A novel disulfidptosis-related lncRNAs index to predict prognosis and therapeutic target in hepatocellular carcinoma

Introduction

The 2020 worldwide oncology statistics highlight liver cancer as a leading malignancy globally. With an estimated 900,000 newly diagnosed cases globally, it holds the sixth position in terms of malignant tumor incidence worldwide. Moreover, hepatocellular carcinoma (HCC) stands out as a significant contributor to cancer-related deaths, ranking fourth globally in terms of prevalence. Constituting approximately 90% of primary liver cancer cases, HCC emerges as the prevailing form within liver malignancies (1-3). Chronic hepatitis B/C infections, alcohol consumption, exposure to aflatoxins, and metabolic factors are identified as the principal risk factors for HCC (4). Despite the array of treatment modalities accessible for HCC, including tumor resection, intervention, radiofrequency ablation, targeted therapy, and others, the overall therapeutic efficacy still lags, as evidenced by a 5-year survival rate hovering around 18%. This is mainly due to the high recurrence rates and the emergence of drug resistance (5,6). Consequently, the exploration of predictive biomarkers for the survival and disease progression of HCC patients assumes paramount importance in augmenting clinical diagnosis and treatment.

Increasing evidence suggests that cancer exhibits resistance to regulated cell death. Regulated cell death, in its various forms, influences both the progression of cancer and its response to treatment (7,8). A recently discovered form of cell death has been identified: disulfidptosis, characterized by an increase in intracellular disulfides induced by the overexpression of SLC7A11 under conditions of glucose deficiency. This process leads to an increase in cystine uptake through SLC7A11, which, when coupled with a lack of glucose, causes significant disulfide stress. Amidst stress conditions, disulfide bonds atypically form within the actin cytoskeletal proteins, causing actin filaments to contract and detach from the plasma membrane. Unlike conventional cell demise routes such as ferroptosis and apoptosis, disulfidptosis is resistant to cell death inhibitors but experiences heightened effects in the presence of thiol-oxidizing agents like diamide (9,10). In-depth exploration of various pathways leading to cell demise in neoplastic cells and gaining insights into the expression patterns and mechanisms governing the formation of disulfide bonds could unveil novel therapeutic approaches for HCC.

Long non-coding RNAs (lncRNAs), transcripts exceeding 200 base pairs falling under the non-coding RNA category, were initially deemed devoid of functionality. Nevertheless, recent studies have demonstrated that lncRNAs exert critical regulatory roles in cellular processes such as the cell cycle, differentiation, invasion, metastasis, and cell death. As exemplified by their modulation of ferroptosis, lncRNAs influence tumor cell death and other oncological processes (11,12). Recent study revealed that lncRNA TEX41 is up-regulated in HCC tissues compared with adjacent non-tumor tissues and correlates with lymph-node metastasis and TNM stage. Silencing TEX41 suppresses proliferation, migration, invasion, and EMT in HCC cell lines, potentially via the miR-200a-3p–BIRC5 axis (13). Moreover, lncRNA TMEM105 promotes pancreatic cancer progression and suppresses disulfidptosis via the β-catenin–c-MYC–GLUT1 axis, thereby representing a potential therapeutic target (14). Yet, there remains considerable uncertainty about the effects and mechanisms of disulfidptosis-related long non-coding RNAs (DRLs) in HCC.

Our study develops a model associated with DRLs utilizing data from both The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) databases. Our model successfully predicted survival rates for HCC patients through thorough investigation. Moreover, this study has explored the intricate interplay among the tumor microenvironment, immune infiltration, immune rejection, and tumorigenesis, thus laying the groundwork for understanding therapeutic intervention mechanisms. By meticulously analyzing this relationship, this study offers valuable insights into treatment and prognostic strategies, thereby providing robust support for enhancing medical interventions. This research not only advances our comprehension of HCC biology but also significantly contributes to future clinical practice and the formulation of personalized treatment plans. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-610/rc).

Methods

Data acquisition and DRG screening

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. In this investigation, we acquired comprehensive datasets encompassing transcriptomic, clinical pathological, and copy number variation (CNV) information from patients diagnosed with HCC. For patient inclusion, the following criteria were applied: (I) histopathologically confirmed HCC; (II) patients with complete survival follow-up data. The transcriptomic data originated from both TCGA and ICGC databases. Specifically, TCGA provided 374 HCC samples, including 50 normal liver tissue samples. Simultaneously, ICGC contributed 243 HCC samples and 202 normal liver tissue samples. Clinical pathological and CNV data originated from TCGA (https://portal.gdc.cancer.gov/). We performed subsequent analysis on the data using perl and R. Differential analysis was performed on tumor tissue and liver cell tissue using screening criteria of P<0.05 and ∣log₂FC∣ ≥1 by limma package (15). An intersection analysis was then carried out on these genes related to disulfidptosis and differential mRNA from both databases, pinpointing key disulfidptosis-related genes (DRGs) for additional exploration (10). The expression levels of HCC DRGs, the frequencies of their CNVs, and the representation of genomic variation sites on chromosomes are depicted through box plots, lollipop charts, and circular plots, respectively. Additionally, mutations linked to DRGs in HCC were illustrated using waterfall plots.

