Abstract

Background and aims

   Polygonum cuspidatum Sieb.et Zucc. (P. cuspidatum) and its active
   components have been clinically proven to have anti-hepatocellular
   carcinoma effects. However, the potential targets of P. cuspidatum for
   these effects have not yet been revealed.

Methods

   We used network pharmacology and single-cell transcriptomic analysis
   with molecular docking to elucidate the active components and targets
   of P. cuspidatum for hepatocellular carcinoma.

Results

   CDK1, ESR1, HSP90A11, and MAPK1 were shown to be the key targets of
   P. cuspidatum for hepatocellular carcinoma. P. cuspidatum was found to
   be likely correlated with the improved abnormal expression of CDK1 and
   ESR1 and the poor prognosis of HSP90AA1 and MAPK1. CDK1 was identified
   as the most potential anti-hepatocellular carcinoma target of
   P. cuspidatum. Among the active components of P. cuspidatum, physcion
   diglucoside was found to have the most potential to treat
   hepatocellular carcinoma by targeting CDK1.

Conclusion

   Our study provides novel insights into the anti-hepatocellular
   carcinoma pharmacological effects of P. cuspidatum, which could serve
   as a scientific basis for its development as a medicinal resource and
   the targeting of CDK1 for hepatocellular carcinoma treatment.

   Keywords: Hepatocellular carcinoma, Network pharmacology, Molecular
   dynamics simulation, Molecular docking, Polygonum cuspidatum,
   Single-cell transcriptomic analysis

1. Introduction

   Polygonum cuspidatum Sieb.et Zucc., (P. cuspidatum) a traditional
   Chinese medicinal herb, is widely used in southern China and Japan
   [[37]1]. Its medicinal value was first recorded in the Mingyi Bielu of
   the Han Dynasty in China and pertains to the liver, gallbladder, and
   lungs. P. cuspidatum was found to have improved on advanced or
   metastatic hepatocellular carcinoma in several clinical studies
   [[38][1], [39][2], [40][3]]. Additionally, the active compounds of
   P. cuspidatum such as emodin, polydatin, and resveratrol have been
   found to have the effect of improving hepatocellular carcinoma
   [[41]3,[42]4]. Therefore, P. cuspidatum may have the potential to treat
   hepatocellular carcinoma. However, studies on P. cuspidatum for the
   treatment of hepatocellular carcinoma are limited.

   Cyclin dependent kinase 1 (CDK1), estrogen receptor 1 (ESR1),
   mitogen-activated protein kinase 1 (MAPK1), and heat shock protein 90 α
   family class AA member 1 (HSP90AA1) are abnormally expressed in
   hepatocellular carcinoma patients, which guides the treatment of
   hepatocellular carcinoma to a certain extent [[43][5], [44][6],
   [45][7]]. CDK1 is a member of the Ser/Thr protein kinase family
   [[46]8]. CDK1 gene expression is significantly increased in
   hepatocellular carcinoma, and high expression of CDK1 was closely
   related to the poor prognosis of hepatocellular carcinoma patients
   [[47]9]. ESR1 gene expression was found to be significantly decreased
   in hepatocellular carcinoma [[48]10]. Decreased expression of ESR1 may
   impair the biological effects of antioxidant, anti-hepatic fibrosis,
   and anticancer drugs, and thus promote tumor cell proliferation and
   inhibit apoptosis [[49]11]. Therefore, targeting to increase ESR1 is
   one of the important strategies for the treatment of hepatocellular
   carcinoma. Inhibiting the activity of MAPK1 or blocking its downstream
   signaling pathway may inhibit the proliferation, invasion, and
   metastatic ability of hepatocellular carcinoma, indicating that this
   strategy may be useful for treating hepatocellular carcinoma [[50]12].
   HSP90AA1 has been shown to be a potent hepatoma marker that can be used
   for early detection, disease monitoring, and treatment efficacy of
   hepatocellular carcinoma [[51]13]. Antiangiogenic therapy targeting
   HSP90AA1 has become an important new avenues in the treatment of
   hepatocellular carcinoma [[52]14]. While inhibitors or agonists of the
   above genes have been found to have potential for the treatment of
   hepatocellular carcinoma, no specific drugs have been identified.
   Therefore, there is still a need to explore drugs that have the
   potential to treat hepatocellular carcinoma.

