Integrated analysis identifies disulfidptosis related tumor antigens and molecular subtypes in hepatocellular carcinoma for mRNA vaccine development

Integrated analysis identifies disulfidptosis related tumor antigens and molecular subtypes in hepatocellular carcinoma for mRNA vaccine development | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Integrated analysis identifies disulfidptosis related tumor antigens and molecular subtypes in hepatocellular carcinoma for mRNA vaccine development Renjie Zhang, Qi Liu, Siqi Liu, Meifeng Li, Weixi Shen This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9072846/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 13 You are reading this latest preprint version Abstract Hepatocellular carcinoma (HCC) remains one of the most lethal malignancies worldwide, yet its molecular heterogeneity continues to limit the efficacy of systemic therapies. Disulfidptosis, a recently characterized form of regulated cell death triggered by disulfide stress under glucose deprivation in SLC7A11-overexpressing cells, represents a promising but poorly characterized therapeutic target in HCC. Here, we established an integrated disulfidptosis-based framework encompassing molecular subtyping, mRNA vaccine design, and prognostic modeling. Transcriptomic data from TCGA-LIHC and GEO datasets were integrated after batch-effect correction. Thirty-two disulfidptosis-related genes (DRGs) defined two molecularly distinct HCC subtypes with significant differences in survival outcomes, immune cell infiltration, and pathway enrichment. Through mutation profiling and survival analysis, tumor-associated antigens WASF2 and LRPPRC were identified as optimal vaccine targets, both exhibiting high mutation frequencies, adverse prognostic associations, and positive correlation with antigen-presenting cell abundance. A multi-epitope mRNA vaccine incorporating B-cell, CTL, and HTL epitopes was computationally designed and validated for strong antigenicity, non-allergenicity, non-toxicity, and favorable physicochemical properties; in silico immunosimulation predicted robust humoral and cellular immune responses. Furthermore, a disulfidptosis-related lncRNA prognostic index (DRI) was constructed via LASSO-Cox regression, effectively stratifying patients by prognosis (AUC > 0.7) and predicted responsiveness to immune checkpoint inhibitors. This multi-dimensional framework linking disulfidptosis biology to HCC subtype classification, immunotherapy-oriented vaccine development, and patient stratification offers a compelling foundation for advancing precision immunotherapy in HCC. hepatocellular carcinoma disulfidptosis mRNA vaccine lncRNA prognostic model tumor microenvironment immunotherapy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1. Introduction Liver cancer is one of the most prevalent and lethal malignancies worldwide. According to World Health Organization data, approximately 865,269 new cases and 757,948 deaths attributable to liver cancer are reported annually, rendering it the sixth leading contributor to the global cancer burden. Histopathologically, liver cancer encompasses hepatocellular carcinoma (HCC), which accounts for approximately 75%–85% of cases, and intrahepatic cholangiocarcinoma, which constitutes 10%–15%[1, 2]. Current therapeutic modalities for HCC are diverse and include surgical resection, radiofrequency ablation, chemotherapy, immunotherapy, and targeted therapy, with surgical intervention remaining the primary curative approach for early-stage HCC. The 1-year and 5-year overall survival (OS) rates following surgical treatment can reach up to 89% and 56%, respectivel[3]. However, HCC is frequently asymptomatic in its early stages and exhibits rapid progression; consequently, the majority of patients are diagnosed at advanced stages. For patients with advanced HCC, the survival benefit from surgery is limited, and treatment relies predominantly on immunotherapy and targeted therapy[4]. In a phase III clinical trial for advanced HCC, atezolizumab (an anti-PD-L1 antibody) combined with bevacizumab (an anti-VEGF agent) demonstrated significant efficacy, achieving an overall response rate (ORR) of 29.8% and a 5.8-month survival benefit over sorafenib. Nevertheless, only a minority of patients derive meaningful benefit from immunotherapy; approximately 70% of patients exhibit no appreciable response, and a subset experience severe adverse event[5, 6]. Therefore, there is an urgent need to develop novel combination therapeutic strategies to reactivate the immunosuppressive microenvironment in HCC. Tumor-associated antigens (TAAs) are protein molecules that are aberrantly expressed or specifically overexpressed in malignant tumor cells. TAAs encompass a broad spectrum of proteins, including those involved in dysregulated cell cycle control, oncogene products, cell surface receptors, and embryo-specific proteins that are normally expressed only during fetal development[7] .In the field of oncology, researchers have strategically leveraged TAAs as therapeutic targets for the development of tumor mRNA vaccines. This innovative approach introduces mRNA sequences encoding TAAs to activate and mobilize the host immune system to recognize and attack tumor cells. The advantages of this strategy include the effective induction of antigen-specific immune recognition and clearance of tumor cells, lower cytotoxicity, and improved patient tolerability, thereby enhancing both therapeutic safety and efficacy[8]. For tumors such as HCC that present significant therapeutic challenges, the development of mRNA vaccines is particularly important [9]. Existing therapies for HCC still have notable limitations, including drug resistance. mRNA vaccines offer a potentially low-toxicity, high-efficacy therapeutic alternative that may overcome these constraint [10]. At the level of the HCC microenvironment, mRNA vaccines also exhibit unique advantages: the HCC microenvironment is typically immunosuppressive, rendering conventional therapies suboptimal. mRNA vaccines can activate and enhance tumor-specific T cell activity within this complex microenvironment, demonstrating considerable therapeutic promise[11]. Programmed cell death (PCD) encompasses multiple non-apoptotic death modalities, including necroptosis, ferroptosis, and oxytosis, which have long been implicated in tumorigenesis and tumor progression [12]. Among these, disulfidptosis is a recently identified form of regulated cell death: under conditions of high SLC7A11 expression and glucose deprivation-induced disulfide stress, aberrant intracellular accumulation of disulfide bonds triggers rapid cell death. SLC7A11 is a membrane-localized amino acid transporter responsible for importing extracellular cystine (a disulfide bond-containing amino acid) into the cell[13, 14]. When cells overexpress SLC7A11, cystine uptake is markedly increased. Under normal conditions, cystine is reduced to cysteine and participates in the synthesis of antioxidant molecules such as glutathione, thereby maintaining cellular redox homeostasis. However, under glucose deprivation, cells are unable to generate sufficient NADPH via the pentose phosphate pathway. NADPH is a critical reducing agent that facilitates the reduction of cystine to cysteine. Consequently, excess cystine accumulates intracellularly, forming highly toxic disulfide compounds that induce disulfide bond-mediated cell death [15]. This death modality has been proposed as a novel therapeutic target in cancer. Long non-coding RNAs (lncRNAs) are transcripts exceeding 200 nucleotides in length that do not encode protein [16]. Once dismissed as transcriptional “junk,” lncRNAs have since been demonstrated to play critical roles in tumor drug resistance and other oncogenic processes, including in breast cancer [17]. In liver cancer, oncogenic lncRNAs such as HOTAIR exert profound effects on tumor growth and metastasis [18]. Furthermore, recent studies have revealed that disulfidptosis-related lncRNAs, such as ZEB1-AS1, promote tumor cell proliferation and metastasis when upregulated [19]. However, the precise molecular mechanisms and clinical significance of disulfidptosis in HCC remain insufficiently elucidated. Therefore, leveraging multi-cohort clinical RNA sequencing datasets to characterize the relationship between disulfidptosis subtypes and lncRNAs in HCC holds the potential to uncover complex mechanisms underlying HCC progression and metastasis, identify promising therapeutic targets, and provide a basis for effective HCC intervention. In the present study, we identified WASF2 and LRPPRC as potential tumor antigens for HCC mRNA vaccine development based on disulfidptosis-related genes and designed a novel mRNA vaccine targeting these two antigens. This not only provides a new strategy for HCC immunotherapy but also lays a foundation for future vaccine development. Additionally, we identified two novel disulfidptosis subtypes, offering new perspectives for understanding HCC pathogenesis. Beyond antigen identification and regulatory pattern analysis, we constructed a prognostic risk scoring model based on disulfidptosis-related lncRNAs. This model not only predicts prognosis in HCC patients but also facilitates patient selection for mRNA vaccine administration, thereby providing more precise and individualized therapeutic guidance for clinical practice. 2. Methods 2.1 Data Collection and Processing RNA-Seq data were retrieved from two public databases: The Cancer Genome Atlas (TCGA, https://www.cancer.gov/tcga ) and the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/ ). After filtering, a total of 493 liver cancer patient samples with complete survival information were included. The microarray dataset (GSE76427) was downloaded directly from the GEO database. Additionally, RNA sequencing data and clinical information for TCGA-LIHC were obtained from the Genomic Data Commons. The GEO and TCGA datasets were merged into a unified expression matrix, and batch effects were removed using the “sva” package and the “ComBat” function in R software. Somatic mutation data were downloaded directly from the TCGA database. Thirty-two disulfidptosis-related genes (SLC7A11, INF2, PDLIM1, CD2AP, MYH9, MYH10, ACTN4, FLNA, FLNB, IQGAP1, TLN1, MYL6, ACTB, DSTN, CAPZB, RPN1, NCKAP1, GYS1, NDUFS1, OXSM, LRPPRC, NDUFA11, NUBPL, SLC3A2, AJAP1, ACTR2, ACTR3, WASF2, CYFIP1, ABI2, SLC2A1, and BRK1) were selected based on previous literature [15]. LncRNAs were annotated using the latest annotation file from the Ensembl database ( http://asia.ensembl.org ). 2.2 Tumor Antigen Identification and Kaplan Meier Survival Analysis The mutation status of disulfidptosis-related genes (DRGs) in TCGA-LIHC patients was analyzed and visualized using the “maftools” package in R, with a mutation rate ≥ 1% considered significant for identifying potential tumor antigens. OS and disease-free survival (DFS) were analyzed using the “survival analysis” module in Gene Expression Profiling Interactive Analysis 2nd version (GEPIA2, http://gepia2.cancer-pku.cn/#index ) [20].DRGs exhibiting mutations (≥ 1%) and significant associations with both OS and DFS were selected for tumor antigen construction. The Tumor Immune Estimation Resource was employed to evaluate correlations between selected tumor antigens and immune cell infiltration [21]. 2.3 Immune Cell Infiltration Analysis The relative proportions of tumor-infiltrating immune cells were estimated using the CIBERSORT algorithm. This algorithm employs support vector regression (ν-SVR) to deconvolve 22 human immune cell subtypes from bulk gene expression profiles [22, 23]. The LM22 signature matrix, comprising 547 genes capable of distinguishing 22 leukocyte phenotypes—including 7 T cell subtypes, naïve/memory B cells, plasma cells, NK cells, and myeloid subsets—was used as the reference. Normalized gene expression data served as input. Permutations were set to 100 to assess statistical significance. Only samples with a CIBERSORT P ≤ 0.05 were retained as reliable deconvolution results for subsequent correlation, survival, and differential abundance analyses. 2.4 Identification of Disulfidptosis Related Subtypes Based on the expression profiles of 32 DRGs, unsupervised consensus clustering was performed on the merged TCGA-LIHC and GSE76427 liver cancer samples to identify potential disulfidptosis subtypes. Clustering analysis was implemented using the “ConsensusClusterPlus” package in R [24]. The specific parameters were as follows: clustering method, k-means; distance metric, Pearson correlation; number of resampling iterations, 1000; and cluster number k ranging from 2 to 9. The optimal number of clusters was determined by comprehensively evaluating the consensus matrix, cumulative distribution function (CDF) curves, proportion of ambiguous clustering (PAC), and item-consensus metrics. 2.5 Gene Set Variation Analysis and Tumor Microenvironment Differential Analysis Gene Set Variation Analysis (GSVA) was employed to evaluate pathway activity differences between disulfidptosis subtypes. GSVA is a non-parametric, unsupervised method that transforms a sample-level gene expression matrix into a gene set enrichment score matrix, thereby quantifying inter-sample variation in pathway activity [25]. The canonical KEGG pathway gene set file “c2.cp.kegg.v7.4.symbols.gmt” was downloaded from the MSigDB database ( https://www.gsea-msigdb.org/gsea/msigdb ), and GSVA enrichment analysis was performed using the “GSVA” package in R. The input data consisted of the batch-corrected normalized expression matrix, with default method parameters, to examine differences in KEGG pathway regulation patterns between disulfidptosis subtypes. Subsequently, Wilcoxon rank-sum tests were used to compare GSVA scores between subtypes, and significantly differential pathways were visualized using heatmaps. Furthermore, to assess immune infiltration levels within the tumor microenvironment, single-sample gene set enrichment analysis (ssGSEA) from the “GSVA” package was employed, using established immune cell marker gene sets to calculate immune cell infiltration enrichment scores for each sample. Simultaneously, the “estimate” package was used to compute ESTIMATE scores for each sample, including ImmuneScore, StromalScore, ESTIMATEScore, and TumorPurity, to comprehensively explore tumor microenvironment heterogeneity across subtype [26]. 2.7 Development of mRNA Vaccine Targeting WASF2 and LRPPRC 2.7.1 Retrieval of Tumor Specific Protein Sequences The amino acid sequences of the two target proteins were retrieved from the UniProt database ( https://www.uniprot.org/ ), with accession numbers WASF2 (Q9Y6W5) and LRPPRC (P42704). 2.7.2 B Cell Epitope Prediction Linear B cell epitopes were predicted using the ABCpred web server ( https://webs.iiitd.edu.in/raghava/abcpred/ ), with a window length of 16 amino acids, a threshold of 0.5, and the overlapping filter enabled to reduce redundant overlapping predictions [27]. This server is trained using a recurrent neural network based on linear B cell epitope data from the BCIPEP database. 2.7.3 T Lymphocyte Epitope Prediction T cell epitope prediction was performed using the IEDB Analysis Resource ( http://tools.iedb.org/ ) T cell prediction tools, employing the NetMHCpan 4.1 EL algorithm (for MHC class I-restricted epitopes) and NetMHCIIpan 4.1 EL algorithm (for MHC class II-restricted epitopes) [28, 29]. These methods integrate extensive binding affinity and mass spectrometry-derived eluted ligand data and represent the current IEDB-recommended first-choice prediction tools. Predictions were conducted using the default HLA allele reference set, and candidate epitopes were ranked by predicted percentile rank, with low percentile epitopes prioritized . 2.7.4 Prediction of Epitope Antigenicity and Toxicity Following identification of linear B cell (LBL), cytotoxic T lymphocyte (CTL), and helper T lymphocyte (HTL) epitopes, candidate epitopes underwent immunobiological safety assessment. Antigenicity was evaluated using the VaxiJen server ( http://www.ddg-pharmfac.net/vaxijen/ ) with the tumor model and a threshold of ≥ 0.5 to enhance specificity [30]. This alignment-free tool predicts protective antigen potential of proteins. Allergenicity screening was performed using AllerTOP v2.1 ( https://www.ddg-pharmfac.net/AllerTOP/ ), employing amino acid propensity-based auto-cross covariance transformation and k-nearest neighbor (kNN) classifier, retaining only epitopes predicted as non-allergens [31]. Toxicity prediction was conducted using ToxinPred3 ( https://webs.iiitd.edu.in/raghava/toxinpred3/ ), an SVM-based model for distinguishing toxic from non-toxic sequences [32]. Only epitopes with VaxiJen scores ≥ 0.5, AllerTOP prediction of non-allergen, and absence of toxicity were retained for subsequent multi-epitope construct design and immune simulation analyses. 2.7.5 Design of the Multi Epitope mRNA Vaccine Construct The multi-epitope mRNA vaccine was constructed by integrating the previously identified LBL, CTL, and HTL epitopes. To ensure correct epitope processing, presentation, and immunogenicity, the following linkers were employed: GPGPG (or GPPGG) spacers between HTL epitopes to maintain domain independence and facilitate MHC class II processing; KK linkers between LBL epitopes to promote lysosomal protease cleavage and preserve independent B cell epitope recognition; and AAY linkers between CTL epitopes to promote proteasomal cleavage and generate precise C-termini favorable for TAP transport and MHC class I presentation. The core antigen amino acid sequence was arranged from N-terminus to C-terminus as follows: GPGPG-linked HTL epitope cluster → KK-linked LBL epitopes → AAY-linked CTL epitopes To achieve efficient mRNA expression, the coding region was flanked by the following regulatory elements: the 5′ end comprised an m7G cap analog, an optimized 5′ untranslated region (5′ UTR), and a Kozak sequence (GCCACC AUG G) to enhance ribosome binding and translation initiation; the C-terminus included an MHC class I trafficking domain (MITD) to redirect antigens to the endosomal compartment and enhance cross-presentation efficiency, a UAA stop codon, a stabilized 3′ UTR, and a 120-nt poly(A) tail to improve mRNA stability and translational fidelity. 2.7.6 Prediction of Antigenicity and Physicochemical Properties of the Vaccine Construct Following assembly of the vaccine sequence, multi-platform bioinformatic validation was performed on the core chimeric antigen (excluding the tPA signal peptide and MITD domain) to evaluate its antigenicity, allergenicity, toxicity, and physicochemical stability. Antigenicity was predicted using VaxiJen and ANTIGENpro (SCRATCH suite, https://scratch.proteomics.ics.uci.edu/ ), both employing alignment-free methods based on amino acid sequences to assess protective antigen potential [30, 33]. Allergenicity was assessed using AllerTop v2.1, utilizing auto-cross covariance transformation and kNN classifier with default parameters [31], retaining sequences predicted as non-allergens. Toxicity was evaluated using ToxinPred3.0, an improved model integrating machine learning, deep learning, and motif analysis, with a toxicity probability < 0.5 classified as non-toxic [32]. Physicochemical properties were calculated using the ProtParam tool ( https://web.expasy.org/protparam/ ), including amino acid composition, molecular weight, theoretical pI, instability index, aliphatic index, and GRAVY value (negative values favoring hydrophilicity and solubility) [34].This cross-validation strategy ensured high immunogenicity, low safety risk, and favorable biophysical properties of the vaccine candidate, providing a robust foundation for subsequent structural modeling and immune simulation. 2.7.7 In Silico Immune Simulation To predict the immune response kinetics of the mRNA vaccine construct in the human body, in silico immune simulation was performed using the C-IMMSIM online platform ( https://kraken.iac.rm.cnr.it/C-IMMSIM/ ). This platform is based on the Celada–Seiden cellular automaton model, which simulates humoral and cellular immune processes in the mammalian immune system [35]. The vaccination regimen was set to three doses, administered at time steps 1, 84, and 168, each comprising 1000 antigen units. In the C-IMMSIM model, each time step corresponds to approximately 8 hours of real time; thus, the three vaccinations were spaced approximately at days 0, 28, and 56. The total simulation length was set to 350 steps, covering approximately 117 days of immune dynamics. All parameters other than vaccination timing and dosage were set to server defaults. This computational framework was used to systematically evaluate the coordination of vaccine-induced humoral and cellular immune responses and the potential for long-term immunological memory formation. 