Building a risk model for DRLs

We utilized the Pearson correlation analysis method to assess the co-expression relationship between lncRNAs and DRGs (P<0.05, |r|>0.4). LncRNAs meeting these criteria were categorized as DRLs, and their association was depicted using Sankey diagrams. To construct a prognostic model, we first applied univariate Cox proportional hazards regression to each lncRNA in the training set (n=185) and retained those with P<0.001. Second, least absolute shrinkage and selection operator-Cox proportional hazards model (LASSO-Cox) regression (glmnet, α=1, family = “cox”) was performed on the retained genes. The penalty parameter λ was chosen via 10-fold cross-validation using the minimum criterion (lambda.min). Third, a multivariate Cox model with backward-forward stepwise selection Akaike Information Criterion (AIC) was fitted to the LASSO-selected genes to obtain the final five DRLs. To mitigate overfitting, all feature selection and parameter tuning were strictly confined to the training set; the validation set (n=185) was used solely for independent assessment. The risk score for each patient was calculated as a linear combination of the expression levels of the five DRLs weighted by their multivariate Cox coefficients. Patients were stratified into high- or low-risk groups using the median risk score derived from the training cohort. The predictive performance was evaluated by time-dependent ROC curves (timeROC) at 1 year and by Kaplan–Meier survival analyses in both the training and validation sets (16-18). This study utilized a boxplot and a heatmap to demonstrate the differential expression of the five DRLs and their co-expression with 11 DRGs.

Validation of the DRLs prognostic signature

In investigating the impact of DRLs model on HCC, this study performed thorough validations across various datasets, encompassing TCGA, training, and testing cohorts. Patients were stratified into high and low-risk groups using the median as a benchmark. Subsequently, this study conducted an analysis of the disparity in survival rates between patient groups categorized as high and low risk, employing the ’survminer’ and ’survival’ packages (19). The creation of risk curves, scatter plots, and heatmaps further confirmed the risk value distribution among various patient risk categories, facilitating the estimation of risk for death related to HCC.

Prognostic independence assessment and nomogram construction

Utilizing packages such as ‘timeROC’, ’survminer’ for Cox regression analysis, this study aims to evaluate the sensitivity and specificity of age, tumor grade, stage, and risk models in predicting survival rates. Forest plots visualized these analyses to confirm the prognostic models’ independence from clinical features. The receiver operating characteristic (ROC) curve and the C-index assess the predictive accuracy of different factors in determining survival rates. Additionally, this study compared overall survival (OS) across different tumor stages (Stage I vs. Stages II–IV) based on the prognostic model’s median scores, utilizing Kaplan-Meier curves for illustration.

Functional enrichment analysis

This study utilized the limma package to analyze the differential genes between two risk groups (∣log₂FC∣≥1, FDR <0.05). Subsequently, this research conducted functional enrichment analysis of these genes using Gene Ontology (GO). Pathway enrichment analysis of the differential genes was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG). Additionally, we employed Gene Set Enrichment Analysis (GSEA) to further elucidate the potential molecular mechanisms of the model (20-23).

Immune-related functional analysis

This study employed the “ESTIMATE” R package to compare the disparities in stromal, immune, and estimate scores among the gene expression data of the two risk groups in the model (24). In addition, Differential analysis of immune cell infiltration and immune-related functions was conducted between high and low-risk groups using the CIBERSORT method (25).

Tumor mutation burden (TMB) analysis

The ‘limma’ R package facilitated the analysis of tumor mutation burden (TMB) disparities between patient groups stratified by high and low risks. In parallel, visualization of mutation frequency was conducted utilizing the ‘maftools’ R package. To assess the influence of TMB on survival analysis, cases were stratified into two groups—high TMB and low TMB—determined by the median TMB value. Concurrently, leveraging both high and low-risk classifications within the model, we employed the ’survival’ and ’survminer’ packages for a comprehensive analysis of the prognostic impact of TMB and the model.

Tumor Immune Dysfunction and Exclusion (TIDE) analysis and drug sensitivity for HCC

Firstly, this study integrated data from the TIDE platform and analyzed the disparities in immune evasion between patient groups categorized as high and low-risk, utilizing the limma package. The application of TIDE scores facilitated the investigation of variations in treatment effectiveness (26). In addition, this study investigated the drug sensitivity of patients across various risk groups using the oncoPredict method with Genomics of Drug Sensitivity in Cancer version 2 (GDSC2) platform data. It evaluates how patients in distinct risk categories respond to drug therapy by examining drug sensitivity scores (27).

Statistical analysis

This study employed R version 4.3.2 and Perl version 5.32.1.1 for data analysis. Differential analysis primarily utilized the limma package, with statistical significance set at a P<0.05 unless stated otherwise.

Results

Expression profiles and mutational landscape of DRGs

In individuals diagnosed with HCC, this study identified 11 significantly differentially expressed genes, namely ACTN2, FLNC, MYL6B, SQSTM1, BOP1, DBN1, FANCI, SLC7A11, PPM1F, ME1, and PPIH, all exhibiting upregulation in tumor tissues. Notably, ACTN2, FLNC, FANCI, and SLC7A11 demonstrated expression levels elevated by more than 8-fold (Figure 1A). Our subsequent exploration delved into CNVs of these 11 genes in HCC. The findings revealed a predominant occurrence of copy number increases in HCC tissues for these genes. In terms of changes in CNV frequency, PPIH demonstrated the most significant DNA copy number losses, whereas BOP1 exhibited the highest frequency of copy number increases (Figure 1B). The chromosomal regions where these gene copy number variations transpired are elucidated (Figure 1C). In order to elucidate the mutational status of these 11 genes within tumor specimens, this research employed a cascade graph, facilitating a thorough comprehension of the varied mutation types and their respective implications in HCC. Among the 431 subjects comprising this HCC cohort, this study identified 66 instances manifesting somatic mutations, with FLNC displaying the highest mutation frequency, primarily marked by missense alterations (Figure 1D).

