In Silico Profiling of Phytochemicals for Modulating Amyloid Beta-Mediated Immune Regulation and Neuroinflammation in Alzheimer’s Disease

Abstract

The etiology of Alzheimer’s disease (AD) has been extensively studied for a long time, primarily associated with multifaceted mechanisms involving aggregation of amyloid beta, tau hyperphosphorylation, neuroinflammation, and immune regulation. Current therapeutics for AD are significantly targets to attenuate the cognitive decline by modulating neurotransmitters and diminishing aggregation of amyloid beta. However, these therapeutics fail to address the neuroimmune dysregulation that critically facilitates disease progressions. In this study, we have delineated the phytochemical potential to modulate the amyloid beta-triggered immune regulation through in silico approaches. Seventy three phytochemicals were selected from established anti-Alzheimer’s plants through literature mining and sequentially administrated to pharmacodynamic and pharmacokinetic investigations. The draggable study has established 14 phytochemicals, and the compounds were further curated based on their molecular interactions with the hub-targets, enriched in the amyloid beta-driven immune regulation. AD-associated proteins were retrieved from different data sets such as GeneCards, DisGeNet, GEO, and Opentargets, and the intersecting targets were curated for downstream analysis. Functional annotation and network pharmacology analysis mapped APOE4, BACE1, TREM2, IL-1β, and TNF-α as key regulatory targets. The molecular interaction analysis revealed that genkwanin and kaempferol exhibited strong binding affinity and stable interactions with the hub-targets as a potent candidates to attenuate the immune regulation in AD. Further molecular mechanics/Poisson-Boltzmannsson-Boltzmann surface area and DFT analysis have revealed their thermodynamic stability and electronic reactivity, highlighting their potential efficacy in mitigating AD progression.

Introduction

In elderly patients, AD is a progressive and irreversible neurodegenerative condition that primarily affects the entorhinal cortex and hypothalamus of the brain environment. These regions are the most prominent for memory formation and spatial navigations. The destruction of nerve cells within the microenvironment leads to the development of hallmark symptoms such as cognitive impairments, memory losses, and behavioral changes. , According to the global prevalence from the World Health Organisation (WHO) (https://www.who.int/news-room/fact-sheets/detail/dementia), the progression is increasing drastically with the rate of 10 million new cases every year and the affected individual numbers would be triple by 2050. Pathologically, AD is driven by aberrant accumulation of amyloid-beta (Aβ) within the extracellular space, the formation of neurofibrillary tangles intracellularly, and neuroinflammation in the brain. Apart from the interplay within the cellular microenvironment, an active new player, microglia, a resident innate immune cells of the brain, has further worsened the disease condition by propagating the pro-inflammatory cytokines and oxidative stress.

Neuropathologically, the transmembrane amyloid precursor protein (APP) is proteolyzed by amyloidogenic mechanism then nonamyloidogenic pathway, which generates Aβ peptides. The generated Aβ peptides subsequently aggregate within the extracellular synaptic cleft of the neurons and lead to neuronal damage. In amyloidogenic mechanism, the proteolytic cleavage of APP is enzymatically catalyzed by β- and γ-secretases, initially the β-secretase mediates the degradation of APP into soluble APP-β (sAPPβ) and a C-terminal fragment (CTF), which contains the Aβ sequence. Further, γ-secretase catabolizes the CTF to generate Aβ_40–42, which are prone to aggregate to from extracellular plaques. Microglia play a critical role in maintaining the cellular homeostasis, clearing cellular debris, and responding to pathogens and abnormal proteins. The aberrant accumulation of Aβ stimulates microglial activation, enduring the phagocytotic clearance and enzymatic degradation of plaques. Initial to the pathogenicity, the microglia supports the homeostasis by facilitating Aβ clearance, but with the sustained and persistent stimulation turned them to adopt the pro-inflammatory phenotype that fosters the neurodegeneration by generating pro-inflammatory cytokines and reactive oxygen species (ROS).

Primarily, therapeutic strategies for AD have focused on hampering Aβ clustering, prevent plaque formation, and depolymerize the Aβ clustering, which are key drivers of pathogenesis. Additionally, incorporating the novel therapeutic advances to modulate the microglia-mediated pathway directs in attenuating chronic inflammatory responses and promoting the Aβ clearance, thereby restoring the brain’s immune homeostasis, diminishing neuroinflammation, and halting the disease progression. The multifaceted pathology constrained the efficacy of antialzheimer drugs in halting the disease or slowing its progression in clinical trials. Currently, the available therapies including cholinesterase inhibitors, NMDA receptor antagonists, antiamyloid therapies, and along with nonpharmacological therapies are typically employed to symptomatic relief only. , Moreover, prolonged use of these medications are often compiled with adverse side effects, including gastrointestinal issues like vomiting, nausea, constipation, and diarrhea, neurological issues such as dizziness, headache, and confusion, and bradycardia, and amyloid-related imaging abnormalities (ARIA).

In this context, phytochemicals represent a potent and mechanistically elucidated alternative due to their intrinsic multitarget effects. Phytochemicals can concurrently modulate the adverse effects stimulated by amyloid beta aggregation, neuroinflammatory signaling and free radicals. , The intricate immunomodulatory effects of phytochemicals can regulate the Aβ clearance and mitigate the pathogenic nature of microglia. Their structural heterogeneity enables them to interact with multiple molecular targets including receptors, enzymes and signaling pathways, their free radical scavenging potential and multitarget regulator mechanisms facilitate them as a board neuroprotective agents. In this study, a comprehensive method were employed to procure the potential phytochemicals for attenuating the disease progression. It encompasses a computational approach such as pharmacodynamic and pharmacokinetic profiling, network pharmacology, molecular docking, and molecular dynamic simulation to systematically investigate their therapeutic efficacy against AD. While previous studies have explored the antiamyloid beta aggregation, antioxidant, and anti-inflammatory effects of phytochemicals. Specifically, limited studies have mechanistically addressed their role in regulating amyloid beta triggered immunomodulation and mitigating neuroinflammation and progression stimulated by microglia. By using comprehensive in silico analysis, this study investigates the molecular interaction of procured phytochemical with key regulators involved in amyloid beta triggered immunomodulation, addressing the critical limitation in AD therapy.

Results and Discussion

Phytochemicals exert neuroprotective effects in Alzheimer’s disease by regulating key pathways including neuroinflammation, oxidative stress, cholinesterase activity, amyloid-beta, and tau cascades. In this study, comprehensive literature analysis was conducted to procure the phytochemicals, already reported plants by analyzing standard research literature. Totally, 162 phytochemicals were retrieved from the literature survey and further scrutinized into 137 compounds by removing the duplicates. Among the 137 evaluated, the available 3D-SDF files were accessible for only 73 compounds. Then, the 73 molecules were involved in the downstream process for further analysis.

Drug-Likeness Analysis

Virtual screening has been emerged as a potent tool in our quest to access successful therapeutic strategy to treat AD. In the early stage of drug discovery, drug-likeness exemplifies a key biochemical properties employed to access whether a molecule contains physiochemical characteristics coherent with pharmacological activity. The physiochemical parameters is based on the Lipinski’s and Veber rules encompassing molecular weight, topological surface area, number of hydrogen bond donor and acceptor, lipophilicity, and number rotatable bonds. Standardly, these parameters were deployed to prefilter the potent drug candidate, thereby excluding molecules with a high probability of clinical failure and potentially enhancing the success rate while diminishing the economic burden for drug discovery. Table represents the physiological parameters over the procured phytochemicals, where 56 compounds were surpassed the criteria from the 73 by complying with Lipinski’s and Veber rules.

