LncRNA NRAV is associated with unfavorable prognosis and immune-related transcriptional features in hepatocellular carcinoma Authors: Xingbao Fang, Yan Cai, Zhaojun Li, Xiongbing Peng, Shaohua Yang, Zhuoyu Zhao, Yuehong Li Published: 2026 ABSTRACT AbstractBackgroundHepatocellular carcinoma (HCC) remains a lethal malignancy, and robust biomarkers are needed to refine risk stratification and support clinical decision-making. Long noncoding RNAs can shape tumor biology and immune states. We investigated the clinical relevance of lncRNA NRAV in HCC and its association with immune-related features.MethodsNRAV expression and clinical associations were assessed using TCGA-LIHC, paired tumor–adjacent tissue comparisons, GTEx normal liver controls, and independent GEO cohorts. Prognostic associations were evaluated by Kaplan–Meier and Cox regression analyses. Immune features were profiled using ssGSEA and correlation analyses with immune checkpoint molecules. Functional effects of NRAV were examined in HCC cells using loss-of-function assays.ResultsNRAV was consistently overexpressed in HCC compared with normal liver tissues across multiple cohorts and paired tumor–normal samples. Higher NRAV expression was associated with adverse clinicopathological features and poorer overall survival. ROC analysis showed that NRAV expression had good discriminatory performance in HCC (AUC = 0.901, 95% CI: 0.865–0.937). Immune analyses showed that NRAV-high tumors were associated with reduced CD8⁺ T-cell-related signatures and increased macrophage- and Th2-associated signatures, together with elevated expression of multiple immune checkpoint molecules (e.g., PDCD1, CD274, CTLA4, LAG3, and HAVCR2). In vitro, NRAV knockdown suppressed colony formation, migration, and invasion in HCC cells.ConclusionNRAV is upregulated in HCC and is associated with unfavorable prognosis, adverse clinicopathological features, and immune-related transcriptional patterns. NRAV may represent a candidate biomarker for risk assessment and warrants further mechanistic and translational investigation.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12672-026-05331-3.Keywords: NRAV, Hepatocellular carcinoma, Prognosis, Biomarker, Immune infiltration, Immune checkpoint Introduction Hepatocellular carcinoma (HCC) remains a major global health burden, ranking among the most prevalent malignancies and leading causes of cancer‑related mortality [1]. Despite notable progress in surgical techniques, systemic therapy, and immunotherapy, the overall survival of patients with HCC remains unsatisfactory [2–4], largely due to late diagnosis and the highly heterogeneous biological behavior of the disease [5]. Therefore, there is a pressing need to identify novel molecular biomarkers that can enhance early detection, refine prognostic stratification, and provide insights into the mechanisms driving hepatocarcinogenesis [6]. Long non‑coding RNAs (lncRNAs), defined as RNA transcripts longer than 200 nucleotides without protein-coding potential [7], have emerged as key regulators of gene expression at multiple levels, including chromatin remodeling, transcriptional control, and post‑transcriptional modification [8]. Increasing evidence suggests that dysregulated lncRNAs contribute to tumor initiation, progression, immune evasion, and therapeutic resistance across multiple cancer types, including HCC [9]. These findings highlight their potential as diagnostic indicators and therapeutic targets. The negative regulator of antiviral response (NRAV) is a recently characterized lncRNA originally implicated in antiviral innate immunity [10]. Although NRAV is downregulated during certain viral infections, emerging studies have reported its aberrant upregulation in several malignancies [11, 12]. Preliminary observations indicate that NRAV may influence tumor biology and immunological processes; however, its functional role, clinical significance, and mechanistic impact in HCC remain poorly defined. To address these gaps, we conducted an integrative analysis combining multi‑cohort transcriptomic data, clinicopathological correlation studies, survival analyses, immune infiltration profiling, and functional experiments. Through this comprehensive approach, we aimed to clarify the expression pattern of NRAV in HCC, determine its prognostic and diagnostic relevance, evaluate its association with immunological features, and experimentally verify its biological effects on tumor cell behavior. This study provides new insights into the potential role of NRAV in liver cancer progression and its association with immune-related transcriptional features in HCC. Materials and methods Data acquisition and preprocessing Transcriptome data and corresponding clinicopathological information for hepatocellular