2.8 Construction and Validation of the Disulfidptosis Risk Index The TCGA-LIHC expression matrix was partitioned into mRNA and lncRNA groups using the Ensembl annotation file. Co-expression analysis was performed on the lncRNA group based on 32 DRGs retaining lncRNAs with correlation coefficients ≥ 0.4 and P < 0.001 for subsequent analysis. After excluding normal samples from TCGA-LIHC, the disulfidptosis-related lncRNA expression data were merged with patient survival data. Samples were randomly divided into training and testing sets at a 1:1 ratio using the “caret” package in R. The prognostic model was constructed in the training set and validated in the testing set. Univariate Cox regression analysis was first performed with a screening threshold of P < 0.001. A Lasso regression model was then constructed using the “glmnet” package in R [36], and feature lncRNAs were selected via cross-validation. These feature lncRNAs were subsequently used to construct a multivariate Cox regression model. The model formula was as follows: Disulfidptosis Risk Index (DRI) = expression of lncRNA₁ × Coef(lncRNA₁) + expression of lncRNA₂ × Coef(lncRNA₂) + … + expression of lncRNAₙ × Coef(lncRNAₙ) where Coef i represents the regression coefficient. Risk scores were calculated for each sample using this formula, and samples were stratified into high-risk and low-risk groups based on the median risk score from the training set. Model validity and stability were verified using ROC curves, concordance index(C-index), and independent prognostic analyses. 2.9 Construction and Validation of the Nomogram Based on the DRI A prognostic nomogram was constructed by integrating the DRI with clinicopathological features to enable individualized survival probability prediction for HCC patients. The nomogram was generated using the “rms” package in R. First, univariate and multivariate Cox proportional hazards regression analyses were performed to confirm whether DRI and clinical variables (age, sex, TNM stage, and tumor grade) were independent prognostic factors. These independent prognostic variables were incorporated into the lrm or cph functions of the “rms” package to generate the nomogram model. The nomogram quantified the contribution of each variable as a point score, and total points were mapped to predicted probabilities of 1-year, 3-year, and 5-year OS. 2.10 Immunotherapy Response Prediction To evaluate potential differential responses to immune checkpoint inhibitors (ICIs) between DRI risk groups, pre-computed immunotherapy response scores (IPS) were obtained from The Cancer Immunome Atlas (TCIA, https://tcia.at/ ). IPS is a computational metric based on tumor transcriptomic data used to predict patient sensitivity to various ICI regimens [37]. Specifically, IPS scores for the following four combinations were included:ips_ctla4_pos_pd1_pos,ips_ctla4_neg_pd1_pos,ips_ctla4_pos_pd1_neg, and ips_ctla4_neg_pd1_neg. Higher IPS scores indicate stronger tumor immunogenicity and better predicted ICI response. Differences in these four IPS scores between high- and low-risk groups were compared using the Wilcoxon rank-sum test. Furthermore, the Tumor Immune Dysfunction and Exclusion (TIDE) score was employed to further predict immunotherapy response. TIDE is a transcriptome-based computational framework that models the two major mechanisms of tumor immune evasion, T cell dysfunction and T cell exclusion, to predict the probability of patient response to ICIs [38]. 2.11 Statistical Analysis All statistical analyses were performed using R software (version 4.3.2). For inter-group comparisons of continuous variables, independent-sample t-tests were used when data were normally distributed with equal variances; otherwise, the non-parametric Wilcoxon rank-sum test was applied. Categorical variables were compared using the chi-square (χ²) test. Survival analysis was conducted using the Kaplan–Meier method, and differences in OS between groups were evaluated using the log-rank test. Multiple comparison correction was performed using the Benjamini–Hochberg false discovery rate method. All statistical tests were two-sided, and P < 0.05 was considered statistically significant. All figures were generated using the “ggplot2,” “ggpubr,” “survminer,” and “pheatmap” packages, with statistical significance annotated as P < 0.05, P < 0.01, ** P < 0.001, and ns: not significant. 3. Results 3.1 Development of mRNA Vaccines Based on Disulfidptosis Related Genes Using the survival analysis module in GEPIA2, we evaluated the association of each DRG's expression levels with OS and disease-free survival DFS in HCC patients.Concurrently, the mutation status of these genes was analyzed, and genes with mutation rates ≥ 1% and significant prognostic associations were selected (Figs. 1 A, 1 B). Ultimately, integrating prognostic value and potential immunogenicity, LRPPRC and WASF2 were identified as candidate tumor antigens for HCC mRNA vaccine development. Kaplan–Meier analysis demonstrated that patients with high expression of LRPPRC and WASF2 exhibited significantly shorter OS and DFS (Figs. 1 C– 1 F, P < 0.05). Antigen-presenting cells (APCs) are critical for activating tumor-specific T cell responses, among which dendritic cells (DCs) are the most potent, efficiently presenting tumor antigens to T cells; macrophages and B cells can also process and present antigens under certain conditions. As shown in Figs. 1 G– 1 H, the expression levels of LRPPRC and WASF2 were significantly positively correlated with APC infiltration ( P < 0.05). These results suggest that LRPPRC and WASF2 may serve as effective shared tumor antigens with potential for HCC mRNA vaccine development. 3.2 DRG Mutation Characteristics and Immune Microenvironment Analysis By integrating transcriptomic data downloaded from the TCGA and GEO databases, we constructed an expression matrix containing survival information. We compared the expression of 32 DRGs between normal and tumor samples. Results revealed that 28 genes exhibited significant differential expression between tumor and normal tissues (Fig. 2 A). Subsequently, a DRG network was constructed to elucidate the interrelationships and prognostic significance of DRGs in HCC. The analysis demonstrated statistically significant correlations among DRGs. The network indicated that, with the exception of NDUFA11 and NUBPL, all remaining genes were risk factors, with statistical evidence supporting significant positive regulatory relationships among them (Fig. 2 B). To investigate whether genetic alterations affect DRG expression levels in HCC, copy number variations (CNVs) were analyzed. Results indicated that CNVs are a significant contributor to DRG expression dysregulation. Most DRGs with copy number gains were significantly upregulated in HCC. All 32 DRGs exhibited CNVs, with copy number gains being more prevalent, while CYFIP1, AJAP1, NDUFA11, WASF2, INF2, MYH10, and CAPZB showed higher frequencies of copy number losses (Fig. 2 C). Figure 2 D illustrates the chromosomal CNV distribution of these DRGs. These findings suggest that dysregulated DRG expression plays a critical role in HCC development and progression, and our study confirms significant differences in the genetic characteristics and expression levels of DRGs between normal and HCC tissues. Furthermore, we conducted an in-depth analysis of the potential tumor antigen LRPPRC. Results demonstrated that the LRPPRC-high expression group exhibited a higher proportion of M0 macrophage infiltration, whereas CD8 + T cell infiltration was significantly increased in the low-expression group (Fig. 2 F). CD8 + T cells are the core effector cells in tumor immunotherapy, and their decreased infiltration is frequently associated with poor immunotherapy response. We also observed that expression of major histocompatibility complex (MHC) molecules, co-stimulatory molecules, and cell adhesion molecules was generally upregulated in the LRPPRC-low expression group (Fig. 2 E). These findings further support the feasibility and validity of LRPPRC as a potential tumor-associated antigen in HCC. 3.3 Identification of Two Disulfidptosis Subtypes Based on 32 DRGs Based on the expression profiles of 32 DRGs, unsupervised consensus clustering analysis was performed, successfully identifying two significantly distinct disulfidptosis molecular subtypes, designated DRGcluster A and DRGcluster B. The consensus matrix exhibited a clear block-diagonal structure, and the CDF curves and delta area curves reached plateaus, indicating stable and reliable clustering results (Figs. 3 A– 3 D). To validate the effectiveness and biological significance of subtype discrimination, principal component analysis (PCA) was performed on the DRG expression matrix. Results showed that samples from the two subtypes were clearly separated in the first two principal component space, with minimal overlap (Fig. 3 E). This confirmed that DRG-based clustering genuinely reflects heterogeneity in disulfidptosis regulatory patterns among HCC patients and captures subtype-specific gene expression signatures. Kaplan–Meier survival analysis revealed that DRGcluster B patients exhibited significantly better overall survival compared to DRGcluster A (Fig. 3 F), indicating a close association between disulfidptosis molecular subtypes and HCC prognosis. Further comparison of immune cell infiltration differences between the two subtypes revealed that DRGcluster A exhibited higher infiltration levels of activated CD4 + T cells, activated dendritic cells, follicular helper T cells, and Th2 helper T cells, whereas DRGcluster B showed higher infiltration of activated B cells, activated CD8 + T cells, eosinophils, macrophages, monocytes, natural killer cells, Th1 helper T cells, and neutrophils (Fig. 3 G). These differences reflect heterogeneity in the tumor immune microenvironment between the two subtypes. A heatmap displaying the expression distribution of 32 DRGs across the cohort, overlaid with clinical parameters, demonstrated that DRGcluster A exhibited higher expression levels for most DRGs, while DRGcluster B showed relatively lower expression (Fig. 3 I). GSVA comparing pathway enrichment differences between the two subtypes revealed that DRGcluster B was predominantly enriched in endocytosis, antidiuretic hormone-regulated water reabsorption, cell cycle, RNA degradation, ubiquitin-mediated proteolysis, and basal transcription factor pathways; DRGcluster A was primarily enriched in nitrogen metabolism, amino acid metabolism, retinol metabolism, and lipid metabolism pathways (Fig. 3 H). In summary, this study identified two distinct disulfidptosis regulatory patterns based on DRGs, which exhibit significant differences in molecular characteristics, patient prognosis, and tumor immune microenvironment infiltration. These findings contribute to elucidating the molecular heterogeneity of HCC and provide potential evidence for developing individualized therapeutic strategies targeting the disulfidptosis pathway. 3.5 Epitope Prediction and Screening Linear B cell epitopes for WASF2 and LRPPRC protein sequences were predicted using the ABCPred server, initially retaining epitopes with scores > 0.85. Candidate epitopes were then subjected to a rigorous three-step screening process: (1) VaxiJen antigenicity assessment (threshold ≥ 0.5; tumor antigen model); (2) AllerTop v2.1 allergenicity screening (non-allergen); and (3) ToxinPred toxicity analysis (non-toxic). This process yielded 11 high-antigenicity, non-allergenic, non-toxic B cell epitopes. For CTL epitopes, the IEDB MHC-I binding prediction tool was employed, retaining high-affinity candidates with percentile rank ≤ 0.01. The same three-step validation was applied, ultimately yielding 6 high-affinity CTL epitopes for vaccine construction to achieve broad HLA coverage and cross-variant immunogenicity. HTL epitopes were comprehensively predicted using the IEDB MHC-II tool and NetMHCIIpan 4.1, retaining high-affinity candidates with percentile rank ≤ 0.2. Following the same three-step screening, 4 high-quality HTL epitopes were prioritized for vaccine inclusion (Supplementary Table S1). To optimize the final multi-epitope vaccine antigen construct, the top two candidate epitopes (encompassing B cell, CTL, and HTL epitopes) from each target protein (WASF2 and LRPPRC) were selected based on VaxiJen antigenicity scores, ensuring maximal immunogenicity and design compactness. 3.6 Design of the Multi Epitope mRNA Vaccine Construct Based on HCC Epitopes The mRNA vaccine construct was designed with the following arrangement from N-terminus to C-terminus: 5′ m7G Cap—5′ UTR—Kozak sequence—Signal peptide (tPA)—GPGPG Linker—ALSFFHMLNGAALRG—GPGPG Linker—AYDIFLNAKEQNIVF—GPGPG Linker—FASRVSSLAERVDRL—GPGPG Linker—KEKMLQDTKDIMKEK—KK Linker—AGIEPGPDTYLALLNA—KK Linker—HYFWPLLVGRRKEKNV—KK Linker—RGSGLAGPKRSSVVSP—KK Linker—VEEQREQEKRDVVGND—AAY Linker—YVSEILEKV—AAY Linker—IPREKTLRL—AAY Linker—DVATILSRR—AAY Linker—RQLGSLSKY—MITD sequence—Stop codon—3′ UTR—Poly(A) tail. 3.7 In Vitro Immunogenicity Assessment of the Multi Epitope Tumor mRNA Vaccine Construct To evaluate the potential immunogenicity of the designed multi-epitope vaccine construct, agent-based immune simulation was performed using the C-IMMSIM online server. The simulation modeled a three-dose vaccination regimen, with injections at time steps 1, 84, and 168 (corresponding to approximately day 1, day 28, and day 56 in real time, with each step representing 8 hours). Key parameters included a simulation volume of 10, total steps of 350, an antigen dose of 1000 particles per injection, and an adjuvant level of 100. The simulation demonstrated effective antigen clearance and robust immune activation (Figs. 4 A– 4 I). Antigen levels (Fig. 4 A) declined rapidly after each injection, approaching zero by days 20–30, accompanied by transient immune complexes. Immunoglobulin production showed a primary IgM response peaking at approximately 10⁴ µg/mL around day 10, followed by class-switching to sustained IgG1 and IgG2 levels (~ 10⁵ µg/mL by day 40), indicative of class switching and long-term humoral memory. Cell population dynamics (Figs. 4 B and 4 H) revealed initial peaks of active and proliferating B cells, plasma cells, T helper cells, and macrophages at approximately 5–20 days post-first injection, with secondary peaks at approximately day 30 following the second injection. The internalization and MHC class II presentation states of DCs and macrophages increased rapidly post-injection, supporting antigen processing and T cell priming; resting states stabilized in later phases, with mild increases in anergic cells suggestive of orderly regulation without excessive tolerance. The cytokine profile (Fig. 4 I) showed early pro-inflammatory peaks (IFN-γ, TNF-α, IL-2 ~ 10⁴–10⁵ pg/mL, days 5–20), with secondary elevations at approximately day 30 following the booster dose. Anti-inflammatory IL-10 subsequently increased to facilitate resolution. Inset analysis confirmed the synergistic action of danger signals and IL-2, supporting leukocyte proliferation without cytokine storm. Overall, the simulation predicted a protective immune profile characterized by robust humoral responses and immunological memory. 3.8 Evaluation of Antigenicity and Physicochemical Properties of the Vaccine Construct Comprehensive in vitro prediction and immune simulation were performed on the vaccine construct. First, physicochemical properties were analyzed using the ExPASy ProtParam tool. Results revealed a total of 200 amino acids with a molecular weight of approximately 21.90 kDa, a molecular formula of C₉₈₂H₁₅₈₂N₂₈₂O₂₇₉S₃, and a theoretical isoelectric point of 9.93. The construct contained 21 negatively charged residues and 36 positively charged residues. The instability index was 45.63, the aliphatic index was 79.55, the GRAVY value was − 0.536, and the estimated in vitro half-life was 30 hours. These parameters indicate favorable solubility and potential for cellular expression, although the slightly elevated instability index suggests that further sequence optimization may be warranted (Table 1 ). Antigenicity predictions using the VaxiJen and ANTIGENpro servers yielded scores of 0.6686 and 0.710346, respectively, indicating that the vaccine construct possesses favorable potential immunogenicity. Allergenicity prediction classified the construct as non-allergenic, and toxicity prediction indicated non-toxicity. These safety assessments support the feasibility of this construct as a candidate tumor mRNA vaccine (Table 1 ). Table 1 Antigenic, allergenic, toxic, and physicochemical assessments of the protein translated from the mRNA vaccine-encoded peptide. Property Measurement Total number of amino acids 200 Molecular weight 21899.27 Formula C982H1582N282O279S3 Theoretical pI 9.93 Total number of negatively charged residues (Asp + Glu) 21 Total number of positively charged residues (Arg + Lys) 36 Total number of atoms 3128 Instability Index (II) 45.63 Aliphatic Index (A.I) 79.55 Grand average of hydropathicity (GRAVY) -0.536 estimated half-life 30 hours (mammalian reticulocytes, in vitro) Antigenicity (using VaxiJen) 0.6686 Antigenicity (using ANTIGENpro) 0.710346 Allergenicity (using AllerTop 2.0) Non-allergenic Toxicity (ToxinPred) Non-toxic 3.9 Construction of a Disulfidptosis Related LncRNA Based Risk Model and Its Diagnostic and Prognostic Value Using the TCGA HCC dataset, we performed co-expression analysis based on previously reported DRGs and identified a total of 1,223 lncRNAs significantly correlated with DRGs (Fig. 5 A). Prognostic risk-associated lncRNAs were screened via univariate Cox regression analysis (Fig. 5 E). Lasso regression combined with cross-validation was subsequently employed to identify the 6 most predictive lncRNAs. Multivariate Cox regression was then performed, ultimately incorporating 4 independently prognostic lncRNAs into the risk scoring model (Disulfidptosis Risk Index, DRI) (Figs. 5 C, 5 D). The DRI was calculated as follows: DRI = (MIR210HG × 0.2977) + (AL031985.3 × 0.3079) + (AC108752.1 × 0.3436) + (AC016717.2 × 0.5874) Correlation analysis revealed significant positive and negative correlations between these 4 lncRNAs and DRGs, suggesting their potential involvement in the disulfidptosis regulatory network (Fig. 5 F). To validate model robustness, the cohort was randomly divided into training and validation sets. The DRI was calculated for each sample based on lncRNA expression levels and regression coefficients, and patients were stratified into high-risk and low-risk groups using the median DRI as the threshold. Kaplan–Meier survival analysis demonstrated significantly worse overall survival in the high-risk group compared to the low-risk group (Figs. 6 J– 6 L); progression-free survival analysis also confirmed longer PFS in the low-risk group (Fig. 5 B). Heatmaps revealed significant expression differences of the 4 lncRNAs between high- and low-risk groups (Figs. 6 A– 6 C). Risk curves and scatter plots further indicated that increasing DRI was associated with more death events and shorter survival times (Figs. 6 D– 6 I). PCA demonstrated that samples could be clearly separated into two distinct clusters based on these 4 risk lncRNAs, supporting the discriminative capacity of the model (Figs. 6 H– 6 K). Time-dependent ROC curve analysis revealed high AUC values for predicting 1-year, 3-year, and 5-year survival rates, indicating favorable predictive performance (Figs. 6 M– 6 O). Compared with clinical indicators including age, sex, TNM stage, and grade, the model exhibited higher ROC curve AUC values (Fig. 7 A); C-index analysis also demonstrated superiority over individual clinical features (Fig. 7 B). Univariate and multivariate Cox regression analyses confirmed that TNM stage ( P < 0.001) and DRI ( P < 0.001) were independent risk factors for HCC prognosis (Figs. 7 D, 7 E). Based on these findings, a nomogram integrating clinical features and DRI was constructed, achieving a C-index of 0.711, indicating satisfactory concordance and clinical predictive value (Fig. 7 C). In summary, the DRI model based on disulfidptosis-related lncRNAs demonstrated favorable prognostic predictive ability in the TCGA cohort, effectively stratifying HCC patient risk and outperforming conventional clinical indicators. This model provides a reference for individualized prognostic assessment and exploration of potential therapeutic targets in HCC, and the role of lncRNAs in disulfidptosis regulation warrants further functional validation. 