Integrative Identification and Validation of Exosome-Related Genes as Diagnostic Biomarkers and Potential Therapeutic Targets in Cholangiocarcinoma

Integrative Identification and Validation of Exosome-Related Genes as Diagnostic Biomarkers and Potential Therapeutic Targets in Cholangiocarcinoma | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Integrative Identification and Validation of Exosome-Related Genes as Diagnostic Biomarkers and Potential Therapeutic Targets in Cholangiocarcinoma Fangfeng Liu, Zhe Wang, Qianchang Wang, Yu Wang, Zhengjian Wang, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8144185/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Cholangiocarcinoma (CCA) is an aggressive biliary malignancy with limited diagnostic tools and poor prognosis. Early detection remains challenging due to nonspecific symptoms and a lack of reliable biomarkers. Exosomes, as stable carriers of molecular cargos, have emerged as promising sources for non-invasive cancer biomarkers. Here, we integrated multiple GEO datasets to identify exosome-related differentially expressed genes (ERDEGs) associated with CCA. Through differential expression analysis, machine-learning feature selection, and immune infiltration profiling, we identified two key exosome-related genes, WNT5A and PFN2 , as potential diagnostic biomarkers. Both genes showed robust diagnostic performance across internal and external validation cohorts. Functional enrichment revealed strong associations with extracellular matrix organization, EMT activation, and immune regulation pathways. Molecular docking suggested potential therapeutic compounds targeting these genes. Immunohistochemistry further confirmed significant overexpression of WNT5A and PFN2 in CCA tissues compared with adjacent controls. Collectively, our findings highlight WNT5A and PFN2 as promising exosome-related biomarkers that may improve early diagnosis and offer new therapeutic opportunities for cholangiocarcinoma. Health sciences/Biomarkers Biological sciences/Cancer Biological sciences/Computational biology and bioinformatics Biological sciences/Immunology Health sciences/Oncology Cholangiocarcinoma Exosomes WNT5A PFN2 Diagnostic biomarkers Tumor Microenvironment Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 1 Introduction Cholangiocarcinoma (CCA) is an aggressive and heterogeneous cancer arising from the biliary epithelium, representing nearly 3% of all gastrointestinal malignancies [ 1 – 3 ]. Despite notable progress in imaging diagnostics and therapeutic interventions, both the incidence and mortality of CCA have continued to increase worldwide in recent decades [ 4 ]. Because of its subtle onset, the absence of specific early clinical manifestations, and the lack of reliable biomarkers, CCA is frequently diagnosed at advanced or metastatic stages [ 5 , 6 ]. Consequently, the prognosis remains dismal, with a five year survival rate below 10% for most patients [ 4 , 7 ]. Surgical resection is still the only treatment offering a potential cure [ 5 ]; however, the majority of individuals are not candidates for surgery at the time of diagnosis due to extensive disease progression. Therefore, it is crucial to identify new molecular indicators and establish effective diagnostic approaches to facilitate early detection and improve patient outcomes in CCA [ 1 , 7 ]. Exosomes are nanosized extracellular vesicles that mediate intercellular communication by transferring proteins, microRNAs, and long noncoding RNAs, and they have been increasingly recognized as vital modulators in tumor progression [ 8 – 10 ]. In CCA, exosomes influence the tumor microenvironment by interacting with immune cells and cancer-associated fibroblasts (CAFs), thereby enhancing cellular proliferation, migration, and invasion through activation of Wnt/β-catenin signaling and cytoskeletal remodeling [ 11 , 12 ]. In addition, exosomal non-coding RNAs regulate gene expression within both tumor and immune cells, suggesting their promise as diagnostic and therapeutic candidates [ 13 – 15 ]. Nevertheless, the precise functions of exosome-related genes (ERGs) in regulating immune infiltration, cytoskeletal rearrangement, and intracellular signaling in CCA remain insufficiently understood [ 16 , 17 ]. Although increasing evidence indicates that exosomes contribute to the remodeling of the immune landscape and oncogenic pathways in CCA [ 14 , 16 , 18 ], comprehensive investigations linking ERGs with immune regulation, cytoskeletal dynamics, and related molecular networks are still limited [ 11 ]. Furthermore, the possible involvement of ERGs in pathways such as “Cytoskeleton in Muscle Cells,” “Protein Digestion and Absorption,” and “Viral Protein Interaction with Cytokine and Cytokine Receptor” has yet to be clarified [ 14 , 19 ], emphasizing a notable gap in our current understanding of CCA pathobiology. In this study, we systematically profiled ERGs associated with CCA and identified two markedly downregulated genes, WNT5A and PFN2, as diagnostic biomarkers. Based on integrated multi-cohort transcriptomic datasets, an ERG-based diagnostic model was constructed using LASSO regression. Functional exploration included enrichment analysis, immune infiltration characterization, gene-interaction mapping, drug-repurposing prediction, molecular docking, and immunohistochemical validation. Collectively, our findings elucidate the molecular roles of ERGs in CCA and highlight WNT5A and PFN2 as promising biomarkers and therapeutic targets. 2 Materials and Methods Data Collection and Preprocessing Gene expression datasets were retrieved from the Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ) [ 20 ]. Raw data were standardized to ensure cross-platform comparability, and all datasets were merged to generate a comprehensive gene dataset (CGD). To correct for unwanted variation and hidden confounding factors, surrogate variable analysis (SVA) [ 21 ] was applied. Principal component analysis (PCA) was subsequently used to verify the effective reduction of batch effects, ensuring the robustness of the dataset for downstream analyses. Identification of Exosome-Related Differentially Expressed Genes The CGD was divided into cholangiocarcinoma (CCA) and normal control groups. Differential expression analysis was performed using the “limma” R package [ 22 ]. Genes meeting the thresholds of |logFC| > 0 and p < 0.05 were identified as differentially expressed genes (DEGs). These DEGs were then intersected with a curated set of exosome-related genes (ERGs) to obtain exosome-related differentially expressed genes (ERDEGs). Gene Ontology (GO) Enrichment Analysis To elucidate the biological significance of the 42 ERDEGs identified in CCA, Gene Ontology (GO) functional enrichment analysis was conducted across the categories of Biological Process (BP), Cellular Component (CC), and Molecular Function (MF) [ 23 ]. The “clusterProfiler” R package (v3.16.1) [24)]was used to perform GO enrichment. Statistically significant enrichment was defined as a corrected p-value (p.adj) < 0.05 and false discovery rate (FDR, Q value) < 0.25. The Benjamini–Hochberg (BH) procedure was applied to adjust for multiple comparisons and reduce false positives [ 25 ]. Gene Set Enrichment Analysis (GSEA) Gene set enrichment analysis was performed using the “clusterProfiler” R package (v3.16.1) [ 24 ] to identify significantly enriched pathways and biological processes between CCA and normal samples based on the full gene expression matrix of exosome-related genes. Pairwise correlations between the key genes and all other genes in the training cohort were calculated and ranked in descending order to form the gene list for enrichment testing. Pathways with |NES| > 1 and p < 0.05 were considered significantly enriched. Diagnostic Model Construction and Validation A diagnostic model for CCA was established based on the 42 ERDEGs. Initially, univariate logistic regression was applied to identify genes with significant diagnostic potential (p < 0.05). The least absolute shrinkage and selection operator (LASSO) [ 26 ] regression was then used to refine candidate genes and construct the diagnostic model. The optimal λ parameter was determined by 10-fold cross-validation [ 27 ], and corresponding gene coefficients were obtained. The diagnostic efficiency of the model was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). A nomogram was generated to provide a clinical visualization of the predictive model, and calibration plots were used to examine the concordance between predicted and observed outcomes [ 28 ]. Decision curve analysis (DCA) was further applied to estimate the clinical benefit of the model. In addition, Circos plots were used to display the chromosomal positions of the key genes, and differential expression between tumor and control tissues was assessed to validate their diagnostic significance. Immune Infiltration Analysis (ssGSEA and MCPcounter) The infiltration levels of 28 immune cell types in the CGD were quantified using the single-sample gene set enrichment analysis (ssGSEA) algorithm implemented in the “GSVA” R package. Differences in immune infiltration between the CCA and normal groups were analyzed, and correlations between the expression of the key genes (WNT5A and PFN2) and immune cell abundance were further examined using the “MCPcounter” algorithm [ 29 ]. Gene Interaction Network Analysis (GeneMANIA) Interaction networks for WNT5A and PFN2 were generated using the GeneMANIA plugin in Cytoscape (v3.10.0). Default parameters were used, integrating evidence from multiple sources such as co-expression, physical interaction, predicted association, co-localization, shared pathways, and genetic interaction. The networks were visualized with each key gene as the central node, and the top 20 functionally associated genes were displayed. Functional enrichment of the resulting network was conducted through GeneMANIA’s built-in enrichment tool, with FDR < 0.05 set as the significance threshold. Drug Enrichment Analysis and Molecular Docking Potential therapeutic compounds targeting CCA-related genes were identified through drug enrichment analysis using Enrichr ( https://maayanlab.cloud/Enrichr/ ) and the Drug–Gene Interaction Database (DGIdb, https://www.dgidb.org/ ) [ 30 ]. The key genes WNT5A and PFN2 were input into both databases, and significantly enriched drug–gene associations were selected based on FDR < 0.05. Candidate compounds were further validated through cross-referencing with DrugBank, ChEMBL, and DrugMatrix databases to confirm biological relevance. Three-dimensional protein structures of target genes were retrieved from the AlphaFold Protein Structure Database ( https://www.alphafold.ebi.ac.uk/ ). Models with a predicted Local Distance Difference Test (pLDDT) confidence score above 70 were selected for analysis. Small-molecule structures of enriched drugs were obtained from PubChem or ChEMBL in SDF format. Molecular docking was carried out using AutoDock Vina, with the grid box size set to 30 × 30 × 30 Å, centered on either the known ligand-binding pocket or the top predicted cavity from AutoSite [ 31 ]. The exhaustiveness parameter was fixed at 8. Compounds with binding free energy ≤ − 6 kcal/mol were regarded as having strong affinity and potential therapeutic value. Immunohistochemistry (IHC) Validation Formalin-fixed, paraffin-embedded CCA and adjacent normal bile duct tissues were sectioned at 4 µm thickness. Sections were deparaffinized in xylene, rehydrated through graded ethanol, and subjected to antigen retrieval in citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was quenched using 3% hydrogen peroxide for 10 min, followed by blocking with 5% bovine serum albumin (BSA) for 30 min. The tissue samples used for IHC validation were obtained from the archived pathology specimens of the Department of Pathology, Shandong Provincial Hospital Affiliated to Shandong First Medical University. These specimens were collected from patients who underwent surgical resection between June 2025 and September 2025, and postoperative histopathology confirmed a diagnosis of cholangiocarcinoma. Three patients were selected for the validation of WNT5A and another three patients for the validation of PFN2, with each patient providing both tumor tissue and paired adjacent normal bile duct tissue (six FFPE slides per gene). The acquisition of human specimens was approved by the Ethics Committee for Biomedical Research Involving Humans of Shandong Provincial Hospital Affiliated to Shandong First Medical University (protocol code SWYX:NO. 2025 − 540, approved on 1 September 2025), and the requirement for written informed consent was waived by the committee. Slides were incubated overnight at 4°C with primary antibodies against WNT5A (Affinity Biosciences, Cat# DF6856, 1:50) and PFN2 (Abcam, Cat# 2H7C12, 1:50). After washing, sections were treated with HRP-conjugated secondary antibodies and visualized using an M&R HRP/DAB Detection Kit (Cat# HC301, DAB dilution 1:10), followed by hematoxylin counterstaining. The stained slides were dehydrated, mounted, and observed under a light microscope. Immunoreactivity was semi-quantitatively assessed by the H-score method, where staining intensity was graded as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong), and the percentage of positive cells was recorded at each intensity. The final H-score was calculated using the formula: $$\:\varvec{H}\text{-}\varvec{s}\varvec{c}\varvec{o}\varvec{r}\varvec{e}={\sum\:}_{\varvec{i}=0}^{3}\left(\varvec{i}\times\:\text{percentage\:of\:positive\:cells\:at\:intensity\:}\varvec{i}\right)$$ yielding a range from 0 to 300. H-scores between CCA and normal tissues were compared to determine differential expression of WNT5A and PFN2. Statistical Analysis All statistical procedures were performed using R software (v4.2.1). The “limma” package was used for differential expression analysis, “clusterProfiler” for functional enrichment, “GSVA” for ssGSEA, and “MCPcounter” for immune cell estimation. P-values were adjusted for multiple testing using the Benjamini–Hochberg method, and adjusted p (FDR) < 0.05 was regarded as statistically significant. Continuous variables were presented as mean ± standard deviation (SD) or median with interquartile range (IQR), depending on distribution normality. Group comparisons between CCA and controls were conducted using the Student’s t-test or Wilcoxon rank-sum test as appropriate. Categorical variables were compared with the χ² test or Fisher’s exact test. Spearman’s correlation was applied to evaluate associations between gene expression and immune infiltration scores, and results were visualized using the “ggplot2” package. For model development, univariate logistic regression and LASSO regression were used to select key genes, with model performance assessed via ROC analysis and AUC values using the “pROC” package. Model calibration was evaluated using calibration plots, and its clinical applicability was examined by decision curve analysis (DCA). For molecular docking, AutoDock Vina calculated binding affinities, and docking poses were visualized in PyMOL. Unless otherwise specified, p < 0.05 was considered statistically significant. Study Workflow To provide an overview of the analytical strategy, we first summarized the study design and workflow (Fig. 1 ). Based on this workflow, the subsequent results are presented in the following sections, including the identification of exosome-related differentially expressed genes (ERDEGs), construction of a diagnostic model, immune infiltration analysis, drug enrichment and molecular docking, and immunohistochemical validation in clinical samples. 