Integrated UHPLC‐Q‐Orbitrap HRMS and Network Pharmacology Approach for Exploring Potential Active Ingredients and Pharmacological Mechanisms of Epimedii Folium Against Alzheimer's Disease

ABSTRACT

Epimedii Folium has been widely used to treat Alzheimer's disease (ad) in China. However, the potential active components of Epimedii Folium and its mechanism against ad are still not clear. In this study, an ultra‐high performance liquid chromatography coupled to quadrupole Orbitrap high‐resolution mass spectrometry (UHPLC‐Q‐OrbitrapHRMS) method was applied to screen the ingredients of Epimedii Folium. Network pharmacology was utilized to explore the potential active components and pharmacological mechanisms of Epimedii Folium against ad. The binding affinity and conformation of key active ingredients and core targets were performed by molecular docking. Consequently, 72 chemical constituents were identified in Epimedii Folium, including four phenolic acids, two quinones, three 9,10‐dihydrophenanthrenes, and 63 flavonoids. Network pharmacology analysis revealed that quercetin, apigenin, kaempferol, and luteolin exhibit favorable pharmacological activities. Molecular docking demonstrated that apigenin–TNF, kaempferol–TNF, kaempferol–TP53, and quercetin‐TP53 represent compound–target pairs with strong binding affinities. These findings help elucidate the material basis and underlying mechanisms of Epimedii Folium against ad and offer valuable evidence supporting the further development and clinical application of Epimedii Folium.

1. Introduction

Alzheimer's disease (ad) is a highly prevalent neurodegenerative disease, which affects the brain and neurons through enormous reduction in nerve cell regenerative capacity (Maniam and Maniam 2024). Currently, approximately 35 million individuals worldwide are affected by ad, and the number is estimated to reach 70 million by 2030, with China accounting for about 30% of the cases (Cheneau and Rapp 2025). The pathological study showed that ad is characterized by extracellular deposition of β‐amyloid and intracellular accumulation of hyperphosphorylated tau, forming senile plaques and neurofibrillary tangles (NFTs), respectively (Hur 2022). Although the amyloid‐β and tau biomarkers are very helpful in tracking ad progression, unfortunately, they have not yet transformed into the treatment for ad. To date, four drugs, including an N‐methyl‐D‐aspartate receptorantagonist (memantine) and three cholinesterase inhibitors (galantamine and donepeziand rivastigmine), had been approved to treat ad by the FDA (Panos 2024). However, these drugs cannot completely cure ad and most of them had serious side effects, such as hepatotoxicity and nephrotoxicity. Therefore, there is an urgent need to seek and develop new drugs for the treatment of ad.

Traditional Chinese Medicine (TCM) has long been considered as potential drug resources in the treatment of ad due to its notable therapeutic effects and minimal side effects (Li and Yang 2024). Epimedii Folium, derived from the dried leaves of Epimedium brevicornum Maxim., E. sagittatum (Sieb. et Zucc.) Maxim., E. pubescens Maxim., and E. koreanum Nakai., is a commonly used TCM (Chen et al. 2024). Chemical investigation indicated that flavonoids were the main ingredients of Epimedii Folium (Yuan et al. 2014). Modern pharmacological studies have revealed that flavonoids in Epimedii Folium can improve ad and has neuroprotective effects (Wang et al. 2024). However, the key effective compounds of Epimedii Folium fortreating ad were still unclear, and the pharmacologic mechanisms against ad have never been clarified.

In recent years, ultra‐high performance liquid chromatography coupled to quadrupole Orbitrap high‐resolution mass spectrometry (UHPLC‐Q‐OrbitrapHRMS) as a modern analytical technique has been widely used in the identification of effective components of TCM (Pang et al. 2025). Ultra‐high‐performance liquid chromatography (UHPLC) has the advantages of high separation capacity, fast analysis speed and high accuracy, while quadrupole Orbitrap high‐resolution mass spectrometry (Q‐OrbitrapHRMS) has the characteristics of high throughput and high resolution, thus showing a superior qualitative ability. The combination of UHPLC with Q‐Orbitrap HRMS allows for the detection of complicated components with higher resolution and sensitivity, which shows more advantages than traditional analytical methods.

Network pharmacology is an important tool for studying complex diseases and revealing the complex relationship between proteins, diseases, and drugs (An et al. 2020). It uses the network visualization method to analyze the complex interaction relationship among diseases, drugs, and targets, which has the characteristics of integrity and systematicness. It has been widely used to predict the mechanism of complex diseases and drugs under the context of the whole biological networks.

Molecular docking is a fast and efficient computational method for the prediction of the binding mode and binding affinity between a ligand and a target protein at the atomic level (Wang et al. 2023). In recent years, molecular docking has emerged as a powerful tool in the study of active sites of the effective components in TCM (Lin et al. 2022).

In this study, we proposed UHPLC‐Q‐Orbitrap HRMS, network pharmacology, and molecular docking to disclose the material basis and potential mechanisms of Epimedii Folium in the treatment of ad. To begin with, the chemical components of Epimedii Folium were screened and identified using UHPLC‐Q‐Orbitrap HRMS. Next, the targets of main bioactive compounds and ad were collected using network pharmacology methods. Then, the key active ingredients‐core targets network as well as the protein to protein interaction (PPI) network was established. The potential mechanisms were further investigated by gene oncology (GO) bioanalysis and kyoto encyclopedia of genes and genomes (KEGG) pathway analysis. Finally, the binding affinity and interactions of bioactive components to ad‐related targets were verified through molecular docking. The results of this study provided a basis for the development and utilization of Epimedii Folium.

2. Materials and Methods

2.1. Reagents and Materials

Epimedii Folium was purchased from Bozhou (Anhui province, China) and identified in accordance with Chinese Pharmacopoeia (2020 Edition). The voucher specimen (20240816) was kept in the institute of Pharmaceutical and Life Sciences, Jiujiang University. Reference standards of apigenin, kaempferol, luteolin, caffeic acid, kaempferitrin and quercetin (all purity > 99%) were purchased from the national institutes for the control of pharmaceutical and biological products (Beijing, China).

HPLC‐grade methanol and acetonitrile were purchased from Merck Company (Darmstadt, Germany), and MS‐grade formic acid was obtained from Sigma‐Aldrich (Sigma‐Aldrich, MO, USA). Water was prepared by the Milli‐Q water purification system (Millipore, Bedford, MA, USA).

2.2. Preparation of Sample

The sample of Epimedii Folium was ground into powder. A 1.0 g portion of the sample powder was accurately weighed and refluxed with 10 mL methanol. The methanol solution was centrifuged at 10000 × g for 15 min, and then the supernatant was filtered through a 0.22 μm membrane for UHPLC‐Q‐Orbitrap HRMS analysis.

