USP20 as a key regulator of immunosuppression and a novel predictor of poor prognosis in lung adenocarcinoma Authors: Longhao Wang, Danni Qiao, Ruixue Gu, Qingyu Zhang, Yan Guo, Lina Liu, Mingqi Zhu, Yaliang Fan, Peng Zhang, Qiongyang Lv, Junya Yan, Shundong Cang Published: 2026 ABSTRACT BackgroundTo systematically delineate the expression landscape, genetic alterations, prognostic value, and immunological relevance of ubiquitin-specific protease 20 (USP20) in lung adenocarcinoma (LUAD). By integrating bulk and single-cell transcriptomic analyses, we aimed to identify the cellular sources of USP20 and clarify its potential roles in remodeling the tumor immune microenvironment (TME), thereby evaluating its utility as a prognostic biomarker and candidate immunotherapeutic target.MethodsUSP20 expression was profiled across The Cancer Genome Atlas (TCGA), Genotype-Tissue Expression (GTEx), and multiple Gene Expression Omnibus (GEO) cohorts. Prognostic significance was assessed using Kaplan-Meier analysis and Cox proportional hazards models. USP20 genetic alterations were interrogated via cBioPortal. Differentially expressed genes (DEGs) associated with USP20 were identified, followed by GO/KEGG enrichment and GSEA to infer biological functions. TME characteristics were evaluated using ESTIMATE and CIBERSORT, and associations with immune checkpoint molecules were examined. Consensus clustering was performed to define USP20-related molecular subtypes and compare their immune landscapes. Single-cell RNA-seq datasets (TISCH and GSE146100) were integrated to resolve USP20 expression across tumor and immune compartments. Finally, immunohistochemistry, multiplex immunofluorescence, flow cytometry, and in vitro/in vivo functional assays were conducted to validate the biological and immunoregulatory roles of USP20 in LUAD.ResultsUSP20 was consistently upregulated in LUAD relative to normal lung tissues across independent cohorts. High USP20 expression was associated with inferior overall survival and remained an independent prognostic risk factor. Genetic profiling showed a low mutation rate of USP20, with copy number amplification as the predominant alteration. Functional analyses indicated that USP20-associated DEGs were enriched in immune-related programs, including T-cell activation, antigen processing and presentation, interferon signaling, and immune checkpoint regulation. High USP20 expression correlated with a distinct immune contexture, featuring reshaped immune cell composition, increased infiltration of immunosuppressive populations, and elevated expression of multiple immune checkpoint molecules. Consensus clustering further revealed USP20-related molecular subtypes with divergent immunosuppressive features and clinical outcomes. Single-cell analyses demonstrated heterogeneous and cell-type-specific USP20 expression across malignant and immune populations, supporting a context-dependent immunomodulatory role. Experimental validation confirmed that USP20 promotes LUAD cell proliferation and migration and is associated with immune dysregulation within the TME.ConclusionsUSP20 is aberrantly overexpressed in LUAD and portends poor prognosis. USP20 may facilitate disease progression by shaping tumor-immune interactions, including immune infiltration patterns, T-cell activation/exhaustion states, and immune checkpoint signaling. USP20 and its related molecular subtypes may serve as prognostic biomarkers and potential indicators of immunotherapeutic responsiveness, offering insights for precision immunotherapy in LUAD.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13062-026-00852-x. Introduction Lung adenocarcinoma (LUAD), the predominant histological subtype of non-small cell lung cancer (NSCLC), accounts for more than 40% of lung cancer cases worldwide, with a continuously increasing incidence and a noticeable trend toward younger onset [1]. Owing to the lack of specific symptoms at early stages, most patients are diagnosed at locally advanced or metastatic stages, resulting in limited therapeutic options and a 5-year survival rate of less than 20% [1–2]. Although targeted therapies against molecular alterations such as EGFR, ALK, ROS1, and KRAS have significantly improved outcomes in selected patient populations, a substantial proportion of LUAD patients lack actionable driver mutations, and acquired resistance to targeted therapy remains inevitable. In recent years, immune checkpoint inhibitors (ICIs) have introduced a major breakthrough in the treatment of advanced LUAD; however, their overall response rate remains approximately 20–30%, and some patients even experience hyperprogressive disease, which limits their broader clinical application [3–4]. Therefore, there is an urgent need to identify novel therapeutic targets and reliable biomarkers from both molecular and immunological perspectives to improve patient survival. The ubiquitin-proteasome system (UPS) plays a critical role in maintaining intracellular protein homeostasis and regulating key biological processes, including signal transduction, cell-cycle progression, and apoptosis [5]. Deubiquitinating enzymes (DUBs), essential components of the UPS, remove ubiquitin chains from substrate proteins, thereby modulating protein stability, subcellular localization, and functional activity. Accumulating evidence indicates that dysregulation of DUBs contributes to tumor initiation, invasion, metastasis, and therapeutic resistance across multiple cancer types [6]. Consequently, DUBs have emerged as promising anticancer targets, among which the ubiquitin-specific protease (USP) family has attracted considerable attention due to its substrate specificity and well-characterized regulatory mechanisms. USP20 (ubiquitin-specific protease 20) is an important member of the USP family, and increasing evidence has highlighted its critical role in tumor biology. On the one hand, USP20 can stabilize key signaling molecules through deubiquitination, including β2-adrenergic receptor, HIF-1α, and p65, thereby regulating NF-κB activation, cell adhesion, metabolic reprogramming, and hypoxic responses [7–9]. On the other hand, previous studies have reported that USP20 promotes migration and invasion of gastric cancer cells [10] and enhances the growth of cervical cancer cells [11–14], suggesting a tumor-promoting role of USP20 in certain malignancies. However, the expression pattern of USP20 in LUAD, its prognostic significance, and its potential involvement in remodeling the tumor immune microenvironment (TME) remain largely unexplored in a systematic and multidimensional manner. The TME plays a central role in determining tumor progression, immune evasion, and responsiveness to immunotherapy. Factors such as immune cell infiltration, expression of immunosuppressive molecules, cell-cell interactions, and inflammatory signaling pathways critically influence patient sensitivity to ICIs [15–17]. Therefore, elucidating the role of USP20 from an immunological perspective is essential for understanding its potential value as an immune-related biomarker. With the rapid development of large-scale databases, including TCGA, GTEx, and GEO, as well as immune deconvolution algorithms such as ESTIMATE, CIBERSORT, and ssGSEA, it has become feasible to systematically evaluate the associations between gene expression, prognosis, and immune infiltration at the tissue level [18]. Moreover, advances in single-cell RNA sequencing (scRNA-seq) and the availability