Prognostic Value of VPS45 in HCC and Its Correlation with Immune Microenvironment Authors: Jiming Duan, Wenjun Zhao, Na Lv, Changhu Xie, Guoping Zhang, Lei Zhang, Renhou Zhi Published: 2026 ABSTRACT AbstractPurposeA growing body of research indicates that Vacuolar Protein Sorting 45 (VPS45) is essential for the development of tumours and the advancement of cancer. Its biological role and expression profile in HCC are still mostly unknown, nevertheless.Materials and MethodsUsing a variety of datasets, including the Tumour Immune Estimation Resource (TIMER) and Gene Expression Profiling Interactive Analysis (GEPIA), among others, we examined the existence and prognostic importance of VPS45. Functional enrichment analysis was employed to elucidate the potential mechanisms through which VPS45 may influence HCC. Additionally, we assessed the relationship between VPS45 expression and immune cell infiltration using single-sample gene set enrichment analysis (ssGSEA) and Estimation of Stromal and Immune cells in Malignant Tumour tissues using Expression data (ESTIMATE). The Tumor Immune Dysfunction and Rejection (TIDE) algorithm was leveraged to assess the sensitivity of predicted gene expression to immunotherapy. Utilizing assays including CCK-8, wound healing, and Transwell migration and invasion methods, we performed in vitro experiments on the MHCC-97H cell line to further investigate the biological significance of VPS45 in the progression of hepatocellular carcinoma (HCC).ResultsAccording to our research, VPS45 expression is markedly elevated in HCC, and elevated VPS45 levels are associated with a worse prognosis for patients. Functional enrichment analysis showed that elevated VPS45 expression is closely linked to the activation of various oncogenic pathways. VPS45 silencing decreases HCC cell invasion, metastasis, and proliferation, according to in vitro investigations. Interestingly, it was shown that VPS45 levels showed a positive link with immunosuppressive cell populations and related genes, but a negative correlation with a number of anti-tumor immune cell types. Immunotherapy resistance is more common in patients with increased VPS45 expression.ConclusionThis study clarifies the biological role of VPS45 in hepatocellular carcinoma (HCC) and, through comprehensive database analyses combined with in vitro experiments, supports its potential as a diagnostic and predictive biomarker for HCC. In addition to its value for disease detection and prognosis, VPS45 may help predict patient response to immunotherapy.Keywords: hepatocellular carcinoma, VPS45, prognosis, immune infiltration, immunotherapy Introduction Hepatocellular carcinoma (HCC) ranks among the most prevalent and deadly malignancies globally, significantly contributing to cancer-related morbidity and mortality.1 It primarily arises in individuals with underlying chronic liver conditions, such as cirrhosis and infections with hepatitis B or C viruses.2 The pathogenesis of HCC is multifaceted, involving a combination of genetic and environmental factors. The accumulation of mutations, genomic instability, and the disruption of vital signalling pathways linked to angiogenesis, apoptosis, and cell proliferation are essential to its development.3 The prognosis for advanced HCC is still poor despite improvements in early identification and treatment options, such as liver transplantation, surgical resection, and systemic treatments.4 This is mainly because of high rates of recurrence and metastasis.5 The tumor microenvironment (TME) of HCC plays a central role in disease progression and is defined by complex interactions among immune cells, extracellular matrix components, and immunosuppressive signaling pathways.6,7 The immune landscape of HCC can be considered in terms of innate and adaptive components.8–10 Innate immune cells—such as tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), neutrophils, and dendritic cells—contribute to the establishment of an immunosuppressive microenvironment.8 Adaptive immunity is largely mediated by CD8+ cytotoxic T lymphocytes (CTLs), which exert anti-tumor effects, while regulatory T cells (Tregs) promote immune suppression; natural killer (NK) cells, although classically innate, also play an important role in anti-tumor responses.8 HCC tumors frequently evade immune surveillance by co-opting immune-regulatory pathways, including immune checkpoints such as PD-1/PD-L1, CTLA-4, LAG-3, and TIM-3.11–13 Although immune checkpoint inhibitors (ICIs) have reshaped the treatment landscape for advanced HCC, many patients exhibit primary or acquired resistance and derive limited clinical benefit.14 This heterogeneity in immunotherapy response highlights the urgent need to dissect the regulatory mechanisms operating within the HCC immune microenvironment and to identify new immune-related biomarkers and therapeutic targets. Additionally, alpha-fetoprotein (AFP) is the most commonly utilized biomarker for diagnosing and monitoring HCC.15 However, its sensitivity and specificity are suboptimal, with many patients presenting normal AFP levels, especially in early-stage HCC or in cases complicated by underlying liver cirrhosis.16 Given these limitations, there is a pressing need to discover innovative biomarkers that can offer more reliable and accurate diagnostic insights, as well as aid in prognostic evaluation and therapeutic decision-making for HCC patients. Vacuolar protein sorting 45 (VPS45) is a gene that encodes a protein crucial for vesicular trafficking, a fundamental process necessary for maintaining cellular homeostasis.17 VPS45 is part of a family of proteins essential for intracellular membrane fusion and endosomal sorting, which are vital for various cellular functions, including receptor trafficking and protein secretion.18 While research on VPS45 has primarily focused on prokaryotic models, its significance in eukaryotic systems, including humans, is gaining increasing recognition. VPS45 has been studied in several cancer types, where it has been implicated