Deciphering the Potential Mechanism of Cordycepin in Alleviating Ulcerative Colitis via the AKT1 Signaling Pathway: An Integrated Approach Combining Network Pharmacology, Molecular Docking, and Experimental Validation

Abstract

Aim

Given the limited availability of safe and effective treatments for inflammatory bowel disease (IBD), we applied an integrated network pharmacology approach to systematically map the targets and pathways of cordycepin, a bioactive compound from Cordyceps militaris, in experimental colitis.

Methods

Cordycepin was administered intraperitoneally during dextran sulfate sodium (DSS) exposure in mice, with efficacy evaluated by the disease activity index (DAI) and histopathological analysis. Network pharmacology analysis (TCMSP, CTD, SEA, BATMAN-TCM, GeneCards, and PharmMapper), molecular docking, and molecular dynamics (MD) simulations were performed to identify and validate potential core targets. AKT1 and tight junction protein ZO-1 expression in colonic tissues was assessed by immunohistochemistry (IHC). The involvement of AKT signaling in cordycepin’s effects on tight junction integrity and mitochondrial function was further investigated in lipopolysaccharide (LPS)-treated Caco-2 cells using the AKT inhibitor MK2206.

Results

Cordycepin (50 mg/kg) significantly attenuated body weight loss and DAI elevation in DSS-treated mice. A total of 361 putative cordycepin-related targets were identified from six public databases, while 2, 072 UC-related targets were obtained from GeneCards, OMIM, and DisGeNET. A total of 199 overlapping targets were functionally enriched in processes including “TNF signaling pathway”, “PI3K-AKT signaling pathway” and “cellular response to lipopolysaccharide”. The PPI network identified 8 core targets, among which AKT1, NFKB1, RELA and TP53 demonstrated strong binding affinity (binding free energy<-6.0 kcal/mol) with cordycepin in molecular docking and were enriched within the PI3K/AKT pathway. IHC analysis showed that cordycepin reversed alterations of colonic AKT1 and ZO-1 levels in DSS mice. In Caco-2 cells, AKT inhibition with MK2206 attenuated the protective effects on tight junction integrity and mitochondrial function against LPS-induced injury.

Conclusion

These findings suggest that prophylactic administration of cordycepin, a promising natural compound, alleviates experimental colitis, potentially through modulation of the PI3K/AKT1 signaling pathway and restoration of epithelial barrier integrity.

Introduction

Ulcerative colitis (UC), one of the two major forms of inflammatory bowel disease (IBD), is a chronic, relapsing-remitting disorder characterized by uncontrolled immune activation and persistent inflammation of the gastrointestinal tract.¹ The pathogenesis of UC involves complex interactions among genetic susceptibility, immune dysregulation, gut microbiota imbalance, and environmental triggers. These factors collectively lead to mucosal autoimmune dysfunction and clinical manifestations such as abdominal pain, chronic diarrhea, urgency to defecate, weight loss and systemic complications.² Despite advances in treatment, current pharmaceutical strategies against UC, including aminosalicylates, corticosteroids, immunomodulators, and biologics (eg, anti-TNF agents), remain limited by variable efficacy, high relapse rates, and significant side effects, such as infections, metabolic disturbances, and increased cancer risk.^3,4 Furthermore, latest evidence highlights that mitochondrial dysfunction, characterized by impaired energy metabolism, oxidative stress, and epithelial apoptosis, plays a critical role in the pathogenesis and perpetuation of IBD, underscoring the importance of therapeutic strategies that target not only immune signaling but also mitochondrial function and epithelial homeostasis.⁵ There is therefore an urgent need to develop novel pharmaceutical options for UC management.

Cordycepin (3’-deoxyadenosine), a bioactive nucleoside derived from the fungus Cordyceps militaris, has emerged as a promising candidate for mitigating colon inflammation, with growing interest in its potential applications in IBD and colitis-associated colorectal cancer (CRC).^6,7 The multifaceted pharmacological properties of cordycepin, including its anti-inflammatory, immunomodulatory, and anti-tumor activities, make it a uniquely versatile agent for targeting the interconnected pathogenic processes underlying these conditions.⁸

A central mechanism driving cordycepin’s therapeutic efficacy lies in its modulation of the immune response, a critical axis in IBD pathogenesis where immune dysregulation fuels persistent intestinal inflammation.^9,10 In a preclinical study of dextran sulfate sodium (DSS)-induced colitis, cordycepin was shown to dose-dependently reduce the disease activity index and attenuate colonic epithelial injury. This protective effect was attributed to its ability to rebalance the pro-inflammatory (Th1/Th17) and anti-inflammatory (Th2/Treg) immune axes, alongside improvements in gut microbial community structure.⁶ Notably, cordycepin exerts further anti-inflammatory effects by inhibiting macrophage pyroptosis, a pro-inflammatory form of programmed cell death, via direct targeting of the NLRP3 inflammasome/Caspase-1/GSDMD signaling pathway in RAW264.7 macrophages, highlighting its capacity to interfere with key inflammatory cascades.¹¹ Moreover, cordycepin strengthens intestinal epithelial barrier function by upregulating tight junction proteins and reshaping the gut microbial composition, specifically by promoting the growth of the beneficial bacterium Akkermansia muciniphila, in a Western diet-induced model of metabolic inflammation and obesity.¹² Beyond IBD, cordycepin exhibits robust anti-tumor activity against CRC, a major sequela of long-standing colitis, by regulating the proliferation and apoptosis of CRC cells and promoting the anti-tumor immune surveillance and reversing the immunosuppressive tumor microenvironment (TME).^13–15 Although growing evidence indicates that cordycepin significantly modulates intestinal immunity and epithelial biological function, both of which play central roles in UC pathogenesis, the precise targets and relevant molecular mechanisms by which cordycepin attenuates intestinal inflammation remain unclear.