3.10 DRI and Tumor Immune Microenvironment and Immunotherapy Response Prediction The tumor immune microenvironment (TME) status is a critical determinant of immunotherapy efficacy. This section focuses on the role of the DRI in the HCC TME and its predictive value for potential immunotherapy response.ESTIMATE algorithm analysis revealed that the low-DRI group exhibited significantly higher ESTIMATEScore, ImmuneScore, and StromalScore compared to the high-DRI group (Fig. 8 D), suggesting that the low-risk group TME is characterized by greater immune infiltration and stromal abundance. Heatmap visualization of immune cell infiltration demonstrated that the low-risk group exhibited higher proportions of B cells, dendritic cells, macrophages, neutrophils, natural killer cells, helper T cells, and tumor-infiltrating lymphocytes, while the high-risk group showed correspondingly lower levels (Figs. 8 A, 8 B). These differences suggest that the low-DRI group may possess more “hot” tumor characteristics. Further comparison of immune checkpoint molecule expression revealed that the low-risk group exhibited higher expression of TMIGD2 and IDO2, whereas the high-risk group showed significant upregulation of multiple checkpoint molecules, including HLA2, TNFRSF4, CD276, TNFRSF14, TNFSF4, CD274, TNFSF9, VTCN1, and TNFSF15 (Fig. 10C). The elevated expression of immunosuppressive checkpoints (such as CD274/PD-L1 and CD276/B7-H3) in the high-risk group suggests the presence of stronger immune evasion mechanisms, and DRI may serve as a reference indicator for evaluating potential benefit from ICIs. Immunotherapy response prediction based on the TIDE algorithm demonstrated that the low-DRI group exhibited lower TIDE scores, T cell exclusion scores, myeloid-derived suppressor cell scores, and interferon-γ -related scores compared to the high-DRI group (Figs. 8 K– 8 N; Figs. 8 E– 8 F). These findings indicate that the low-risk group has lower degrees of immune dysfunction and exclusion, suggesting better potential IPS analysis further confirmed that the low-DRI group had significantly higher IPS than the high-DRI group, suggesting the potential for superior clinical benefit following ICI treatment (Figs. 8 G– 8 I). In summary, the DRI model effectively reflects the immune microenvironment heterogeneity of HCC: the low-risk group demonstrates more active immune infiltration and lower immunosuppressive characteristics, with potential value in predicting immunotherapy response. This finding provides a bioinformatic basis for individualized immunotherapy decision-making in HCC patients and lays a foundation for subsequent immunotherapy strategies integrating disulfidptosis regulation. Through DRI subtype stratification, the response of HCC patients to immunotherapy can be more accurately predicted, thereby facilitating the selection of individuals suitable for mRNA immunotherapy. In conclusion, the DRI model effectively identifies the immune microenvironment heterogeneity of HCC: the low-DRI group exhibits more active immune infiltration and lower immunosuppressive features, with value in predicting general immunotherapy response. Given that the present study designed a multi-epitope mRNA vaccine construct based on the disulfidptosis-related targets LRPPRC and WASF2, DRI may serve as a potential biomarker to assist in selecting immunologically “hot” patient subgroups for prioritized mRNA vaccine or combination immunotherapy strategies. These findings provide a bioinformatic basis for integrating disulfidptosis regulation with personalized mRNA vaccines. 3.11 Association Analysis Between DRI and the Somatic Mutation Landscape of HCC To explore the relationship between the DRI and tumor genomic features, we analyzed the somatic mutation profiles of high- and low-DRI groups in the TCGA-HCC cohort using the maftools package in R. Waterfall plots revealed that the tumor mutation burden (TMB) was overall higher in the high-DRI group than in the low-DRI group (Figs. 9 A, 9 B). Specifically, TP53 was the most frequently mutated gene in the high-DRI group (~ 38%), while CTNNB1 was more frequently mutated in the low-DRI group (~ 25%). Further analysis demonstrated that TMB was significantly higher in the high-DRI group compared to the low-DRI group, indicating an association between high DRI and elevated somatic mutation burden (Fig. 9 C). Survival analysis revealed that among low-DRI patients, those with high TMB derived greater survival benefit; in the overall cohort, patients with low TMB exhibited relatively better prognosis (Fig. 9 D). Combined DRI and TMB stratification analysis revealed significant survival differences among four groups (Fig. 9 E): the low-TMB + low-DRI group had the best prognosis, the high-TMB + high-DRI group had the worst prognosis, and the high-TMB + low-DRI and low-TMB + high-DRI groups exhibited intermediate outcomes. These results suggest that the high TMB in the high-DRI group may be accompanied by increased neoantigen production, but combined with its immunosuppressive microenvironment, the overall immunotherapy response remains poor; conversely, although the low-DRI group has lower TMB and fewer neoantigens, its immunologically “hot” phenotype may be more conducive to effector cell activation and antigen presentation. In summary, combined analysis of DRI and TMB more precisely reflects the genomic-immune heterogeneity of HCC. The high-TMB + low-DRI subgroup may have advantages in generating tumor-specific antigens, while the immunologically active microenvironment of the low-DRI group provides potential synergistic effects for mRNA vaccines. Our results suggest that the combined application of DRI and TMB may improve the accuracy of HCC immunotherapy efficacy prediction and provide a potential reference for individualized therapeutic decisions. 4. Discussion HCC is one of the most prevalent and lethal malignancies worldwide, responsible for hundreds of thousands of deaths annually. Because early-stage liver cancer is typically asymptomatic, the majority of patients are diagnosed at advanced stages. The principal therapeutic modalities for advanced HCC consist of targeted therapy and immunotherapy. However, only a minority of patients benefit from these treatments, and some experience severe adverse events. Consequently, there is an urgent need to develop novel therapeutic approaches to improve clinical outcomes for HCC patients. In recent years, mRNA vaccines have garnered extensive attention as a novel class of therapeutic vaccines in oncology. The mechanism of action involves the design and synthesis of mRNA molecules encoding specific tumor antigens. Upon cellular uptake, these mRNA molecules direct the intracellular protein synthesis machinery to produce the specified tumor antigen proteins. These tumor antigens are recognized by the host immune system as foreign entities, thereby activating immune responses against cancer cells. Compared with conventional immunotherapy and targeted therapy, mRNA vaccines offer advantages including high efficacy, favorable safety profiles, and low toxicity [10, 39]. Regulated necrosis differs from traditional necrosis caused by external injury in that it is precisely controlled by intracellular signaling pathways and exhibits programmatic features. This form of cell death involves multiple intracellular signaling pathways and signaling molecules and has long been implicated in tumorigenesis and tumor progression. In liver cancer, RIPK3, a key regulator of necroptosis, has been shown to promote HCC progression through activation of innate immunity [40, 41].Zhou et al. employed bioinformatic approaches to identify tumor antigens, including CARD8, NAIP, NLRP1, and NLRP3, based on pyroptosis-related genes, demonstrating the feasibility of developing mRNA vaccines from regulated necrosis-related genes[42]. Disulfidptosis, as a newly identified form of regulated necrosis, has been proposed as a novel therapeutic target in cancer. Based on transcriptomic data, this study screened DRGs significantly associated with HCC prognosis and identified LRPPRC and WASF2 as potential immunotherapy targets. Previous studies have also confirmed the critical role of LRPPRC in tumor development. liu et al. found that LRPPRC functions as a scaffold protein binding JAK2 and STAT3, enhancing JAK2–STAT3 complex stability and thereby promoting JAK2/STAT3/MYC axis activation and esophageal squamous cell carcinoma progression. Disruption of the LRPPRC–JAK2–STAT3 and JAK2–STAT3–CDK1 interactions suppressed tumorigenesis in 4-nitroquinoline N-oxide-induced ESCC mouse models and inhibited tumor growth in patient-derived xenograft models [43]. Similarly, Yu et al. confirmed through Western blot and RT-qPCR that LRPPRC overexpression in triple-negative breast cancer enhances glycolysis and promotes tumor progression [44]. WASF2 has also been shown to play important roles in the development and progression of multiple cancers. In ovarian cancer, high WASF2 expression is closely associated with poor patient survival, with WASF2 promoting tumor cell invasion and migration, thereby exacerbating ovarian cancer malignancy [45]. In liver cancer, WASF2 overexpression promotes normal hepatocyte proliferation; inactivation of WASF2 by inducing G2/M phase arrest reduces the viability, growth, proliferation, migration, and invasion of Huh-7 and SNU475 liver cancer cells . In the present study, we employed an immunoinformatics-driven reverse vaccinology strategy to design a novel multi-epitope mRNA vaccine targeting the HCC-associated antigens WASF2 and LRPPRC. Through rigorous epitope prediction and multi-step screening, high-antigenicity B cell, CTL, and HTL epitopes were identified, and the top two candidates from each target protein were selected based on antigenicity scores for compact assembly. The final vaccine construct integrates the tPA signal peptide, GPGPG/KK/AAY linker peptides, and the MITD sequence. C-IMMSIM agent-based immune simulation predicted robust humoral and cellular responses, accompanied by a balanced pro-inflammatory and regulatory cytokine profile without cytokine storm. Physicochemical analysis confirmed favorable physicochemical properties with no allergenicity or toxicity. The incorporation of CTL, HTL, and B cell epitopes, combined with MITD optimization, is designed to overcome the immunosuppressive tumor microenvironment (TME) of HCC, including T cell exhaustion, low CD8 + infiltration, and Treg/M2 macrophage dominance. C-IMMSIM demonstrated efficient post-booster antigen clearance, IgG persistence, and secondary immune cell activation peaks, suggesting the induction of long-term memory and protective immunity. The cytokine profile is consistent with an effective Th1-biased response, potentially capable of “heating” immunologically cold tumors, which is particularly critical for HCC. Compared with prior HCC multi-epitope designs targeting AFP, GPC3, or NY-ESO-1, our construct uniquely focuses on WASF2 and LRPPRC, potentially providing novel epitopes less susceptible to immune escape [10]. Rigorous screening and compact selection minimize off-target risks while maximizing immunogenicity, distinguishing it from broad-spectrum epitope constructs. Given that mRNA vaccines demonstrate efficacy only in specific cancer patient subpopulations, we stratified HCC patients into high-risk and low-risk subgroups based on disulfidptosis-related lncRNA expression profiles to identify the most promising vaccine candidate populations. Furthermore, because DRG markers encompass multiple genes and the tumor microenvironment is highly heterogeneous, their utility in predicting HCC prognosis has certain limitations. Accordingly, we established a DRI-based prognostic model. Among the prognostically relevant DRGs we identified, accumulating evidence indicates that the lncRNAs used to construct the DRI risk model play critical roles in cancer. For example, MIR210HG is associated with disease progression in multiple cancers. In endometrial cancer, it is enriched in the Wnt and β/Smad3 signaling pathways and promotes cancer development [46]; in breast cancer, MIR210HG induces cancer cell migration and invasion by regulating the epithelial–mesenchymal transition process [47]. AL031985.3, AC108752.1, and AC016717.2 have also been demonstrated to exhibit significant differential expression between HCC and normal tissues and are significantly associated with patient prognosis [48, 49]. Our bioinformatic analyses similarly identified these lncRNAs as disulfidptosis-related prognostic markers. These findings deepen our understanding of the disulfidptosis-related tumor microenvironment in HCC and support the practical utility of the DRI model. To validate model performance, Kaplan–Meier survival analysis and ROC curve analysis were performed in the study cohort. Results demonstrated that this signature possesses robust capability in predicting HCC prognosis. Furthermore, significant correlations were observed between DRI risk scores and immune cell infiltration as well as immunotherapy efficacy. The DRI risk score serves as an effective tool for characterizing the heterogeneity of tumor disulfidptosis regulatory patterns and identifying TME infiltration features. Comprehensive analysis confirmed that the risk score is an effective predictor of liver cancer prognosis. Results across the overall, training, and validation cohorts consistently demonstrated that patients in the low-risk group exhibited longer overall survival. ROC curve analysis indicated that this lncRNA-based signature model possesses high sensitivity and specificity, demonstrating excellent predictive performance. To enhance clinical applicability, a nomogram integrating DRI scores with pathological stage, clinical stage, age, and sex was constructed. Combined ROC analysis showed that this integrated model outperformed the DRI scoring system alone. Consequently, this nomogram can more accurately predict 1-year, 2-year, and 3-year survival rates in HCC patients, with a favorable C-index. Subsequently, we analyzed immune cell infiltration differences between high- and low-risk DRI groups. The low-risk group exhibited significantly higher immune cell types and quantities compared to the high-risk group, along with better survival rates, further supporting the stronger immune profile of the low-risk group. Tumor responsiveness to immune checkpoint therapy has been demonstrated to be closely associated with the degree of immune cell infiltration and tumor mutation burden. We investigated the relationship between DRI and immune checkpoint signals, and results indicated that disulfidptosis-related lncRNA expression can influence the efficacy of immune checkpoint therapy. Finally, our results revealed a significant correlation between DRI and TMB: the low-DRI risk group exhibited lower TMB levels and higher immune cell infiltration, resulting in superior survival prognosis compared to the high-DRI group. It is generally believed that tumors with high mutation burden have better prognosis with immunogenic cell death (ICD) therapy. However, if ICD efficacy is limited by the immunosuppressive microenvironment or stromal barriers in HCC, combining immunotherapy with tumor neoantigen mining to develop tumor antigen vaccines may help improve the efficacy of existing treatments in the DRI high-risk population. For example, BNT111, an mRNA vaccine targeting four common TAAs (NY-ESO-1, MAGE-A3, tyrosinase, and TPTE), induced new and enhanced pre-existing immune responses in over 90% of melanoma patients in a phase I clinical trial (NCT02410733) [50]. Developing tumor antigen vaccines in combination with existing cancer treatment modalities and DRI-based stratification strategies may represent a promising approach for achieving synergistic immune stimulation and significant clinical benefit. This study has several limitations. First, the feasibility and efficacy of the antigen vaccine have not been experimentally validated; due to the lack of RNA sequencing data and clinical data from HCC immunotherapy cohorts, further follow-up studies are needed to verify its potential value. Additionally, the model could be further optimized using machine learning and feature selection algorithms to improve liver cancer vaccine development and patient prognosis prediction. Although the in vitro simulation results are encouraging, this study is limited to computational predictions and lacks experimental validation (e.g., in vitro T cell ELISpot assays, in vivo HCC mouse models, or human PBMC stimulation). The slightly elevated instability index (45.63) suggests the need for further codon optimization or stabilizing mutations. Future work should include wet laboratory validation of epitope–MHC binding (tetramer staining), LNP-mRNA expression in DCs, immunogenicity in HLA-transgenic mice, and synergy testing with ICIs. Clinical translation will require GMP manufacturing and phase I safety trials in patients with advanced HCC. 4.1 Conclusion This study identified the disulfidptosis-related genes LRPPRC and WASF2 as potential tumor antigens for HCC mRNA vaccine development. Based on these two tumor antigens, a novel HCC mRNA vaccine was designed. Furthermore, we identified two disulfidptosis-related HCC subtypes and analyzed the differences in enriched pathways between the two disulfidptosis subtypes. We also constructed a novel disulfidptosis-related lncRNA-based scoring model that can be used to assess immunotherapy response and identify patients most suitable for vaccine treatment. Declarations Availability of Data and Materials The datasets analyzed during the current study are publicly available. The transcriptomic and clinical data were obtained from The Cancer Genome Atlas (TCGA) repository (https://portal.gdc.cancer.gov/, TCGA-LIHC project) and the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/, accession number GSE76427). Protein sequence information for WASF2 (accession number Q9Y6W5) and LRPPRC (accession number P42704) was retrieved from the UniProt database (https://www.uniprot.org/). The KEGG pathway gene set (c2.cp.kegg.v7.4.symbols.gmt) was downloaded from the Molecular Signatures Database (MSigDB) repository (https://www.gsea-msigdb.org/gsea/msigdb). Immune simulations were performed using the C-IMMSIM platform (https://kraken.iac.rm.cnr.it/C-IMMSIM/). All data used in this study are publicly accessible from the above repositories. Author Contributions RZ and QL designed the research study. RZ and SL performed the research and analyzed the data. ML and WS provided help and advice on methodology and data interpretation. RZ wrote the manuscript. QL, SL, ML, and WS contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. Ethics Approval and Consent to Participate Not applicable. Acknowledgment Not Applicable. Funding Not Applicable. Conflict of Interest The authors declare no conflict of interest. Consent to publish Not applicable. Declaration of AI and AI-assisted Technologies in the Writing Process During the preparation of this manuscript, the authors used Claude Opus 4.6 for the purpose of translating the original Chinese manuscript into English and refining the language to meet academic publishing standards. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. References Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024; 74: 229-263. https://doi.org/10.3322/caac.21834 N A, J A, A S, NN N, N A. Hepatocellular Carcinoma: A Comprehensive Review. Diseases (Basel, Switzerland). 2025; 13. https://doi.org/10.3390/diseases13070207 Reveron-Thornton RF, Teng MLP, Lee EY, Tran A, Vajanaphanich S, Tan EX, et al. Global and regional long-term survival following resection for HCC in the recent decade: A meta-analysis of 110 studies. Hepatol Commun. 2022; 6: 1813-1826. https://doi.org/10.1002/hep4.1923 AG S, F K, JM L. Global trends in hepatocellular carcinoma epidemiology: implications for screening, prevention and therapy. Nature reviews. Clinical oncology. 2023; 20: 864-884. https://doi.org/10.1038/s41571-023-00825-3 RS F, S Q, M I, PR G, M D, TY K, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. The New England Journal of Medicine. 2020; 382: 1894-1905. https://doi.org/10.1056/NEJMoa1915745 AL C, S Q, M I, PR G, M D, TY K, et al. Updated efficacy and safety data from IMbrave150: atezolizumab plus bevacizumab vs. sorafenib for unresectable hepatocellular carcinoma. Journal of Hepatology. 2021. https://doi.org/10.1016/j.jhep.2021.11.030 OJ F, HG R. Is It Possible to Develop Cancer Vaccines to Neoantigens, What Are the Major Challenges, and How Can These Be Overcome? Neoantigens: Nothing New in Spite of the Name. Cold Spring Harbor perspectives in biology. 2018; 10. https://doi.org/10.1101/cshperspect.a028829 U S, P O, E D, RA J, M V, M G, et al. An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma. Nature. 2020; 585: 107-112. https://doi.org/10.1038/s41586-020-2537-9 Foerster F, Gairing SJ, Ilyas SI, Galle PR. Emerging immunotherapy for HCC: A guide for hepatologists. Hepatology. 2022; 75: 1604-1626. https://doi.org/10.1002/hep.32447 Tojjari A, Saeed A, Singh M, Cavalcante L, Sahin IH, Saeed A. A Comprehensive Review on Cancer Vaccines and Vaccine Strategies in Hepatocellular Carcinoma. Vaccines (Basel). 2023; 11. https://doi.org/10.3390/vaccines11081357 I L, W W, R M, C R, F T, L H. In Situ Vaccination as a Strategy to Modulate the Immune Microenvironment of Hepatocellular Carcinoma. Frontiers in Immunology. 2021; 12: 650486. https://doi.org/10.3389/fimmu.2021.650486 Messmer MN, Snyder AG, Oberst A. Comparing the effects of different cell death programs in tumor progression and immunotherapy. Cell Death Differ. 2019; 26: 115-129. https://doi.org/10.1038/s41418-018-0214-4 Koppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell. 2021; 12: 599-620. https://doi.org/10.1007/s13238-020-00789-5 Tang X, Chen W, Liu H, Liu N, Chen D, Tian D, et al. Research progress on SLC7A11 in the regulation of cystine/cysteine metabolism in tumors. Oncol Lett. 2022; 23: 47. https://doi.org/10.3892/ol.2021.13165 Liu X, Nie L, Zhang Y, Yan Y, Wang C, Colic M, et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol. 2023; 25: 404-414. https://doi.org/10.1038/s41556-023-01091-2 Beňačka R, Szabóová D, Guľašová Z, Hertelyová Z, Radoňak J. Non-Coding RNAs in Human Cancer and Other Diseases: Overview of the Diagnostic Potential. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242216213 Ahmadpour ST, Orre C, Bertevello PS, Mirebeau-Prunier D, Dumas JF, Desquiret-Dumas V. Breast Cancer Chemoresistance: Insights into the Regulatory Role of lncRNA. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242115897 Wu L, Zhang L, Zheng S. Role of the long non-coding RNA HOTAIR in hepatocellular carcinoma. Oncol Lett. 2017; 14: 1233-1239. https://doi.org/10.3892/ol.2017.6312 Chi H, Huang J, Yan Y, Jiang C, Zhang S, Chen H, et al. Unraveling the role of disulfidptosis-related LncRNAs in colon cancer: a prognostic indicator for immunotherapy response, chemotherapy sensitivity, and insights into cell death mechanisms. Front Mol Biosci. 2023; 10: 1254232. https://doi.org/10.3389/fmolb.2023.1254232 Tang Z, Kang B, Li C, Chen T, Zhang Z. GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis. Nucleic Acids Res. 2019; 47: W556-W560. https://doi.org/10.1093/nar/gkz430 Li T, Fan J, Wang B, Traugh N, Chen Q, Liu JS, et al. TIMER: A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells. Cancer Res. 2017; 77: e108-e110. https://doi.org/10.1158/0008-5472.Can-17-0307 Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018; 1711: 243-259. https://doi.org/10.1007/978-1-4939-7493-1_12 Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015; 12: 453-457. https://doi.org/10.1038/nmeth.3337 Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010; 26: 1572-1573. https://doi.org/10.1093/bioinformatics/btq170 Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013; 14: 7. https://doi.org/10.1186/1471-2105-14-7 Yoshihara K, Shahmoradgoli M, Martínez E, Vegesna R, Kim H, Torres-Garcia W, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013; 4: 2612. https://doi.org/10.1038/ncomms3612 Saha S, Raghava GP. Prediction of continuous B-cell epitopes in an antigen using recurrent neural network. Proteins. 2006; 65: 40-48. https://doi.org/10.1002/prot.21078 Yan Z, Kim K, Kim H, Ha B, Gambiez A, Bennett J, et al. Next-generation IEDB tools: a platform for epitope prediction and analysis. Nucleic Acids Res. 2024; 52: W526-W532. https://doi.org/10.1093/nar/gkae407 Reynisson B, Alvarez B, Paul S, Peters B, Nielsen M. NetMHCpan-4.1 and NetMHCIIpan-4.0: improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data. Nucleic Acids Res. 2020; 48: W449-W454. https://doi.org/10.1093/nar/gkaa379 Doytchinova IA, Flower DR. VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines. BMC Bioinformatics. 2007; 8: 4. https://doi.org/10.1186/1471-2105-8-4 Dimitrov I, Bangov I, Flower DR, Doytchinova I. AllerTOP v.2--a server for in silico prediction of allergens. J Mol Model. 2014; 20: 2278. https://doi.org/10.1007/s00894-014-2278-5 Rathore AS, Choudhury S, Arora A, Tijare P, Raghava GPS. ToxinPred 3.0: An improved method for predicting the toxicity of peptides. Comput Biol Med. 2024; 179: 108926. https://doi.org/10.1016/j.compbiomed.2024.108926 Magnan CN, Zeller M, Kayala MA, Vigil A, Randall A, Felgner PL, et al. High-throughput prediction of protein antigenicity using protein microarray data. Bioinformatics. 2010; 26: 2936-2943. https://doi.org/10.1093/bioinformatics/btq551 Gasteiger E, Gattiker A, Hoogland C, Ivanyi I, Appel RD, Bairoch A. ExPASy: The proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 2003; 31: 3784-3788. https://doi.org/10.1093/nar/gkg563 Rapin N, Lund O, Bernaschi M, Castiglione F. Computational immunology meets bioinformatics: the use of prediction tools for molecular binding in the simulation of the immune system. PLoS One. 2010; 5: e9862. https://doi.org/10.1371/journal.pone.0009862 Simon N, Friedman J, Hastie T, Tibshirani R. Regularization Paths for Cox's Proportional Hazards Model via Coordinate Descent. J Stat Softw. 2011; 39: 1-13. https://doi.org/10.18637/jss.v039.i05 P C, F F, M A, C M, M E, D R, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell reports. 2017; 18: 248-262. https://doi.org/10.1016/j.celrep.2016.12.019 Jiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018; 24: 1550-1558. https://doi.org/10.1038/s41591-018-0136-1 Lorentzen CL, Haanen JB, Met Ö, Svane IM. Clinical advances and ongoing trials on mRNA vaccines for cancer treatment. Lancet Oncol. 2022; 23: e450-e458. https://doi.org/10.1016/s1470-2045(22)00372-2 Tayir M, Wang YW, Chu T, Wang XL, Fan YQ, Cao L, et al. The function of necroptosis in liver cancer. Biochim Biophys Acta Mol Basis Dis. 2025; 1871: 167828. https://doi.org/10.1016/j.bbadis.2025.167828 Afonso MB, Rodrigues PM, Mateus-Pinheiro M, Simão AL, Gaspar MM, Majdi A, et al. RIPK3 acts as a lipid metabolism regulator contributing to inflammation and carcinogenesis in non-alcoholic fatty liver disease. Gut. 2021; 70: 2359-2372. https://doi.org/10.1136/gutjnl-2020-321767 F Z, M W, Z W, W L, X L. Screening of novel tumor-associated antigens for lung adenocarcinoma mRNA vaccine development based on pyroptosis phenotype genes. BMC Cancer. 2024; 24: 28. https://doi.org/10.1186/s12885-023-11757-7 H L, Y Z, M F, P N, Y G, R W, et al. Targeting leucine-rich PPR motif-containing protein/LRPPRC by 5,7,4'-trimethoxyflavone suppresses esophageal squamous cell carcinoma progression. International journal of biological macromolecules. 2024; 269: 131966. https://doi.org/10.1016/j.ijbiomac.2024.131966 Yu Y, Deng H, Wang W, Xiao S, Zheng R, Lv L, et al. LRPPRC promotes glycolysis by stabilising LDHA mRNA and its knockdown plus glutamine inhibitor induces synthetic lethality via m 6 A modification in triple-negative breast cancer. Clin Transl Med. 2024; 14: e1583. https://doi.org/10.1002/ctm2.1583 Yang X, Ding Y, Sun L, Shi M, Zhang P, He A, et al. WASF2 Serves as a Potential Biomarker and Therapeutic Target in Ovarian Cancer: A Pan-Cancer Analysis. Front Oncol. 2022; 12: 840038. https://doi.org/10.3389/fonc.2022.840038 J M, FF K, D Y, H Y, C W, R C, et al. lncRNA MIR210HG promotes the progression of endometrial cancer by sponging miR-337-3p/137 via the HMGA2-TGF-β/Wnt pathway. Molecular therapy. Nucleic acids. 2021; 24: 905-922. https://doi.org/10.1016/j.omtn.2021.04.011 Shi W, Tang Y, Lu J, Zhuang Y, Wang J. MIR210HG promotes breast cancer progression by IGF2BP1 mediated m6A modification. Cell Biosci. 2022; 12: 38. https://doi.org/10.1186/s13578-022-00772-z Guo Z, Xie Y, Zhang L, Liu S, Jiang W. A novel disulfidptosis-related lncRNAs signature for predicting survival and immune response in hepatocellular carcinoma. Aging (Albany NY). 2024; 16: 267-284. https://doi.org/10.18632/aging.205367 Jia X, Wang Y, Yang Y, Fu Y, Liu Y. Constructed Risk Prognosis Model Associated with Disulfidptosis lncRNAs in HCC. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242417626 Sahin U, Oehm P, Derhovanessian E, Jabulowsky RA, Vormehr M, Gold M, et al. An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma. Nature. 2020; 585: 107-112. https://doi.org/10.1038/s41586-020-2537-9 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 27 Apr, 2026 Reviews received at journal 07 Apr, 2026 Reviewers agreed at journal 07 Apr, 2026 Reviews received at journal 07 Apr, 2026 Reviews received at journal 07 Apr, 2026 Reviewers agreed at journal 06 Apr, 2026 Reviews received at journal 06 Apr, 2026 Reviewers agreed at journal 02 Apr, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers invited by journal 31 Mar, 2026 Editor assigned by journal 17 Mar, 2026 Submission checks completed at journal 12 Mar, 2026 First submitted to journal 12 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9072846","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":616681760,"identity":"14a0219d-8115-4173-947f-dccd09e7cfea","order_by":0,"name":"Renjie Zhang","email":"","orcid":"","institution":"The Second Affiliated Hospital of Harbin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Renjie","middleName":"","lastName":"Zhang","suffix":""},{"id":616681762,"identity":"1d64ab7a-6b65-4684-b2db-08902619bf0e","order_by":1,"name":"Qi Liu","email":"","orcid":"","institution":"Peking University Third Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qi","middleName":"","lastName":"Liu","suffix":""},{"id":616681764,"identity":"de8445dd-6ffc-4c60-9975-2a54cb9c5c40","order_by":2,"name":"Siqi Liu","email":"","orcid":"","institution":"The Second Affiliated Hospital of Harbin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Siqi","middleName":"","lastName":"Liu","suffix":""},{"id":616681765,"identity":"3447d502-81a3-4a1d-9eb9-46609cfc5e0e","order_by":3,"name":"Meifeng Li","email":"","orcid":"","institution":"The Second Affiliated Hospital of Harbin Medical University","correspondingAuthor":false,"prefix":"","firstName":"Meifeng","middleName":"","lastName":"Li","suffix":""},{"id":616681766,"identity":"ff57199a-375c-410b-a702-caff15280f58","order_by":4,"name":"Weixi Shen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYBAC+RlgSoKBjZn54IMPFRJy8oS0GNyAaWFnSzacccbC2LCBkBYJGIufx0yat60ikeEAIS3Szcce81RYJPYx8xgb8M6TSGBsYH746AY+v8w5lm7Mc0bCmI2ZrfCB5DaJPHYGNmPjHHzW3Mgxk85tk5ADen+zgeE2iWLGBh42afxa8r9J5/6T4GFjZjCTSJwjkdhwgKCWHDbp3AaQLSxmEgcbiNBicCPNTPrPMbBfkg0bgAzDZgJ+kZ+R/ExyRk1d4vz+wwcf/6mpk5Nnb374GK/DMAEzacpHwSgYBaNgFGABAMXOQ2Ic0n5SAAAAAElFTkSuQmCC","orcid":"","institution":"The Second Affiliated Hospital of Harbin Medical University","correspondingAuthor":true,"prefix":"","firstName":"Weixi","middleName":"","lastName":"Shen","suffix":""}],"badges":[],"createdAt":"2026-03-09 12:24:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9072846/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9072846/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106189460,"identity":"44bd917c-8724-4882-92c9-674485a6fb4c","added_by":"auto","created_at":"2026-04-05 17:09:57","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3150349,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScreening of tumor antigens based on disulfidptosis-related genes (DRGs) in hepatocellular carcinoma (HCC).\u003c/strong\u003e(A) Venn diagram illustrating the identification of HCC tumor antigens. Two genes were identified as highly expressed, frequently mutated, and significantly associated with overall survival (OS) and disease-free survival (DFS).(B) Waterfall plot showing the mutation status of 32 DRGs in HCC samples.(C-F) Kaplan-Meier survival curves demonstrating the association between the expression levels of LRPPRC and WASF2 and OS (C, E) as well as DFS (D, F) in HCC patients.(G, H) Scatter plots illustrating the correlations between the expression of the two antigens (LRPPRC and WASF2), tumor purity, and antigen-presenting cells (APCs).\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/b4dc9b1e47cc840abd6424b9.png"},{"id":106189324,"identity":"95dffa17-7a55-42e4-be24-a1a15150c6d4","added_by":"auto","created_at":"2026-04-05 17:09:44","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2621308,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of DRGs mutations and the tumor immune microenvironment in HCC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Differential expression of DRGs between tumor and normal tissues in HCC.(B) Interaction network among the DRGs.(C) Copy number variation (CNV) landscape of the 32 DRGs in HCC.(D) Chromosomal locations of DRGs with CNVs.(E) Expression levels of major histocompatibility complex (MHC) molecules, co-stimulatory molecules, and adhesion molecules in the tumor microenvironment of groups with high versus low LRPPRC expression.(F) Infiltration levels of immune cells in the tumor microenvironment of groups with high versus low LRPPRC expression.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/5ff789034b6171f28bd01275.png"},{"id":106189459,"identity":"628221c6-2f6e-4821-b731-6edd3a043d3f","added_by":"auto","created_at":"2026-04-05 17:09:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3028804,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacterization of DRG subtypes in hepatocellular carcinoma HCC.\u003c/strong\u003e(A–D) Unsupervised consensus clustering of 32 DRGs in the meta-cohort (TCGA + GEO datasets), identifying two distinct subtypes.(E) Principal component analysis (PCA) demonstrating separation between the two disulfidptosis-related gene subtypes.(F) Kaplan-Meier survival analysis showing overall survival differences between the two disulfidptosis subtypes in the meta-cohort.(G) Differential immune cell infiltration in the tumor microenvironment between the two subtype groups.(H) GSVA enrichment heatmap of hallmark pathways, with red indicating activation (high enrichment scores) and blue indicating suppression (low enrichment scores) in the two subtypes.(I) Heatmap of unsupervised clustering of DRGs, annotated with clinical features including age, gender, tumor stage, and the two disulfidptosis subtypes.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/d62b8d326f8355fb6248f20a.png"},{"id":106189466,"identity":"c3a34b76-1ae0-4567-b409-249cee462c2b","added_by":"auto","created_at":"2026-04-05 17:09:58","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2248038,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIn silico immune simulation of the designed mRNA vaccine via the C-ImmSim server.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) C-ImmSim simulation of antigen and antibody levels after three-dose vaccination. Antigen (black line) peaks after each dose (vertical black lines) and clears rapidly; IgM (green) dominates the primary response, while IgG1 (purple), IgG2 (yellow), and IgG1+IgG2 (cyan) rise sharply and persist longer after booster doses, showing enhanced booster effect.(B) Changes in B cell subpopulations after three-dose vaccination. Total B cells (black) and memory B cells (light blue) increase post-vaccination; IgM-type (red) peaks early then declines; IgG1 (purple) and IgG2 (orange) rise significantly after second and third doses, illustrating class switching and memory formation.(C)B cell states over time. Active state (purple) shows sharp peaks after each dose then declines; presenting (black) and duplicating (cyan) increase briefly during response; internalized (yellow) and anergic (orange) remain low, demonstrating B cell activation, differentiation, and regulation dynamics.(D) Helper T cell (TH) subpopulations. Total TH cells (black) peak after each dose (up to ~9000 cells/mm³) then decline; non-memory TH (light blue) dominate primary response; memory TH (cyan) increase substantially and persist after boosters, reflecting TH activation, expansion, and memory formation.(E) TH cell states. Active state (purple) peaks sharply after each vaccination then declines; duplicating (green) shows brief increases; resting (red) and anergic (black) stay low, illustrating TH cell activation, proliferation, and regulation with stronger responses on subsequent doses.(F) Cytotoxic T cell (TC) states. Active state (purple) rises mainly after primary dose then gradually declines; duplicating (green), resting (cyan), and anergic (yellow) remain low, showing early TC activation followed by attenuation.(G) Dendritic cell (DC) states. Total (black) and active (purple) peak after each dose then decline; internalized (yellow) and presenting states (orange/red) fluctuate, demonstrating DC antigen uptake, presentation, and activation supporting adaptive immunity initiation.(H) Macrophage (MA) states. Total (black) and active (purple) show fluctuating peaks after each dose then decline; internalized (yellow) and presenting-2 (cyan) increase briefly, reflecting MA antigen processing, presentation, and regulatory dynamics during multiple immunizations.(I) Cytokine concentration time course. Peaks occur after each dose; IFN-γ (purple) and IL-4 (cyan) show the highest amplitudes, with others (TNF-α, IL-10, TGF-β, etc.) lower, illustrating activation and balance of Th1 (IFN-γ, TNF-α) and Th2 (IL-4, IL-10) responses, with enhanced responses after boosters.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/30d4b466bfb19d1e3b9cf3c8.png"},{"id":106189461,"identity":"30d64a2c-fc4e-4d6b-9df7-167b9aa26057","added_by":"auto","created_at":"2026-04-05 17:09:57","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2227539,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrognostic modeling of disulfidptosis-related lncRNAs in HCC\u003c/strong\u003e.(A) Sankey diagram showing the correspondence between DRGs and associated lncRNAs.(B) Kaplan-Meier survival analysis of progression-free survival (PFS) between the high-risk and low-risk groups stratified by the prognostic model.(C, D) Lasso regression for constructing the disulfidptosis-related lncRNA prognostic model (E) Forest plot of univariate Cox regression analysis demonstrating the prognostic significance of disulfidptosis-related lncRNAs in HCC.(F) Correlation heatmap showing the relationships between the prognostic lncRNAs incorporated in the model and disulfidptosis-related genes.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/3f309371c3833ef41a22cb3d.png"},{"id":106189359,"identity":"058b74aa-0851-437e-9e70-ddf4e94346b5","added_by":"auto","created_at":"2026-04-05 17:09:50","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":2022711,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePerformance evaluation of the disulfidptosis-related LncRNA prognostic model across the training and validation cohorts.\u003c/strong\u003e (A-O) Visualization of the risk profiles, survival status, OS curves, and area under the curve (AUC) for the model\u003c/p\u003e\n\u003cp\u003ePCA demonstrated that samples could be clearly separated into two distinct clusters based on these 4 risk lncRNAs, supporting the discriminative capacity of the model (Figures 6H–6K). Time-dependent ROC curve analysis revealed high AUC values for predicting 1-year, 3-year, and 5-year survival rates, indicating favorable predictive performance (Figures 6M–6O). Compared with clinical indicators including age, sex, TNM stage, and grade, the model exhibited higher ROC curve AUC values (Figure 7A); C-index analysis also demonstrated superiority over individual clinical features (Figure 7B).\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/db85001c931db871738da3c0.png"},{"id":106189330,"identity":"90bb17ac-645a-402b-ba21-3710b6a3d1d6","added_by":"auto","created_at":"2026-04-05 17:09:44","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1643406,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssessment of the discriminative ability, predictive accuracy, and clinical utility of the disulfidptosis-related lncRNA prognostic model in HCC.\u003c/strong\u003e(A) ROC curves comparing the discriminative ability of the risk score with individual clinical features(B) Time-dependent C-index curves evaluating the predictive accuracy of the risk score and clinical features over time.(C) Calibration curves demonstrating the agreement between predicted and actual 1-, 3-, and 5-year OS probabilities for the nomogram.(D,E) Univariate and multivariate Cox regression analyses to identify key features.(F) Nomogram constructed by integrating the risk score and independent clinical prognostic factors for predicting 1-, 3-, and 5-year OS in HCC patients.(H–K) PCA plots illustrating the separation of different sample groups based on: (H) all genes, (I) DRGs, (J) disulfidptosis-related lncRNAs, and (K) the prognostic disulfidptosis-related lncRNAs incorporated in the model.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/16967ed86e50a53d065388d3.png"},{"id":106189416,"identity":"acd35a18-8ba3-4165-9722-3462d523bfaf","added_by":"auto","created_at":"2026-04-05 17:09:54","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2686474,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune Landscape and Tumor Microenvironment Associated with DRI in HCC. \u003c/strong\u003e(A,B) Differential analysis of immune cell infiltration between high and low-risk groups. (C) Differential analysis of immune checkpoint expression between high and low-risk groups.(D)Differences in the tumor microenvironment between the DRI high-risk group and the DRI low-risk group. (E-K) Correlation analysis of TIDE, dysfunction, ips, TAM.M2, exclusion, IFNG, and MDSC in high and low-risk groups of DRI.\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/b0bb00a2edc58f9a40651be9.png"},{"id":106189331,"identity":"4138d55e-b1da-4235-ad38-40d24183f587","added_by":"auto","created_at":"2026-04-05 17:09:44","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":1689724,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSomatic Mutation Landscape and Its Prognostic Relevance in DRI of HCC.\u003c/strong\u003e (A, B) Somatic mutation maps of the DRI high-risk and low-risk groups.(C) Analysis of the correlation between tumor mutation burden and prognosis. (D) Prognostic analysis of patients stratified by high and low tumor mutation burden in combination with high and low-risk DRI groups.\u003c/p\u003e","description":"","filename":"image9.png","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/086964284923dc88b97c74ed.png"},{"id":106404168,"identity":"f4692ff8-6746-4712-a917-39f914082504","added_by":"auto","created_at":"2026-04-08 09:15:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":22176356,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9072846/v1/6f01f22a-8dc6-4042-a612-5e3f338daea4.