Figure 1 Characteristics of disulfidptosis-related genes in HCC. (A) Differential expression of disulfidptosis-related genes in HCC and normal liver tissue. (B) CNVs of 11 disulfidptosis-related genes in HCC. (C) The site of CNVs in HCC on chromosomes. (D) Somatic mutation landscape of disulfidptosis-related genes. ***, P<0.001. CNV, copy number variation; HCC, hepatocellular carcinoma.

Construction of the DRLs prognostic model

This study identified significant correlations between 796 lncRNAs and 11 DRGs using Pearson correlation analysis (Figure 2A). After conducting Cox regression analysis on 796 lncRNAs associated with disulfidptosis, 85 lncRNAs with distinct expression patterns were unveiled, serving as significant prognostic indicators (Figure 2B). This study amalgamated transcriptome data from the TCGA dataset with clinical data and randomly allocated them into training and testing groups. No significant differences in clinical pathological features were observed between the three groups (Table 1). The final risk model was determined through LASSO-Cox regression analysis, which incorporated five lncRNAs: TMCC1-AS1, LINC01224, MKLN1-AS, MIR210HG, and DANCR. Figure 2C,2D illustrate the Cvft curve and (λ) curve. This research demonstrated the link between these five lncRNAs and DRGs, predominantly showing significant positive correlations with FLNC, MYL6B, SQSTM1, BOP1, DBN1, FANCI, SLC7A11, PPM1F, ME1, and PPIH (Figure 2E). The expression of these five lncRNAs exhibited a notable upregulation trend in tumor tissues (Figure 2F).

Figure 2 Identification of disulfidptosis-related lncRNAs and construction of prognostic model in HCC. (A) The Sankey diagram illustrates 796 lncRNAs significantly associated with disulfidptosis-related genes. (B) The forest plot displays 85 disulfidptosis-related lncRNAs associated with hepatocellular carcinoma survival. (C) LASSO coefficient profiles of model. (D) Optimal λ value of model. (E) The heatmap shows the correlation between 11 disulfidptosis-related genes and 5 disulfidptosis-related lncRNAs in the model. (F) Differential expression of disulfidptosis-related lncRNAs in HCC and normal liver tissue. HCC, hepatocellular carcinoma; LASSO, least absolute shrinkage and selection operator; lncRNA, long non-coding RNA.

Evaluation and verification of the risk prognostic model

To evaluate the prognostic model’s predictive power, prognostic analyses were performed on three separate cohorts: TCGA, training, and testing. The study categorized cases into two groups based on the median score of the risk model. Across three cohorts, low-risk group exhibited significantly better OS compared to high-risk group (Figure 3A-3C). Furthermore, investigation unveiled a heightened mortality risk among individuals classified in high-risk category across all cohorts (Figure 3D-3F). Finally, this study carefully assessed the progression-free survival (PFS) of HCC patients, supporting our OS predictions, highlighting better PFS outcomes for those identified as low-risk (Figure 3G).

Figure 3 Prognostic model’s predictive power. (A-C) Low-risk group exhibited significantly better overall survival in the TCGA, test, and training cohorts. (D-F) A heightened mortality risk among individuals classified in the high-risk category across the TCGA, test, and training cohorts. (G) Low-risk group exhibited significantly better progression free survival in the TCGA cohort. TCGA, The Cancer Genome Atlas.

Independent hazard attribution to DRLs signature

This study aimed to examine the association between age, gender, tumor grade, stage, and risk models with the survival time and status of patients diagnosed with HCC, utilizing Cox regression analysis. The notable point is that there are significant differences in tumor staging and risk model between different groups (P<0.001). This implies that the predictive capacities of the risk model and tumor staging for prognosis stand independently of other clinical features (Figure 4A,4B). The results of ROC curve analysis validate that the area under the curve (AUC) for the risk model is the highest, measuring 0.720, surpassing both tumor staging (0.680) and other clinical features (Figure 4C). Moreover, the risk model demonstrates robust performance in accurately predicting prognosis at various time intervals. Specifically, its AUC values at 1, 3, and 5 years stand at 0.778, 0.720, and 0.664, respectively (Figure 4D). Corroborating these findings, the C-index curve exhibits values surpassing those of other clinical features (Figure 4E). In addition, this study meticulously examined the survival rates, conducting thorough comparisons between different risk groups throughout tumor stages. The results consistently exhibited heightened survival rates within the low-risk cohort across various disease stages (Figure 4F,4G).

Figure 4 Validating the clinical pathological features and model’s accuracy in predicting prognosis. (A) Univariate Cox regression analysis. (B) Multivariate regression analysis. (C) The ROC curve illustrates the precision of age, gender, tumor grade, tumor stage, and risk model. (D) The ROC curve for 1-, 3-, and 5-year survival. (E) C-index illustrates the precision of risk model in predicting the 1-, 3-, and 5-year survival. (F,G) The Kaplan-Meier curve demonstrates that across various stages. AUC, area under the curve; CI, confidence interval; HR, hazard ratio; ROC, receiver operating characteristic.