carcinoma were obtained from The Cancer Genome Atlas liver hepatocellular carcinoma cohort (TCGA-LIHC) through the UCSC Xena platform. RNA-seq data from TCGA tumors and GTEx normal liver tissues were downloaded from the UCSC Xena Toil recompute unified dataset, which processes both resources through a consistent computational workflow. Expression values were represented in transcripts per million (TPM) format and transformed as log2(TPM + 1) for downstream analyses and visualization. Because residual batch effects may persist despite unified processing, no additional batch-correction procedure such as ComBat was applied, and comparisons between TCGA tumors and GTEx normal tissues were interpreted cautiously. In this study, paired tumor–adjacent normal comparisons within TCGA and external validation using independent Gene Expression Omnibus (GEO) datasets were considered the primary evidence for differential expression, whereas TCGA–GTEx comparisons were used as supportive analyses. Independent GEO datasets, including GSE113850, GSE112613, and GSE67260, were downloaded from the GEO database and processed according to their corresponding platform annotation files. Survival analysis Patients in the TCGA-LIHC cohort were classified into NRAV-high and NRAV-low groups according to the median NRAV expression level. Overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI) were defined according to the TCGA Pan-Cancer Clinical Data Resource [13] and evaluated using Kaplan–Meier analysis with log-rank tests. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Multivariable Cox regression analyses were performed using the variables retained in the current prognostic model, including NRAV expression and TNM-related clinicopathological parameters. Cases with missing values in any covariate included in the multivariable model were excluded from the corresponding analysis. The proportional hazards assumption was assessed using Schoenfeld residual tests. In addition to median-based dichotomization, NRAV expression was also analyzed as a continuous variable in sensitivity analyses. Clinicopathological correlation Clinicopathological variables (e.g., age, grade, TNM stage, vascular invasion, AFP level, and tumor status) were extracted from TCGA-LIHC. Associations between NRAV expression and categorical variables were assessed using chi-square or Fisher’s exact tests, and continuous variables were compared using Wilcoxon rank-sum tests. Identification of differentially expressed genes (DEGs) Differential expression analysis between NRAV-high and NRAV-low groups was performed using DESeq2 with TCGA-LIHC count data [14]. A relatively stringent threshold (|log2FC| > 1.5) was used to prioritize genes with larger expression changes for downstream functional analyses. Genes with |log2 fold change| > 1.5 and adjusted P < 0.05 were considered DEGs. Volcano plots and heatmaps were used for visualization. Protein–protein interaction (PPI) network construction The STRING database was used to explore functional interactions among the top NRAV-associated DEGs [15]. Interactions with a confidence score > 0.40 (medium confidence) were retained. Network visualization was performed using the iGraph package [16]. Functional enrichment analysis and GSEA Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using the clusterProfiler package [17]. Gene set enrichment analysis (GSEA) was conducted using Molecular Signatures Database (MSigDB) gene sets [18]. Adjusted P < 0.05 was considered statistically significant for over-representation analyses. For GSEA, pathways with a false discovery rate (FDR) < 0.25 and |normalized enrichment score| > 1 were considered significantly enriched [19, 20]. Nomogram construction and ROC analysis A prognostic nomogram incorporating NRAV expression and significant clinicopathological variables was constructed using the rms package [21]. Calibration curves were used to assess the agreement between predicted and observed survival probabilities. Receiver operating characteristic (ROC) curves were generated to evaluate the discriminatory performance of NRAV expression in the present cohort. An AFP-related ROC curve was displayed in the same figure for contextual reference; however, because AFP and NRAV were derived from different specimen types, no formal direct comparison was used for interpretation in the revised manuscript. Immune infiltration assessment Immune cell infiltration in TCGA-LIHC was estimated using single-sample gene set enrichment analysis (ssGSEA) implemented in the GSVA package. Immune-cell signature gene sets were obtained from published immune gene-signature resources. Associations between NRAV expression and immune-cell infiltration scores, as well as immune checkpoint gene expression, were