2.3. UHPLC Conditions

UHPLC analysis was performed on an Ultimate 3000 RSLCnano system; meanwhile, a hyper gold C18 column (100 × 2.1 mm, 1.9 μm) was applied to separate the mixed compounds, and the column temperature was set at 30°C. The mobile phases consisted of water containing 0.1% formic acid (A) and acetonitrile (B), and the elution gradient was set as follows: 0.0–3.0 min, 5.0% – 20.0% B; 3.0–8.0 min, 20.0% B; 8.0–13.0 min, 20.0%– 50.0% B; 13.0–18.0 min, 50.0% B; 18.0–22.0 min, 50.0%– 80.0% B; 22.0–26.0 min, 80.0% B; 26.0–35.0 min, 80.0%– 95.0% B; 35.0–40.0 min, 95.0% B. The flow rate was set at 0.3 mL/min, and the injection volume was 2 μL.

2.4. Mass Spectrometry Conditions

HRMS experiments were conducted on a Q‐Exactive hybrid quadrupole‐orbitrap mass spectrometer (Thermo Scientific, San Jose, CA, USA), equipped with heat ESI operating in positive and negative ionization modes. The ESI source parameters were set as follows: spray voltage, 3.6 kV (+); spray voltage, 3.3 kV (−); capillary temperature, 350°C; capillary voltage, 35 V; tube lens voltage, 110 V; sheath gas (nitrogen) flowrate of 30 arb, aux gas (nitrogen) flow rate of 5 arb. The mass scan range was between m/z 100 and 1500.

2.5. Acquisition of Compound Targets

The compounds identified from Epimedii Folium by UHPLC‐Q‐Orbitrap HRMS were evaluated using the 3D structures from PubChem (data accessed July 2024) and then saved in SMILES format. Further compounds were collected from TCMSP database version 2.3 (2014). Compounds were filtered by oral bioavailability (OB) ≥ 20%, drug‐likeness (DL) ≥ 0.1, and redundant entries were abandoned. These structures were analyzed for potential human targets using the Swiss Target Prediction server.

2.6. Acquisition of ad Targets

The keyword “Alzheimer's disease” was used to search disease targets by the online platforms of GeneCards (version 5.21), DisGeNET (version 24.2), and OMIM (2024 version). A Venn diagram identified overlaps between the compound targets of Epimedii Folium and the ad‐related targets.

2.7. Protein–Protein Interaction Analysis

Intersection targets were uploaded to the STRING database version 12.0 and the organism was set to Homo sapiens before submission. Construct a network of protein–protein interactions utilizing H. sapiens as the protein species, with the protein interaction confidence threshold set at the highest level of confidence ≥ 0.9. The information from the STRING database was imported into the Cytoscape 3.7.2 software for data analysis to find the core targets.

2.8. KEGG Pathway and GO Enrichment Analysis

The intersection targets obtained from the Venn diagrams were imported into DAVID database 2021 for GO enrichment and KEGG pathway analysis. As part of the enrichment analysis, “official gene symbol” was selected as the identifier, “gene list” as the list type, and the organism was set to H. sapiens . Under the condition of adjusted p‐value (FDR) < 0.05, the top 20 enriched GO terms, including molecular function, biological process and cellular components, were visualized on a bar chart and the first 20 enriched KEGG pathways were visualized on a bubble map to reveal the multipathway and multifunction mechanisms of Epimedii Folium in treating ad.

2.9. Drug‐Compound‐Target‐Pathway‐Disease Network

The top 20 KEGG pathways, corresponding pathway targets, and active ingredients were used to construct the “compound‐target‐pathway” network by Cytoscape software. In the network, active ingredients, targets, and edges represent the interactions area, which was used to seek the key ingredients of Epimedii Folium against ad.

2.10. Molecular Docking

Molecular docking simulations were carried out via AutoDock Vina 1.1.2 to predict the binding affinity and optimal binding conformations between core bioactive compounds and hub target proteins. The SDF files of key active constituents were retrieved from the PubChem database, then imported and converted into PDB files using PyMOL 2.3.2. After ligand structures were loaded into AutoDock Tools, atomic types were assigned and partial atomic charges were computed; the preprocessed ligands were subsequently exported as PDBQT files for downstream docking calculations. Crystal structures of receptor proteins were downloaded from the RCSB Protein Data Bank (data accessed July 2024), and all crystal water molecules as well as heteroatoms were eliminated within PyMOL 2.3.2. The cleaned receptor structures were imported into AutoDock Tools for atomic type assignment and charge calculation, before being saved in PDBQT format. Preprocessed receptor and ligand PDBQT files were loaded into AutoDock Vina 1.1.2 to define the docking search space. All ligands were treated as flexible molecules during simulations, whereas receptors were kept rigid. The geometric centroid of each target protein was designated as the central coordinate of the grid box. The X/Y/Z dimensions and central X/Y/Z coordinates of each grid box were fine‐tuned to ensure the complete binding pocket and full protein backbone of every receptor were fully enclosed within the search region.

3. Result and Discussion

3.1. Chemical Components Analysis of Epimedii Folium

The chemical components in Epimedii Folium were characterized by the reliable UHPLC‐Q‐Orbitrap HRMS method. The components in Epimedii Folium showed higher MS response in negative ion mode than that in positive ion mode, so the negative ion mode was applied in this study. The base peak chromatogram (BPC) of Epimedii Folium by UHPLC‐Q‐Orbitrap HRMS was shown in Figure 1. A total of 72 compounds, including four phenolic acids, two quinones, three 9, 10‐dihydrophenanthrenes and 63 flavonoids, were identified from Epimedii Folium based on the retention time, accurate mass, fragmentation pattern and MS/MS spectra, as shown in Table 1.

FIGURE 1.

The base peak chromatograms (BPC) of Epimedii Folium extract. (A) Negative ion mode; (B) Positive ion mode.

TABLE 1.

Characterisation of chemical constituents in Epimedii Folium by UHPLC‐Q‐Orbitrap HRMS.