Functional enrichment analysis of the DRLs model in HCC

Genes exhibiting distinct expression patterns between different groups were analyzed through GO analysis. The results highlighted their involvement in chromosome segregation, nuclear division, chromosomal region, organelle fission, and microtubule binding (Figure 5A). Additionally, the analysis of KEGG underscored their involvement in vital pathways related to cell cycle, cytokine-cytokine receptor interaction, motor proteins, and the intricate dysregulation of transcriptional misregulation in cancer progression (Figure 5B). Furthermore, the GSEA results reveal significant enrichment in cell cycle regulation, ECM receptor interaction, and neuroactive ligand-receptor interaction. Conversely, low-risk group demonstrates predominant enrichment in fatty acid metabolism, primary bile acid biosynthesis, and retinol metabolism (Figure 5C,5D). These discoveries offer potential molecular insights into the varying roles of genes in HCC across different risk groups.

Figure 5 The GO, KEGG, and GSEA analysis of the model. (A) GO analysis. (B) KEGG analysis. (C,D) GSEA analysis. BP, biological process; CC, cellular component; GO, Gene Ontology; GSEA, Gene Set Enrichment Analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

Immune landscape profiling of the risk model

This study employed the estimation algorithm to compare the tumor microenvironment across various cohorts within the liver cancer risk model. Our findings reveal a substantial increase in both stromal scores and Estimation of STromal and Immune cells in MAlignant Tumours using Expression data (ESTIMATE) scores within the low-risk group (Figure 6A). This suggests a diminished infiltration of immune cells in high-risk HCC. Recent investigations reveal notable alterations in immune cell populations within the high-risk HCC group. Specifically, CD4 memory resting T cells, monocytes, and resting mast cells exhibit a significant decrease. Conversely, resting NK cells and M0 macrophages demonstrate a substantial increase. Remarkably, the remaining immune cells display no significant variations (Figure 6B). Additionally, the analysis of immune function reveals ten noteworthy disparities across various risk groups of HCC. Compared with the high-risk group, the low-risk group exhibited significantly elevated levels of B cells, cytolytic activity, dendritic cells (DCs), mast cells, neutrophils, natural killer (NK) cells, tumor-infiltrating lymphocytes (TIL), and both type I and type II IFN responses, whereas macrophages were markedly decreased. (Figure 6C).

Figure 6 Different immune cell infiltration levels in model. (A) The violin plot illustrates the disparities in stromal, immune, and estimate scores between the two risk groups. (B) The variance in immune cell infiltration within the risk model. (C) The variance in immune cell functions within the risk model. *, P<0.05; **, P<0.01; ***, P<0.001. TME, tumor microenvironment.

Mutational landscape of the risk model

The TMB significantly influences tumor prognosis. In this investigation, the TMB for high and low-risk cohorts was 87.08% and 75.41%, respectively. Although the high-risk cohort displayed a higher mutation rate, this disparity lacked statistical significance. However, in the high-risk group, the mutation frequencies of TP53 and CTNNB1 show a significant increase (Figure 7A-7C). Next, this study conducted a comparison of OS among various groups stratified by TMB. The findings indicated a notably superior OS within the low TMB group (Figure 7D). Ultimately, by integrating the risk model with TMB grouping, this study stratified the cases into four distinct groups. Remarkably, the cohort characterized by both low risk and elevated TMB exhibited the most favorable survival rates among HCC cases (Figure 7E).

Figure 7 Differential TMB between high and low risk groups. (A) Mutation frequencies in the high-risk group. (B) Mutation frequencies in the low-risk group. (C) Differential analysis was conducted to compare the TMB between high and low-risk groups. (D) Kaplan-Meier curves showing the difference in OS between different TMB groups. (E) Kaplan-Meier curves showing the difference in OS between four groups. OS, overall survival; TMB, tumor mutational burden.

Predicting tumor response to therapy

Drug resistance emerges as the main factor for cancer recurrence and mortality related to cancer. Remarkably, in certain instances of HCC, immune checkpoint blockers (ICBs) have shown significant therapeutic success. Hence, this study conducted an analysis to examine the immune evasion status among distinct risk model cohorts by employing TIDE scores. Observations revealed that individuals diagnosed with low-risk HCC displayed notably reduced TIDE scores (Figure 8A), indicative of enhanced effectiveness of immune checkpoint inhibitors in this subgroup. Furthermore, an investigation was conducted to probe the nexus between HCC risk scores and susceptibility to alternative anti-tumor agents. The drug sensitivity score was computed based on data from the GDSC2 database, revealing that patients in the low-risk group exhibit greater sensitivity to numerous anticancer medications, including Bcl-2 inhibitors (such as ABT737, Navitoclax, WEHI-539) (Figure 8B), inhibitors of EGFR tyrosine kinase (like gefitinib, erlotinib, afatinib, lapatinib, osimertinib, sapitinib) (Figure 8C), and PI3K inhibitors (including alpelisib, AMG-319, CZC24832, GNE-317, pictilisib, taselisib) (Figure 8D). This indicates that the risk model developed in our study holds significant long-term potential for enhancing drug treatment strategies for HCC. Additionally, this study investigated the relationship between the expression levels of lncRNAs included in the prognostic model and drug sensitivity. The results demonstrated that low expression of DANCR, MIR210HG, MKLN1-AS, and TMCC1-AS1 was associated with increased sensitivity to the EGFR tyrosine kinase inhibitor afatinib. Similarly, reduced expression of DANCR, MIR210HG, and TMCC1-AS1 was linked to enhanced sensitivity to another EGFR tyrosine kinase inhibitor, sapitinib. Furthermore, low expression of all lncRNAs within the risk model correlated with heightened sensitivity to 5-fluorouracil (Figure S1).

