Investigating the prognostics and theranostic cell cycle-related genes of Hinokinin in high-grade glioma: an integrated network pharmacology, bioinformatics, and molecular docking approach Authors: Oukhdouch Anass, Nadir Youssef, Salman El mostafa, Jebhi Khalid, Zinbi Basma, Sellami Souad, Rais Hanane Published: 2026 ABSTRACT AbstractBackgroundGlioblastoma (GBM) is the most malignant and prevalent primary brain tumor in adults, accounting for up to 60% of all adult brain tumor, classified as WHO grade 4. Uncontrolled cell-cycle progression is one of the hallmarks of GBM. The patient prognosis is still poor after multimodal treatment, underscoring the need for new therapeutic strategies. Derived from Piper cubeba, hinokinin is a bioactive dibenzylbutyrolactone lignan that has demonstrated interesting anticancer effects. However, little is known about its molecular processes in GBM.MethodsTo understand the therapeutic potential of Hinokinin in GBM, we used an integrated network pharmacology, transcriptomic and molecular docking method. GBM-related genes and differentially expressed genes (DEGs) were found in the GeneCards and GEPIA2 databases. Hinokinin-associated targets were predicted using SwissTargetPrediction. Protein-protein interaction (PPI) networks, Gene Ontology (GO), and KEGG pathway enrichment analyses were used to further evaluate the core targets that were identified by intersecting these datasets. Expression profiling, survival analysis, immune infiltration analysis, genetic alteration analysis, and scRNA-sequencing datasets were used to determine clinical significance. Hinokinin's binding affinity and drug-likeness were evaluated using molecular docking and ADMET analyses.ResultsHinokinin's targets in GBM were found to be 36 overlapping genes with enrichment in cell-cycle regulation, kinase activity, and carcinogenic pathways. CDK2, PLK1, CHEK1, AURKA, and AURKB were identified as important hub genes by PPI network analysis. These genes were related in tumors with p53/Rb alterations, markedly elevated in GBM tissues, linked to poor overall survival (OS), and mostly expressed at the single-cell level in malignant cell types. Strong negative relationships between hub gene expression and several immune cell types were found by immune infiltration analysis. Frequent transcriptional dysregulation of these kinases was shown by genetic alteration analysis, and this was associated with a poor prognosis. Hinokinin interacts to all five hub proteins with strong binding energies. Drug-likeness and blood brain barrier (BBB) permeability, were validated by ADMET.ConclusionOur findings highlight that Hinokinin may have antitumor effects in GBM targeting key cell-cycle and mitotic kinases involved in tumor growth and poor prognosis. These results provide a strong mechanistic rationale for Hinokinin as a promising multitarget therapeutic for GBM.Keywords: Glioblastoma, Hinokinin, Cell-cycle kinases, Network pharmacology, Molecular docking Introduction Glioblastoma (GBM) is the most malignant and prevalent primary brain tumor in adults, accounting for up to 60% of all adult brain tumor, classified as WHO grade 4.1 Even though surgery, radiation therapy, and temozolomide chemotherapy have all gotten better, patients still don't live very long, with a median of only 14 to 16 months.2 Histologically, GBM is characterized by high cell proliferation, necrosis, and microvascular proliferation (MVP). Clinical treatment decisions are not made based on histological classification. Nonetheless, the prognosis of patients is closely correlated with the diversity of molecular alterations within GBM. In particular, IDH-WT glioblastoma typically has greater levels of epidermal growth factor receptor (EGFR) amplification, TERT promoter mutation and PTEN loss. Patients with MGMT promoter methylation, might have a better prognosis and respond better to treatment of Temozolomide (TMZ)3,4.4 The cell cycle is controlled by a several cell-cycle-related kinases that regulate DNA replication, chromosome segregation, and mitotic progression. One of the important actor of this regulatory system include the cyclin-dependent kinases (CDKs) and a group of mitotic kinases, such as Polo-like kinases and Aurora kinases, which promote orderly transitions through the G1, S, G2, and M phases.5 In parallel, checkpoint kinases maintain genomic integrity by stopping cell-cycle progression in response to replication stress or DNA damage.6 One of the major oncogenic mechanisms underlying glioma progression is dysregulation of these pathways, particularly through overactivation of CDKs and checkpoint regulators leads to uncontrolled proliferation and reduction of apoptosis.7 Large-scale genomic studies of GBM have consistently shown that alterations affecting cell-cycle regulation represent some of the most frequent and biologically significant events in these tumors.8 In addition to chemotherapy, targeted therapies have been explored to inhibit molecular alterations such as EGFR amplification, VEGF-driven angiogenesis, and PI3K/AKT/mTOR signaling; however, most targeted agents have shown limited efficacy due to the high genetic heterogeneity and adaptive resistance of GBM.9 Immunotherapy has emerged as a major research direction, including immune checkpoint inhibitors, therapeutic vaccine, and CAR-T cell strategies. Although these approaches have demonstrated encouraging preclinical activity, clinical success has been limited due to the highly immunosuppressive glioma microenvironment.10 In addition to those strategies currently used in GBM therapy, several medicinal plants traditionally used in folk medicine have attracted increasing scientific interest. Among these, Piper cubeba has been employed for centuries in traditional medicine and is now emerging as a promising source of bioactive compounds and played an important role in the search for novel therapeutic agents for various diseases.11 Piper cubeba, originally native to Java widely referred to as Java pepper, is well established in several regions. Notable populations of this species are found across Indonesia, India, parts of North Africa, and both