1. Drug-Likeness Properties for the Screened Phytochemicals.

S. No	compounds	MW	log P	HBA	HBD	TPSA	nRB	nHB
1	12-O-methylcarnosic acid	346.21	5.366	4	2	66.76	3	6
2	Apigenin	270.05	1.138	5	3	86.99	1	8
3	asiatic acid	488.35	7.439	5	4	97.99	2	9
4	biochanin A	284.07	1.364	5	2	75.99	2	7
5	carnosic acid	332.2	5.474	4	3	77.76	2	7
6	carnosol	330.18	5.108	4	2	66.76	1	6
7	cirsimaritin	314.08	1.699	6	2	85.22	3	8
8	genkwanin	284.07	1.459	5	2	75.99	2	7
9	kaempferol	286.05	1.486	6	4	107.22	1	10
10	magnoflorine	342.17	0.924	4	2	58.92	2	6
11	mucronulatol	302.12	1.662	5	2	68.15	3	7
12	pectolinarigenin	314.08	1.699	6	2	85.22	3	8
13	rosmadial	344.16	3.73	5	1	80.67	3	6
14	rosmaridiphenol	316.2	5.509	3	2	57.53	1	5

Pharmacokinetic Analysis

The dietary phytochemicals were computationally screened pharmacodynamic, pharmacokinetic, physiochemical, and medicinal chemistry based on certain rules before being considered as a potent drug candidate. Pharmacokinetic analysis is crucial in estimating the potential of new compounds as effective drugs and investigating their kinetics in the human body. The pharmacokinetic analysis involves the estimation of absorption, distribution, metabolism, excretion, and toxicity (ADMET) in the virtual environment.

The resulted compounds from the pharmacodynamic analysis were subjected to pharmacokinetic analysis by surpassed the drug-likeness validation, which were further sequentially evaluated by predicting their Caco2 permeability, steady state volume of distribution (VDss), AMES toxicity, Maximum tolerated Dose in human, and hepatotoxicity which are represented in Table . Initially, toxicity were accessed by estimating their hepatoxicity­(yes/no), AMES mutagenesis­(yes/no), and Maximum tolerated Dose in human­(≤0.550 log mg/kg/day). From that 53 compounds, 37 compounds are nonhepatotoxic, 33 compounds are nonmutagenic, 20 compounds are potent druggable dose. Subsequently, CNS permeability (logPS ≥ −2.500) demonstrated 14 compounds within acceptable CNS penetration. These compounds also exhibited moderate blood–brain barrier (BBB) permeability and were used for the downstream process.

2. ADMET Analysis for the Screened Phytochemicals.

S. no	compound	Caco2 permeability (log P app in 10–6 cm/s)	VDss (human) (log L/kg)	BBB permeability	CNS permeability	AMES toxicity	max. tolerated dose (human) (log mg/kg/day)	hepatotoxicity
1	12-O-methylcarnosic acid	0.604	–1.087	0.002	–1.779	No	0.221	no
2	apigenin	1.007	0.822	–0.734	–2.061	No	0.328	no
3	asiatic acid	0.479	–1.6	–0.64	–1.848	No	0.078	no
4	biochanin A	0.897	–0.341	–0.221	–2.115	No	0.4	no
5	carnosic acid	0.803	–1.027	–0.545	–1.998	No	0.345	no
6	carnosol	0.572	0.819	–0.096	–1.816	No	0.227	no
7	cirsimaritin	1.022	0.001	–0.59	–2.324	No	0.033	no
8	genkwanin	0.917	0.011	–0.23	–2.033	No	0.032	no
9	kaempferol	0.032	1.274	–0.939	–2.228	No	0.531	no
10	magnoflorine	1.575	1.211	–0.651	–2.101	No	0.198	no
11	mucronulatol	1.056	0.12	–0.281	–2.216	No	0.136	no
12	pectolinarigenin	1.128	–0.133	–0.58	–2.314	No	0.323	no
13	rosmadial	1.248	0.276	–0.275	–2.083	No	0.383	no
14	rosmaridiphenol	1.356	0.553	0.16	–1.501	No	0.219	no

Target Extraction

The target extraction was conducted by extracting the gene data sets from the online servers containing 15,715, 210, 22,185, and 4626 from the different platforms such as GeneCards, DisGeNet, GEO, and Open Targets, respectively. The genes retrieved from GEO data set are represented in Figure S1. Further, the data sets were curated using the Venny server to map the overlapping enriched regulatory proteins across the multiple sources, resulting in the curation 154 regulatory proteins were potentially associated with AD (Figure A).

1.

AD-associated gene mining and pathway enrichment analysis. (A) Venn diagram generated from Venny tool analysis depicting the intersecting genes form the retrieved databases. (B) KEGG pathway enrichment analysis exhibiting the selected genes significantly associated with AD pathology. (C) Schematic representation of key signaling pathways enriched through KEGG analysis.

Gene Ontology and KEGG Analysis

The functional annotation for the retrieved genes was conducted using the DAVID bioinformatics tool. The 154 genes were enriched in the key pathways including Alzheimer’s disease, pathways of neurodegeneration, cholesterol metabolism, neurotrophin signaling pathways, HIF-1 signaling pathway, PI3K-Akt signaling pathway, MAPK signaling pathway, etc., represented in Figure B. Further, the biological processes were employed to map the key pathological process involved in the disease progression. The retrieved genes were further subjected to ShinyGo Web server for additional confirmation and KEGG enrichment analysis. The primary genes enriched in the AD were represented as KEGG representation in Figure C.

Network Pharmacology

In this study, a network pharmacology-based approach was applied to construct and analyze the complex interaction between biological homeostasis. This approach enabled the profiling of key molecular interactions and potent regulatory proteins, which are enriched in disease mechanisms and provide a holistic understanding for novel discovery of therapeutic targets. The annotated genes mapped from the enrichment analysis was subjected to the STRING databases to generate protein–protein interactions (PPI) network, applying determined confidence interval thresholds to ensure biologically relevant associations (Figure S2 and Table S1). In the network pharmacological representation, individual proteins were denoted as nodes, while the edges correspond to the interaction between the two connected notes. The PPI profiled data from the STRING database was extracted in TSV format and uploaded to Cytoscape for further analysis (Table ). The network constituted 153 nodes, with an average node degree value of 24.5. Figure A exemplifies the PPI network for the mapped and enriched key biological process such as amyloid beta cascade (Figure S3), microglial activation (Figure S4) and pro-inflammatory mechanisms (Figure S5) from the retrieved genes. Consequent to network formation, the topological analysis was accessed using centrality matrices such as degree, closeness, betweenness, and average shortest path length (ASPL) to investigate the nodal significance within the interactome, thereby the five regulatory targets were mapped (Figure B). The mapped key regulatory proteins enriched in the above-mentioned pathways were identified as Apolipoprotein E4 (APOE4) and β-site APP cleaving enzyme 1 (BACE1) in amyloid beta cascade, Triggering receptor expressed on myeloid cells 2 (TREM2) in microglial activation, and Interleukin 1 beta (IL-1β) and Tumor Necrosis Factor-alpha (TNF-α) in pro-inflammatory mechanisms (Figure C). Moreover, the enrichment analysis was proceeded for the hub-targets to annotate the biological process which resulted in the targets were enriched in regulation of amyloid beta fibril responses such as aggregation and clearance, microglial activation and their proliferation, and regulation of neuroinflammatory and immune responses (Figure D).