   Network pharmacology is a comprehensive strategy that incorporates
   elements from pharmacology, network biology, systems biology,
   bioinformatics, and computational science [[53]15]. This approach
   facilitates the screening of synergistic compounds derived from TCM
   formulas using high-throughput methods and elucidates the principles of
   combinations and the regulatory effects on networks associated with
   these traditional remedies. Single-cell transcriptomic analysis, which
   can analyze specific gene expression at the single-cell level, has been
   used to reveal the heterogeneity of cells [[54]16,[55]17]. These
   approaches are used not only to examine the complex mechanism of TCMs
   in the treatment of diseases but also to identify active compounds of
   TCMs. Molecular docking technology is a frequently used method for
   investigating the interaction sites between protein targets and the
   bioactive constituents of TCMs. Molecular dynamics simulates the
   physical trajectories and states of atoms and molecules, grounded in
   the principles of Newtonian mechanics [[56]18]. MDs can further reveal
   the interactions between active compounds and the targets of TCMs.

   This study proposed to comprehensively explore the potential components
   and targets underlying the action of P. cuspidatum in hepatocellular
   carcinoma. We probed the compounds and targets of P. cuspidatum against
   hepatocellular carcinoma using a combination of network pharmacology,
   single-cell transcriptomic analysis, and molecular docking.

2. Material and methods

2.1. Network pharmacology

2.1.1. Screening of the active compounds and targets of P. cuspidatum

   The active compounds of P. cuspidatum were searched in the TCMSP
   ([57]http://tcmspw.com/tcmsp.php) [[58]19] and ChEMBL
   ([59]https://www.ebi.ac.uk/chembl, accessed January 1, 2024) databases
   [[60]20]. Active compounds of P. cuspidatum were screened using
   Lipinski's rule of five (five indicators “MW, HAcc, HDon, AlogP, and
   RBN” were fulfilled) [[61]21]. The target proteins corresponding to the
   compounds were extracted in TCMSP and ChEMBL, and the selected targets
   were converted into the SwissProt identifiers in the UniProt database
   ([62]https://www.uniprot.org/, accessed January 1, 2024) [[63]22].

2.1.2. Screening of hepatocellular carcinoma–related genes

   The GeneCards Suite ([64]https://www.genecards.org/, accessed January
   3, 2024) [[65]23], OMIM database ([66]https://www.omim.org/. accessed
   January 3, 2024) [[67]24], and TTD database
   ([68]http://db.idrblab.net/ttd/) [[69]25] were used to search for
   hepatocellular carcinoma–related genes.

2.1.3. Screening of drug-disease targets

   The label information was separated from the drug target and disease
   target files for later use. The R language Venn Diagram program package
   was used to process the data and generate Venn diagrams.

2.1.4. Network construction

   Disease–drug interaction proteins were imported into the STRING
   database, and the human database was selected [[70]26]. After adjusting
   and setting the score value (confidence level) > 0.9, the free nodes
   were hidden to generate the PPI protein interaction network node map.
   The Cytoscape 3.10.1 software was used for further analysis.

2.1.5. GO analysis and KEGG pathway enrichment analysis

   The R language BiocManager data packet was used to transform the data
   of disease drug targets to show their gene IDs, which was convenient
   for subsequent data enrichment analysis. KEGG and GO enrichment
   analyses were performed with bioinformatics tools
   ([71]http://www.bioinformatics.com.cn, accessed January 4, 2024)
   [[72]27].

2.2. Expression and survival analysis of key targets in hepatocellular
carcinoma

   The Single Gene Analysis column was selected in the GEPIA database
   ([73]http://gepia2.cancer-pku.cn/, accessed January 5, 2024) [[74]28].
   The key targets were entered into the “enter gene name” field to obtain
   the expression of the targets and their effects on the overall survival
   rate of hepatocellular carcinoma patients.