	tR				Measured	Accuracy	Error	MS/MS fragment ions
No.	(min)	Formula	Compound	Selected ion	Mass ( m/z )	Mass ( m/z )	(ppm)	( m/z )
1	2.52	C7H6O4	Protocatechuic acid	[M‐H]−	153.0185	153.0182	1.99	109.0285
2	3.41	C9H8O3	Naringeninic acid	[M‐H]−	163.0393	163.0390	2.15	119.0492
3	4.16	C7H6O3	4‐Hydroxybenzoic acid	[M‐H]−	137.0235	137.0233	1.30	119.0128, 93.0335
4	9.65	C18H24O10	Sagittatoside M	[M‐H]−	399.1301	399.1286	3.82	219.0661, 193.0865
5	9.82	C21H20O12	Hyperoside	[M‐H]−	463.0890	463.0871	4.07	300.0277, 271.0251, 255.0300
6	10.34	C32H38O18	Sagittatin A	[M‐H]−	709.1978	709.1974	0.44	563.1416, 283.0250, 255.0301
7	10.37	C27H30O14	Kaempferitrin	[M‐H]−	577.1569	577.1563	1.04	431.0988, 285.0406, 255.0299
8	10.68	C39H32O14	Kaempferol‐3‐O‐(2″‐E‐p‐coumaroyl, 4″‐Z‐p‐coumaroyl)‐α‐L‐rhamnopyranoside	[M‐H]−	723.2157	723.2131	3.59	269.0458
9	10.81	C33H42O16	Icarisid B	[M‐H]−	693.2044	693.2019	3.60	297.0407
10	10.96	C21H20O11	Quercitrin	[M‐H]−	447.0939	447.0922	3.90	300.0279, 285.0409
11	11.14	C24H30O10	Epimedoicarisoside A	[M‐H]−	477.1044	477.1031	2.72	314.0435, 299.0200, 271.0250
12	11.44	C9H8O4	Caffeic acid	[M‐H]−	179.0344	179.0339	2.75	151.0393, 109.0241
13	11.63	C15H10O6	Luteolin	[M‐H]−	285.0404	285.0394	3.65	267.0300, 229.0504
14	11.80	C32H38O16	3, 4, 5‐Trihydroxy‐8‐prenylflavone 7‐O‐[β‐D‐glucopyranosyl(1 → 2)‐β‐D‐glucopyranoside]	[M‐H]−	677.2092	677.2076	2.30	515.1564, 353.1035, 323.0928
15	11.81	C18H24O10	Sagittatoside N	[M‐H]−	399.1301	399.1286	3.89	219.0660, 193.0865
16	11.95	C38H48O20	Diphylloside A	[M‐H]−	823.2673	823.2655	2.20	661.2145, 353.1034
17	12.11	C21H20O10	Kaempferol‐3‐O‐rhamnoside	[M‐H]−	431.0989	431.0973	3.72	285.0407, 255.0300, 227.0348
18	12.20	C28H32O12	Koreanoside D	[M‐H]−	559.1827	559.1810	3.11	413.1245, 323.0563
19	12.27	C37H46O19	Epimedoside E	[M‐H]−	793.2554	793.2550	0.52	631.2037, 351.0876
20	12.52	C32H38O15	Epimedin B	[M‐H]−	661.2143	661.2127	2.43	515.1563, 353.1035
21	12.62	C33H40O16	Sagittasine C	[M + HCOO]−	737.2309	737.2298	1.49	529.1721, 383.1141
22	13.44	C16H12O6	6‐Demethoxy‐7‐methylcapillarisi	[M‐H]−	299.0564	299.0550	4.52	284.0331, 256.0576
23	13.68	C39H48O20	4, 5‐Dihydroxyl‐8‐(3, 3‐dimethylallyl)‐flavonol 3‐O‐[β‐D‐xylopyranosyl‐(1 → 3)‐4‐O‐acetyl‐α‐L‐rhamnopyranoside]‐7‐O‐β‐D‐glucopyranoside	[M‐H]−	835.2645	835.2655	−1.19	673.2146, 353.1034
24	14.58	C39H50O20	Epimedin A	[M‐H]−	837.2828	837.2812	1.92	675.2296, 513.1780, 366.1111, 351.0873
25	14.71	C32H38O14	Sagittatoside B or isomer	[M‐H]−	645.2197	645.2178	2.92	366.1109, 351.0877, 323.0928
26	14.75	C38H48O19	Epimedins B	[M‐H]−	807.2723	807.2717	0.74	645.2194, 366.1115
27	14.92	C39H50O19	Epimedin C	[M‐H]−	821.2880	821.2874	0.73	659.2355, 366.1111, 351.0876
28	14.95	C40H50O21	Koreanoside H	[M + HCOO]−	911.2144	911.2123	2.30	865.2048, 805.1857, 749.1359
29	14.99	C33H40O14	2″‐O‐Rhamnosyl icarisoside II or isomer	[M‐H]−	659.2356	659.2334	3.28	366.1113, 351.0878, 323.0929
30	15.12	C32H40O16	Epimedkoresides A	[M‐H]−	679.3693	679.3672	3.09	517.3182, 370.1008
31	15.14	C27H30O10	Icariside II or isomer	[M‐H]−	513.1776	513.1755	3.95	366.1111, 351.0876, 323.0927
32	15.19	C33H40O15	Icariinor isomer	[M‐H]−	675.2299	675.2284	2.24	513.1772, 367.1190, 351.0876, 323.0929
33	15.23	C27H30O11	Sagittatosides D‐N7	[M‐H]−	529.1718	529.1715	0.57	367.1192
34	15.51	C27H32O10	Icaritin‐3‐O‐rhamnoside	[M‐H]−	531.1880	531.1861	3.58	384.1219, 369.0979, 341.1034
35	15.56	C17H14O5	5‐Hydroxy‐7, 4′‐dimethoxyflavone	[M‐H]−	297.0769	297.0758	3.89	281.0457, 267.0304
36	15.76	C39H48O19	8‐(3, 3‐Dimethylallyl)‐5, 7‐dihydroxy‐4′‐methoxyflavonol 3‐[O‐2‐C‐carboxy‐3, 5‐di‐deoxy‐β‐D‐erythro‐pentofuranosyl‐(1 → 2)‐α‐L‐rhamnopyranoside] 7‐(β‐D‐glucopyranoside)	[M‐H]−	819.2726	819.2717	1.10	657.2194, 513.1772, 367.1189
37	16.76	C37H44O17	Korepimedoside A	[M‐H]−	759.2501	759.2495	0.85	367.1190, 352.0955
38	16.98	C37H48O19	Epimedin B	[M + HCOO]−	841.4601	841.4597	0.47	633.4014
39	17.05	C33H38O15	Sagittatosides D‐N3	[M‐H]−	673.2144	673.2127	2.19	381.0983, 366.0748, 337.0722
40	17.19	C40H50O20	Epimedin B	[M‐H]−	849.2837	849.2812	2.37	687.2300, 367.1189, 352.0954
41	17.45	C31H36O14	Icarisoside F	[M‐H]−	631.2042	631.2021	3.22	352.0955
42	17.47	C40H50O20	Sempervirenoside B	[M‐H]−	849.2836	849.2812	2.90	687.2302, 367.1190, 352.0955
43	17.61	C15H10O7	Quercetin	[M‐H]−	301.2022	301.2017	1.66	283.1917, 265.1815, 221.1910
44	17.68	C17H14O7	Morelosin	[M‐H]−	329.2338	329.2319	5.77	311.2238, 293.2144
45	17.97	C35H42O16	Epimedokoreanoside II	[M + HCOO]−	763.2461	763.2455	0.79	555.1872, 367.1187
46	18.00	C26H28O10	Epimedoside C	[M‐H]−	499.1616	499.1610	1.20	353.1034, 309.0409, 281.0460
47	18.34	C39H46O18	Korepimeoside A	[M‐H]−	801.2621	801.2600	2.55	367.1191, 352.0955
48	19.16	C32H40O15	Epimedoside A	[M‐H]−	663.3758	663.3746	1.81	501.3229
49	19.45	C33H40O15	Icariinor isomer	[M‐H]−	675.2297	675.2284	1.97	367.1190, 352.0955, 323.0920
50	19.63	C21H16O5	Epimesatines A–I7	[M‐H]−	347.1867	347.1855	3.46	303.1967
51	19.84	C15H10O5	Apigenin	[M‐H]−	269.1314	269.1309	1.86	226.8879, 205.1590
52	20.01	C32H38O14	Sagittatoside B or isomer	[M‐H]−	645.2199	645.2178	3.30	366.1121, 351.0873, 323.0928
53	20.20	C33H40O14	2″‐O‐Rhamnosyl icarisoside II or isomer	[M‐H]−	659.2356	659.2334	3.28	366.1111, 351.0877, 323.0920
54	20.93	C20H16O6	Yinyanghuo E or isomer	[M‐H]−	351.0879	351.0874	1.42	307.0624, 279.0663
55	20.94	C27H30O10	Icariside II or isomer	[M‐H]−	513.1774	513.1755	3.59	366.1111, 351.0878, 323.0928
56	21.14	C21H20O6	Icaritin	[M‐H]−	367.1190	367.1176	3.91	337.0726, 309.0408, 297.0406
57	22.31	C32H36O14	Koreanoside A	[M‐H]−	643.2039	643.2039	2.78	365.1034, 350.0799
58	22.93	C20H18O5	Yinyanghuo D or isomer	[M‐H]−	337.1086	337.1070	4.63	281.0459, 253.0504
59	23.24	C20H16O5	Yinyanghuo C or isomer	[M‐H]−	335.0928	335.0914	4.19	305.0454, 291.1027
60	23.39	C20H18O5	Yinyanghuo D or isomer	[M‐H]−	337.1085	337.1070	4.36	281.0458, 253.0514
61	23.85	C25H26O6	Epimedokoreanin B	[M‐H]−	421.1662	421.1646	3.81	301.1084, 119.0492
62	23.95	C18H18O5	Epicornunins C	[M‐H]−	313.2387	313.2373	6.38	201.1128
63	24.73	C20H18O6	8‐Prenylkaempferol	[M‐H]−	353.1398	353.1382	4.53	309.1499
64	25.09	C20H16O5	Yinyanghuo C or isomer	[M‐H]−	335.0929	335.0914	4.46	305.0458, 291.0674
65	25.23	C25H26O6	Epimesatine D	[M‐H]−	421.1661	421.1646	3.66	301.1083, 119.0492
66	25.52	C15H10O6	Kaempferol	[M‐H]−	285.2074	285.2056	6.31	267.1698, 223.2065
67	25.57	C25H26O5	5, 7, 4′‐Trihydroxy‐8, 3′‐prenylflavone or isomer	[M‐H]−	405.1716	405.1707	2.22	387.1606, 119.0493
68	25.72	C18H20O5	4‐Hydroxy‐2, 3, 6, 7‐tetramethoxy‐9, 10‐dihydrophenanthrene	[M‐H]−	315.2542	315.2527	4.75	297.2437, 279.2332
69	26.84	C25H26O5	5, 7, 4′‐Trihydroxy‐8, 3′‐prenylflavone or isomer	[M‐H]−	405.1714	405.1697	4.29	387.1619, 119.0492
70	30.77	C25H24O6	Yinyanghuo A	[M‐H]−	419.1502	419.1489	3.16	391.1555, 375.1611
71	31.10	C25H24O5	Epimesatine A	[M‐H]−	403.1553	403.1540	3.23	373.1037, 347.0930
72	34.21	C20H16O6	Yinyanghuo E or isomer	[M‐H]−	351.0877	351.0863	3.97	307.0973