3 Results Data Collection and Correction Following the normalization and integration of the two cholangiocarcinoma datasets into the combined gene dataset (CGD), surrogate variable analysis (SVA) was applied to remove hidden batch effects. Principal component analysis (PCA) plots (Fig. 2 ) demonstrated that the batch effects among samples were substantially reduced after correction, confirming the reliability of the integrated dataset. Exosome-Related Differentially Expressed Genes in Cholangiocarcinoma The identified differentially expressed genes (DEGs) were visualized in a heatmap (Fig. 3 A). A total of 443 DEGs were identified (|logFC| > 0; p 0) and 225 downregulated genes (logFC < 0). These results were further illustrated using a volcano plot (Fig. 3 B).By intersecting the DEGs with a curated list of exosome-related genes (ERGs) [ 32 ], 42 exosome-related differentially expressed genes (ERDEGs) were identified (Table S2). The overlap was visualized using a Venn diagram (Fig. 3 C). GO enrichment analysis The results (Fig. 4 , Table S3) showed that ERDEGs were enriched in biological process (BP), cellular component (CC), and molecular function (MF) terms. In the BP group, enriched terms mainly involved negative regulation of cell migration, locomotion, and epithelial cell development. In the CC group, the top terms were membrane raft, focal adhesion, and collagen-containing extracellular matrix. In the MF group, ERDEGs were linked to integrin binding, cadherin binding, actin binding, and oxidoreductase activity. These findings suggest that ERDEGs in cholangiocarcinoma are mainly associated with pathways regulating cell migration, adhesion, signal transduction, and cytoskeletal remodeling [ 33 ]. These functions respectively correspond to key biological processes such as tumor invasion and metastasis, modulation of the tumor microenvironment, and cytoskeletal dynamics. Collectively, the results suggest that these ERDEGs may promote cholangiocarcinoma progression through exosome-mediated intercellular signaling. Gene set enrichment analysis (GSEA) The GSEA results(Table S4) indicated that the pathways “Cytoskeleton in Muscle Cells,” “Protein Digestion and Absorption,” and “Viral Protein Interaction with Cytokine and Cytokine Receptor” were significantly downregulated in the CCA group compared to the healthy controls. The normalized enrichment scores (NES) for these pathways were − 1.98, − 1.83, and − 1.80, respectively, all of which were statistically significant (P < 0.001, adjusted P ≤ 0.04)(Fig. 5 .). These findings suggest that the aforementioned biological processes may be suppressed in patients with CCA, potentially correlating with disease progression or therapeutic response. Diagnostic model for CCA 42 ERDEGs were statistically significant (P < 0.05; Table S2). Univariate logistic regression was used to assess the diagnostic value of 42 ERDEGs in CCA, and 3 genes showed statistical significance (P < 0.05, Table 1 ).Based on these results, a LASSO regression model was constructed to develop a diagnostic signature for CCA, with visualization provided by the LASSO regression plot (Fig. 6 A) and the coefficient profile plot (Fig. 6 B). In the cross-validation plot (binomial deviance vs. log λ), the optimal λ was determined, and the model at this λ retained two genes. The coefficient profile plot illustrates the variation in the regression coefficients of each gene across a range of λ values. Consequently, two genes, WNT5A and PFN2, were identified as key genes for subsequent analysis. Table 1 Univariate Logistic Regression Analysis of ERDEGs for Diagnostic Value in CCA. OR OR.95L OR.95H P value WNT5A 5.64 × 10⁻² 9.92 × 10⁻⁴ 4.61 × 10⁻¹ 4.45 × 10⁻² PFN2 7.18 × 10⁻² 1.45 × 10⁻³ 4.82 × 10⁻¹ 4.58 × 10⁻² BASP1 4.48 × 10⁻¹ 1.46 × 10⁻¹ 8.63 × 10⁻¹ 4.86 × 10⁻² Diagnostic model validation To assess key gene expression differences, boxplots compared control and CCA samples. WNT5A and PFN2 showed markedly reduced expression in CCA tissues relative to controls (P < 0.05). This finding suggests a downregulation trend of these two genes under CCA conditions, indicating their potential suppressive roles in the pathogenesis of CCA (Fig. 7 A). Notably, when examining the chromosomal locations of these key genes, the constructed Circos plot (Fig. 7 B) showed that WNT5A and PFN2 are both located on chromosome 3, further indicating that these genes may exert synergistic effects during the occurrence and progression of CCA. To validate the CCA diagnostic model constructed based on the two key ERDEGs, ROC curves (Fig. 7 C) and a nomogram (Fig. 7 D) were generated. The ROC analysis demonstrated that WNT5A (AUC = 0.857) and PFN2 (AUC = 0.841) both had AUC values greater than 0.8, indicating good diagnostic performance of these genes in the CCA diagnostic model. The effectiveness of the key genes (WNT5A and PFN2) in the diagnostic model was significantly higher than that of other variables. Finally, to assess the clinical applicability and predictive stability of the diagnostic model, a calibration curve (Fig. 7 E) and a decision curve analysis (DCA) plot (Fig. 7 F) were generated. The calibration curve was used to evaluate the model’s ability to predict actual outcomes under different conditions. Although the calibration curve showed slight deviations from the ideal diagonal, it demonstrated an overall good fit, indicating reliable predictive performance. The DCA plot was used to evaluate the clinical utility of the diagnostic model, showing that the model curve remained stable within a certain threshold range and was higher than the strategies of treating all cases as positive or negative, indicating that the model offers high net benefits and demonstrates favorable clinical applicability. Immune infiltration analysis (ssGSEA and MCPCounter) Immune cell infiltration was evaluated in the integrated dataset (CGD) using the ssGSEA algorithm, yielding infiltration scores for 28 immune cell types (Table S5). Boxplot analysis revealed significant differences (P < 0.05) in 18 immune cell types between the CCA and normal groups, including activated CD4 T cells, activated dendritic cells, CD56dim natural killer cells, central memory CD4 T cells, central memory CD8 T cells, eosinophils, immature dendritic cells, macrophages, mast cells, MDSCs, memory B cells, natural killer T cells, neutrophils, plasmacytoid dendritic cells, regulatory T cells, follicular helper T cells, Th1 cells, and Th17 cells (Fig. 8 A). Correlation analysis using the MCPcounter algorithm demonstrated that WNT5A was significantly positively correlated with activated CD8 T cells, CD56dim natural killer cells, central memory CD4 T cells, macrophages, mast cells, monocytes, regulatory T cells, and Th17 cells (P 0.5), but negatively correlated with activated CD4 T cells, central memory CD8 T cells, memory B cells, and type 2 T helper cells (P < 0.05, r < − 0.5). PFN2 exhibited a similar correlation profile, with minor differences (Fig. 8 B). (Correlation is represented by r.) Combined with our previous differential expression analysis showing significant downregulation of WNT5A and PFN2 in CCA patients, we hypothesize that these genes may act as tumor suppressors (protective genes) in cholangiocarcinoma. Immune infiltration analysis using MCPcounter revealed that immune cell types positively correlated with WNT5A and PFN2 expression, including activated CD8 T cells, CD56dim natural killer cells, central memory CD4 T cells, macrophages, mast cells, monocytes, and regulatory T cells, which are generally associated with antitumor immune responses. Their positive correlation with protective genes suggests their potential roles in immune surveillance and tumor clearance in CCA. Conversely, reduced WNT5A and PFN2 expression in CCA was associated with increased activity of several immune cell types, such as activated CD4 T cells, central memory CD8 T cells, memory B cells, and type 2 helper T cells. This alteration may drive immune tolerance or evasion, facilitating tumor growth. These results suggest that WNT5A and PFN2 modulate the CCA immune microenvironment by affecting immune cell infiltration and composition, thus contributing to disease development. GeneMANIA The potential interaction gene network constructed with WNT5A as the core is shown in Fig. 11 A. Functional enrichment analysis of the WNT5A-associated gene network revealed significant overrepresentation of pathways related to non-canonical Wnt signaling, planar cell polarity, and G protein-coupled receptor binding (FDR < 1e-12). These findings support the role of WNT5A in regulating cell polarity, tissue morphogenesis, and signal transduction, consistent with its known involvement in tumor progression and developmental processes[ 34 – 38 ] (Table S6). The constructed gene interaction network, with PFN2 at its nexus, is depicted in Fig. 11 B. The most significantly enriched functional pathway in its associated gene network is cytoskeletal regulation. Genes such as PFN1, PFN3, PFN4, VASP, and FHOD1 not only interact with PFN2 but are also co-expressed with it, all of which are closely involved in cytoskeletal remodeling processes(Table S7). Drug Enrichment Analysis and Molecular Docking A total of 16 significantly associated compounds were identified for the two key genes. Among them, 14 compounds were significantly linked to WNT5A, while pentadecafluorooctanoic acid showed a significant association exclusively with PFN2. Notably, 3,3′,4,4′,5-pentachlorobiphenyl was significantly associated with both WNT5A and PFN2(Table 2 ). Among the enriched results, 3,3',4,4',5-Pentachlorobiphenyl showed the smallest p-value and was enriched in both WNT5A and PFN2. However, considering its environmental pollutant nature and practical limitations, the focus was shifted to drug candidates with established clinical safety profiles and known antitumor potential, including gemcitabine [ 39 – 42 ] and temozolomide (43–48) (P < 0.03), both of which were enriched with WNT5A in the analysis. The identification of both gemcitabine and temozolomide as significantly enriched compounds associated with WNT5A suggests a potential shared mechanism beyond their canonical roles in DNA damage. Given WNT5A's involvement in non-canonical Wnt signaling pathways regulating cytoskeletal dynamics and cell motility[ 36 , 37 , 49 ], it is plausible that these compounds may exert antitumor effects, at least in part, by modulating WNT5A-mediated signaling[ 34 , 35 ]. This interaction could impair cellular processes critical for cholangiocarcinoma progression, such as epithelial–mesenchymal transition (EMT), invasion, and metastasis. Further experimental validation is warranted to elucidate whether these drugs directly interfere with WNT5A function or its downstream pathways. To further evaluate the binding affinity of these candidate compounds to WNT5A, gemcitabine and temozolomide(p.adjust < 0.04)were subjected to molecular docking analysis, examining their binding modes and molecular interactions with CCA.The molecular docking results confirmed that these compounds can potentially bind to WNT5A and modulate its function. In Figure.9A, gemcitabine demonstrates a deep binding conformation within the hydrophobic cavity of WNT5A, forming multiple hydrogen bonds and occupying a stable spatial niche, indicating a relatively high binding affinity and potential to modulate WNT5A function.In contrast, Figure.9B shows that temozolomide interacts more superficially with WNT5A, with fewer stabilizing interactions and a looser binding pattern, suggesting lower affinity and possibly limited direct regulatory capacity. But their Vina docking score are similar(Fig. 10 ). Table 2 Compounds significantly associated with the key genes WNT5A and PFN2. Compound (Description) P -value Adjusted P -value Gene Count 3,3',4,4',5-Pentachlorobiphenyl < 0.001 0.002 WNT5A/PFN2 2 p-benzoquinone 0.005 0.029 WNT5A 1 DL-Homocysteine 0.006 0.029 WNT5A 1 NSC94017 0.008 0.029 WNT5A 1 isoflupredone 0.009 0.029 WNT5A 1 hydroquinone 0.021 0.038 WNT5A 1 Pentadecafluorooctanoic acid 0.021 0.038 PFN2 1 2,2',4,4',5,5'-Hexachlorobiphenyl 0.022 0.038 WNT5A 1 [6-[6-(butanoylamino)purin-9-yl]-2-hydroxy-2-oxo-4a,6,7,7a-tetrahydro-4H-furo[3,2-d][ 1 , 3 , 2 ]dioxaphosphinin-7-yl] butanoate 0.024 0.038 WNT5A 1 gemcitabine 0.027 0.04 WNT5A 1 temozolomide 0.036 0.04 WNT5A 1 folic acid 0.037 0.04 WNT5A 1 mitomycin C 0.039 0.04 WNT5A 1 5-azacytidine 0.04 0.04 WNT5A 1 Medroxyprogesterone acetate 0.04 0.04 WNT5A 1 mifepristone < 0.001 0.002 WNT5A 1 Immunohistochemistry Validation of Key Genes To validate the two exosome-related genes identified by bioinformatics, immunohistochemistry (IHC) was performed on paraffin-embedded CCA tissues and matched adjacent normal bile ducts. Both WNT5A and PFN2 proteins were predominantly localized in the cytoplasm, with partial distribution along the cell membrane, consistent with their functional roles in intracellular signaling and cytoskeletal regulation. As shown in Fig. 12 and Fig. 13 , adjacent normal bile duct tissues ex-hibited strong brownish staining for WNT5A and PFN2, whereas CCA tissues displayed markedly weaker staining intensity. Semi-quantitative evaluation using the H-score method revealed that WNT5A expression was significantly reduced in CCA tissues (mean H-score: 17.89 ± 2.09) com-pared with adjacent normal tissues (mean H-score: 38.36 ± 6.29; P < 0.05, Fig. 12 B). Similarly, PFN2 expression was markedly lower in CCA tissues (mean H-score: 24.53 ± 2.61) than in normal tissues (mean H-score: 59.86 ± 5.24; P < 0.01, Fig. 13 B). In addition, quantitative image analysis was per-formed using FIJI software, and the proportions of strongly positive, positive, weakly positive, and negative cells, together with the mean optical density (Mean OD) values, are summarized in the supplementary materials (Table S8, S9). These findings are consistent with the transcriptome analysis, confirming that WNT5A and PFN2 are downregulated at the protein level in CCA. Col-lectively, these results support their potential roles as tumor suppressors and highlight their diag-nostic value as candidate biomarkers in cholangiocarcinoma. 4 Discussion Cholangiocarcinoma (CCA) is a highly aggressive malignancy arising from the biliary epithelium, characterized by poor prognosis, limited therapeutic options, and high molecular heterogeneity(1, 50). Despite advances in surgery and chemotherapy, early diagnosis and effective treatment remain major challenges[ 51 ]. Exosomes, as key mediators of intercellular communication, have gained increasing attention for their ability to transfer proteins, lipids, and nucleic acids, thereby influencing tumor proliferation, metastasis, immune evasion, and drug resistance[ 32 , 52 ]. In CCA, exosomes have been implicated in the regulation of oncogenic pathways, including Wnt/β-catenin, PI3K/AKT, and TGF-β signaling[ 53 ], and show promise as stable, non-invasive biomarkers[ 54 ]. However, the specific exosome-associated genes driving CCA progression and their diagnostic potential remain poorly defined. In this study, integrated bioinformatics analyses identified 42 exosome-related differentially expressed genes (ERDEGs) in CCA. Functional enrichment indicated strong associations with cytoskeletal organization, cell adhesion, migration, and membrane-related pathways. Using univariate logistic and LASSO regression, we constructed a diagnostic model and identified two key downregulated genes—WNT5A and PFN2—both demonstrating robust diagnostic performance across validation analyses. Notably, both genes reside on chromosome 3p14 and exhibited highly overlapping expression patterns and immune correlation profiles, suggesting potential functional synergy in CCA biology. WNT5A, a prototypical non-canonical Wnt ligand, regulates cell polarity, migration, and adhesion through pathways such as Wnt/PCP and Wnt/Ca²⁺, acting upstream of cytoskeletal remodeling by activating RhoA, Rac1, and related signaling molecules [ 34 – 36 ]. PFN2, a profilin family member, plays a direct role in actin cytoskeleton dynamics, influencing membrane tension and cell motility through its interactions with actin-binding partners [ 37 , 55 ]. While WNT5A primarily transduces extracellular signals, PFN2 executes the structural reorganization of the cytoskeleton. The co-downregulation of these genes in CCA suggests that WNT5A may exert part of its effects via PFN2-dependent cytoskeletal remodeling, forming a coordinated WNT5A–PFN2 axis that integrates signal transduction with structural regulation. This axis could influence exosome-mediated communication, tumor cell invasiveness, and responsiveness to microenvironmental cues. Immune infiltration profiling revealed profound alterations in the CCA microenvironment [ 29 ], with significant differences in 18 immune cell types between tumor and normal tissues. Both WNT5A and PFN2 expression correlated positively with cytotoxic and regulatory immune subsets—such as activated CD8⁺ T cells, NK cells, macrophages, and regulatory T cells—typically linked to antitumor immunity [ 56 , 57 ]. Conversely, negative correlations were observed with immune populations associated with tumor-promoting phenotypes, including memory B cells and Th2 cells[ 58 ]. These findings suggest that downregulation of WNT5A and PFN2 may weaken immune surveillance and facilitate immune evasion, highlighting their potential roles as both diagnostic biomarkers and immune modulators. Given the growing interest in immunotherapy for CCA, these results provide a rationale for exploring strategies to restore WNT5A and PFN2 expression or function to enhance antitumor immunity. Drug enrichment analysis identified 16 compounds significantly associated with the key genes, among which gemcitabine and temozolomide were prioritized for their established clinical use and safety [ 39 , 48 ]. Traditionally considered DNA-damaging agents, both drugs were enriched in WNT5A-related pathways, suggesting additional roles in modulating non-canonical Wnt signaling. Molecular docking supported this hypothesis: gemcitabine exhibited deep hydrophobic pocket binding within WNT5A with multiple hydrogen bonds, while temozolomide demonstrated weaker, surface-level binding. Comparable docking scores suggest that both may modulate WNT5A function, potentially impairing EMT, metastasis, and immune-related processes. These findings underscore the translational potential of drug repurposing in CCA [ 59 ], offering a feasible approach to target WNT5A without the need for novel drug development. Immunohistochemistry (IHC) further validated the reduced protein expression of WNT5A and PFN2 in CCA tissues relative to normal bile ducts, consistent with transcriptomic data. Histopathological analysis revealed that low PFN2 expression often coincided with disorganized cellular architecture, suggestive of enhanced migratory capacity, while low WNT5A expression was associated with epithelial–mesenchymal transition (EMT)-like features. These observations provide histological evidence linking gene downregulation to functional tumor behaviors, strengthening the credibility of our bioinformatic findings [ 60 ]. This study offers several innovations: integration of multiple datasets to enhance robustness; an exosome-centered approach to gene discovery; construction and validation of a two-gene diagnostic model; and combination of drug enrichment with molecular docking to identify clinically available agents with potential CCA relevance. The inclusion of experimental validation at both mRNA and protein levels forms a complete discovery-to-validation pipeline, enhancing translational relevance. Nonetheless, limitations remain: reliance on public datasets with limited sample sizes, the need for mechanistic studies to dissect the WNT5A–PFN2 axis in cytoskeletal and immune regulation, and the necessity for in vitro and in vivo confirmation of predicted drug–target interactions. Future studies should expand clinical cohorts, explore the therapeutic restoration of WNT5A and PFN2 function, and evaluate the efficacy of repurposed agents in preclinical CCA models. 