of resources such as the TISCH database enable high-resolution characterization of gene expression across distinct cellular populations, providing powerful tools to dissect gene-specific regulatory roles within the complex TME [19–20]. Based on these considerations, the present study integrates multi-database and multi-omics approaches to comprehensively evaluate the expression patterns, genetic alterations, and prognostic significance of USP20 in LUAD, and to systematically investigate its associations with the TME, including immune infiltration patterns, immune checkpoint expression, immune scores, and TME subtypes. Furthermore, single-cell transcriptomic data were employed to delineate the cellular sources of USP20 expression across immune and tumor cell populations, thereby exploring its potential immunoregulatory mechanisms. This study aims to provide a theoretical basis for USP20 as a potential prognostic biomarker and immunotherapeutic target, offering new insights for precision treatment strategies in LUAD. Materials and methods Public database data acquisition and preprocessing Transcriptomic data and corresponding clinical information of LUAD were obtained from The Cancer Genome Atlas (TCGA) database. Normal lung tissue transcriptomic data were retrieved from the Genotype-Tissue Expression (GTEx) database. TCGA and GTEx datasets were uniformly normalized via the UCSC Xena platform using the Toil pipeline, and transcripts per million values were used for subsequent analyses. External validation datasets (GSE31210, GSE30219, GSE72094, GSE41271, and GSE50081) were downloaded from the Gene Expression Omnibus (GEO) database and processed using the limma R package for background correction and batch effect removal. scRNA-seq data were obtained from the GSE146100 dataset and the TISCH database to analyze USP20 expression patterns across different cell types. Differential expression and pan-cancer analysis of USP20 Differential expression of USP20 between tumor and normal tissues was analyzed using the Wilcoxon rank-sum test. Pan-cancer expression analysis of USP20 was performed based on TCGA multi-cancer datasets, and the results were visualized using boxplots. Mutation spectrum analysis Genetic alteration data of USP20 were retrieved from cBioPortal (TCGA Pan-Cancer Atlas). Analyses included mutation frequency, mutation types, and copy number variations. The maftools R package was used to generate Oncoplots and lollipop plots to illustrate the distribution of mutation sites across protein domains. Survival analysis Kaplan-Meier survival analysis was performed to evaluate the association between USP20 expression and overall survival (OS) as well as disease-specific survival (DSS). The optimal cutoff value for USP20 expression was determined using the maxstat method, and patients were stratified into high- and low-expression groups accordingly. Differentially expressed genes (DEGs) and functional enrichment analysis DEGs between high- and low-USP20 expression groups were identified using the limma package, with thresholds of |log2 fold change| ≥ 1 and false discovery rate (FDR) < 0.05. Gene Ontology (GO; Biological Process) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted using the clusterProfiler package, and results were visualized as bubble plots and bar charts. Gene Set Enrichment Analysis (GSEA) was additionally performed to identify immune-related signaling pathways enriched in the USP20 high-expression group, with gene sets obtained from the Molecular Signatures Database. TME analysis The ESTIMATE algorithm was applied to calculate ImmuneScore, StromalScore, and ESTIMATEScore for LUAD samples, reflecting the relative proportions of immune and stromal components. The CIBERSORT algorithm, using the LM22 signature matrix with 1,000 permutations, was employed to estimate the relative abundance of 22 immune cell types within tumor tissues. Expression levels of immune checkpoint-related genes, including PDCD1, CD274, CTLA4, LAG3, and TIGIT, were compared between high- and low-USP20 expression groups. Consensus clustering analysis Consensus clustering analysis was performed based on USP20-related gene expression profiles using the ConsensusClusterPlus R package. A total of 1,000 resampling iterations were conducted, and the optimal number of clusters (K = 4) was determined according to the cumulative distribution function (CDF) curves and delta area plots. Principal component analysis (PCA) and heatmaps were used to validate the stability and rationality of the identified molecular subtypes. scRNA-seq analysis scRNA-seq data were analyzed using the Seurat R package for quality control, normalization, dimensionality reduction, and clustering. Cell types were annotated based on established marker genes, and the distribution of USP20 expression was examined across different immune cell subsets, including CD8⁺ T cells, exhausted CD8⁺ T cells, CD4⁺ T cells, natural killer (NK) cells, and monocytes/macrophages. Human cohort We collected peripheral blood samples from patients diagnosed with lung adenocarcinoma at Henan Provincial People’s Hospital. All patients signed written informed consent forms, and this study was approved by the review committee of Henan Provincial People’s Hospital. Special medical staff collected the peripheral blood samples from the patients after obtaining their informed consent. The collected blood samples were then transported under 4 °C conditions to the laboratory. Gradient centrifugation was used to obtain the peripheral blood mononuclear cells (PBMC) and the supernatant of the patients. The obtained PBMC were cultured in a 10% RPMI 1640 medium. Formalin-fixed paraffin-embedded (FFPE) tumor specimens were used for immunohistochemistry and multiplex immunofluorescence analyses. Informed consent was obtained from all patients, and the study was approved by the Institutional Ethics Committee in accordance with the Declaration of Helsinki. Immunohistochemistry (IHC) Tissue samples used for IHC were obtained from Henan Provincial People’s Hospital with approval from the institutional ethics committee. USP20 expression was detected by IHC staining using a specific USP20 antibody (1:100 dilution). Briefly, the tissue array section (4 μm) was dehydrated, and peroxidase was blocked. A 0.01 mol/L citric acid buffer (pH 6.0) and a pressure cooker were used for antigen recovery. Tissues were then incubated with primary antibodies overnight at 4 °C, then stained with tissue staining kits (SP-9000) and 3,3’-diaminobenzidine tetrahydrochloride (DAB) (ZLI-9032; ZSGB-BlO, Beijing, China). The slides were counterstained with hematoxylin and eosin (HE). The stained glass slides were observed and an image was obtained through a microscope. Based on the staining intensity, the samples were divided into four groups, from the lowest intensity (-) to the highest intensity (+++) groups. Cell culture Human LUAD cell lines A549 and H3122 were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured at 37 °C in a humidified atmosphere containing 5% CO₂ using high-glucose Dulbecco’s Modified Eagle Medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% penicillin/streptomycin (Gibco). USP20-targeting shRNA was introduced into cells via lentiviral transduction, and stable knockdown cell lines were selected using puromycin. Colony formation assay A total of 800 control or USP20-knockdown cells were seeded into 6-well plates and cultured for 12–14 days. Colonies were