Figure 8 A comparative analysis was conducted to assess the differential sensitivity to ICB and other anti-tumor drugs across various risk groups. (A) Difference in TIDE scores between the two risk groups. (B) Drug sensitivity analysis of Bcl-2 inhibitors. (C) Drug sensitivity analysis of EGFR tyrosine kinase inhibitors. (D) Drug sensitivity analysis of PI3K inhibitors. ICB, immune checkpoint blockade; TIDE, Tumor Immune Dysfunction and Exclusion.

Discussion

HCC greatly impacts global health, characterized by its high occurrence and death rates. Despite the advancements made in therapeutic interventions, including surgical procedures, ablative methods, radiation therapy, and targeted pharmacotherapies, the clinical outcomes for individuals diagnosed with HCC persistently demonstrate suboptimal results. This underscores the urgent necessity to explore new biomarkers and therapeutic agents (3,5,28). In recent years, a growing body of researchers has focused on investigating the association between lncRNAs and tumors. Their aim is to address the limitations of current tumor treatments through comprehensive exploration of lncRNAs. These investigations underscore the important role of lncRNAs in driving tumor proliferation, invasion, and metastasis. Numerous investigators have adeptly leveraged lncRNAs to construct prognostic models, resulting in encouraging predictive accuracies (29,30). In 2023, Liu and colleagues introduced a new mechanism linked to metabolism-driven cell death, named disulfidptosis. Under glucose scarcity, the enhanced expression of SLC7A11 leads to abnormal buildup of intracellular disulfides, which in turn promotes unusual disulfide bond creation with cytoskeletal proteins (9,10). Previous investigations suggest that disulfidptosis markers can prognosticate various cancer types, including bladder cancer and HCC, offering a novel avenue for cancer therapy (31,32). Currently, the study of DRLs relating to the development, advancement, signaling pathways, prognosis, and treatment of HCC is still insufficient. This study identified five distinct DRLs in HCC and formulated a risk model based on these lncRNAs, thereby offering novel insights into the prognosis and therapeutic strategies for individuals with HCC.

Bioinformatics analysis is extensively utilized in cancer research, encompassing detection, diagnosis, prognosis assessment, treatment, and drug screening, thus expanding the array of methods available for cancer management (33-35). This investigation delves into the function of DRLs in HCC, aiming to devise a prognostic framework capable of predicting outcomes. In this investigation, 11 differential DRGs were identified through differential analysis, namely ACTN2, FLNC, MYL6B, SQSTM1, BOP1, DBN1, FANCI, SLC7A11, PPM1F, ME1, and PPIH. Significantly, BOP1 showed the most notable increase in copy number, whereas PPIH experienced the most pronounced decrease. FLNC had the highest mutation frequency, primarily missense mutations. Previous research indicates a link between high BOP1 expression and advanced tumor stage, microvascular invasion, and shorter PFS in HCC. The heightened expression of BOP1 fosters the migratory and invasive tendencies of liver cancer cells. PPIH’s significant overexpression in HCC correlates with tumor stage, differentiation, and TP53 mutation status. Filamin C (FLNC) protein’s substantial upregulation in HCC suggests a strong association with invasion and metastasis (36-38). Through comprehensive analysis of 11 significantly distinct DRGs, a prognostic model for DRLs was constructed, encompassing TMCC1-AS1, LINC01224, MKLN1-AS, MIR210HG, and DANCR. Prior studies have connected TMCC1-AS1 and MIR210HG expression with copper metabolism and N6-methyladenosine (m6A) regulation, and a negative correlation with OS in HCC. Silencing of LINC01224 results in the reduction of CHEK1 expression by constraining its interaction with miR-330-5p, thereby attenuating HCC cell clonogenicity, proliferation, migration, and invasion. Contrarily, the overexpression of LINC01224 in nude mice stimulates tumor proliferation by elevating the expression levels of CHEK1. MKLN1-AS elevates HDGF expression by capturing miR-654-3p, and its silencing reduces HCC cell invasion, migration, and proliferation, induces apoptosis in vitro, and decreases in vivo cell proliferation. DANCR, recognized for its abnormal expression in various cancers, exerts substantial effects on apoptosis, proliferation, invasion, migration, chemotherapy resistance in vitro (39-42). These insights align with our findings, showing TMCC1-AS1, LINC01224, MKLN1-AS, MIR210HG, and DANCR upregulation in HCC tissues and association with HCC patient survival.

Our research endeavors led to the development of a pioneering clinical prognostic framework through the meticulous analysis of the expression profiles of five distinct lncRNAs. Noteworthy, individuals identified as low-risk exhibited markedly elevated rates of OS and PFS in contrast to those categorized as high-risk. Additionally, our observations underscore the efficacy of this prognostic framework as an autonomous prognostic determinant, surpassing conventional clinical variables in prognostic precision. Additionally, our analysis of HCC patients at varying disease stages consistently demonstrated higher survival rates among those classified as low-risk. By conducting a detailed examination of relevant studies (39-42). Our investigation delved into the intricate role of five DRLs (TMCC1-AS1, LINC01224, MKLN1-AS, MIR210HG, DANCR) in influencing fundamental biological functions such as apoptosis, proliferation, invasion, and migration in HCC. As a result, this study highlights the significant diagnostic and prognostic potential of newly developed clinical prognostic model for individuals diagnosed with HCC.