in the development of tumours and immunological regulation. For instance, elevated VPS45 expression levels are linked to a bad prognosis and promote malignant progression in lung cancer.19 Similarly, in ovarian cancer, VPS45 expression correlates with unfavorable outcomes and regulates the proportion of side population (SP) cells, thereby promoting tumor growth.20 In addition, our previous study demonstrated that VPS45 is significantly overexpressed in HCC tissues and cell lines. Mechanistic investigations indicate that VPS45 promotes HCC cell proliferation, migration, and invasion by activating the Wnt/β‑catenin signaling pathway.21 Beyond its direct effects on tumor cell malignancy, VPS45 has emerged as a regulator of immune processes. In NAFLD-related HCC mouse models, VPS45 expression correlated positively with the proportion of CD4+ T cells in liver tissue, suggesting a role in shaping immune cell infiltration within the HCC microenvironment.22 VPS45 is also a component of endocytosis-related prognostic gene signatures for HCC, and expression patterns of endocytosis-associated genes have been linked to the abundance of CD4+ T lymphocytes, natural killer cells, and macrophages.23 Notably, biallelic mutations in VPS45 underlie a congenital immunodeficiency characterized by defective endosomal protein trafficking and impaired neutrophil function, providing strong evidence that VPS45 plays an intrinsic role in immune cell biology.24 Despite mounting evidence—from mechanistic HCC studies, endocytosis-related prognostic models, exosome research linking VPS45 to immune infiltration, and congenital immunodeficiency caused by VPS45 mutations—no study has systematically evaluated the clinical-translational relevance of VPS45 in HCC from an integrated perspective. In particular, the prognostic significance of VPS45, its relationship with the tumor immune microenvironment, and its potential to predict immunotherapy response remain unclear. Therefore, this study was designed to address four key questions: (1) What is the expression pattern of VPS45 in HCC and how does it relate to patient prognosis? (2) How does VPS45 expression correlate with the tumor immune microenvironment and immune checkpoint molecules? (3) Can VPS45 expression predict response to immunotherapy? and (4) What is the biological role of VPS45 in HCC cell malignancy? In this study, we systematically investigated—for the first time—the role of VPS45 in HCC by integrating analyses of multiple public databases with in vitro experiments. We defined VPS45’s prognostic significance, elucidated its oncogenic mechanisms and immunoregulatory functions, examined its associations with the tumor immune microenvironment and immunotherapy resistance, and identified potential biomarkers for predicting response to HCC immunotherapy. Materials and Methods Patients and Samples This study examined postoperative pathological paraffin specimens from 82 patients who underwent surgical resection for hepatocellular carcinoma (HCC) at the Shanxi Bethune Hospital between March 2023 and January 2025 (Supplementary Document 1: Baseline clinical pathology characteristics). All participants had not received any non-surgical treatments prior to their surgeries and supplied written informed consent to participate in this investigation. The study was carried out in compliance with the protocol authorised by Shanxi Bethune Hospital’s Ethics Committee and the ethical guidelines set forth in the International Medical Association’s Declaration of Helsinki. The ethical approval number for this study is YXLL-2023-076. Research Data Sources The publicly available database The Cancer Genome Atlas (TCGA) (http://cancergenome.nih.gov/) included RNA sequencing (RNA-Seq) expression data for hepatocellular carcinoma (HCC).25 Alongside the gene expression data, relevant clinical pathology information was also downloaded. Gene expression (RNA-seq) data for TCGA-LUAD (release date: 07–20-2019) were obtained from the UCSC Xena platform (https://xenabrowser.net/datapages/). The sequencing was performed on the Illumina HiSeq 2000. For each sample, measurements from multiple aliquots were averaged and merged into a single gene-expression matrix. Expression values were normalized to FPKM (fragments per kilobase of transcript per million mapped reads) to account for gene length and library size, and then transformed using log2(x + 1).Following data sorting and filtering, 419 samples in all—369 HCC samples and 50 samples of normal liver tissue—were included in the analysis. TIMER and GEPIA Databases The TIMER2.0 database (Tumor Immune Estimation Resource 2.0) integrates a wealth of clinical data, including patient survival and disease status, enabling correlation analyses between immune cell infiltration and patient prognosis. This resource is instrumental in examining the relationship between tumor immune escape and clinical outcomes.26 TIMER2.0 encompasses multiple tumor types, including lung, breast, gastric, and colon cancers, providing extensive data on tumor immunity. Researchers can utilize this database to analyze the expression of genes linked to immunity, gaining insights into the molecular mechanisms underlying immunological reactions, as well as evaluating efficacy in immunotherapy, including immune checkpoint inhibitors. The GEPIA (Gene Expression Profiling Interactive Analysis) database serves as a public resource for cancer gene expression analysis, offering gene expression data across various tumor types through an interactive interface.27 It is built upon data from TCGA and the Genotype-Tissue Expression (GTEx) projects, providing comprehensive gene expression information. This resource supports tumor researchers in conducting differential gene expression analyses, prognostic evaluations, and functional research. CPTAC Database The Clinical Proteomic Tumor Analysis Consortium (CPTAC) is a research initiative dedicated to generating high-quality tumor proteome data for the cancer research community.28 Its database encompasses