5 Conclusions This study identified two exosome-related genes, WNT5A and PFN2, as consistently downregulated in cholangiocarcinoma and demonstrated their strong diagnostic value through integrative bioinformatics and experimental validation. Functional analyses revealed that their loss is closely linked to impaired cytoskeletal remodeling, enhanced tumor migration and EMT-like features, as well as dysregulated immune infiltration, highlighting their dual roles in structural and immune regulation. Drug enrichment and molecular docking further indicated gemcitabine and temozolomide as promising therapeutic agents potentially targeting WNT5A. Together, these findings establish WNT5A and PFN2 as robust diagnostic biomarkers and potential therapeutic targets, providing a translational framework for improving early diagnosis and drug repurposing strategies in cholangiocarcinoma. Declarations Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Funding This research was funded by the Shandong Provincial Health Commission General Program, grant number 202304010282. Author Contribution Conceptualization, F.L.; methodology, Q.W. and Y.W.; software, Q.W.; validation, Z.W., Q.W. and Y.W.; formal analysis, Q.W.; investigation, Z.W. and H.C.; resources, F.L.; data curation, Q.W.; writing—original draft preparation, Z.W.; writing—review and editing, F.L.; visualization, Q.W.; supervision, F.L.; project administration, F.L.; funding acquisition, F.L. All authors have read and agreed to the published version of the manuscript. Acknowledgments We are grateful to the technical staff and colleagues at Shandong First Medical University for their helpful support throughout this research. Data Availability The datasets analyzed for this study are publicly available in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). The accession numbers are GSE26566 and GSE107943. All data utilized in this research are open-access and can be freely obtained from the GEO repository. Additional information is available from the corresponding author upon reasonable request. References Banales, J. M. et al. Cholangiocarcinoma 2020: the next horizon in mechanisms and management. Nat. Rev. Gastroenterol. Hepatol. 17 , 557–588. https://doi.org/10.1038/s41575-020-0310-z (2020). Dixon, M. E., Williams, M., Pappas, S. G. & Cholangiocarcinoma Cancer Treat. Res. 192 , 165–184 ; https://doi.org/10.1007/978-3-031-61238-1_9 (2024). Kidanemariam, S. et al. Cholangiocarcinoma: epidemiology and imaging-based review. R. I. Med. J. 107, 43–48 (2024). (2013). Harrison, J. M., Visser, B. C. & Cholangiocarcinoma Surg. Clin. North. Am. 104 , 1281–1293 ; https://doi.org/10.1016/j.suc.2024.04.003 (2024). Sarcognato, S. et al. Cholangiocarcinoma Pathologica 113 , 158–169 ; https://doi.org/10.32074/1591-951X-252 (2021). Vithayathil, M. & Khan, S. A. Current epidemiology of cholangiocarcinoma in Western countries. J. Hepatol. 77 , 1690–1698. https://doi.org/10.1016/j.jhep.2022.07.022 (2022). Qurashi, M., Vithayathil, M. & Khan, S. A. Epidemiology of cholangiocarcinoma. Eur . J. Surg. Oncol. 51 , 107064. https://doi.org/10.1016/j.ejso.2023.107064 (2025). A, N. V. L. K., Subramanian, S. & Ramakrishna, S. Therapeutic applications of exosomes in various diseases: a review. Biomater. Adv. 134 , 112579. https://doi.org/10.1016/j.msec.2021.112579 (2022). Jafari, A., Babajani, A., Abdollahpour-Alitappeh, M., Ahmadi, N. & Rezaei-Tavirani, M. Exosomes and cancer: from molecular mechanisms to clinical applications. Med. Oncol. 38 , 45. https://doi.org/10.1007/s12032-021-01491-0 (2021). Kimiz-Gebologlu, I. & Oncel, S. S. Exosomes: large-scale production, isolation, drug loading efficiency, and biodistribution and uptake. J. Control Release . 347 , 533–543. https://doi.org/10.1016/j.jconrel.2022.05.027 (2022). Cao, L., Hong, J. & Wu, J. Potential of extracellular vesicles and exosomes as diagnostic markers for cholangiocarcinoma. Hepatobiliary Surg. Nutr. 11 , 436–438. https://doi.org/10.21037/hbsn-2022-02 (2022). Fabris, L., Sato, K., Alpini, G. & Strazzabosco, M. The tumor microenvironment in cholangiocarcinoma progression. Hepatology 73 Suppl. 1 , 75–85. https://doi.org/10.1002/hep.31410 (2021). Lapitz, A. et al. Liquid biopsy-based protein biomarkers for risk prediction, early diagnosis, and prognostication of cholangiocarcinoma. J. Hepatol. 79 , 93–108. https://doi.org/10.1016/j.jhep.2023.02.027 (2023). Luo, C. et al. Tumor-derived exosomes induce immunosuppressive macrophages to foster intrahepatic cholangiocarcinoma progression. Hepatology 76 , 982–999. https://doi.org/10.1002/hep.32387 (2022). Shu, L. et al. Bile exosomal miR-182/183-5p increases cholangiocarcinoma stemness and progression by targeting HPGD and increasing PGE2 generation. Hepatology 79 , 307–322. https://doi.org/10.1097/hep.0000000000000437 (2024). Zhang, M. et al. Single-cell transcriptomic architecture and intercellular crosstalk of human intrahepatic cholangiocarcinoma. J. Hepatol. 73 , 1118–1130. https://doi.org/10.1016/j.jhep.2020.05.039 (2020). Zhao, K. et al. Exosomes in the tumor microenvironment of cholangiocarcinoma: current status and future perspectives. J. Transl Med. 20 , 117. https://doi.org/10.1186/s12967-022-03294-x (2022). Wang, J. et al. Exosome-derived miR-182-5p promoted cholangiocarcinoma progression and vasculogenesis by regulating ADK/SEMA5A/PI3K pathway. Liver Int. 44 , 370–388. https://doi.org/10.1111/liv.15773 (2024). Jiang, T. Y. et al. PTEN deficiency facilitates exosome secretion and metastasis in cholangiocarcinoma by impairing TFEB-mediated lysosome biogenesis. Gastroenterology 164 , 424–438. https://doi.org/10.1053/j.gastro.2022.11.025 (2023). Edgar, R., Domrachev, M. & Lash, A. E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res. 30 , 207–210. https://doi.org/10.1093/nar/30.1.207 (2002). Leek, J. T., Johnson, W. E., Parker, H. S., Jaffe, A. E. & Storey, J. D. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28 , 882–883. https://doi.org/10.1093/bioinformatics/bts034 (2012). Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43 , e47. https://doi.org/10.1093/nar/gkv007 (2015). Ashburner, M. et al. Gene ontology: tool for the unification of biology. Nat. Genet. 25 , 25–29. https://doi.org/10.1038/75556 (2000). Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 16 , 284–287. https://doi.org/10.1089/omi.2011.0118 (2012). Glickman, M. E., Rao, S. R. & Schultz, M. R. False discovery rate control is a recommended alternative to Bonferroni-type adjustments in health studies. J. Clin. Epidemiol. 67 , 850–857. https://doi.org/10.1016/j.jclinepi.2014.03.012 (2014). Alhamzawi, R. & Ali, H. T. M. The Bayesian adaptive lasso regression. Math. Biosci. 303 , 75–82. https://doi.org/10.1016/j.mbs.2018.06.004 (2018). Kucheryavskiy, S., Rodionova, O. & Pomerantsev, A. Procrustes cross-validation of multivariate regression models. Anal. Chim. Acta . 1255 , 341096. https://doi.org/10.1016/j.aca.2023.341096 (2023). Vickers, A. J., Van Calster, B. & Steyerberg, E. W. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ 352 , i6. https://doi.org/10.1136/bmj.i6 (2016). Becht, E. et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 17 , 218. https://doi.org/10.1186/s13059-016-1070-5 (2016). Cotto, K. C. et al. DGIdb 3.0: a redesign and expansion of the drug-gene interaction database. Nucleic Acids Res. 46 , D1068–D1073. https://doi.org/10.1093/nar/gkx1143 (2018). Ravindranath, P. A., Forli, S., Goodsell, D. S., Olson, A. J. & Sanner, M. F. AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility. PLoS Comput. Biol. 11 , e1004586. https://doi.org/10.1371/journal.pcbi.1004586 (2015). Kalluri, R. & LeBleu, V. S. The biology, function, and biomedical applications of exosomes. Science 367 , eaau6977. https://doi.org/10.1126/science.aau6977 (2020). Hoshino, A. et al. Tumour exosome integrins determine organotropic metastasis. Nature 527 , 329–335. https://doi.org/10.1038/nature15756 (2015). Tufail, M. & Wu, C. WNT5A: a double-edged sword in colorectal cancer progression. Mutat. Res. Rev. Mutat. Res. 792 , 108465. https://doi.org/10.1016/j.mrrev.2023.108465 (2023). Bueno, M. L. P., Saad, S. T. O. & Roversi, F. M. WNT5A in tumor development and progression: a comprehensive review. Biomed. Pharmacother . 155 , 113599. https://doi.org/10.1016/j.biopha.2022.113599 (2022). Astudillo, P. An emergent Wnt5a/YAP/TAZ regulatory circuit and its possible role in cancer. Semin Cell. Dev. Biol. 125 , 45–54. https://doi.org/10.1016/j.semcdb.2021.10.001 (2022). Zhou, W. et al. Wnt5a: a promising therapeutic target in ovarian cancer. Pathol. Res. Pract. 219 , 153348. https://doi.org/10.1016/j.prp.2021.153348 (2021). Liu, Q. et al. Wnt5a-induced M2 polarization of tumor-associated macrophages via IL-10 promotes colorectal cancer progression. Cell. Commun. Signal. 18 , 51. https://doi.org/10.1186/s12964-020-00557-2 (2020). Natu, J. & Nagaraju, G. P. Gemcitabine effects on tumor microenvironment of pancreatic ductal adenocarcinoma: special focus on resistance mechanisms and metronomic therapies. Cancer Lett. 573 , 216382. https://doi.org/10.1016/j.canlet.2023.216382 (2023). Chitty, J. L. et al. A first-in-class pan-lysyl oxidase inhibitor impairs stromal remodeling and enhances gemcitabine response and survival in pancreatic cancer. Nat. Cancer . 4 , 1326–1344. https://doi.org/10.1038/s43018-023-00614-y (2023). Pandit, B. & Royzen, M. Recent development of prodrugs of gemcitabine. Genes (Basel) . 13 , 466. https://doi.org/10.3390/genes13030466 (2022). Miao, H., Chen, X. & Luan, Y. Small molecular gemcitabine prodrugs for cancer therapy. Curr. Med. Chem. 27 , 5562–5582. https://doi.org/10.2174/0929867326666190816230650 (2020). Dutra, J. A. P. et al. Temozolomide: an overview of biological properties, drug delivery nanosystems, and analytical methods. Curr. Pharm. Des. 28, 2073–; (2088). https://doi.org/10.2174/1381612828666220603152918 (2022). Zou, Y. et al. Brain co-delivery of temozolomide and cisplatin for combinatorial glioblastoma chemotherapy. Adv. Mater. 34 , e2203958. https://doi.org/10.1002/adma.202203958 (2022). Rao, V. et al. Temozolomide resistance: a multifarious review on mechanisms beyond O-6-methylguanine-DNA methyltransferase. CNS Neurol. Disord Drug Targets . 22 , 817–831. https://doi.org/10.2174/1871527321666220404180944 (2023). Yu, Z. C. et al. Temozolomide sensitizes ARID1A-mutated cancers to PARP inhibitors. Cancer Res. 83 , 2750–2762. https://doi.org/10.1158/0008-5472.CAN-22-3646 (2023). Hui-Zheng, H. & Lei, L. Temozolomide for aggressive pituitary tumors: a literature review. Asian J. Surg. 47 , 1300–1301. https://doi.org/10.1016/j.asjsur.2023.11.106 (2024). Jezierzański, M. et al. Temozolomide (TMZ) in the treatment of glioblastoma multiforme: a literature review and clinical outcomes. Curr. Oncol. 31 , 3994–4002. https://doi.org/10.3390/curroncol31070296 (2024). Prasad, C. P., Chaurasiya, S. K., Guilmain, W. & Andersson, T. WNT5A signaling impairs breast cancer cell migration and invasion via mechanisms independent of the epithelial-mesenchymal transition. J. Exp. Clin. Cancer Res. 35 , 144. https://doi.org/10.1186/s13046-016-0421-0 (2016). Ilyas, S. I., Khan, S. A., Hallemeier, C. L., Kelley, R. K. & Gores, G. J. Cholangiocarcinoma: evolving concepts and therapeutic strategies. Nat. Rev. Clin. Oncol. 15 , 95–111. https://doi.org/10.1038/nrclinonc.2017.157 (2018). Andreyev, J. et al. Guidance on the management of diarrhoea during cancer chemotherapy. Lancet Oncol. 15 , e447–e460. https://doi.org/10.1016/S1470-2045(14)70006-3 (2014). Altorki, N. K. et al. The lung microenvironment: an important regulator of tumour growth and metastasis. Nat. Rev. Cancer . 19 , 9–31. https://doi.org/10.1038/s41568-018-0081-9 (2019). Liao, G. et al. Bevacizumab treatment of radiation-induced brain necrosis: a systematic review. Front. Oncol. 11 , 593449. https://doi.org/10.3389/fonc.2021.593449 (2021). Thery, C. et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J. Extracell. Vesicles . 7 , 1535750. https://doi.org/10.1080/20013078.2018.1535750 (2018). Nam, S., Chung, J. W. & Yang, J. Y. WNT5A correlates with clinicopathological characteristics in gastric cancer: a meta-analysis. Cell. Physiol. Biochem. 41 , 33–40. https://doi.org/10.1159/000455934 (2017). Soerens, A. G. et al. Functional T cells are capable of supernumerary cell division and longevity. Nature 614 , 762–766. https://doi.org/10.1038/s41586-022-05626-9 (2023). Rui, R., Zhou, L. & He, S. Cancer immunotherapies: advances and bottlenecks. Front. Immunol. 14 , 1212476. https://doi.org/10.3389/fimmu.2023.1212476 (2023). Gentles, A. J. et al. The prognostic landscape of genes and infiltrating immune cells across human cancers. Nat. Med. 21 , 938–945. https://doi.org/10.1038/nm.3909 (2015). Pushpakom, S. et al. Drug repurposing: progress, challenges and recommendations. Nat. Rev. Drug Discov . 18 , 41–58. https://doi.org/10.1038/nrd.2018.168 (2019). Longuespee, R. et al. MALDI mass spectrometry imaging: a cutting-edge tool for fundamental and clinical histopathology. Proteom. Clin. Appl. 10 , 701–719. https://doi.org/10.1002/prca.201500140 (2016). Additional Declarations No competing interests reported. Supplementary Files SupplementaryTable.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8144185","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":560796296,"identity":"a9b2a27e-076b-4a5e-879c-aec8815d98ec","order_by":0,"name":"Fangfeng Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIiWNgGAWjYBACPmYGhg9Amh+IGR8wsMmBRSXwaWFjZmCcAaQlGxgYmA0Y2IyJ0MKA0MImQZwWdh7Dhp87aiX4pduvVf4oM4g2OMB88DYPg10ebofxGDb2njkuITnnTNltnnMGuRsOsCVb8zAkF+PRYv6At+1YncGNnLTbjG1/gFp4zKR5GA4kNuCz5W/bMQl7oJbCn20gW/i/EdTSzNtWI2EgkX6MgReshYeNgBa2wmbZtgMSEjdymKVBfpl5mM3Yco5BMk4t/PyHNza+bauT4J+R/vAjMMRy+443P7zxpsIOpxYoOAzEPAYQNjOIMMCvHgjqgJj9AUFlo2AUjIJRMDIBAChxU/3/KgpsAAAAAElFTkSuQmCC","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":true,"prefix":"","firstName":"Fangfeng","middleName":"","lastName":"Liu","suffix":""},{"id":560796297,"identity":"c9d793de-5e49-465c-8b14-bc9ffd399a2b","order_by":1,"name":"Zhe Wang","email":"","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhe","middleName":"","lastName":"Wang","suffix":""},{"id":560796298,"identity":"a9f87e1a-455a-415b-9bff-a090cedd5bed","order_by":2,"name":"Qianchang Wang","email":"","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qianchang","middleName":"","lastName":"Wang","suffix":""},{"id":560796299,"identity":"a0e22f78-62c6-4aca-babc-6ab79f486ace","order_by":3,"name":"Yu Wang","email":"","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Wang","suffix":""},{"id":560796300,"identity":"a7049266-292a-4a61-b1ec-5bf71c511fd2","order_by":4,"name":"Zhengjian Wang","email":"","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhengjian","middleName":"","lastName":"Wang","suffix":""},{"id":560796302,"identity":"973220d2-8c7f-4706-b232-22833b93624b","order_by":5,"name":"Hong Chang","email":"","orcid":"","institution":"Shandong First Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Chang","suffix":""}],"badges":[],"createdAt":"2025-11-18 10:23:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8144185/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8144185/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98440668,"identity":"87442eca-82f9-4d48-af83-8e56b0afd686","added_by":"auto","created_at":"2025-12-17 17:04:10","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63448,"visible":true,"origin":"","legend":"","description":"","filename":"SRManuscriptExosomeCCAV2.docx","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/f0ce7c03f50985faee84c3c4.docx"},{"id":98370510,"identity":"701e190d-44d0-468b-b10b-ea600f390211","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133941,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/8f33c5e3d97cdca1e2b6fbc4.tif"},{"id":98439410,"identity":"922ee787-8575-4acc-9a1c-0d44e935e951","added_by":"auto","created_at":"2025-12-17 17:01:48","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4517416,"visible":true,"origin":"","legend":"","description":"","filename":"Fig10.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/b46b039db801fe688aa77d83.tif"},{"id":98370518,"identity":"123b570b-5b6c-4f63-8bc0-b09a9d625e4b","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":330424,"visible":true,"origin":"","legend":"","description":"","filename":"Fig11.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/9570d66053fb7330ea6febfc.tif"},{"id":98439615,"identity":"d34eb2ae-8bbe-4214-8814-418263e6ff23","added_by":"auto","created_at":"2025-12-17 17:02:13","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10701,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/9c9ee955f1a55ba1c4729719.xlsx"},{"id":98440085,"identity":"802067d1-fc50-4d08-9f76-98ef12b90945","added_by":"auto","created_at":"2025-12-17 17:03:16","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3733260,"visible":true,"origin":"","legend":"","description":"","filename":"Fig12.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/c0daa8c64ffb2cfb525b5c54.tif"},{"id":98440574,"identity":"a19e22e6-aadf-454a-8422-aa013e7a60fb","added_by":"auto","created_at":"2025-12-17 17:04:01","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11236,"visible":true,"origin":"","legend":"","description":"","filename":"Table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/b26f4bd17750450dd44f831d.xlsx"},{"id":98370532,"identity":"54a00cee-8691-4cd8-accc-7b958f1c8fc2","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3326400,"visible":true,"origin":"","legend":"","description":"","filename":"Fig13.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/86d75885a8172182b72e35e7.tif"},{"id":98370528,"identity":"a3f782e3-0de3-4d24-8037-041e12bf0b0a","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":639488,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/c574f6d3c61fc2b453c89d34.tif"},{"id":98440675,"identity":"a4e5d58d-7e21-437d-a762-f5b3bc5a2334","added_by":"auto","created_at":"2025-12-17 17:04:11","extension":"tif","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2651852,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/483df82e4db84d61a2249380.tif"},{"id":98370524,"identity":"ae209288-665c-4143-824d-cef63b9e1739","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4100784,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/506480836c0248b7d8d7a2ac.tif"},{"id":98440545,"identity":"8d952666-d06f-4de2-9453-2b9cd7566e99","added_by":"auto","created_at":"2025-12-17 17:03:59","extension":"tif","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78199,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/f63f8b24667b792e0c0a861d.tif"},{"id":98439645,"identity":"1fd050cf-8a6a-49ce-aae0-7410b5408efe","added_by":"auto","created_at":"2025-12-17 17:02:21","extension":"tif","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":998000,"visible":true,"origin":"","legend":"","description":"","filename":"Fig6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/6e077d7ff590caa028f987a6.tif"},{"id":98440256,"identity":"ca8dcd5c-fb56-4b7a-8ed6-3f7ddf2a2f07","added_by":"auto","created_at":"2025-12-17 17:03:39","extension":"tif","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3474720,"visible":true,"origin":"","legend":"","description":"","filename":"Fig8.