Network pharmacology, an emerging interdisciplinary field at the intersection of systems biology, bioinformatics, and computational pharmacology, offers a holistic framework to dissect the complex interplay between bioactive compounds, biological systems, and disease processes.¹⁶ In contrast to traditional pharmacology, which focuses on single-target drug actions, the network pharmacology prioritizes the characterization of “multi-target-multi-pathway” interaction networks, leveraging computational tools (eg, database mining, molecular docking, and pathway enrichment analysis) to map the pharmacological profiles of natural products or synthetic agents.¹⁷

By systematically identifying potential targets through database screening, predicting molecular interactions via docking simulations, and reconstructing regulatory networks using pathway enrichment analysis, network pharmacology excels at unraveling layers of biological complexity. This integrative strategy not only overcomes the limitations of single-target approaches but also provides a mechanistic framework to interpret how bioactive components modulate multi-level biological systems. Such capabilities make it uniquely suited for investigating the therapeutic mechanisms of compounds in contexts where disease pathology or drug action involves multiple interconnected molecular and cellular processes, thereby bridging the gap between pharmacological observations and the underlying biological networks.¹⁸

Given that cordycepin represents a promising natural compound for preventing and ameliorating intestinal inflammation, further investigation into its mechanisms and therapeutic applications is warranted. In the present study, we employed an integrated approach combining network pharmacology, a murine model of colitis and in vitro Caco-2 cell model of colonic inflammation were employed to investigate and validate the potential targets and mechanisms of cordycepin attenuating IBD (Figure 1).

Figure 1.

The flowchart of investigating the effects and mechanisms of cordycepin in UC.

Materials and Methods

Induction of Experimental Colitis and Cordycepin Treatment

All male C57BL/6 mice (age: 6–8 weeks, body weight: 18–20 g) purchased from Cavens Model Animal Co., Ltd. (Changzhou, China, License No. SCXK (Su) 2021–0013) and housed in Jiangsu Kerbio Medical Technology Group Co., Ltd. (Changzhou, China, License No. SYXK (Su) 2021–0013). This study adopted a DSS-induced experimental colitis model in mice. The modeling methods, sample sizes, and commonly used groups were based on our previous research and slightly modified.¹⁹ After the mice adapted to the environment, they were randomly assigned to the control group, the DSS group, the DSS+Cor_L group, the DSS+Cor_M group and the DSS+Cor_H group. Random grouping is carried out using computer-generated random sequences. Briefly, mice in the DSS group received 3.0% DSS (w/v; MW: 36,000–50,000, MP Biochemicals; dissolved in autoclaved drinking water, ad libitum) for 7 consecutive days (day 0 to day 7) to induce experimental colitis. Mice in the DSS+Cor groups were administered cordycepin at three distinct doses (10, 50, and 200 mg/kg/day, dissolved in saline, i.p.) daily during DSS exposure, corresponding to the low- (DSS+Cor_L), moderate- (DSS+Cor_M), and high-dose (DSS+Cor_H) groups, respectively. Intraperitoneal administration was deliberately chosen over the oral route to minimize confounding from gut microbiota-related mechanisms, given that we have previously demonstrated that cordycepin modulates intestinal microbial community structure.⁶ This approach is also more consistent with the network pharmacology paradigm, which predicts direct molecular bindings and interactions between the small molecular compound and its targets. Cordycepin (purity>98%) used in this study was provided by Nanjing University of Technology. Mice in both the Control and DSS groups received an equivalent volume of saline (i.p.) daily. General health status, including body weight and stool characteristics, was monitored daily. The disease activity index (DAI) was calculated as the sum of three parameters: body weight loss (<1%: 0; 1–5%: 1; 6–10%: 2; 11–18%: 3; >18%: 4), stool consistency (normal: 0; soft: 1; unformed: 2; loose stool: 3; watery diarrhea: 4), and fecal occult blood (none: 0; mild: 1; severe: 2; bloody stool: 3; rectal bleeding: 4). The total DAI score represented the sum of these three individual scores. At the experimental endpoint (day 7), mice were anesthetized via intraperitoneal injection of Zoletil^® 50 (60 mg/kg; Virbac, Carros, France). Subsequently, mice were sacrificed by cervical dislocation. Colon tissues were collected, measured for length, and fixed in 4% paraformaldehyde for subsequent analysis. Investigators responsible for DAI scoring, colon length measurement, histological evaluation, and data analysis were blinded to group allocation whenever feasible. Predefined exclusion criteria included death before the experimental endpoint, development of severe unrelated illness, or technical failure during treatment administration or sample collection. Unless otherwise specified, all animals were included in the final analysis. All animal procedures were approved by the Ethics Committee of Changzhou Children’s Hospital (Approval No: 2023–011).

Histomorphological Examination

The paraformaldehyde-fixed mouse colon tissues were dehydrated through a graded ethanol series, cleared in xylene, and embedded in paraffin, and sectioned continuously at a thickness of 4 μm. Sections were floated on a constant-temperature water bath, dried, and two randomly selected sections from each sample were subjected to hematoxylin and eosin (H&E) staining.

Immunohistochemistry (IHC)

Paraffin-embedded colon sections of mice were deparaffinized, rehydrated, and underwent antigen retrieval. Endogenous peroxidase activity was neutralized with 3% H₂O₂, followed by a blocking step with 5% bovine serum albumin (BSA). Sections were incubated overnight at 4°C with primary antibodies against pan-AKT (Cell signaling technology, 4685, 1:200) or AKT1 (Abcam, ab81283, 1:100) or zonula occludens-1 (ZO-1) (Proteintech, 25055-1-AP, 1:100). After washing with PBS, sections were incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature. Immunoreactivity was visualized using a diaminobenzidine (DAB) substrate kit. Ultimately, the sections were counterstained with hematoxylin, dehydrated, mounted, and imaged under a light microscope (Olympus, Japan).