3. Overall PPI Table for Enriched Genes in Alzheimer’s Disease.

network parameters	values
number of nodes	153
number of edges	1890
average number of neighbors	24.706
characteristic path length	2.034
clustering coefficient	0.537
network density	0.163
network centralization	0.469
network diameter	4
network radius	2
network heterogeneity	0.759

2.

Protein–protein interaction (PPI) analysis for the enriched genes in AD. (A) The cytoscape network exhibits interaction among the key regulatory protein enriched in AD pathogenesis such as amyloid-beta cascade, microglial activation, and pro-inflammatory mechanism. (B) Venn diagram illustrate the screened hub-targets based on annotated by passing the criteria Degree, Closeness, Betweenness, and ASPL. (C) The mapped hub-targets such as APOE, BACE1, IL-1β, TNF, and TREM2 in STRING. (D) Biological process analysis for the screened hub-targets in disease patho-mechanism in AD.

Molecular Docking

Molecular docking is a potential approach to predict the bioactive potential of derived compounds against regulatory targets, enriched in the disease microenvironment. The active sites parameters for the proteins were APOE4X:-6.7849, Y:-8.8387, Z:-5.2788; BACE1X: -14.438, Y: -43.4465, Z- 2.1641; IL-1βX:-14.1597, Y:13.9497, Z:-0.7676; TREM2X:7.3917, Y:-48.4357, Z:-19.869 and TNFαX:-5.8138, Y:-15.8407, Z:-2.1820 mapped. In this study, the procured 12 phytoconstituents were analyzed using molecular docking and their evaluated binding affinities are depicted in Table . Among the docked complexes, genkwanin and kaempferol have exhibited strong binding affinities, while the coumaric acid-4-O-glucoside, isotectorigenin, 7-methyl ether, and mucronulatol have showed the least binding affinities with the mapped regulatory proteins. Meanwhile, galantamine has depicted negligible binding affinity with the key regulatory targets. The key atomic interactions such as hydrogen bonds, and other interactions like van der Waals bond, pi-alkyl bond, and hydrophobic bonds are exhibited in Table . The docking representations are displayed in Figure .

4. Binding Energy Values for the Phytochemicals against the Hub-Targets.

S. No	compound	APOE4	BACE1	IL-1β	TNF-α	TREM2
1	12-O-methylcarnosic acid	–5.9	–7.7	4.3	–5.9	–5.8
2	apigenin	–6.5	–8.4	–5.3	–6.3	–6
3	asiatic acid	–5.8	–6.1	–5.8	–6.3	–5.5
4	biochanin A	–6.2	–7.9	–4.5	–6.4	–5.6
5	carnosic acid	–5.6	–6.3	–5.7	–6.4	–5.8
6	carnosol	–5.6	–6.2	–5.5	–6	–6.3
7	cirsimaritin	–6.1	–8.1	–4.8	–5.3	–6
8	genkwanin	–6.8	–8.1	–5.4	–7.2	–6
9	kaempferol	–6.8	–8.1	–5.5	–6.7	–6.2
10	magnoflorine	–5.4	–8	–5.2	–5.6	–5.8
11	mucronulatol	–5.7	–7.9	–4.2	–5.9	–5.6
12	pectolinarigenin	–6.3	–8.1	–4.1	–6	–5.8
13	rosmadial	–5.2	–8.2	–5.4	–5.9	–5.8
14	rosmaridiphenol	–5.5	–7.5	–5.1	–6.3	–5.9
15	galantamine	–6.4	–6.4	–5	–4.7	–4.8

5. Interacting Atoms between the Key Regulatory Proteins and Scrutinized Phytochemicals.

compound	protein	hydrogen bonds interactions atoms	other interaction atoms
genkwanin	APOE4	ARG145	LYS1, ALA5, VAL6, THR8, GLY31, TRP34, ASP35, ARG38.
	BACE1	TRP137, PHE169	ASN98, VAL130, TYR132, ILE171, TRP176, ILE179, ARG189
	IL-1ß	LEU134	TYR24, GLU25, LEU26, LYS77, PRO78, THR79, LEU80, LEU82, VAL132, PHE133.
	TNF-α	GLN47, ALA134	GLN25, LEU26, ASP45, ASN46, LYS59, SER133, GLU135.
	TREM2	SER81, THR82, HIS103	ARG52, LEU54, GLN61, VAL63, ASP104.
kaempferol	APOE4	ASP35, ARG145	ALA5, THR8, GLY31, TRP34, ARG38.
	BACE1	PHE169	ASP93, SER96, TYR132, GLN134, TRP137, LYS168, ILE179.
	IL-1ß	TYR24, LEU26, LEU134	GLU25, LYS77, PRO78, THR79, LEU80, VAL132, PHE133.
	TNF-α	ALA134	GLY24, LEU26, ASN46, ILE57, LYS59, GLU135, ILE136, PRO139.
	TREM2	ASN20, THR22, PRO37, ASP39, HIS43, ARG122	LEU121.
galantamine	APOE4	GLN249	ARG150, ILE250, GLN253, ALA257, GLN279, LYS282, VAL283.
	BACE1	TYR132, LYS168, PHE169	GLN134, TRP137, ASP289, GLN291, THR292.
	IL-1ß	PRO131	GLU25, THR79, LEU80, VAL132, PHE133.
	TNF-α		GLN25, LEU26, ASP45, ASN46, SER133, ALA134, GLU135, ILE136, PRO139.
	TREM2	ARG52, TYR108	VAL63, ASN79, GLY80, SER81, THR82, HIS103, ASP104.

3.

Molecular interaction profiles of genkwanin and kaempferol with AD-associated hub-targets. (A) Binding interaction of genkwanin with APOE4, BACE1, IL-1β, TNF, and TREM2. (B) Binding interaction of kaempferol with APOE4, BACE1, IL-1β, TNF, and TREM2. (C) Binding profiling of galantamine with APOE4, BACE1, IL-1β, TNF, and TREM2. The left side depicts the 3D representation, while the right side shows the 2D interaction map.

Molecular Dynamic Simulation

From the molecular docking studies, phytochemicals such as genkwanin and kaempferol have exhibited strong binding affinity against the regulatory targets enriched in the disease pathology. The conformational stability and several intramolecular interactions were extensively studied by a molecular dynamic simulation approach using Gromacs-2023.1 version. This approach extensively provides the quantitative evaluation of intramolecular architecture of protein–ligand complexes when the molecular assembly undergoes under simulated microenvironment. In this study, a 300 ns molecular dynamic simulation of the docked complexes were conducted to elucidate the conformational dynamics and intramolecular structural rearrangements induced by ligand binding.

Substantially, the trajectory files were extracted from the terminal frame of the 300 ns simulation to capture the equilibrated conformational state and intramolecular behavior of the docked complexes.