3.1.1. Phenolic Acids

Four phenolic acids (1, 2, 3, 12) were characterized in Epimedii Folium. Peak 1 exhibited [M − H]⁻ion at m/z 153.0185 due to the neutral loss of CO₂, as shown in Figure 2. Thus, Peak 1 was characterized as protocatechuic acid by comparison with the literature (Su et al. 2018). Peak 2 gave a molecular ion [M − H]⁻ at m/z 163.0393 and diagnostic product ion at m/z 119.0492 by the neutral loss of CO₂, which were consistent with the MS data in the literature, as shown in Figure 2. Thus, it was characterized as naringeninic acid (Su et al. 2018). Peak 3 showed [M − H]⁻ ion at m/z 137.0235 and fragment ions at m/z 119.0128 [M − H − H₂O]⁻ and 93.0335 [M − H − CO₂]⁻, which were in agreement with the previous literature, as shown in Figure 2. Therefore, it was identified as 4‐hydroxybenzoic acid (Su et al. 2018). Peak 12 exhibited a precursor ion [M − H]⁻ at m/z 179.0344 and produced fragment ions at m/z 151.0393 [M − H − CO]⁻and109.0241 [M‐H‐C₃H₂O₂]⁻, as shown in Figure 2. Comparing with the reported literature, Peak 12 was thus identified as caffeic acid (Su et al. 2018).

FIGURE 2.

MS/MS spectra and proposed fragmentation pathways of Compounds 1, 2, 3, and 12.

3.1.2. Quinones

Two quinones were characterized in Epimedii Folium (4 and 15). Peak 4 exhibited a precursor ion [M − H]⁻at m/z 399.1301, and the fragment ions were observed at m/z 219.0661[M‐H‐Glc‐H₂O]⁻and 193.0865 [M‐H‐Glc‐CO₂]⁻. Based on the molecular weight and fragmentation pattern, Peak 4 was tentatively identified as sagittatoside M (Lin et al. 2022). Peak 15 was judged to be the isomer of Compound 4 due to the same precursor ion and fragment ions, and was thus identified as sagittatoside N (Lin et al. 2022).

3.1.3. 9,10‐Dihydrophenanthrenes

Three 9, 10‐dihydrophenanthrenes were characterized in Epimedii Folium (11, 62, and 68). Peak 11 exhibited an [M − H]⁻ ion at m/z 477.1044, and the characteristic ions at m/z 314.0435 [M‐H‐Glc]⁻, 299.0200 [M‐H‐Glc‐CH₃]⁻and 271.0250 [M‐H‐Glc‐CH₃‐CO]⁻were observed in the MS² spectrum. Based on the molecular weight and fragmentation pattern, Peak 11 was tentatively identified as epimedoicarisoside A (Li et al. 1995). Peaks 62 and 68 were respectively identified as epicornunins C and 4‐hydroxy‐2, 3, 6, 7‐tetramethoxy‐9, 10‐dihydrophenanthrene (Pang et al. 2018).

