Abstract Cuproptosis was characterized as a novel type of programmed cell death. Recently, however, the role of cuproptosis-related long noncoding RNAs (CRLs) in tumors has not yet been studied. Identifying a predictive CRL signature in hepatocellular carcinoma (HCC) and investigating its putative molecular function were the goals of this work. Initially, Pearson’s test was used to assess the relationship between lncRNAs and cuproptosis-associated genes obtained from HCC data of The Cancer Genome Atlas (TCGA). By implementing differential expression and univariate Cox analysis, 61 prognostic CRLs were subsequent to the least absolute shrinkage and selection operator (LASSO) Cox regression analysis. A prognostic risk score model was then constructed to evaluate its ability to predict patients’ survival when combined with clinicopathological parameters in HCC. The five-lncRNA prognostic signature categorized the HCC patients into high- and low-risk groups. The low-risk group exhibited more sensitivity to elesclomol than the high-risk one. Surprisingly, distinct mitochondrial metabolism pathways connected to cuproptosis and pivotal immune-related pathways were observed between the two groups via gene set enrichment analysis (GSEA). Meanwhile, there were substantial differences between the high-risk group and the low-risk group in terms of tumor-infiltrating immune cells (TIICs). Furthermore, a positive relationship was shown between the risk score and the expression of immune checkpoints. Additionally, differential expression of the five lncRNAs was confirmed in our own HCC samples and cell lines via RT-qPCR. Finally, in vitro assays confirmed that WARS2-AS1 and MKLN1-AS knockdown could sensitize HCC cells to elesclomol-induced cuproptosis. Overall, our predictive signature may predict the prognosis of HCC patients in an independent manner, give a better understanding of how CRLs work in HCC, and offer therapeutic reference for patients with HCC. Keywords: lncRNA, hepatocellular carcinoma, immune microenvironment, survival analysis, cuproptosis Introduction Hepatocellular carcinoma remains one of the most common malignancies globally, ranking as the third cause of cancer-related mortality and sixth with regard to the incidence of all tumor types ([37]1). A number of risk factors for the occurrence of HCC include hepatitis B virus (HBV), hepatitis C virus (HCV), alcoholic liver disease, and nonalcoholic steatohepatitis ([38]2). Due to the low rate of early diagnosis in HCC, the majority of cases are not detected until they have progressed to an advanced stage of cancer ([39]3). Over the past decades, many different treatment modalities have been developed to combat HCC, but high post-operative recurrence rates and drug resistance continue to be barriers to cure HCC ([40]4, [41]5). As a result, a thorough understanding of the networks involved in HCC development and progression is essential for improving detection efficiency and developing more effective therapies. It is known that copper (Cu) is a cofactor for enzymes that control a variety of vital cellular activities, including mitochondrial respiration, antioxidant defense, and the manufacture of hormones ([42]6). According to recent findings, although Cu is increasingly implicated in cell proliferation ([43]7), dysregulation of Cu stores can also induce cytotoxicity via multiple pathways. One mechanism proposed by Masazumi Nagai et al. demonstrated that the Cu-binding drug elesclomol preferentially chelated Cu outside of cells and selectively transported the Cu to mitochondria as elesclomol-Cu (II), thereby triggering mitochondrial reactive oxygen species (ROS) induction ([44]8). Another Cu chelators disulfiram (DSF) formed a DSF/Cu complex with Cu to severely disrupt mitochondrial homeostasis and increase the free iron pool, ultimately provoking lipid peroxidation and causing ferroptotic cell death ([45]9). However, these traditional views of copper-induced cytotoxicity have been challenged by emerging evidence that Cu-dependent death occurs via direct binding of Cu to lipoylated components of the TCA cycle, which leads to proteotoxic stress and ultimately cell death ([46]10). In addition, treatment with inhibitors of other well-known cell death pathways, such as apoptosis (Z-VAD-FMK), ferroptosis (ferrostatin-1), necroptosis (necrostatin-1), and oxidative stress (N-acetyl cysteine), could not reverse Cu ionophore-induced cell death. Therefore, scientists proposed this previously uncharacterized cell death mechanism as cuproptosis. So far, there have only been a few cuproptosis-related genes discovered in cancer. It is urgent for us to find novel regulators of cuproptosis for the purpose of improving the diagnosis and treatment of cancer. In recent years, advances in sequencing technologies have led to the discovery of a multitude of non-coding RNA (ncRNA) species, which are a class of RNAs lacking potential to encode proteins. Pervasive transcription produces a vast repertoire of ncRNAs of all sizes and shapes, including short ncRNAs (such as microRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (cirRNAs) ([47]11). ncRNA containing more than 200 nucleotides was defined as lncRNA, which can act as miRNA sponges, RNA-binding protein sequestering factors, as well as regulators of gene expression by controlling mRNA transcription ([48]12). LncRNAs have been implicated in tumorigenesis-associated biological functions such as metastasis and programmed death, according to a growing body of research over the last several decades ([49]13, [50]14). Despite the fact that lncRNAs have been studied extensively in other types of cell death ([51]15), no study on cuproptosis-associated lncRNAs has yet been documented. It is thus a great challenge for us to explore how lncRNAs work during the process of cuproptosis especially in cancer. Intriguingly, studies have also shown that cell death-related lncRNA could exert a pivotal role in the regulation of immune cells. For example, three lncRNA A2M-AS1, C2orf27A, and ZNF667-AS1 were identified