Over the course of the simulation, the RMSD measures the average positional displacement of atomic coordinates and offers the conformational drift and structural stability of the system. During the simulation of APOE4, for the initial equilibration phase (50 ns), it rapidly drifted from its initial geometry, amplifying form 0.5 to 1.5 nm within the first 30 ns, whereas genkwanin revealed a gradual increase and stabilized near 1.0 nm by 40 ns, and the kaempferol exhibited only a minor deviation to 0.6 nm, reaching equilibrium before 25 ns. After the midsimulation time (50–200 ns), APOE4 endured extensive structural changes, changing periodically between 1.5 and 2.5 nm with random fluctuations near 3 nm, while genkwanin maintained within a confined 0.9–1.3 nm fluctuations and the kaempferol expressed nearly stable at 0.7 nm. At the final phase (200–300 ns), APOE4 showed a slight but persistent drift toward 3 nm, demonstrating progressive expansion and a lack of stabilization. For the galantamine complex, the compound equilibrated within the initial 30–40 ns and sustained low and stable backbone fluctuations (0.7–0.9 nm) throughout the simulation. The genkwanin, galantamine, and kaempferol stabilized its conformation and were retained with limited significant deviations. The RMSD trajectory for the docked complexes have revealed strong structural stability when compared to the protein, which is represented in Figure A.

4.

Molecular dynamic simulation analysis for genkwanin and kaempferol with regulatory targets. (A) APOE4 complexes, (B) BACE1 complexes, and (C) IL-1β complexes.

RMSF analysis quantitively delineates residue-specific flexibility by measuring the mean positional deviation of individual atom or residue from its mean coordinates throughout the simulation trajectory, facilitating accurate assessment of local confirmational dynamics. The increased RMSF value indicates the residue signify with intrinsic flexibility or conformational mobility, while the decreased value denotes structurally stable and spatially constrained segments of the proteins. Figure A has represented the RMSF trajectory for the simulated complexes. In APOE4, the residues 230–280 reveals significant flexibility, with RMSF values surpassing 2 nm. This amplified fluctuations demonstrate a weak intramolecular constraints and susceptibility to huge conformational rearrangements. The binding effect of genkwanin declines these fluctuations within approximately 0.5–1.0 nm, expressing moderate intramolecular stabilization, which still permit for some conformational flexibility. In contrast, kaempferol has maintained fluctuation around 0.5 nm throughout the simulation, offering a reduction in both backbone and side-chain dynamic points to strengthen the rearrangement by ligand-induced stabilization. The galantamine docked complex have exhibited the lowest RMSF values across multiple residues, which conferring the compound induces maximum local stability followed by kaempferol.

The radius of the gyration profile provides the global molecular compactness during the course of molecular dynamic simulation. The APOE4 depicted gradual increases in R _g amplitude from 2.3 to 3.0 nm over the 260 ns, representing the tertiary structure destabilization and partial unfolding. The genkwanin equilibrates the fluctuation about 2.3 nm within the first 40 ns and substantially maintains only ±0.1 nm peaks, indicating a sustained but moderate dynamic fold. In the early simulation, kaempferol retains the compactness around 2.1 nm R _g and holds the negligible drift, signifying the restoration of the tightly packed tertiary architecture of the protein Figure A. Likewise, galantamine remains highly compact within 2.1–2.2 nm amplitudes throughout the simulation.

The SASA was evaluated to quantify the dynamic alterations in solvent exposure throughout the simulation Figure A. APOE4 upheld an average surface area of 210–220 nm² approximately, consistent with more surface area exposure. Genkwanin reduced the exposure to within 200 nm², exemplifying moderate solvent shielding. The kaempferol revealed decreased fluctuations around 170–180 nm², resulting in minimal solvent exposure and enhanced structural rearrangement. The galantamine has exhibited 190–200 nm² with minimal fluctuations and reduced solvent exposure area.

The hydrogen bond analysis revealed the interaction profiles between the ligand and APOE4. During the simulation, kaempferol preserved up to 6–7 interactions and persistently holds 2–4 bonds but the genkwanin exhibited limited hydrogen bonds, typically ranging from 0 to 2 interactions. The galantamine-APOE4 complex maintained 2–3 hydrogen bonds with minimal interruptions, suggesting a stable binding with minimal interactions count than kaempferol.

Figure B exhibits the molecular dynamic simulation trajectory files for BACE1 and their phytochemical docked complexes. The RMSD analysis have revealed that the BACE1 structure rapidly amplified from its initial conformation, raising from 0.5 to 0.8 nm within the initial equilibration phase, continues with minor amplitudes fluctuating between 0.6 and 0.8 nm. Genkwanin equilibrated the conformational rearrangement within 20 ns of RMSD fluctuations around 0.25–0.35 nm, while the kaempferol established an intermediate stabilizing behavior with 0.35–0.55 amplitudes after 30 ns. The galantamine exhibited minimal RMSD fluctuations within 0.28–0.38 nm, which equilibrates the major conformation drift. The protein depicted substantial relaxation in its structural conformation, whereas ligand-induced rigidification were addressed, especially genkwanin pronounced the strongest stabilization when compared to kaempferol binding induced changes. The RMSF profile for the BACE1 depicted that amplitudes have maintained at 0.4 nm fluctuations among the interval of 60–223 residues. The elevated amplitude were exhibited over 0.8 nm with 235–246 residues that corresponds to the exposure of solvents and the active site underwent conformational changes. The flexible fluctuations were observed throughout the simulation indicate conformational changes of the flap and C-terminal loop, which modulate exposure to the catalytic site. Both genkwanin and kaempferol significantly attenuate backbone amplitudes within the catalytically relevant regions, thereby imparting enhanced conformational rigidity. Likewise, galantamine have effectively amplified the local flexibility. From that genkwanin expressed tight regulation of local stabilization and modulate the exposure, while kaempferol existed minor terminal flexibility, providing more dynamic interactions.

The overall gyration was maintained between 2.2 and 2.3 nm, while the protein and kaempferol were expressed the similar R _g. Meanwhile genkwanin and galantamine have shown the lowest and stable profile of R _g around 2.15–2.2 nm with minor fluctuations and stabilized with more compact conformation, preserving the folded core with minimal structural fluctuations. Overall structural equilibration for all the complexes have expressed within a range of 190–215 nm² during the simulation. In relation with the protein and kaempferol bounded complex, the genkwanin-bounded complex has possessed minor fluctuations over 5–10 nm², indicating the slight reduction in the SASA. Simultaneously, with regard to the gyration analysis, the galantamine reduced the solvent exposure comparable to the genkwanin complex. This reduction for the genkwanin complex aligns with the gyration, demonstrating enhanced compactness and slight burial of hydrophobic residues upon the ligand. The hydrogen profile for the kaempferol have maintained 2–4 with minor fluctuations around 6–7 during the simulation. In contrast, genkwanin have exhibited 0–2 interactions with the protein, which provides subtle stability without extensive bonding. Although, the galantamine produced fewer hydrogen bond interactions over the protein.

The molecular dynamic simulation profiles were exhibited in Figure C for the docked complexes and IL-1β. The RMSD graphs have exhibited 0.2–0.40 nm amplitudes for IL-1β and its docked complexes. While the genkwanin-bound complex indicated slight deviation noticed during the course of simulation with 0.45 nm amplitudes, which indicated intermittent local conformation rearrangements. Followed by genkwanin, galantamine slightly lower fluctuations ranged within 0.20–0.27 nm. Conversely, kaempferol equilibrated the system with 0.2–0.25 nm fluctuations, resulting in stabilizing the protein with highly stable conformations. The RMSF values for the protein and its docked complexes were remained with 0.6 nm fluctuations, while major fluctuations have been observed around the residues such as 25–32, 40–60 and 90–100 for the genkwanin-bound complex. Meanwhile, the kaempferol complex closely followed the native protein with minor fluctuations observed within 110–115 amino acid residues, resulting in kaempferol having exhibited better stabilization, whereas genkwanin increased the flexibility and partially altered the conformational stability. The galantamine bound complex displayed marginal fluctuations relative to the apo protein throughout the simulation.