The subsequent investigation, GSEA results reveal significant enrichment in cell cycle regulation, ECM receptor interaction, and neuroactive ligand-receptor interaction within the high-risk group. Conversely, the low-risk group demonstrates predominant enrichment in fatty acid metabolism, primary bile acid biosynthesis, and retinol metabolism. Existing studies suggest a tight link between the ECM and immune system, providing essential structural backing for the regular physiological functions of tissue cells. Its rich array of protein constituents and immunoreactive molecules is instrumental in governing immune regulation, both in maintaining the body’s homeostasis and in response to pathological conditions. Conversely, the immune system is tasked with preserving the microenvironmental equilibrium of the matrix and facilitating matrix restoration post-injury. There is a reciprocal dependence between ECM and immune cells, and exploring this complex interplay aids in the treatment of diseases and slowing down the aging process (43). The elimination of 18S RNA m6A modification by the methyltransferase METTL5 disrupts the formation of 80S ribosome complexes, which leads to a reduction in the expression of mRNAs associated with fatty acid metabolism. Conversely, the modulation of METTL5’s activity by ACSL4 via the regulation of fatty acid metabolism is noteworthy. Targeting of both ACSL4 and METTL5 could potentially decelerate the onset and advancement of HCC (44). The increasing recognition of NAFLD/NASH as a significant risk factor for HCC is closely linked to the global surge in obesity and NAFLD due to shifts in diet and lifestyle. This could potentially contribute to a higher prevalence of NAFLD/NASH-associated HCC. Suppressing the fatty acid metabolism signaling axis can forestall the conversion of NAFLD into HCC, offering novel approaches for preventing NASH-related HCC (45). This is consistent with the findings from our GSEA pathway analysis. As a result, this study suggests that the prognostic model could impact the incidence and progression of HCC via factors such as the matrix microenvironment, immune system, and fatty acid metabolism, offering new directions for immunotherapy in this area.

The occurrence and progression of tumors are highly intricate processes, wherein variations in the extent of immune cell infiltration, shifts in the expression of diverse functional immune cells, and modifications in the tumor microenvironment significantly impact the prognosis of various tumors, including HER2-positive breast cancer (46-49). Our study highlights the variations in microenvironmental and immune factors across distinct subgroups within the risk model for HCC DRLs. The significant finding is that high-risk HCC group demonstrates reduced levels of CD4 memory resting T cells, monocytes, and resting macrophages. Systemically distributed CD4⁺ memory T cells serve a protective function in defense against reinfection and cancer. Recent investigations underscore their significance in the efficacy of ICB therapy. Monocytes, capable of producing tumor-killing mediators and stimulating NK cells, further underscore the intricate immune dynamics within the tumor microenvironment, potentially informing future immunotherapeutic strategies for HCC (50,51). TMB emerges as a critical determinant affecting the effectiveness of ICB therapy in malignancies. High TMB tumors, rich in neoantigens, may induce strong ICB responses, yet the effectiveness of the anti-tumor immune reaction primarily depends on the interactions between tumors and lymphocytes (52). Our evaluation revealed no notable disparities in overall TMB between different risk HCC cohorts. However, mutations in the TP53 and CTNNB1 genes exhibit a higher frequency in high-risk group, corroborating prior research that positions mutations in the TERT promoter, CTNNB1, and TP53 at the forefront of HCC advancement. The low-risk group’s reduced TIDE scores hint at a greater likelihood of responding to immunotherapy, incorporating the differences among specific immune cell subpopulations, patients in the low-risk group may achieve better responses to immunotherapy (53). Nevertheless, the mechanistic underpinnings linking DRLs to the observed disparities in immune-cell subpopulations remain largely undefined; we therefore propose to employ both cellular and in vivo models to rigorously delineate the functional roles of DRLs in modulating immune evasion and immune infiltration. Furthermore, there was a noticeable decrease in responsiveness to other anti-cancer treatments, like EGFR tyrosine kinase inhibitors, Bcl-2 inhibitors, and PI3K pathway inhibitors, in the high-risk HCC patients. Previous research has shown the effectiveness of combining EGFR inhibitors, such as gefitinib, post-oral lenvatinib failure in patients with EGFR-positive liver cancer, achieving significant control of HCC progression and the objective response rate (ORR) reached 33.3%. The PI3K/AKT/mTOR pathway plays a pivotal role in tumor initiation and progression, with its significance underscored by its involvement in HCC development, thereby emphasizing the therapeutic potential of DHW-208, an innovative pan-PI3K inhibitor, which has shown efficacy in curbing tumor growth while presenting a well-tolerated safety profile. Similarly, BCL-2 inhibitors, notably Venetoclax, exhibit promising anti-tumor effects by selectively targeting BCL-2, thus impeding tumor survival (54-56).

Conclusions

To summarize, our research pinpointed 796 DRLs linked to HCC, leading to the creation of a predictive model that incorporates 5 DRLs, developed through LASSO-COX regression analysis. This model independently predicts OS and PFS among HCC patients, with subsequent pathway analysis shedding light on its potential implications for tumor immune microenvironment modulation, TMB variations, and drug resistance mechanisms. Our results indicate that this model could substantially impact the development and progression of HCC, potentially informing personalized approaches to immunotherapy and targeted treatments. Although this study validated the results through analyses of multiple datasets, it is primarily based on data analysis and has certain limitations, as all data were obtained from public databases without experimental verification. Given the inherent limitations in bioinformatics research, future investigations may yield novel insights or modify current interpretations. Therefore, continuous scrutiny and validation of the research findings remain imperative to maintain scientific rigor, and further experimental investigations are necessary to validate and explore the underlying mechanisms.