subsequently fixed with 4% paraformaldehyde (PFA; Biosharp) and stained with 0.1% crystal violet (Solarbio). After air-drying, plates were photographed and colonies were quantified using ImageJ software. Cell migration assays Cell migration assays were performed using Transwell chambers. A total of 3 × 10⁴ cells were seeded into the upper chambers in 200 µL serum-free DMEM, while the lower chambers were filled with 500 µL DMEM containing 10% FBS. After incubation for 24–48 h, non-migrated cells on the upper surface were removed with cotton swabs. Migrated cells were fixed with 4% PFA, stained with crystal violet, photographed, and counted in high-power fields under a microscope. Western blotting Cells were lysed using radioimmunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) supplemented with a protease inhibitor cocktail (Thermo Scientific, Waltham, MA, USA) and phenylmethanesulfonyl fluoride (PMSF, Beyotime, Shanghai, China). The bicinchoninic acid (BCA) method was used to measure protein concentration. Cell lysates (40 µg) were subjected to sodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE), transferred to polyvinyldene fluoride (PVDF) membrane (Sigma, St Louis, MO, USA). Autoradiograms were densitometrically quantified using Quantity One software (Bio-Rad), with β-actin/GAPDH (Beyotime, Shanghai, China) serving as internal controls. Animal experiments A murine syngeneic tumor model was established by subcutaneous injection of LLC cells into immunocompetent C57BL/6 mice. All animals were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. For tumor implantation, 2 × 10⁶ LLC, LLC-shUSP20#1, or LLC-shUSP20#2 cells were subcutaneously injected into the flanks of C57BL/6 mice. Tumor-bearing mice were randomly assigned into three groups, with five mice per group (n = 5). Tumor growth was monitored regularly, and body weight was recorded throughout the experiment. Tumor size was measured using digital calipers, and tumor volume was calculated using the ellipsoid formula: volume = (length × width²) / 2. At the experimental endpoint, all mice were euthanized, and tumors were surgically excised. Tumor tissues were collected and either snap-frozen in liquid nitrogen or fixed in 4% PFA for subsequent analyses. Investigators were blinded to group allocation during data collection and subsequent data analysis. All animal experiments were conducted under specific pathogen-free (SPF) conditions with ad libitum access to food and water. All experimental procedures involving animals were approved by the Institutional Animal Care and Use Committee of Zhengzhou University and performed in accordance with institutional guidelines. Flow cytometric analysis Cells were stained with FITC anti-human CD4 Antibody, APC anti-human CD25 Antibody, PE anti-human FOXP3 Antibody, PerCP/Cyanine5.5 anti-human CD8a Antibody and Brilliant Violet 421™ anti-human CD279 (PD-1) Antibody according to the manufacturer’s instructions (BioLegend, USA). Among them, FOXP3 were used for intracellular staining. Then, the samples were collected by flow cytometry (BD FACSCanto l, USA) according to the manufacturer’s instructions, and data analysis was conducted with FlowJO software (version 10.0, Tree Star, Inc. Ashland, OR, USA). Multiplex immunofluorescence (mIF) We selected typical representatives with low and high expression of USP20 from the tissue sections and performed multiplex immunofluorescence. Tissue section (4 μm) were deparaffinized in xylene and rehydrated through graded ethanol. Antigen retrieval was performed using citrate buffer (pH 6.0) or EDTA buffer (pH 9.0) according to the antibody requirements. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide, followed by incubation with blocking buffer for 30 min at room temperature. Multiplex immunofluorescence staining was performed using a tyramide signal amplification (TSA)-based multiplex staining kit according to the manufacturer’s instructions. Briefly, tissue sections were sequentially incubated overnight at 4 °C with the following primary antibodies: anti-CD4 (1:100), anti-CD8a (1:100), anti-CD279 (PD-1) (1:100), and anti-FOXP3 (1:100). After washing, sections were incubated with HRP-conjugated secondary antibodies, followed by fluorophore-conjugated TSA reagents. Heat-mediated stripping was performed between each staining cycle before incubation with the next primary antibody. After completion of all staining procedures, nuclei were counterstained with DAPI and slides were mounted with antifade mounting medium. Multispectral images were acquired using a fluorescence imaging system and analyzed using image analysis software. Statistical analysis Quantitative data are presented as mean ± standard deviation (SD). Differences between two groups were analyzed using Student’s t-test. Cox proportional hazards regression analysis and the log-rank test were applied to evaluate hazard ratios (HRs) and corresponding p values for survival analyses. A p value < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism version 9.0 and R software (version 4.3.0). Results USP20 is upregulated in multiple cancer types and associated with prognosis We first evaluated the expression pattern of USP20 in normal and tumor tissues using data from TCGA, GTEx, and GEO databases. In the GTEx dataset, USP20 expression was detected across 31 different normal tissue types. Pan-cancer analysis revealed that USP20 was significantly upregulated in eight cancer types, including LUAD, breast invasive carcinoma (BRCA), liver hepatocellular carcinoma (LIHC), colon adenocarcinoma (COAD), stomach adenocarcinoma (STAD), esophageal carcinoma (ESCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), and cholangiocarcinoma (CHOL). In contrast, USP20 expression was lower in tumor tissues compared with corresponding normal tissues in several cancers, such as kidney renal clear cell carcinoma (KIRC), thyroid carcinoma (THCA), bladder urothelial carcinoma (BLCA), and uterine corpus endometrial carcinoma (UCEC). For certain tumor types, including mesothelioma (MESO), sarcoma (SARC), and ovarian cancer (OV), the lack of matched normal tissues precluded definitive conclusions (Fig. 1A). Overall, these findings indicate that USP20 expression is upregulated in the majority of tumor types relative to normal tissues. Fig. 1USP20 is upregulated in multiple cancer types and associated with prognosis. (A) Pan-cancer expression distribution of USP20 in TCGA and GTEx datasets. Red indicates tumor tissues and blue indicates normal tissues (P < 0.05, P < 0.01, P < 0.001). (B) Boxplot showing USP20 expression differences between LUAD tumor tissues and normal lung tissues in the TCGA dataset. (C) Boxplots illustrating USP20 expression in LUAD tumors and normal tissues across multiple GEO datasets (e.g., GSE19804, GSE31210). (D-F) Kaplan-Meier survival curves depicting the association between USP20 expression and patient survival in different cancer types. (G) Determination of the optimal cutoff value for USP20 expression using maximally selected rank statistics; Kaplan-Meier overall survival (OS) and disease-specific survival (DSS) curves stratified by USP20 expression level USP20 is upregulated in multiple cancer types and associated with prognosis. (A) Pan-cancer expression distribution of USP20 in TCGA and GTEx datasets. Red indicates tumor tissues and blue indicates normal tissues (P < 0.05, P < 0.01, P < 0.001). (B) Boxplot showing USP20 expression differences between LUAD tumor tissues and normal