evaluated using Spearman correlation analysis. P values for multiple correlation analyses were adjusted using the Benjamini–Hochberg method where applicable. Because ssGSEA is a signature-based computational approach, these immune associations were interpreted as correlative rather than causal. Cell culture and siRNA transfection Human hepatocellular carcinoma cell lines HepG2 and Huh7 were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The immortalized human hepatocyte-derived cell line HHL-5 was used as a non-tumor control for expression comparison. HepG2, Huh7, and HHL-5 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% penicillin–streptomycin in a humidified incubator at 37 °C with 5% CO2. Two independent small interfering RNAs targeting NRAV (si-NRAV-1 and si-NRAV-2) and a negative control siRNA (si-NC) were synthesized by GenePharma (Shanghai, China). Cells were transfected using Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer’s instructions. Unless otherwise specified, cells were harvested 48 h after transfection for subsequent RT-qPCR, colony formation, migration, invasion, and protein-expression analyses. Knockdown efficiency was verified by RT-qPCR before functional assays. Because NRAV is a long non-coding RNA without protein-coding capacity, knockdown efficiency was evaluated by RT-qPCR, whereas Western blot analysis was used to assess downstream protein changes after NRAV silencing. RNA extraction, RT-qPCR and Western blot analysis Total RNA was extracted using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s instructions. Complementary DNA was synthesized using the PrimeScript RT reagent kit (Takara, Japan). Quantitative real-time PCR (RT-qPCR) was performed using SYBR Green Master Mix (Takara, Japan) on a QuantStudio real-time PCR system (Applied Biosystems, USA). Relative RNA expression levels were calculated using the 2−ΔΔCt method. GAPDH was used as the internal reference for total RNA expression analysis. For subcellular localization analysis, U6 and GAPDH were used as nuclear and cytoplasmic reference genes, respectively. Primer sequences are listed in Supplementary Table 1. Total cellular proteins were extracted using RIPA lysis buffer (Beyotime, China) supplemented with protease inhibitors. Protein concentrations were measured using a BCA Protein Assay Kit (Beyotime, China). Equal amounts of protein were separated by SDS-PAGE and transferred onto polyvinylidene fluoride membranes (Millipore, USA). After blocking, membranes were incubated with primary antibodies overnight at 4 °C, followed by incubation with appropriate horseradish peroxidase-conjugated secondary antibodies. Protein bands were visualized using an enhanced chemiluminescence detection system. GAPDH was used as the loading control. Detailed antibody information is provided in Supplementary Table 2. Colony formation assay Transfected cells were seeded into 6-well plates at low density and cultured for approximately 10–14 days until visible colonies formed. Colonies were fixed with paraformaldehyde, stained with crystal violet, photographed, and counted. Each experiment was performed in at least three independent biological replicates. Transwell migration and invasion assays Cell migration and invasion assays were performed using Transwell chambers with 8-µm pore-size membranes. For invasion assays, the upper chamber was pre-coated with Matrigel; migration assays were performed without Matrigel coating. Transfected cells suspended in serum-free medium were seeded into the upper chambers, whereas medium containing 10% FBS was added to the lower chambers as a chemoattractant. After incubation, non-migrated or non-invaded cells on the upper surface were gently removed, and cells on the lower surface were fixed, stained, and imaged under a light microscope. For each insert, cells were quantified in five randomly selected microscopic fields. Each condition was assessed in at least three independent biological replicates. Quantification was performed using ImageJ. Blinding was not applied unless otherwise stated. Nuclear–cytoplasmic fractionation and FISH To preliminarily assess the subcellular localization of NRAV, nuclear and cytoplasmic RNA fractions were isolated using a nuclear/cytoplasmic separation approach according to the manufacturer’s protocol. NRAV abundance in each fraction was measured by RT-qPCR. U6 and GAPDH were used as nuclear and cytoplasmic controls, respectively. Fluorescence in situ hybridization (FISH) was further performed to visualize the intracellular distribution of NRAV in HCC cells according to the manufacturer’s instructions. Statistical analyses All statistical analyses were performed using R software and GraphPad Prism. Continuous variables between two groups were compared using the