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrated analysis identifies disulfidptosis related tumor antigens and molecular subtypes in hepatocellular carcinoma for mRNA vaccine development","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eLiver cancer is one of the most prevalent and lethal malignancies worldwide. According to World Health Organization data, approximately 865,269 new cases and 757,948 deaths attributable to liver cancer are reported annually, rendering it the sixth leading contributor to the global cancer burden. Histopathologically, liver cancer encompasses hepatocellular carcinoma (HCC), which accounts for approximately 75%\u0026ndash;85% of cases, and intrahepatic cholangiocarcinoma, which constitutes 10%\u0026ndash;15%[1, 2]. Current therapeutic modalities for HCC are diverse and include surgical resection, radiofrequency ablation, chemotherapy, immunotherapy, and targeted therapy, with surgical intervention remaining the primary curative approach for early-stage HCC. The 1-year and 5-year overall survival (OS) rates following surgical treatment can reach up to 89% and 56%, respectivel[3]. However, HCC is frequently asymptomatic in its early stages and exhibits rapid progression; consequently, the majority of patients are diagnosed at advanced stages. For patients with advanced HCC, the survival benefit from surgery is limited, and treatment relies predominantly on immunotherapy and targeted therapy[4]. In a phase III clinical trial for advanced HCC, atezolizumab (an anti-PD-L1 antibody) combined with bevacizumab (an anti-VEGF agent) demonstrated significant efficacy, achieving an overall response rate (ORR) of 29.8% and a 5.8-month survival benefit over sorafenib. Nevertheless, only a minority of patients derive meaningful benefit from immunotherapy; approximately 70% of patients exhibit no appreciable response, and a subset experience severe adverse event[5, 6]. Therefore, there is an urgent need to develop novel combination therapeutic strategies to reactivate the immunosuppressive microenvironment in HCC.\u003c/p\u003e \u003cp\u003eTumor-associated antigens (TAAs) are protein molecules that are aberrantly expressed or specifically overexpressed in malignant tumor cells. TAAs encompass a broad spectrum of proteins, including those involved in dysregulated cell cycle control, oncogene products, cell surface receptors, and embryo-specific proteins that are normally expressed only during fetal development[7] .In the field of oncology, researchers have strategically leveraged TAAs as therapeutic targets for the development of tumor mRNA vaccines. This innovative approach introduces mRNA sequences encoding TAAs to activate and mobilize the host immune system to recognize and attack tumor cells. The advantages of this strategy include the effective induction of antigen-specific immune recognition and clearance of tumor cells, lower cytotoxicity, and improved patient tolerability, thereby enhancing both therapeutic safety and efficacy[8]. For tumors such as HCC that present significant therapeutic challenges, the development of mRNA vaccines is particularly important [9]. Existing therapies for HCC still have notable limitations, including drug resistance. mRNA vaccines offer a potentially low-toxicity, high-efficacy therapeutic alternative that may overcome these constraint [10]. At the level of the HCC microenvironment, mRNA vaccines also exhibit unique advantages: the HCC microenvironment is typically immunosuppressive, rendering conventional therapies suboptimal. mRNA vaccines can activate and enhance tumor-specific T cell activity within this complex microenvironment, demonstrating considerable therapeutic promise[11].\u003c/p\u003e \u003cp\u003eProgrammed cell death (PCD) encompasses multiple non-apoptotic death modalities, including necroptosis, ferroptosis, and oxytosis, which have long been implicated in tumorigenesis and tumor progression [12]. Among these, disulfidptosis is a recently identified form of regulated cell death: under conditions of high SLC7A11 expression and glucose deprivation-induced disulfide stress, aberrant intracellular accumulation of disulfide bonds triggers rapid cell death. SLC7A11 is a membrane-localized amino acid transporter responsible for importing extracellular cystine (a disulfide bond-containing amino acid) into the cell[13, 14]. When cells overexpress SLC7A11, cystine uptake is markedly increased. Under normal conditions, cystine is reduced to cysteine and participates in the synthesis of antioxidant molecules such as glutathione, thereby maintaining cellular redox homeostasis. However, under glucose deprivation, cells are unable to generate sufficient NADPH via the pentose phosphate pathway. NADPH is a critical reducing agent that facilitates the reduction of cystine to cysteine. Consequently, excess cystine accumulates intracellularly, forming highly toxic disulfide compounds that induce disulfide bond-mediated cell death [15]. This death modality has been proposed as a novel therapeutic target in cancer.\u003c/p\u003e \u003cp\u003eLong non-coding RNAs (lncRNAs) are transcripts exceeding 200 nucleotides in length that do not encode protein [16]. Once dismissed as transcriptional \u0026ldquo;junk,\u0026rdquo; lncRNAs have since been demonstrated to play critical roles in tumor drug resistance and other oncogenic processes, including in breast cancer [17]. In liver cancer, oncogenic lncRNAs such as HOTAIR exert profound effects on tumor growth and metastasis [18]. Furthermore, recent studies have revealed that disulfidptosis-related lncRNAs, such as ZEB1-AS1, promote tumor cell proliferation and metastasis when upregulated [19]. However, the precise molecular mechanisms and clinical significance of disulfidptosis in HCC remain insufficiently elucidated. Therefore, leveraging multi-cohort clinical RNA sequencing datasets to characterize the relationship between disulfidptosis subtypes and lncRNAs in HCC holds the potential to uncover complex mechanisms underlying HCC progression and metastasis, identify promising therapeutic targets, and provide a basis for effective HCC intervention.\u003c/p\u003e \u003cp\u003eIn the present study, we identified WASF2 and LRPPRC as potential tumor antigens for HCC mRNA vaccine development based on disulfidptosis-related genes and designed a novel mRNA vaccine targeting these two antigens. This not only provides a new strategy for HCC immunotherapy but also lays a foundation for future vaccine development. Additionally, we identified two novel disulfidptosis subtypes, offering new perspectives for understanding HCC pathogenesis. Beyond antigen identification and regulatory pattern analysis, we constructed a prognostic risk scoring model based on disulfidptosis-related lncRNAs. This model not only predicts prognosis in HCC patients but also facilitates patient selection for mRNA vaccine administration, thereby providing more precise and individualized therapeutic guidance for clinical practice.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data Collection and Processing\u003c/h2\u003e \u003cp\u003eRNA-Seq data were retrieved from two public databases: The Cancer Genome Atlas (TCGA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cancer.gov/tcga\u003c/span\u003e\u003cspan address=\"https://www.cancer.gov/tcga\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and the Gene Expression Omnibus (GEO, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). After filtering, a total of 493 liver cancer patient samples with complete survival information were included. The microarray dataset (GSE76427) was downloaded directly from the GEO database. Additionally, RNA sequencing data and clinical information for TCGA-LIHC were obtained from the Genomic Data Commons. The GEO and TCGA datasets were merged into a unified expression matrix, and batch effects were removed using the \u0026ldquo;sva\u0026rdquo; package and the \u0026ldquo;ComBat\u0026rdquo; function in R software. Somatic mutation data were downloaded directly from the TCGA database. Thirty-two disulfidptosis-related genes (SLC7A11, INF2, PDLIM1, CD2AP, MYH9, MYH10, ACTN4, FLNA, FLNB, IQGAP1, TLN1, MYL6, ACTB, DSTN, CAPZB, RPN1, NCKAP1, GYS1, NDUFS1, OXSM, LRPPRC, NDUFA11, NUBPL, SLC3A2, AJAP1, ACTR2, ACTR3, WASF2, CYFIP1, ABI2, SLC2A1, and BRK1) were selected based on previous literature [15]. LncRNAs were annotated using the latest annotation file from the Ensembl database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://asia.ensembl.org\u003c/span\u003e\u003cspan address=\"http://asia.ensembl.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Tumor Antigen Identification and Kaplan Meier Survival Analysis\u003c/h2\u003e \u003cp\u003eThe mutation status of disulfidptosis-related genes (DRGs) in TCGA-LIHC patients was analyzed and visualized using the \u0026ldquo;maftools\u0026rdquo; package in R, with a mutation rate\u0026thinsp;\u0026ge;\u0026thinsp;1% considered significant for identifying potential tumor antigens. OS and disease-free survival (DFS) were analyzed using the \u0026ldquo;survival analysis\u0026rdquo; module in Gene Expression Profiling Interactive Analysis 2nd version (GEPIA2,\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://gepia2.cancer-pku.cn/#index\u003c/span\u003e\u003cspan address=\"http://gepia2.cancer-pku.cn/#index\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [20].DRGs exhibiting mutations (\u0026ge;\u0026thinsp;1%) and significant associations with both OS and DFS were selected for tumor antigen construction. The Tumor Immune Estimation Resource was employed to evaluate correlations between selected tumor antigens and immune cell infiltration [21].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Immune Cell Infiltration Analysis\u003c/h2\u003e \u003cp\u003eThe relative proportions of tumor-infiltrating immune cells were estimated using the CIBERSORT algorithm. This algorithm employs support vector regression (ν-SVR) to deconvolve 22 human immune cell subtypes from bulk gene expression profiles [22, 23]. The LM22 signature matrix, comprising 547 genes capable of distinguishing 22 leukocyte phenotypes\u0026mdash;including 7 T cell subtypes, na\u0026iuml;ve/memory B cells, plasma cells, NK cells, and myeloid subsets\u0026mdash;was used as the reference. Normalized gene expression data served as input. Permutations were set to 100 to assess statistical significance. Only samples with a CIBERSORT \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026le;\u0026thinsp;0.05 were retained as reliable deconvolution results for subsequent correlation, survival, and differential abundance analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Identification of Disulfidptosis Related Subtypes\u003c/h2\u003e \u003cp\u003eBased on the expression profiles of 32 DRGs, unsupervised consensus clustering was performed on the merged TCGA-LIHC and GSE76427 liver cancer samples to identify potential disulfidptosis subtypes. Clustering analysis was implemented using the \u0026ldquo;ConsensusClusterPlus\u0026rdquo; package in R [24]. The specific parameters were as follows: clustering method, k-means; distance metric, Pearson correlation; number of resampling iterations, 1000; and cluster number k ranging from 2 to 9. The optimal number of clusters was determined by comprehensively evaluating the consensus matrix, cumulative distribution function (CDF) curves, proportion of ambiguous clustering (PAC), and item-consensus metrics.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Gene Set Variation Analysis and Tumor Microenvironment Differential Analysis\u003c/h2\u003e \u003cp\u003eGene Set Variation Analysis (GSVA) was employed to evaluate pathway activity differences between disulfidptosis subtypes. GSVA is a non-parametric, unsupervised method that transforms a sample-level gene expression matrix into a gene set enrichment score matrix, thereby quantifying inter-sample variation in pathway activity [25]. The canonical KEGG pathway gene set file \u0026ldquo;c2.cp.kegg.v7.4.symbols.gmt\u0026rdquo; was downloaded from the MSigDB database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/gsea/msigdb\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/gsea/msigdb\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and GSVA enrichment analysis was performed using the \u0026ldquo;GSVA\u0026rdquo; package in R. The input data consisted of the batch-corrected normalized expression matrix, with default method parameters, to examine differences in KEGG pathway regulation patterns between disulfidptosis subtypes. Subsequently, Wilcoxon rank-sum tests were used to compare GSVA scores between subtypes, and significantly differential pathways were visualized using heatmaps.\u003c/p\u003e \u003cp\u003eFurthermore, to assess immune infiltration levels within the tumor microenvironment, single-sample gene set enrichment analysis (ssGSEA) from the \u0026ldquo;GSVA\u0026rdquo; package was employed, using established immune cell marker gene sets to calculate immune cell infiltration enrichment scores for each sample. Simultaneously, the \u0026ldquo;estimate\u0026rdquo; package was used to compute ESTIMATE scores for each sample, including ImmuneScore, StromalScore, ESTIMATEScore, and TumorPurity, to comprehensively explore tumor microenvironment heterogeneity across subtype [26].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Development of mRNA Vaccine Targeting WASF2 and LRPPRC\u003c/h2\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.7.1 Retrieval of Tumor Specific Protein Sequences\u003c/h2\u003e \u003cp\u003eThe amino acid sequences of the two target proteins were retrieved from the UniProt database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.uniprot.org/\u003c/span\u003e\u003cspan address=\"https://www.uniprot.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), with accession numbers WASF2 (Q9Y6W5) and LRPPRC (P42704).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.7.2 B Cell Epitope Prediction\u003c/h2\u003e \u003cp\u003eLinear B cell epitopes were predicted using the ABCpred web server (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://webs.iiitd.edu.in/raghava/abcpred/\u003c/span\u003e\u003cspan address=\"https://webs.iiitd.edu.in/raghava/abcpred/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), with a window length of 16 amino acids, a threshold of 0.5, and the overlapping filter enabled to reduce redundant overlapping predictions [27]. This server is trained using a recurrent neural network based on linear B cell epitope data from the BCIPEP database.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.7.3 T Lymphocyte Epitope Prediction\u003c/h2\u003e \u003cp\u003eT cell epitope prediction was performed using the IEDB Analysis Resource (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://tools.iedb.org/\u003c/span\u003e\u003cspan address=\"http://tools.iedb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) T cell prediction tools, employing the NetMHCpan 4.1 EL algorithm (for MHC class I-restricted epitopes) and NetMHCIIpan 4.1 EL algorithm (for MHC class II-restricted epitopes) [28, 29]. These methods integrate extensive binding affinity and mass spectrometry-derived eluted ligand data and represent the current IEDB-recommended first-choice prediction tools. Predictions were conducted using the default HLA allele reference set, and candidate epitopes were ranked by predicted percentile rank, with low percentile epitopes prioritized .\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e2.7.4 Prediction of Epitope Antigenicity and Toxicity\u003c/h2\u003e \u003cp\u003eFollowing identification of linear B cell (LBL), cytotoxic T lymphocyte (CTL), and helper T lymphocyte (HTL) epitopes, candidate epitopes underwent immunobiological safety assessment. Antigenicity was evaluated using the VaxiJen server (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ddg-pharmfac.net/vaxijen/\u003c/span\u003e\u003cspan address=\"http://www.ddg-pharmfac.net/vaxijen/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) with the tumor model and a threshold of \u0026ge;\u0026thinsp;0.5 to enhance specificity [30]. This alignment-free tool predicts protective antigen potential of proteins. Allergenicity screening was performed using AllerTOP v2.1 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ddg-pharmfac.net/AllerTOP/\u003c/span\u003e\u003cspan address=\"https://www.ddg-pharmfac.net/AllerTOP/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), employing amino acid propensity-based auto-cross covariance transformation and k-nearest neighbor (kNN) classifier, retaining only epitopes predicted as non-allergens [31]. Toxicity prediction was conducted using ToxinPred3 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://webs.iiitd.edu.in/raghava/toxinpred3/\u003c/span\u003e\u003cspan address=\"https://webs.iiitd.edu.in/raghava/toxinpred3/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), an SVM-based model for distinguishing toxic from non-toxic sequences [32]. Only epitopes with VaxiJen scores\u0026thinsp;\u0026ge;\u0026thinsp;0.5, AllerTOP prediction of non-allergen, and absence of toxicity were retained for subsequent multi-epitope construct design and immune simulation analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e2.7.5 Design of the Multi Epitope mRNA Vaccine Construct\u003c/h2\u003e \u003cp\u003eThe multi-epitope mRNA vaccine was constructed by integrating the previously identified LBL, CTL, and HTL epitopes. To ensure correct epitope processing, presentation, and immunogenicity, the following linkers were employed: GPGPG (or GPPGG) spacers between HTL epitopes to maintain domain independence and facilitate MHC class II processing; KK linkers between LBL epitopes to promote lysosomal protease cleavage and preserve independent B cell epitope recognition; and AAY linkers between CTL epitopes to promote proteasomal cleavage and generate precise C-termini favorable for TAP transport and MHC class I presentation.\u003c/p\u003e \u003cp\u003eThe core antigen amino acid sequence was arranged from N-terminus to C-terminus as follows:\u003c/p\u003e \u003cp\u003eGPGPG-linked HTL epitope cluster \u0026rarr; KK-linked LBL epitopes \u0026rarr; AAY-linked CTL epitopes\u003c/p\u003e \u003cp\u003eTo achieve efficient mRNA expression, the coding region was flanked by the following regulatory elements: the 5\u0026prime; end comprised an m7G cap analog, an optimized 5\u0026prime; untranslated region (5\u0026prime; UTR), and a Kozak sequence (GCCACC AUG G) to enhance ribosome binding and translation initiation; the C-terminus included an MHC class I trafficking domain (MITD) to redirect antigens to the endosomal compartment and enhance cross-presentation efficiency, a UAA stop codon, a stabilized 3\u0026prime; UTR, and a 120-nt poly(A) tail to improve mRNA stability and translational fidelity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003e2.7.6 Prediction of Antigenicity and Physicochemical Properties of the Vaccine Construct\u003c/h2\u003e \u003cp\u003eFollowing assembly of the vaccine sequence, multi-platform bioinformatic validation was performed on the core chimeric antigen (excluding the tPA signal peptide and MITD domain) to evaluate its antigenicity, allergenicity, toxicity, and physicochemical stability. Antigenicity was predicted using VaxiJen and ANTIGENpro (SCRATCH suite, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://scratch.proteomics.ics.uci.edu/\u003c/span\u003e\u003cspan address=\"https://scratch.proteomics.ics.uci.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), both employing alignment-free methods based on amino acid sequences to assess protective antigen potential [30, 33]. Allergenicity was assessed using AllerTop v2.1, utilizing auto-cross covariance transformation and kNN classifier with default parameters [31], retaining sequences predicted as non-allergens. Toxicity was evaluated using ToxinPred3.0, an improved model integrating machine learning, deep learning, and motif analysis, with a toxicity probability\u0026thinsp;\u0026lt;\u0026thinsp;0.5 classified as non-toxic [32]. Physicochemical properties were calculated using the ProtParam tool (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://web.expasy.org/protparam/\u003c/span\u003e\u003cspan address=\"https://web.expasy.org/protparam/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), including amino acid composition, molecular weight, theoretical pI, instability index, aliphatic index, and GRAVY value (negative values favoring hydrophilicity and solubility) [34].This cross-validation strategy ensured high immunogenicity, low safety risk, and favorable biophysical properties of the vaccine candidate, providing a robust foundation for subsequent structural modeling and immune simulation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section3\"\u003e \u003ch2\u003e2.7.7 In Silico Immune Simulation\u003c/h2\u003e \u003cp\u003eTo predict the immune response kinetics of the mRNA vaccine construct in the human body, in silico immune simulation was performed using the C-IMMSIM online platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://kraken.iac.rm.cnr.it/C-IMMSIM/\u003c/span\u003e\u003cspan address=\"https://kraken.iac.rm.cnr.it/C-IMMSIM/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). This platform is based on the Celada\u0026ndash;Seiden cellular automaton model, which simulates humoral and cellular immune processes in the mammalian immune system [35].