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/f11e7bbef706ba8d0cca9e5c.tif"},{"id":98440215,"identity":"7ef62096-9baa-49f2-b7c1-3209acf89251","added_by":"auto","created_at":"2025-12-17 17:03:31","extension":"tif","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7233488,"visible":true,"origin":"","legend":"","description":"","filename":"Fig9.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/ab0ac4ad916b5d2663e3356a.tif"},{"id":98370536,"identity":"c8316b53-fefd-4bc7-b95a-a83082b250e9","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3045808,"visible":true,"origin":"","legend":"","description":"","filename":"Fig7.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/acebf92b1be8eb6c1d8d00fc.tif"},{"id":98440877,"identity":"271b732a-2bc8-41b9-8784-8467a84a026f","added_by":"auto","created_at":"2025-12-17 17:04:33","extension":"json","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7574,"visible":true,"origin":"","legend":"","description":"","filename":"d5cf9898b9444200bb074f38829753fa.json","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/d3c53593cf715b19497b9ff7.json"},{"id":98370534,"identity":"e79342cb-1590-46bb-9c1b-8efc00857e4e","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"xlsx","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119775,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/88be647c4665a7c725f652c4.xlsx"},{"id":98370543,"identity":"4c50c19d-1190-49fc-a027-888a07b55119","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"xml","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":157354,"visible":true,"origin":"","legend":"","description":"","filename":"d5cf9898b9444200bb074f38829753fa1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/159a689e078f2f52947fe329.xml"},{"id":98440805,"identity":"ef95dee6-eac0-44bc-889e-f104b1e931c5","added_by":"auto","created_at":"2025-12-17 17:04:23","extension":"tif","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133941,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/1dfc9d44b9f3eacd379b24fa.tif"},{"id":98370542,"identity":"d9ef04be-8dd9-471e-bca9-917e284ff6f1","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4517416,"visible":true,"origin":"","legend":"","description":"","filename":"Fig10.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/80297da4560813e471d61066.tif"},{"id":98370537,"identity":"d19bf549-b509-4e5d-a1d0-9b964889f753","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":330424,"visible":true,"origin":"","legend":"","description":"","filename":"Fig11.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/304eab4cce2b8ad261707ea1.tif"},{"id":98439716,"identity":"eba2df26-204e-4cc2-90b7-0533c53bdc6e","added_by":"auto","created_at":"2025-12-17 17:02:48","extension":"tif","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3733260,"visible":true,"origin":"","legend":"","description":"","filename":"Fig12.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/07cf360dd6b9d66cf166538d.tif"},{"id":98439824,"identity":"1473d5ec-f6e9-48b0-849b-b31b7f0d6e9f","added_by":"auto","created_at":"2025-12-17 17:02:59","extension":"tif","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3326400,"visible":true,"origin":"","legend":"","description":"","filename":"Fig13.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/fec052f5e517122f2d1810b4.tif"},{"id":98439919,"identity":"25a518eb-1b88-4073-8df0-747ea2469d8a","added_by":"auto","created_at":"2025-12-17 17:03:04","extension":"tif","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":639488,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/c70a60244f0bc17e557d4985.tif"},{"id":98370548,"identity":"bec43853-4a82-437a-90e1-04792bb630e4","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"tif","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2651852,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/a14a55087aa3d28815a50143.tif"},{"id":98370552,"identity":"a65c6a06-7131-428c-bbec-97325fd86860","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"tif","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4100784,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/bfe616150d4e7b52e2a90f26.tif"},{"id":98370540,"identity":"2ee2f40d-4c5c-491d-a7c5-26a8e03c8e03","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"tif","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78199,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/4c96220bc95a81cc78908b10.tif"},{"id":98370549,"identity":"72e531d1-f4d4-490e-82ca-ca342f243e5b","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"tif","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":998000,"visible":true,"origin":"","legend":"","description":"","filename":"Fig6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/bcbad592ef0cd2cd17913e60.tif"},{"id":98439663,"identity":"07d2a435-998e-4d0b-88b7-0ae98ee865a6","added_by":"auto","created_at":"2025-12-17 17:02:28","extension":"tif","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3474720,"visible":true,"origin":"","legend":"","description":"","filename":"Fig8.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/c93e545fb9a692ced3c4bd6d.tif"},{"id":98370555,"identity":"4323f14a-2394-424f-b079-9f333a145247","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"tif","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7233488,"visible":true,"origin":"","legend":"","description":"","filename":"Fig9.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/47a39ecf2d056357b6a69f2e.tif"},{"id":98370551,"identity":"2fffee34-c5f6-45f6-9979-0af0694de0a2","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"tif","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3045808,"visible":true,"origin":"","legend":"","description":"","filename":"Fig7.tif","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/979eb93e79daf8a9c31c94e2.tif"},{"id":98370546,"identity":"3e6fcef5-1a57-4f34-9363-306992c65f84","added_by":"auto","created_at":"2025-12-17 05:29:27","extension":"xml","order_by":45,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":154764,"visible":true,"origin":"","legend":"","description":"","filename":"d5cf9898b9444200bb074f38829753fa1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/6682d8798b932120e9f5c4a9.xml"},{"id":98439637,"identity":"b95f8e4a-4e02-48c5-906a-c63372074c5e","added_by":"auto","created_at":"2025-12-17 17:02:16","extension":"html","order_by":46,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174015,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/b0bb978c1a812cd649b0aa4b.html"},{"id":98370512,"identity":"1b3ea6a0-afe2-46cc-bfc9-67223e0df7ba","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":24676,"visible":true,"origin":"","legend":"\u003cp\u003eStudy Workflow. PCA principal component analysis, GSEA gene set enrichment analysis, CGD combined gene datasets, DEGs differentially expressed genes, ERGs exosome-related genes, ERDEGs exosome-related differentially expressed genes, GO Gene Ontology, LASSO least absolute shrinkage and selection operator, ROC receiver operating characteristic, DCA decision curve analysis.\u003c/p\u003e","description":"","filename":"OnlineFig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/0cd8a2e7d89b6bb0fc5d2983.png"},{"id":98370508,"identity":"c6d33fd8-6b6a-44f3-becb-11b91c20f126","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":63590,"visible":true,"origin":"","legend":"\u003cp\u003eDebatching of the dataset. PCA plots of CGD before (A) and after (B) batch effect removal. Red represents the dataset GSE144521, and blue represents the dataset GSE77984. PCA principal component analysis, GEO gene expression omnibus, CGD combined GEO datasets.\u003c/p\u003e","description":"","filename":"OnlineFig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/3e824f98247d7e293fa92665.png"},{"id":98440002,"identity":"3efe4419-229b-4c28-a659-896547275649","added_by":"auto","created_at":"2025-12-17 17:03:12","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":279048,"visible":true,"origin":"","legend":"\u003cp\u003eDifferential gene expression analysis. (A) The heatmap shows 443 DEGs (218 upregulated, 225 downregulated) across GSE144521 and GSE77984. Columns indicate samples and rows indi-cate genes. The color gradient (blue–red) reflects expression from low to high (z-score normalized). Samples are annotated by group above the heatmap (Control group: blue; CCA group: red) and by dataset source (GSE144521: green; GSE77984: pink). Hierarchical clustering was applied to both genes and samples, demonstrating clear segregation between CCA and control groups, indicating distinct gene expression patterns associated with CCA. (B) Volcano map of differentially expressed genes between CCA and Normal group in CGD. (C) Venn diagram of DEGs and ERGs in CGD. DEGs differentially expressed genes, ERGs exosome-related genes, ERDEGs exosome-related dif-ferentially expressed genes, GEO gene expression omnibus, CGD combined GEO datasets.\u003c/p\u003e","description":"","filename":"OnlineFig3.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/9c8457d71d902417c6d724b2.png"},{"id":98440899,"identity":"22bd2147-28be-46a0-aecb-e70bdac659fc","added_by":"auto","created_at":"2025-12-17 17:04:35","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":484880,"visible":true,"origin":"","legend":"\u003cp\u003eGO enrichment analysis of ERDEGs. (A,B) Bubble and bar plots showing enriched GO terms in BP, CC, and MF. In the bubble plot, the x-axis shows the gene ratio, bubble size reflects gene counts, and color depth indicates adjusted p-values (red = smaller p-value, blue = larger p-value). (C) Circular plot of GO enrichment, with outer rings showing GO categories (red: BP; green: CC; blue: MF) and inner rings indicating gene counts and rich factor (0–1). The criteria for enrichment were P \u0026lt; 0.05 and FDR \u0026lt; 0.05. ERDEGs: exosome-related DEGs; GO: Gene Ontology; BP: biological process; CC: cellular component; MF: molecular function.\u003c/p\u003e","description":"","filename":"OnlineFig4.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/fc7b7f301741ccfc733fc658.png"},{"id":98440434,"identity":"b2d5a2ec-353b-4567-8e09-48e4a30079e6","added_by":"auto","created_at":"2025-12-17 17:03:51","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":20633,"visible":true,"origin":"","legend":"\u003cp\u003eGSEA of exosome-related gene sets in CGD. Mountain plots show enrichment curves for three pathways: Cytoskeleton in Muscle Cells (green), Protein Digestion and Absorption (red), and Viral Protein Interaction with Cytokine and Cytokine Receptor (purple). All three were downreg-ulated in CCA compared with normal tissues, with NES values of −1.98, −1.83, and −1.80 (P \u0026lt; 0.001, adjusted P ≤0.04). Thresholds for significance were P \u0026lt; 0.05 and FDR \u0026lt; 0.25. CGD: combined GEO datasets; CCA: cholangiocarcinoma.\u003c/p\u003e","description":"","filename":"OnlineFig5.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/26e971859954d1257431a231.png"},{"id":98440659,"identity":"e905fffa-76e0-4477-9371-07ab6c9d53e4","added_by":"auto","created_at":"2025-12-17 17:04:09","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":80477,"visible":true,"origin":"","legend":"\u003cp\u003eConstruction of the LASSO regression model for key gene selection in CCA. (A) LASSO regression with 10-fold cross-validation was applied to select the optimal λ, based on the minimum binomial deviance. The dashed vertical line marks this λ, where two genes were kept in the model. (B) Coefficient profile plot of candidate genes across the log λ sequence. The plot displays how the regression coefficients of each gene change with increasing regularization, with two genes (WNT5A and PFN2) ultimately selected as key genes for the diagnostic model. LASSO, least absolute shrinkage and selection operator; CCA, cholangiocarcinoma.\u003c/p\u003e","description":"","filename":"OnlineFig6.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/3849ec23a1d87aabe898d8f8.png"},{"id":98439622,"identity":"a4b98e17-7baa-42c9-9171-b0d918f06cde","added_by":"auto","created_at":"2025-12-17 17:02:13","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2080000,"visible":true,"origin":"","legend":"\u003cp\u003eDiagnostic model validation and performance assessment in cholangiocarcinoma (CCA). (A) Boxplot showing the expression of key genes in the cholangiocarcinoma group versus the control group. (B) Circos plot showing chromosomal positions of key genes. (C) ROC curves of lo-gistic regression predictors. (D) Nomogram of the model genes in the CGD for the diagnostic model of CCA. (E, F) Calibration curve plot (E) and DCA plot (F) of the model genes in the CGD for the diagnostic model of CCA. The ordinate of the calibration curve represents the net benefit, while the abscissa corresponds to the threshold probability. An AUC value greater than 0.9 indicates a high level of accuracy. AUC, area under the curve; DCA, decision curve analysis; ROC, receiver oper-ating characteristic; GEO, Gene Expression Omnibus; CCA, cholangiocarcinoma, CGD combined GEO datasets.\u003c/p\u003e","description":"","filename":"OnlineFig7.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/7739991182e31d2fb883b545.png"},{"id":98439533,"identity":"fcc194b5-610a-4f98-b9b9-81a0d8a3ffb6","added_by":"auto","created_at":"2025-12-17 17:02:03","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":182779,"visible":true,"origin":"","legend":"\u003cp\u003eImmune Infiltration analysis (ssGSEA and MCPCounter). (A) Group comparison graph for 28 types of immune cells in different groups in CGD by ssGSEA. (B) Heat map of correlation analysis between Key Genes and immune cell infiltration abundance by MCPCounter. In the cor-relation heat map, the red circle represents the positive correlation between the genes and the in-filtration abundance of immune cells. The symbol ns is equivalent to P \u0026lt;0.05,indicating no statistical significance. The symbol * is equivalent to P\u0026lt;0.05, indicating statistical significance. The symbol ** is equivalent to P \u0026lt;0.01, indicating a highly statistical significance. The symbol *** is equivalent to P \u0026lt;0.001, indicating a extremely statistical significance. ssGSEAsingle-sample gene-set enrichment analysis, MCPCounter microenvironment cell populations-counter, CCA, cholangiocarcinoma, CGD combined GEO datasets.\u003c/p\u003e","description":"","filename":"OnlineFig8.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/5b6274be504f35ab36d16b81.png"},{"id":98440664,"identity":"44e427d3-5cfe-4eed-aa10-686db395744e","added_by":"auto","created_at":"2025-12-17 17:04:09","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":313591,"visible":true,"origin":"","legend":"\u003cp\u003eGeneMania network analysis. (A) Network centered on WNT5A. (B) Network centered on PFN2.\u003c/p\u003e","description":"","filename":"OnlineFig9.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/51663489046cb719c89e3674.png"},{"id":98370519,"identity":"7c98403b-738f-4370-96ff-978560c0ab0a","added_by":"auto","created_at":"2025-12-17 05:29:26","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":289087,"visible":true,"origin":"","legend":"\u003cp\u003eMolecular docking analysis of drug compounds with WNT5A. (A)Docking of gemcita-bine with WNT5A. (B) Docking of temozolomide with WNT5A.\u003c/p\u003e","description":"","filename":"OnlineFig10.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/25ebb9c3292ceaaed48f47dc.png"},{"id":98440601,"identity":"a77341b6-bf40-4070-9209-73841b3e45d3","added_by":"auto","created_at":"2025-12-17 17:04:05","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":29801,"visible":true,"origin":"","legend":"\u003cp\u003ePocket-level molecular docking outcomes of gemcitabine, and temozolomide with WNT5A. (A) Docking results of gemcitabine with WNT5A, showing the best binding affinity at pocket C1 (Vina score = –6.5 kcal/mol). (B) Docking results of temozolomide with WNT5A, showing the best binding affinity at pocket C1 (Vina score = –6.1 kcal/mol).\u003c/p\u003e","description":"","filename":"OnlineFig11.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/1a35254adcec6ea3882291af.png"},{"id":98439489,"identity":"e8367fcc-6c82-4a56-b9bb-379c87327436","added_by":"auto","created_at":"2025-12-17 17:01:58","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":296372,"visible":true,"origin":"","legend":"\u003cp\u003eImmunohistochemical analysis of WNT5A expression in CCA and adjacent non-tumor tissues. Left: Representative IHC staining of WNT5A in CCA tissues.Middle: Representative IHC staining of WNT5A in adjacent non-tumor bile duct tissues.Right: Semi-quantitative analysis using the H-score method, showing significantly lower WNT5A expression in CCA tissues compared with adjacent non-tumor tissues (*P \u0026lt; 0.05). Scale bar = 20 μm.\u003c/p\u003e","description":"","filename":"OnlineFig12.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/bdbcb1638c185492b10ff04c.png"},{"id":98439319,"identity":"b2dd7ee6-fbb2-4ed7-a48c-a82fa9a600af","added_by":"auto","created_at":"2025-12-17 17:01:35","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":249141,"visible":true,"origin":"","legend":"\u003cp\u003eImmunohistochemical analysis of PFN2 expression in CCA and adjacent non-tumor tissues. Left: Representative IHC staining of PFN2 in CCA tissues.Middle: Representative IHC staining of PFN2 in adjacent non-tumor bile duct tissues.Right: Semi-quantitative analysis using the H-score method, showing significantly lower PFN2 expression in CCA tissues compared with ad-jacent non-tumor tissues (***P \u0026lt; 0.001). Scale bar = 20 μm.