Student’s t-test or Wilcoxon rank-sum test, as appropriate. Comparisons among three or more groups were performed using one-way analysis of variance (ANOVA) followed by appropriate post hoc testing [22, 23]. Correlations were evaluated using Spearman correlation analysis. Survival analyses were conducted using Kaplan–Meier methods with log-rank tests and Cox proportional hazards regression models. Unless otherwise specified, all in vitro experiments were repeated independently at least three times. P values for multiple testing were adjusted using the Benjamini–Hochberg method where applicable. A two-sided P < 0.05 was considered statistically significant. Results NRAV is upregulated in hepatocellular carcinoma across independent cohorts To characterize NRAV expression in hepatocellular carcinoma, we first analyzed transcriptomic data from TCGA-LIHC together with normal liver tissues from the GTEx project processed through the UCSC Xena Toil recompute unified dataset. A pan-cancer overview suggested that NRAV was dysregulated across multiple tumor types (Fig. 1A). In the liver cohort, NRAV expression was higher in TCGA-LIHC tumors than in GTEx normal liver tissues (Fig. 1B). Given the potential residual batch effects inherent to cross-cohort comparisons, this analysis was interpreted as supportive evidence. We therefore further evaluated NRAV expression within the TCGA-LIHC cohort and in independent GEO datasets. NRAV remained significantly upregulated in unpaired and paired tumor-versus-adjacent normal comparisons in TCGA-LIHC (Fig. 1C, D). Consistent findings were observed in external GEO validation datasets (Fig. 1E-G). Together, these results support that NRAV is upregulated in HCC across multiple independent cohorts. NRAV is upregulated in hepatocellular carcinoma across multiple cohorts. A Pan-cancer overview of NRAV expression in tumor and normal tissues from TCGA, as indicated. B NRAV expression in hepatocellular carcinoma (HCC) tissues from TCGA-LIHC compared with normal liver tissues from GTEx. C NRAV expression in unpaired tumor and normal samples in TCGA-LIHC. D NRAV expression in paired tumor–normal samples in TCGA-LIHC. E–G Differential expression analyses of NRAV in independent GEO cohorts, including GSE113850 (E), GSE112613 (F), and GSE67260 (G) High NRAV expression is associated with unfavorable survival outcomes in HCC Kaplan–Meier survival analyses showed that patients with high NRAV expression had significantly poorer overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI) than those with low NRAV expression in the TCGA-LIHC cohort (Fig. 2A-C). Subgroup analyses further suggested that the adverse survival association of high NRAV expression was observable in several clinical strata (Fig. 2D-I). These findings indicate that elevated NRAV expression is associated with unfavorable clinical outcome in HCC. In the present study, the survival results are interpreted as evidence of prognostic association, whereas further optimized multivariable modeling and external validation will be required to more rigorously define its independent prognostic value. High NRAV expression is associated with unfavorable survival outcomes in HCC. A–C Kaplan–Meier survival curves comparing overall survival (OS) (A), disease-specific survival (DSS) (B), and progression-free interval (PFI) (C) between NRAV-high and NRAV-low groups in TCGA-LIHC. D–I Subgroup Kaplan–Meier analyses of OS stratified by clinical characteristics, including N stage (D), M stage (E), pathological stage (F), histological grade (G), AFP level (H), and vascular invasion (I) Clinical relevance of NRAV expression in hepatocellular carcinoma To further assess the clinical relevance of NRAV expression, we examined its distribution across clinicopathological subgroups in the TCGA-LIHC cohort. NRAV expression differed significantly across several adverse clinicopathological categories, including T stage, pathological stage, tumor status, vascular invasion, and residual tumor status (Fig. 3A–F). These findings support that elevated NRAV expression is associated with more aggressive clinical features in HCC. Detailed clinicopathological characteristics stratified by NRAV expression are summarized in Supplementary Tables 3, and the corresponding univariate logistic regression results are provided in Supplementary Table 4. Associations between NRAV expression and clinicopathological features in HCC. A–F Violin plots showing NRAV expression levels across clinicopathological subgroups in TCGA-LIHC, including T stage (A), pathological stage (B), N stage (C), tumor status (D), vascular invasion (E), and residual tumor status (F) Preliminary prognostic modeling and discriminatory performance of NRAV in HCC To further explore the potential clinical relevance