2.3. Single-cell RNA-seq data analysis

   For the single-cell RNA-seq data, the Single Cell Portal database
   ([75]https://singlecell.broadinstitute.org/single_cell, accessed
   January 6, 2024) [[76]29] was searched for the term “hepatocellular
   carcinoma.” The visualization focused on the discovery of key genes
   specific for hepatocellular carcinoma (key targets were entered in
   “search genes and find plots” to obtain visualizations). The GSCA
   online tool ([77]http://bioinfo.life.hust.edu.cn/GSCA/#/, accessed
   January 6, 2024) [[78]29] was used to explore the biological processes
   of single-gene expression and GDSC and CTRP drug sensitivity and the
   expression of the key genes.

2.4. Molecular docking

   The active components of P. cuspidatum were retrieved in Mol2 format
   from the PubChem database ([79]https://pubchem.ncbi.nlm.nih.gov/,
   accessed January 11, 2024) [[80]30]. Key genes were identified through
   network pharmacology and single-cell transcriptomic analysis and
   subsequently downloaded from the Protein Data Bank
   ([81]http://www.rcsb.org, accessed January 11, 2024) [[82]31].
   Molecular docking studies involving active components of P. cuspidatum
   and target proteins were conducted using Schrödinger (version 12.4;
   Schrodinger LLC, New York, NY). PyMOL software (PyMOL Molecular
   Graphics System, Version 2.3.2) was used for visualization of the
   docking structures.

2.5. Molecular dynamics simulation

   Molecular dynamics simulations of protein–drug complexes were conducted
   using GROMACS 2023.2. AMBER and GAFF force fields characterized
   proteins and drugs, respectively. Energy minimization under vacuum
   preceded solvation and relaxation in SPC water, with positional
   constraints initially applied. After constraint release, 50 ns
   simulations under NPT ensemble were performed. LINCS constrained C–H,
   O–H bonds, while PME treated long-range electrostatics. The resp charge
   was used to describe the atomic charge of the drug small molecule, and
   the excess charge of the system was neutralized using Na^+ or Cl^−
   ions. Trajectories were analyzed, processed, and visualized using the
   Gromacs and VMD 1.9.3 software packages.

2.6. Statistical analysis

   Statistical analysis was performed using SPSS 16.0 software (SPSS Inc.,
   Chicago, IL, USA). Comparisons between the two treatments were
   performed using paired or unpaired t-tests. Comparisons between
   multiple treatments were performed using one-way ANOVA. p < 0.05
   indicated statistical significance.

3. Results

3.1. Compound-target network analysis

   The study flowchart is presented in [83]Fig. 1. Twelve active
   components in P. cuspidatum were identified using Lipinski's rule of
   five ([84]Table 1). A total of 6843 hepatocellular carcinoma targets
   were obtained from the GeneCards and OMIM databases, and 257
   drug-disease cross-targets were obtained as potential targets for
   further studies ([85]Fig. 2A). The compound-target network analysis was
   constructed using Cytoscape software to visualize the complex
   associations between the active components and their potential targets
   ([86]Fig. 2B).

Fig. 1.

   [87]Fig. 1
   [88]Open in a new tab

   Overall workflow of the study.

Table 1.