3.1.4. Flavonoids

Flavonoids were the major chemical constituents of Epimedii Folium. A total of 63 flavonoids (5–10, 13, 14, 16–61, 63–67, and 69–72) were identified in Epimedii Folium. Peak 7 had [M‐H]⁻ ion at m/z 577.1569 with the retention time at 10.37 min. It produced the fragment ions at m/z 431.0988 [M‐H‐146 Da]⁻ and 285.0406 [M‐H‐2 × 146 Da]⁻ by the successive losses of Rha moiety. The fragment ion at m/z 255.0299 was formed by losing CH₂O from the fragment ion at m/z 285.0406, as shown in Figure 3. Therefore, we defined Compound 7 as kaempferitrin by comparison to the reference standard (Han and Lee 2017). Peak 16 showed a molecular ion at m/z 823.2678 with the molecular formula C₃₈H₄₈O₂₀. The fragment ions at m/z 661.2145 [M‐H‐162 Da]⁻ and 353.1034 [M‐H‐2 × 162 Da–146 Da]⁻ were yielded by the successive losses of Glc and Rha residues, as shown in Figure 3. Thus, Peak 16 was identified as diphylloside A by comparison with the reference standard (Han and Lee 2017). Peak 20 gave [M‐H]⁻ ion at m/z 661.2147 (C₃₂H₃₈O₁₅). The fragment ions at m/z 515.1563 and 353.1034 were produced by the successive losses of Rha and Glc residues, as shown in Figure 3. After reference to the standard compound, Peak 20 was unambiguously identified as epimedin B (Kim et al. 2021). Peak 46 gave [M‐H]⁻ ion at m/z 499.1616 with the molecular formula of C₂₆H₂₈O₁₀. MS/MS investigation revealed the major fragment ions at m/z 353.1034 [M‐H‐146 Da]⁻, m/z 309.0409 [M‐H‐146 Da‐44 Da]⁻, and m/z 281.0460 [M‐H‐146 Da‐44 Da‐28 Da]⁻, indicating the successive losses of Rha, CO₂, and CO residues, as shown in Figure 3. Thus, Peak 46 was unambiguously identified as epimedoside C by comparison to the standard (Li et al. 2016).

FIGURE 3.

MS/MS spectra and proposed fragmentation pathways of Compounds 7, 16, 20, and 46.

3.2. Network Pharmacology

3.2.1. Potential Target Identification of Main Components and ad

After merging and removing duplicates, 2946 targets associated with ad were obtained from OMIM, GeneCard, and DisGeNET databases, respectively. After supplementing targets from UHPLC‐Q‐Orbitrap HRMS data and the TCMSP database, merging datasets and eliminating duplicates, 232 potential targets linked to the characterized compounds were acquired. A total of 161 intersection targets associated with Epimedii Folium in the treatment of ad were displayed by the Vennytool, as shown in Figure 4. The chemical constituents of Epimedii Folium were screened based on potential therapeutic targets of Epimedii Folium for the treatment of Alzheimer's disease. Compounds without corresponding target associations were excluded, and potential active ingredients that interact with disease targets were retained. The screening results are shown in Table 2.

FIGURE 4.

Venn diagram of the targets of Epimedii Folium and ad.

TABLE 2.

Key active compounds found in Epimedii Folium.

NO.	Molecule name
Epi1	Luteolin
Epi2	Apigenin
Epi3	Quercetin
Epi4	Oleanolic acid
Epi5	(2R,3R)‐2‐(3,4‐dimethoxyphenyl)‐7‐methoxy‐3‐methyl‐5‐[(E)‐prop‐1‐enyl]‐2,3‐dihydrobenzofuran
Epi6	Sitogluside
Epi7	Sitosterol
Epi8	Kaempferol
Epi9	Emodin
Epi10	Magnograndiolide
Epi11	Patchouli alcohol
Epi12	(+)‐Cycloolivil
Epi13	24‐epicampesterol
Epi14	Copaene
Epi15	Linoleyl acetate
Epi16	Poriferast‐5‐en‐3beta‐ol
Epi17	Isoliquiritigenin
Epi18	DFV
Epi19	Tricin
Epi20	Junipene
Epi21	Chryseriol
Epi22	Flavone der.
Epi23	8‐Isopentenyl‐kaempferol
Epi24	Olivil
Epi25	Anhydroicaritin
Epi26	C‐Homoerythrinan, 1,6‐didehydro‐3,15,16‐trimethoxy‐, (3.beta.)—
Epi27	Besigomsin
Epi28	Yinyanghuo A
Epi29	Yinyanghuo C
Epi30	Yinyanghuo E
Epi31	6‐hydroxy‐11,12‐dimethoxy‐2,2‐dimethyl‐1,8‐dioxo‐2,3,4,8‐tetrahydro‐1H‐isochromeno[3,4‐h]isoquinolin‐2‐ium
Epi32	8‐(3‐methylbut‐2‐enyl)‐2‐phenyl‐chromone
Epi33	Anhydroicaritin
Epi34	1,2‐bis(4‐hydroxy‐3‐methoxyphenyl)propan‐1,3‐diol
Epi35	Icariside A7
Epi36	Icariside I

3.2.2. Construction and Analysis of PPI Network

The PPI network is shown in Figure 5. After hiding the disconnected nodes in the network, 144 nodes and 924 edges are shown in the network. According to the results of network analyzer, the average degree value of PPI network was 6.78, and 5 target genes (Table 3) were greater than three times the average degree value, indicative of them as the potential genes related to ad. At the same time, these potential genes are higher than the median of BetweennessCentrality and ClosenessCentrality.

FIGURE 5.

The PPI network of Epimedii Folium and ad core targets.

TABLE 3.

Genes with degree values more than three times the average.

No.	Protein	Gene	Uniprot ID	Degree
1	Cellular tumor antigen p53	TP53	P04637	39
2	RAC‐alpha serine/threonine‐protein kinase	AKT1	P31749	29
3	Tumor necrosis factor	TNF	P01375	28
4	Interleukin‐6	IL6	P05231	27
5	Transcription factor Jun	JUN	P05412	26

3.2.3. GO Analysis and KEGG Pathway Analysis

The 161 intersection targets were used for Gene Ontology and KEGG pathway enrichment by inputting them into the DAVID database. With the condition of adjusted p‐value (FDR) < 0.05, the top 20 enriched results of biological process, cellular component, and molecular function with the minimum p‐value involved are shown in Figure 6. In the biological process, the terms mainly involved positive regulation of gene expression, response to xenobiotic stimulus, and positive regulation of MAPK cascade. In the cellular component, the terms mainly involved extracellular space, caveola, and extracellular region. In the molecular function, terms mainly involved enzyme binding, identical protein binding, and protein binding. As a result, under the condition of adjusted p‐value (FDR) < 0.05, 20 most significantly enriched GO terms were obtained, and the Top 3 were the pathways in cancer, PI3K‐Akt signaling pathway, and lipid and atherosclerosis, as shown in Figure 7. The top 20 enriched pathways of common targets between Epimedii Folium and ad are shown in Table 4.

FIGURE 6.

GO enrichment analysis of core targets.

FIGURE 7.

KEGG pathway enrichment analysis of core targets.

TABLE 4.

Top 20 enriched pathways of common targets between Epimedii Folium and Alzheimer's disease.