Acknowledgments

We would like to thank TCGA and ICGC databases for making the data available.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-610/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-610/prf

Funding: This study was supported by Department of Education of Guangdong Province (No. 2022KTSCX021).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-610/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Jemal A, Ward EM, Johnson CJ, et al. Annual Report to the Nation on the Status of Cancer, 1975-2014, Featuring Survival. J Natl Cancer Inst 2017;109:djx030. [Crossref] [PubMed]
3. Llovet JM, Kelley RK, Villanueva A, et al. Hepatocellular carcinoma. Nat Rev Dis Primers 2021;7:6. [Crossref] [PubMed]
4. McGlynn KA, Petrick JL, El-Serag HB. Epidemiology of Hepatocellular Carcinoma. Hepatology 2021;73:4-13. [Crossref] [PubMed]
5. Cucarull B, Tutusaus A, Rider P, et al. Hepatocellular Carcinoma: Molecular Pathogenesis and Therapeutic Advances. Cancers (Basel) 2022;14:621. [Crossref] [PubMed]
6. Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33. [Crossref] [PubMed]
7. Hellmann MD, Ciuleanu TE, Pluzanski A, et al. Nivolumab plus Ipilimumab in Lung Cancer with a High Tumor Mutational Burden. N Engl J Med 2018;378:2093-104. [Crossref] [PubMed]
8. Peng F, Liao M, Qin R, et al. Regulated cell death (RCD) in cancer: key pathways and targeted therapies. Signal Transduct Target Ther 2022;7:286. [Crossref] [PubMed]
9. Koppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell 2021;12:599-620. [Crossref] [PubMed]
10. Liu X, Nie L, Zhang Y, et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol 2023;25:404-14. [Crossref] [PubMed]
11. Chodurska B, Kunej T. Long non-coding RNAs in humans: Classification, genomic organization and function. Noncoding RNA Res 2025;11:313-27. [Crossref] [PubMed]
12. Ju Y, Lv Y, Liu X, et al. Role of long non-coding RNAs in the regulation of ferroptosis in tumors. Front Immunol 2025;16:1568567. [Crossref] [PubMed]
13. Li Z, Xu H, Chen H, et al. The high expression of Long noncoding RNA TEX41 promotes the proliferation, migration, and invasion of hepatocellular carcinoma. Transl Cancer Res 2025;14:3175-85. [Crossref] [PubMed]
14. Yin Y, Sun Y, Yao H, et al. TMEM105 modulates disulfidptosis and tumor growth in pancreatic cancer via the β-catenin-c-MYC-GLUT1 axis. Int J Biol Sci 2025;21:1932-48. [Crossref] [PubMed]
15. Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47. [Crossref] [PubMed]
16. Simon N, Friedman J, Hastie T, et al. Regularization Paths for Cox's Proportional Hazards Model via Coordinate Descent. J Stat Softw 2011;39:1-13. [Crossref] [PubMed]
17. Tibshirani R. The lasso method for variable selection in the Cox model. Stat Med 1997;16:385-95. [Crossref] [PubMed]
18. Zhu T, Zheng J, Hu S, et al. Construction and validation of an immunity-related prognostic signature for breast cancer. Aging (Albany NY) 2020;12:21597-612. [Crossref] [PubMed]
19. Blanche P, Dartigues JF, Jacqmin-Gadda H. Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med 2013;32:5381-97. [Crossref] [PubMed]
20. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
21. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 2016;32:2847-9. [Crossref] [PubMed]
22. Wu T, Hu E, Xu S, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb) 2021;2:100141. [Crossref] [PubMed]
23. Yu G, Wang LG, Yan GR, et al. DOSE: an R/Bioconductor package for disease ontology semantic and enrichment analysis. Bioinformatics 2015;31:608-9. [Crossref] [PubMed]
24. Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612. [Crossref] [PubMed]
25. Chen B, Khodadoust MS, Liu CL, et al. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol 2018;1711:243-59. [Crossref] [PubMed]
26. Jiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018;24:1550-8. [Crossref] [PubMed]
27. Maeser D, Gruener RF, Huang RS. oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief Bioinform 2021;22:bbab260. [Crossref] [PubMed]
28. Villanueva A. Hepatocellular Carcinoma. N Engl J Med 2019;380:1450-62. [Crossref] [PubMed]
29. Wang D, Zhang L, Zhang Y, et al. A ferroptosis-associated lncRNAs signature predicts the prognosis of hepatocellular carcinoma. Medicine (Baltimore) 2022;101:e29546. [Crossref] [PubMed]
30. Chen F, Yang J, Fang M, et al. Necroptosis-related lncRNA to establish novel prognostic signature and predict the immunotherapy response in breast cancer. J Clin Lab Anal 2022;36:e24302. [Crossref] [PubMed]
31. Chen H, Yang W, Li Y, et al. Leveraging a disulfidptosis-based signature to improve the survival and drug sensitivity of bladder cancer patients. Front Immunol 2023;14:1198878. [Crossref] [PubMed]
32. Wang T, Guo K, Zhang D, et al. Disulfidptosis classification of hepatocellular carcinoma reveals correlation with clinical prognosis and immune profile. Int Immunopharmacol 2023;120:110368. [Crossref] [PubMed]