   The information on the active components of Polygonum cuspidatum.
   NO Molecule ID Compounds MW AlogP Hdon Hacc OB (%) DL RBN
   1 MOL000006 Luteolin 286.25 2.07 4 6 36.16 0.25 1
   2 MOL000098 Quercetin 302.25 1.50 5 7 46.43 0.28 1
   3 MOL000358 Beta-sitosterol 414.79 8.08 1 1 36.91 0.75 6
   4 MOL000492 Catechin 290.29 1.92 5 6 54.83 0.28 1
   5 MOL002259 Physciondiglucoside 608.60 −0.91 8 15 41.65 0.63 7
   6 MOL002268 Rhein 284.23 1.88 3 6 47.07 0.28 1
   7 MOL002280 Torachrysone-8-O-beta-D-(6′-oxayl)-glucoside 480.46 0.64 5
   12 43.02 0.74 8
   8 MOL012744 Resveratrol 228.26 3.01 3 3 19.07 0.11 2
   9 MOL013281 6,8-Dihydroxy-7-methoxyxanthone 258.24 2.41 2 5 35.83 0.21
   1
   10 MOL013287 Physovenine 262.34 2.08 1 5 106.21 0.19 2
   11 MOL013288 Picralinal 366.45 1.80 1 6 58.01 0.75 3
   12 MOL013289 Polydatin 390.42 1.11 6 8 21.44 0.50 5
   [89]Open in a new tab

Fig. 2.

   [90]Fig. 2
   [91]Open in a new tab

   Key ingredients and key targets for hepatocellular carcinoma. (A)
   Common targets of Polygonum cuspidatum compounds and hepatocellular
   carcinoma. (B) Polygonum cuspidatum-compositions-targets network. (C)
   PPI network diagram of common targets. A circle with a larger size and
   darker color represents a larger degree. (D) The degree ordering of PPI
   networks.

3.2. PPI network construction and analysis

   PPI networks were generated to investigate the interactions between the
   targets of P. cuspidatum. As shown in [92]Fig. 2C and D, 15 targets had
   a high degree of interaction, including TP53 (degree = 86), AKT1
   (degree = 56), HSP90AA1 (degree = 50), TNF (degree = 44), ESR1
   (degree = 42), IL6 (degree = 40), BCL2 (degree = 36), CASP3
   (degree = 36), CCND1 (degree = 36), MAPK1 (degree = 36), IL1B
   (degree = 32), CDK1 (degree = 30), EGFR (degree = 30), MDM2
   (degree = 30), and MYC (degree = 30).

3.3. GO enrichment and KEGG pathway analyses

   The potential biological effects of P. cuspidatum on hepatocellular
   carcinoma were analyzed by functional annotation and enrichment
   analysis. A total of 1543 BP, 35 CC, and 102 MF terms were enriched. A
   bubble plot of the 10 top terms identified by p-value and counts was
   visualized. GO biological processes analysis indicated that the target
   genes of P. cuspidatum were associated with the regulation of the
   cellular response to chemical stress, response to lipopolysaccharide,
   and response to molecule of bacterial origin ([93]Fig. 3A). GO cellular
   component analysis indicated that the target genes were involved in the
   membrane raft, membrane microdomain, and membrane region ([94]Fig. 3B).
   GO molecular function analysis suggested that the target genes of
   P. cuspidatum were involved in DNA-binding transcription factor
   binding, RNA polymerase II–specific DNA-binding transcription factor
   binding, and nuclear receptor activity ([95]Fig. 3C).

Fig. 3.

   [96]Fig. 3
   [97]Open in a new tab

   GO enrichment and KEGG analysis. (A) Analysis of molecular function
   (MF). (B) Analysis of biological process (BP). (C) Analysis of cellular
   component (CC). (D) Enrichment analysis of KEGG signaling pathways.

   The enrichment analysis of KEGG signaling pathways included 143 terms.
   Most genes were involved in lipid and atherosclerosis (n = 37, n
   represents the number of targets.), human cytomegalovirus infection
   (n = 34), chemical carcinogenesis - receptor activation (n = 32),
   hepatitis B (n = 31), and hepatitis C (n = 29) ([98]Fig. 3D). As shown
   in [99]Table 2, the top 10 pathways were related to lipid and
   atherosclerosis, prostate cancer, AGE-RAGE signaling pathway, bladder
   cancer, and hepatitis B.

Table 2.