Pathway ID	Pathways name	Adjusted p	Genes number	Genes
hsa05160	Hepatitis C	2.04 × 10−18	28	RB1, GSK3B, CDKN1A, TNF, RELA, EGFR, CASP9, IKBKB, CASP8, RXRA, CCND1, MYC, CASP3, E2F1, AKT1, MAPK1, CHUK, EGF, STAT1, CFLAR, NFKBIA, CXCL10, IFNG, BAX, CYCS, RAF1, PPARA, TP53
hsa05161	Hepatitis B	1.24 × 10−23	33	RB1, CDKN1A, PCNA, CXCL8, ELK1, TNF, RELA, CASP9, IKBKB, MAPK8, CASP8, MYC, CASP3, E2F1, AKT1, MAPK1, JAK2, JUN, TGFB1, CHUK, PRKCB, STAT1, PRKCA, FOS, MAPK14, MMP9, NFKBIA, IL6, BCL2, BAX, CYCS, RAF1, TP53
hsa05163	Human cytomegalovirus infection	2.28 × 10−19	33	RB1, GSK3B, CDKN1A, CXCL8, PTGER3, PTGS2, ELK1, TNF, RELA, EGFR, CASP9, IKBKB, PPP3CA, CASP8, CCND1, MYC, CASP3, E2F1, CCL2, AKT1, MAPK1, CHUK, PRKCB, PRKCA, MAPK14, VEGFA, NFKBIA, IL6, IL1B, BAX, CYCS, RAF1, TP53
hsa05167	Kaposi sarcoma‐associated herpesvirus infection	2.57 × 10−22	34	RB1, GSK3B, CDKN1A, CSF2, CXCL8, PTGS2, HIF1A, RELA, PIK3CG, ICAM1, CASP9, IKBKB, PPP3CA, MAPK8, CASP8, CCND1, MYC, CASP3, E2F1, AKT1, MAPK1, JAK2, JUN, CHUK, STAT1, FOS, MAPK14, VEGFA, NFKBIA, IL6, BAX, CYCS, RAF1, TP53
hsa05200	Pathways in cancer	4.81 × 10−37	66	RB1, GSK3B, CDKN1A, CXCL8, FLT4, PTEN, SLC2A1, ELK1, IGF1R, CASP9, IKBKB, CASP7, CASP8, CCND1, MYC, CASP3, AKT1, JAK2, CHUK, PRKCB, MMP1, MMP2, IL13, PRKCA, FOS, MMP9, AR, IFNG, PPARG, RAF1, MET, TP53, PPARD, GSTP1, PTGER3, PTGS2, HIF1A, EGFR, RELA, MAPK8, RXRA, ERBB2, E2F1, HMOX1, MAPK1, RUNX1T1, NQO1, JUN, GSTM1, TGFB1, NOS2, EGF, STAT1, IGF2, ESR1, ESR2, IL2, VEGFA, NFKBIA, IL4, IL6, APC, BCL2, CYCS, BAX, NFE2L2
hsa05205	Proteoglycans in cancer	1.15 × 10−16	29	CDKN1A, ELK1, HIF1A, TNF, EGFR, IGF1R, ERBB3, CCND1, PLAU, MYC, CASP3, ERBB2, KDR, AKT1, MAPK1, TGFB1, PRKCB, CAV1, MMP2, IGF2, PRKCA, MAPK14, MMP9, ESR1, VEGFA, COL1A1, RAF1, MET, TP53
hsa05207	Chemical carcinogenesis—receptor activation	3.79 × 10−16	29	RB1, NR1I3, ADRB1, ADRB2, CYP3A4, RELA, EGFR, RXRA, CCND1, MYC, E2F1, AKT1, MAPK1, JAK2, JUN, GSTM1, PRKCB, EGF, PRKCA, FOS, ESR1, ESR2, VEGFA, AR, CYP1A2, CYP1A1, BCL2, RAF1, PPARA
hsa05212	Pancreatic cancer	3.36 × 10−16	20	RB1, CDKN1A, TGFB1, CHUK, STAT1, EGF, EGFR, RELA, VEGFA, CASP9, IKBKB, MAPK8, CCND1, ERBB2, E2F1, BAX, AKT1, MAPK1, RAF1, TP53
hsa05215	Prostate cancer	1.15 × 10−22	27	RB1, GSK3B, CDKN1A, GSTP1, PTEN, PLAT, RELA, EGFR, IGF1R, INS, CASP9, IKBKB, CCND1, PLAU, ERBB2, E2F1, AKT1, MAPK1, CHUK, EGF, MMP3, MMP9, NFKBIA, AR, BCL2, RAF1, TP53
hsa05417	Basal cell carcinoma	1.59 × 10−27	5	GSK3B, CDKN1A, APC, BAX, TP53
hsa05418	Fluid shear stress and atherosclerosis	2.38 × 10−24	32	NCF1, GSTP1, PLAT, TNF, RELA, ICAM1, IKBKB, THBD, MAPK8, KDR, CCL2, AKT1, HMOX1, NQO1, JUN, GSTM1, VCAM1, CHUK, NOS3, CAV1, MMP2, FOS, MAPK14, SELE, MMP9, VEGFA, IL1A, IFNG, IL1B, BCL2, TP53, NFE2L2
hsa04010	MAPK signaling pathway	1.48 × 10−18	36	FLT1, FLT4, HSPB1, ELK1, TNF, RELA, EGFR, IGF1R, INS, IKBKB, PPP3CA, MAPK8, ERBB3, MYC, CASP3, ERBB2, KDR, AKT1, MAPK1, JUN, TGFB1, CHUK, PRKCB, EGF, INSR, IGF2, PRKCA, FOS, MAPK14, VEGFA, IL1A, RASA1, IL1B, RAF1, MET, TP53
hsa04066	HIF‐1 signaling pathway	1.15 × 10−16	23	CDKN1A, FLT1, NOS2, PRKCB, NOS3, EGF, INSR, SERPINE1, SLC2A1, PRKCA, HIF1A, EGFR, RELA, INS, IGF1R, VEGFA, IL6, IFNG, ERBB2, BCL2, AKT1, HMOX1, MAPK1
hsa04151	PI3K‐Akt signaling pathway	1.89 × 10−19	40	CHRM2, GSK3B, CDKN1A, FLT1, CHRM1, FLT4, PTEN, RELA, EGFR, PIK3CG, IGF1R, INS, CASP9, IKBKB, RXRA, ERBB3, CCND1, MYC, ERBB2, KDR, SPP1, AKT1, MAPK1, JAK2, MCL1, CHUK, EGF, NOS3, INSR, IGF2, PRKCA, IL2, VEGFA, IL4, COL1A1, IL6, BCL2, RAF1, MET, TP53
hsa04657	IL‐17 signaling pathway	8.74 × 10−22	26	GSK3B, CSF2, CXCL8, PTGS2, TNF, RELA, IKBKB, MAPK8, CASP8, CASP3, CCL2, MAPK1, JUN, CHUK, MMP1, IL13, MMP3, FOS, MAPK14, MMP9, IL4, NFKBIA, CXCL10, IL6, IFNG, IL1B
hsa04668	TNF signaling pathway	2.65 × 10−19	26	CSF2, PTGS2, TNF, RELA, ICAM1, IKBKB, CASP7, MAPK8, CASP8, CASP3, CCL2, AKT1, MAPK1, JUN, VCAM1, CHUK, MMP3, CFLAR, FOS, MAPK14, SELE, MMP9, NFKBIA, CXCL10, IL6, IL1B
hsa04933	AGE‐RAGE signaling pathway in diabetic complications	5.52 × 10−29	32	CXCL8, SERPINE1, TNF, RELA, ICAM1, THBD, MAPK8, CCND1, CASP3, CCL2, AKT1, MAPK1, JAK2, JUN, TGFB1, VCAM1, PRKCB, NOS3, STAT1, MMP2, PRKCA, MAPK14, SELE, F3, VEGFA, COL1A1, IL1A, COL3A1, IL6, IL1B, BCL2, BAX
hsa05140	Leishmaniasis	3.36 × 10−16	20	IL10, JUN, TGFB1, NOS2, NCF1, PRKCB, STAT1, FOS, PTGS2, MAPK14, ELK1, TNF, RELA, IL4, NFKBIA, IL1A, IFNG, IL1B, MAPK1, JAK2
hsa05142	Chagas disease	3.44 × 10−17	23	IL10, JUN, TGFB1, CXCL8, NOS2, CHUK, SERPINE1, CFLAR, FOS, MAPK14, TNF, IL2, RELA, NFKBIA, IKBKB, IL6, MAPK8, CASP8, IFNG, IL1B, CCL2, AKT1, MAPK1
hsa05145	Toxoplasmosis	1.57 × 10−16	23	IL10, TGFB1, NOS2, CHUK, STAT1, MAPK14, TNF, PIK3CG, RELA, CASP9, NFKBIA, IKBKB, MAPK8, CASP8, CD40LG, IFNG, CASP3, ALOX5, BCL2, AKT1, MAPK1, CYCS, JAK2