The R _g analysis has observed that all the complex remain within a stable and narrow range of 1.45–1.60 nm fluctuations, providing significantly maintained without large-scale unfolding during the simulation. With respect to the reduction in RMSD and RMSF trajectories, kaempferol has equilibrated the complex to stabilize its compactness and structural packing consistently over the reduction within the R _g of 1.48 nm. However, genkwanin-docked complexes have exhibited within 1.52–1.55 nm. Galantamine has resulted in consistently maintained structural compactness similar to kaempferol with minor fluctuations (1.49–1.52 nm). The SASA were steady between 85 and 105 nm² for the protein and docked complexes. From that kaempferol expressed lowest peaks between 85 and 95 nm² with minor fluctuations, which pronounced enhanced compactness by reducing its exposure to the solvent and also the R _g-derived compactness of the complex. On the other hand, genkwanin exhibited slight higher values (95–105 nm²) when compared to protein ranging 95–100 nm², with an upward drift after 150 ns. The galantamine revealed mildly higher SASA values with moderate stabilization. The hydrogen bond interaction analysis has demonstrated that kaempferol retained its interaction ranging 1–4 bonds with the IL-1β, but genkwanin maintained 2 interaction and frequently dropped to zero during the simulation and galantamine formed few interactions within a stable mood. The trajectories have exhibited lowest RMSD, compact R _g, and reduced SASA for the kaempferol-bound protein, resulting in kaempferol-stabilized IL-1β followed by galantamine and genkwanin reducing its structural integrity with weaker anchoring with the protein.

The trajectory files for the TNF-α and the docked phytochemical complexes are represented in Figure A. The backbone RMSD values for the TNF-α and its docked complexes have remained within 0.18–0.42 nm across the simulation. The proteins have maintained values near 0.28 nm, whereas genkwanin complex exhibited slight higher fluctuation with 0.30–0.35 nm with minimal fluctuations. However, the kaempferol complexes have observed with 0.25–0.30 nm and settling with smallest deviations. In parallel to the phytochemicals, galantamine expressed significant smooth fluctuations in RMSD analysis with 0.23–0.35 nm. Upon the ligand binding, kaempferol enhanced the protein stability and conserved its conformational changes, but genkwanin offered only negligible stabilization relative to the normal state. The RMSF analyses have exhibited that TNF-α underwent major fluctuations in the central loop of the protein (80–115 amino acid residues) with 0.587 nm. For the docked complexes, kaempferol allowed higher flexibility when compared to genkwanin with ±0.15 nm shifts. For the galantamine-bound complex, it produced higher fluctuation (0.7 nm) over the central loop of the protein, which induced the region-specific conformation mobility, instead of uniform rigidification of the protein. During the simulation, minor shifts were also observed at the N- and C- terminal regions, with genkwanin exerting stronger stabilization.

5.

Molecular dynamic simulation profile for genkwanin and kaempferol with regulatory targets. (A) TNF-α, (B) TREM2, and (C) comparative binding free energy (ΔG, kcal/mol) of genkwanin (red), kaempferol (green), and galantamine (blue).

The R _g attributed to the protein spiked around 1.68–1.75 nm during the simulation. Upon the ligand binding, genkwanin expressed slightly elevated spikes in R _g (1.72–1.75 nm), which resulted in marginal expansion and conformation reorganization around the binding pockets. Conversely, kaempferol binding reduced the R _g with 1.67–1.70 nm, resulting in enhanced compactness and their stability. However, galantamine sustained with minimal and stable R _g values, mimicking the global compactness packing. For the SASA analysis, genkwanin expressed increased exposure to the solvent area relative to protein. Upon the binding, genkwanin induced the conformation adjustment in the central loop of the protein. However, kaempferol exhibited tight packing with the protein and decreased the solvent exposure. Galantamine slightly extended the fluctuations in SASA analysis with minor drift during 240–265 ns and balanced throughout the simulations. The hydrogen bond analysis revealed that kaempferol binding generated 2–5 hydrogen interactions, but genkwanin and galantamine build only 1–3 interactions. The combined analyses exhibited that kaempferol enhanced its compactness and diminished the solvent exposure but failed to suppress the structural flexibility in the central loop. While the genkwanin binding was more effective in regulating TNF-α dynamics.

Figure B illustrates the molecular dynamic analysis for the TREM2 protein and the docked ligand complexes. The RMSD plot for all the complexes were equilibrated within 50 ns and constantly stabilized over the 300 ns trajectory. The TREM2 protein have expressed within 0.30–0.35 nm, genkwanin maintained slight higher amplitude over 0.35–0.45 nm but kaempferol exhibited lower RMSD (0.26–0.32 nm) compared to both the protein and genkwanin. Galantamine have displayed moderate fluctuations ranging from 0.2 to 0.40 nm. The kaempferol have exhibited to stabilize the native protein into most favorable conformational changes with reduced RMSD amplitudes among genkwanin and galantamine. The RMSF illustrated the compactness for the docked complexes and the proteins, which revealed that the entire complexes remained within the 0.2–0.7 nm throughout the simulation. For the genkwanin complex, major fluctuations were noticed within the 60–80 amino acid residues, which raised due to the destabilization of flexible loops within the binding site. However, the kaempferol and galantamine bound complexes elevated up to 0.3 nm amplitudes and maintained over loops with the protein.

Further, the R _g representation addressed the key differences in global compactness of the proteins and their docked complexes. For TREM2, the peaks resided within a range of 1.55–1.60 nm, the genkwanin induced elevated R _g with fluctuations maintained at 1.50–1.65 nm, but the kaempferol initiated the subtle global stabilization and enhanced the compactness during the simulation with the reduced R _g into 1.45–1.55 nm. For the galantamine complex, the amplitudes maintained intermediately within 1.48–1.55 nm range and exhibited conversed global compactness. The SASA for the protein remained intermediate along with the genkwanin and kaempferol bounded complex, which depicted around 80–90 nm². Due to the genkwanin-induced structural changes within the central loop of the protein, genkwanin expressed higher solvent exposure of 90–105 nm². The kaempferol reduced the solvent exposure and maintained values within 75–85 nm², which represented the stabilization of the protein. Galantamine minimized the solvent exposure by maintaining in the 80–90 nm² range. The hydrogen bond interactions analysis outlined that kaempferol maintained 1–3 interactions persistently and occasionally 4–5 interactions during the binding. Meanwhile, genkwanin and galantamine maintained minimal hydrogen interactions with the proteins and sometimes with 2. The results exhibited that both the compounds were potent to regulate the multitarget medication to attenuate the disease progressions further.

The Molecular Mechanics/Poisson-Boltzmann Surface Area

The MM-PBSA analysis was employed to calculate the net free energy difference between the bound and unbound states of the docked protein–ligand complexes throughout the 300 ns simulation. Figure C shows the net binding energies of genkwanin and kaempferol over the regulatory targets. Among the targets, genkwanin revealed the most potent affinity with APOE4 (−18.08 kcal/mol) and TREM2 (−16.43 kcal/mol), followed by BACE1, IL-1β and TNF-α. Whereas kaempferol exhibited comparable binding with BACE1 (−13.79 kcal/mol) and APOE4 (−13.91 kcal/mol), followed by TREM2, IL-1β, and TNF-α. The galantamine system revealed potent binding energy with APOE4 (−14.70 kcal/mol) and BACE1 (−12.50 kcal/mol), followed by TREM2, IL-1β, and TNF-α. The integrated analysis revealed that both the compounds articulated the strongest binding affinity with the regulatory proteins in AD, resulting as a more effective multitarget regulator in the disease progression.