\u003c/p\u003e","description":"","filename":"OnlineFig13.png","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/45dc1a3d5b1754553edd790f.png"},{"id":99317845,"identity":"22ff27ad-73e4-483e-bd75-76fd933f91f9","added_by":"auto","created_at":"2025-12-31 16:30:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4319991,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/4d116279-8f73-4d41-90da-1e1349618d93.pdf"},{"id":98440513,"identity":"bda9a4e6-d57d-4305-b59f-d4436dc79b43","added_by":"auto","created_at":"2025-12-17 17:03:56","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":119775,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8144185/v1/967d129a60952ee4df5ef70b.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrative Identification and Validation of Exosome-Related Genes as Diagnostic Biomarkers and Potential Therapeutic Targets in Cholangiocarcinoma","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eCholangiocarcinoma (CCA) is an aggressive and heterogeneous cancer arising from the biliary epithelium, representing nearly 3% of all gastrointestinal malignancies [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Despite notable progress in imaging diagnostics and therapeutic interventions, both the incidence and mortality of CCA have continued to increase worldwide in recent decades [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Because of its subtle onset, the absence of specific early clinical manifestations, and the lack of reliable biomarkers, CCA is frequently diagnosed at advanced or metastatic stages [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Consequently, the prognosis remains dismal, with a five year survival rate below 10% for most patients [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Surgical resection is still the only treatment offering a potential cure [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]; however, the majority of individuals are not candidates for surgery at the time of diagnosis due to extensive disease progression. Therefore, it is crucial to identify new molecular indicators and establish effective diagnostic approaches to facilitate early detection and improve patient outcomes in CCA [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eExosomes are nanosized extracellular vesicles that mediate intercellular communication by transferring proteins, microRNAs, and long noncoding RNAs, and they have been increasingly recognized as vital modulators in tumor progression [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. In CCA, exosomes influence the tumor microenvironment by interacting with immune cells and cancer-associated fibroblasts (CAFs), thereby enhancing cellular proliferation, migration, and invasion through activation of Wnt/β-catenin signaling and cytoskeletal remodeling [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. In addition, exosomal non-coding RNAs regulate gene expression within both tumor and immune cells, suggesting their promise as diagnostic and therapeutic candidates [\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Nevertheless, the precise functions of exosome-related genes (ERGs) in regulating immune infiltration, cytoskeletal rearrangement, and intracellular signaling in CCA remain insufficiently understood [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAlthough increasing evidence indicates that exosomes contribute to the remodeling of the immune landscape and oncogenic pathways in CCA [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], comprehensive investigations linking ERGs with immune regulation, cytoskeletal dynamics, and related molecular networks are still limited [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Furthermore, the possible involvement of ERGs in pathways such as \u0026ldquo;Cytoskeleton in Muscle Cells,\u0026rdquo; \u0026ldquo;Protein Digestion and Absorption,\u0026rdquo; and \u0026ldquo;Viral Protein Interaction with Cytokine and Cytokine Receptor\u0026rdquo; has yet to be clarified [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], emphasizing a notable gap in our current understanding of CCA pathobiology.\u003c/p\u003e \u003cp\u003eIn this study, we systematically profiled ERGs associated with CCA and identified two markedly downregulated genes, WNT5A and PFN2, as diagnostic biomarkers. Based on integrated multi-cohort transcriptomic datasets, an ERG-based diagnostic model was constructed using LASSO regression. Functional exploration included enrichment analysis, immune infiltration characterization, gene-interaction mapping, drug-repurposing prediction, molecular docking, and immunohistochemical validation. Collectively, our findings elucidate the molecular roles of ERGs in CCA and highlight WNT5A and PFN2 as promising biomarkers and therapeutic targets.\u003c/p\u003e"},{"header":"2 Materials and Methods","content":"\u003cp\u003eData Collection and Preprocessing\u003c/p\u003e \u003cp\u003eGene expression datasets were retrieved from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Raw data were standardized to ensure cross-platform comparability, and all datasets were merged to generate a comprehensive gene dataset (CGD). To correct for unwanted variation and hidden confounding factors, surrogate variable analysis (SVA) [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] was applied. Principal component analysis (PCA) was subsequently used to verify the effective reduction of batch effects, ensuring the robustness of the dataset for downstream analyses.\u003c/p\u003e \u003cp\u003eIdentification of Exosome-Related Differentially Expressed Genes\u003c/p\u003e \u003cp\u003eThe CGD was divided into cholangiocarcinoma (CCA) and normal control groups. Differential expression analysis was performed using the \u0026ldquo;limma\u0026rdquo; R package [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Genes meeting the thresholds of |logFC| \u0026gt; 0 and p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were identified as differentially expressed genes (DEGs). These DEGs were then intersected with a curated set of exosome-related genes (ERGs) to obtain exosome-related differentially expressed genes (ERDEGs).\u003c/p\u003e \u003cp\u003eGene Ontology (GO) Enrichment Analysis\u003c/p\u003e \u003cp\u003eTo elucidate the biological significance of the 42 ERDEGs identified in CCA, Gene Ontology (GO) functional enrichment analysis was conducted across the categories of Biological Process (BP), Cellular Component (CC), and Molecular Function (MF) [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. The \u0026ldquo;clusterProfiler\u0026rdquo; R package (v3.16.1) [24)]was used to perform GO enrichment. Statistically significant enrichment was defined as a corrected p-value (p.adj)\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and false discovery rate (FDR, Q value)\u0026thinsp;\u0026lt;\u0026thinsp;0.25. The Benjamini\u0026ndash;Hochberg (BH) procedure was applied to adjust for multiple comparisons and reduce false positives [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eGene Set Enrichment Analysis (GSEA)\u003c/p\u003e \u003cp\u003eGene set enrichment analysis was performed using the \u0026ldquo;clusterProfiler\u0026rdquo; R package (v3.16.1) [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] to identify significantly enriched pathways and biological processes between CCA and normal samples based on the full gene expression matrix of exosome-related genes. Pairwise correlations between the key genes and all other genes in the training cohort were calculated and ranked in descending order to form the gene list for enrichment testing. Pathways with |NES| \u0026gt; 1 and p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered significantly enriched.\u003c/p\u003e \u003cp\u003eDiagnostic Model Construction and Validation\u003c/p\u003e \u003cp\u003eA diagnostic model for CCA was established based on the 42 ERDEGs. Initially, univariate logistic regression was applied to identify genes with significant diagnostic potential (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The least absolute shrinkage and selection operator (LASSO) [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] regression was then used to refine candidate genes and construct the diagnostic model. The optimal λ parameter was determined by 10-fold cross-validation [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e], and corresponding gene coefficients were obtained. The diagnostic efficiency of the model was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). A nomogram was generated to provide a clinical visualization of the predictive model, and calibration plots were used to examine the concordance between predicted and observed outcomes [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Decision curve analysis (DCA) was further applied to estimate the clinical benefit of the model. In addition, Circos plots were used to display the chromosomal positions of the key genes, and differential expression between tumor and control tissues was assessed to validate their diagnostic significance.\u003c/p\u003e \u003cp\u003eImmune Infiltration Analysis (ssGSEA and MCPcounter)\u003c/p\u003e \u003cp\u003eThe infiltration levels of 28 immune cell types in the CGD were quantified using the single-sample gene set enrichment analysis (ssGSEA) algorithm implemented in the \u0026ldquo;GSVA\u0026rdquo; R package. Differences in immune infiltration between the CCA and normal groups were analyzed, and correlations between the expression of the key genes (WNT5A and PFN2) and immune cell abundance were further examined using the \u0026ldquo;MCPcounter\u0026rdquo; algorithm [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eGene Interaction Network Analysis (GeneMANIA)\u003c/p\u003e \u003cp\u003eInteraction networks for WNT5A and PFN2 were generated using the GeneMANIA plugin in Cytoscape (v3.10.0). Default parameters were used, integrating evidence from multiple sources such as co-expression, physical interaction, predicted association, co-localization, shared pathways, and genetic interaction. The networks were visualized with each key gene as the central node, and the top 20 functionally associated genes were displayed. Functional enrichment of the resulting network was conducted through GeneMANIA\u0026rsquo;s built-in enrichment tool, with FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05 set as the significance threshold.\u003c/p\u003e \u003cp\u003eDrug Enrichment Analysis and Molecular Docking\u003c/p\u003e \u003cp\u003ePotential therapeutic compounds targeting CCA-related genes were identified through drug enrichment analysis using Enrichr (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://maayanlab.cloud/Enrichr/\u003c/span\u003e\u003cspan address=\"https://maayanlab.cloud/Enrichr/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and the Drug\u0026ndash;Gene Interaction Database (DGIdb, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.dgidb.org/\u003c/span\u003e\u003cspan address=\"https://www.dgidb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. The key genes WNT5A and PFN2 were input into both databases, and significantly enriched drug\u0026ndash;gene associations were selected based on FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Candidate compounds were further validated through cross-referencing with DrugBank, ChEMBL, and DrugMatrix databases to confirm biological relevance. Three-dimensional protein structures of target genes were retrieved from the AlphaFold Protein Structure Database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.alphafold.ebi.ac.uk/\u003c/span\u003e\u003cspan address=\"https://www.alphafold.ebi.ac.uk/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Models with a predicted Local Distance Difference Test (pLDDT) confidence score above 70 were selected for analysis. Small-molecule structures of enriched drugs were obtained from PubChem or ChEMBL in SDF format. Molecular docking was carried out using AutoDock Vina, with the grid box size set to 30 \u0026times; 30 \u0026times; 30 \u0026Aring;, centered on either the known ligand-binding pocket or the top predicted cavity from AutoSite [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. The exhaustiveness parameter was fixed at 8. Compounds with binding free energy \u0026le; \u0026minus;\u0026thinsp;6 kcal/mol were regarded as having strong affinity and potential therapeutic value.\u003c/p\u003e \u003cp\u003eImmunohistochemistry (IHC) Validation\u003c/p\u003e \u003cp\u003eFormalin-fixed, paraffin-embedded CCA and adjacent normal bile duct tissues were sectioned at 4 \u0026micro;m thickness. Sections were deparaffinized in xylene, rehydrated through graded ethanol, and subjected to antigen retrieval in citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was quenched using 3% hydrogen peroxide for 10 min, followed by blocking with 5% bovine serum albumin (BSA) for 30 min. The tissue samples used for IHC validation were obtained from the archived pathology specimens of the Department of Pathology, Shandong Provincial Hospital Affiliated to Shandong First Medical University. These specimens were collected from patients who underwent surgical resection between June 2025 and September 2025, and postoperative histopathology confirmed a diagnosis of cholangiocarcinoma. Three patients were selected for the validation of WNT5A and another three patients for the validation of PFN2, with each patient providing both tumor tissue and paired adjacent normal bile duct tissue (six FFPE slides per gene). The acquisition of human specimens was approved by the Ethics Committee for Biomedical Research Involving Humans of Shandong Provincial Hospital Affiliated to Shandong First Medical University (protocol code SWYX:NO. 2025\u0026thinsp;\u0026minus;\u0026thinsp;540, approved on 1 September 2025), and the requirement for written informed consent was waived by the committee. Slides were incubated overnight at 4\u0026deg;C with primary antibodies against WNT5A (Affinity Biosciences, Cat# DF6856, 1:50) and PFN2 (Abcam, Cat# 2H7C12, 1:50). After washing, sections were treated with HRP-conjugated secondary antibodies and visualized using an M\u0026amp;R HRP/DAB Detection Kit (Cat# HC301, DAB dilution 1:10), followed by hematoxylin counterstaining. The stained slides were dehydrated, mounted, and observed under a light microscope.\u003c/p\u003e \u003cp\u003eImmunoreactivity was semi-quantitatively assessed by the H-score method, where staining intensity was graded as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong), and the percentage of positive cells was recorded at each intensity. The final H-score was calculated using the formula:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\varvec{H}\\text{-}\\varvec{s}\\varvec{c}\\varvec{o}\\varvec{r}\\varvec{e}={\\sum\\:}_{\\varvec{i}=0}^{3}\\left(\\varvec{i}\\times\\:\\text{percentage\\:of\\:positive\\:cells\\:at\\:intensity\\:}\\varvec{i}\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eyielding a range from 0 to 300. H-scores between CCA and normal tissues were compared to determine differential expression of WNT5A and PFN2.\u003c/p\u003e \u003cp\u003eStatistical Analysis\u003c/p\u003e \u003cp\u003eAll statistical procedures were performed using R software (v4.2.1). The \u0026ldquo;limma\u0026rdquo; package was used for differential expression analysis, \u0026ldquo;clusterProfiler\u0026rdquo; for functional enrichment, \u0026ldquo;GSVA\u0026rdquo; for ssGSEA, and \u0026ldquo;MCPcounter\u0026rdquo; for immune cell estimation. P-values were adjusted for multiple testing using the Benjamini\u0026ndash;Hochberg method, and adjusted p (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was regarded as statistically significant. Continuous variables were presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD) or median with interquartile range (IQR), depending on distribution normality. Group comparisons between CCA and controls were conducted using the Student\u0026rsquo;s t-test or Wilcoxon rank-sum test as appropriate. Categorical variables were compared with the χ\u0026sup2; test or Fisher\u0026rsquo;s exact test.\u003c/p\u003e \u003cp\u003eSpearman\u0026rsquo;s correlation was applied to evaluate associations between gene expression and immune infiltration scores, and results were visualized using the \u0026ldquo;ggplot2\u0026rdquo; package. For model development, univariate logistic regression and LASSO regression were used to select key genes, with model performance assessed via ROC analysis and AUC values using the \u0026ldquo;pROC\u0026rdquo; package. Model calibration was evaluated using calibration plots, and its clinical applicability was examined by decision curve analysis (DCA). For molecular docking, AutoDock Vina calculated binding affinities, and docking poses were visualized in PyMOL. Unless otherwise specified, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003cp\u003eStudy Workflow\u003c/p\u003e \u003cp\u003eTo provide an overview of the analytical strategy, we first summarized the study design and workflow (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Based on this workflow, the subsequent results are presented in the following sections, including the identification of exosome-related differentially expressed genes (ERDEGs), construction of a diagnostic model, immune infiltration analysis, drug enrichment and molecular docking, and immunohistochemical validation in clinical samples.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"3 Results","content":"\u003cp\u003eData Collection and Correction\u003c/p\u003e \u003cp\u003eFollowing the normalization and integration of the two cholangiocarcinoma datasets into the combined gene dataset (CGD), surrogate variable analysis (SVA) was applied to remove hidden batch effects. Principal component analysis (PCA) plots (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) demonstrated that the batch effects among samples were substantially reduced after correction, confirming the reliability of the integrated dataset.