lung tissues in the TCGA dataset. (C) Boxplots illustrating USP20 expression in LUAD tumors and normal tissues across multiple GEO datasets (e.g., GSE19804, GSE31210). (D-F) Kaplan-Meier survival curves depicting the association between USP20 expression and patient survival in different cancer types. (G) Determination of the optimal cutoff value for USP20 expression using maximally selected rank statistics; Kaplan-Meier overall survival (OS) and disease-specific survival (DSS) curves stratified by USP20 expression level Focusing on LUAD, analysis of the TCGA dataset demonstrated that USP20 expression was significantly higher in LUAD tumor tissues compared with normal lung tissues (Fig. 1B). Consistent results were obtained from independent GEO cohorts, including GSE19804 and GSE31210, in which USP20 was also markedly upregulated in LUAD samples relative to normal controls (Fig. 1C). To further investigate the prognostic relevance of USP20 across cancers, Kaplan-Meier survival analyses were performed. High USP20 expression was associated with poorer OS in lung squamous cell carcinoma (LUSC), BRCA, and THCA, suggesting that USP20 may serve as an unfavorable prognostic factor in these tumor types (Fig. 1D-F). Using the TCGA-LUAD cohort, the optimal cutoff value for USP20 expression was determined by maximally selected rank statistics (cutoff = 9.43), and patients were accordingly stratified into high- and low-expression groups (Fig. 1G). Kaplan-Meier survival analysis demonstrated that patients with high USP20 expression had significantly shorter OS compared with those in the low-expression group (p < 0.001, Fig. 1G). Similarly, high USP20 expression was associated with unfavorable DSS (p < 0.01, Fig. 1G), indicating a strong correlation between USP20 overexpression and poor clinical outcomes in LUAD. Mutation and copy number alteration characteristics of USP20 Comprehensive analysis of genetic alterations using cBioPortal (TCGA Pan-Cancer Atlas) demonstrated that USP20 alterations were present across multiple cancer types but occurred at a relatively low overall frequency. Among 32 cancer types, the total alteration frequency of USP20 was below 6%, with higher frequencies observed in UCEC, melanoma, and LUAD (Fig. 2A). Missense mutations constituted the predominant alteration type, followed by gene amplification and deep deletion events. Fig. 2Mutation and copy number alteration characteristics of USP20. (A) Pan-cancer distribution of USP20 genetic alterations in TCGA, including mutations, structural variants, amplifications, and deletions. (B) Waterfall plot showing somatic mutation profiles of USP20 in LUAD samples. (C) Schematic representation of USP20 protein domains. (D) Lollipop plot illustrating the distribution and frequency of USP20 mutation sites along the amino acid sequence. (E) Scatter plot showing USP20 mRNA expression across different copy number alteration states Mutation and copy number alteration characteristics of USP20. (A) Pan-cancer distribution of USP20 genetic alterations in TCGA, including mutations, structural variants, amplifications, and deletions. (B) Waterfall plot showing somatic mutation profiles of USP20 in LUAD samples. (C) Schematic representation of USP20 protein domains. (D) Lollipop plot illustrating the distribution and frequency of USP20 mutation sites along the amino acid sequence. (E) Scatter plot showing USP20 mRNA expression across different copy number alteration states Within the LUAD cohort, 532 cases (86.36%) harbored at least one genomic alteration, while the somatic mutation frequency of USP20 was approximately 1.3%. Missense mutations were the most common variant, with occasional nonsense and splice-site mutations also detected (Fig. 2B). Lollipop plot analysis revealed that USP20 mutation sites were distributed throughout the protein sequence, with enrichment near the zinc finger ubiquitin-specific protease (zf-UBP) domain and the catalytic UCH/Peptidase_C19 domain (Fig. 2C, D). These mutations may potentially affect the deubiquitinating activity of USP20 and thereby alter its regulatory role in protein stability and tumor progression. Given the prominent upregulation of USP20 in LUAD, we further assessed the relationship between copy number alterations (CNAs) and transcriptomic expression. Based on GISTIC analysis, samples were classified into deep deletion, shallow deletion, diploid, copy number gain, and amplification groups. USP20 mRNA expression exhibited a clear copy number-dependent pattern: expression levels were significantly increased in gain and amplification groups compared with diploid samples, whereas shallow and deep deletion groups showed markedly reduced expression (Fig. 2E). Collectively, these findings indicate that the aberrant overexpression of USP20 in LUAD is closely associated with copy number amplification rather than high-frequency point mutations, suggesting that CNAs represent a major genetic mechanism driving USP20 dysregulation. DEGs and pathway enrichment associated with USP20 expression in LUAD To explore the potential biological functions of USP20 in LUAD, differential expression analysis was performed based on USP20 high- and low-expression groups in the TCGA-LUAD cohort. The volcano plot revealed a large number of DEGs, with upregulated genes shown in red and downregulated genes shown in blue, indicating that USP20 expression is closely associated with the activation of multiple signaling pathways (Fig. 3A). Fig. 3Differentially expressed genes and pathway enrichment associated with USP20 expression in LUAD. (A) Volcano plot showing differentially expressed genes (DEGs) between the USP20 high- and low-expression groups. Red dots indicate upregulated genes, blue dots indicate downregulated genes, and black dots indicate genes without significant differences. (B-C) GO enrichment analysis showing that DEGs are mainly enriched in biological processes related to extracellular structure organization, cell adhesion, immune regulation, and inflammatory responses. (D-E) KEGG pathway enrichment analysis indicating that these DEGs are primarily involved in cytokine-cytokine receptor interaction, cancer-related pathways, adherens junction, and immune-related signaling pathways. (F-G) GSEA showing that the USP20 high-expression group is enriched in multiple tumor-related and immune regulatory pathways Differentially expressed genes and pathway enrichment associated with USP20 expression in LUAD. (A) Volcano plot showing differentially expressed genes (DEGs) between the USP20 high- and low-expression groups. Red dots indicate upregulated genes, blue dots indicate downregulated genes, and black dots indicate genes without significant differences. (B-C) GO enrichment analysis showing that DEGs are mainly enriched in biological processes related to extracellular structure organization, cell adhesion, immune regulation, and inflammatory responses. (D-E) KEGG pathway enrichment analysis indicating that these DEGs are primarily involved in cytokine-cytokine receptor interaction, cancer-related pathways, adherens junction, and immune-related signaling pathways. (F-G) GSEA showing that the USP20 high-expression group is enriched in multiple tumor-related and immune regulatory pathways GO enrichment analysis demonstrated that upregulated genes in the USP20 high-expression group were mainly enriched in immune-related biological processes, including regulation of immune response, extracellular structure organization, and