of NRAV, we evaluated its prognostic association and discriminatory performance in the TCGA-LIHC cohort. Cox regression analysis indicated that higher NRAV expression was associated with unfavorable survival outcome in the TCGA-LIHC cohort (Fig. 4A). Receiver operating characteristic analysis showed that NRAV expression had good discriminatory performance for distinguishing tumor from non-tumor samples in the present cohort, with an AUC of 0.901 (95% CI: 0.865–0.937) (Fig. 4B). For contextual reference within the same panel, the AFP-related ROC curve is also displayed; however, because AFP is a blood-based clinical marker whereas NRAV was evaluated using tissue transcriptomic data, the two readouts were not interpreted as directly comparable in the revised manuscript. Data completeness and evaluable cases for survival-related analyses are summarized in Supplementary Table 5. In addition, a nomogram integrating NRAV expression with clinicopathological variables was constructed to estimate 1-, 3-, and 5-year survival probabilities, and calibration plots showed general agreement between predicted and observed outcomes within the TCGA cohort (Fig. 4C, D). Given that the model was developed and assessed within a single cohort, these analyses should be regarded as preliminary and hypothesis-generating, and further internal optimization and external validation will be required before clinical application. Collectively, these findings suggest that NRAV may have potential value for prognostic stratification and tissue-level discrimination in HCC. Prognostic model construction and discriminatory performance of NRAV in HCC. A Forest plot of Cox regression analysis evaluating the prognostic value of NRAV expression and clinical variables in TCGA-LIHC. B Receiver operating characteristic (ROC) curves displayed in the same panel for contextual reference. NRAV expression showed an AUC of 0.901 (95% CI: 0.865–0.937) for distinguishing tumor from non-tumor samples in the TCGA-LIHC cohort. Because AFP is a blood-based clinical marker whereas NRAV was evaluated using tissue transcriptomic data, the two curves were not interpreted as directly comparable in the revised manuscript. C Nomogram integrating NRAV expression and clinicopathological variables for estimating 1-, 3-, and 5-year survival probabilities. D Calibration curves evaluating the agreement between predicted and observed survival probabilities at 1, 3, and 5 years NRAV-associated transcriptional programs in HCC To explore the transcriptional programs associated with NRAV expression, we performed differential expression analysis between NRAV-high and NRAV-low tumors in the TCGA-LIHC cohort. A distinct pattern of differentially expressed genes was observed between the two groups, as illustrated by the volcano plot and heatmap (Fig. 5A, B). A protein–protein interaction network constructed from these differentially expressed genes further suggested coordinated functional relationships among NRAV-associated genes (Fig. 5C). Functional enrichment analyses indicated that these genes were enriched in biological processes and pathways related to developmental programs, membrane-associated components, receptor activity, and neuroactive ligand–receptor interaction (Fig. 5D, E; Supplementary Fig. 1). Given the prognostic relevance of NRAV and the enrichment results observed above, we next examined its association with immune-related features in the TCGA-LIHC cohort. ssGSEA-based analyses suggested that NRAV-high tumors were associated with lower CD8 + T-cell-related signatures and higher macrophage- and Th2-related signatures. NRAV expression also showed positive correlations with multiple immune checkpoint molecules, including PDCD1, CD274, CTLA4, LAG3, and HAVCR2. Correlations between NRAV expression and representative immune cell marker genes are summarized in Supplementary Tables 6, and additional immune- and genome-related associations are shown in Supplementary Fig. 2. Because these analyses were derived from computational immune estimation and transcript-level correlations, they were interpreted as associative rather than causal and should be considered as preliminary evidence that warrants further experimental validation. NRAV-associated transcriptional programs and functional enrichment in HCC. A Volcano plot showing differentially expressed genes (DEGs) between NRAV-high and NRAV-low tumors in TCGA-LIHC. B Heatmap of representative DEGs between NRAV-high and NRAV-low groups. C Protein–protein interaction (PPI) network of NRAV-associated DEGs, highlighting hub genes. D GO and KEGG enrichment analyses of NRAV-associated DEGs. E Functional enrichment network visualization showing major enriched terms and their relationships NRAV knockdown suppresses malignant phenotypes