Morocco and Western Europe.12 (−)-Hinokinin, one of the lignans extracted from Piper cubeba seeds or is obtained by partial synthesis of (−)-cubebin. (−)- Hinokinin has biological activities include anti-inflammatory, anti-oxidant and anti-tumoral.13 Recently, showed that Hinokinin can inhibit tumor cell proliferation by inducing apoptosis and modulating key signaling cascades such as CDKs and PI3K/Akt.14 Network pharmacology is a method to discover new drug targets and molecular mechanisms systems biology approaches by combining computational analysis with in vitro experiments and integrating a large amount of information. Recent advances in network pharmacology and bioinformatics have revolutionized the exploration of drug-cancer interactions by integrating multi-omics data such as genomic, transcriptomic and proteomic.15 Moreover, molecular docking simulations provide structural insights into ligand-protein interactions, helping to validate potential targets at the atomic level.16 In this study, we employed an integrated network pharmacology, bioinformatics, and molecular docking approach to explore the molecular mechanisms of (−)-hinokinin targeting cell-cycle related kinases in GBM. By combining multi-database target prediction, functional enrichment, protein-protein interaction analysis, survival assessment, and structural validation, we aimed to identify the key molecular targets and signaling pathways through which Hinokinin may exert its therapeutic potential in GBM (Fig. 1). Graphical abstract. Materials and methods Screening candidate targets of Hinokinin and GBM The SMILES and 3D structures of Hinokinin were downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/). The potential targets of Hinokinin were retrieved from SwissTargetPrediction (http://www.swisstargetprediction.ch/) accessed on [06/09/2025]. To identify disease-associated targets, GBM-related genes were obtained from the GeneCards database (https://www.genecards.org/) accessed on [02/09/2025]. “Glioblastoma multiforme” was used as the keyword, and genes related GBM were selected. The UniProt database (https://www.uniprot.org/) was used to standardize each gene. Duplicated entries were removed. To further refine GBM-associated genes, we used GEPIA2 (http://gepia2.cancer-pku.cn/), an interactive online tool based on TCGA and GTEx transcriptomic data, to identify differentially expressed genes (DEGs) between GBM tumor and normal brain tissues. DEGs were defined using the default thresholds (|log2 fold change| ≥ 1 and p-value <0.05). The DEG list was downloaded from GEPIA2 and visualized using a volcano plot generated in R (version 2025.09.2 + 418) based on the following criteria: |Fold Change(log2) | > 2 and an adjusted p < 0.05. Identification of overlapping genes and CTD network construction The overlapping targets between Hinokinin-related genes, GBM-associated genes, and DEGs were identified using a Venn diagram tool available at Bioinformatics & Evolutionary Genomics (http://www.bioinformatics.com.cn/). These intersecting genes were considered as candidate therapeutic targets for subsequent network pharmacology and functional analyses. A compound-target-disease (CTD) network was then constructed and visualized in cytoscap (Version 3.10.4). Gene ontology (GO) and KEGG pathway enrichment analysis To investigate the functional roles and signaling pathways of the intersecting genes, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted using the R softwear. The clusterProfiler package was employed to perform enrichment analysis. GO analysis was carried out across three categories: biological process (BP), molecular function (MF), and cellular component (CC). KEGG pathway enrichment was performed to identify significantly associated signaling pathways. Adjusted p-values were calculated, and terms with adjusted p < 0.05 were considered statistically significant. Protein-protein interaction network and identification of hub genes The 36 identified intersecting genes considered as target genes were uploaded to the STRING database (https://string-db.org/) to construct a protein-protein interaction (PPI) network, with a confidence score threshold set to 0.4. The resulting PPI network was exported and subsequently visualized using Cytoscape software. To identify top core genes with the highest interaction strength, we applied the CytoHubba plugin in Cytoscape. Four topological analysis algorithms, including Maximal Clique Centrality (MCC), Maximum Neighborhood Component (MNC), Edge Percolated Component (EPC), and Degree, were employed. The top-ranked genes from each algorithm were subsequently compared, and a Venn diagram was generated to determine the intersection among the four ranking methods. The overlapping genes were defined as the hub genes. Survival analysis of hub genes To investigate the prognostic value of hub genes, we used the GEPIA2 web server (http://gepia2.cancer-pku.cn/). Gene expression data were derived from The Cancer Genome Atlas (TCGA) with matched normal samples from GTEx. Expression was compared between tumor and normal tissues using boxplot function. Kaplan–Meier survival curves were generated and statistical significance was evaluated using the log-rank test (p < 0.05). UALCAN database analysis The UALCAN database (http://ualcan.path.uab.edu/analysisprot.html) is an interactive web resource used for the analysis of cancer Omics data. In the present study, UALCAN was employed to compare the differential expression levels of the five core genes between GBM tissues and corresponding normal brain tissues. Additionally, we analyzed the expression profiles of these genes across various cancer types stratified according to p53/Rb pathway status alterations. Expression of hub genes in multiples tumor type and IHC staining The differential expression levels of CDK2, PLK1, CHEK1, AURKA and AURKB in tumor versus normal tissues were evaluated using the TIMER2.0 (http://timer.cistrome.org/). TIMER2.0 integrates RNA-seq expression profiles from The Cancer Genome Atlas (TCGA) across 33 cancer types. Furthermore, the Human Protein Atlas (HPA) database (https://www.proteinatlas.org/) was