cell adhesion (Fig. 3B-C). In contrast, downregulated genes were primarily involved in cell cycle processes and metabolism-related pathways. These findings suggest that USP20 may regulate LUAD progression by modulating the TME and intercellular interactions. Further KEGG pathway enrichment analysis revealed that DEGs were significantly enriched in several classical tumor- and immune-related pathways, including PD-L1 expression and PD-1 checkpoint pathway, TNF signaling pathway, pathways in cancer, leukocyte transendothelial migration, and the Hippo signaling pathway (Fig. 3D-E). Consistently, GSEA further validated these observations, showing that USP20 high-expression samples exhibited significant positive enrichment (NES > 1.5, FDR < 0.05) in multiple immune-related and signaling pathways, particularly the PD-1/PD-L1 immune checkpoint pathway (p = 0.008) and the TNF signaling pathway (p = 0.033) (Fig. 3F-G). Collectively, these differential expression and enrichment analyses indicate that high USP20 expression is closely associated with immune regulation, cell adhesion, and signal transduction, and may promote LUAD progression through activation of immune checkpoint and inflammatory pathways. Association between USP20 expression and the TME in LUAD To investigate the relationship between USP20 and the TME, immune infiltration patterns in LUAD samples were systematically evaluated using the CIBERSORT and ESTIMATE algorithms. Overall immune cell composition analysis showed that M2 macrophages, resting CD4 memory T cells, and M0 macrophages were the dominant infiltrating immune cell types in LUAD tumors, suggesting a predominantly immunosuppressive TME (Fig. 4A-B). Fig. 4Association between USP20 expression and the tumor immune microenvironment in LUAD. (A) Relative abundance of 22 immune cell subsets estimated by CIBERSORT across individual samples. Each dot represents the abundance of a given immune cell type in a single sample; different colors denote different immune cell subsets. (B) Heatmap showing normalized abundance of immune cell subsets across samples. (C) Comparison of StromalScore, ImmuneScore, ESTIMATEScore, and tumor purity (TumorPurity) between the high (green) and low (red) USP20 expression groups. Error bars represent interquartile ranges; “ns” indicates no statistical significance (P > 0.05). (D) Distribution of ImmuneScore in the high (red) and low (blue) USP20 expression groups. (E) Relative proportions of the 22 CIBERSORT-estimated immune cell subsets in the high (blue) and low (red) USP20 expression groups. (F) Pearson correlation heatmap of the 22 immune cell subsets. Both axes represent immune cell subsets. Positive values (0.2–0.8) indicate positive correlations, whereas negative values (− 0.2 to − 0.8) indicate negative correlations. (G) Distribution of correlation coefficients between USP20 expression and immune checkpoint molecules (e.g., CD27, PDCD1) Association between USP20 expression and the tumor immune microenvironment in LUAD. (A) Relative abundance of 22 immune cell subsets estimated by CIBERSORT across individual samples. Each dot represents the abundance of a given immune cell type in a single sample; different colors denote different immune cell subsets. (B) Heatmap showing normalized abundance of immune cell subsets across samples. (C) Comparison of StromalScore, ImmuneScore, ESTIMATEScore, and tumor purity (TumorPurity) between the high (green) and low (red) USP20 expression groups. Error bars represent interquartile ranges; “ns” indicates no statistical significance (P > 0.05). (D) Distribution of ImmuneScore in the high (red) and low (blue) USP20 expression groups. (E) Relative proportions of the 22 CIBERSORT-estimated immune cell subsets in the high (blue) and low (red) USP20 expression groups. (F) Pearson correlation heatmap of the 22 immune cell subsets. Both axes represent immune cell subsets. Positive values (0.2–0.8) indicate positive correlations, whereas negative values (− 0.2 to − 0.8) indicate negative correlations. (G) Distribution of correlation coefficients between USP20 expression and immune checkpoint molecules (e.g., CD27, PDCD1) ESTIMATE analysis revealed that the USP20 high-expression group exhibited significantly higher ImmuneScore, whereas StromalScore and overall ESTIMATEScore showed no significant differences, indicating that high USP20 expression is closely associated with increased immune cell infiltration (Fig. 4C-D). Comparative analysis of immune cell proportions between USP20 high- and low-expression groups demonstrated that the high-expression group had increased infiltration of CD8⁺ T cells, regulatory T cells, and M1 macrophages, accompanied by decreased proportions of M2 macrophages (Fig. 4E). These findings suggest that USP20 may participate in regulating immune cell polarization and antitumor immune activity. Correlation analysis further revealed that USP20 expression was positively correlated with the infiltration of CD8⁺ T cells and M1 macrophages, while negatively correlated with M2 macrophages and resting CD4 T cells, supporting a positive association between USP20 and immune activation status (Fig. 4F). Finally, immune checkpoint analysis showed that PD-1(PDCD1), PD-L1(CD274), LAG3, and other immune checkpoint molecules were significantly upregulated in USP20 high-expression samples (Fig. 4G), indicating that USP20 may promote immune activation accompanied by T-cell exhaustion, thereby shaping an “immune-hot but exhausted” TME phenotype. Overall, these results demonstrate that high USP20 expression is closely associated with enhanced immune cell infiltration and upregulation of immune checkpoint pathways in LUAD, suggesting a role for USP20 in remodeling the TME through regulation of T-cell activation and exhaustion. Molecular subtyping based on USP20-related genes and validation of immune features USP20 as an adverse prognostic factor suggested its tumor-promoting roles in LUAD. Next, the function module of LinkedOmics was used to analyze RNA-seq data from 517 LUAD patients in TCGA Firehose Legacy.The volcano plot showed that genes were significantly correlated with USP20 in LUAD (Supplementary Fig. 1A, Supplementary Table 1). The 50 significant gene sets were positively and negatively correlated with USP20 as shown in the heat map (Supplementary Fig. 1B). This result suggests a widespread impact of USP20 on the transcriptome. Consistently, the LinkInterpreter module showed significant enrichment of NF-κB, proteasome and immune associated pathways, suggesting a critical role for USP20 co-expressed genes in mediating these biological processes (Supplementary Fig. 1C). To systematically investigate the molecular heterogeneity and immunological significance of USP20 in LUAD, consensus clustering analysis was performed based on USP20 co-expressed genes. The optimal number of clusters was determined by consensus clustering, with CDF curves becoming stable as the number of clusters increased. When k = 4, the delta area showed minimal growth, indicating that four clusters represented the most stable classification (Fig. 5A-B). Fig. 5Molecular subtyping based on USP20-related genes and validation of immune features. (A) Consensus cumulative distribution function (CDF) curves. (B) Delta area plot. (C) Consensus matrix heatmaps across different values of k. (D) Cluster tracking plot. (E) Consensus clustering dendrogram Molecular subtyping based on USP20-related genes and validation of immune