in HCC cells To investigate the functional relevance of NRAV in HCC cells, NRAV was silenced in HepG2 cells using two independent siRNAs, and knockdown efficiency was confirmed by RT-qPCR (Fig. 6A). As NRAV is a non-coding transcript, silencing efficiency was assessed at the RNA level by RT-qPCR, while downstream protein changes after NRAV depletion were further examined by Western blotting (Fig. 6I). Functional assays showed that NRAV depletion reduced colony formation, migration, and invasion in HepG2 cells (Fig. 6B, C). In addition, NRAV knockdown altered the mRNA expression of several NRAV-associated genes, including ADAMTS4, B3GAT3, BTG1, CCN2, and DCAF15 (Fig. 6D-H). Consistently, western blot analysis further confirmed corresponding changes in protein expression following NRAV silencing (Fig. 6I). These data support that NRAV contributes to malignant cellular phenotypes in HepG2 cells and is associated with transcriptional and protein-level changes in multiple NRAV-related genes. NRAV knockdown suppresses malignant phenotypes in HCC cells. A RT-qPCR analysis of NRAV knockdown efficiency in HepG2 cells transfected with si-NC, si-NRAV-1, or si-NRAV-2. B Representative images of colony formation in HepG2 cells after NRAV knockdown. C Representative images of Transwell migration and invasion assays in HepG2 cells after NRAV knockdown. D–H RT-qPCR validation of NRAV-regulated genes, including ADAMTS4 (D), B3GAT3 (E), BTG1 (F), CCN2 (G), and DCAF15 (H). I Western blot analysis of ADAMTS4, B3GAT3, BTG1, CCN2, and DCAF15 protein expression following NRAV knockdown, with GAPDH as the loading control Additional validation in Huh7 cells and subcellular localization of NRAV To further assess the robustness of the in vitro findings, NRAV was silenced in Huh7 cells using two independent siRNAs, and knockdown efficiency was confirmed by RT-qPCR (Fig. 7A). Functional assays showed that NRAV depletion reduced colony formation, migration, and invasion in Huh7 cells (Fig. 7B, C). In addition, nuclear–cytoplasmic fractionation analysis indicated that NRAV was predominantly enriched in the nuclear fraction (Fig. 7D), which was further supported by FISH in both HepG2 and Huh7 cells (Fig. 7E). These results provide additional support for the growth- and motility-related effects of NRAV in HCC cells and suggest that NRAV is preferentially localized in the nucleus. Functional validation and subcellular localization of NRAV in HCC cells. A RT-qPCR analysis of NRAV knockdown efficiency in Huh7 cells transfected with si-NC, si-NRAV-1, or si-NRAV-2. B Representative images of colony formation in Huh7 cells after NRAV knockdown. C Representative images of Transwell migration and invasion assays in Huh7 cells after NRAV knockdown. D Nuclear–cytoplasmic fractionation followed by RT-qPCR analysis of NRAV localization, with GAPDH and U6 used as cytoplasmic and nuclear controls, respectively. E Fluorescence in situ hybridization (FISH) showing the intracellular localization of NRAV in HepG2 and Huh7 cells. DAPI was used for nuclear staining Discussion Primary liver cancer, particularly HCC, remains a major global health challenge because of its high incidence, late-stage presentation, and limited therapeutic efficacy [24, 25]. Although systemic and immunotherapeutic strategies have improved outcomes for selected patients, the overall survival rate for most individuals with HCC remains unsatisfactory [26, 27]. Therefore, identifying robust biomarkers that enhance diagnostic precision, improve prognostic stratification, and potentially guide therapeutic decisions is of substantial clinical importance. lncRNAs have emerged as key regulators of numerous biological and pathological processes [28], and increasing evidence has confirmed their roles in tumor initiation and progression [29]. Dysregulated lncRNAs have been implicated in several malignancies, including gastric cancer [30], lung cancer [31], liver cancer [32], and breast cancer [33]. In the present study, we found that negative regulator of antiviral response (NRAV) was consistently upregulated in HCC across TCGA-LIHC and independent GEO datasets, with the TCGA–GTEx comparison serving as supportive evidence and the paired tumor–adjacent analysis within TCGA and external GEO validation providing the main support. The consistent elevation across independent cohorts and the confirmation in paired tumor-normal samples support NRAV upregulation as a tumor-associated alteration rather than a cohort-specific artifact. In addition, the prognostic nomogram was developed using the TCGA cohort, and further validation using independent external cohorts will be necessary to assess its generalizability and reduce potential overfitting. Given the residual batch-effect concern inherent to TCGA–GTEx