proteomic data from multiple cancer types, including tumor samples, paired normal tissues, and associated clinical information. By providing comprehensive datasets on protein expression profiles, mutations, and epigenetic features, CPTAC serves as a valuable resource for elucidating the molecular mechanisms of cancer, identifying potential biomarkers, and guiding the creation of innovative treatment approaches. Ultimately, CPTAC aims to advance personalized cancer medicine through the systematic analysis of protein-level alterations in tumors. UALCAN Database UALCAN is a bioinformatics platform designed for cancer research, enabling researchers to analyze gene expression differences across cancer types and investigate their correlations with clinicopathological features.29 This database offers important insights into cancer biology, supporting efforts in prognostic prediction and therapeutic target discovery. Establishing a Nomogram for Prognostic Prediction Using the R package “RMS,” a prognostic algorithm utilising both univariate and multivariate Cox regression analyses was created to estimate overall survival (OS) probabilities for hepatocellular carcinoma (HCC) patients at different years. The model’s prediction accuracy was evaluated using calibration curves. VPS45 expression and all available major clinicopathological variables (age, gender, AFP, TNM stage, vascular invasion, and ethnicity) were first assessed by univariate Cox regression. Variables with P < 0.05 in the univariate analysis were then entered into a multivariate Cox regression. Using these statistical criteria, independent prognostic factors for HCC were identified and incorporated into the final prognostic model. GO and KEGG Analysis Gene Ontology (GO) is a widely adopted bioinformatics resource that provides a standardized framework for annotating genes and their products. The Kyoto Encyclopedia of Genes and Genomes (KEGG) is another key bioinformatics database that integrates biological information, with a focus on pathways, diseases, and drugs. Differentially expressed genes between high- and low-expression groups based on VPS45 gene expression were then mapped to KEGG pathways to uncover potential biological mechanisms involved. This also provides a hypothetical basis for subsequent functional mechanism research. ESTIMATE Performs Estimation of Corresponding Components in the Tumor Microenvironment The Estimation of Stromal and Immune cells in Malignant Tumor tissues using Expression data (ESTIMATE) algorithm estimates the levels of infiltrating stromal and immune cells in tumor samples by leveraging gene expression data to evaluate tumor purity.30 Through enrichment analysis of sample-specific gene expression profiles, the algorithm generates three key scores: matrix, immune and estimated score. We employed the single-sample Gene Set Enrichment Analysis (ssGSEA) to assess the presence of immune cells in tumor samples and quantify the extent of immune cell infiltration.31 By analyzing the TCGA-LIHC dataset, we explored the relationship between VPS45 expression levels and the infiltration degrees of 28 distinct immune cell types. To visually represent these relationships, we employed R software packages, specifically Heatmap and ggplot2, to create correlation heat maps based on calculated correlation coefficients. Analysis of TIDE Algorithm in Evaluating Immunotherapy TIDE (Tumor Immune Dysfunction and Exclusion) is a computational framework that predicts the likelihood of tumor immune escape and response to immunotherapy.32 It assesses immune activity in the tumor microenvironment by integrating large-scale transcriptomic data with immune-related gene expression profiles. The method focuses on two key components: immune dysfunction, which reflects T cell exhaustion or impaired effector function, and immune rejection (also referred to as immune exclusion), which reflects physical or biochemical barriers that prevent immune cells from infiltrating the tumor. To investigate the potential role of VPS45 in predicting therapeutic response, some correlation analyses was performed using information from the GDSC (Genomics of Drug Sensitivity in Cancer) database.33 The R package “pRophetic” was employed for this purpose; it applies machine learning algorithms, trained on existing drug sensitivity datasets, to estimate drug response in tumor samples.34 Specifically, a Ridge regression model was implemented to predict the half-maximal inhibitory concentration (IC50) of various compounds for each patient sample. Based on the median expression level of VPS45, samples were divided into high- and low-expression groups. The variations in projected drug sensitivity between these groups were then examined. Cell Function Experiments CCK-8 Detection: MHCC-97H cells (purchased from Wuhan Procell Life Science and Technology Co., Ltd.) and their VPS45 knockdown variants were seeded in a 96-well plate at a density of 1 × 103 cells per well, with three replicate wells for each condition. At different hours post-seeding, 10 µL of CCK-8 solution was added to each well, followed by a 2-hour incubation. The intensity was measured at 450 nm using a Victor 31420 multi-label detector (PerkinElmer), and the proliferation curve was subsequently generated. Wound Healing Assay: When the cells in the 6-well plate reached 95–100% confluence, a sterile pipette tip was used to create a linear scratch across the cell monolayer. Following a wash with PBS, fresh medium was added, and the wound healing rate was recorded using a microscope at both 0 hours and 24 hours post-scratch. Transwell Assay: The upper chamber was pre-coated with Matrigel for the invasion test or left untreated for the migration test. A microscope was used to count the cells that passed through the membrane in eight randomly chosen fields. All in vitro functional assays (CCK‑8, wound‑healing, and Transwell migration and invasion) were conducted in three independent biological replicates, with three technical replicates performed for each biological