Identifying the Targets of Cordycepin

To enhance target coverage and minimize bias associated with reliance on a single database, biological targets associated with cordycepin were systematically integrated from six complementary resources: the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP), the Comparative Toxicogenomics Database (CTD), the Similarity Ensemble Approach (SEA), GeneCards (http://www.genecards.org/), BATMAN-TCM (with a score threshold>0.84), and PharmMapper (with a standardized fit score>0.9). TCMSP and BATMAN-TCM use structural similarity and pharmacological activity scoring based on traditional medicine compound databases; CTD provides experimentally curated chemical-gene interaction data; SEA predicts targets based on ligand set ensemble similarity; PharmMapper employs reverse pharmacophore mapping; and GeneCards integrates multi-source genomic and pharmacological annotation data to retrieve compound-associated gene targets based on curated literature evidence. The species was limited to “Homo sapiens”. Candidate targets retrieved from six databases were uniformly converted to official human gene symbols using the UniProt database (https://www.uniprot.org/), followed by deduplication, achieved by removing redundant entries in Microsoft Excel 2021 (Figure S1).

Identifying the Underlying Targets of UC

Targets associated with UC were obtained from the GeneCards, the Online Mendelian Inheritance in Man (OMIM, http://OMIM.org/), and the DisGeNET (http://www.disgenet.org/) databases. The species was limited to “Homo sapiens”. The search utilized “Colitis, Ulcerative” as the keyword, and targets were subsequently screened. Candidate targets were standardized within the UniProt database. Subsequently, all targets were consolidated, duplicates were eliminated. The overlapping targets associated with cordycepin and UC were identified and visualized via a Venn diagram using the OEBiotech online platform (https://cloud.oebiotech.cn/#/bio/tools, Figure S1).

Protein-Protein Interaction (PPI) Network Analysis

The identified overlapping targets were imported into the STRING database (https://string-db.org), with the species set to “Homo sapiens”. Interactions with a confidence score>0.9 were selected, and unconnected nodes were hidden. The resulting data were exported in TSV format and imported into Cytoscape (version 3.10.3) for topological analysis. The CytoHubba plugin was utilized to identify the top 15 targets based on three parameters: Degree centrality, Closeness centrality, and Betweenness centrality. The intersection of these top 15 targets across all three algorithms was defined as the core hub targets.

Functional Enrichment Analysis

The DAVID database (https://davidbioinformatics.nih.gov/) was utilized for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analysis. A significance threshold of p<0.05 was established. The top 10 GO terms and the top 30 KEGG pathways based on enrichment scores were visualized using bubble charts generated on the bioinformatics platform (https://www.bioinformatics.com.cn/).

Construction of the “Drug-Disease-Target-Pathway” Network

Cordycepin, UC, overlapping targets and KEGG pathways (top 10) were imported into Cytoscape 3.10.3. The topological properties of the Network were calculated using the Network Analyzer tool within the software. Following adjustments, a multi-dimensional regulatory relationship diagram was constructed to illustrate the “drug-disease-target-pathway” interactions.

Molecular Docking

The three-dimensional structure file of cordycepin in mol2 format was sourced from TCMSP database. The small molecule underwent full hydrogen addition using AutoDock 1.5.6 software, was designated as a ligand, and subsequently saved in pdbqt format. The three-dimensional structure file of the core targets in pdb format was retrieved from the Protein Data Bank (PDB, http://www.rcsb.org) and preprocessed with Pymol 3.03 and AutoDock 1.5.6 software to eliminate water molecules, remove any original ligands, add complete hydrogen atoms, and designate it as the receptor before saving it in pdbqt format. Following this preparation, molecular docking between protein molecules and drug ligands was conducted using AutoDock 1.5.6 software to obtain binding energies. In this study, binding energies<-5 kcal/mol were considered indicative of strong binding affinity. Interactions were visualized using PyMOL 3.03.

Molecular Dynamics (MD) Simulations

To evaluate the stability of the core compound-target complex exhibiting the highest binding affinity, MD simulations were performed. On a Linux platform, GROMACS (https://www.gromacs.org/) was employed to carry out 100-ns, all-atom MD simulations of the selected complex. The protein was parameterized using the AMBER99SB-ILDN force field, while ligand parameters were generated with the Sobtop tool (Tian Lu, Sobtop, Version 1.0 dev5, http://sobereva.com/soft/Sobtop). The entire system was solvated using the TIP3P water model; an ionic strength of 10 Å was applied as the cutoff distance for electrostatic interactions. The simulation box was constructed as a cubic solvation cell, and Na⁺ and Cl⁻ ions were added to neutralize the system’s net charge. Energy minimization was conducted using the steepest descent and conjugate gradient algorithms. The system was then equilibrated under the NVT ensemble at 300 K for 100 ps, followed by NPT equilibration at 1 bar for 100 ps. Finally, a 100 ns production MD simulation was carried out. The resulting trajectories were analyzed and visualized using DuIvyTools (https://duivytools.readthedocs.io/en/v0.6.0/DIT.html), including root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), hydrogen bond analysis, solvent-accessible surface area, and free energy landscape (FEL).

Caco-2 Cell Culture and in vitro Inflammation Model

The human colon adenocarcinoma cell line Caco-2 was originally obtained from Abbkine Scientific Co., Ltd (Wuhan, China). Cells were cultured in 60-mm dishes (NEST) using DMEM medium (Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomycin. Cells were maintained at 37°C with 5% CO₂, with medium changes every 2–3 days. At 80% confluence, cells were washed with PBS and passaged using 0.25% trypsin (Beyotime). Experimental groups were established: (1) Control (<1‰ DMSO); (2) LPS (10 μg/mL); (3) LPS+Cor (10 μM); (4) LPS+Cor+MK2206. For the inflammation model, confluent Caco-2 cells were stimulated with 10 μg/mL LPS in serum-free (SF) medium for 48 h. Cordycepin (10 μM) and MK2206 (1 μM, Aladdin) were co-administered during LPS stimulation to investigate the role of AKT signaling.