utilized to investigate the protein expression patterns and tissue distribution of the core genes through immunohistochemical (IHC) analysis in both normal brain tissues and GBM samples. Analysis of tumor-immune cell interactions using the TISIDB database The TISIDB database (http://cis.hku.hk/TISIDB/) is a web-based integrated repository that explores the interactions between tumors and the immune system across multiple datasets. In this study, the TISIDB portal was used to analyze the correlation between CDK2, PLK1, CHEK1, AURKA and AURKB mRNA expression and the abundance of tumor-infiltrating lymphocytes (TILs) and TME-related cells across various immune subtypes and molecular subtypes of GBM. Hub genes validation To compare the mRNA expression levels of the hub genes between GBM and normal tissues, publicly available gene expression datasets were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) and analyzed using R Studio software. The GSE4290 dataset, generated using the GPL570 [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array platform, includes 23 non-tumor and 81 GBM samples. Genetic alterations in the hub genes were analyzed using the cBioPortal for Cancer Genomics (https://www.cbioportal.org), a comprehensive, open-access resource for exploring multidimensional cancer genomics data. The dataset for diffuse glioma (TCGA, GDC) was selected for this analysis. Genetic alterations, including mutations, copy number variations (amplifications and deep deletions), and mRNA expression changes, were retrieved and analyzed. The hub genes were queried in the cBioPortal platform, and the resulting data were used to assess the frequency and distribution of alterations across the cohort which contain 516 patients. Additionally, overall survival was extracted and associated with the presence or absence of genetic alterations in the hub genes. Single-cell RNA analysis of hub genes using TISCH2 The Tumor Immune Single-cell Hub 2 (TISCH2, https://tisch.comp-genomics.org/) is an online platform that provides single-cell transcriptomic profiles of the TME across multiple cancer types. In this study, TISCH2 was utilized to investigate the expression landscape of the hub genes at single-cell resolution in GBM. Two GBM single-cell RNA sequencing datasets were selected for analysis: Glioma_GSE131928_Smartseq2,17 and Glioma_GSE163108_Smartseq2.18 The expression patterns of the hub genes were explored across different cell populations within the TME, including malignant cells, immune cells, and stromal cells. Molecular docking Protein structure retrieval and preparation The X-ray crystal structures of CDK2, PLK1, CHEK1, AURKA and AURKB were obtained from the Protein Data Bank (PDB). Structures were inspected and prepared using AutoDock Tools (MGLTools). For each structure, non-standard residues, crystallographic water molecules and co-crystallized ligands were removed. Missing side chains or residues were modeled where necessary using AutoDock Tools. Polar hydrogens were added and Gasteiger partial charges were assigned. The protein structures were saved in PDBQT format for docking. Ligand preparation Ligand structures were obtained from PubChem and converted to three-dimensional structures. Each ligand geometry was energy-minimized in Avogadro, major protonation states at physiological pH (7.4) were considered where relevant. Final ligand files were saved in PDBQT format using AutoDock Tools. Docking protocol and grid definition Molecular docking analysis was performed using AutoDock to predict the binding affinity and interaction mode between the target protein and selected ligands. For targets with co-crystallized ligands. The active site was defined from the literature. The grid box center and size used for each protein are reported in Table 1. As a general guideline the grid box dimensions were chosen to fully encompass the binding pocket (typical box size: 40 × 40 × 40 Å), with slight adjustments per protein to accommodate pocket shape. The docking procedure was conducted using the VinaDock tool to estimate the most favorable binding conformations. The binding affinity and interaction profiles were analyzed to identify key amino acid residues. Visualization and interpretation of protein–ligand complexes were performed using PyMOL and Discovery Studio. Table 1Grid box docking parameters for molecular docking.TargetsPDB IDSpacing (angstrom)Center grid boxLiterature referencesX CenterY CenterZ CenterCDK22C6O0.375−11.0721.71−17.08549PLK14J520.37535.989−31.302−11.47150CHEK14FTK0.37514.8982.07123.36249AURKA5DT00.375−17.022−32.3447.97449AURKB4AF30.37516.434−17.71−1.31649 Grid box docking parameters for molecular docking. ADMET analysis The small-molecule Hinokinin was subjected to absorption, distribution, metabolism, excretion, and toxicity (ADMET) in the SWISS ADME (http://www. swissadme.ch) accessed on [12/11/2025] to evaluate ADMET. The following parameters were studied: drug physiochemical properties, Lipophilicity, Water Solubility, Pharmacokinetics, Druglikeness and Medicinal Chemistry. Statistical description P < 0.05 was considered statistically significant. The following P values were considered: *P < 0.05, **P < 0.01, and ***P < 0.001, with statistical methods as described in materials section. Results Targets of Hinokinin against GBM The 2D chemical structure of Hinokinin is shown in Fig. 2A. A total of 100 potential target genes of Hinokinin were obtained from the SwissTargetPrediction database. Additionally, 9681 GBM-related genes were retrieved from the GeneCards database. Differentially expressed genes (DEGs) in GBM were identified using the GEPIA2 database, including 1336 upregulated and 943 downregulated genes. By intersecting the Hinokinin targets, GBM-related genes and DEGs, 36 overlapping genes were identified and considered as core target genes potentially mediating the therapeutic effects of Hinokinin against GBM (Fig. 2B). These overlapping genes were subsequently used to construct a compound-target-disease interaction network, revealing strong associations between Hinokinin and key GBM-related targets (Fig. 2C). Furthermore, enrichment analysis of