integration, we interpreted the cross-cohort comparison as supportive rather than standalone evidence. Clinically, NRAV expression was associated with tumor aggressiveness, including advanced T stage, higher pathological stage, vascular invasion, and elevated AFP levels. These correlations are consistent with previously described oncogenic lncRNAs that promote HCC invasion and metastasis [34]. Moreover, high NRAV expression was associated with poorer overall survival [35], disease-specific survival [36, 37], and progression-free interval, supporting its potential relevance as a prognostic indicator [38]. These observations extend earlier findings linking dysregulated lncRNAs to cancer progression and adverse outcomes in multiple malignancies [39–41]. They are also compatible with the broader concept that cytoskeleton-related remodeling contributes to tumor-cell motility and invasion [42]. To further explore the biological programs associated with high NRAV expression, we performed co-expression, enrichment, and PPI analyses in the TCGA-LIHC cohort. NRAV-associated genes were enriched in multiple biological processes and pathways, indicating broad transcriptional reprogramming in NRAV-high tumors. Notably, the PPI network highlighted genes such as CDK1, CCNB1, TOP2A, and KIF11, which are broadly involved in cell-cycle progression and mitotic regulation. These computational findings were broadly consistent with our experimental observations that NRAV knockdown suppressed colony formation and reduced migratory and invasive phenotypes in HCC cells. Immune-related analyses revealed notable associations between NRAV expression and immune-related transcriptional features in HCC. NRAV expression positively correlated with Th2-cell-, TFH-cell-, macrophage-, NK CD56bright-cell-, and activated dendritic-cell-related signatures, several of which have been implicated in tumor-associated immune modulation [43]. Conversely, negative correlations with CD8⁺ T-cell-, Th17-cell-, neutrophil-, and plasmacytoid dendritic-cell-related signatures [44, 45] were compatible with a less favorable antitumor immune context [46]. Furthermore, NRAV expression showed positive associations with multiple immune checkpoint molecules, supporting an association between higher NRAV expression and an immune-regulatory transcriptomic context. Because these analyses were based on ssGSEA-derived signatures and transcript-level correlations, they should be interpreted as associative rather than causal. In addition, the marker-gene correlation results provided complementary support for the immune-related associations observed in the ssGSEA analysis. Functional experiments further supported that NRAV is not merely associated with clinical and transcriptomic changes, but is also linked to malignant phenotypes in HCC cells. Silencing NRAV markedly suppressed colony formation, migration, and invasion in HCC cells. These phenotypic changes were accompanied by altered expression of ADAMTS4, B3GAT3, CCN2, BTG1, and DCAF15, indicating that NRAV is associated with multiple downstream molecular changes in HCC cells. Collectively, our results suggest that NRAV is associated with proliferative, motility-related, and immune-linked features in HCC, supporting its candidacy as a biomarker for further investigation. Its correlation with immune checkpoint expression raises the possibility that NRAV-associated tumor states may be relevant to immunotherapy-related stratification, although this remains speculative in the absence of treatment-response data. Although subcellular localization analyses suggest that NRAV is predominantly enriched in the nucleus, the precise molecular mechanisms underlying NRAV-mediated regulation remain to be elucidated. Future studies will be required to identify interacting proteins and downstream regulatory pathways. Because immune infiltration was estimated using a signature-based computational approach, the results may vary across deconvolution methods and were not explicitly adjusted for tumor purity in the present study. Future studies incorporating additional immune infiltration estimation methods may further strengthen the robustness of these findings. In addition, functional experiments were conducted in HepG2 cells and further validated in Huh7 cells. The present study focused on loss-of-function assays, and complementary gain-of-function experiments were not performed. Future overexpression studies will be needed to further define the bidirectional functional role of NRAV in HCC. Nevertheless, validation in additional HCC models, particularly in vivo systems, will be necessary to further assess generalizability. Recent studies provide further context for the potential biological relevance of NRAV in HCC. A 2024 ferroptosis-related