Cell Viability Assay (MTT)

Following treatment, medium was aspirated, and 20 μL MTT solution (5 mg/mL in SF medium) was added to each well. Cells were incubated at 37°C for 4 h, after which the formazan crystals were dissolved in 150 μL of DMSO. Absorbance at 490 nm was measured using a microplate reader to calculate cell viability.

ROS Detection

Intracellular ROS levels were detected using the DCFH-DA probe (Beyotime). Cells were washed with PBS, incubated with DCFH-DA for 20 min at 37°C. Fluorescence images were captured using a Zeiss fluorescence microscope.

Mitochondrial Membrane Potential Assay

The TMRE fluorescent probe (Beyotime) was employed to evaluate the mitochondrial membrane potential. Cells were incubated with TMRE for 40 min at 37°C in the dark, and counterstained with Hoechst 33342. Red fluorescence intensity was quantified using ImageJ.

Wound-Healing Assay

Caco-2 cells were seeded in 6-well plates and grown to 90% confluence. A linear wound was created using a sterile 200 μL pipette tip. Cells were then cultured in SF medium with or without cordycepin. Wound closure was monitored, and images were captured at 0, 6, 24, and 48 h. The closure rate was calculated using ImageJ: Closure%=[(A₀-AT)/A₀]×100%, where A₀ represents the average scratch length at 0 hours and A_t represents the average scratch length after T hours of culture.

Statistical Analysis

The data are presented as mean±SD. For normally distributed data, one-way ANOVA with Tukey’s post-hoc test was used for multiple comparisons. For non-normally distributed data, the Kruskal–Wallis H-test with Dunn’s post-hoc test was used. All statistical analyses were conducted by GraphPad Prism 8.0.2 (GraphPad Software, USA). A two-sided p-value<0.05 was considered statistically significant.

Results

The Effects of Cordycepin on DSS-Induced Experimental Colitis in Mice

To evaluate the therapeutic effect of cordycepin on UC, a mouse model of UC was established by administering 3% DSS in drinking water for 7 consecutive days (day 0-day 7). Mice in the cordycepin treatment group received intraperitoneal injections of cordycepin (10, 50, 200 mg/kg/day) concomitantly during the DSS induction period (Figure 2A). On day 7, compared with the control group, mice in the DSS group exhibited significant body weight loss, a marked decrease in colon length and an increase in spleen weight. The DAI score in DSS group reached 9.0±2.14. Notably, administration of cordycepin at moderate dose (50 mg/kg/day, i.p.) markedly reversed these pathological changes in DSS mice: body weight loss was reduced (Figure 2B and C); DAI score was decreased (Figure 2D); the colon length was extended (Figure 2E); and the spleen weight declined (Figure 2F).

Figure 2.

Cordycepin ameliorates disease severity in a murine model of UC. (A) Schematic diagram of the experimental design for UC induction and cordycepin treatment (L: low dose, 10 mg/kg/day; M: moderate dose, 50 mg/kg/day; H: high dose, 200 mg/kg/day; all i.p.). (B) Body weight over the course of the experiment. (C) Percentage of body weight loss at the end of the experiment (day 7). (D) DAI scores. (E) Colon length. (F) Spleen weight. (G) Representative hematoxylin and eosin (H&E) stained colon sections (100×). Control group: n=7; DSS group: n=8; DSS+Cordycepin groups: n=7 per group. Data are presented as mean±SD. Statistical significance was determined by Kruskal–Wallis test followed by Dunn’s post hoc test for multiple comparisons. *p<0.05, **p<0.01, ***p<0.001.

Histopathological examination of H&E-stained colon tissues from day 7 revealed that, compared to the control group, colon tissues from DSS group exhibited significant ulcer-like pathological alterations characterized by severe intestinal mucosal destruction, including epithelial cell necrosis, goblet cell loss, and massive inflammatory cell infiltration. These pathological features indicate that DSS induced acute injury and inflammatory responses in the mouse colon, consistent with the typical histopathological manifestations of clinical UC. Cordycepin administration significantly restored the structural integrity and reduced inflammatory infiltration in colon tissues of DSS-treated mice (Figure 2G). These findings support the role of cordycepin as an effective anti-inflammatory agent in experimental UC.

Identification of Overlapping Targets Between Cordycepin and UC

Based on the databases of TCMSP, CTD, SEA, Genecards, BATMAN-TCM and Pharmapper, 361 targets related to cordycepin were obtained after summarization and removal of duplicate targets (Figure 3A, Table S1). By conducting a thorough search of the OMIM, GeneCards, and DisGenet databases, followed by systematic screening, merging, and deduplication processes, we identified a total of 2, 072 targets that are closely associated with UC (Table S2). A total of 199 overlapping targets of cordycepin targets and UC targets were obtained (Figure 3B, Table S3) and subsequently subjected to a protein-protein interaction (PPI) network analysis to elucidate their functional interconnectivity (Figure 3C).

Figure 3.

Identification of cordycepin targets and functional enrichment analysis in UC. (A) UpSet plot illustrating the number of putative cordycepin targets identified from six databases. (B) Venn diagram showing the overlapping targets between cordycepin and UC. (C) PPI network of the overlapping targets. Node size and color intensity represent the degree of connectivity. (D) GO enrichment analysis of the overlapping targets presented as a bubble chart. Terms are categorized into biological process (BP), cellular component (CC), and molecular function (MF). (E) KEGG pathway enrichment analysis of the overlapping targets. (F) An integrated “compound-target-pathway-disease” network. Nodes are color-coded: Orange for cordycepin, green for targets, purple for pathways, and blue for UC.

