An Innovative Ferroptosis-related Genes Signature for Prediction of Prognosis and Immune Microenvironment in Patients with Hepatocellular Carcinoma

doi:10.21203/rs.3.rs-2315131/v1

Download PDF

Article

An Innovative Ferroptosis-related Genes Signature for Prediction of Prognosis and Immune Microenvironment in Patients with Hepatocellular Carcinoma

https://doi.org/10.21203/rs.3.rs-2315131/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Hepatocellular carcinoma (HCC) is a lethal malignancy with high heterogeneity, which impedes prognostic prediction. Cuproptosis is an innovative copper-trigged modality of mitochondrial cell death. Ferroptosis is a unique form of programmed cell death with iron-dependent phospholipid peroxidation. The target of this study was to establish a new signature with ferroptosis-related genes conjoining cuproptosis-related genes for the prediction of prognosis and tumor immunity in patients with HCC.

Method: The expression profiles and corresponding clinical data of HCC patients were retrieved from TCGA. Pearson correlation analysis was conducted to identify the cuproptosis-related ferroptosis genes (CRFGs). A risk prognostic signature of 5-gene was constructed based on the CRFGs via the Cox regression and the least absolute shrinkage and selection operator (LASSO) method. Subsequently, the prognostic performance of the signature was evaluated by survival analysis, nomogram, Cox regression, clinicopathological characteristics correlation analysis, and the receiver operating characteristic (ROC) curve. Furthermore, the GO, KEGG pathway, GSEA, protein-protein interaction (PPI), the correlation between the signature risk score and immune cell infiltration, immune checkpoint, immune function, TMB, TIDE score, ESTIMATE score, and anticancer drug sensitivity were analyzed.

Results: The 5-CRFGs signature comprising ATG13, BAP1, ELAVL1, SLC38A1, and YY1AP1 were explored to ameliorate the prognosis prediction of HCC patients. Kaplan-Meier analysis and ROC curve revealed the high-risk group associated with a poorer prognosis compared to the low-risk group. Cox regression and stratified survival analysis demonstrated that this signature was a risk factor independent of various clinical parameters. Moreover, immune cell infiltration, immune function, immune checkpoint, increased TMB, decreased TIDE score, ESTIMATE score, and half-inhibitory concentration significantly differed between high-and low-risk subtypes, which implied that the signature had acceptable assessing potency in the clinical efficacy of chemotherapy and immunotherapy.

Conclusion: The novel cuproptosis-related ferroptosis gene signature indicated extraordinary predictive potency in prognosis and the immune microenvironment, which can provide a potentially promising therapeutic strategy for HCC patients.

Biological sciences/Cancer

Biological sciences/Immunology

Health sciences/Diseases

Health sciences/Gastroenterology

cuproptosis

ferroptosis

hepatocellular carcinoma

prognosis

biomarkers

immune microenvironment

Targeting regulated cell death (RCD) is a revolutionary tumor treatment modality, and its regulation in hepatocellular carcinoma remains unknown. The purpose of this study was to figure out the potential molecular biological roles of cuproptosis-related ferroptosis genes in hepatocellular carcinoma. The 5-CRFGs signature was successfully constructed and its value in predicting prognosis based on the clinicopathological features was identified series of analyses. Furthermore, we performed validation of the correlation between the signature risk score and immune cell infiltration, immune checkpoint, immune function, TMB, TIDE score, ESTIMATE score, and anticancer drug sensitivity. In this paper, the 5-CRFGs model was developed with promising application prospects for predicting the clinical outcomes and treatment strategies in hepatocellular carcinoma.

Liver cancer is one of the lethal gastrointestinal malignancies with a dire prognosis due to its delayed diagnosis, lacking effective treatment, high recurrence, and obstinate resistance to chemotherapy (Anwanwan et al., 2020). Epidemiologically, Hepatocellular carcinoma (HCC), accounting for approximately 90% of primary liver cancer, represents the fourth-leading cause of cancer-associated death and ranks sixth in terms of incidence among malignancies worldwide (Villanueva, 2019). Although genetics and aging are considered irresistible risk factors for this fatal disease, several modifiable risk factors include chronic hepatitis B or C virus infection, nonalcoholic fatty liver disease, alcohol abuse, smoking, obesity, iron overload, and various dietary exposures (Center and Jemal, 2011). HCC is endowed with refractoriness owing to its highly heterogeneous character documented in interpatient, intertumoral, and intratumoral levels (Craig et al., 2020). The pivot to harvest an improved prognosis for patients with HCC is to explore new diagnostic and prognostic biomarkers, establish a convenient and effective prediction system, and actively conduct a scientific clinical intervention.