as the upstream transcriptional regulators of several hub ferroptosis-associated genes (FAGs) while these FAGs had a significant effect on immune cell infiltration in gastric cancer, indicating that lncRNA might affect immune response via mediating ferroptosis process ([52]16). Moreover, another literature has reported that the construction of a ferroptosis-related lncRNA model could contribute to the immune status and response to immunotherapy of lung cancer, which built a link between cell death-related lncRNAs and immune regulation ([53]17). Such findings have provided us with new perspectives on the investigation of the relationship between the cuproptosis-related lncRNAs and tumor-associated pathways in the future. In this work, we developed a predictive model based on cuproptosis-associated lncRNAs in HCC. The performance of predicting survival on the basis of the signature-based risk score was analyzed together with other standard clinicopathological parameters. In addition, its significance in assessing the sensitivity of cuproptosis inducer elesclomol has been further evaluated between the high- and low- risk groups. Moreover, internal cohorts were then carried out to verify the results above. Further investigation into the mechanism of action of CRLs in HCC was conducted using gene set enrichment analysis (GSEA). and immune infiltration analysis. Finally, the role of three lncRNAs in regulating cuproptosis was validated via in vitro experiments. Overall, our findings provided valuable clues into the underlying mechanisms of cuproptosis in HCC and may help predict patients’ survival more accurately. Methods Patients and datasets The fragments per kilobase of transcript per million mapped reads (FPKM)-standardized RNA-seq data of 424 samples, including 50 normal hepatic tissues and 374 tumor samples, and the corresponding clinical and prognostic data were downloaded from the TCGA website ([54]https://portal.gdc.cancer.gov/projects/TCGA-LIHC). Patients with unknown clinical information and an overall survival time of less than 30 days were excluded. Then, Ensembl IDs were processed and converted to official gene symbol, including lncRNAs, protein-coding genes, miRNAs, etc. Identification of cuproptosis-related lncRNAs To identify the CRLs, a total of 13 cuproptosis-related genes (CRGs) were summarized from recently published literature (10) ([55] Supplementary Table S1 ). According to previous documents, Pearson analysis was considered an accepted method to investigate the correlation between coding genes and lncRNAs in the RNA-seq data of TCGA HCC samples ([56]18, [57]19). Before obtaining enough cuproptosis-related lncRNAs, we have set various R values based on the published documents ([58]18, [59]20, [60]21) and chosen R > 0.4 and P < 0.001 as the best cutoff value eventually. Then, the associations between CRLs and CRGs were initially filtered. Differential expression analysis of lncRNAs The expression levels of CRLs between HCC and normal hepatic tissues were examined using the Wilcoxon test. A false discovery rate (FDR) < 0.01 and |Fold change (FC)| > 2 were set as screening criteria to obtain differentially expressed lncRNAs (DELs). The construction of the co-expression network The co-expression network between CRLs and CRGs was constructed by Cytoscape software (version 3.7.2). Then, the ggalluvial R software package was used to plot a Sankey diagram in order to demonstrate the degree of correlation between CRLs (risk/protect) and their corresponding CRGs. The construction of cuproptosis-related prognostic signature By using the ‘survival’ R package and defining p < 0.05 as screening criteria, the intersecting lncRNAs of CRLs and DELs were subsequent to univariate cox analysis for obtaining prognostic CRLs in HCC patients. least absolute shrinkage and selection operator (LASSO) Cox regression analysis was applied to construct CRLs predictive signature ([61]22, [62]23). Initially, 18 prognostic lncRNAs were screened out on the basis of the optimal penalty parameter λ determined by tenfold cross-validation following the minimum criteria. Afterwards, multivariate cox regression analysis was conducted for the establishment of a five-lncRNA predictive model. The computational formula used for cuproptosis-related prognostic risk score was as follows: Risk score = Coef[i] lncRNA1 × lncRNA1 expression + Coef[i] lncRNA2 × lncRNA2 expression + · ···· +Coef[i] lncRNAn × lncRNAn expression. Coef[i] represents the coefficient value of the corresponding lncRNA. Based on the median value of the risk score, patients were divided into low-risk and high-risk groups. The R package ‘survminer’ was used to generate the Kaplan–Meier curve with a log-rank test to compare the prognostic significance of cuproptosis-related five-lncRNA risk model. To assess the predictive ability of the lncRNA-based prognostic risk signature, the R package “timeROC” was used to examine the receiver operating characteristic (ROC) of 1/3/5-year survival ([63]24). Moreover, univariate and multivariate Cox regression methods were performed to evaluate the prognostic prediction power of this risk score model. Construction of nomogram A nomogram for predicting the 1-, 3-, and 5-year survival of HCC patients was developed using the risk model in conjunction with clinicopathological parameters such as age, gender, grade, stage, metastasis (M), positive lymph node (N) and vascular invasion. To determine if the anticipated survival rate was congruent with the observed survival rate, we employed a calibration curve. Function enrichment analysis For the purpose of investigating the molecular mechanism and biological process involved in the cuproptosis-related lncRNA signature, GSEA was performed to discover which pathway genes were mainly enriched between the high/low risk groups using the h.all.v7.4 symbols.gmt [Hallmarks], and the Kyoto Encyclopedia of Genes and Genomes (KEGG) dataset c2.cp.kegg.v7.4.symbols.gmt from the molecular signature dataset ([64]https://www.gsea-msigdb.org/gsea/msigdb) as references ([65]25).