GO and KEGG Enrichment Analysis of Overlapping Targets

The functional annotation of 199 overlapping targets based on GO and KEGG enrichment analysis were performed using the DAVID database. The top 10 items from the GO analysis were selected to create enrichment bubble plots (Figure 3D). Among these, biological processes (BP) primarily include positive regulation of gene expression, response to xenobiotic stimulus, negative regulation of apoptotic process, inflammatory response, cellular response to lipopolysaccharide, negative regulation of gene expression, positive regulation of cell migration, positive regulation of transcription by RNA polymerase II, and positive regulation of interleukin-6 production. In terms of cellular components (CC), key categories encompass the extracellular region, cytosol, cytoplasm, and extracellular space. Molecular functions (MF) predominantly involve identical protein binding, protein binding in general, enzyme binding activities as well as growth factor activity. The enrichment bubble chart was generated based on the results from the top 30 KEGG pathways (Figure 3E). These pathways mainly included pathways in cancer, lipid metabolism and atherosclerosis, AGE-RAGE signaling pathway in diabetic complications, TNF signaling pathway, IL-17 signaling pathway, and PI3K-AKT signaling pathway among others.

Construction of the “Drug-Disease-Target-Pathway” Network

The network files and attribute files, which contain information on cordycepin, UC, overlapping targets, and the top 10 KEGG pathways, were imported into Cytoscape version 3.10.3 to construct the “compound-target-disease” interaction network of cordycepin in UC (Figure 3F). This visualization illustrates the interrelationships among cordycepin, overlapping targets, diseases, and signaling pathways.

Construction of PPI Networks for Overlapping Targets and Screening of Core Targets

The 199 overlapping targets identified were imported into the STRING database for analysis, resulting in a PPI network graph. Subsequently, visualization was performed using Cytoscape 3.10.3. The top 15 targets from each dataset were selected based on three key topological parameters provided by the CytoHubba plugin: Degree, Closeness, and Betweenness. By taking the intersection of these selections, a total of 8 core hub targets were identified: interleukin6 (IL6), tumor necrosis factor (TNF), protein kinase B (AKT1), JUN proto-oncogene (JUN), tumor suppressor gene p53 (TP53), nuclear factor κB p65 (RELA), nuclear factor κB subunit 1 (NFKB1), and catenin β1 (CTNNB1) (Figure 4A). They were subsequently functionally enriched into biological processes including “salmonella infection”, “toll-like receptor signaling pathway”, “TNF signaling pathway” and “apoptosis” in KEGG analysis and into pathways including “regulation of apoptotic process”, “inflammatory response”, “positive regulation of leukocyte adhesion to vascular endothelial cell” and “NF-kappaB p50/p65 complex” in the GO analysis (Figure 4B and C).

Figure 4.

The screening and functional enrichment for the core targets of cordycepin in UC. (A) Screening of core hub targets from 199 overlapping targets using PPI network analysis. (B) Functional annotation of the core hub targets based on KEGG (left) and GO (right) analysis. (C) Core hub targets enriched in the “TNF signaling pathway” (left) and “toll-like receptor signaling pathway” (right) via KEGG enrichment.

Molecular Docking of Cordycepin with Core Targets

Molecular docking verification of cordycepin with core targets (IL6, TNF, AKT1, JUN, TP53, RELA, NFKB1, and CTNNB1) was performed using AutoDock 1.5.6 software. Model visualization was conducted with PyMOL 3.03 software, generating molecular docking diagrams that illustrate the interactions between cordycepin and the aforementioned core hub targets (Figure 5A–H). Docked energies between the drug small molecule (cordycepin) and each target protein were subsequently calculated. Among the 8 core targets, 4 (AKT1, TP53, RELA, and NFKB1) exhibited binding energies ≤ −6 kcal/mol, indicating a high potential for stable binding (Figure 5I). Notably, these four targets (AKT1, TP53, RELA, and NFKB1) were all functionally annotated into the PI3K-AKT signaling pathway in the KEGG enrichment analysis (Figure 5I).

Figure 5.

Molecular docking of cordycepin with core targets and pathway mapping. (A–H) Three-dimensional molecular docking models depicting the binding poses and interactions between cordycepin and the core protein targets: (A) TNF, (B) AKT1, (C) TP53, (D) RELA, (E) NFKB1, (F) CTNNB1, (G) IL-6, (H) JUN. The docked energy for each complex is indicated. Cordycepin is represented as green sticks. Hydrogen bonds and other key interactions are depicted as yellow dashed lines. (I) KEGG pathway analysis highlights the significant enrichment of the core targets (including TP53, AKT1, and NF-κB subunits) in the PI3K-AKT signaling pathway.

Cordycepin Stably Binds to the Active Pocket of AKT1

MD simulations were conducted to evaluate the stability of the AKT1-cordycepin complex. The root mean square deviation (RMSD) analysis showed that the complex reached equilibrium after approximately 20 ns and remained stable around 0.2 nm throughout the simulation period, indicating a stable binding conformation (Figure 6A). The root mean square fluctuation (RMSF) analysis revealed that most residues exhibited low fluctuations (0.1–0.3 nm), indicating structural rigidity in the core regions, while higher fluctuations were observed in terminal and loop regions, reflecting intrinsic protein flexibility (Figure 6B). The radius of gyration (Rg) remained stable at approximately 2.18 nm, suggesting that the overall compactness of the protein structure was well maintained without significant conformational changes (Figure 6C). The free energy landscape (FEL) revealed a well-defined global minimum energy basin, indicating that the AKT1-cordycepin complex converged into a stable low-energy conformation. The presence of a dominant energy basin suggests a favorable and persistent binding mode with limited conformational variability (Figure 6D). Collectively, these results demonstrate that cordycepin can stably bind to the active pocket of AKT1 and maintain structural stability throughout the simulation, supporting its potential as a functional ligand.