\u003c/p\u003e \u003cp\u003eThe vaccination regimen was set to three doses, administered at time steps 1, 84, and 168, each comprising 1000 antigen units. In the C-IMMSIM model, each time step corresponds to approximately 8 hours of real time; thus, the three vaccinations were spaced approximately at days 0, 28, and 56. The total simulation length was set to 350 steps, covering approximately 117 days of immune dynamics. All parameters other than vaccination timing and dosage were set to server defaults. This computational framework was used to systematically evaluate the coordination of vaccine-induced humoral and cellular immune responses and the potential for long-term immunological memory formation.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Construction and Validation of the Disulfidptosis Risk Index\u003c/h2\u003e \u003cp\u003eThe TCGA-LIHC expression matrix was partitioned into mRNA and lncRNA groups using the Ensembl annotation file. Co-expression analysis was performed on the lncRNA group based on 32 DRGs retaining lncRNAs with correlation coefficients\u0026thinsp;\u0026ge;\u0026thinsp;0.4 and \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001 for subsequent analysis. After excluding normal samples from TCGA-LIHC, the disulfidptosis-related lncRNA expression data were merged with patient survival data. Samples were randomly divided into training and testing sets at a 1:1 ratio using the \u0026ldquo;caret\u0026rdquo; package in R. The prognostic model was constructed in the training set and validated in the testing set. Univariate Cox regression analysis was first performed with a screening threshold of P\u0026thinsp;\u0026lt;\u0026thinsp;0.001. A Lasso regression model was then constructed using the \u0026ldquo;glmnet\u0026rdquo; package in R [36], and feature lncRNAs were selected via cross-validation. These feature lncRNAs were subsequently used to construct a multivariate Cox regression model. The model formula was as follows:\u003c/p\u003e \u003cp\u003eDisulfidptosis Risk Index (DRI) = expression of lncRNA₁ \u0026times; Coef(lncRNA₁) + expression of lncRNA₂ \u0026times; Coef(lncRNA₂) + \u0026hellip; + expression of lncRNAₙ \u0026times; Coef(lncRNAₙ)\u003c/p\u003e \u003cp\u003ewhere Coef\u003csub\u003ei\u003c/sub\u003e represents the regression coefficient. Risk scores were calculated for each sample using this formula, and samples were stratified into high-risk and low-risk groups based on the median risk score from the training set. Model validity and stability were verified using ROC curves, concordance index(C-index), and independent prognostic analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Construction and Validation of the Nomogram Based on the DRI\u003c/h2\u003e \u003cp\u003eA prognostic nomogram was constructed by integrating the DRI with clinicopathological features to enable individualized survival probability prediction for HCC patients. The nomogram was generated using the \u0026ldquo;rms\u0026rdquo; package in R. First, univariate and multivariate Cox proportional hazards regression analyses were performed to confirm whether DRI and clinical variables (age, sex, TNM stage, and tumor grade) were independent prognostic factors. These independent prognostic variables were incorporated into the lrm or cph functions of the \u0026ldquo;rms\u0026rdquo; package to generate the nomogram model. The nomogram quantified the contribution of each variable as a point score, and total points were mapped to predicted probabilities of 1-year, 3-year, and 5-year OS.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Immunotherapy Response Prediction\u003c/h2\u003e \u003cp\u003eTo evaluate potential differential responses to immune checkpoint inhibitors (ICIs) between DRI risk groups, pre-computed immunotherapy response scores (IPS) were obtained from The Cancer Immunome Atlas (TCIA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tcia.at/\u003c/span\u003e\u003cspan address=\"https://tcia.at/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). IPS is a computational metric based on tumor transcriptomic data used to predict patient sensitivity to various ICI regimens [37]. Specifically, IPS scores for the following four combinations were included:ips_ctla4_pos_pd1_pos,ips_ctla4_neg_pd1_pos,ips_ctla4_pos_pd1_neg, and ips_ctla4_neg_pd1_neg. Higher IPS scores indicate stronger tumor immunogenicity and better predicted ICI response. Differences in these four IPS scores between high- and low-risk groups were compared using the Wilcoxon rank-sum test.\u003c/p\u003e \u003cp\u003eFurthermore, the Tumor Immune Dysfunction and Exclusion (TIDE) score was employed to further predict immunotherapy response. TIDE is a transcriptome-based computational framework that models the two major mechanisms of tumor immune evasion, T cell dysfunction and T cell exclusion, to predict the probability of patient response to ICIs [38].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Statistical Analysis\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed using R software (version 4.3.2). For inter-group comparisons of continuous variables, independent-sample t-tests were used when data were normally distributed with equal variances; otherwise, the non-parametric Wilcoxon rank-sum test was applied. Categorical variables were compared using the chi-square (χ\u0026sup2;) test. Survival analysis was conducted using the Kaplan\u0026ndash;Meier method, and differences in OS between groups were evaluated using the log-rank test. Multiple comparison correction was performed using the Benjamini\u0026ndash;Hochberg false discovery rate method. All statistical tests were two-sided, and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. All figures were generated using the \u0026ldquo;ggplot2,\u0026rdquo; \u0026ldquo;ggpubr,\u0026rdquo; \u0026ldquo;survminer,\u0026rdquo; and \u0026ldquo;pheatmap\u0026rdquo; packages, with statistical significance annotated as \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01, **\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, and ns: not significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Development of mRNA Vaccines Based on Disulfidptosis Related Genes\u003c/h2\u003e \u003cp\u003eUsing the survival analysis module in GEPIA2, we evaluated the association of each DRG's expression levels with OS and disease-free survival DFS in HCC patients.Concurrently, the mutation status of these genes was analyzed, and genes with mutation rates\u0026thinsp;\u0026ge;\u0026thinsp;1% and significant prognostic associations were selected (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Ultimately, integrating prognostic value and potential immunogenicity, LRPPRC and WASF2 were identified as candidate tumor antigens for HCC mRNA vaccine development. Kaplan\u0026ndash;Meier analysis demonstrated that patients with high expression of LRPPRC and WASF2 exhibited significantly shorter OS and DFS (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC\u0026ndash;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eAntigen-presenting cells (APCs) are critical for activating tumor-specific T cell responses, among which dendritic cells (DCs) are the most potent, efficiently presenting tumor antigens to T cells; macrophages and B cells can also process and present antigens under certain conditions. As shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG\u0026ndash;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH, the expression levels of LRPPRC and WASF2 were significantly positively correlated with APC infiltration (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). These results suggest that LRPPRC and WASF2 may serve as effective shared tumor antigens with potential for HCC mRNA vaccine development.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.2 DRG Mutation Characteristics and Immune Microenvironment Analysis\u003c/h2\u003e \u003cp\u003eBy integrating transcriptomic data downloaded from the TCGA and GEO databases, we constructed an expression matrix containing survival information. We compared the expression of 32 DRGs between normal and tumor samples. Results revealed that 28 genes exhibited significant differential expression between tumor and normal tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Subsequently, a DRG network was constructed to elucidate the interrelationships and prognostic significance of DRGs in HCC. The analysis demonstrated statistically significant correlations among DRGs. The network indicated that, with the exception of NDUFA11 and NUBPL, all remaining genes were risk factors, with statistical evidence supporting significant positive regulatory relationships among them (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). To investigate whether genetic alterations affect DRG expression levels in HCC, copy number variations (CNVs) were analyzed. Results indicated that CNVs are a significant contributor to DRG expression dysregulation. Most DRGs with copy number gains were significantly upregulated in HCC. All 32 DRGs exhibited CNVs, with copy number gains being more prevalent, while CYFIP1, AJAP1, NDUFA11, WASF2, INF2, MYH10, and CAPZB showed higher frequencies of copy number losses (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD illustrates the chromosomal CNV distribution of these DRGs. These findings suggest that dysregulated DRG expression plays a critical role in HCC development and progression, and our study confirms significant differences in the genetic characteristics and expression levels of DRGs between normal and HCC tissues.\u003c/p\u003e \u003cp\u003eFurthermore, we conducted an in-depth analysis of the potential tumor antigen LRPPRC. Results demonstrated that the LRPPRC-high expression group exhibited a higher proportion of M0 macrophage infiltration, whereas CD8\u0026thinsp;+\u0026thinsp;T cell infiltration was significantly increased in the low-expression group (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF). CD8\u0026thinsp;+\u0026thinsp;T cells are the core effector cells in tumor immunotherapy, and their decreased infiltration is frequently associated with poor immunotherapy response. We also observed that expression of major histocompatibility complex (MHC) molecules, co-stimulatory molecules, and cell adhesion molecules was generally upregulated in the LRPPRC-low expression group (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). These findings further support the feasibility and validity of LRPPRC as a potential tumor-associated antigen in HCC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Identification of Two Disulfidptosis Subtypes Based on 32 DRGs\u003c/h2\u003e \u003cp\u003eBased on the expression profiles of 32 DRGs, unsupervised consensus clustering analysis was performed, successfully identifying two significantly distinct disulfidptosis molecular subtypes, designated DRGcluster A and DRGcluster B. The consensus matrix exhibited a clear block-diagonal structure, and the CDF curves and delta area curves reached plateaus, indicating stable and reliable clustering results (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA\u0026ndash;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003eTo validate the effectiveness and biological significance of subtype discrimination, principal component analysis (PCA) was performed on the DRG expression matrix. Results showed that samples from the two subtypes were clearly separated in the first two principal component space, with minimal overlap (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE). This confirmed that DRG-based clustering genuinely reflects heterogeneity in disulfidptosis regulatory patterns among HCC patients and captures subtype-specific gene expression signatures. Kaplan\u0026ndash;Meier survival analysis revealed that DRGcluster B patients exhibited significantly better overall survival compared to DRGcluster A (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF), indicating a close association between disulfidptosis molecular subtypes and HCC prognosis.\u003c/p\u003e \u003cp\u003eFurther comparison of immune cell infiltration differences between the two subtypes revealed that DRGcluster A exhibited higher infiltration levels of activated CD4\u0026thinsp;+\u0026thinsp;T cells, activated dendritic cells, follicular helper T cells, and Th2 helper T cells, whereas DRGcluster B showed higher infiltration of activated B cells, activated CD8\u0026thinsp;+\u0026thinsp;T cells, eosinophils, macrophages, monocytes, natural killer cells, Th1 helper T cells, and neutrophils (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG). These differences reflect heterogeneity in the tumor immune microenvironment between the two subtypes. A heatmap displaying the expression distribution of 32 DRGs across the cohort, overlaid with clinical parameters, demonstrated that DRGcluster A exhibited higher expression levels for most DRGs, while DRGcluster B showed relatively lower expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI).\u003c/p\u003e \u003cp\u003eGSVA comparing pathway enrichment differences between the two subtypes revealed that DRGcluster B was predominantly enriched in endocytosis, antidiuretic hormone-regulated water reabsorption, cell cycle, RNA degradation, ubiquitin-mediated proteolysis, and basal transcription factor pathways; DRGcluster A was primarily enriched in nitrogen metabolism, amino acid metabolism, retinol metabolism, and lipid metabolism pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH). In summary, this study identified two distinct disulfidptosis regulatory patterns based on DRGs, which exhibit significant differences in molecular characteristics, patient prognosis, and tumor immune microenvironment infiltration. These findings contribute to elucidating the molecular heterogeneity of HCC and provide potential evidence for developing individualized therapeutic strategies targeting the disulfidptosis pathway.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Epitope Prediction and Screening\u003c/h2\u003e \u003cp\u003eLinear B cell epitopes for WASF2 and LRPPRC protein sequences were predicted using the ABCPred server, initially retaining epitopes with scores\u0026thinsp;\u0026gt;\u0026thinsp;0.85. Candidate epitopes were then subjected to a rigorous three-step screening process: (1) VaxiJen antigenicity assessment (threshold\u0026thinsp;\u0026ge;\u0026thinsp;0.5; tumor antigen model); (2) AllerTop v2.1 allergenicity screening (non-allergen); and (3) ToxinPred toxicity analysis (non-toxic). This process yielded 11 high-antigenicity, non-allergenic, non-toxic B cell epitopes.\u003c/p\u003e \u003cp\u003eFor CTL epitopes, the IEDB MHC-I binding prediction tool was employed, retaining high-affinity candidates with percentile rank\u0026thinsp;\u0026le;\u0026thinsp;0.01. The same three-step validation was applied, ultimately yielding 6 high-affinity CTL epitopes for vaccine construction to achieve broad HLA coverage and cross-variant immunogenicity.\u003c/p\u003e \u003cp\u003eHTL epitopes were comprehensively predicted using the IEDB MHC-II tool and NetMHCIIpan 4.1, retaining high-affinity candidates with percentile rank\u0026thinsp;\u0026le;\u0026thinsp;0.2. Following the same three-step screening, 4 high-quality HTL epitopes were prioritized for vaccine inclusion (Supplementary Table S1).\u003c/p\u003e \u003cp\u003eTo optimize the final multi-epitope vaccine antigen construct, the top two candidate epitopes (encompassing B cell, CTL, and HTL epitopes) from each target protein (WASF2 and LRPPRC) were selected based on VaxiJen antigenicity scores, ensuring maximal immunogenicity and design compactness.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Design of the Multi Epitope mRNA Vaccine Construct Based on HCC Epitopes\u003c/h2\u003e \u003cp\u003eThe mRNA vaccine construct was designed with the following arrangement from N-terminus to C-terminus: 5\u0026prime; m7G Cap\u0026mdash;5\u0026prime; UTR\u0026mdash;Kozak sequence\u0026mdash;Signal peptide (tPA)\u0026mdash;GPGPG Linker\u0026mdash;ALSFFHMLNGAALRG\u0026mdash;GPGPG Linker\u0026mdash;AYDIFLNAKEQNIVF\u0026mdash;GPGPG Linker\u0026mdash;FASRVSSLAERVDRL\u0026mdash;GPGPG Linker\u0026mdash;KEKMLQDTKDIMKEK\u0026mdash;KK Linker\u0026mdash;AGIEPGPDTYLALLNA\u0026mdash;KK Linker\u0026mdash;HYFWPLLVGRRKEKNV\u0026mdash;KK Linker\u0026mdash;RGSGLAGPKRSSVVSP\u0026mdash;KK Linker\u0026mdash;VEEQREQEKRDVVGND\u0026mdash;AAY Linker\u0026mdash;YVSEILEKV\u0026mdash;AAY Linker\u0026mdash;IPREKTLRL\u0026mdash;AAY Linker\u0026mdash;DVATILSRR\u0026mdash;AAY Linker\u0026mdash;RQLGSLSKY\u0026mdash;MITD sequence\u0026mdash;Stop codon\u0026mdash;3\u0026prime; UTR\u0026mdash;Poly(A) tail.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.7 In Vitro Immunogenicity Assessment of the Multi Epitope Tumor mRNA Vaccine Construct\u003c/h2\u003e \u003cp\u003eTo evaluate the potential immunogenicity of the designed multi-epitope vaccine construct, agent-based immune simulation was performed using the C-IMMSIM online server. The simulation modeled a three-dose vaccination regimen, with injections at time steps 1, 84, and 168 (corresponding to approximately day 1, day 28, and day 56 in real time, with each step representing 8 hours). Key parameters included a simulation volume of 10, total steps of 350, an antigen dose of 1000 particles per injection, and an adjuvant level of 100. The simulation demonstrated effective antigen clearance and robust immune activation (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA\u0026ndash;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eI).\u003c/p\u003e \u003cp\u003eAntigen levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) declined rapidly after each injection, approaching zero by days 20\u0026ndash;30, accompanied by transient immune complexes. Immunoglobulin production showed a primary IgM response peaking at approximately 10⁴ \u0026micro;g/mL around day 10, followed by class-switching to sustained IgG1 and IgG2 levels (~\u0026thinsp;10⁵ \u0026micro;g/mL by day 40), indicative of class switching and long-term humoral memory.\u003c/p\u003e \u003cp\u003eCell population dynamics (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eH) revealed initial peaks of active and proliferating B cells, plasma cells, T helper cells, and macrophages at approximately 5\u0026ndash;20 days post-first injection, with secondary peaks at approximately day 30 following the second injection. The internalization and MHC class II presentation states of DCs and macrophages increased rapidly post-injection, supporting antigen processing and T cell priming; resting states stabilized in later phases, with mild increases in anergic cells suggestive of orderly regulation without excessive tolerance.\u003c/p\u003e \u003cp\u003eThe cytokine profile (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eI) showed early pro-inflammatory peaks (IFN-γ, TNF-α, IL-2\u0026thinsp;~\u0026thinsp;10⁴\u0026ndash;10⁵ pg/mL, days 5\u0026ndash;20), with secondary elevations at approximately day 30 following the booster dose. Anti-inflammatory IL-10 subsequently increased to facilitate resolution. Inset analysis confirmed the synergistic action of danger signals and IL-2, supporting leukocyte proliferation without cytokine storm. Overall, the simulation predicted a protective immune profile characterized by robust humoral responses and immunological memory.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.8 Evaluation of Antigenicity and Physicochemical Properties of the Vaccine Construct\u003c/h2\u003e \u003cp\u003eComprehensive in vitro prediction and immune simulation were performed on the vaccine construct. First, physicochemical properties were analyzed using the ExPASy ProtParam tool. Results revealed a total of 200 amino acids with a molecular weight of approximately 21.90 kDa, a molecular formula of C₉₈₂H₁₅₈₂N₂₈₂O₂₇₉S₃, and a theoretical isoelectric point of 9.93. The construct contained 21 negatively charged residues and 36 positively charged residues. The instability index was 45.63, the aliphatic index was 79.55, the GRAVY value was \u0026minus;\u0026thinsp;0.536, and the estimated in vitro half-life was 30 hours. These parameters indicate favorable solubility and potential for cellular expression, although the slightly elevated instability index suggests that further sequence optimization may be warranted (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAntigenicity predictions using the VaxiJen and ANTIGENpro servers yielded scores of 0.6686 and 0.710346, respectively, indicating that the vaccine construct possesses favorable potential immunogenicity. Allergenicity prediction classified the construct as non-allergenic, and toxicity prediction indicated non-toxicity. These safety assessments support the feasibility of this construct as a candidate tumor mRNA vaccine (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAntigenic, allergenic, toxic, and physicochemical assessments of the protein translated from the mRNA vaccine-encoded peptide.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProperty\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMeasurement\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal number of amino acids\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e200\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMolecular weight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21899.