replicate. The siRNA sequence of vps45 is provided in Supplementary Document 2. Immunohistochemistry (IHC) Paraffin-embedded specimens from 82 patients with HCC and adjacent non-tumor tissues were collected to assess VPS45 and PD-L1 expression using immunohistochemistry. The 82‑patient IHC cohort used in this study is the same cohort that was partially utilized in our previous publication (Zhi et al, Molecular Carcinogenesis, 2025) for limited VPS45 expression verification. In the present study, all IHC staining (including PD‑L1), scoring, and correlation analyses were performed de novo and are entirely new. The IHC assessment was scored by two independent pathologists blinded to the patient’s clinical data and was in good agreement (ICC=0.85, 95% CI =0.79–0.90).VPS45 staining intensity was categorized into four grades: 0 (no staining/−), 1 (weak staining/+), 2 (moderate staining/++), and 3 (strong staining/+++). The scoring for stained areas was defined as follows: 0 (0–10%), 1 (11–30%), 2 (31–69%), and 3 (over 70% stained areas). The expression score for VPS45 protein in each patient was determined by combining the staining region score by the strength of the stain score.35 For PD-L1, staining intensity was classified according to the following criteria: 0 (no staining), 1/+ (weak staining), 2/++ (moderate staining), and 3/+++ (strong staining). The area scores for PD-L1 were defined as 0 (non-stained area), 1 (1–5%), 2 (6–10%), and 3 (greater than 10%). Similar to VPS45, the expression of PD-L1 was quantified by combining the staining region score by the strength of the stain score for each patient.36 Ultimately, the relationship among VPS45 and PD-L1 expression was examined based on the calculated expression scores for both proteins. The antibodies used in this study included anti-VPS45 (catalog number 12006-1-AP, diluted 1:5000) and anti-PD-L1 (catalog number 28076-1-AP, diluted 1:2000), both sourced from Proteintech. Statistical Analysis This study utilized SPSS version 23.0 for statistical analysis, GraphPad Prism version 7 for visualization, and R language version 3.6.7 for additional data processing. The Student’s t-test was implemented to compare the variations in VPS45 expression levels across groups. Correlations between VPS45 gene expression and clinical pathological traits were examined using the chi-square test. Spearman correlation analysis assessed the association between VPS45 and other genes, as well as with ESTIMATE scores, TIDE scores, and immune cell infiltration. For correlation analyses involving multiple pairs of genes, we used the Benjamini-Hochberg (BH) procedure to control the hair-loss rate (FDR).A two-way analysis of variance was utilised to compare continuous variables between many groups. Post-hoc testing for two-way analysis of variance uses Tukey’s HSD test for multiple pairwise comparisons. The association between VPS45 and PD-L1 expression was assessed using Spearman correlation analysis. This was the definition of statistical significance: *P<0.05; **P<0.01; ***P<0.001, where <0.05 indicates a statistically significant difference between the data. Results Expression of VPS45 mRNA in Multiple Human Tumors and Its Prognosis The expression levels of VPS45 in different tumour categories were investigated using the TIMER 2.0 program. As shown in Figure 1A, the results showed that VPS45 was markedly overexpressed in malignant tumour tissues relative to their corresponding normal tissues in a variety of tumour types, including hepatocellular carcinoma (HCC). Notably, VPS45 expression was markedly heightened in cases of invasive breast cancer, bile duct carcinoma, head and neck squamous cell carcinoma, esophageal cancer, lung squamous cell carcinoma, lung adenocarcinoma, endometrial cancer, and HCC, with statistical significance (P< 0.001). VPS45 Expression in Human Cancers and Prognostic Impact. (A) VPS45 expression was analyzed in various cancers using the TIMER 2.0 tool. GEPIA database analysis of 33 cancers showed that high VPS45 expression was associated with poor overall survival (OS) (B) and disease-free survival (DFS) (C) in LIHC. Kaplan-Meier and Log rank tests from GEPIA showed that high VPS45 expression in LIHC was associated with significantly worse overall survival (P<0.001) (D) and disease-free survival (P<0.05) (E).(*p < 0.05, **p < 0.01, and ***p < 0.001). Further analysis using the GEPIA database assessed overall survival (OS) and disease-free survival (DFS) for VPS45 expression across 33 common human tumors, as depicted in Figure 1B and C. High levels of VPS45 expression in HCC (LIHC), low-grade glioma (LGG), adrenal cortical carcinoma (ACC), renal chromophobe cell carcinoma (KICH), and rectal adenocarcinoma (READ) were correlated with poor overall survival. Similarly, high VPS45 expression in ACC, acute myeloid leukemia (LAML), HCC, pancreatic carcinoma (PAAD), and gastric carcinoma (STAD) was linked to unfavorable disease-free survival outcomes. Additionally, the GEPIA database facilitated the use of the Log rank test and Kaplan-Meier analysis to compare survival outcomes between different VPS45 expression groups within HCC. Our findings indicate that elevated VPS45 expression is significantly associated with poorer prognosis in HCC patients: higher VPS45 levels correlated with reduced overall survival (P < 0.001) and shorter disease-free survival (P < 0.05) (Figure 1D–E). Differential Expression of VPS45 Gene in HCC and Its Correlation with Clinicopathological Characteristics Hepatocellular carcinoma (HCC) had considerably higher VPS45 gene expression than normal tissues, according to our analysis of the TCGA-LIHC dataset (Figure 2A and B, P < 0.001). Furthermore, VPS45 expression increased progressively with the advancement of tumor grade and pathological stage (Figure 2C and D, P < 0.001). Expression of VPS45 gene in HCC and its correlation with clinicopathological characteristics. Analysis of the TCGA-LIHC dataset shows significantly higher expression of VPS45 mRNA (A) in HCC compared to normal tissues (P<0.001).Analysis of the CPTAC dataset showed that the expression level of VPS45 protein (B) was significantly higher in HCC than in normal tissues (P<0.001).VPS45 expression progressively increases with higher tumor grade (C) and pathological stage (D) (P<0.001).