Figure 6.

Molecular dynamics simulation reveals the structural stability and binding characteristics of the AKT1-cordycepin complex. (A) Root mean square deviation (RMSD). (B) Root mean square fluctuation (RMSF) of AKT1 residues. (C) Radius of gyration (Rg). (D) Free energy landscape (FEL).

The Expression of AKT and Tight Junction Proteins in Cordycepin-Treated Experimental Colitis

Given that the moderate dose demonstrated robust efficacy comparable to the high dose but with a lower drug load, the 50 mg/kg/day group was selected for mechanistic validation. Our IHC examination showed that in contrast to pan-AKT, the expression pattern of the AKT isoform AKT1 showed a distinct trend (Figure 7A–D). Compared with the Control group, the DSS model group exhibited a significant reduction in colonic AKT1 expression, suggesting that DSS-induced intestinal inflammation may specifically downregulate the expression of the AKT1 isoform (Figure 7C and D). Notably, cordycepin treatment at the dose of 50 mg/kg/day significantly reversed the DSS-induced downregulation of AKT1. IHC staining of ZO-1, a key tight junction protein, showed that in the DSS group, ZO-1 expression was significantly downregulated (discontinuous and weak staining of ZO-1 at the intestinal epithelial tight junction; Figure 7E and F). In response to cordycepin treatment, the ZO-1 signals at the epithelial tight junction were restored, indicating that cordycepin effectively mitigates DSS-induced tight junction damage.

Figure 7.

Cordycepin (50 mg/kg/day) modulates the AKT signaling pathway and restores the tight junction protein ZO-1 in UC mice. (A, C and E) Representative IHC images of (A) pan-AKT (200×), (C) AKT1 (200×), and (E) ZO-1 (400×) in colon tissues. (B, D and F) Quantitative analysis of the average optical density (AOD) for (B) pan-AKT, (D) AKT1, and (F) ZO-1 from corresponding images. Data are presented as mean±SD (n=5). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. *p<0.05, **p<0.01.

Effects of Cordycepin on LPS-Induced Cytotoxicity, Redox Shift, and Mitochondrial Dysfunction in Caco-2 Epithelial Cells via the AKT Signaling Pathway

Treatment with LPS for 48 h significantly reduced cell viability compared to the control (p<0.001). Notably, cordycepin co-treatment markedly restored cell viability, suggesting that cordycepin effectively counteracts LPS-induced cytotoxicity and promotes cell survival (p<0.05) (Figure 8A). However, the addition of MK2206, an allosteric AKT inhibitor, completely reversed this protective effect (p<0.01), resulting in a viability level comparable to that of the LPS-only group. This indicates that the activation of AKT signaling is indispensable for cordycepin’s pro-survival action.

Figure 8.

Cordycepin protects against LPS-induced cytotoxicity, oxidative stress and mitochondrial dysfunction in Caco-2 cells via the AKT pathway. (A) Cell viability was assessed by MTT assay in LPS-stimulated Caco-2 cells treated with cordycepin with or without the AKT inhibitor MK2206. (B) Intracellular reactive oxygen species (ROS) levels were detected by the DCFH-DA fluorescent probe under the same treatment conditions and quantified by mean fluorescence intensity. (C) Mitochondrial membrane potential (ΔΨm) was measured using the TMRE fluorescent probe under the same treatment conditions. Caco-2 cells were stimulated with 10 μg/mL LPS in serum-free (SF) medium for 48 h. Cordycepin (10 μM) and MK2206 (1 μM) were co-administered during LPS stimulation (n=5). Statistical significance was determined by ANOVA test followed by Tukey’s post hoc test. *p<0.05, **p<0.01, ***p<0.001.

Cell viability is closely associated with mitochondrial function: mitochondrial dysfunction induces the accumulation of oxidative stress, and the excessive activation of oxidative stress further impairs mitochondrial structure and function, forming a vicious cycle that ultimately compromises cell survival.²⁰ In this study, intracellular ROS level measured by the DCFH-DA fluorescent probe showed that LPS stimulation significantly increased ROS generation (p<0.001), indicating a redox shift toward oxidative stress. Cordycepin treatment effectively attenuated this LPS-induced ROS accumulation (p<0.001), while MK2206 co-treatment abolished the antioxidant effect of cordycepin (p<0.001) (Figure 8B). We further evaluated mitochondrial membrane potential (ΔΨm) using TMRE fluorescence. LPS treatment caused a pronounced loss of ΔΨm (p<0.001), reflecting mitochondrial dysfunction. In contrast, cordycepin co-treatment largely restored ΔΨm (p<0.001), an effect that was again reversed by MK2206 (p<0.001) (Figure 8C). These data imply that cordycepin facilitates the recovery of mitochondrial function through activation of the AKT pathway under LPS-induced stress. These findings demonstrate that cordycepin mitigates LPS-induced oxidative stress, prevents ΔΨm loss, protects against LPS-induced cytotoxicity, effects that are dependent on the AKT signaling pathway.

Cordycepin Promoted Wound Healing and Barrier Integrity in LPS-Stimulated Caco-2 Cells via AKT Signaling

The scratch wound assay showed that LPS significantly reduced scratch closure rates at 48 h (p<0.001). The improved scratch closure rates were observed in LPS+Cor group compared with LPS only group (p<0.01), an effect that was attenuated by MK2206 (p<0.05) (Figure 9A). ZO-1 immunofluorescence staining of Caco-2 monolayers showed that in the control group, ZO-1 proteins were continuously and regularly distributed on the cell membrane. LPS led to a disruption of ZO-1 expression on the cell membrane (p<0.001), which was restored in response to cordycepin treatment (p<0.001) (Figure 9B). This restoration of tight junction was inhibited by MK2206 (p<0.001).