these core genes identified the top 20 significant term (Supplementary Table 1) mainly associated with tumor progression, lymphoid malignancies, enzyme inhibition, and cellular proliferation (Fig. 2D). Network pharmacology analysis of Hinokinin in glioblastoma. A. Chemical structure of Hinokinin. B. Venn diagram showing the overlap between glioblastoma-associated genes, differentially expressed genes (DEGs), and predicted Hinokinin targets, identifying shared candidate genes. C. Hinokinin-glioblastoma target interaction network illustrating the relationships between Hinokinin and common GBM-related genes. D. Bubble plot of the top enriched disease-related terms for Hinokinin-glioblastoma overlapping genes. Gene ontology (GO) and KEGG pathway enrichment analysis GO enrichment and KEGG pathway analysis of the 36 intersecting target genes was performed to explore their potential biological roles and mechanism. The biological process (BP), the enriched terms were mainly associated with innate immune response, signal release, activation of immune signaling pathways, and the ERK1/ERK2 cascade. In the cellular component (CC), target genes were primarily enriched in the synaptic membrane, chromosomal region, protein kinase complex, and spindle microtubule, indicating their involvement in signal transduction and cell division. In molecular function (MF), the genes were significantly enriched in protein serine/threonine kinase activity, protein tyrosine kinase activity, and transmembrane receptor protein kinase activity, highlighting the importance of kinases regulatory (Fig. 3A). These findings suggest that Hinokinin may exert anti-GBM effects through modulation of immune response pathways and inhibition of kinase-related to cell cycle. KEGG pathway enrichment analysis as shown in Fig. 3B, the significantly enriched pathways included the FoxO signaling pathway, ErbB signaling pathway, Fc epsilon RI signaling pathway, Apoptosis, Cell cycle, and Ras signaling pathway. These pathways are closely associated with cell proliferation, apoptosis regulation, and oncogenic signaling.19 Among them, the FoxO and ErbB signaling pathways showed the highest enrichment scores and adjusted p-values <0.01. The pathway-gene network (Fig. 3C) further highlighted key hub genes such as PLK1, PLK2, CDK2, MAPK10, RAF1, AURKA and ERBB2. Functional enrichment and pathway gene network analysis of Hinokinin-GBM overlapping genes. A. Gene Ontology (GO) enrichment analysis of overlapping genes, showing significantly enriched biological processes, cellular components, and molecular functions. B. KEGG pathway enrichment analysis associated with Hinokinin-GBM target genes. C. Pathway-gene interaction network illustrating the relationships between enriched KEGG pathways and their corresponding genes. PPI network of Hinokinin and GBM We used STRING database to establish a PPI network. A total of 36 nodes and 97 edges was identified (Supplementary Fig. 1). Then four topological algorithms (MCC, MNC, EPC, and Degree) were applied via the CytoHubba plugin in cytoscap to rank the most influential nodes. Each algorithm identified a subset of top eight high-scoring genes (Supplementary Table 2), with CDK2, PLK1, CHEK1, AURKA, and AURKB consistently appearing among the top-ranked candidates (Fig. 4A and B). The Venn diagram analysis demonstrated that these five genes overlapped across all four topological methods, thereby defining them as the key hub genes in the PPI network (Fig. 4C). Protein-protein interaction (PPI) network construction and identification of hub genes. A PPI networks generated by four CytoHubba topological algorithms: Maximal Clique Centrality (MCC), Maximum Neighborhood Component (MNC), Edge Percolated Component (EPC), and Degree. Nodes with warmer colors indicate higher ranking scores. B. Visualization of the core PPI network highlighting the major hub genes (CDK2, PLK1, CHEK1, AURKA, and AURKB) based on interaction strength. Node size corresponds to degree centrality. C. Venn diagram showing the intersection of top-ranked genes among the four algorithms, identifying five overlapping genes as the key hub genes. Survival analysis of hub genes in GBM To evaluate the clinical relevance of the identified hub genes, we performed expression and survival analyses using the GEPIA2 platform. CDK2, PLK1, CHEK1, AURKA, and AURKB were significantly upregulated in GBM tumor samples compared with normal brain tissues (Fig. 5A, B, C, D and E). Kaplan Meier analysis revealed that high expression of each hub gene was associated with significantly worse overall survival. Specifically, high expression of hub genes correlated with reduced survival (CDK2, PLK1, CHEK1, AURKA and AURKB: p < 0.01). These results confirmed also using UALCAN database, CDK2, PLK1, CHEK1, AURKA, and AURKB were significantly upregulated (p < 0.01) in GBM primary tumor samples compared with normal brain tissues (Supplementary Fig. 2). The expression of hub genes across cancers type in somatic alterations based on p53/Rb-related pathway status revealed the protein expression levels of CDK2, CHEK1, AURKA, and AURKB varied among cancer types depending on the p53/Rb pathway status. In GBM, all four proteins exhibited higher expression in the p53/Rb-altered group compared to the other group (Supplementary Fig. 3). Differential expression in GBM and normal tissue of the hub genes and their overall survival analysis. Differential expression, correlation of hub genes across human cancers and IHC staining To further validate the clinical relevance of the identified hub genes and to confirm their protein-level expression. We examined the expression patterns of CDK2, PLK1, CHEK1, AURKA, and AURKB across 33 tumor types using the TIMER2.0 database. As shown in Fig. 6A, all fives genes exhibited significantly higher expression levels in pan-cancers compared with their corresponding normal tissues. The analysis revealed the hub genes were all significantly (p < 0.001) upregulated in GBM compared with normal brain tissues. Correlation analysis was performed between hub genes using TCGA transcriptomic data. As shown in Fig. 6B, CDK2 expression exhibited a