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFormula\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eC982H1582N282O279S3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTheoretical pI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.93\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal number of negatively charged residues (Asp\u0026thinsp;+\u0026thinsp;Glu)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal number of positively charged residues (Arg\u0026thinsp;+\u0026thinsp;Lys)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal number of atoms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3128\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInstability Index (II)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45.63\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAliphatic Index (A.I)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79.55\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrand average of hydropathicity (GRAVY)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.536\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eestimated half-life\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 hours (mammalian reticulocytes, in vitro)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntigenicity (using VaxiJen)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.6686\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntigenicity (using ANTIGENpro)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.710346\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAllergenicity (using AllerTop 2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNon-allergenic\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eToxicity (ToxinPred)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNon-toxic\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e3.9 Construction of a Disulfidptosis Related LncRNA Based Risk Model and Its Diagnostic and Prognostic Value\u003c/h2\u003e \u003cp\u003eUsing the TCGA HCC dataset, we performed co-expression analysis based on previously reported DRGs and identified a total of 1,223 lncRNAs significantly correlated with DRGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003ePrognostic risk-associated lncRNAs were screened via univariate Cox regression analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). Lasso regression combined with cross-validation was subsequently employed to identify the 6 most predictive lncRNAs. Multivariate Cox regression was then performed, ultimately incorporating 4 independently prognostic lncRNAs into the risk scoring model (Disulfidptosis Risk Index, DRI) (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). The DRI was calculated as follows:\u003c/p\u003e \u003cp\u003eDRI = (MIR210HG \u0026times; 0.2977) + (AL031985.3 \u0026times; 0.3079) + (AC108752.1 \u0026times; 0.3436) + (AC016717.2 \u0026times; 0.5874)\u003c/p\u003e \u003cp\u003eCorrelation analysis revealed significant positive and negative correlations between these 4 lncRNAs and DRGs, suggesting their potential involvement in the disulfidptosis regulatory network (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo validate model robustness, the cohort was randomly divided into training and validation sets. The DRI was calculated for each sample based on lncRNA expression levels and regression coefficients, and patients were stratified into high-risk and low-risk groups using the median DRI as the threshold. Kaplan\u0026ndash;Meier survival analysis demonstrated significantly worse overall survival in the high-risk group compared to the low-risk group (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eJ\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eL); progression-free survival analysis also confirmed longer PFS in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Heatmaps revealed significant expression differences of the 4 lncRNAs between high- and low-risk groups (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). Risk curves and scatter plots further indicated that increasing DRI was associated with more death events and shorter survival times (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePCA demonstrated that samples could be clearly separated into two distinct clusters based on these 4 risk lncRNAs, supporting the discriminative capacity of the model (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eK). Time-dependent ROC curve analysis revealed high AUC values for predicting 1-year, 3-year, and 5-year survival rates, indicating favorable predictive performance (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eM\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eO). Compared with clinical indicators including age, sex, TNM stage, and grade, the model exhibited higher ROC curve AUC values (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA); C-index analysis also demonstrated superiority over individual clinical features (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003eUnivariate and multivariate Cox regression analyses confirmed that TNM stage (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and DRI (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were independent risk factors for HCC prognosis (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD,\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE). Based on these findings, a nomogram integrating clinical features and DRI was constructed, achieving a C-index of 0.711, indicating satisfactory concordance and clinical predictive value (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eIn summary, the DRI model based on disulfidptosis-related lncRNAs demonstrated favorable prognostic predictive ability in the TCGA cohort, effectively stratifying HCC patient risk and outperforming conventional clinical indicators. This model provides a reference for individualized prognostic assessment and exploration of potential therapeutic targets in HCC, and the role of lncRNAs in disulfidptosis regulation warrants further functional validation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e3.10 DRI and Tumor Immune Microenvironment and Immunotherapy Response Prediction\u003c/h2\u003e \u003cp\u003eThe tumor immune microenvironment (TME) status is a critical determinant of immunotherapy efficacy. This section focuses on the role of the DRI in the HCC TME and its predictive value for potential immunotherapy response.ESTIMATE algorithm analysis revealed that the low-DRI group exhibited significantly higher ESTIMATEScore, ImmuneScore, and StromalScore compared to the high-DRI group (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eD), suggesting that the low-risk group TME is characterized by greater immune infiltration and stromal abundance. Heatmap visualization of immune cell infiltration demonstrated that the low-risk group exhibited higher proportions of B cells, dendritic cells, macrophages, neutrophils, natural killer cells, helper T cells, and tumor-infiltrating lymphocytes, while the high-risk group showed correspondingly lower levels (Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA, \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB). These differences suggest that the low-DRI group may possess more \u0026ldquo;hot\u0026rdquo; tumor characteristics.\u003c/p\u003e \u003cp\u003eFurther comparison of immune checkpoint molecule expression revealed that the low-risk group exhibited higher expression of TMIGD2 and IDO2, whereas the high-risk group showed significant upregulation of multiple checkpoint molecules, including HLA2, TNFRSF4, CD276, TNFRSF14, TNFSF4, CD274, TNFSF9, VTCN1, and TNFSF15 (Fig.\u0026nbsp;10C). The elevated expression of immunosuppressive checkpoints (such as CD274/PD-L1 and CD276/B7-H3) in the high-risk group suggests the presence of stronger immune evasion mechanisms, and DRI may serve as a reference indicator for evaluating potential benefit from ICIs.\u003c/p\u003e \u003cp\u003eImmunotherapy response prediction based on the TIDE algorithm demonstrated that the low-DRI group exhibited lower TIDE scores, T cell exclusion scores, myeloid-derived suppressor cell scores, and interferon-γ -related scores compared to the high-DRI group (Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eK\u0026ndash;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eN; Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eE\u0026ndash;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eF). These findings indicate that the low-risk group has lower degrees of immune dysfunction and exclusion, suggesting better potential IPS analysis further confirmed that the low-DRI group had significantly higher IPS than the high-DRI group, suggesting the potential for superior clinical benefit following ICI treatment (Figs.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eG\u0026ndash;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eI).\u003c/p\u003e \u003cp\u003eIn summary, the DRI model effectively reflects the immune microenvironment heterogeneity of HCC: the low-risk group demonstrates more active immune infiltration and lower immunosuppressive characteristics, with potential value in predicting immunotherapy response. This finding provides a bioinformatic basis for individualized immunotherapy decision-making in HCC patients and lays a foundation for subsequent immunotherapy strategies integrating disulfidptosis regulation. Through DRI subtype stratification, the response of HCC patients to immunotherapy can be more accurately predicted, thereby facilitating the selection of individuals suitable for mRNA immunotherapy.\u003c/p\u003e \u003cp\u003eIn conclusion, the DRI model effectively identifies the immune microenvironment heterogeneity of HCC: the low-DRI group exhibits more active immune infiltration and lower immunosuppressive features, with value in predicting general immunotherapy response. Given that the present study designed a multi-epitope mRNA vaccine construct based on the disulfidptosis-related targets LRPPRC and WASF2, DRI may serve as a potential biomarker to assist in selecting immunologically \u0026ldquo;hot\u0026rdquo; patient subgroups for prioritized mRNA vaccine or combination immunotherapy strategies. These findings provide a bioinformatic basis for integrating disulfidptosis regulation with personalized mRNA vaccines.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e3.11 Association Analysis Between DRI and the Somatic Mutation Landscape of HCC\u003c/h2\u003e \u003cp\u003eTo explore the relationship between the DRI and tumor genomic features, we analyzed the somatic mutation profiles of high- and low-DRI groups in the TCGA-HCC cohort using the maftools package in R. Waterfall plots revealed that the tumor mutation burden (TMB) was overall higher in the high-DRI group than in the low-DRI group (Figs.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB). Specifically, TP53 was the most frequently mutated gene in the high-DRI group (~\u0026thinsp;38%), while CTNNB1 was more frequently mutated in the low-DRI group (~\u0026thinsp;25%). Further analysis demonstrated that TMB was significantly higher in the high-DRI group compared to the low-DRI group, indicating an association between high DRI and elevated somatic mutation burden (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eC). Survival analysis revealed that among low-DRI patients, those with high TMB derived greater survival benefit; in the overall cohort, patients with low TMB exhibited relatively better prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eD). Combined DRI and TMB stratification analysis revealed significant survival differences among four groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eE): the low-TMB\u0026thinsp;+\u0026thinsp;low-DRI group had the best prognosis, the high-TMB\u0026thinsp;+\u0026thinsp;high-DRI group had the worst prognosis, and the high-TMB\u0026thinsp;+\u0026thinsp;low-DRI and low-TMB\u0026thinsp;+\u0026thinsp;high-DRI groups exhibited intermediate outcomes. These results suggest that the high TMB in the high-DRI group may be accompanied by increased neoantigen production, but combined with its immunosuppressive microenvironment, the overall immunotherapy response remains poor; conversely, although the low-DRI group has lower TMB and fewer neoantigens, its immunologically \u0026ldquo;hot\u0026rdquo; phenotype may be more conducive to effector cell activation and antigen presentation.\u003c/p\u003e \u003cp\u003eIn summary, combined analysis of DRI and TMB more precisely reflects the genomic-immune heterogeneity of HCC. The high-TMB\u0026thinsp;+\u0026thinsp;low-DRI subgroup may have advantages in generating tumor-specific antigens, while the immunologically active microenvironment of the low-DRI group provides potential synergistic effects for mRNA vaccines. Our results suggest that the combined application of DRI and TMB may improve the accuracy of HCC immunotherapy efficacy prediction and provide a potential reference for individualized therapeutic decisions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eHCC is one of the most prevalent and lethal malignancies worldwide, responsible for hundreds of thousands of deaths annually. Because early-stage liver cancer is typically asymptomatic, the majority of patients are diagnosed at advanced stages. The principal therapeutic modalities for advanced HCC consist of targeted therapy and immunotherapy. However, only a minority of patients benefit from these treatments, and some experience severe adverse events. Consequently, there is an urgent need to develop novel therapeutic approaches to improve clinical outcomes for HCC patients.\u003c/p\u003e \u003cp\u003eIn recent years, mRNA vaccines have garnered extensive attention as a novel class of therapeutic vaccines in oncology. The mechanism of action involves the design and synthesis of mRNA molecules encoding specific tumor antigens. Upon cellular uptake, these mRNA molecules direct the intracellular protein synthesis machinery to produce the specified tumor antigen proteins. These tumor antigens are recognized by the host immune system as foreign entities, thereby activating immune responses against cancer cells. Compared with conventional immunotherapy and targeted therapy, mRNA vaccines offer advantages including high efficacy, favorable safety profiles, and low toxicity [10, 39].\u003c/p\u003e \u003cp\u003eRegulated necrosis differs from traditional necrosis caused by external injury in that it is precisely controlled by intracellular signaling pathways and exhibits programmatic features. This form of cell death involves multiple intracellular signaling pathways and signaling molecules and has long been implicated in tumorigenesis and tumor progression. In liver cancer, RIPK3, a key regulator of necroptosis, has been shown to promote HCC progression through activation of innate immunity [40, 41].Zhou et al. employed bioinformatic approaches to identify tumor antigens, including CARD8, NAIP, NLRP1, and NLRP3, based on pyroptosis-related genes, demonstrating the feasibility of developing mRNA vaccines from regulated necrosis-related genes[42].\u003c/p\u003e \u003cp\u003eDisulfidptosis, as a newly identified form of regulated necrosis, has been proposed as a novel therapeutic target in cancer. Based on transcriptomic data, this study screened DRGs significantly associated with HCC prognosis and identified LRPPRC and WASF2 as potential immunotherapy targets. Previous studies have also confirmed the critical role of LRPPRC in tumor development. liu et al. found that LRPPRC functions as a scaffold protein binding JAK2 and STAT3, enhancing JAK2\u0026ndash;STAT3 complex stability and thereby promoting JAK2/STAT3/MYC axis activation and esophageal squamous cell carcinoma progression. Disruption of the LRPPRC\u0026ndash;JAK2\u0026ndash;STAT3 and JAK2\u0026ndash;STAT3\u0026ndash;CDK1 interactions suppressed tumorigenesis in 4-nitroquinoline N-oxide-induced ESCC mouse models and inhibited tumor growth in patient-derived xenograft models [43]. Similarly, Yu et al. confirmed through Western blot and RT-qPCR that LRPPRC overexpression in triple-negative breast cancer enhances glycolysis and promotes tumor progression [44]. WASF2 has also been shown to play important roles in the development and progression of multiple cancers. In ovarian cancer, high WASF2 expression is closely associated with poor patient survival, with WASF2 promoting tumor cell invasion and migration, thereby exacerbating ovarian cancer malignancy [45]. In liver cancer, WASF2 overexpression promotes normal hepatocyte proliferation; inactivation of WASF2 by inducing G2/M phase arrest reduces the viability, growth, proliferation, migration, and invasion of Huh-7 and SNU475 liver cancer cells .\u003c/p\u003e \u003cp\u003eIn the present study, we employed an immunoinformatics-driven reverse vaccinology strategy to design a novel multi-epitope mRNA vaccine targeting the HCC-associated antigens WASF2 and LRPPRC. Through rigorous epitope prediction and multi-step screening, high-antigenicity B cell, CTL, and HTL epitopes were identified, and the top two candidates from each target protein were selected based on antigenicity scores for compact assembly. The final vaccine construct integrates the tPA signal peptide, GPGPG/KK/AAY linker peptides, and the MITD sequence. C-IMMSIM agent-based immune simulation predicted robust humoral and cellular responses, accompanied by a balanced pro-inflammatory and regulatory cytokine profile without cytokine storm. Physicochemical analysis confirmed favorable physicochemical properties with no allergenicity or toxicity. The incorporation of CTL, HTL, and B cell epitopes, combined with MITD optimization, is designed to overcome the immunosuppressive tumor microenvironment (TME) of HCC, including T cell exhaustion, low CD8\u0026thinsp;+\u0026thinsp;infiltration, and Treg/M2 macrophage dominance. C-IMMSIM demonstrated efficient post-booster antigen clearance, IgG persistence, and secondary immune cell activation peaks, suggesting the induction of long-term memory and protective immunity. The cytokine profile is consistent with an effective Th1-biased response, potentially capable of \u0026ldquo;heating\u0026rdquo; immunologically cold tumors, which is particularly critical for HCC. Compared with prior HCC multi-epitope designs targeting AFP, GPC3, or NY-ESO-1, our construct uniquely focuses on WASF2 and LRPPRC, potentially providing novel epitopes less susceptible to immune escape [10]. Rigorous screening and compact selection minimize off-target risks while maximizing immunogenicity, distinguishing it from broad-spectrum epitope constructs.\u003c/p\u003e \u003cp\u003eGiven that mRNA vaccines demonstrate efficacy only in specific cancer patient subpopulations, we stratified HCC patients into high-risk and low-risk subgroups based on disulfidptosis-related lncRNA expression profiles to identify the most promising vaccine candidate populations. Furthermore, because DRG markers encompass multiple genes and the tumor microenvironment is highly heterogeneous, their utility in predicting HCC prognosis has certain limitations. Accordingly, we established a DRI-based prognostic model.\u003c/p\u003e \u003cp\u003eAmong the prognostically relevant DRGs we identified, accumulating evidence indicates that the lncRNAs used to construct the DRI risk model play critical roles in cancer. For example, MIR210HG is associated with disease progression in multiple cancers. In endometrial cancer, it is enriched in the Wnt and β/Smad3 signaling pathways and promotes cancer development [46]; in breast cancer, MIR210HG induces cancer cell migration and invasion by regulating the epithelial\u0026ndash;mesenchymal transition process [47]. AL031985.3, AC108752.1, and AC016717.2 have also been demonstrated to exhibit significant differential expression between HCC and normal tissues and are significantly associated with patient prognosis [48, 49]. Our bioinformatic analyses similarly identified these lncRNAs as disulfidptosis-related prognostic markers. These findings deepen our understanding of the disulfidptosis-related tumor microenvironment in HCC and support the practical utility of the DRI model. To validate model performance, Kaplan\u0026ndash;Meier survival analysis and ROC curve analysis were performed in the study cohort. Results demonstrated that this signature possesses robust capability in predicting HCC prognosis. Furthermore, significant correlations were observed between DRI risk scores and immune cell infiltration as well as immunotherapy efficacy.\u003c/p\u003e \u003cp\u003eThe DRI risk score serves as an effective tool for characterizing the heterogeneity of tumor disulfidptosis regulatory patterns and identifying TME infiltration features. Comprehensive analysis confirmed that the risk score is an effective predictor of liver cancer prognosis. Results across the overall, training, and validation cohorts consistently demonstrated that patients in the low-risk group exhibited longer overall survival. ROC curve analysis indicated that this lncRNA-based signature model possesses high sensitivity and specificity, demonstrating excellent predictive performance. To enhance clinical applicability, a nomogram integrating DRI scores with pathological stage, clinical stage, age, and sex was constructed. Combined ROC analysis showed that this integrated model outperformed the DRI scoring system alone. Consequently, this nomogram can more accurately predict 1-year, 2-year, and 3-year survival rates in HCC patients, with a favorable C-index. Subsequently, we analyzed immune cell infiltration differences between high- and low-risk DRI groups. The low-risk group exhibited significantly higher immune cell types and quantities compared to the high-risk group, along with better survival rates, further supporting the stronger immune profile of the low-risk group.