Figure 9.

Cordycepin enhances epithelial wound healing and tight junction integrity via AKT signaling in LPS-stimulated Caco-2 cells. (A) Cell migration was evaluated by a scratch wound assay. (B) Tight junction integrity was assessed by immunocytochemistry for the ZO-1 protein. During the 48-hour stimulation of Caco-2 cells with 10 μg/mL LPS in SF medium, cordycepin (10 μM) and MK2206 (1 μM) were simultaneously administered (n=5). Statistical significance was determined by ANOVA test followed by Tukey’s post hoc test. *p<0.05, **p<0.01, ***p<0.001.

These findings suggest that cordycepin enhances cell repair and improves barrier integrity in LPS challenged colonic epithelial cells via AKT signaling, a process specifically blocked by MK2206.

Discussion

Due to the biological complexity, significant side effects, and limited disease-modifying effects of existing options, the need for effective management of colonic inflammation remains unmet.²¹ Consequently, the development of novel anti-inflammatory therapeutics is urgently needed. In this study, we employed an integrated approach combining network pharmacology, molecular docking, and experimental validation to systematically elucidate the protective effects and underlying mechanisms of cordycepin against UC. Our findings demonstrate that prophylactic administration of cordycepin significantly alleviates DSS-induced colitis in mice, and its protective action is closely associated with the regulation of the PI3K/AKT1 signaling pathway and the enhancement of intestinal epithelial barrier integrity.

Our in vivo experiments unequivocally confirmed the efficacy of cordycepin in ameliorating experimental colitis. Cordycepin treatment markedly improved the disease phenotype, as evidenced by attenuated body weight loss, reduced DAI scores, restored colon length, and diminished splenomegaly. Histopathological examination further revealed that cordycepin preserved colonic architecture, reduced inflammatory cell infiltration, and crucially, promoted the restoration of goblet cells.

Naturally derived compounds possess the capacity for multi-targeting, enabling the simultaneous modulation of several signaling pathways.²² The network pharmacology and molecular docking are advancing drug discovery by enabling the prediction of multi-target interactions and deconvolution of complex mechanisms, which is fundamental to developing modern therapeutics.¹⁸ By identifying 199 overlapping targets between cordycepin and UC, we revealed that cordycepin likely exerts its effects through a “multi-target, multi-pathway” mode of action. The enrichment analyses (GO and KEGG) of these overlapped targets strongly suggested that they are critically involved in key pathological processes of UC, including the inflammatory response (eg, TNF and IL-17 signaling pathways), apoptosis, and cellular responses to lipopolysaccharide. The convergence of these pathways highlights the potential of cordycepin to simultaneously modulate multiple facets of UC pathogenesis. Furthermore, PPI network analysis of the overlapping targets identified 8 core hubs. Among these, 4 targets (AKT1, TP53, RELA, and NFKB1) exhibited strong binding potential with cordycepin in molecular docking, providing a structural basis for these interactions and suggesting the formation of stable complexes. Notably, AKT1, TP53, RELA, and NFKB1 were all enriched in the PI3K/AKT signaling pathway, directing our subsequent experimental validation towards this critical axis. Since molecular docking revealed strong binding affinity between AKT1 and cordycepin, a 100-ns MD simulation was performed to further assess the stability of the predicted AKT1-cordycepin complex. The simulation results demonstrated that the system reached equilibrium, as evidenced by RMSD and Rg profiles over time. Moreover, the free energy landscape exhibited a predominant low-energy basin, indicating conformational stability of the complex. These findings collectively support the formation of a stable AKT1-cordycepin complex, thereby validating the docking predictions and providing a mechanistic foundation for prioritizing AKT1-mediated signaling pathways in subsequent functional analyses.

The PI3K/AKT pathway is a central signaling axis that governs epithelial survival, proliferation, metabolism, migration, and barrier function.²³ Its balanced activity is essential for epithelial health, while its dysregulation is a pathogenic feature in diverse diseases, ranging from barrier dysfunction in IBD to hyper-proliferation in carcinoma.²⁴ Notably, our earlier bioinformatic analyses predicted this potential role, as the overlapping targets of UC and cordycepin were functionally enriched in “inflammatory response” and “cellular response to lipopolysaccharide”, while the core hub targets were implicated in the “toll-like receptor signaling pathway”. Our in vitro data showed that cordycepin significantly mitigated the cellular damage induced by LPS (lipopolysaccharide, an endotoxin from Gram-negative bacteria and a canonical TLR4 agonist), as evidenced by restored cell viability and mitochondrial membrane potential, attenuated redox imbalance, and enhanced tight junction integrity in Caco-2 cells. In addition, cordycepin facilitated cell migration in wound healing assay. These protective effects were abolished by MK2206, a pan-AKT inhibitor with the highest affinity for the AKT1 subtype, suggesting that AKT signaling is involved in the protective effects of cordycepin. Our immunohistochemistry examination of the colon tissues from the mice with DSS-induced colitis provided a crucial link between the computational predictions and the phenotypic improvements. We found that while total AKT expression showed a complex pattern, the specific isoform AKT1 was significantly downregulated in DSS-induced colitis, a suppression that was reversed by cordycepin. Concurrently, cordycepin treatment robustly enhanced the expression and membrane localization of the tight junction protein ZO-1, indicating a strengthened intestinal epithelial barrier. This functional restoration of the barrier aligns with the initial GO enrichment of core targets in processes such as “regulation of apoptotic process” and “inflammatory response”, demonstrating a coherent mechanism from network prediction to phenotypic protection.