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eExosome-Related Differentially Expressed Genes in Cholangiocarcinoma\u003c/p\u003e \u003cp\u003eThe identified differentially expressed genes (DEGs) were visualized in a heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). A total of 443 DEGs were identified (|logFC| \u0026gt; 0; p\u0026thinsp;\u0026lt;\u0026thinsp;0.05)(Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e), including 218 upregulated genes (logFC\u0026thinsp;\u0026gt;\u0026thinsp;0) and 225 downregulated genes (logFC\u0026thinsp;\u0026lt;\u0026thinsp;0). These results were further illustrated using a volcano plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB).By intersecting the DEGs with a curated list of exosome-related genes (ERGs) [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], 42 exosome-related differentially expressed genes (ERDEGs) were identified (Table S2). The overlap was visualized using a Venn diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eGO enrichment analysis\u003c/p\u003e \u003cp\u003eThe results (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, Table S3) showed that ERDEGs were enriched in biological process (BP), cellular component (CC), and molecular function (MF) terms.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn the BP group, enriched terms mainly involved negative regulation of cell migration, locomotion, and epithelial cell development. In the CC group, the top terms were membrane raft, focal adhesion, and collagen-containing extracellular matrix. In the MF group, ERDEGs were linked to integrin binding, cadherin binding, actin binding, and oxidoreductase activity.\u003c/p\u003e \u003cp\u003eThese findings suggest that ERDEGs in cholangiocarcinoma are mainly associated with pathways regulating cell migration, adhesion, signal transduction, and cytoskeletal remodeling [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. These functions respectively correspond to key biological processes such as tumor invasion and metastasis, modulation of the tumor microenvironment, and cytoskeletal dynamics. Collectively, the results suggest that these ERDEGs may promote cholangiocarcinoma progression through exosome-mediated intercellular signaling.\u003c/p\u003e \u003cp\u003eGene set enrichment analysis (GSEA)\u003c/p\u003e \u003cp\u003eThe GSEA results(Table S4) indicated that the pathways \u0026ldquo;Cytoskeleton in Muscle Cells,\u0026rdquo; \u0026ldquo;Protein Digestion and Absorption,\u0026rdquo; and \u0026ldquo;Viral Protein Interaction with Cytokine and Cytokine Receptor\u0026rdquo; were significantly downregulated in the CCA group compared to the healthy controls. The normalized enrichment scores (NES) for these pathways were \u0026minus;\u0026thinsp;1.98, \u0026minus;\u0026thinsp;1.83, and \u0026minus;\u0026thinsp;1.80, respectively, all of which were statistically significant (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001, adjusted P\u0026thinsp;\u0026le;\u0026thinsp;0.04)(Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.). These findings suggest that the aforementioned biological processes may be suppressed in patients with CCA, potentially correlating with disease progression or therapeutic response.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDiagnostic model for CCA\u003c/p\u003e \u003cp\u003e42 ERDEGs were statistically significant (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Table S2). Univariate logistic regression was used to assess the diagnostic value of 42 ERDEGs in CCA, and 3 genes showed statistical significance (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).Based on these results, a LASSO regression model was constructed to develop a diagnostic signature for CCA, with visualization provided by the LASSO regression plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA) and the coefficient profile plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). In the cross-validation plot (binomial deviance vs. log λ), the optimal λ was determined, and the model at this λ retained two genes. The coefficient profile plot illustrates the variation in the regression coefficients of each gene across a range of λ values. Consequently, two genes, WNT5A and PFN2, were identified as key genes for subsequent analysis.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariate Logistic Regression Analysis of ERDEGs for Diagnostic Value in CCA.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026times;\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026times;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026times;\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026times;\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eOR.95L\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOR.95H\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c2\"\u003e \u003cp\u003e5.64 \u0026times; 10⁻\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e \u003cp\u003e9.92 \u0026times; 10⁻⁴\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c4\"\u003e \u003cp\u003e4.61 \u0026times; 10⁻\u0026sup1;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c5\"\u003e \u003cp\u003e4.45 \u0026times; 10⁻\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePFN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c2\"\u003e \u003cp\u003e7.18 \u0026times; 10⁻\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e \u003cp\u003e1.45 \u0026times; 10⁻\u0026sup3;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c4\"\u003e \u003cp\u003e4.82 \u0026times; 10⁻\u0026sup1;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c5\"\u003e \u003cp\u003e4.58 \u0026times; 10⁻\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBASP1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c2\"\u003e \u003cp\u003e4.48 \u0026times; 10⁻\u0026sup1;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e \u003cp\u003e1.46 \u0026times; 10⁻\u0026sup1;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c4\"\u003e \u003cp\u003e8.63 \u0026times; 10⁻\u0026sup1;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c5\"\u003e \u003cp\u003e4.86 \u0026times; 10⁻\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDiagnostic model validation\u003c/p\u003e \u003cp\u003eTo assess key gene expression differences, boxplots compared control and CCA samples. WNT5A and PFN2 showed markedly reduced expression in CCA tissues relative to controls (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). This finding suggests a downregulation trend of these two genes under CCA conditions, indicating their potential suppressive roles in the pathogenesis of CCA (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). Notably, when examining the chromosomal locations of these key genes, the constructed Circos plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB) showed that WNT5A and PFN2 are both located on chromosome 3, further indicating that these genes may exert synergistic effects during the occurrence and progression of CCA.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo validate the CCA diagnostic model constructed based on the two key ERDEGs, ROC curves (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC) and a nomogram (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD) were generated. The ROC analysis demonstrated that WNT5A (AUC\u0026thinsp;=\u0026thinsp;0.857) and PFN2 (AUC\u0026thinsp;=\u0026thinsp;0.841) both had AUC values greater than 0.8, indicating good diagnostic performance of these genes in the CCA diagnostic model. The effectiveness of the key genes (WNT5A and PFN2) in the diagnostic model was significantly higher than that of other variables.\u003c/p\u003e \u003cp\u003eFinally, to assess the clinical applicability and predictive stability of the diagnostic model, a calibration curve (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE) and a decision curve analysis (DCA) plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF) were generated. The calibration curve was used to evaluate the model\u0026rsquo;s ability to predict actual outcomes under different conditions. Although the calibration curve showed slight deviations from the ideal diagonal, it demonstrated an overall good fit, indicating reliable predictive performance. The DCA plot was used to evaluate the clinical utility of the diagnostic model, showing that the model curve remained stable within a certain threshold range and was higher than the strategies of treating all cases as positive or negative, indicating that the model offers high net benefits and demonstrates favorable clinical applicability.\u003c/p\u003e \u003cp\u003eImmune infiltration analysis (ssGSEA and MCPCounter)\u003c/p\u003e \u003cp\u003eImmune cell infiltration was evaluated in the integrated dataset (CGD) using the ssGSEA algorithm, yielding infiltration scores for 28 immune cell types (Table S5). Boxplot analysis revealed significant differences (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in 18 immune cell types between the CCA and normal groups, including activated CD4 T cells, activated dendritic cells, CD56dim natural killer cells, central memory CD4 T cells, central memory CD8 T cells, eosinophils, immature dendritic cells, macrophages, mast cells, MDSCs, memory B cells, natural killer T cells, neutrophils, plasmacytoid dendritic cells, regulatory T cells, follicular helper T cells, Th1 cells, and Th17 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). Correlation analysis using the MCPcounter algorithm demonstrated that WNT5A was significantly positively correlated with activated CD8 T cells, CD56dim natural killer cells, central memory CD4 T cells, macrophages, mast cells, monocytes, regulatory T cells, and Th17 cells (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, r\u0026thinsp;\u0026gt;\u0026thinsp;0.5), but negatively correlated with activated CD4 T cells, central memory CD8 T cells, memory B cells, and type 2 T helper cells (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, r \u0026lt; \u0026minus;\u0026thinsp;0.5). PFN2 exhibited a similar correlation profile, with minor differences (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB). (Correlation is represented by r.)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCombined with our previous differential expression analysis showing significant downregulation of WNT5A and PFN2 in CCA patients, we hypothesize that these genes may act as tumor suppressors (protective genes) in cholangiocarcinoma. Immune infiltration analysis using MCPcounter revealed that immune cell types positively correlated with WNT5A and PFN2 expression, including activated CD8 T cells, CD56dim natural killer cells, central memory CD4 T cells, macrophages, mast cells, monocytes, and regulatory T cells, which are generally associated with antitumor immune responses. Their positive correlation with protective genes suggests their potential roles in immune surveillance and tumor clearance in CCA.\u003c/p\u003e \u003cp\u003eConversely, reduced WNT5A and PFN2 expression in CCA was associated with increased activity of several immune cell types, such as activated CD4 T cells, central memory CD8 T cells, memory B cells, and type 2 helper T cells. This alteration may drive immune tolerance or evasion, facilitating tumor growth. These results suggest that WNT5A and PFN2 modulate the CCA immune microenvironment by affecting immune cell infiltration and composition, thus contributing to disease development.\u003c/p\u003e \u003cp\u003eGeneMANIA\u003c/p\u003e \u003cp\u003eThe potential interaction gene network constructed with WNT5A as the core is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e11\u003c/span\u003eA. Functional enrichment analysis of the WNT5A-associated gene network revealed significant overrepresentation of pathways related to non-canonical Wnt signaling, planar cell polarity, and G protein-coupled receptor binding (FDR\u0026thinsp;\u0026lt;\u0026thinsp;1e-12). These findings support the role of WNT5A in regulating cell polarity, tissue morphogenesis, and signal transduction, consistent with its known involvement in tumor progression and developmental processes[\u003cspan additionalcitationids=\"CR35 CR36 CR37\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] (Table S6). The constructed gene interaction network, with PFN2 at its nexus, is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e11\u003c/span\u003eB. The most significantly enriched functional pathway in its associated gene network is cytoskeletal regulation. Genes such as PFN1, PFN3, PFN4, VASP, and FHOD1 not only interact with PFN2 but are also co-expressed with it, all of which are closely involved in cytoskeletal remodeling processes(Table S7).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eDrug Enrichment Analysis and Molecular Docking\u003c/p\u003e \u003cp\u003eA total of 16 significantly associated compounds were identified for the two key genes. Among them, 14 compounds were significantly linked to WNT5A, while pentadecafluorooctanoic acid showed a significant association exclusively with PFN2. Notably, 3,3\u0026prime;,4,4\u0026prime;,5-pentachlorobiphenyl was significantly associated with both WNT5A and PFN2(Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Among the enriched results, 3,3',4,4',5-Pentachlorobiphenyl showed the smallest p-value and was enriched in both WNT5A and PFN2. However, considering its environmental pollutant nature and practical limitations, the focus was shifted to drug candidates with established clinical safety profiles and known antitumor potential, including gemcitabine [\u003cspan additionalcitationids=\"CR40 CR41\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e] and temozolomide (43\u0026ndash;48) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.03), both of which were enriched with WNT5A in the analysis. The identification of both gemcitabine and temozolomide as significantly enriched compounds associated with WNT5A suggests a potential shared mechanism beyond their canonical roles in DNA damage. Given WNT5A's involvement in non-canonical Wnt signaling pathways regulating cytoskeletal dynamics and cell motility[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e], it is plausible that these compounds may exert antitumor effects, at least in part, by modulating WNT5A-mediated signaling[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. This interaction could impair cellular processes critical for cholangiocarcinoma progression, such as epithelial\u0026ndash;mesenchymal transition (EMT), invasion, and metastasis. Further experimental validation is warranted to elucidate whether these drugs directly interfere with WNT5A function or its downstream pathways. To further evaluate the binding affinity of these candidate compounds to WNT5A, gemcitabine and temozolomide(p.adjust\u0026thinsp;\u0026lt;\u0026thinsp;0.04)were subjected to molecular docking analysis, examining their binding modes and molecular interactions with CCA.The molecular docking results confirmed that these compounds can potentially bind to WNT5A and modulate its function. In Figure.9A, gemcitabine demonstrates a deep binding conformation within the hydrophobic cavity of WNT5A, forming multiple hydrogen bonds and occupying a stable spatial niche, indicating a relatively high binding affinity and potential to modulate WNT5A function.In contrast, Figure.9B shows that temozolomide interacts more superficially with WNT5A, with fewer stabilizing interactions and a looser binding pattern, suggesting lower affinity and possibly limited direct regulatory capacity. But their Vina docking score are similar(Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e10\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCompounds significantly associated with the key genes WNT5A and PFN2.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompound (Description)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAdjusted \u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGene\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCount\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3,3',4,4',5-Pentachlorobiphenyl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A/PFN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ep-benzoquinone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDL-Homocysteine\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNSC94017\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eisoflupredone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehydroquinone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePentadecafluorooctanoic acid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePFN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2,2',4,4',5,5'-Hexachlorobiphenyl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[6-[6-(butanoylamino)purin-9-yl]-2-hydroxy-2-oxo-4a,6,7,7a-tetrahydro-4H-furo[3,2-d][\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]dioxaphosphinin-7-yl] butanoate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003egemcitabine\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003etemozolomide\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.036\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003efolic acid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.037\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003emitomycin C\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.039\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5-azacytidine\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedroxyprogesterone acetate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003emifepristone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWNT5A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eImmunohistochemistry Validation of Key Genes\u003c/p\u003e \u003cp\u003eTo validate the two exosome-related genes identified by bioinformatics, immunohistochemistry (IHC) was performed on paraffin-embedded CCA tissues and matched adjacent normal bile ducts. Both WNT5A and PFN2 proteins were predominantly localized in the cytoplasm, with partial distribution along the cell membrane, consistent with their functional roles in intracellular signaling and cytoskeletal regulation. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003e, adjacent normal bile duct tissues ex-hibited strong brownish staining for WNT5A and PFN2, whereas CCA tissues displayed markedly weaker staining intensity. Semi-quantitative evaluation using the H-score method revealed that WNT5A expression was significantly reduced in CCA tissues (mean H-score: 17.89\u0026thinsp;\u0026plusmn;\u0026thinsp;2.09) com-pared with adjacent normal tissues (mean H-score: 38.36\u0026thinsp;\u0026plusmn;\u0026thinsp;6.29; P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eB). Similarly, PFN2 expression was markedly lower in CCA tissues (mean H-score: 24.53\u0026thinsp;\u0026plusmn;\u0026thinsp;2.61) than in normal tissues (mean H-score: 59.86\u0026thinsp;\u0026plusmn;\u0026thinsp;5.24; P\u0026thinsp;\u0026lt;\u0026thinsp;0.01, Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eB). In addition, quantitative image analysis was per-formed using FIJI software, and the proportions of strongly positive, positive, weakly positive, and negative cells, together with the mean optical density (Mean OD) values, are summarized in the supplementary materials (Table S8, S9). These findings are consistent with the transcriptome analysis, confirming that WNT5A and PFN2 are downregulated at the protein level in CCA. Col-lectively, these results support their potential roles as tumor suppressors and highlight their diag-nostic value as candidate biomarkers in cholangiocarcinoma.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eCholangiocarcinoma (CCA) is a highly aggressive malignancy arising from the biliary epithelium, characterized by poor prognosis, limited therapeutic options, and high molecular heterogeneity(1, 50). Despite advances in surgery and chemotherapy, early diagnosis and effective treatment remain major challenges[\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Exosomes, as key mediators of intercellular communication, have gained increasing attention for their ability to transfer proteins, lipids, and nucleic acids, thereby influencing tumor proliferation, metastasis, immune evasion, and drug resistance[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. In CCA, exosomes have been implicated in the regulation of oncogenic pathways, including Wnt/β-catenin, PI3K/AKT, and TGF-β signaling[\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e], and show promise as stable, non-invasive biomarkers[\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e]. However, the specific exosome-associated genes driving CCA progression and their diagnostic potential remain poorly defined.\u003c/p\u003e \u003cp\u003eIn this study, integrated bioinformatics analyses identified 42 exosome-related differentially expressed genes (ERDEGs) in CCA. Functional enrichment indicated strong associations with cytoskeletal organization, cell adhesion, migration, and membrane-related pathways. Using univariate logistic and LASSO regression, we constructed a diagnostic model and identified two key downregulated genes\u0026mdash;WNT5A and PFN2\u0026mdash;both demonstrating robust diagnostic performance across validation analyses. Notably, both genes reside on chromosome 3p14 and exhibited highly overlapping expression patterns and immune correlation profiles, suggesting potential functional synergy in CCA biology.\u003c/p\u003e \u003cp\u003eWNT5A, a prototypical non-canonical Wnt ligand, regulates cell polarity, migration, and adhesion through pathways such as Wnt/PCP and Wnt/Ca\u0026sup2;⁺, acting upstream of cytoskeletal remodeling by activating RhoA, Rac1, and related signaling molecules [\u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. PFN2, a profilin family member, plays a direct role in actin cytoskeleton dynamics, influencing membrane tension and cell motility through its interactions with actin-binding partners [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. While WNT5A primarily transduces extracellular signals, PFN2 executes the structural reorganization of the cytoskeleton. The co-downregulation of these genes in CCA suggests that WNT5A may exert part of its effects via PFN2-dependent cytoskeletal remodeling, forming a coordinated WNT5A\u0026ndash;PFN2 axis that integrates signal transduction with structural regulation. This axis could influence exosome-mediated communication, tumor cell invasiveness, and responsiveness to microenvironmental cues.\u003c/p\u003e \u003cp\u003eImmune infiltration profiling revealed profound alterations in the CCA microenvironment [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], with significant differences in 18 immune cell types between tumor and normal tissues. Both WNT5A and PFN2 expression correlated positively with cytotoxic and regulatory immune subsets\u0026mdash;such as activated CD8⁺ T cells, NK cells, macrophages, and regulatory T cells\u0026mdash;typically linked to antitumor immunity [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. Conversely, negative correlations were observed with immune populations associated with tumor-promoting phenotypes, including memory B cells and Th2 cells[\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. These findings suggest that downregulation of WNT5A and PFN2 may weaken immune surveillance and facilitate immune evasion, highlighting their potential roles as both diagnostic biomarkers and immune modulators. Given the growing interest in immunotherapy for CCA, these results provide a rationale for exploring strategies to restore WNT5A and PFN2 expression or function to enhance antitumor immunity.\u003c/p\u003e \u003cp\u003eDrug enrichment analysis identified 16 compounds significantly associated with the key genes, among which gemcitabine and temozolomide were prioritized for their established clinical use and safety [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Traditionally considered DNA-damaging agents, both drugs were enriched in WNT5A-related pathways, suggesting additional roles in modulating non-canonical Wnt signaling. Molecular docking supported this hypothesis: gemcitabine exhibited deep hydrophobic pocket binding within WNT5A with multiple hydrogen bonds, while temozolomide demonstrated weaker, surface-level binding. Comparable docking scores suggest that both may modulate WNT5A function, potentially impairing EMT, metastasis, and immune-related processes. These findings underscore the translational potential of drug repurposing in CCA [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e], offering a feasible approach to target WNT5A without the need for novel drug development.\u003c/p\u003e \u003cp\u003eImmunohistochemistry (IHC) further validated the reduced protein expression of WNT5A and PFN2 in CCA tissues relative to normal bile ducts, consistent with transcriptomic data. Histopathological analysis revealed that low PFN2 expression often coincided with disorganized cellular architecture, suggestive of enhanced migratory capacity, while low WNT5A expression was associated with epithelial\u0026ndash;mesenchymal transition (EMT)-like features. These observations provide histological evidence linking gene downregulation to functional tumor behaviors, strengthening the credibility of our bioinformatic findings [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study offers several innovations: integration of multiple datasets to enhance robustness; an exosome-centered approach to gene discovery; construction and validation of a two-gene diagnostic model; and combination of drug enrichment with molecular docking to identify clinically available agents with potential CCA relevance. The inclusion of experimental validation at both mRNA and protein levels forms a complete discovery-to-validation pipeline, enhancing translational relevance. Nonetheless, limitations remain: reliance on public datasets with limited sample sizes, the need for mechanistic studies to dissect the WNT5A\u0026ndash;PFN2 axis in cytoskeletal and immune regulation, and the necessity for in vitro and in vivo confirmation of predicted drug\u0026ndash;target interactions. Future studies should expand clinical cohorts, explore the therapeutic restoration of WNT5A and PFN2 function, and evaluate the efficacy of repurposed agents in preclinical CCA models.\u003c/p\u003e"},{"header":"5 Conclusions","content":"\u003cp\u003eThis study identified two exosome-related genes, WNT5A and PFN2, as consistently downregulated in cholangiocarcinoma and demonstrated their strong diagnostic value through integrative bioinformatics and experimental validation. Functional analyses revealed that their loss is closely linked to impaired cytoskeletal remodeling, enhanced tumor migration and EMT-like features, as well as dysregulated immune infiltration, highlighting their dual roles in structural and immune regulation. Drug enrichment and molecular docking further indicated gemcitabine and temozolomide as promising therapeutic agents potentially targeting WNT5A. Together, these findings establish WNT5A and PFN2 as robust diagnostic biomarkers and potential therapeutic targets, providing a translational framework for improving early diagnosis and drug repurposing strategies in cholangiocarcinoma.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eConflict of Interest\u003c/h2\u003e \u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis research was funded by the Shandong Provincial Health Commission General Program, grant number 202304010282.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization, F.L.; methodology, Q.W. and Y.W.; software, Q.W.; validation, Z.W., Q.W. and Y.W.; formal analysis, Q.W.; investigation, Z.W. and H.C.; resources, F.L.; data curation, Q.W.; writing\u0026mdash;original draft preparation, Z.W.; writing\u0026mdash;review and editing, F.L.; visualization, Q.W.; supervision, F.L.; project administration, F.L.; funding acquisition, F.L. All authors have read and agreed to the published version of the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eWe are grateful to the technical staff and colleagues at Shandong First Medical University for their helpful support throughout this research.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets analyzed for this study are publicly available in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). The accession numbers are GSE26566 and GSE107943. All data utilized in this research are open-access and can be freely obtained from the GEO repository. Additional information is available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBanales, J. M. et al. Cholangiocarcinoma 2020: the next horizon in mechanisms and management. \u003cem\u003eNat. Rev. Gastroenterol. Hepatol.\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e, 557\u0026ndash;588. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41575-020-0310-z\u003c/span\u003e\u003cspan address=\"10.1038/s41575-020-0310-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDixon, M. E., Williams, M., Pappas, S. G. \u0026amp; Cholangiocarcinoma \u003cem\u003eCancer Treat. Res.\u003c/em\u003e \u003cb\u003e192\u003c/b\u003e, 165\u0026ndash;184 ; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-61238-1_9\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-61238-1_9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKidanemariam, S. et al. Cholangiocarcinoma: epidemiology and imaging-based review. \u003cem\u003eR. I. Med. J.\u003c/em\u003e 107, 43\u0026ndash;48 (2024). (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarrison, J. M., Visser, B. C. \u0026amp; Cholangiocarcinoma \u003cem\u003eSurg. Clin. North. Am.\u003c/em\u003e \u003cb\u003e104\u003c/b\u003e, 1281\u0026ndash;1293 ; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.suc.2024.04.003\u003c/span\u003e\u003cspan address=\"10.1016/j.suc.2024.04.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSarcognato, S. et al. \u003cem\u003eCholangiocarcinoma Pathologica\u003c/em\u003e \u003cb\u003e113\u003c/b\u003e, 158\u0026ndash;169 ; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32074/1591-951X-252\u003c/span\u003e\u003cspan address=\"10.32074/1591-951X-252\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVithayathil, M. \u0026amp; Khan, S. A. Current epidemiology of cholangiocarcinoma in Western countries. \u003cem\u003eJ. Hepatol.\u003c/em\u003e \u003cb\u003e77\u003c/b\u003e, 1690\u0026ndash;1698. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jhep.2022.07.022\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2022.07.022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQurashi, M., Vithayathil, M. \u0026amp; Khan, S. A. Epidemiology of cholangiocarcinoma. \u003cem\u003eEur\u003c/em\u003e. \u003cem\u003eJ. Surg. Oncol.\u003c/em\u003e \u003cb\u003e51\u003c/b\u003e, 107064. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejso.2023.107064\u003c/span\u003e\u003cspan address=\"10.1016/j.ejso.2023.107064\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA, N. V. L. K., Subramanian, S. \u0026amp; Ramakrishna, S. Therapeutic applications of exosomes in various diseases: a review. \u003cem\u003eBiomater. Adv.\u003c/em\u003e \u003cb\u003e134\u003c/b\u003e, 112579. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.msec.2021.112579\u003c/span\u003e\u003cspan address=\"10.1016/j.msec.2021.112579\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJafari, A., Babajani, A., Abdollahpour-Alitappeh, M., Ahmadi, N. \u0026amp; Rezaei-Tavirani, M. Exosomes and cancer: from molecular mechanisms to clinical applications. \u003cem\u003eMed. Oncol.\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12032-021-01491-0\u003c/span\u003e\u003cspan address=\"10.1007/s12032-021-01491-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKimiz-Gebologlu, I. \u0026amp; Oncel, S. S. Exosomes: large-scale production, isolation, drug loading efficiency, and biodistribution and uptake. \u003cem\u003eJ. Control Release\u003c/em\u003e. \u003cb\u003e347\u003c/b\u003e, 533\u0026ndash;543. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jconrel.2022.05.027\u003c/span\u003e\u003cspan address=\"10.1016/j.jconrel.2022.05.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao, L., Hong, J. \u0026amp; Wu, J. Potential of extracellular vesicles and exosomes as diagnostic markers for cholangiocarcinoma. \u003cem\u003eHepatobiliary Surg. Nutr.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 436\u0026ndash;438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21037/hbsn-2022-02\u003c/span\u003e\u003cspan address=\"10.21037/hbsn-2022-02\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFabris, L., Sato, K., Alpini, G. \u0026amp; Strazzabosco, M. The tumor microenvironment in cholangiocarcinoma progression. \u003cem\u003eHepatology 73 Suppl.\u003c/em\u003e \u003cb\u003e1\u003c/b\u003e, 75\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/hep.31410\u003c/span\u003e\u003cspan address=\"10.1002/hep.31410\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLapitz, A. et al. Liquid biopsy-based protein biomarkers for risk prediction, early diagnosis, and prognostication of cholangiocarcinoma. \u003cem\u003eJ. Hepatol.\u003c/em\u003e \u003cb\u003e79\u003c/b\u003e, 93\u0026ndash;108. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jhep.2023.02.027\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2023.02.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo, C. et al. Tumor-derived exosomes induce immunosuppressive macrophages to foster intrahepatic cholangiocarcinoma progression. \u003cem\u003eHepatology\u003c/em\u003e \u003cb\u003e76\u003c/b\u003e, 982\u0026ndash;999. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/hep.32387\u003c/span\u003e\u003cspan address=\"10.1002/hep.32387\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShu, L. et al. Bile exosomal miR-182/183-5p increases cholangiocarcinoma stemness and progression by targeting HPGD and increasing PGE2 generation. \u003cem\u003eHepatology\u003c/em\u003e \u003cb\u003e79\u003c/b\u003e, 307\u0026ndash;322. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/hep.0000000000000437\u003c/span\u003e\u003cspan address=\"10.1097/hep.0000000000000437\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, M. et al. Single-cell transcriptomic architecture and intercellular crosstalk of human intrahepatic cholangiocarcinoma. \u003cem\u003eJ. Hepatol.\u003c/em\u003e \u003cb\u003e73\u003c/b\u003e, 1118\u0026ndash;1130. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jhep.2020.05.039\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2020.05.039\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, K. et al. Exosomes in the tumor microenvironment of cholangiocarcinoma: current status and future perspectives. \u003cem\u003eJ. Transl Med.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e, 117. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12967-022-03294-x\u003c/span\u003e\u003cspan address=\"10.1186/s12967-022-03294-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, J. et al. Exosome-derived miR-182-5p promoted cholangiocarcinoma progression and vasculogenesis by regulating ADK/SEMA5A/PI3K pathway. \u003cem\u003eLiver Int.\u003c/em\u003e \u003cb\u003e44\u003c/b\u003e, 370\u0026ndash;388. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/liv.15773\u003c/span\u003e\u003cspan address=\"10.1111/liv.15773\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, T. Y. et al. PTEN deficiency facilitates exosome secretion and metastasis in cholangiocarcinoma by impairing TFEB-mediated lysosome biogenesis. \u003cem\u003eGastroenterology\u003c/em\u003e \u003cb\u003e164\u003c/b\u003e, 424\u0026ndash;438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2022.11.025\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2022.11.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEdgar, R., Domrachev, M. \u0026amp; Lash, A. E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e30\u003c/b\u003e, 207\u0026ndash;210. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/30.1.207\u003c/span\u003e\u003cspan address=\"10.1093/nar/30.1.207\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeek, J. T., Johnson, W. E., Parker, H. S., Jaffe, A. E. \u0026amp; Storey, J. D. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cb\u003e28\u003c/b\u003e, 882\u0026ndash;883. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/bioinformatics/bts034\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/bts034\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRitchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e43\u003c/b\u003e, e47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkv007\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkv007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshburner, M. et al. Gene ontology: tool for the unification of biology. \u003cem\u003eNat. Genet.\u003c/em\u003e \u003cb\u003e25\u003c/b\u003e, 25\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/75556\u003c/span\u003e\u003cspan address=\"10.1038/75556\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, G., Wang, L. G., Han, Y. \u0026amp; He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. \u003cem\u003eOMICS\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e, 284\u0026ndash;287. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1089/omi.2011.0118\u003c/span\u003e\u003cspan address=\"10.1089/omi.2011.0118\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGlickman, M. E., Rao, S. R. \u0026amp; Schultz, M. R. False discovery rate control is a recommended alternative to Bonferroni-type adjustments in health studies. \u003cem\u003eJ. Clin. Epidemiol.\u003c/em\u003e \u003cb\u003e67\u003c/b\u003e, 850\u0026ndash;857. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jclinepi.2014.03.012\u003c/span\u003e\u003cspan address=\"10.1016/j.jclinepi.2014.03.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlhamzawi, R. \u0026amp; Ali, H. T. M. The Bayesian adaptive lasso regression. \u003cem\u003eMath. Biosci.\u003c/em\u003e \u003cb\u003e303\u003c/b\u003e, 75\u0026ndash;82. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.mbs.2018.06.004\u003c/span\u003e\u003cspan address=\"10.1016/j.mbs.2018.06.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKucheryavskiy, S., Rodionova, O. \u0026amp; Pomerantsev, A. Procrustes cross-validation of multivariate regression models. \u003cem\u003eAnal. Chim. Acta\u003c/em\u003e. \u003cb\u003e1255\u003c/b\u003e, 341096. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.aca.2023.341096\u003c/span\u003e\u003cspan address=\"10.1016/j.aca.2023.341096\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVickers, A. J., Van Calster, B. \u0026amp; Steyerberg, E. W. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. \u003cem\u003eBMJ\u003c/em\u003e \u003cb\u003e352\u003c/b\u003e, i6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj.i6\u003c/span\u003e\u003cspan address=\"10.1136/bmj.i6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecht, E. et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. \u003cem\u003eGenome Biol.\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e, 218. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13059-016-1070-5\u003c/span\u003e\u003cspan address=\"10.1186/s13059-016-1070-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCotto, K. C. et al. DGIdb 3.0: a redesign and expansion of the drug-gene interaction database. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e46\u003c/b\u003e, D1068\u0026ndash;D1073. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkx1143\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkx1143\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRavindranath, P. A., Forli, S., Goodsell, D. S., Olson, A. J. \u0026amp; Sanner, M. F. AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility. \u003cem\u003ePLoS Comput. Biol.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, e1004586. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pcbi.1004586\u003c/span\u003e\u003cspan address=\"10.1371/journal.pcbi.1004586\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalluri, R. \u0026amp; LeBleu, V. S. The biology, function, and biomedical applications of exosomes. \u003cem\u003eScience\u003c/em\u003e \u003cb\u003e367\u003c/b\u003e, eaau6977. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.aau6977\u003c/span\u003e\u003cspan address=\"10.1126/science.aau6977\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoshino, A. et al. Tumour exosome integrins determine organotropic metastasis. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e527\u003c/b\u003e, 329\u0026ndash;335. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nature15756\u003c/span\u003e\u003cspan address=\"10.1038/nature15756\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTufail, M. \u0026amp; Wu, C. WNT5A: a double-edged sword in colorectal cancer progression. \u003cem\u003eMutat. Res. Rev. Mutat. Res.\u003c/em\u003e \u003cb\u003e792\u003c/b\u003e, 108465. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.mrrev.2023.108465\u003c/span\u003e\u003cspan address=\"10.1016/j.mrrev.2023.108465\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBueno, M. L. P., Saad, S. T. O. \u0026amp; Roversi, F. M. WNT5A in tumor development and progression: a comprehensive review. \u003cem\u003eBiomed. Pharmacother\u003c/em\u003e. \u003cb\u003e155\u003c/b\u003e, 113599. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.biopha.2022.113599\u003c/span\u003e\u003cspan address=\"10.1016/j.biopha.2022.113599\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAstudillo, P. An emergent Wnt5a/YAP/TAZ regulatory circuit and its possible role in cancer. \u003cem\u003eSemin Cell. Dev. Biol.\u003c/em\u003e \u003cb\u003e125\u003c/b\u003e, 45\u0026ndash;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.semcdb.2021.10.001\u003c/span\u003e\u003cspan address=\"10.1016/j.semcdb.2021.10.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, W. et al. Wnt5a: a promising therapeutic target in ovarian cancer. \u003cem\u003ePathol. Res. Pract.\u003c/em\u003e \u003cb\u003e219\u003c/b\u003e, 153348. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.prp.2021.153348\u003c/span\u003e\u003cspan address=\"10.1016/j.prp.2021.153348\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Q. et al. Wnt5a-induced M2 polarization of tumor-associated macrophages via IL-10 promotes colorectal cancer progression. \u003cem\u003eCell. Commun. Signal.\u003c/em\u003e \u003cb\u003e18\u003c/b\u003e, 51. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12964-020-00557-2\u003c/span\u003e\u003cspan address=\"10.1186/s12964-020-00557-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNatu, J. \u0026amp; Nagaraju, G. P. Gemcitabine effects on tumor microenvironment of pancreatic ductal adenocarcinoma: special focus on resistance mechanisms and metronomic therapies. \u003cem\u003eCancer Lett.\u003c/em\u003e \u003cb\u003e573\u003c/b\u003e, 216382. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.canlet.2023.216382\u003c/span\u003e\u003cspan address=\"10.1016/j.canlet.2023.216382\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChitty, J. L. et al. A first-in-class pan-lysyl oxidase inhibitor impairs stromal remodeling and enhances gemcitabine response and survival in pancreatic cancer. \u003cem\u003eNat. Cancer\u003c/em\u003e. \u003cb\u003e4\u003c/b\u003e, 1326\u0026ndash;1344. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s43018-023-00614-y\u003c/span\u003e\u003cspan address=\"10.1038/s43018-023-00614-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePandit, B. \u0026amp; Royzen, M. Recent development of prodrugs of gemcitabine. \u003cem\u003eGenes (Basel)\u003c/em\u003e. \u003cb\u003e13\u003c/b\u003e, 466. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/genes13030466\u003c/span\u003e\u003cspan address=\"10.3390/genes13030466\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiao, H., Chen, X. \u0026amp; Luan, Y. Small molecular gemcitabine prodrugs for cancer therapy. \u003cem\u003eCurr. Med. Chem.\u003c/em\u003e \u003cb\u003e27\u003c/b\u003e, 5562\u0026ndash;5582. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2174/0929867326666190816230650\u003c/span\u003e\u003cspan address=\"10.2174/0929867326666190816230650\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDutra, J. A. P. et al. Temozolomide: an overview of biological properties, drug delivery nanosystems, and analytical methods. \u003cem\u003eCurr. Pharm. Des.\u003c/em\u003e 28, 2073\u0026ndash;; (2088). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2174/1381612828666220603152918\u003c/span\u003e\u003cspan address=\"10.2174/1381612828666220603152918\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZou, Y. et al. Brain co-delivery of temozolomide and cisplatin for combinatorial glioblastoma chemotherapy. \u003cem\u003eAdv. Mater.\u003c/em\u003e \u003cb\u003e34\u003c/b\u003e, e2203958. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/adma.202203958\u003c/span\u003e\u003cspan address=\"10.1002/adma.202203958\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRao, V. et al. Temozolomide resistance: a multifarious review on mechanisms beyond O-6-methylguanine-DNA methyltransferase. \u003cem\u003eCNS Neurol. Disord Drug Targets\u003c/em\u003e. \u003cb\u003e22\u003c/b\u003e, 817\u0026ndash;831. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2174/1871527321666220404180944\u003c/span\u003e\u003cspan address=\"10.2174/1871527321666220404180944\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, Z. C. et al. Temozolomide sensitizes ARID1A-mutated cancers to PARP inhibitors. \u003cem\u003eCancer Res.\u003c/em\u003e \u003cb\u003e83\u003c/b\u003e, 2750\u0026ndash;2762. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1158/0008-5472.CAN-22-3646\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-22-3646\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHui-Zheng, H. \u0026amp; Lei, L. Temozolomide for aggressive pituitary tumors: a literature review. \u003cem\u003eAsian J. Surg.\u003c/em\u003e \u003cb\u003e47\u003c/b\u003e, 1300\u0026ndash;1301. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.asjsur.2023.11.106\u003c/span\u003e\u003cspan address=\"10.1016/j.asjsur.2023.11.106\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJezierzański, M. et al. Temozolomide (TMZ) in the treatment of glioblastoma multiforme: a literature review and clinical outcomes. \u003cem\u003eCurr. Oncol.\u003c/em\u003e \u003cb\u003e31\u003c/b\u003e, 3994\u0026ndash;4002. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/curroncol31070296\u003c/span\u003e\u003cspan address=\"10.3390/curroncol31070296\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrasad, C. P., Chaurasiya, S. K., Guilmain, W. \u0026amp; Andersson, T. WNT5A signaling impairs breast cancer cell migration and invasion via mechanisms independent of the epithelial-mesenchymal transition. \u003cem\u003eJ. Exp. Clin. Cancer Res.\u003c/em\u003e \u003cb\u003e35\u003c/b\u003e, 144. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13046-016-0421-0\u003c/span\u003e\u003cspan address=\"10.1186/s13046-016-0421-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIlyas, S. I., Khan, S. A., Hallemeier, C. L., Kelley, R. K. \u0026amp; Gores, G. J. Cholangiocarcinoma: evolving concepts and therapeutic strategies. \u003cem\u003eNat. Rev. Clin. Oncol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 95\u0026ndash;111. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrclinonc.2017.157\u003c/span\u003e\u003cspan address=\"10.1038/nrclinonc.2017.157\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndreyev, J. et al. Guidance on the management of diarrhoea during cancer chemotherapy. \u003cem\u003eLancet Oncol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, e447\u0026ndash;e460. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S1470-2045(14)70006-3\u003c/span\u003e\u003cspan address=\"10.1016/S1470-2045(14)70006-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAltorki, N. K. et al. The lung microenvironment: an important regulator of tumour growth and metastasis. \u003cem\u003eNat. Rev. Cancer\u003c/em\u003e. \u003cb\u003e19\u003c/b\u003e, 9\u0026ndash;31. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41568-018-0081-9\u003c/span\u003e\u003cspan address=\"10.1038/s41568-018-0081-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiao, G. et al. Bevacizumab treatment of radiation-induced brain necrosis: a systematic review. \u003cem\u003eFront. Oncol.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 593449. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fonc.2021.593449\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2021.593449\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThery, C. et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. \u003cem\u003eJ. Extracell. Vesicles\u003c/em\u003e. \u003cb\u003e7\u003c/b\u003e, 1535750. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/20013078.2018.1535750\u003c/span\u003e\u003cspan address=\"10.1080/20013078.2018.1535750\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNam, S., Chung, J. W. \u0026amp; Yang, J. Y. WNT5A correlates with clinicopathological characteristics in gastric cancer: a meta-analysis. \u003cem\u003eCell. Physiol. Biochem.\u003c/em\u003e \u003cb\u003e41\u003c/b\u003e, 33\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1159/000455934\u003c/span\u003e\u003cspan address=\"10.1159/000455934\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoerens, A. G. et al. Functional T cells are capable of supernumerary cell division and longevity. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e614\u003c/b\u003e, 762\u0026ndash;766. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41586-022-05626-9\u003c/span\u003e\u003cspan address=\"10.1038/s41586-022-05626-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRui, R., Zhou, L. \u0026amp; He, S. Cancer immunotherapies: advances and bottlenecks. \u003cem\u003eFront. Immunol.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 1212476. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2023.1212476\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2023.1212476\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGentles, A. J. et al. The prognostic landscape of genes and infiltrating immune cells across human cancers. \u003cem\u003eNat. Med.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 938\u0026ndash;945. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nm.3909\u003c/span\u003e\u003cspan address=\"10.1038/nm.3909\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePushpakom, S. et al. Drug repurposing: progress, challenges and recommendations. \u003cem\u003eNat. Rev. Drug Discov\u003c/em\u003e. \u003cb\u003e18\u003c/b\u003e, 41\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrd.2018.168\u003c/span\u003e\u003cspan address=\"10.1038/nrd.2018.168\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLonguespee, R. et al. MALDI mass spectrometry imaging: a cutting-edge tool for fundamental and clinical histopathology. \u003cem\u003eProteom. Clin. Appl.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e, 701\u0026ndash;719. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/prca.201500140\u003c/span\u003e\u003cspan address=\"10.1002/prca.201500140\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Cholangiocarcinoma, Exosomes, WNT5A, PFN2, Diagnostic biomarkers, Tumor Microenvironment","lastPublishedDoi":"10.21203/rs.3.rs-8144185/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8144185/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCholangiocarcinoma (CCA) is an aggressive biliary malignancy with limited diagnostic tools and poor prognosis. Early detection remains challenging due to nonspecific symptoms and a lack of reliable biomarkers. Exosomes, as stable carriers of molecular cargos, have emerged as promising sources for non-invasive cancer biomarkers. Here, we integrated multiple GEO datasets to identify exosome-related differentially expressed genes (ERDEGs) associated with CCA. Through differential expression analysis, machine-learning feature selection, and immune infiltration profiling, we identified two key exosome-related genes, \u003cb\u003eWNT5A\u003c/b\u003e and \u003cb\u003ePFN2\u003c/b\u003e, as potential diagnostic biomarkers. Both genes showed robust diagnostic performance across internal and external validation cohorts. Functional enrichment revealed strong associations with extracellular matrix organization, EMT activation, and immune regulation pathways. Molecular docking suggested potential therapeutic compounds targeting these genes. Immunohistochemistry further confirmed significant overexpression of WNT5A and PFN2 in CCA tissues compared with adjacent controls. Collectively, our findings highlight WNT5A and PFN2 as promising exosome-related biomarkers that may improve early diagnosis and offer new therapeutic opportunities for cholangiocarcinoma.\u003c/p\u003e","manuscriptTitle":"Integrative Identification and Validation of Exosome-Related Genes as Diagnostic Biomarkers and Potential Therapeutic Targets in Cholangiocarcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-17 05:29:17","doi":"10.21203/rs.3.rs-8144185/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"08c9f9fe-7a34-4004-9d6c-6c7179fa4f6e","owner":[],"postedDate":"December 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":59706110,"name":"Health sciences/Biomarkers"},{"id":59706111,"name":"Biological sciences/Cancer"},{"id":59706112,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":59706113,"name":"Biological sciences/Immunology"},{"id":59706114,"name":"Health sciences/Oncology"}],"tags":[],"updatedAt":"2025-12-30T07:54:36+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-17 05:29:17","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8144185","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8144185","identity":"rs-8144185","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}