   KEGG enrichment analysis.
   ID Description GeneRatio p-value p adjust
   Hsa05417 Lipid and atherosclerosis 37/141 1.60 × 10^−27 4.15 × 10^−25
   Hsa05215 Prostate cancer 27/141 4.37 × 10^−26 5.60 × 10^−24
   Hsa04933 AGE-RAGE signaling pathway in diabetic complications 27/141
   1.09 × 10^−25 9.43 × 10^−24
   Hsa05219 Bladder cancer 20/141 2.40 × 10^−25 1.55 × 10^−23
   Hsa05161 Hepatitis B 31/141 1.70 × 10^−24 8.85 × 10^−23
   Hsa05163 Human cytomegalovirus infection 34/141 2.51 × 10^−23
   1.08 × 10^−21
   Hsa05418 Fluid shear stress and atherosclerosis 28/141 8.54 × 10^−23
   3.16 × 10^−21
   Hsa05160 Hepatitis C 29/141 1.82 × 10^−22 5.90 × 10^−21
   Hsa05207 Chemical carcinogenesis - receptor activation 32/141
   6.18 × 10^−22 1.77 × 10^−20
   Hsa04657 IL-17 signaling pathway 23/141 6.91 × 10^−21 1.79 × 10^−19
   [100]Open in a new tab

   Significance and survival analyses were performed for the 15 key
   targets obtained from PPI networks. The results showed that CDK1 was
   significantly decreased while ESR1 was significantly increased in
   hepatocellular carcinoma. (p < 0.05) ([101]Fig. 4E and F). Survival
   analysis was used to examine the association of the genes with
   hepatocellular carcinoma survival. The results showed that CDK1, ESR1,
   HSP90AA1, and MAPK1 genes were correlated with overall survival
   ([102]Fig. 5E, F, H, and K). These results suggest that the
   anti-hepatocellular carcinoma effect of P. cuspidatum may be related to
   the aberrant expression of CDK1 and ESR1 and the poor prognosis of
   HSP90AA1 and MAPK1. These results suggest that CDK1, ESR1, HSP90AA1,
   and MAPK1 are key targets of P. cuspidatum for its effects against
   hepatocellular carcinoma.

Fig. 4.

   [103]Fig. 4
   [104]Open in a new tab

   The expression of the key genes in hepatocellular carcinoma. (A–O)
   Expression of the top 15 targets obtained from the PPI network in
   hepatocellular carcinoma.

Fig. 5.

   [105]Fig. 5
   [106]Open in a new tab

   Survival analysis of the key genes. (A–O) Expression of the top 15
   targets obtained from the PPI network in hepatocellular carcinoma.

3.4. Expression of key genes in cell subsets

   To explore the distribution of the expression of CDK1, ESR1, HSP90AA1,
   and MAPK1 in hepatocellular carcinoma, single-cell analysis was
   performed to investigate the enrichment of key genes. As shown in
   [107]Fig. 6A–E, HSP90AA1 and MAPK1 were enriched in almost all
   hepatocellular carcinoma subsets, while CDK1 and ESR1 were mainly
   expressed in individual hepatocellular carcinoma subsets. CDK1 was
   mainly expressed in RBCs, while ESR1 was mainly expressed in
   endothelial cells. These results confirmed that hepatocellular
   carcinoma has metabolic heterogeneity and suggest that P. cuspidatum
   may exert anti-hepatocellular carcinoma effects by targeting multiple
   cell types.

Fig. 6.

   [108]Fig. 6
   [109]Open in a new tab

   Expression of the key genes in liver cancer cell subsets. (A)
   Pathological phenotypes of CDK1, ESR1, HSP90A11, and MAPK1 in
   hepatocellular carcinoma. (B) Expression of CDK1 in hepatocellular
   carcinoma cell subsets. (C) Expression of ESR1 in hepatocellular
   carcinoma cell subsets. (D) Expression of HSP90A11 in hepatocellular
   carcinoma cell subsets. (E) Expression of MAPK1 in hepatocellular
   carcinoma cell subsets.