It is worth noting that prior mechanistic studies on cordycepin in intestinal inflammation have primarily focused on the suppression of NF-κB signaling, the modulation of gut microbiota composition and associated metabolites, and the regulation of intestinal immune cell homeostasis.^6,25,26 While these are well-established drivers of mucosal inflammation, they do not fully account for epithelial barrier dysfunction and impaired mucosal repair, which are equally central to UC pathogenesis. In contrast, the AKT1 signaling axis, emerging from our unbiased network pharmacology screen, governs not only inflammatory signaling but also epithelial survival, proliferation, and tight junction integrity, representing a complementary and previously underexplored dimension of cordycepin’s therapeutic action. The observation that cordycepin restored AKT1 expression suggests a nuanced molecular mechanism within the broader PI3K/AKT pathway, particularly given the divergent roles of AKT isoforms in intestinal integrity.²⁷ Research indicates that the balance between AKT1 and AKT2 critically regulates epithelial barrier function and mucosal healing. Specifically, in DSS-induced colitis, these isoforms exert opposing regulatory effects: AKT1 inhibition exacerbates colitis and disorganizes tight junctions, whereas targeted AKT2 inhibition significantly ameliorates disease severity by promoting repair and enhancing the epithelial barrier.²⁸ This contrasting effect has been associated with the differential regulation of key repair pathways, such as the WNT/β-Catenin signaling, which is essential for tissue regeneration.²⁹ Therefore, the increased AKT1 expression observed after cordycepin treatment is consistent with a protective mechanism involving epithelial survival and repair, standing in contrast to its AKT-inhibiting role often observed in cancer therapy. It should be noted, however, that MK2206 is a pan-AKT inhibitor and does not selectively target the AKT1 isoform; therefore, the isoform-specific contribution of AKT1 warrants further investigation using more selective genetic approaches. Collectively, our findings support a context-dependent effect of cordycepin on AKT-related signaling in experimental colitis, a phenomenon frequently observed in multi-target bioactive natural compounds where the pharmacological outcome is dictated by the specific pathological landscape.^30–32

Notably, other core hub targets identified in our screen align with well-established pathogenic drivers of UC. IL-6 and TNF are pro-inflammatory cytokines whose overexpression in the colonic mucosa correlates with disease severity,³³ and their neutralization is a clinical mainstay for moderate-to-severe UC.^34,35 Cordycepin has been shown to effectively inhibit NF-κB signal transduction by suppressing the activities of NF-κB, IκB, and IKK.³⁶ The fact that cordycepin targets these cytokines, consistent with its binding affinity to RELA (NF-κB p65, a subunit of the NF-κB complex) and NFKB1 (key subunits of the NF-κB complex, a master regulator of IL-6/TNF transcription),³⁷ highlights its ability to interfere with the upstream inflammatory signaling cascade, rather than merely neutralizing downstream cytokines. This mechanism may confer a broader anti-inflammatory effect, as NF-κB dysregulation is central to the sustained mucosal inflammation in UC, and targeting it addresses a common node in multiple pro-inflammatory pathways.³⁸

Limitations

This study is a preliminary exploratory research with certain limitations. Firstly, while molecular docking predicted favorable binding between cordycepin and AKT1, and MD simulations further supported the stability of this interaction, direct experimental binding evidence from biophysical assays remains lacking and warrants further investigation. Secondly, the isoform-specific contribution of AKT1 has not been fully established; AKT1-specific genetic validation using conditional knockout models is needed for future mechanistic work. Thirdly, although Caco-2 cells are widely used and well-validated as an in vitro model for intestinal epithelial barrier research in colitis, their transformed colorectal adenocarcinoma origin means they cannot fully recapitulate the physiological signaling landscape of normal intestinal epithelium; future studies incorporating IBD patient-derived organoid systems would provide a more physiologically relevant cellular context for validating the findings of this study.

Conclusion

In this study, an integrated strategy combining network pharmacology, bioinformatic molecular docking, and both in vivo and in vitro experimental validation was employed to systematically identify the potential core targets and key biological processes of cordycepin in ameliorating colonic inflammation. This approach provides insights into the therapeutic potential and possible mechanisms of cordycepin in ulcerative colitis (UC). Our findings suggest that cordycepin may alleviate intestinal inflammation and promote barrier repair, at least in part, through modulation of the AKT1 signaling pathway. Collectively, these results highlight cordycepin as a promising lead compound for the development of novel IBD therapeutics and underscore the value of integrating computational and experimental approaches to explore the complex mechanisms of natural products.

Funding Statement

This work was supported by China Postdoctoral Science Foundation (2021M700546); the Key R&D Project of Jiangsu Province (BE2022719); Scientific Research Project of Jiangsu Commission of Health (MQ2024030); Changzhou Sci&Tech Program (CJ20245049); Key Project of Clinical Medicine Research of Nanjing University (2025JZ014); Scientific Research Project of Changzhou Health Commission (QY202505).

Data Sharing Statement

All data generated or analyzed during this study are included in this article. And the data that support the findings of this study are available on reasonable request from the corresponding author.

Ethics Approval

The animal study protocol was approved by the Ethics Committee of Affiliated Changzhou Children’s Hospital of Nantong University for studies involving animals (Approval No: 2023-011). And all animal experiments were performed in accordance with the guidelines of Nantong University Laboratory Animal Management Measures and regulations and the ARRIVE guidelines.

Author Contributions

Wenting Zhang and Minyan Qian are co-first authors. Wenting Zhang: Conceptualization, Formal analysis, Investigation, Methodology, Writing-original draft, Writing-review & editing. Nan Hu: Conceptualization, Investigation, Writing-original draft, Writing-review & editing. Jingting Jiang: Conceptualization, Investigation, Writing-original draft, Writing-review & editing. Minyan Qian: Formal analysis, Methodology, Visualization. Wenwei Jiang: Methodology, Formal Analysis, Writing-review & editing. Jie Chen: Methodology, Formal Analysis, Writing-review & editing. All the listed authors have read and approved the submitted manuscript. All authors took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that there are no conflicts of interest in this study.