lncRNA study reported that NRAV is associated with oxidative stress and iron metabolism pathways and correlates with ferroptosis-related gene signatures and immune infiltration patterns [47]. In parallel, a comprehensive 2025 review highlighted the broad roles of lncRNAs in shaping the HCC immune microenvironment, including immune cell recruitment and immune checkpoint regulation [48]. These reports are consistent with our observation that NRAV expression correlates with immune-related signatures and checkpoint molecules, supporting the possibility that NRAV may be involved in broader tumor–immune interactions. However, the present data do not distinguish whether these immune-related associations reflect direct microenvironmental remodeling by NRAV or secondary consequences of intrinsic tumor-cell states. In addition, virus-associated ncRNAs have been proposed to reshape lncRNA expression landscapes during chronic HBV infection and contribute to hepatocarcinogenesis [49]. Given that NRAV was initially described as a regulator of antiviral responses, its upregulation in HCC may reflect virus-associated immune or epigenetic alterations, which merits further investigation. Based on these findings, future studies should prioritize defining the direct molecular mechanisms underlying NRAV function, including its interacting partners and downstream regulatory networks, and evaluating its role in vivo. It will also be important to determine whether NRAV is linked to ferroptosis-related programs and whether NRAV-associated immune features have value for patient stratification in future immunotherapy-oriented studies. Conclusion This study shows that NRAV is upregulated in HCC and is associated with unfavorable prognosis, aggressive clinicopathological features, and immune-related transcriptional patterns. Functional assays further support a role for NRAV in malignant cellular phenotypes in vitro. Together, these findings identify NRAV as a candidate biomarker that warrants further mechanistic and translational investigation in HCC. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Author contributions X.F. and Y.C. contributed equally to this work. X.F. and Y.L. conceived and designed the study. Y.C. and Z.L. performed data acquisition and bioinformatics analyses. X.P., S.Y., and Z.Z. contributed to data interpretation and visualization. Y.L. supervised the study. All authors reviewed and approved the final manuscript. Funding This work was supported by the Joint Special Project for Basic Research of Yunnan Provincial Department of Science and Technology and Kunming Medical University (Key Program, No. 202301AY070001-019), the Joint Special Project for Basic Research of Yunnan Provincial Department of Science and Technology and Kunming Medical University (General Program, No. 202401AY070001-260), and the Scientific Research Fund of Qujing Medical College (No. 2025ZD07). Data availability The datasets analyzed in this study are publicly available. The TCGA-LIHC dataset is available from The Cancer Genome Atlas (TCGA) via the Genomic Data Commons portal (https://portal.gdc.cancer.gov/). The GTEx dataset can be accessed through the UCSC Xena platform (https://xenabrowser.net/). The unified TCGA and GTEx RNA-seq data (Toil recompute) are available from UCSC Xena. Gene expression datasets from the Gene Expression Omnibus (GEO) are available under accession numbers GSE113850, GSE112613, and GSE67260 (https://www.ncbi.nlm.nih.gov/geo/). All data used in this study are publicly accessible from the above repositories.The experimental data generated in this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate This study used publicly available, de-identified datasets and established cell lines. Therefore, ethical approval was not required. Consent for publication Not applicable. This manuscript does not contain any individual person’s data in any form. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xingbao Fang and Yan Cai have contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement The datasets analyzed in this study are publicly available. The TCGA-LIHC dataset is available from The Cancer Genome Atlas (TCGA) via the Genomic Data Commons portal (https://portal.gdc.cancer.gov/). The GTEx dataset can be accessed through the UCSC Xena platform (https://xenabrowser.net/). The unified TCGA and GTEx RNA-seq data (Toil recompute) are available from UCSC Xena. Gene expression datasets from the Gene Expression Omnibus (GEO) are available under accession numbers GSE113850, GSE112613, and GSE67260 (https://www.ncbi.nlm.nih.gov/geo/). All data used in this study are publicly accessible from the above repositories.The experimental data generated in this study are available from the corresponding author upon reasonable request.