3.5. Molecular docking

   From the results of network pharmacology and single-cell transcriptomic
   analysis, luteolin (Molecule ID: MOL000006), quercetin (Molecule ID:
   MOL000098), beta-sitosterol (Molecule ID: MOL000358), catechin
   (Molecule ID: MOL000492), physcion diglucoside (Molecule ID:
   MOL002259), rhein (Molecule ID: MOL002268),
   torachrysone-8-O-beta-D-(6′-oxayl)-glucoside (Molecule ID: MOL002280),
   resveratrol (Molecule ID: MOL01274), 6,8-dihydroxy-7-methoxyxanthone
   (Molecule ID: MOL013281), physovenine (Molecule ID: MOL013287),
   picralinal (Molecule ID: MOL013288) and, polydatin (Molecule ID:
   MOL013289) were conjugated with CDK1 (PDB ID: [110]5LQF), ESR1 (PDB ID:
   [111]1L2I), HSP90AA1 (PDB ID: [112]1UYD), and MAPK1 (PDB ID:
   [113]3W55). The total score was used to evaluate the results of
   ligand–protein interactions. A total score equal to or greater than 7
   indicated a good ligand–protein interaction [[114]32]. The total scores
   of the 4 key targets docking with the 12 key components are listed in
   [115]Table 3. ESR1, HSP90AA1, and MAPK1 and the components
   beta-sitosterol, 6,8-dihydroxy-7-methoxyxanthone, physovenine, and
   picralinal had no significant effect on the key ingredients or targets
   screened. The four targets showed excellent binding effects with eight
   other compounds, and a typical docking diagram is shown in
   [116]Fig. 7A–D. As shown in [117]Fig. 7E–H, the main types of
   interactions are hydrogen bonding, alkyl Pi-alkyl interactions,
   hydrophobic interactions, and π-π stacking forces.

Table 3.

   Total scores of 4 key targets docking with 12 key components.
   Components Targets
     __________________________________________________________________

                CDK1   ESR1 HSP90AA1  MAPK1
    MOL000006  9.622  9.231    8.232  7.215
    MOL000098  9.962  8.232    9.323  9.011
    MOL000358  8.232  5.221   10.221 −6.331
    MOL000492 10.413 11.300    9.626  8.991
    MOL002259 13.836 10.221   16.028  3.991
    MOL002268 10.066  9.221    9.331 −9.134
    MOL002280  7.321  8.331    8.332  4.551
    MOL012744 11.232  8.212   11.223  8.323
    MOL013281  7.321  5.331    6.991  8.231
    MOL013287  7.252  6.881    5.221  4.341
    MOL013288  8.321  6.331    8.232  6.018
    MOL013289  9.232   9.55   11.231  7.661
   [118]Open in a new tab

Fig. 7.

   [119]Fig. 7
   [120]Open in a new tab

   Docking complex 3D and 2D diagrams of four key targets along with their
   strongest binding partners. (A), (E), CDK1 (PDB ID: [121]5LQF) with
   physcion diglucoside. (B), (F), ESR1 (PDB ID: [122]1L2I) with catechin.
   (C), (G), HSP90A11 (PDB ID: [123]1UYD) with physcion diglucoside. (D),
   (H), MAPK1 (PDB ID: [124]3W55) with rhein.

3.6. CDK1 has the most potential as an anti-hepatocellular carcinoma target
for the active components of P. cuspidatum

   CDK1, ESR1, HSP90A11, and MAPK1 are involved in apoptosis, cell cycle,
   PI3K/AKT, and other biological processes that are closely related to
   hepatocellular carcinoma ([125]Fig. 8A). The four genes were abnormally
   expressed in the pathological stages of hepatocellular carcinoma and
   two cancers related to the digestive organs (gastric and esophageal
   cancers) ([126]Fig. 8B, C). CDK1, HSP90A11A, and MAPK1 showed
   fluctuating changes in the pathological progression of hepatocellular
   carcinoma, while ESR1 decreased gradually with disease development. The
   correlation analysis between GDSC and CTRP drug sensitivity and the
   expression of the four genes showed that CDK1 has the most potential as
   an anti-hepatocellular carcinoma target among the active components of
   P. cuspidatum ([127]Fig. 8D and E). CDK1 showed excellent binding with
   12 compounds of P. cuspidatum ([128]Table 3); a typical docking diagram
   is shown in [129]Fig. 7F. The total score of CDK1 docking with physcion
   diglucoside was the highest, and the protein–ligand complex was
   selected for molecular dynamics simulation and further verification.