(ns p ≥ 0.05, ***p < 0.001). Clinical Application Value of VPS45 Gene in HCC We discovered VPS45 as a possible independent prognostic predictor by incorporating additional clinicopathological traits linked to HCC prognosis and using both univariate and multivariate Cox regression models (Table 1).To further this analysis, we developed a nomogram incorporating race, TNM stage, and VPS45 expression to predict overall survival (OS) at 1, 3, and 5 years (Figure 3A). The calibration curve validated its model’s efficiency in estimating HCC mortality (Figure 3B). ROC curve analysis yielded low AUC values for both VPS45 and AFP, indicating limited utility as standalone diagnostic markers (Figure 3C). Notably, their AUCs were comparable, suggesting VPS45 performs at a level similar to AFP; however, neither achieved clinically acceptable diagnostic accuracy. Table 1Univariate and Multivariate Analyses of VPS45 Expression Influencing the Overall Survival Rate of Liver Cancer Patients in the TCGA-LIHC CohortCharacteristicsUnivariate AnalysesMultivariate AnalysesHazard RatioP valueHazard RatioP valueAge (year) > 60 vs ≤ 601.651 (1.014–2.688)0.043721.413 (0.8502–2.348)0.1823Gender Male vs Female0.6376 (0.3928–1.035)0.06861RACE2.357 (1.393–3.987)0.001392.296 (1.295–4.07)0.004438TNM stage III–IV vs I–II2.188 (1.322 −3.622)0.002331.777 (1.043–3.027)0.03443Vascular invasion1.704 (1.039–2.793)0.034751.43 (0.8582–2.382)0.1699AFP (ng)0.8315 (0.4888–1.414)0.496VPS45 expression High vs low2.03 (1.235–3.335)0.0052072.547 (1.505–4.31)0.0004933 Univariate and Multivariate Analyses of VPS45 Expression Influencing the Overall Survival Rate of Liver Cancer Patients in the TCGA-LIHC Cohort Establishment and application value of a prognostic model related to VPS45 gene expression in HCC. (A) Nomogram combining race, TNM stage, and VPS45 expression for predicting 1-, 3-, and 5-year overall survival (OS) in HCC. (B) Calibration curve validating the accuracy of the nomogram in predicting HCC(HCC) mortality. (C) ROC curve demonstrating that VPS45 has diagnostic ability comparable to AFP for HCC. Potential Molecular Mechanism of VPS45 Gene in the Progression of HCC We examined VPS45 expression data from the TCGA database using the R package edgeR to find differentially expressed genes (DEGs) between the VPS45 low and high expression groups in order to gain a deeper understanding of the signalling pathways in which VPS45 may be implicated in HCC regulation. The criteria for DEG identification included |Log2FoldChange| > 1 and p-value < 0.05. After DEGs were identified, possible pathways associated with the VPS45 gene were identified by Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis using the clusterProfiler program in R software. According to the GO enrichment study results, humoral immune responses, unsaturated fatty acid metabolism, small molecule catabolic processes, acute inflammatory responses, and fatty acid metabolism were among the biological processes (BP) that significantly enriched DEGs. Kinesin complexes, plasma lipoprotein particles, high-density lipoprotein particles, endocytic vesicle cavities, and the basal plasma membrane all showed notable enrichments in terms of cellular components (CC). Notable molecular functions (MF) identified included microtubule binding, aromatase activity, carboxylic acid binding, cargo receptor activity, and fatty acid binding (Figure 4A). Furthermore, KEGG pathway analysis highlighted key pathways associated with these DEGs, notably the Wnt/β-catenin pathway and cellular endocytosis (Figure 4B). Potential biological functions and carcinogenic mechanisms of VPS45 in the progression of HCC. (A) GO enrichment analysis of differentially expressed genes (DEGs) between VPS45Low and VPS45High groups reveals significant enrichment in biological processes (BP), Key cellular components (CC) and Molecular functions (MF). (B) KEGG pathway analysis highlights the key pathways associated with VPS45: Wnt/β-catenin signaling and endocytosis. Relationship Between VPS45 Gene Expression and Immune Microenvironment in HCC Based on the results of our previous analysis, the biological process (BP) analysis revealed that genes with differential expression (DEGs) are markedly favoured in humoral immune responses and acute inflammatory responses, suggesting a potential regulatory role for this gene in immune processes. To delve deeper into the specific relationship between VPS45 and the tumour immunity ecosystem, we applied the ESTIMATE algorithm to explore immune infiltration in hepatocellular carcinoma (HCC) samples. The ESTIMATE algorithm analysis demonstrated the degree of immune cell infiltration and the total ESTIMATE score, as well as the expression levels of VPS45 and different matrix components inside the tumour microenvironment, were negatively correlated (see Figure 5A). Notably, samples with high VPS45 expression exhibited significantly lower stroma scores, immune scores, and ESTIMATE scores in contrast to individuals whose VPS45 expression is low (Figure 5B). This finding underscores the critical role that VPS45 may play in modulating the tumor microenvironment, particularly in relation to invasion of immunological cells. Effect of VPS45 expression in HCC on tumor immune microenvironment. (A) ESTIMATE algorithm analysis shows a negative correlation between VPS45 expression and tumor microenvironment components, immune cell infiltration, and overall ESTIMATE scores. (B) VPS45 high expression in HCC(HCC) samples is associated with significantly lower stromal scores, immune scores, and ESTIMATE scores compared to VPS45 low expression samples. (C) ssGSEA analysis reveals that VPS45 expression is correlated with immune cells. (D) Comparison of immune cell infiltration between VPS45 high and low expression groups.