33. Su K, Shen Q, Tong J, et al. Construction and validation of a nomogram for HBV-related hepatocellular carcinoma: A large, multicenter study. Ann Hepatol 2023;28:101109. [Crossref] [PubMed]
34. Su K, Wang F, Li X, et al. Effect of external beam radiation therapy versus transcatheter arterial chemoembolization for non-diffuse hepatocellular carcinoma (≥ 5 cm): a multicenter experience over a ten-year period. Front Immunol 2023;14:1265959. [Crossref] [PubMed]
35. Su K, Huang W, Li X, et al. Evaluation of Lactate Dehydrogenase and Alkaline Phosphatase as Predictive Biomarkers in the Prognosis of Hepatocellular Carcinoma and Development of a New Nomogram. J Hepatocell Carcinoma 2023;10:69-79. [Crossref] [PubMed]
36. Chung KY, Cheng IK, Ching AK, et al. Block of proliferation 1 (BOP1) plays an oncogenic role in hepatocellular carcinoma by promoting epithelial-to-mesenchymal transition. Hepatology 2011;54:307-18. [Crossref] [PubMed]
37. Li M, Liu Z, Wang J, et al. Systematic Analysis Identifies a Specific RNA-Binding Protein-Related Gene Model for Prognostication and Risk-Adjustment in HBV-Related Hepatocellular Carcinoma. Front Genet 2021;12:707305. [Crossref] [PubMed]
38. Qi Y, Xu F, Chen L, et al. Quantitative proteomics reveals FLNC as a potential progression marker for the development of hepatocellular carcinoma. Oncotarget 2016;7:68242-52. [Crossref] [PubMed]
39. Zhu Y, Tan JK, Goon JA. Cuproptosis- and m6A-Related lncRNAs for Prognosis of Hepatocellular Carcinoma. Biology (Basel) 2023;12:1101. [Crossref] [PubMed]
40. Gong A, Luo X, Tan Y, et al. High expression of C10orf91 and LINC01224 in hepatocellular carcinoma and poor prognosis. Am J Transl Res 2022;14:2567-79.
41. Wang M, Gu J, Zhang X, et al. Long Non-coding RNA DANCR in Cancer: Roles, Mechanisms, and Implications. Front Cell Dev Biol 2021;9:753706. [Crossref] [PubMed]
42. Gao W, Chen X, Chi W, et al. Long non‑coding RNA MKLN1‑AS aggravates hepatocellular carcinoma progression by functioning as a molecular sponge for miR‑654‑3p, thereby promoting hepatoma‑derived growth factor expression. Int J Mol Med 2020;46:1743-54. [Crossref] [PubMed]
43. Sutherland TE, Dyer DP, Allen JE. The extracellular matrix and the immune system: A mutually dependent relationship. Science 2023;379:eabp8964. [Crossref] [PubMed]
44. Peng H, Chen B, Wei W, et al. N(6)-methyladenosine (m(6)A) in 18S rRNA promotes fatty acid metabolism and oncogenic transformation. Nat Metab 2022;4:1041-54. [Crossref] [PubMed]
45. Yang H, Deng Q, Ni T, et al. Targeted Inhibition of LPL/FABP4/CPT1 fatty acid metabolic axis can effectively prevent the progression of nonalcoholic steatohepatitis to liver cancer. Int J Biol Sci 2021;17:4207-22. [Crossref] [PubMed]
46. Salgado R, Denkert C, Campbell C, et al. Tumor-Infiltrating Lymphocytes and Associations With Pathological Complete Response and Event-Free Survival in HER2-Positive Early-Stage Breast Cancer Treated With Lapatinib and Trastuzumab: A Secondary Analysis of the NeoALTTO Trial. JAMA Oncol 2015;1:448-54. [Crossref] [PubMed]
47. Hwang C, Lee SJ, Lee JH, et al. Stromal tumor-infiltrating lymphocytes evaluated on H&E-stained slides are an independent prognostic factor in epithelial ovarian cancer and ovarian serous carcinoma. Oncol Lett 2019;17:4557-65. [Crossref] [PubMed]
48. Shaban M, Khurram SA, Fraz MM, et al. A Novel Digital Score for Abundance of Tumour Infiltrating Lymphocytes Predicts Disease Free Survival in Oral Squamous Cell Carcinoma. Sci Rep 2019;9:13341. [Crossref] [PubMed]
49. Paijens ST, Vledder A, de Bruyn M, et al. Tumor-infiltrating lymphocytes in the immunotherapy era. Cell Mol Immunol 2021;18:842-59. [Crossref] [PubMed]
50. Künzli M, Masopust D. CD4(+) T cell memory. Nat Immunol 2023;24:903-14. [Crossref] [PubMed]
51. de Visser KE, Joyce JA. The evolving tumor microenvironment: From cancer initiation to metastatic outgrowth. Cancer Cell 2023;41:374-403. [Crossref] [PubMed]
52. Anagnostou V, Bardelli A, Chan TA, et al. The status of tumor mutational burden and immunotherapy. Nat Cancer 2022;3:652-6. [Crossref] [PubMed]
53. Lam YK, Yu J, Huang H, et al. TP53 R249S mutation in hepatic organoids captures the predisposing cancer risk. Hepatology 2023;78:727-40. [Crossref] [PubMed]
54. Jin H, Shi Y, Lv Y, et al. EGFR activation limits the response of liver cancer to lenvatinib. Nature 2021;595:730-4. [Crossref] [PubMed]
55. Wang S, Wu Y, Liu M, et al. DHW-208, A Novel Phosphatidylinositol 3-Kinase (PI3K) Inhibitor, Has Anti-Hepatocellular Carcinoma Activity Through Promoting Apoptosis and Inhibiting Angiogenesis. Front Oncol 2022;12:955729. [Crossref] [PubMed]
56. Xu J, Dong X, Huang DCS, et al. Current Advances and Future Strategies for BCL-2 Inhibitors: Potent Weapons against Cancers. Cancers (Basel) 2023;15:4957. [Crossref] [PubMed]