DFT Analysis

Further, DFT analysis was performed to investigate the quantum chemical reactivity and electronic structure for the curated phytochemicals. The Frontier molecular orbital (FMO) was carried out using the optimized structures such as HOMO (highest occupied molecular orbital) and LUMO (lowest unoccupied molecular orbital), which play a crucial role in determining the electronic behavior, charge transfer characteristics, and overall chemical reactivity of the molecule. The molecules contain minimal HOMO–LUMO energy gaps which are electronically softer, possessing significant reactivity, lower stability, and greater polarizability, due to their spontaneous electron density redistribution. However, larger gap molecules exhibit increased kinetic stability and decreased susceptibility to charge-transfer interactions. The screened compounds genkwanin and kaempferol were subjected to DFT analysis, the optimized structures are depicted in Figure A,B.

6.

Comparative quantum chemical and electrostatic analysis of genkwanin and kaempferol. (A,B) Optimized 3D molecular structure of genkwanin and kaempferol, (C,D) FMO distribution exhibiting HOMO and LUMO energy levels, along with calculated energy gaps, (E,F) MEP surface maps illustrating charge distribution over the molecular frame, (G,H) MAC distribution depicting the atomic charge density of the molecules.

For the compound kaempferol, EHOMO and ELUMO are shown as −5.870 eV and −1.887, respectively, and its energy gap (ΔE) is 3.983 eV. Genkwanin revealed −2.488 eV and −1.322 eV as EHOMO and ELUMO with ΔE of 1.166 eV (Figure C,D). In contrast, kaempferol exhibited larger gap, which enables greater hardness confer higher kinetic stability and selectivity, reducing nonspecific interactions. Mechanistically, FMO analysis exhibited that genkwanin is predicted to possess better electronic-driven potency in receptor site that favor donor or highly polarizable ligand when compared with kaempferol. However, the compounds were carried with certain limitations that may hinder their druggability, specifically high reactive and stability characteristics. The compounds with narrow HOMO–LUMO gaps exhibit enhanced capacity for charge redistribution and intramolecular charge transfer within the environment, which readily increased the metabolic susceptibility and off-target effects. In this context, nanoencapsulation methods provide a promising strategy to mitigate these limitations by enhancing their pharmacokinetic characteristics, improving stability, and reducing the off-target effects, thereby actively reinforcing the translational effects of the compounds. The predicted chemical potential (μ), chemical hardness (η), electrophilicity index (ω), electron affinity (A), electronegativity (χ), ionization potential (I), and global softness (S) are represented in Table .

6. Quantum Chemical Properties of the Screened Phytochemicals.

compound	chemical potential (μ)	chemical hardness (η)	electrophilicity index (ω)	electron affinity (A)	electronegativity (χ)	ionization potential (I)	global softness (S)
genkwanin	–1.905	0.583	3.113	1.322	1.905	2.488	0.857
kaempferol	–3.879	1.992	3.776	1.887	3.879	5.87	0.251

The molecular electrostatic potential (MEP) analysis was further carried out to describe the spatial distribution of electron density within the molecule, resulted to predict the regions susceptible to electrophilic or nucleophilic attack. Critically, this analysis estimates the reactive sites and provides the comprehensive visualization of intramolecular interactions in molecular recognitions. The electron cloud mapped the electron density of the molecule, representing the spatial charge distribution where the red regions (negative potential) indicate electron-rich sites corresponding to electrophilic attack or hydrogen bond acceptor regions, whereas blue regions (positive potential) denote electron-deficient areas susceptible to nucleophilic attack or hydrogen bond acceptor regions. The MEP for the screened phytochemicals is depicted in Figure E,F.

The Mulliken Atomic Charge (MAC) of the ligand exhibited amplified charge distribution ranging from −2.168 to 1.890 and −0.987 to 1.160 for genkwanin and kaempferol, respectively (Figure G,H). Genkwanin expressed wider charge ranges and was capable for stronger and more electrostatic interaction oriented with complementary charged residues such as Lys/Arg and Glu/Asp within the binding pocket. Kaempferol depicted constrained charge polarization likely lead to weak electrostatic complementarity, significantly limiting the specificity and interaction strength.

The screened phytochemicals, such as genkwanin and kaempferol, would enhance the neuroprotective effects by preserving neural integrity and synaptic stability. Further, these bioactive compounds would modulate the immune dysfunction by amplifying the microglial dynamics throughout the progressions. Figure delineates the phytochemicals role in the AD disease environment and their mechanistic insights into amyloid beta aggregation and microglial dynamics.

7.

Schematic representation of phytochemicals role in neuroprotective mechanism. (A) Interplay of pathogenic microglia and neuron in the diseased brain environment. (B) Role of phytochemicals in attenuating the effects such as ROS and inflammatory cytokines in progression. (C) Phytochemical interactions with amyloid-beta aggregates and activated microglia to diminish the progression.

Both compounds such as genkwanin and kaempferol are small nonglycosylated flavonoid extensively studied with antioxidant, anti-inflammatory, and enzyme-modulating properties. In this study, these compounds outperformed with the most favorable target modulating effects and were screened among the pharmacokinetic and pharmacodynamic analysis. Even though many flavonoids have limited BBB permeability, for the compound kaempferol, multiple preclinical studies demonstrated that kaempferol exhibits potent neuroprotective agent and can modulate neuroinflammation and BBB integrity. In an kunming mice model, Lie et al. (2012) demonstrated that kaempferol can attenuate the d-gal-induced cognitive impairment by upregulating the ERK1/2-CREB pathway expression, stimulating the Na⁺,K⁺-ATPase activities, and scavenging the free radicals in hippocampus. Genkwanin is a O-methylated flavone with multiple therapeutic benefits such as anti-inflammatory, antioxidant, anticancer, and neuroprotective properties. In an in vitro analysis, Li et al. (2021) elucidated that genkwanin mitigates the neuroinflammation and neurotoxicity by downregulating the TLR4/MyD88/NLRP3 inflammasome pathway in the MMP⁺-induced PD (Parkinson’s disease) cellular model. Despite the favorable mechanistic profiles of genkwanin and kaempferol against the key regulatory targets enriched in AD, their translational potential is constrained by moderate BBB permeability. To overcome this limitation, encapsulation or loading to a carrier-based strategy would enhance BBB permeability and delivery to the site of the disease microenvironment. In the future, we are currently working on encapsulation of genkwanin and kaempferol into a carrier-based delivery system to overcome the limitation. This approach would counteract and enhance the BBB permeability and pharmacokinetic properties that restrict their therapeutic efficacy. Nanovesicle encapsulation is postulated to improve compound stability, increase specific delivery mechanisms, and promote efficient cellular uptake by endogenous vesicular transport mechanisms. In a study, Ongtanasup et al. (2024) successfully developed a niosomal-based drug delivery system by encapsulating the Zingiber officinale extract within a palm wax-based gel, resulting in improved drug loading efficacy, sustained release, and significantly enhanced therapeutic performance when compared to the unencapsulated extract. In an in vivo study, Guo et al. (2021) effectively encapsulated the quercetin within a GAP43 (monoclonal antibody)-conjugated exosome to actively target the GAP43-expressed damaged neurons to improve the neuronal survival by suppressing the ROS production through upregulating the Nrf2/HO-1 mechanism.