strong positive correlation with PLK1 (r = 0.434, p < 0.001), CHEK1 (r = 0.687, p < 0.001), AURKA (r = 0.555, p < 0.001), and AURKB (r = 0.436, p < 0.001), suggesting that CDK2 is co-regulated with multiple key cell cycle checkpoint kinases. Immunohistochemistry (IHC) slides of hub genes from the HPA were analyzed. As illustrated in Fig. 7, the proteins CDK2, PLK1, AURKA, and AURKB showed low or undetectable expression in normal brain tissues (M-00100), whereas a marked increase in staining intensity was observed in high-grade glioma malignant tissues (M-938033). Notably, PLK1 and AURKA displayed strong cytoplasmic and nuclear staining in GBM samples. Expression of hub genes in several tumor types and normal tissue, and hub genes correlation (*p < 0.05, **p < 0.01, ***p < 0.001). Immunohistochemical expression of CDK2, PLK1, AURKA, and AURKB in normal brain and glioblastoma tissues. Tumor-infiltrating immune cells The correlation between the expression of hub genes and the abundance of tumor-infiltrating immune cells in GBM was analyzed using the TISIDB database. As shown in Fig. 8A, CDK2 expression was negatively correlated with the abundance of CD8+ T cells (ρ = −0.236, p = 0.00227), Tregs (ρ = −0.193, p = 0.013), macrophages (ρ = −0.232, p = 0.00273), and activated dendritic cells (aDCs) (ρ = −0.241, p = 0.00178), while no significant correlation was observed with NKT. Similarly, PLK1 expression exhibited a strong negative association with Tregs (ρ = −0.211, p = 0.00677), macrophages (ρ = −0.310, p < 0.001), aDCs (ρ = −0.314, p < 0.001), and NKT cells (ρ = −0.384, p < 0.001), whereas the correlation with CD8+ T cells was not significant (ρ = −0.032, p = 0.679) (Fig. 8B). CHEK1 expression exhibited significant negative correlations with Tregs (ρ = −0.337, p < 0.001), macrophages (ρ = −0.379, p < 0.001), aDCs (ρ = −0.458, p < 0.001), and NKT cells (ρ = −0.343, p < 0.001). A weaker but significant negative correlation was also observed with CD8+ T cells (ρ = −0.186, p = 0.0163) (Fig. 8C). AURKA expression showed significant inverse associations with Tregs (ρ = −0.238, p = 0.00206), macrophages (ρ = −0.319, p < 0.001), aDCs (ρ = −0.321, p = 0.001), and NKT cells (ρ = −0.347, p = 0.001), whereas no correlation was observed with CD8+ T cells (Fig. 8D). As shown in Fig. 8E, AURKB expression exhibited significant negative correlations with Tregs (ρ = −0.317, p < 0.001), macrophages (ρ = −0.324, p < 0.001), aDCs (ρ = −0.354, p < 0.001), and NKT (ρ = −0.461, p < 0.001). A weaker significant correlation was also observed with CD8+ T cells (ρ = −0.198, p = 0.0107). In addition, the expression levels of the five hub genes were evaluated across immune and molecular subtypes in GBM, the results revealed significant expression differences were observed in immun subtypes (p < 0.05, Kruskal-Wallis test), with the C4 subtype exhibiting the highest expression for all five genes (Supplementary Fig. 4 A). Regarding molecular subtype, the higher expression observed in the Mesenchymal-like and LGm6-GBM subtypes (Supplementary Fig. 4B). Expression of CDK2, PLK1, CHEK1, AURKA, and AURKB is negatively correlated with tumor-infiltrating lymphocytes, including Tregs, macrophages, activated dendritic cells, and NKT cells in glioblastoma. Expression level analysis of hub genes Differential expression analysis revealed that CDK2, PLK1, CHEK1, AURKA and AURKB were markedly overexpressed in GBM compared to normal brain controls (Fig. 9A). The upregulation of these genes was statistically significant (p < 0.0001). Consistently, the volcano plot (Fig. 9B) highlighted these genes among the significantly upregulated, further validating their involvement in GBM pathogenesis and their potential as key regulators of cell cycle progression and mitotic control. Differential expression analysis of hub genes in glioblastoma. A. Boxplots showing the expression of hub genes in GBM tissues compared with normal brain samples. Significant upregulation of all five genes in GBM (**** p < 0.0001) compared to normal tissues. B. Volcano plot of differentially expressed genes in GBM versus normal samples, with the five hub genes highlighted. All hub genes were located in the significantly upregulated region (red dots). Genetic alterations in hub genes The TCGA diffuse glioma cohort (n = 516) showed genetic alterations in at least one of the five hub genes in 49 patients (9% of the cohort), according to genetic alteration analysis (Fig. 10). The largest rates of modification were shown by CHEK1 and AURKA, followed by CDK2, PLK1, and AURKB. The most prevalent alteration type across these genes was mRNA upregulation. This alteration was consistent with immunohistochemical analysis confirmed markedly increased protein expression of CDK2, PLK1, AURKA, and AURKB in GBM tissues compared to normal brain tissues (Fig. 7). Survival analysis demonstrated that patients harboring genetic alterations in any of the five hub genes exhibited a significantly worse overall survival compared to those without alterations (log-rank p = 4.55e-7). The Kaplan-Meier curve clearly shows a steep decline in survival probability among the altered group, suggesting that genomic and transcriptional dysregulation of these mitotic regulators is strongly associated with poor prognosis in glioma. Genetic alterations and prognostic impact of hub genes in glioma. A. Oncoplot for fives hub genes of different genetic alterations. Each column represents a sample/patient (n = 516). One or more queried genes were altered in 49 patients. B. Distribution of alteration types across the cohort. C. Kaplan-Meier survival analysis according to alteration status of the hub genes. Hub genes expression detected in malignant cell clusters at single-cell level Single-cell clustering revealed distinct populations, including malignant cells (AC-like, MES-like, NPC-like, and OPC-like), immune cells (CD8+ T exhausted cells and monocytes/macrophages), and stromal cells by using the Glioma_GSE131928_Smartseq2 dataset (Fig. 11A). The expression patterns of CDK2, CHEK1, PLK1, AURKA, and AURKB were predominantly enriched in malignant cell clusters, MES-like and NPC-like malignant cells