\u003c/p\u003e \u003cp\u003eTumor responsiveness to immune checkpoint therapy has been demonstrated to be closely associated with the degree of immune cell infiltration and tumor mutation burden. We investigated the relationship between DRI and immune checkpoint signals, and results indicated that disulfidptosis-related lncRNA expression can influence the efficacy of immune checkpoint therapy. Finally, our results revealed a significant correlation between DRI and TMB: the low-DRI risk group exhibited lower TMB levels and higher immune cell infiltration, resulting in superior survival prognosis compared to the high-DRI group.\u003c/p\u003e \u003cp\u003eIt is generally believed that tumors with high mutation burden have better prognosis with immunogenic cell death (ICD) therapy. However, if ICD efficacy is limited by the immunosuppressive microenvironment or stromal barriers in HCC, combining immunotherapy with tumor neoantigen mining to develop tumor antigen vaccines may help improve the efficacy of existing treatments in the DRI high-risk population. For example, BNT111, an mRNA vaccine targeting four common TAAs (NY-ESO-1, MAGE-A3, tyrosinase, and TPTE), induced new and enhanced pre-existing immune responses in over 90% of melanoma patients in a phase I clinical trial (NCT02410733) [50]. Developing tumor antigen vaccines in combination with existing cancer treatment modalities and DRI-based stratification strategies may represent a promising approach for achieving synergistic immune stimulation and significant clinical benefit.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, the feasibility and efficacy of the antigen vaccine have not been experimentally validated; due to the lack of RNA sequencing data and clinical data from HCC immunotherapy cohorts, further follow-up studies are needed to verify its potential value. Additionally, the model could be further optimized using machine learning and feature selection algorithms to improve liver cancer vaccine development and patient prognosis prediction. Although the in vitro simulation results are encouraging, this study is limited to computational predictions and lacks experimental validation (e.g., in vitro T cell ELISpot assays, in vivo HCC mouse models, or human PBMC stimulation). The slightly elevated instability index (45.63) suggests the need for further codon optimization or stabilizing mutations. Future work should include wet laboratory validation of epitope\u0026ndash;MHC binding (tetramer staining), LNP-mRNA expression in DCs, immunogenicity in HLA-transgenic mice, and synergy testing with ICIs. Clinical translation will require GMP manufacturing and phase I safety trials in patients with advanced HCC.\u003c/p\u003e \u003cdiv id=\"Sec32\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Conclusion\u003c/h2\u003e \u003cp\u003eThis study identified the disulfidptosis-related genes LRPPRC and WASF2 as potential tumor antigens for HCC mRNA vaccine development. Based on these two tumor antigens, a novel HCC mRNA vaccine was designed. Furthermore, we identified two disulfidptosis-related HCC subtypes and analyzed the differences in enriched pathways between the two disulfidptosis subtypes. We also constructed a novel disulfidptosis-related lncRNA-based scoring model that can be used to assess immunotherapy response and identify patients most suitable for vaccine treatment.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eAvailability of Data and Materials\u003c/p\u003e\n\u003cp\u003eThe datasets analyzed during the current study are publicly available. The transcriptomic and clinical data were obtained from The Cancer Genome Atlas (TCGA) repository (https://portal.gdc.cancer.gov/, TCGA-LIHC project) and the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/, accession number GSE76427). Protein sequence information for WASF2 (accession number Q9Y6W5) and LRPPRC (accession number P42704) was retrieved from the UniProt database (https://www.uniprot.org/). The KEGG pathway gene set (c2.cp.kegg.v7.4.symbols.gmt) was downloaded from the Molecular Signatures Database (MSigDB) repository (https://www.gsea-msigdb.org/gsea/msigdb). Immune simulations were performed using the C-IMMSIM platform (https://kraken.iac.rm.cnr.it/C-IMMSIM/). All data used in this study are publicly accessible from the above repositories.\u003c/p\u003e\n\u003cp\u003eAuthor Contributions\u003c/p\u003e\n\u003cp\u003eRZ and QL designed the research study. RZ and SL performed the research and analyzed the data. ML and WS provided help and advice on methodology and data interpretation. RZ wrote the manuscript. QL, SL, ML, and WS contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003eEthics Approval and Consent to Participate\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003eAcknowledgment\u003c/p\u003e\n\u003cp\u003eNot Applicable.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eNot Applicable.\u003c/p\u003e\n\u003cp\u003eConflict of Interest\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003eConsent to publish\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003eDeclaration of AI and AI-assisted Technologies in the Writing Process\u003c/p\u003e\n\u003cp\u003eDuring the preparation of this manuscript, the authors used Claude Opus 4.6 for the purpose of translating the original Chinese manuscript into English and refining the language to meet academic publishing standards. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024; 74: 229-263. https://doi.org/10.3322/caac.21834\u003c/li\u003e\n\u003cli\u003eN A, J A, A S, NN N, N A. Hepatocellular Carcinoma: A Comprehensive Review. Diseases (Basel, Switzerland). 2025; 13. https://doi.org/10.3390/diseases13070207\u003c/li\u003e\n\u003cli\u003eReveron-Thornton RF, Teng MLP, Lee EY, Tran A, Vajanaphanich S, Tan EX, et al. Global and regional long-term survival following resection for HCC in the recent decade: A meta-analysis of 110 studies. Hepatol Commun. 2022; 6: 1813-1826. https://doi.org/10.1002/hep4.1923\u003c/li\u003e\n\u003cli\u003eAG S, F K, JM L. Global trends in hepatocellular carcinoma epidemiology: implications for screening, prevention and therapy. Nature reviews. Clinical oncology. 2023; 20: 864-884. https://doi.org/10.1038/s41571-023-00825-3\u003c/li\u003e\n\u003cli\u003eRS F, S Q, M I, PR G, M D, TY K, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. The New England Journal of Medicine. 2020; 382: 1894-1905. https://doi.org/10.1056/NEJMoa1915745\u003c/li\u003e\n\u003cli\u003eAL C, S Q, M I, PR G, M D, TY K, et al. Updated efficacy and safety data from IMbrave150: atezolizumab plus bevacizumab vs. sorafenib for unresectable hepatocellular carcinoma. Journal of Hepatology. 2021. https://doi.org/10.1016/j.jhep.2021.11.030\u003c/li\u003e\n\u003cli\u003eOJ F, HG R. Is It Possible to Develop Cancer Vaccines to Neoantigens, What Are the Major Challenges, and How Can These Be Overcome? Neoantigens: Nothing New in Spite of the Name. Cold Spring Harbor perspectives in biology. 2018; 10. https://doi.org/10.1101/cshperspect.a028829\u003c/li\u003e\n\u003cli\u003eU S, P O, E D, RA J, M V, M G, et al. An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma. Nature. 2020; 585: 107-112. https://doi.org/10.1038/s41586-020-2537-9\u003c/li\u003e\n\u003cli\u003eFoerster F, Gairing SJ, Ilyas SI, Galle PR. Emerging immunotherapy for HCC: A guide for hepatologists. Hepatology. 2022; 75: 1604-1626. https://doi.org/10.1002/hep.32447\u003c/li\u003e\n\u003cli\u003eTojjari A, Saeed A, Singh M, Cavalcante L, Sahin IH, Saeed A. A Comprehensive Review on Cancer Vaccines and Vaccine Strategies in Hepatocellular Carcinoma. Vaccines (Basel). 2023; 11. https://doi.org/10.3390/vaccines11081357\u003c/li\u003e\n\u003cli\u003eI L, W W, R M, C R, F T, L H. In Situ Vaccination as a Strategy to Modulate the Immune Microenvironment of Hepatocellular Carcinoma. Frontiers in Immunology. 2021; 12: 650486. https://doi.org/10.3389/fimmu.2021.650486\u003c/li\u003e\n\u003cli\u003eMessmer MN, Snyder AG, Oberst A. Comparing the effects of different cell death programs in tumor progression and immunotherapy. Cell Death Differ. 2019; 26: 115-129. https://doi.org/10.1038/s41418-018-0214-4\u003c/li\u003e\n\u003cli\u003eKoppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell. 2021; 12: 599-620. https://doi.org/10.1007/s13238-020-00789-5\u003c/li\u003e\n\u003cli\u003eTang X, Chen W, Liu H, Liu N, Chen D, Tian D, et al. Research progress on SLC7A11 in the regulation of cystine/cysteine metabolism in tumors. Oncol Lett. 2022; 23: 47. https://doi.org/10.3892/ol.2021.13165\u003c/li\u003e\n\u003cli\u003eLiu X, Nie L, Zhang Y, Yan Y, Wang C, Colic M, et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol. 2023; 25: 404-414. https://doi.org/10.1038/s41556-023-01091-2\u003c/li\u003e\n\u003cli\u003eBeňačka R, Szab\u0026oacute;ov\u0026aacute; D, Guľa\u0026scaron;ov\u0026aacute; Z, Hertelyov\u0026aacute; Z, Radoňak J. Non-Coding RNAs in Human Cancer and Other Diseases: Overview of the Diagnostic Potential. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242216213\u003c/li\u003e\n\u003cli\u003eAhmadpour ST, Orre C, Bertevello PS, Mirebeau-Prunier D, Dumas JF, Desquiret-Dumas V. Breast Cancer Chemoresistance: Insights into the Regulatory Role of lncRNA. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242115897\u003c/li\u003e\n\u003cli\u003eWu L, Zhang L, Zheng S. Role of the long non-coding RNA HOTAIR in hepatocellular carcinoma. Oncol Lett. 2017; 14: 1233-1239. https://doi.org/10.3892/ol.2017.6312\u003c/li\u003e\n\u003cli\u003eChi H, Huang J, Yan Y, Jiang C, Zhang S, Chen H, et al. Unraveling the role of disulfidptosis-related LncRNAs in colon cancer: a prognostic indicator for immunotherapy response, chemotherapy sensitivity, and insights into cell death mechanisms. Front Mol Biosci. 2023; 10: 1254232. https://doi.org/10.3389/fmolb.2023.1254232\u003c/li\u003e\n\u003cli\u003eTang Z, Kang B, Li C, Chen T, Zhang Z. GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis. Nucleic Acids Res. 2019; 47: W556-W560. https://doi.org/10.1093/nar/gkz430\u003c/li\u003e\n\u003cli\u003eLi T, Fan J, Wang B, Traugh N, Chen Q, Liu JS, et al. TIMER: A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells. Cancer Res. 2017; 77: e108-e110. https://doi.org/10.1158/0008-5472.Can-17-0307\u003c/li\u003e\n\u003cli\u003eChen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018; 1711: 243-259. https://doi.org/10.1007/978-1-4939-7493-1_12\u003c/li\u003e\n\u003cli\u003eNewman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015; 12: 453-457. https://doi.org/10.1038/nmeth.3337\u003c/li\u003e\n\u003cli\u003eWilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010; 26: 1572-1573. https://doi.org/10.1093/bioinformatics/btq170\u003c/li\u003e\n\u003cli\u003eH\u0026auml;nzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013; 14: 7. https://doi.org/10.1186/1471-2105-14-7\u003c/li\u003e\n\u003cli\u003eYoshihara K, Shahmoradgoli M, Mart\u0026iacute;nez E, Vegesna R, Kim H, Torres-Garcia W, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013; 4: 2612. https://doi.org/10.1038/ncomms3612\u003c/li\u003e\n\u003cli\u003eSaha S, Raghava GP. Prediction of continuous B-cell epitopes in an antigen using recurrent neural network. Proteins. 2006; 65: 40-48. https://doi.org/10.1002/prot.21078\u003c/li\u003e\n\u003cli\u003eYan Z, Kim K, Kim H, Ha B, Gambiez A, Bennett J, et al. Next-generation IEDB tools: a platform for epitope prediction and analysis. Nucleic Acids Res. 2024; 52: W526-W532. https://doi.org/10.1093/nar/gkae407\u003c/li\u003e\n\u003cli\u003eReynisson B, Alvarez B, Paul S, Peters B, Nielsen M. NetMHCpan-4.1 and NetMHCIIpan-4.0: improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data. Nucleic Acids Res. 2020; 48: W449-W454. https://doi.org/10.1093/nar/gkaa379\u003c/li\u003e\n\u003cli\u003eDoytchinova IA, Flower DR. VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines. BMC Bioinformatics. 2007; 8: 4. https://doi.org/10.1186/1471-2105-8-4\u003c/li\u003e\n\u003cli\u003eDimitrov I, Bangov I, Flower DR, Doytchinova I. AllerTOP v.2--a server for in silico prediction of allergens. J Mol Model. 2014; 20: 2278. https://doi.org/10.1007/s00894-014-2278-5\u003c/li\u003e\n\u003cli\u003eRathore AS, Choudhury S, Arora A, Tijare P, Raghava GPS. ToxinPred 3.0: An improved method for predicting the toxicity of peptides. Comput Biol Med. 2024; 179: 108926. https://doi.org/10.1016/j.compbiomed.2024.108926\u003c/li\u003e\n\u003cli\u003eMagnan CN, Zeller M, Kayala MA, Vigil A, Randall A, Felgner PL, et al. High-throughput prediction of protein antigenicity using protein microarray data. Bioinformatics. 2010; 26: 2936-2943. https://doi.org/10.1093/bioinformatics/btq551\u003c/li\u003e\n\u003cli\u003eGasteiger E, Gattiker A, Hoogland C, Ivanyi I, Appel RD, Bairoch A. ExPASy: The proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 2003; 31: 3784-3788. https://doi.org/10.1093/nar/gkg563\u003c/li\u003e\n\u003cli\u003eRapin N, Lund O, Bernaschi M, Castiglione F. Computational immunology meets bioinformatics: the use of prediction tools for molecular binding in the simulation of the immune system. PLoS One. 2010; 5: e9862. https://doi.org/10.1371/journal.pone.0009862\u003c/li\u003e\n\u003cli\u003eSimon N, Friedman J, Hastie T, Tibshirani R. Regularization Paths for Cox's Proportional Hazards Model via Coordinate Descent. J Stat Softw. 2011; 39: 1-13. https://doi.org/10.18637/jss.v039.i05\u003c/li\u003e\n\u003cli\u003eP C, F F, M A, C M, M E, D R, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell reports. 2017; 18: 248-262. https://doi.org/10.1016/j.celrep.2016.12.019\u003c/li\u003e\n\u003cli\u003eJiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018; 24: 1550-1558. https://doi.org/10.1038/s41591-018-0136-1\u003c/li\u003e\n\u003cli\u003eLorentzen CL, Haanen JB, Met \u0026Ouml;, Svane IM. Clinical advances and ongoing trials on mRNA vaccines for cancer treatment. Lancet Oncol. 2022; 23: e450-e458. https://doi.org/10.1016/s1470-2045(22)00372-2\u003c/li\u003e\n\u003cli\u003eTayir M, Wang YW, Chu T, Wang XL, Fan YQ, Cao L, et al. The function of necroptosis in liver cancer. Biochim Biophys Acta Mol Basis Dis. 2025; 1871: 167828. https://doi.org/10.1016/j.bbadis.2025.167828\u003c/li\u003e\n\u003cli\u003eAfonso MB, Rodrigues PM, Mateus-Pinheiro M, Sim\u0026atilde;o AL, Gaspar MM, Majdi A, et al. RIPK3 acts as a lipid metabolism regulator contributing to inflammation and carcinogenesis in non-alcoholic fatty liver disease. Gut. 2021; 70: 2359-2372. https://doi.org/10.1136/gutjnl-2020-321767\u003c/li\u003e\n\u003cli\u003eF Z, M W, Z W, W L, X L. Screening of novel tumor-associated antigens for lung adenocarcinoma mRNA vaccine development based on pyroptosis phenotype genes. BMC Cancer. 2024; 24: 28. https://doi.org/10.1186/s12885-023-11757-7\u003c/li\u003e\n\u003cli\u003eH L, Y Z, M F, P N, Y G, R W, et al. Targeting leucine-rich PPR motif-containing protein/LRPPRC by 5,7,4'-trimethoxyflavone suppresses esophageal squamous cell carcinoma progression. International journal of biological macromolecules. 2024; 269: 131966. https://doi.org/10.1016/j.ijbiomac.2024.131966\u003c/li\u003e\n\u003cli\u003eYu Y, Deng H, Wang W, Xiao S, Zheng R, Lv L, et al. LRPPRC promotes glycolysis by stabilising LDHA mRNA and its knockdown plus glutamine inhibitor induces synthetic lethality via m 6 A modification in triple-negative breast cancer. Clin Transl Med. 2024; 14: e1583. https://doi.org/10.1002/ctm2.1583\u003c/li\u003e\n\u003cli\u003eYang X, Ding Y, Sun L, Shi M, Zhang P, He A, et al. WASF2 Serves as a Potential Biomarker and Therapeutic Target in Ovarian Cancer: A Pan-Cancer Analysis. Front Oncol. 2022; 12: 840038. https://doi.org/10.3389/fonc.2022.840038\u003c/li\u003e\n\u003cli\u003eJ M, FF K, D Y, H Y, C W, R C, et al. lncRNA MIR210HG promotes the progression of endometrial cancer by sponging miR-337-3p/137 via the HMGA2-TGF-\u0026beta;/Wnt pathway. Molecular therapy. Nucleic acids. 2021; 24: 905-922. https://doi.org/10.1016/j.omtn.2021.04.011\u003c/li\u003e\n\u003cli\u003eShi W, Tang Y, Lu J, Zhuang Y, Wang J. MIR210HG promotes breast cancer progression by IGF2BP1 mediated m6A modification. Cell Biosci. 2022; 12: 38. https://doi.org/10.1186/s13578-022-00772-z\u003c/li\u003e\n\u003cli\u003eGuo Z, Xie Y, Zhang L, Liu S, Jiang W. A novel disulfidptosis-related lncRNAs signature for predicting survival and immune response in hepatocellular carcinoma. Aging (Albany NY). 2024; 16: 267-284. https://doi.org/10.18632/aging.205367\u003c/li\u003e\n\u003cli\u003eJia X, Wang Y, Yang Y, Fu Y, Liu Y. Constructed Risk Prognosis Model Associated with Disulfidptosis lncRNAs in HCC. Int J Mol Sci. 2023; 24. https://doi.org/10.3390/ijms242417626\u003c/li\u003e\n\u003cli\u003eSahin U, Oehm P, Derhovanessian E, Jabulowsky RA, Vormehr M, Gold M, et al. An RNA vaccine drives immunity in checkpoint-inhibitor-treated melanoma. Nature. 2020; 585: 107-112. https://doi.org/10.1038/s41586-020-2537-9\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"hepatocellular carcinoma, disulfidptosis, mRNA vaccine, lncRNA, prognostic model, tumor microenvironment, immunotherapy","lastPublishedDoi":"10.21203/rs.3.rs-9072846/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9072846/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eHepatocellular carcinoma (HCC) remains one of the most lethal malignancies worldwide, yet its molecular heterogeneity continues to limit the efficacy of systemic therapies. Disulfidptosis, a recently characterized form of regulated cell death triggered by disulfide stress under glucose deprivation in SLC7A11-overexpressing cells, represents a promising but poorly characterized therapeutic target in HCC. Here, we established an integrated disulfidptosis-based framework encompassing molecular subtyping, mRNA vaccine design, and prognostic modeling. Transcriptomic data from TCGA-LIHC and GEO datasets were integrated after batch-effect correction. Thirty-two disulfidptosis-related genes (DRGs) defined two molecularly distinct HCC subtypes with significant differences in survival outcomes, immune cell infiltration, and pathway enrichment. Through mutation profiling and survival analysis, tumor-associated antigens WASF2 and LRPPRC were identified as optimal vaccine targets, both exhibiting high mutation frequencies, adverse prognostic associations, and positive correlation with antigen-presenting cell abundance. A multi-epitope mRNA vaccine incorporating B-cell, CTL, and HTL epitopes was computationally designed and validated for strong antigenicity, non-allergenicity, non-toxicity, and favorable physicochemical properties; in silico immunosimulation predicted robust humoral and cellular immune responses. Furthermore, a disulfidptosis-related lncRNA prognostic index (DRI) was constructed via LASSO-Cox regression, effectively stratifying patients by prognosis (AUC\u0026thinsp;\u0026gt;\u0026thinsp;0.7) and predicted responsiveness to immune checkpoint inhibitors. This multi-dimensional framework linking disulfidptosis biology to HCC subtype classification, immunotherapy-oriented vaccine development, and patient stratification offers a compelling foundation for advancing precision immunotherapy in HCC.\u003c/p\u003e","manuscriptTitle":"Integrated analysis identifies disulfidptosis related tumor antigens and molecular subtypes in hepatocellular carcinoma for mRNA vaccine development","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-05 17:09:36","doi":"10.21203/rs.3.rs-9072846/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-04-27T11:57:36+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-08T01:30:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"103137259349064744403264483811396763021","date":"2026-04-08T01:17:04+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-07T13:36:09+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-07T09:15:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"264668410602411409374993812584476731386","date":"2026-04-07T03:21:39+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-06T18:05:19+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"32445979626889827391274037762449488271","date":"2026-04-02T08:47:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"308548153704380107719556204390772693214","date":"2026-03-31T10:01:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-31T08:31:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-17T05:37:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-12T15:23:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oncology","date":"2026-03-12T09:55:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f9b338de-8d4f-4ded-9087-fcdfe826172e","owner":[],"postedDate":"April 5th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-04-27T12:09:49+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-05 17:09:36","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9072846","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9072846","identity":"rs-9072846","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}