Conclusion

Alzheimer’s disease has multifaceted pathophysiological mechanisms comprising aggregation of misfolded proteins, mitochondrial dysfunction, neuroinflammation, and oxidative stress following the activation of innate cells in the brain microenvironment. In conclusion, this study was incorporated the in silico approaches to systematically evaluate the phytochemicals and their therapeutic regulation in AD. Employing the pharmacokinetic and pharmacodynamic evaluations, the procured compounds were curated into 14 phytochemicals based on their druggable principles. The curated compounds were further evaluated against the key regulatory proteins such as APOE4, BACE1, IL-1β, TNF-α, and TREM2, enriched in key patho-mechanisms like amyloid beta cascade, microglial activation, and inflammatory responses in disease progression. Applying a structure-based drug design strategy, integrated with molecular docking, molecular dynamics, and free energy calculations established that genkwanin and kaempferol exhibited significant insights over their binding, stability, and affinities with respect to the regulatory targets. Further, the quantum chemical analysis revealed the electronic properties for indicating the reactivity and stability of the phytochemicals. The resulting analysis over the procured phytochemicals genkwanin and kaempferol established significantly druggable activity to modulate the immune regulation mediated by amyloid beta plaques.

Materials and Methods

Ligand Preparation

Bioactive active compounds with potential antialzheimeric activity were extracted through a comprehensive literature review of recently published scientific articles. The selection criteria was relied on established phytochemical profiling data for the plants which are extensively studied for antialzheimeric properties. The compiled phytochemicals were then further curated by merging and removing duplicates for in silico analysis and further validation.

Pharmacodynamic Analysis

The pharmacodynamic profile for the selected compounds were evaluated using DruLito (Drug-Likeness Tool), an open cheminformatics software that evaluate molecular structure against multiple rule-based filters to predict as orally active drug candidates. The compounds SDF (Structural Data Files) were extracted from the NCBI-PubChem database, and the retrieved SDF was subjected to analysis based on the criteria including Lipinski’s rule of five and Veber’s rule. The core physiochemical parameters including molecular weight, hydrogen bond acceptor/donor, number of rotatable bonds, TPSA (Topological Surface Area), and calculated partition coefficient (c Log P) were quantitatively analyzed using this tool.

Pharmacokinetic Analysis

Pharmacokinetic properties of the extracted compounds were predicted using the online tool pKCSM (https://biosig.lab.uq.edu.au/pkcsm/), employed graph-based signatures to predict a wide range of ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) parameters. The SMILES (simplified molecular input line entry system) representation inputs were computed to the pKCSM Web server to predict key pharmacokinetics characteristics such as Caco2 permeability, VDss, BBB permeability, CNS permeability AMES mutagenesis, maximum tolerated dose human, and hepatotoxicity.

Target Extraction

The Alzheimer’s associated targeted proteins were obtained from the GeneCards (https://www.genecards.org/) (accessed on 28.09.2024), DisGeNet (https://disgenet.com/) (accessed on 12.09.2024), open targets (https://www.opentargets.org/) (accessed on 28.09.2024), and the NCBI Gene Expression Omnibus (GEO) (GSE5281) (https://www.ncbi.nlm.nih.gov/geo/) (accessed on 24.09.2024) by using the key word Alzheimer’s disease. The retrieved genes from GEO were further analyzed to extract differentially expressed genes via GEO2R. The retrieved genes were filtered by using criteria including p < 0.005 and log 2-fold change | < 2. The retrieved DEGs and the genes from the database were subjected to Venny 2.1 mapping tool (https://bioinfogp.cnb.csic.es/tools/venny/) to extract the overlapping genes from the each data sets. The intersecting genes were extracted for further analysis.

Gene Ontology and KEGG Analysis

The gene ontology (GO) and KEGG (Kyoto Encyclopaedia of Genes and Genomes) pathway enrichment analysis were elucidated to screen the potentially enriched genes in Alzheimer’s disease and pathways significantly mapped with the retrieved data sets. The DAVID (Database for Annotation, Visualization, and Integrated Discovery) ver. 6.8 (https://davidbioinformatics.nih.gov/) is an online-based bioinformatic platform that encompasses multiple annotation databases to map the enriched genes via Gene ontology terms like cellular components, molecular functions, and biological processes. The curated gene lists of official gene symbols was utilized as an input, and the background species were limited as Homo sapiens to ensure species-specific annotation. The overlapped genes were utilized to profile the functional enrichment analysis. DAVID was implemented to screen statistically significant clustered genetic profiles to map biological processes and signaling pathways. Further, genes enriched in Alzheimer’s disease were subsequently visualized and mapped to secondary enrichment analysis using the ShinyGo ver. 0.77 (https://bioinformatics.sdstate.edu/go77/).

Network Pharmacology

The enriched genes were imported into the Search Tool for the Retrieval of Integrating Genes and Proteins (STRING) ver. 12.0 (https://string-db.org/). A protein–protein interaction network was fabricated with a species limited to H. sapiens and confidence threshold of ≥0.4. Finally, the detached nodes were eliminated, and the network was retrieved into the TSV format and uploaded to Cytoscape ver. 3.10.1. The hub-targets were scrutinized by evaluating the key topological parameters like degree of centrality, closeness, betweenness, and ASPL. Nodes with high centrality scores were considered topologically significant and highlighted as regulatory targets. Further, the screened targets that surpassed the criteria were evaluated using a molecular docking approach.

Molecular Docking

The objective of molecular docking is to evaluate the binding affinities and interactions of the screened compounds with the hub-targets. In this study, molecular docking was carried out using an open-source virtual screening software PyRx (ver. 0.8), which integrates AutoDock Vina for effective ligand–receptor interacting simulation. Initially the 3D structures of screened ligands and regulatory targeted proteins were procured form the PubChem database and RCSB protein data bank (https://www.rcsb.org/), respectively. The imported PDBQT structures were uploaded to Discovery Studio to eliminate the nonessential compounds such as water molecules, bounded ligands, and other heteroatom entities to establish a clean docking environment. Before docking, the energy minimization for the ligand geometrics was carried out by using the mmff94 (Merk Molecular Force Field) algorithm to generate the most stable low-energy conformers. Subsequently, the energy-minimized ligands were modified into a PDBQT file to abide with AutoDock Vina criteria.

The AutoDock Vina engine was implemented to simulate the docking, which predicts the binding conformations and computes the binding affinities (kcal/mol). The stronger binding affinity was predicted by computing the ligand–receptor complexes which satisfy the lowest binding affinity (more negative). , In this study, a reference standard drug (galantamine, an FDA approved drug) was simultaneously implemented alongside the tested compounds in the molecular docking and further downstream process to elucidate the interaction profiles and targets specificity. Further, the docked complexes were elucidated using Discovery Studio Visualizer 2017 to investigate the binding orientations, binding patterns, and the integration of key amino acid residues.

Molecular Dynamic Simulation

The molecular dynamic (MD) simulations were performed on protein–ligand complexes using GROMACS software [package 2023.1 version] to elucidate the structural and dynamic behavior of the docked complexes. The initial phase of preparing protein–ligand complexes for MD simulations involved a molecular docking analysis. Typically, molecular docking was conducted under energy-minimized and static conditions to anticipate the optimal binding orientation of the ligand within the active site of the regulatory proteins. However, this approach captures only the fixed depiction of the molecular binding conformation within the active site. Consequently, simulations integrate the Newton’s classical equation of motion to compute the temporal evolution of atomistic insights into the structural stability, conformational dynamics, and intramolecular interactions within a fully solvated system that closely replicates physiological conditions over time.