(**p < 0.01, and ***p < 0.001). We used the single-sample gene set enrichment analysis (ssGSEA) tool to analyse 28 different immune cell groups in the tumour microenvironment in order to better understand the relationship amongst VPS45 expression and immune activity. According to our findings, active CD4 T cells and type 2 helper T cells were positively correlated with VPS45 expression. Conversely, A number of immune cell types showed a negative correlation with VPS45 expression, including eosinophils, immature dendritic cells, γδ T cells, effector memory CD8 T cells, monocytes, type 1 helper T cells, activated CD8 T cells, mast cells, natural killer (NK) cells, neutrophils, CD56bright NK cells, and myeloid suppressor cells (Figure 5C). A comparative evaluation of immune cell expression between the VPS45 high expression group and the low expression group indicated that type 2 helper T cells and activated CD4 T cells were considerably increased in the high expression cohort, whereas the numbers of eosinophils, immature dendritic cells, γδ T cells, type 1 helper T cells, effector memory CD8 T cells, monocytes, activated CD8 T cells, mast cells, NK cells, neutrophils, CD56bright NK cells, and myeloid suppressor cells were noticeably reduced (Figure 5D). These results suggest that VPS45 gene is involved in immune cell infiltration in the tumor microenvironment, and its elevated expression level may be related to the formation of an immunosuppressive environment in HCC. Immune-Related Genes and VPS45 Gene Expression in HCC We conducted a comprehensive analysis of immune-related genes to investigate the relationship between VPS45 expression and various immune-associated gene categories, including chemokine genes, receptor genes, major histocompatibility complex (MHC) genes, immune checkpoint genes, and immune activation genes.37 Our outcomes show a positive correlation between VPS45 and a majority of immune-activating genes, particularly CD276, ENTPD1, MICB, and TNFSF4 (Figure 6A). Additionally, VPS45 exhibited a strong positive correlation with most MHC genes, especially TAP1 (Figure 6B). Furthermore, we observed notable positive correlations with various receptor genes, including significant associations with CCR10, CCR3, CX3CR1, and CXCR4 (Figure 6C). Among chemokine genes, VPS45 displayed significant positive correlations with CCL26, CCL28, CXCL1, and CXCL5, whereas revealing a essential negative correlation with CCL14 and CCL16 (Figure 6D). These results suggest that VPS45 is intricately linked with multiple immune-related genes. Correlation analysis between VPS45 expression and immune-related genes in HCC. (A) VPS45 expression is positively correlated with most immune activation genes. (B) VPS45 shows a positive correlation with most major histocompatibility complex (MHC) genes. (C) VPS45 expression is positively correlated with most receptor genes. (D) VPS45 correlates with chemokine genes. (E) VPS45 expression is positively correlated with most immune checkpoint genes. (F) Expression of immune checkpoint inhibitors, such as CD274 (PD-L1), ADORA2A, CD276, HAVCR2, LGALS3, and VTCN1, is significantly higher in the VPS45 high expression group compared to the VPS45 low expression group.(*p < 0.05, **p < 0.01, and ***p < 0.001). We conducted a correlation study among VPS45 and 20 inhibitory immune checkpoint genes in order to further evaluate the significance of VPS45 in the context of immunotherapy. The findings suggested a positive correlation among the VPS45 gene and several immune checkpoint genes, including CD274 (PD-L1), PDCD1 (PD-1), ADORA2A, CD276, CTLA4, HARAVCR2, LGALS3, and VTCN1 (Figure 6E). Notably, we discovered that immune checkpoint suppressor genes like ADORA2A, CD274 (PD-L1), CD276, HARAVCR2, LGALS3, and VTCN1 showed differential expression between the low and high VPS45 expression groups, with noticeably higher levels seen in the high VPS45 expression group (Figure 6F). Correlation Between VPS45 and Prediction of Immunotherapy for HCC and Screening of Sensitive Drugs Immunotherapy has become a major therapeutic option in addition to traditional surgical resection for the treatment of hepatocellular carcinoma (HCC). To investigate the connection among VPS45 expression and tumor response to immunotherapy, this study employed the Tumor Immune Disorder and Exclusion Characteristics (TIDE) algorithm, utilizing VPS45 expression levels to forecast how well HCC patients will respond to immune checkpoint medications. The results showed a positive relationship between TIDE scores and VPS45 gene expression (Figure 7A). Subsequent investigation revealed that patients with high VPS45 expression had a much lower immunotherapy response rate than those with low VPS45 expression (Figure 7B). Additionally, our analysis indicated that the VPS45-high expression group displayed higher TIDE and Exclusion scores compared to the low VPS45 expression cohort (Figure 7C and E, P<0.001), while the VPS45-high group showed lower Dysfunction scores than the VPS45-low group (Figure 7D, P<0.001). TIDE analysis suggested that VPS45-overexpressing group may have a higher risk of immunotherapy resistance. Moreover, we assessed the relationship among VPS45 gene expression levels and the sensitivity to four antitumor small molecule drugs: cisplatin, methotrexate, mitomycin C, and imatinib. Our findings suggested that HCC patients with higher VPS45 expression displayed lower IC50 values for cisplatin, methotrexate, and mitomycin C compared to those with lower VPS45 expression (Figure 7F–H). In contrast, patients with high VPS45 expression demonstrated higher IC50 values for imatinib relative to their low-expression counterparts (Figure 7I). Previous studies have confirmed that both VPS45 mRNA and protein are highly expressed in the MHCCC97H cell line, which we utilized in this study to further validate the malignant phenotype of HCC. Predicting the impact of VPS45 expression in HCC on immunotherapy response. (A) VPS45 expression is positively correlated