exhibited particularly high expression signals (Supplementary Fig. 5). the UMAP plot from the Glioma_GSE163108_Smartseq2 dataset (Fig. 11B) demonstrates distinct clustering of major immune cell subtypes, including CD4+ conventional T cells (CD4Tconv), CD8+ T cells (CD8T), and regulatory T cells (Tregs). CDK2, PLK1, CHEK1, AURKA and AURKB exhibited low expression levels within CD4Tconv and CD8 + T. Notably, the expression of AURKA and CDK2 appeared more prominent compared to the other hub genes. Single-cell transcriptomic analysis of hub gene expression in GBM. UMAP plots from (A) Glioma_GSE131928_Smartseq2 and (B) Glioma_GSE163108_Smartseq2 datasets showing hub gene expression (CDK2, CHEK1, PLK1, AURKA, AURKB) across malignant, immune, and stromal cell populations. Hub genes were predominantly enriched in malignant clusters, particularly MES-like and NPC-like subtypes, while showing low expression in immune cell populations. Molecular docking analysis Molecular docking studies were performed to evaluate the binding affinity and interaction modes of Hinokinin with the five hub protein targets: CDK2, PLK1, CHEK1, AURKA, and AURKB. The binding energies and key molecular interactions are summarized in Table 2. All five hub proteins exhibited favorable binding energies with Hinokinin, with values ranging from −7.2 to −9.5 kcal/mol, indicating strong and stable binding complexes. The compound-target interactions were visualized in 2D and 3D using Discovery Studio and PyMOL (Fig. 12A-E). Hydrogen bonding, hydrophobic interactions, and π-π stacking were identified as the predominant stabilizing forces across all docking complexes. Table 2Molecular docking prediction of binding energy, hydrogen bond and covalent bond.TargetsPDB IDResidue involved in H bondingH-bond length (Å)Residue involved in covalent bondingCovalent-bond length (Å)Binding energy (kcal/Mol)Hinokinin (PubChem CID442,879)CDK22C6OARG-214, LEU-202, LYS-250, THR-2182.16, 3.52, 2.36, 2.39ARG-217, VAL-2514.21, 4.87−6.6PLK14J52ALA-80, ARG-57, ARG-134, ARG-136, CYS-133, LEU-59, LEU-1324.31, 4.03, 4.81, 5.48, 4.90, 5.29, 5.42−9.5CHEK14FTKASP-148, SER-1473.03, 2.84LEU-15, LEU-84, LEU-137, VAL-683.74, 3.58, 4.90, 5.03−9.8AURKA5DT0LYS-1623.59ALA-160, ALA-213, LEU-139, LEU-263, THR-217, VAL-1475.17, 5.23, 3.91, 3.81, 3.70, 5.30−8.8AURKB4AF3ASN-205, GLU-1613.31, 2.55ALA-217, LEU-83, LEU-207, LYS-1064.32, 5.07, 3.78, 4.46−9.6 Molecular docking prediction of binding energy, hydrogen bond and covalent bond. Molecular docking of Hinokinin with hub proteins. 2D and 3D binding interaction visualizations of Hinokinin with A: CDK2, B: PLK1, C: CHEK1, D: AURKA, and E: AURKB, showing hydrogen bonds, hydrophobic interactions, and π–π stacking contacts. For CDK2, Hinokinin formed multiple hydrogen bonds with residues ARG-214, LEU-202, LYS-250, and THR-218, with bond lengths ranging from 2.16 to 3.52 Å, and additional stabilizing interactions with ARG-217 and VAL-251. In CHEK1, hydrogen bonds were observed with ASP-148 and SER-147 (bond lengths: 3.03 and 2.84 Å, respectively), along with hydrophobic interactions involving LEU-15, LEU-84, LEU-137, and VAL-68. AURKA formed a hydrogen bond with LYS-162 (3.59 Å) and additional hydrophobic contacts with ALA-160, ALA-213, LEU-139, LEU-263, THR-217, and VAL-147. In AURKB, Hinokinin formed hydrogen bonds with ASN-205 and GLU-161 (bond lengths: 3.31 and 2.55 Å, respectively), with further stabilization by ALA-217, LEU-83, LEU-207, and LYS-106. Although no hydrogen bond formation was detected between Hinokinin and PLK1, the complex was stabilized by multiple nonpolar interactions involving ALA-80, ARG-57, ARG-134, ARG-136, CYS-133, LEU-59, and LEU-132, yielding the strongest binding energy among all targets (−9.5 kcal/mol). Hinokinin demonstrated favorable ADMET properties (Table 3) with a molecular weight of 354.40 g/mol and optimal physicochemical balance (TPSA 63.22 Å2). It showed moderate solubility (Log S = −4.41) and full compliance with Lipinski and Ghose drug-likeness rules. Predicted oral bioavailability was good (score = 0.55) with high gastrointestinal absorption and blood-brain barrier (BBB) permeability. The molecule is not a P-gp substrate but exhibits CYP450 inhibition, suggesting possible metabolic interactions. No PAINS or structural alerts were identified, and its synthetic accessibility (3.53) indicates ease of synthesis. Table 3ADMET analysis of Hinokinin.CategoryParameterPredicted ValuePhysicochemical PropertiesMolecular weight (g/mol)354.40Hydrogen bond donors0Hydrogen bond acceptors6Topological Polar Surface Area (TPSA, Å2)63.22Water SolubilityLog S (ESOL)−4.41Solubility1.37e-02 mg/mlClassModerately solubleDruglikenessLipinskiYes, 0 violationGhoseYesBioavailability score0.55PharmacokineticsGastrointestinal (GI) absorptionHighBlood–brain barrier (BBB) permeabilityYesP-glycoprotein (P-gp) substrateNoCYP450 inhibition (CYP1A2, CYP2C19, CYP2C9, CYP2D6, CYP3A4)YesMedicinal ChemistryPAINS (Pan-Assay Interference Compounds)0 alertBrenk (Structural alert)0 alertSynthetic accessibility3.53 ADMET analysis of Hinokinin. Discussion Hinokinin is a bioactive lignan primarily derived from Piper cubeba, and has been increasingly investigated for its potential anticancer effect. (−)-Hinokinin is a member of dibenzylbutyrolactone lignan family, which are known for diverse biological activities, including anti-tumor effects by blocking proliferation pathways and induced apoptosis.20 (−)-Hinokinin has demonstrated cytotoxic activity against a range of human tumor cell lines, including HT-29 (human colon adenocarcinoma), A-549 (human lung adenocarcinoma), MCF-7 (human breast adenocarcinoma) and HeLa (human cervical cancer), with ED50 values below 4 μg/ml in sensitive lines, a threshold conventionally associated with antitumoral activity.21 (−)-Hinokinin has been shown to induce G2/M cell cycle arrest and contribute to the antiproliferative effects of doxorubicin in breast cancer cell lines MCF-7 and SKBR-3, without affecting normal MCF-10 A breast cells, suggesting a degree of tumor selectivity and safety.13 Hinokinin has shown no genotoxic or mutagenic activity in various experimental