The force-fields for the each ligands were created using the Charmm-GUI online server-based graphical interface and the CHARMM36m force field. The complexes were positioned in a rectangular simulation box, maintaining a buffer spacing of 10 units in all directions along the Cartesian axes. The Monte Carlo-based ion platform method was incorporated to solvate the systems by immersing in a TIP3P water environment, and Na⁺ or Cl^– ions were introduced to neutralize the system. , Each simulation was subjected to energy minimization using the steepest descent algorithm to eliminate steric hindrance and optimize the system geometry. Subsequently, the system equilibrated for 1 ns with gradual increase in temperature from 100 to 300 K. The equilibrated systems were used to simulate the molecular dynamics for 300 ns under NPT ensemble conditions, maintaining the constant number of particles (N), pressure (P), and temperature (T). The velocity-rescale thermostat algorithm was used to regulate the system temperature at 310.15 K, whereas Berendsen barostat algorithm was implemented to control the pressure at 1 atm. The Particle-Mesh-Ewald (PME) algorithm was employed to compute long-range electrostatic interactions under periodic boundary condition for the precisely solvated system, and a 1 nm cutoff distance was limited to compute the short-range interactions. , The P-LINCS algorithm was employed to improve the computational efficiency by constraining covalent bond geometry through a paralyzed linear constraint solver, allowing the integration of Newton’s equations of motion with a 2 fs time step. After the completion of simulations, solvent water molecules were stripped from the trajectory files, and global translation and rotational motions were eliminated to ensure precise investigation of internal structural dynamics such as root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), radius of gyration (R _g), solvent-accessible surface area (SASA), and hydrogen bond analysis.

The Molecular Mechanics/Poisson-Boltzmann Surface Area

The MM-PBSA were incorporated to calculate the interaction free energies (ΔG _bind) of the docked complexes using the gmx_MMPBSA approach. Computationally, MM-PBSA offers a reasonably accurate estimation of binding free energy when compared to systemic approaches such as free-energy perturbation or thermodynamic integration by decomposing the free energy into molecular mechanics and solvation contribution. This approach incorporates three major energetic components including: (1) molecular mechanics (MM) relies on classical force-field energies (angle, torsion, bond, electrostatics, van der Waals), (2) Poison-Boltzmann (PB) evaluates the electrostatic solvation energy, and (3) surface area (SA) investigates the nonpolar solvation contribution. The binding free energy (ΔG _bind) was estimated using the following relationships

$${\mathrm{\Delta }G}_{\mathrm{b}\mathrm{i}\mathrm{n}\mathrm{d}}=G_{\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\mathrm{x}}-(G_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}}+G_{\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{d}})$$

where G _complextotal free energy of the docked complex, G _protein and G _ligandtotal free energies of the isolated protein and ligand in the solvent, respectively. Additionally, the free energy for each entity was expressed from

$$G_x=⟨E_{\mathrm{M}\mathrm{M}}⟩+⟨G_{\mathrm{s}\mathrm{o}\mathrm{l}}⟩-T\mathrm{\Delta }S$$

where xligand/protein/protein–ligand complex, Ttemperature, Sentropy, TΔSentropic contribution to the free energy under vacuum, and ⟨E _MM⟩molecular mechanics under vacuum, this term encompasses bonded and nonbonded interactions, where ⟨E _MM⟩ is derived from the MM force field parameters as represented

$$⟨E_{\mathrm{M}\mathrm{M}}⟩=⟨E_{\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{d}}⟩+⟨E_{\mathrm{n}\mathrm{o}\mathrm{n}‐\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{d}}⟩=⟨E_{\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{d}}⟩+(E_{\mathrm{v}\mathrm{d}\mathrm{W}}+E_{\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}})$$

where E _non‑bondedbonded interactions including angle, bond, dihedral, and improper interactions, E _non‑bondednonbonded interactions consisting both (E _vdW)van der Waals and (E _elec)electrostatic interactions and are simulated using Lennard-Jones (LJ) and Coulombic potential functions, respectively. The free energy of solvation ⟨G _solv⟩ includes G _non‑polar and G _polar and can be predicted using

$$⟨G_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{v}}⟩=G_{\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}}+G_{\mathrm{n}\mathrm{o}\mathrm{n}‐\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}}=G_{\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r}}+(\gamma \times \mathrm{S}\mathrm{A}\mathrm{S}\mathrm{A}+b)$$

where γcoefficient related to surface tension, SASAsolvent-accessible surface area, bfitting parameter, and TΔSentropic contribution to free energy. The MD simulation trajectories for the 300 ns simulation were employed to calculate the binding energy for the docked complexes. The solute and solvent dielectric constants were 2 and 80, respectively, γ was 0.0227, and the Poison-Boltzmann (PB) was derived by using the linear PBsolver.

Density Functional Theory Analysis

Following the molecular docking and molecular dynamic simulations, quantum chemical calculations were performed to map the electronic and reactive properties of the screened ligands. All quantum calculations were computed by using DFT as predicted in Gaussian16 software. The molecular geometry of the ligands was completely calibrated using the Becke’s three-parameter hybrid exchange functional (B3LYP) combined with the 6-31­(d, p). Geometry optimizations were conducted without symmetry constraints, and the resulting geometries were verified to be local minima by imaging all vibrational frequencies.

The FMO analysis was performed to investigate the HOMO and LUMO energies for the phytochemicals. The energy gap (ΔE) between the nucleophilic and nucleophilic effects was estimated to access the reactivity, stability, and electron transfer efficiency of the ligand. To map the nucleophilic and electrophilic regions, the spatial orientation of the HOMO and LUMO orbitals were visualized using GaussView software. Further, the chemical potentials (μ), chemical hardness (η), chemical softness (σ), and electrophilicity index (ω) were predicted using Koopman’s theorem.

$$\mathrm{\Delta }E=E_{\mathrm{L}\mathrm{U}\mathrm{M}\mathrm{O}}-E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}};\eta =\frac{\mathrm{\Delta }E}{2};S=\frac{1}{\eta }$$

$$\mu =\frac{[L+H]}{2};\chi =-\frac{[L+H]}{2};\omega =\frac{{\mu }^2}{2\eta }$$

The MEP technique was implemented to characterize the spatial distribution of electron density within a molecule and compute its electrostatic interaction potential with external positive or negative charge distributions. It derives the molecular insights into electrophilic and nucleophilic reactive regions, which determine the molecular recognition process and noncovalent interactions.

MAC analysis was computed to evaluate partial atomic charges and the charge polarization within the ligand. , The predicted atomic charges were utilized to interpret the electrostatic contributions and hydrogen-bonding capabilities of individual atoms, complementing the binding interactions observed in docking and MD simulations.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c11503.

• Volcano plot generated from GEO data set analysis to visualize significantly differentially expressed genes; STRING-based protein–protein interaction networks to illustrate the overall interaction landscape, as well as pathway-specific networks associated with the amyloid-β cascade, microglial activation, and inflammatory mechanisms; and comprehensive table summarizing STRING network characteristics, interaction scores, and hub proteins to complement the network visualization (PDF)

Bharathi Kalidassconceptualization, performing, analyzing, writingoriginal draft and editing; Gothandam Kodiveri Muthukaliannanproject supervision, review, and editing.

The authors declare no competing financial interest.