with TIDE (Tumor Immune Dysfunction and Exclusion) scores in HCC. (B) Patients with high VPS45 expression show significantly lower response rates to immunotherapy compared to those with low VPS45 expression. Higher VPS45 expression was associated with higher TIDE (C), lower dysfunction score (D), and higher Exclusion (E) score compared with the low expression group (P<0.001).Patients with high VPS45 expression show lower IC50 values for cisplatin (F), methotrexate (G), and mitomycin C (H). (I) In contrast, high VPS45 expression is associated with higher IC50 values for imatinib. Knockdown of VPS45 Inhibits Malignant Biological Behavior of HCC Cells Using the CCK-8 assay, we assessed the impact of VPS45 on the proliferation of HCC cells, demonstrating that, in comparison to the sh-NC group, the ability of cells infected with sh-VPS45 to proliferate was much reduced (Figure 8A). Furthermore, scratch assays demonstrated a marked inhibition of hepatic fine cell migration following VPS45 knockdown (Figure 8B). Transwell assays further corroborated these findings, showing that VPS45 knockdown notably impaired the invasion and migration capabilities of HCC cells (Figure 8C). The knockdown effect of MHCC-97H is shown in Supplementary Figure. Tumor promoting function of VPS45 in HCC. (A) CCK-8 assay shows significantly reduced proliferation in sh-VPS45-infected HCC cells compared to the sh-NC group. (B) Scratch assay demonstrates a marked inhibition of cell migration following VPS45 knockdown. (C) Transwell assay confirms that VPS45 knockdown significantly reduces migration and invasion capabilities of HCC cells.(*p < 0.05, **p < 0.01, and ***p < 0.001). Expression of VPS45 in HCC and Its Correlation with Immunosuppressive Gene PD-L1 Immunohistochemical analyses uncovered an increase in VPS45 and PD-L1 expression in HCC tissues correlated with tumor grade, thereby supporting the biological analysis conducted earlier in this study (Figure 9A). Additionally, Spearman correlation analysis confirmed VPS45 and PD-L1 expression in HCC are positively correlated (Figure 9B). Correlation analysis of VPS45 and PD-L1 expression in HCC. (A) Immunohistochemistry results show that VPS45 and PD-L1 expression levels progressively increase with tumor grade in HCC, confirming previous bioinformatic analysis. (B) Pearson correlation analysis reveals a positive correlation between VPS45 and PD-L1 expression in HCC. Discussion Our research systematically elucidates the function of the VPS45 gene in hepatocellular carcinoma (HCC) through comprehensive bioinformatics analysis and basic experimental validation, focusing on its carcinogenic effects, prognostic significance, influence on the tumor immune microenvironment, and clinical therapeutic implications. Our initial findings, derived from multiple public databases (including TIMER2.0 and GEPIA), confirm that VPS45 is consistently overexpressed not only in HCC but also in various other cancers. Moreover, this overexpression correlates significantly with poor overall survival (OS) and disease-free survival (DFS) among HCC patients. This preliminary observation suggests that VPS45 may become a candidate target for HCC treatment, but its clinical transformation value requires more comprehensive functional verification (including multiple cell lines, in vivo models, and targeted intervention experiments). Further analysis of data from TCGA-LIHC revealed that VPS45 is markedly elevated in HCC at both the protein and mRNA levels. We also found a favourable relationship between VPS45 expression and both pathological stage and tumor grade, reinforcing the notion that VPS45 contributes to the malignant progression of HCC. Crucially, VPS45 was found to be an independent predictive factor for HCC using multivariate Cox regression analysis. The nomogram developed based on these findings exhibited strong predictive capabilities, with ROC curve analysis indicating that VPS45 has diagnostic potential comparable to that of alpha-fetoprotein (AFP) in HCC. To further explore the molecular mechanisms basis of VPS45-driven HCC progression, we conducted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. The results highlighted significant enrichment in the Wnt/β-catenin signaling pathway and cellular endocytosis, aligning with our previous research findings. The Wnt/β-catenin pathway is among the most frequently activated signaling cascades in HCC;38 its aberrant activation is known to promote cell proliferation, invasion, and stem cell characteristics. Our data suggest that VPS45 may be involved in activating this pathway to play a regulatory role, thereby promoting the malignant phenotype of HCC. Additionally, VPS45 functions as a key regulatory protein in the cellular endocytosis process; dysregulation of its expression could directly impact the internalization and signaling of cell membrane receptors, such as growth factor receptors, providing new insights into its role as a promoter of cancer progression. One of the most enlightening findings of this study is the intricate relationship between VPS45 and the hepatocellular carcinoma (HCC) tumor immune microenvironment (TIME). Analysis utilizing the ESTIMATE algorithm showed that high VPS45 expression levels were negatively linked to overall immune infiltration scores, including matrix and immune scores. This observation suggests that tumors characterized by elevated VPS45 expression may be in an “immune desert” or exhibit an “immune rejection” state. Detailed analysis through single-sample Gene Set Enrichment Analysis (ssGSEA) across 28 immune cell subsets paints a more precise picture: high VPS45 expression is significantly connected to reduced infiltration of various immune cells with antitumor activity, such as effector memory CD8 T cells and type 1 helper T cells, while showing a positive correlation with certain immune-suppressive cell types, including type 2 helper T cells. This distinct pattern of immune cell infiltration strongly indicates that HCC with high VPS45 expression fosters