3.2.4. Drug‐Compound–Target–Pathway‐Disease Network Analysis

In order to evaluate the key active ingredients of Epimedii Folium for treating ad, the top 20 KEGG corresponding targets, active compound, herb, intersection targets, and disease were entered into Cytoscapeto to make drug‐compound–target–pathway‐disease network. The network had 2356 edges and 218 nodes, as shown in Figure 8. The 20 green circles were pathways, the 34 yellow octagons were active ingredients, the orange v was herb, the green triangle was disease, and the 161 blue round rectangles were considered as the targets of ad. Each edge shows the interaction between nodes. The average degree value of the network was 16. Quercetin, apigenin, kaempferol, and luteolin, whose degree values were higher than three times the average degree, were identified as the core compounds of Epimedii Folium against ad.

FIGURE 8.

Drug‐compounds‐ad targets‐pathways‐therapeutic effects network of Epimedii Folium on ad.

3.2.5. Molecular Docking

A systematic molecular docking method was used to further verify the binding affinity between the active compound and the target. According to the analysis of the core targets and key active ingredients, quercetin, apigenin, kaempferol, and luteolin were selected to perform molecular docking with AKT1 (PDB ID: 3MVH), TP53 (PDB ID: 6GGC), TNF (PDB ID: 5M2J), and IL6 (PDB ID: 1ALU), respectively, as shown in Table 5. The lower the energy indicates the greater the affinity between the compound and the target and indicates higher structural stability. All the number of energies are lower than −6 kcal/mol, indicating that the affinity between the compound and the target was strong. Of those, the results of apigenin–TNF, kaempferol–TNF, and kaempferol–TP53 quercetin‐TP53 were better, and the visualization docking, as shown in Figure 9.

TABLE 5.

The binding free energy of each energy component.

Ligands receptors	Receptors binding	Energy (kcal/mol)
Apigenin	TNF	−8.09
	TP53	−7.07
	AKT1	−7.34
	IL6	−7.2
Kaempferol	TNF	−8.1
	TP53	−8
	AKT1	−6.68
	IL6	−6.85
Luteolin	TNF	−7.52
	TP53	−7.85
	AKT1	−7.42
	IL6	−6.71
Quercetin	TNF	−7.31
	TP53	−7.95
	AKT1	−7.94
	IL6	−6.07

FIGURE 9.

Molecular docking of the core targets and the corresponding compounds (A: Kaempferol acts on TNF, B: Kaempferol acts on TP53, C: Quercetin acts on TP53, D: Apigenin acts on TNF).

As illustrated in Figure 9A, the active compound occupies the target protein's binding pocket via hydrophilic and hydrophobic contacts and is tightly stabilized by extensive hydrogen‐bond networks with bond lengths spanning 1.6–3.3 Å. Notably, a 1.6 Å hydrogen bond formed between ASN‐34 and the ligand contributes the most robust intermolecular attraction.

Figure 9B reveals that the active ingredient interacts with the target protein's active cavity through hydrophilic moieties and hydrophobic scaffolds. Its hydroxyl moieties generate two short, strong hydrogen bonds with LEU‐145 and PRO‐152, with bond distances below 2.2 Å. Meanwhile, weak aliphatic van der Waals forces involving PRO‐152 consolidate the binding complex.

Figure 9C demonstrates that the active ingredient occupies the target protein's binding pocket through hydrophilic polar moieties, with hydrogen bond distances falling within 1.8–2.1 Å. The small molecule's hydroxyl groups form hydrogen bonds with CYS‐220 and LEU‐145. In addition, the aliphatic side chain of LEU‐145 engages in hydrophobic contacts with the ligand's aromatic ring, enhancing the ligand–receptor binding affinity.

Figure 9D illustrates that the active compound occupies the target protein's binding pocket via concurrent hydrophilic and hydrophobic contacts and achieves steady binding through massive hydrogen‐bond networks, with bond distances spanning 1.7–3.5 Å. A short hydrogen bond is detected at the ASN‐34 site, while π–π stacking occurs between the ligand aromatic ring and hydrophobic residue PHE‐144. These dual forces synergistically strengthen ligand–receptor binding.

4. Discussion

To systematically analyze and characterize the chemical constituents of Epimedii Folium and screen its bioactive components, we utilized the UHPLC‐Q‐Orbitrap HRMS platform, a rapid and high‐efficiency analytical workflow. Benefiting from its superior chromatographic separation capacity and precise qualitative identification performance, we comprehensively characterized a broad spectrum of chemical constituents from Epimedii Folium. Compound identification was accomplished by matching exact molecular masses, molecular formulas, and fragment ion cleavage patterns acquired via high‐resolution mass spectrometry with chromatographic retention times. Supplementary compound structural information was retrieved from TCMSP database version 2.3 (released in 2014), which provided robust foundational data for the subsequent network pharmacology workflow implemented in this research.