systems, with preclinical studies indicating no in vivo genotoxicity at doses up to 40 mg/kg.22The purpose of this work was to conduct an integrated approach to investigate the molecular mechanisms of Hinokinin and the potential to which it may have regulatory effects on cell cycle-related genes implicated in GBM progression, by combining network pharmacology, transcriptomic data, and molecular docking. We identified Hinokinin targets using the SwissTargetPrediction database. We then screened GBM-related expression genes from GeneCards and identified differentially expressed genes (DEGs) using GEPIA2. We analyzed these DEGs using a volcano plot in R and, by integrating Hinokinin-related targets, GBM-associated genes and DEGs, identified 36 overlapping genes representing the most biologically relevant core targets. The enrichment results emphasise that the shared targets are functionally enriched in pathways related to tumor progression, cellular proliferation and enzyme regulation. These processes are central to GBM progression and are characterized by uncontrolled cell-cycle progression and enhanced DNA replication.23 The PPI network revealed CDK2, PLK1, CHEK1, AURKA, and AURKB as key targets of Hinokinin's therapeutic activity against GBM. These genes involved in the cell-cycle machinery, particularly checkpoints controlling DNA replication, mitotic entry, and chromosomal segregation.24 The GO enrichment analysis indicates that these hub genes involved in immune regulation, kinases activities and cell-cycle control.21 The strong representation of kinases in molecular functions, particularly serine/threonine and tyrosine kinase activities support the involvement of this genes in GBM progression. KEGG highlight the FoxO, ErbB, Ras, and cell-cycle pathways as key regulatory pathways potentially modulated by these genes. These pathways recognized as major drivers of proliferation, apoptosis resistance, and therapeutic failure in GBM, with FoxO inactivation and ErbB hyperactivation being strongly associated with tumor progression and poor outcomes.25, 26 The FoxO signaling is essential for controlling apoptosis, tumor suppression, and cell cycle arrest. It has been demonstrated that FOXO transcription factors are often dysregulated in GBM, with FOXO1 being a crucial driver of the glial-mesenchymal transition linked to the most aggressive GBM subtype and a convergence site of receptor tyrosine kinase signaling.27 However, in GBM cell lines, FOXO1 inhibition was demonstrated to cause apoptosis.28 The differential expression and survival analysis demonstrate that CDK2, PLK1, CHEK1, AURKA, and AURKB were upregulated in GBM tissues compared with normal brain, additionally these genes were associated with unfavorable prognostic outcome. The consistent association between high expression and overall survival indicates that those genes not only drive tumor progression but may also serve as prognostic markers. Hengrui Liu, Jieling Weng, indicate that overexpression or hyperactivation of CDK2 has been associated with tumor aggressiveness, poor prognosis in glioma.29 According to Luo et al., glioma patients with high PLK1 levels had a worse prognosis than those with low PLK1 levels.30 Chinyama, Harold A et al., showed that high expression level of CHEK1 promotes GBM tumor aggressiveness, immunosuppression, and progression.31 Barton, Valerie N et al., demonstrated that AURKA expression correlates with poor survival in glioblastoma patients and that AURKA inhibition significantly reduced glioma cell proliferation.32 The p53-p21-RB signaling pathway plays a central role in cell cycle regulation, where p53 activation induces p21 expression, which in turn inhibits CDK2 activity and maintains RB, thereby blocking E2F-dependent transcription of cell cycle genes.33, 34 The elevated expression of CDK2, CHEK1, AURKA, and AURKB in GBM tumors with p53/Rb pathway alterations highlights the dependence of p53/Rb-defective gliomas on cell-cycle promoting kinases. This is consistent with the established molecular profile of GBM, in which loss of p53 or retinoblastoma (Rb) function increases dependence on checkpoint and mitotic regulators for survival.35 The expression analysis in multi-cancer showed that CDK2, PLK1, CHEK1, AURKA, and AURKB are upregulated across multiple tumor types. Such higher overexpression is consistent with previous studies showing that cell-cycle and mitotic checkpoint kinases are among the most frequently activated oncogenic drivers across human cancers, including gliomas.36 In GBM, the significant upregulation of all five hub genes reinforces their relevance to glioma process. The strong positive correlations observed in this study between CDK2 and other genes, indicate a coordinated regulation of cell-cycle progression in GBM. This co-expression pattern suggests a close functional link between CDK2 in cell cycle entry, CHEK1 in checkpoint regulation, and Aurora kinases in mitotic progression, which collectively enable GBM cells to sustain rapid proliferation despite genomic instability.5, 37 Moreover, the IHC validation using HPA database, the strong staining of CDK2, PLK1, AURKA, and AURKB in high-grade glioma specimens, underscoring their tumor-specific activation. The prominent nuclear and cytoplasmic localization of PLK1 and AURKA in GBM samples is consistent with their established roles in centrosome maturation, spindle assembly, and mitotic checkpoint regulation.32, 38 A growing body of literature has demonstrated that immune-related mechanisms play a pivotal role in antitumor responses and may represent promising diagnostic and therapeutic targets in oncology.39 Therefore, the association between hub gene expression and tumor-infiltrating immune cells was analyzed. The negative correlations observed between the high expression of CDK2, PLK1, CHEK1, AURKA, and AURKB and CD8+ T cells, Tregs, macrophages, aDCs, and NKT cells indicate that overexpression of cell-cycle and mitotic checkpoint kinases is linked to reduced immune infiltration or inhibition immune activation in GBM. These results suggest that these hub genes may contribute to the establishment of an immunosuppressive