Fig. 8.

   [130]Fig. 8
   [131]Open in a new tab

   Drug sensitivity analysis of key genes. (A) Pathway analysis of key
   genes. (B–C) Expression of key genes in different cancers during
   pathological periods. (D–E) Drug sensitivity analysis of key genes.

3.7. Molecular dynamics simulation

   As shown in [132]Fig. 9A, the RMSD of the protein and small molecule
   complex remains relatively stable at around 0.5 Å after 25 ns,
   indicating the stability of the binding between physcion diglucoside
   and CDK1. As shown in [133]Fig. 9B, the α-helix and β-fold structures
   of the protein have small RMSF values (with a maximum value of
   approximately 0.97 Å), indicating a stable structure throughout the MD
   trajectory. The RG is commonly used to measure the average radius of
   mass weight in a simulation and can also indicate the closeness of
   molecules. There was no significant change in RG, indicating the stable
   binding effect between physcion diglucoside and CDK1 ([134]Fig. 9C). In
   molecular dynamics simulations, the binding free energy of a
   protein-ligand complex is often used as an index to measure its
   stability. Generally, a binding free energy of a complex < −60 kcal/mol
   indicates high stability. [135]Figure 9D showed the binding free energy
   and related energy decomposition data of the physcion diglucoside–CDK1
   complex. The binding free energy of the physcion diglucoside–CDK1
   complex was approximately −72.5 kcal/mol, confirming its high
   stability.

Fig. 9.

   [136]Fig. 9
   [137]Open in a new tab

   Analysis of molecular dynamics simulation for physcion diglucoside and
   CDK1 (PDB ID: [138]5LQF) performed by Schrödinger. (A) RMSD (root mean
   square deviation); (B) RMSF (root mean square fluctuation); (C) RG; (D)
   binding free energy.

4. Discussion

   This study was the first to reveal the components and potential targets
   of P. cuspidatum for hepatocellular carcinoma using network
   pharmacology, single-cell transcriptomics analysis, and molecular
   docking. CDK1, ESR1, HSP90A11, and MAPK1 were found to be the key
   targets of P. cuspidatum in hepatocellular carcinoma, and the
   anti-hepatocellular carcinoma effect of P. cuspidatum correlated with
   the improved abnormal expression of CDK1 and ESR1 and the poor
   prognosis of HSP90AA1 and MAPK1. This study also found that CDK1 is the
   most potential anti-hepatocellular carcinoma target of P. cuspidatum.
   Molecular docking and molecular dynamics simulation results showed that
   physcion diglucoside in P. cuspidatum and CDK1 stably bind, and that
   physcion diglucoside is expected to be a lead compound for targeting
   CDK1 in treating hepatocellular carcinoma.

   The anti-hepatocellular carcinoma effects of P. cuspidatum have been
   confirmed by researchers [[139]33]. Several studies have indicated that
   P. cuspidatum exhibits anti-cancer activity in hepatocellular carcinoma
   through AKT, JNK, ERK, STAT3, BAX, and other targets
   [[140]3,[141]4,[142]34,[143]35]. In this study, we found that CDK1,
   ESR1, HSP90A11, and MAPK1 were the key targets of P. cuspidatum against
   hepatocellular carcinoma, and CDK1 was the most promising
   anti-hepatocellular carcinoma target. CDK1 plays an important
   regulatory role in hepatocellular carcinoma, and its high expression
   promotes the proliferation of hepatocellular carcinoma cells [[144]36].
   Because of the important role of CDK1 in hepatocellular carcinoma, CDK1
   inhibitors have become an important research direction in
   hepatocellular carcinoma treatment. miR-195-5p was found to target
   CDK1, regulate de novo DNA synthesis, and inhibit the proliferation of
   hepatocellular carcinoma. Metformin and cucurbitacin B were found to
   have similar effects [[145]37,[146]38]. However, no drug that targets
   CDK1 for hepatocellular carcinoma has been identified. Therefore, the
   discovery of potential CDK1-targeting drugs for hepatocellular
   carcinoma remains an area of active research.