features. (A) Consensus cumulative distribution function (CDF) curves. (B) Delta area plot. (C) Consensus matrix heatmaps across different values of k. (D) Cluster tracking plot. (E) Consensus clustering dendrogram Consensus matrices, clustering tracking plots, and dendrograms further confirmed the robustness and internal consistency of the four molecular subtypes (C1–C4) (Fig. 5C-E). Heatmap visualization revealed distinct transcriptomic expression patterns among the four subtypes, with clear stratification between C1–C4 (Fig. 6A-B). PCA further validated subtype stability, as samples from different subtypes exhibited relatively independent distributions along principal component 1, with significant differences in PC1 scores among the four groups (Fig. 6C). Differential expression heatmaps further demonstrated marked differences in gene expression patterns between C1/C2 and C3/C4 subtypes, supporting the biological rationality of this classification (Fig. 6D). Fig. 6Validation of USP20-related molecular subtypes and immune characteristics. (A) Item-consensus likelihood (ICL) plot for determining the optimal number of clusters. (B) Consensus heatmap of clustering results. (C) Principal component analysis (PCA) plot. (D) Heatmap of differentially expressed genes (DEGs) across the subtypes. (E) Boxplots showing differences in TME cell composition and immune checkpoint expression among the subtypes Validation of USP20-related molecular subtypes and immune characteristics. (A) Item-consensus likelihood (ICL) plot for determining the optimal number of clusters. (B) Consensus heatmap of clustering results. (C) Principal component analysis (PCA) plot. (D) Heatmap of differentially expressed genes (DEGs) across the subtypes. (E) Boxplots showing differences in TME cell composition and immune checkpoint expression among the subtypes Analysis of immune microenvironment characteristics revealed that C1 and C2 subtypes were enriched with M2 macrophages, representing an immune-cold phenotype. In contrast, C3 and C4 subtypes exhibited increased infiltration of CD8⁺ T cells, M1 macrophages, and regulatory T cells (Tregs), indicating an immune-hot but exhausted phenotype (Fig. 6E). Consistently, immune checkpoint gene expression analysis showed significant upregulation of PDCD1 (PD-1), CD274 (PD-L1), CTLA4, LAG3, and TIGIT in the C3/C4 subtypes, suggesting more active immune checkpoint signaling and potential sensitivity to immunotherapy. Taken together, consensus clustering based on USP20-related genes identified four stable LUAD subtypes with distinct molecular and immunological features. The immune-hot but exhausted phenotype observed in C3/C4 subtypes suggests that USP20 may influence antitumor immune responses and prognosis by modulating the TME. Single-cell transcriptomic analysis reveals predominant USP20 expression in immune cell populations To further delineate the cellular sources and expression characteristics of USP20 within the LUAD tumor microenvironment, single-cell transcriptomic analyses were performed. UMAP visualization of the GSE146100 dataset demonstrated that tumor tissues consisted of immune cells, stromal cells, and epithelial/malignant cells, with immune cell clusters exhibiting marked heterogeneity and widespread distribution (Fig. 7A). Fig. 7Single-cell transcriptomic analysis of USP20 expression in LUAD/NSCLC. (A) UMAP plot showing cell-type distribution in the LUAD GSE146100 dataset. (B) Feature plot showing USP20 expression distribution across cell populations. (C) USP20 expression levels across multiple GEO datasets (e.g., GSE127465, GSE146100) and major cell categories (immune cells, malignant cells, other cells, and stromal cells). (D) USP20 expression levels across multiple GEO datasets (e.g., GSE127465, GSE146100) and detailed cell subpopulations (e.g., CD4 T cells, Tregs, macrophages) Single-cell transcriptomic analysis of USP20 expression in LUAD/NSCLC. (A) UMAP plot showing cell-type distribution in the LUAD GSE146100 dataset. (B) Feature plot showing USP20 expression distribution across cell populations. (C) USP20 expression levels across multiple GEO datasets (e.g., GSE127465, GSE146100) and major cell categories (immune cells, malignant cells, other cells, and stromal cells). (D) USP20 expression levels across multiple GEO datasets (e.g., GSE127465, GSE146100) and detailed cell subpopulations (e.g., CD4 T cells, Tregs, macrophages) Overlaying USP20 expression onto the UMAP space revealed that USP20 signals were predominantly localized within immune cell clusters, whereas expression in stromal cells and malignant epithelial cells was minimal (Fig. 7B). Consistently, integrated analyses across multiple datasets in the TISCH database showed that USP20 expression was significantly higher in immune cells compared with stromal and malignant cells across several independent LUAD/NSCLC cohorts (Fig. 7C). At finer cellular resolution, USP20 expression was most prominent in CD8⁺ T cells (including exhausted CD8Tex cells), CD4⁺ T cells, NK cells, and monocytes/macrophages, while remaining low in fibroblasts, endothelial cells, and malignant epithelial cells (Fig. 7D). Cross-dataset validation across seven independent NSCLC single-cell datasets further confirmed that USP20 expression was consistently enriched in immune cells, particularly exhausted CD8⁺ T cells and macrophages. Collectively, single-cell transcriptomic analyses demonstrate that USP20 is primarily expressed in immune cell populations rather than tumor epithelial cells, suggesting that USP20 may participate in LUAD immune microenvironment remodeling by regulating immune cell activity and immune checkpoint signaling. USP20 expression promotes malignant phenotypes and immunosuppressive T-cell infiltration in LUAD To characterize USP20 expression and its biological function in LUAD, immunohistochemical analysis was performed on 75 tumor tissues and 70 adjacent normal tissues. USP20 exhibited markedly stronger staining intensity in tumor tissues compared with adjacent normal tissues, with expression levels classified as ±, +, ++, and +++ (Fig. 8A). Quantitative IHC scoring confirmed significantly higher USP20 expression in LUAD tumors (Fig. 8B). Kaplan-Meier survival analysis further demonstrated that patients with high USP20 expression had significantly shorter OS than those with low expression (Fig. 8C). Fig. 8USP20 expression in LUAD tissues and its effects on tumor cell malignant behaviors. (A) Immunohistochemistry (IHC) showing that USP20 expression is significantly higher in LUAD tumor tissues than in adjacent normal tissues. Staining intensity is categorized as ±, +, ++, and +++. (B) Quantitative analysis of USP20 IHC scores in tumor tissues and adjacent normal tissues. (C) Kaplan-Meier survival analysis showing that patients with high USP20 expression have significantly lower overall survival than those with low expression. (D-E) Colony formation assays showing that USP20 knockdown (shUSP20#1, shUSP20#2) markedly suppresses colony-forming ability in A549 and H3122 cells, with quantitative analysis. (F-G) Transwell migration assays showing that USP20 knockdown significantly inhibits migration of A549 and H3122 cells, with quantitative analysis. (H) Western blot analysis showing that USP20 knockdown upregulates apoptosis-related proteins c-PARP and c-CASP3, while total PARP and CASP3 are decreased in A549 and H3122 cells. (I) Representative images of tumors from different treatment groups in the xenograft