TME in GBM. Luo L, Zhang XY, Zhen YW, et al., have been found PLK1 prevented immune cells from infiltrating GBM, allowing the tumor cells to avoid the immune system.30 Among the hub genes, PLK1, CHEK1, AURKA, and AURKB showed particularly strong negative correlations with antigen-presenting cells of macrophages and aDCs, which play critical roles in initiating antitumor immune responses. This observation aligns with previous studies demonstrating that PLK1 and AURKA promote immune evasion by suppressing macrophage infiltration and maintaining an immunosuppressive cytokine milieu in glioma.30, 40 Sen T et al., revealed that by enabling tumor cells to withstand DNA damage without causing immunogenic cell death, CHEK1 has been linked to restricting immune recognition.41 Our results were further validated by the enrichment of CDK2, PLK1, CHEK1, AURKA, and AURKB expression in the C4 immune subtype, typically linked to a mesenchymal and known to show increased macrophage/microglia infiltration with dysfunctional antitumor immunity.42 Mesenchymal-like and LGm6-GBM molecular subtypes consistently exhibit increased hub gene expression, which supports their correlation with tumor aggressiveness, immunological dysregulation, and treatment resistance.43 Genetic alteration analysis demonstrated that CDK2, PLK1, CHEK1, AURKA, and AURKB are frequently dysregulated in gliomas, with upregulation of expression being the most common alteration type. Patients with alterations in any of these hub genes exhibited significantly worse overall survival, highlighting their prognostic significance. The transcriptional overexpression of these genes likely facilitates unregulated cell-cycle progression by passing DNA damage checkpoints and enhancing abnormal mitotic kinase activity, contributing to the aggressive proliferative phenotype of GBM. Previous study have shown that aberrant activation of CDK2 and PLK1 promotes abnormal cell cycle progression,5 while CHEK1, AURKA and AURKB overexpression enhances tumor aggressiveness, particularly in p53/RB deficient glioma.44, 45 Interestingly, our results indicate that alterations in the expression of mitotic kinases contributes to glioma malignancy and supports their significance as potential targets. Single-cell transcriptomic analysis showed that CDK2, CHEK1, PLK1, AURKA and AURKB were enriched in malignant glioma cell clusters, particularly within MES-like and NPC-like. In contrast, these hub genes showed less expression in immune cell population, including CD4+ and CD8+ T cells, indicating that their dysregulation is largely tumor intrinsic.46 Overall, these findings reinforce the critical involvement of cell cycle-related kinases in malignant GBM cell states and provide single-cell level support for Hinokinin's predicted therapeutic targeting of these pathways. Molecular docking analysis of Hinokinin exhibits strong binding affinities to CDK2, PLK1, CHEK1, AURKA and AURKB, may considered as multitarget inhibitory potential against the cell-cycle kinases as we demonstrate in this work. The variability in binding energies of Hinokinin between the protein's kinase domains reflects to structural and physicochemical differences (Table 2). Hinokinin exhibited the strongest binding affinities to PLK1, CHEK1, and AURKB (−9.5 to −9.8 kcal/mol) due to favorable accommodation within hydrophobic subpockets and the formation of multiple stabilizing interactions with conserved catalytic and hinge-region residues. Previous studies have shown that PLK1 possess well-defined hydrophobic grooves and flexible activation loops that enhance ligand stabilization through van der Waals contacts and hydrogen bonding, confirming higher binding affinities for small-molecule inhibitors.47 ADMET analysis revealed favorable drug-likeness features, including compliance with Lipinski and Ghose rules, good oral bioavailability, high gastrointestinal absorption, and blood-brain barrier permeability, an essential requirement for effective GBM therapeutics.48 Predicted CYP450 inhibition suggests potential metabolic interactions, the absence of PAINS alerts and acceptable synthetic accessibility further support Hinokinin as a promising, brain-penetrant small-molecule candidate targeting GBM. Conclusion In summary, this used an integrated bioinformatics and network pharmacology approach to identify Hinokinin as a potential therapeutic candidate in GBM. CDK2, PLK1, CHEK1, AURKA, and AURKB were identified as key hub genes associated with GBM progression, poor prognosis, immune suppression. Molecular docking and ADMET analyses further supported the strong binding affinity and favorable drug-like properties of Hinokinin, including blood brain barrier (BBB) permeability. Nevertheless, these results are based exclusively on computational analyses and need experimental validation, functional assays using glioma cell lines and in vivo systems are necessary to confirm the biological activity, molecular targets, and therapeutic efficacy of Hinokinin. Despite these limitations, our results provide a strong mechanistic foundation for future experimental and translational studies exploring Hinokinin as a novel therapeutic strategy for GBM. Ethics declarations No applicable. CRediT authorship contribution statement Oukhdouch Anass: Supervision, Software, Methodology, Formal analysis, Conceptualization. Nadir Youssef: Visualization, Data curation. Salman El mostafa: Docking analysis validation and confirmation. Jebhi Khalid: Writing – review & editing, Supervision. Zinbi Basma: Writing – review & editing, Validation. Sellami Souad: Conceptualization. Rais Hanane: Project administration, Methodology, Investigation, Conceptualization. Consent to participate No applicable. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declaration of competing interest The authors declare no competing interests. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.jgeb.2026.100724. Contributor Information Oukhdouch Anass, Email: a.oukhdouch.ced@uca.ac.ma. Rais Hanane, Email: ha.rais@uca.ac.ma. Supplementary data Supplementary material References Associated Data Supplementary Materials Supplementary material