Traditional Chinese medicine (TCM) adopts a holistic view of the human organism and follows a systemic therapeutic logic. For decades, decoding the biological mechanisms underlying TCM pharmacological activities at the molecular level has remained a major analytical challenge. Network pharmacology integrates bioinformatics, systems biology, and computational biology to dissect therapeutic mechanisms across full biological signaling networks. Unlike conventional single‐target research paradigms, which fail to recapitulate TCM's characteristic multi‐target, multi‐pathway regulatory modes, network pharmacology breaks through such analytical bottlenecks. By quantitatively mapping the interactive associations between TCM bioactive molecules and disease‐associated protein targets, this systematic methodology elucidates complex pharmacological mechanisms from a global biological perspective, overcomes the limitations of single‐target screening strategies, and delivers innovative analytical frameworks to scientifically interpret the molecular basis of TCM efficacy (Zhao et al. 2023). In the present work, we combined network pharmacology prediction with molecular docking validation to dissect the anti‐ad molecular mechanisms of Epimedii Folium, establishing a theoretical basis for subsequent mechanistic and translational studies of its anti‐ad therapeutic activity.

In total, 72 distinct chemical constituents were tentatively identified from Epimedii Folium via UHPLC‐Q‐Orbitrap HRMS analysis. Construction and topological analysis of the drug–compound–target–pathway–disease network revealed the top three significantly enriched KEGG pathways, namely Pathways in cancer, the PI3K‐Akt signaling pathway, and lipid and atherosclerosis. GO biological process enrichment further indicated that Epimedii Folium primarily exerts anti‐ad effects by modulating the pathways in cancer, AGE‐RAGE signaling pathway in diabetic complications, and lipid and atherosclerosis signaling axes to ameliorate ad pathological phenotypes. Following screening of core bioactive molecules and hub targets, apigenin, kaempferol, luteolin, and quercetin were selected as ligand candidates, while TP53, AKT1, TNF, and IL6 were defined as core target proteins for molecular docking assays. Docking results confirmed strong binding affinities between these candidate flavonoids and their corresponding hub targets.

Accumulating pharmacological studies demonstrate that kaempferol confers protective effects against inflammatory disorders, including cardiovascular lesions, malignancies, and neurodegenerative diseases, via the suppression of aberrantly activated protein kinases and transcription factors (Ren et al. 2019). Quercetin mediates robust neuroprotection by activating the PI3K/Akt cascade to alleviate neuronal injury and apoptosis; it simultaneously inhibits tau hyperphosphorylation and acetylcholinesterase enzymatic activity. Collectively, these regulatory functions retard ad pathological progression and endow neuroprotective capacity, highlighting quercetin's promising therapeutic potential for neurodegenerative disorders (Jiang et al. 2016).

Multiple in vivo animal studies have validated apigenin's capacity to suppress neurodegeneration and rescue deficits in learning, memory, and neurovascular function. The neuroprotective properties of apigenin have been well documented in prior preclinical models. Notably, one study utilizing a transgenic ad fruit fly model reported that apigenin treatment extended the lifespan of neurodegenerative fly strains (Kramer and Johnson 2024). Luteolin has also been repeatedly verified to reverse impaired spatial learning and memory. Mechanistically, luteolin functions as a potent reactive oxygen species (ROS) scavenger across diverse cell and tissue types, mitigating memory loss and preventing the atrophy of CA1 pyramidal neurons in ad model animals. Its neuroprotective activity arises from the clearance of cytotoxic free radicals and enhanced antioxidant defense against β‐amyloid plaque aggregation, which together contribute to its anti‐ad efficacy (Pu et al. 2018).

Protein–protein interaction (PPI) network topological analysis identified TP53 (39 interaction nodes), AKT1 (29 nodes), TNF (28 nodes), and IL6 (27 nodes) as the central hub targets in our dataset. TP53 serves as a master transcriptional regulator governing glycolytic metabolism and neuronal apoptosis; it triggers activation of the pentose phosphate pathway to boost intracellular nicotinamide adenine dinucleotide phosphate production, ultimately restraining excessive neuroinflammatory cascades (Li et al. 2021). Mounting evidence indicates that ROS‐triggered oxidative modification of AKT1 drives synaptic dysfunction in ad. This oxidative post‐translational modification disrupts synaptic transmission and activity‐dependent protein translation. Consistently, boosting synaptic AKT1‐mTOR signaling mitigates neurodegeneration and cognitive impairment, two hallmark pathological features of ad (Ahmad et al. 2017).

Tumor necrosis factor (TNF), a secreted polypeptide predominantly released by tissue macrophages, participates in a wide array of cellular signaling events, including NF‐κB cascade activation, apoptotic signal transduction, and systemic immune homeostasis modulation. Therapeutic suppression of TNF‐α signaling has exhibited encouraging efficacy for alleviating ad‐associated pathological damage (Wang et al. 2017). Neuroinflammation represents a core pathogenic driver of ad, and interleukin‐6 (IL6) is a pro‐inflammatory cytokine secreted by monocytes. Together with IL1β and IL10, IL6 amplifies neuroinflammatory responses and exacerbates oxidative stress, ultimately triggering neuronal apoptosis and cerebral neuronal atrophy (Gao et al. 2021).

This study bears several inherent limitations. All mechanistic conclusions drawn herein rely solely on computational network pharmacology prediction and molecular docking simulation, without complementary functional validation via in vitro cellular experiments or in vivo animal models. Therefore, the reliability and translational potential of our computational findings require comprehensive experimental verification in follow‐up research.

Future work will prioritize systematic in vitro and in vivo functional validation to confirm the core bioactive compounds of Epimedii Folium and dissect their precise molecular signaling cascades underlying anti‐ad therapeutic effects.

5. Conclusion

This study constructed a reliable rapid UHPLC‐Q‐Orbitrap HRMS method to profile phytochemicals in Epimedii Folium, combined with network pharmacology to preliminarily screen its candidate anti‐ad components and related potential pathways. Seventy‐two compounds were identified, with four flavonoids (quercetin, apigenin, kaempferol, and luteolin) recognized as key bioactive substances, and TP53, AKT1, TNF, and IL6 as core therapeutic targets. Molecular docking validated stable binding between these flavonoids and hub targets. Notably, all regulatory pathways predicted in silico require further biological verification to avoid overinterpreting actual pharmacological mechanisms. This research establishes a compound‐target correlation framework to interpret the multi‐target anti‐ad characteristics of Epimedii Folium, supplying abundant candidate molecules for subsequent cellular and in vivo pharmacodynamic validation. It provides solid methodological and theoretical support for the translational research and neuroprotective drug development of Epimedii Folium, and offers a referable integrated analytical strategy for mechanism exploration of other TCM treating neurodegenerative diseases.

Funding

This work was supported by the National Science Foundation of China (82560886), the Jiangxi Provincial Natural Science Foundation (20252BAC240556), and the Science and the Educational Department of Jiangxi Province of China (GJJ2401823).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.