model, showing reduced tumor size in USP20 knockdown groups. (J) Statistical analysis of tumor weights showing that USP20 knockdown significantly suppresses tumor growth. (K) Tumor growth curves showing reduced tumor growth rate in USP20 knockdown groups. Data are presented as mean ± SD; *P < 0.05, **P < 0.01, ***P < 0.001 USP20 expression in LUAD tissues and its effects on tumor cell malignant behaviors. (A) Immunohistochemistry (IHC) showing that USP20 expression is significantly higher in LUAD tumor tissues than in adjacent normal tissues. Staining intensity is categorized as ±, +, ++, and +++. (B) Quantitative analysis of USP20 IHC scores in tumor tissues and adjacent normal tissues. (C) Kaplan-Meier survival analysis showing that patients with high USP20 expression have significantly lower overall survival than those with low expression. (D-E) Colony formation assays showing that USP20 knockdown (shUSP20#1, shUSP20#2) markedly suppresses colony-forming ability in A549 and H3122 cells, with quantitative analysis. (F-G) Transwell migration assays showing that USP20 knockdown significantly inhibits migration of A549 and H3122 cells, with quantitative analysis. (H) Western blot analysis showing that USP20 knockdown upregulates apoptosis-related proteins c-PARP and c-CASP3, while total PARP and CASP3 are decreased in A549 and H3122 cells. (I) Representative images of tumors from different treatment groups in the xenograft model, showing reduced tumor size in USP20 knockdown groups. (J) Statistical analysis of tumor weights showing that USP20 knockdown significantly suppresses tumor growth. (K) Tumor growth curves showing reduced tumor growth rate in USP20 knockdown groups. Data are presented as mean ± SD; *P < 0.05, **P < 0.01, ***P < 0.001 In vitro functional assays showed that stable knockdown of USP20 using shRNA in A549 and H3122 cell lines significantly suppressed colony formation ability (Fig. 8D-E). Transwell migration assays further revealed that USP20 knockdown markedly reduced migratory capacity in both cell lines, as confirmed by quantitative analysis (Fig. 8F-G). Western blot analysis demonstrated that USP20 knockdown led to increased levels of apoptosis-related proteins c-PARP and c-CASP3, accompanied by decreased total PARP and CASP3 expression, indicating enhanced apoptosis in LUAD cells following USP20 depletion (Fig. 8H). In vivo xenograft experiments showed that tumors derived from USP20-knockdown cells exhibited significantly reduced tumor volume, lower tumor weight, and slower growth rates compared with control tumors (Fig. 8I-K). To further evaluate the relationship between USP20 expression and T-cell infiltration, multiplex immunofluorescence analysis revealed that USP20 high-expression tumors exhibited enhanced PD-1 and FOXP3 signals with reduced CD4 and CD8 staining intensity compared with USP20 low-expression tumors (Fig. 9A). Flow cytometric analysis of peripheral blood samples from LUAD patients further demonstrated significantly higher proportions of PD-1⁺CD8⁺ T cells and FOXP3⁺CD4⁺ Tregs in the USP20 high-expression group (Fig. 9B-C). Fig. 9USP20 is associated with immunosuppressive T-cell enrichment and regulates T-cell infiltration in vivo. (A) Multiplex immunofluorescence (mIF) showing immune cell distribution in LUAD tissues with low versus high USP20 expression. (B-C) Flow cytometry analysis of PD-1⁺CD8⁺ T-cell and FOXP3⁺CD4⁺ regulatory T-cell (Treg) proportions in peripheral blood from LUAD patients with low versus high USP20 expression, with quantitative statistics. (D) IHC staining in murine subcutaneous tumors after USP20 knockdown, showing expression of USP20, CD8, PD-1, CD4, and FOXP3. (E-F) Flow cytometry analysis of PD-1⁺CD8⁺ T-cell and FOXP3⁺CD4⁺ Treg proportions in tumor tissues after USP20 knockdown, with quantitative statistics. Data are presented as mean ± SD; *P < 0.05, **P < 0.01, ***P < 0.001 USP20 is associated with immunosuppressive T-cell enrichment and regulates T-cell infiltration in vivo. (A) Multiplex immunofluorescence (mIF) showing immune cell distribution in LUAD tissues with low versus high USP20 expression. (B-C) Flow cytometry analysis of PD-1⁺CD8⁺ T-cell and FOXP3⁺CD4⁺ regulatory T-cell (Treg) proportions in peripheral blood from LUAD patients with low versus high USP20 expression, with quantitative statistics. (D) IHC staining in murine subcutaneous tumors after USP20 knockdown, showing expression of USP20, CD8, PD-1, CD4, and FOXP3. (E-F) Flow cytometry analysis of PD-1⁺CD8⁺ T-cell and FOXP3⁺CD4⁺ Treg proportions in tumor tissues after USP20 knockdown, with quantitative statistics. Data are presented as mean ± SD; *P < 0.05, **P < 0.01, ***P < 0.001 Functional validation using USP20 knockdown in murine tumor models showed reduced PD-1 and FOXP3 staining intensity with increased CD4 and CD8 expression in tumor tissues (Fig. 9D). Flow cytometry further confirmed that USP20 depletion significantly decreased the proportions of PD-1⁺CD8⁺ T cells and FOXP3⁺CD4⁺ Tregs within tumor tissues (Fig. 9E-F). Taken together, these in vitro and in vivo experiments demonstrate that aberrant USP20 overexpression in LUAD promotes tumor cell proliferation, migration, and resistance to apoptosis, and is closely associated with enrichment of immunosuppressive T-cell populations. Conversely, USP20 knockdown suppresses tumor growth and alleviates immunosuppressive features within the tumor microenvironment, highlighting a critical role for USP20 in LUAD progression and immune regulation. Discussion In this study, we integrated multi-omics bioinformatic analyses with in vitro and in vivo functional experiments to characterize the expression profile, genetic basis, biological functions, and potential immunoregulatory roles of the deubiquitinating enzyme USP20 in LUAD. The results showed that USP20 was consistently upregulated in LUAD and was significantly associated with unfavorable clinical outcomes. Notably, USP20 upregulation appeared to be mainly driven by copy number amplification rather than recurrent somatic mutations [21]. Pan-cancer and LUAD-specific mutation analyses showed that USP20 had a relatively low somatic mutation frequency, with sporadic missense mutations as the predominant alteration and no apparent functional mutation hotspots. These findings suggest that USP20 is unlikely to act as a classical mutation-driven oncogene. Instead, its oncogenic relevance may be more closely related to aberrant expression dosage or altered protein stability. This pattern is consistent with previous observations for several DUBs, which often contribute to tumorigenesis through dysregulated expression or post-translational regulation rather than high-frequency mutational events [22–25]. Copy number variation analysis further supported this interpretation. USP20 mRNA expression showed a clear dose-dependent relationship with copy number status: samples with copy number gain or amplification displayed significantly higher USP20 expression, whereas those with shallow or deep deletions exhibited markedly reduced expression. These results indicate that copy number amplification may represent an important genetic mechanism underlying USP20 overexpression in LUAD. Copy number–driven oncogene activation is a common molecular event in LUAD and other solid tumors, where it can promote tumor progression by increasing transcriptional output or reshaping regulatory chromatin architecture [26–28]. The clinical relevance of USP20 was further reflected in