBVS of 1880 USFDA-approved drugs led to identification of domperidone as an identified hit. Domperidone was further corroborated in vitro using anticholinesterase-based assessment, keeping donepezil as the positive control. The analysis revealed that the identified lead (domperidone) induces an inhibitory effect on AChE in a dose-dependent manner with an IC₅₀ of 3.67 μM as compared to donepezil, which exhibited an IC₅₀ of 1.37 μM. However, as domperidone is known to have poor BBB permeability, we rationally proposed new analogues utilizing the principles of bioisosterism. The lead optimization was carried out primarily, a) to improve the domperidone containing analogue's ability to penetrate the blood–brain barrier (BBB); b) to improve the potency and selectivity of domperidone analogues against AChE; c) to optimize the pharmacokinetic and pharmacodynamic properties and reduce the off-target effects of domperidone and its designed analogue. The bioisostere-clubbed analogues were found to have better BBB permeability, affinity, and stability within the catalytic domain of AChE as deduced from in silico based studies. The proposed bioisosteres may be synthesized in the future and plausibly be explored for their implication in the development of new anti-Alzheimer agents.

Animal use

All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of NMIMS University, Shirpur campus, Maharashtra, India and approved by the Animal Ethics Committee of School of Pharmacy & Technology Management Shirpur with project proposal no. SPTM/09/2023/IAEC/24.

Data availability

This study was carried out using publicly available data from Drug Approvals and Databases. 2024. Available from: https://www.fda.gov/drugs/development-approval-process-drugs/drug-approvals-and-databases.

Author contributions

Ashish Ranjan Dwivedi, Devendra Kumar: conceptualization, supervision methodology, data curation, Nidhi Nainwal, Pramod Rawat, Kapil Kumar Goel, Shivani Jaiswal: writing – original draft preparation. Savita Yadav, Prachi Fartyal: visualization, Sandhya Chahal: molecular dynamics study, Rahul Singh, Shriya Mahajan, Vibhu Jha: in vivo study, Ashish Ranjan Dwivedi, Devendra Kumar: writing – reviewing and editing.

Conflicts of interest

There are no conflicts to declare.