   The other finding of this study is that physcion diglucoside from
   P. cuspidatum may target CDK1. Physcion diglucoside is a glycoside of
   a substance consisting of emodin and quercetin 3-O-gentobioside. Emodin
   was reported to have anti-hepatocellular carcinoma properties
   [[147]39]. Emodin promotes apoptosis and inhibits the proliferation and
   migration of hepatocellular carcinoma [[148]40,[149]41]. The target of
   action involves STAT3, PI3K, and AKT. However, emodin has not been
   reported to target CDK1. Quercetin 3-O-gentobioside, a glycosidized
   derivative of quercetin, has been found to target CDK1 in some studies.
   The combination of these two compounds in physcion diglucoside
   may result in a novel mechanism of action that specifically targets
   CDK1. Quercetin 3-O-gentobioside is a glycosidized derivative of
   quercetin. Quercetin is a star compound for hepatocellular carcinoma
   [[150]42]. Quercetin 3-O-gentobioside may modify the bioavailability,
   stability, or subcellular localization of emodin, allowing it to access
   and inhibit CDK1 in a manner that would not be possible with
   emodin alone. Additionally, the glycosidic linkage between emodin and
   quercetin 3-O-gentobioside may create a new molecular entity
   with unique properties that enable it to bind to CDK1 with higher
   affinity or specificity. These speculations need to be evaluated in
   future studies.

   This study has several limitations. First, this study obtained the
   active components and targets of hepatocellular carcinoma
   anti-hepatocellular carcinoma. However, the key results have not been
   validated. Therefore, we need to validate the results in future
   analyses. Additionally, there are several active components in
   P. cuspidatum that exert anti-hepatocellular carcinoma effects.
   Therefore, it is still worthwhile to explore the mechanisms and targets
   of other components in P. cuspidatum against hepatocellular carcinoma.
   These questions will be investigated in future research.

5. Conclusion

   This study identified the active components of P. cuspidatum that may
   exert anti-hepatocellular carcinoma activity using network pharmacology
   and single-cell transcriptomic analysis with molecular docking. The
   results showed P. cuspidatum may be correlated with the improved
   abnormal expression of CDK1 and ESR1 and the poor prognosis of HSP90AA1
   and MAPK1. Additionally, CDK1 was identified as the most potential
   anti-hepatocellular carcinoma target for P. cuspidatum. Molecular
   docking and molecular dynamics simulation results showed that physcion
   diglucoside from P. cuspidatum has the potential to treat
   hepatocellular carcinoma by targeting CDK1. The above findings provided
   new insights into the treatment of hepatocellular carcinoma and the
   development of new drugs.

Funding

   Supported by National Training Program of Innovation and
   Entrepreneurship for Undergraduates (202410145118).

CRediT authorship contribution statement

   Wenze Wu: Conceptualization, Data curation, Formal analysis, Funding
   acquisition, Investigation, Methodology, Project administration,
   Resources, Software, Supervision, Validation, Visualization, Writing –
   original draft, Writing – review & editing. Yuzhu Shi:
   Conceptualization, Formal analysis, Investigation, Resources, Software.
   Yongzi Wu: Conceptualization, Data curation, Formal analysis, Funding
   acquisition, Project administration. Rui Zhang: Conceptualization.
   Xinyan Wu: Conceptualization, Investigation, Project administration,
   Software. Weidi Zhao: Conceptualization, Funding acquisition, Project
   administration, Visualization. Zhiyuan Chen: Conceptualization,
   Investigation, Project administration. Gang Ye: Conceptualization,
   Project administration, Resources.

Acknowledgments