survival analyses. Based on TCGA and multiple independent GEO cohorts, high USP20 expression was significantly associated with shorter OS and DSS in patients with LUAD. These findings suggest that USP20 may not merely reflect tumor-associated molecular alterations, but may also be linked to tumor aggressiveness and disease progression. This is in line with previous studies showing that DUB dysregulation has important prognostic implications in cancer [6]. Functional experiments provided additional evidence for the tumor-promoting role of USP20 in LUAD. In vitro, USP20 knockdown significantly suppressed colony formation and migratory capacity of LUAD cells and promoted apoptosis, as indicated by increased levels of c-PARP and c-CASP3. In vivo, USP20 depletion markedly inhibited xenograft tumor growth. Collectively, these results suggest that USP20 contributes to LUAD progression by sustaining tumor cell survival, maintaining malignant phenotypes, and suppressing apoptotic signaling. These findings are also consistent with prior studies showing that DUBs can promote tumor growth by stabilizing key oncogenic proteins or interfering with apoptotic pathways [29–30]. Importantly, our study further uncovered a potential link between USP20 expression and the TME. The preliminary analysis on the correlation between USP20 expression and the tumor immune microenvironment (TIME) in lung adenocarcinoma showed that the group with high USP20 expression exhibited immunological active phenotypic characteristics, characterized by increased infiltration of CD8+T cells and Treg cells, and a higher overall immune infiltration level. However, further single-cell sequencing analysis revealed that USP20 was mainly highly expressed in the CD8+T cell population, especially enriched in the CD8+ exhausted T cell (CD8 Tex) subgroup. Subsequent functional experiments further confirmed that after knockdown of USP20, the expression of inhibitory receptor PD-1 in CD8+ exhausted T cells was significantly reduced, and the expression of the key transcription factor FOXP3 in Treg cells was also significantly downregulated. These results suggest that although high USP20 expression is associated with immune activation, it may induce the formation of an immune exhaustion-type tumor immune microenvironment by promoting CD8+T cell exhaustion and enhancing the immunosuppressive effect mediated by Treg cells, ultimately leading to impaired anti-tumor immune function [31]. Extensive evidence has established that CD8⁺T-cell exhaustion and Treg accumulation are key mechanisms underlying tumor immune evasion and resistance to immunotherapy [32–34]. These observations were further supported by experimental validation using multiplex immunofluorescence, flow cytometry, and immunohistochemistry. USP20 knockdown significantly reduced the proportions of PD-1⁺ CD8⁺ T cells and FOXP3⁺ Tregs, while enhancing CD8⁺ T-cell infiltration [21]. These results suggest that USP20 may facilitate immune evasion by modulating immune checkpoint–related pathways and T-cell functional states. Consistently, pathway enrichment and GSEA analyses demonstrated that USP20-associated genes were significantly enriched in immune- and tumor-related pathways, including the PD-1/PD-L1 signaling pathway, TNF signaling pathway, leukocyte transendothelial migration, and the Hippo signaling pathway. These pathways are known to play central roles in regulating antitumor immune responses, inflammatory signaling, and tumor cell fate decisions [35–37]. It is worth noting that the single-cell transcriptome analysis of this study revealed that USP20 is mainly expressed in immune cell populations, particularly enriched in exhausted CD8+T cells, macrophages, and regulatory T cells, while its expression is relatively low in malignant epithelial cells. However, further in vitro functional experiments confirmed that knocking down USP20 in lung adenocarcinoma cells still significantly inhibits cell proliferation and migration and promotes apoptosis, suggesting that USP20 not only participates in the regulation of the immune microenvironment but also has an intrinsic oncogenic effect in tumor cells. Therefore, we speculate that USP20 may jointly promote the progression of LUAD through a “tumor cell-immune cell” dual mechanism: on the one hand, USP20 maintains cell survival and malignant phenotype in tumor cells; on the other hand, its high expression in immune cells may be involved in T cell exhaustion and the formation of an immunosuppressive microenvironment. Several limitations should be acknowledged. The analyses were mainly based on bulk transcriptomic datasets, which limited the ability to fully characterize intratumoral heterogeneity and subtype-specific cellular distribution of USP20 in LUAD. Although single-cell analysis was included to provide additional insight, further studies with larger single-cell cohorts and spatial transcriptomic validation are needed to clarify the cell-type-specific functions of USP20 within the LUAD ecosystem [38]. In addition, different cohort-specific cutoff values were used to define high- and low-expression groups, which may affect the reproducibility and clinical generalizability of the prognostic findings. Future studies using predefined fixed cutoffs or continuous-variable modeling may help improve model robustness [39]. As a single-gene biomarker, USP20-based prognostic analysis may also be influenced by platform variability and batch effects; gene-pair expression ranking approaches may provide a more stable strategy for cross-cohort prognostic prediction [40]. Moreover, although USP20 expression was significantly associated with prognosis and immune characteristics, the present study mainly demonstrates correlation rather than causality. Future investigations integrating Mendelian randomization, multi-omics approaches, and mechanistic validation may help establish a clearer causal relationship between USP20 dysregulation and LUAD progression [41]. In summary, this study proposes a comprehensive model in which copy number amplification-driven USP20 overexpression contributes to LUAD progression by promoting tumor cell proliferation, survival, and migration, while also shaping an immunosuppressive tumor microenvironment characterized by enhanced T-cell exhaustion and Treg enrichment. These findings expand the current understanding of the biological functions of USP20 and highlight its potential value as a prognostic biomarker and candidate therapeutic target, particularly in the context of combinatorial immunotherapeutic strategies for LUAD. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Supplementary Fig. 1. (A) A volcano plot shows genes (red dots) that positively correlate with USP20 and genes (green dots) that negatively correlate with USP20 in LUAD. (B) Heat maps show genes positively and negatively correlated with USP20 mRNA expression in LUAD (top 50). (C) Gene set enrichment analysis (GSEA) was performed though the LinkInterpreter module. Supplementary Material 1: Supplementary Fig. 1. (A) A volcano plot shows genes (red dots) that positively correlate with USP20 and genes (green dots) that negatively correlate with USP20 in LUAD. (B) Heat maps show genes positively and negatively correlated with USP20 mRNA expression in LUAD (top 50). (C) Gene set enrichment analysis (GSEA) was performed though the LinkInterpreter module. Supplementary Material 2 Supplementary Material 2 Supplementary Material 3 Supplementary Material 3