an immunosuppressive microenvironment. This inference is further validated at the mechanistic level through our examination of immune-related genes. VPS45 shows a significant positive correlation with multiple critical immune checkpoint receptors and ligands, such as PDCD1, CD274/PD-L1, and CTLA4, all of which are upregulated in VPS45-overexpressing groups. The binding of receptor PD-1 to PD-L1 represents a central mechanism of tumor immune evasion, and immune checkpoint inhibitors (ICIs) that target this pathway have become pivotal in HCC treatment.39–41 Our wet laboratory experiments corroborated this bioinformatics finding through immunohistochemistry, demonstrating that VPS45 and PD-L1 expression levels in HCC tissues not only increased with tumor grade but also exhibited a significant positive correlation. This suggests that VPS45 may work synergistically to promote immune evasion in HCC by upregulating various immune checkpoints, including PD-L1. Given these findings, we were naturally inclined to explore whether VPS45 expression levels influence ICI efficacy. TIDE analysis yielded affirmative results, indicating that high VPS45 expression corresponds with raised TIDE scores and exclusion scores, which suggest a decreased response rate to ICI treatment among these patients. This finding presents a potential biomarker for clinical screening of individuals who might benefit from immunotherapy. Furthermore, in our drug sensitivity analysis, we discovered that HCC cells with elevated VPS45 expression exhibited increased sensitivity to cisplatin, methotrexate, and mitomycin C, while showing reduced sensitivity to imatinib. This indicates that VPS45 expression levels could serve as a supplementary tool in guiding chemotherapy and targeted therapy for HCC, with patients exhibiting high VPS45 levels potentially deriving greater benefit from cisplatin-based regimens. In conclusion, we confirmed the biological function of VPS45 through in vitro experiments, wherein targeted knockdown of VPS45 in the MHCC97H cell line significantly impaired cell proliferation, migration, and invasive capabilities. These results directly underscore the necessity of VPS45 in sustaining the malignant phenotype of HCC cells, providing strong experimental support for its potential as a therapeutic target. Alpha-fetoprotein (AFP) is the most commonly used serum marker for hepatocellular carcinoma (HCC), but its prognostic utility is limited: only about 50–60% of HCC patients have elevated AFP, and its specificity is suboptimal.42–44 Emerging molecular markers—such as miRNAs,45,46 lncRNAs,47,48 circRNAs,49,50 and various protein biomarkers51,52—have shown promising prognostic potential, yet most remain at the research stage and lack validation in large, multicenter prospective studies. Therefore, VPS45 could potentially help address current diagnostic and therapeutic challenges in HCC. This study provides novel insights into VPS45 as a potential biomarker for HCC by integrating database analyses with in vitro experiments to characterize its expression, prognostic significance, association with the immune microenvironment, and biological functions. However, several limitations should be noted: (1)TCGA‑LIHC data were normalized using RSEM with a log2 transformation and were not batch‑corrected. Because the analysis was confined to a single, uniformly processed LIHC cohort, any batch effects are expected to be minimal and batch correction was deemed unnecessary.(2) Incomplete recording of key clinical variables (eg, HBV/HCV status and Child–Pugh classification) reduces the statistical power of multivariate analyses and limits the model’s generalizability.(3) Immune infiltration and predicted immunotherapy response were inferred from transcriptome‑based algorithms (ssGSEA, ESTIMATE, TIDE), providing indirect estimates that may not fully reflect the true tumor immune microenvironment.(4) Immunohistochemical validation was performed on a small, single‑center cohort; larger, multi‑ethnic, multicenter studies are needed for confirmation. Unlike our prior mechanistic study of VPS45/Wnt–β‑catenin (Zhi et al, 2025), this work focuses on the clinical‑translational utility of VPS45 as a prognostic and immune biomarker. Although both studies used the same 82‑patient IHC cohort and the TCGA‑LIHC dataset, the present analyses are entirely new and more comprehensive: we performed de novo IHC staining with composite scoring, systematic clinicopathological and immune‑correlation analyses, multi‑database cross‑validation, immune microenvironment characterization, and predictive modeling—none of which were reported previously. Thus, the two studies are complementary rather than overlapping, extending understanding of VPS45 from mechanism to clinical relevance. Conclusion We established, for the first time, a systematic framework linking VPS45 expression, prognosis, biological function, and immune contexture in HCC, introducing new candidate nodes to the regulatory network of the tumor immune microenvironment. Our findings suggest VPS45 may contribute to HCC progression: elevated VPS45 expression associates with poorer patient outcomes, indicating potential utility as a prognostic marker. Additionally, VPS45 expression correlates with immunosuppressive features and activation of oncogenic pathways, which could collectively promote resistance to immunotherapy. Although these observations are exploratory, they point to VPS45 as a possible therapeutic target; nevertheless, confirmation in larger patient cohorts and mechanistic studies in experimental models is required before clinical translation. Acknowledgments This study was funded by the Fundamental Research Program of Shanxi Province (No. 202403021221259). Abbreviations VPS45, Vacuolar Protein Sorting 45; HCC, Hepatocellular carcinoma. Disclosure Jiming Duan reports Support for the manuscript from Provision of study, materialsMedical writing, during the conduct of the study. The authors report no other conflicts of interest in this work. References