Ferroptosis, an original form of programmed cell death with its distinguishing morphological, biochemical, pathophysiological, genetical, and functional characteristics, is distinct from other types of cell death such as apoptosis, various forms of necrosis, pyroptosis, and autophagy to name a few(Chen et al., 2021). Numerous cellular metabolic pathways swarm into the hub of this unique cell death process–the lethal accumulation of iron-dependent phospholipid peroxidation(Hadian and Stockwell, 2020). Mounting shreds of evidence from miscellaneous scientific research fields have witnessed the great potential of ferroptosis as a silver bullet for solid tumors. Ferroptosis, the nonapoptotic cell death, can provide an alternative strategy for HCC cells characterized by apoptosis resistance.

Recently, cuproptosis, first coined by Tsvetkov et al, and published in Science, is an innovative copper-trigged modality of mitochondrial cell death(Tsvetkov et al., 2022). By analyzing phenomena, mechanisms, and disease models, the researchers found that copper-induced cell death occurs is closely associated with protein lipoylation concentrated during the tricarboxylic acid (TCA) cycle for enzymatic function. This results in the accumulation of lipoylated proteins, the loss of iron-sulfur clusters, and the induction of heat shock protein 70 (HSP 70) which triggers proteotoxic stress and ultimately cell death(Tsvetkov et al., 2022).

Here, our present research work aimed to perform a comprehensive prognostic model for HCC analysis based on cuproptosis-related ferroptosis differentially expressed genes that could be utilized for prognostic prediction and selection of patients for further personalized therapies.

2.1 Dataset Acquisition and Processing

The public mRNA expression and corresponding clinical information of a total of 424 samples, including 374 HCC tissues and 50 normal hepatic tissues were obtained from the public database base up to May 23, 2022 (the Cancer Genome Atlas, TCGA, https://portal.gdc.cancer.gov/). Ferroptosis-related genes (FRGs) were downloaded from the public database (FerrDb, http://www.zhounan.org/ ). A total of 259 ( Diver: 108; Suppressor: 69; Marker: 111 ) FRGs were amalgamated and those duplicated genes were selectively deleted. Cuproptosis-related genes (CRGs) including FDX1, LIAS, LIPT1, DLD, DLAT, PDHA1, PDHB, MTF1, GLS, CDKN2A, SLC31A1, ATP7A, and ATP7B were retrieved from the scientific research ( Science. 2022 Mar18; 375(6586): 1254–1261). According to the criteria of the correlation coefficient, the Cuproptosis-related Ferroptosis genes (CRFGs) were identified and extracted by relevant R packages based on R software. Since these data were all available online with usage allowance, an extra ethical approval was not necessary.

2.2 Construction and Validation of Risk Score Based on CRFGs

We constructed a univariate Cox regression analysis to further screen the prognostic value of CRGs. Preliminarily screened prognosis-related CRGs were further narrowed down utilizing the least absolute shrinkage and selection operator (LASSO) analysis through the R package of glmnet. We calculated the risk scores through the following tool: Risk score = (Coef1* expression mRNA1) +( Coef2* expression mRNA2) +…+ (Coef n* expression mRNA n), in which Coef represents lasso Cox regression model coefficient of the corresponding mRNA. On the basis of the prognostic model, the prognostic risk score of each pancreatic cancer patient was calculated and the medium was delimitated as the edge between high-risk and low-risk groups. We performed survival analysis by Kaplan-Meier curve to evaluate the prognosis in two groups and applied time-related ROC analysis to evaluate the prognostic ability of the risk model via survival ROC package.

2.3 Enrichment analyses and protein-protein interaction (PPI)

The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed through the Cluster Profiler package. Gene Ontology (GO) analysis, including molecular function (MF), cellular components (CC), and biological process (BP) was conducted with the same method. We investigated the underlying molecular mechanisms among high-risk and low-risk groups by conducting Gene Set Enrichment Analysis (GSEA). False discovery rate (FDR) and p < 0.05 were regarded as significantly enriched. The screened CRFGs were submitted to the STRING database (https://string-db.org) in order to acquire protein-protein interaction information. We implemented the Cytohubba plug-in to select the top 10 significant modules of the PPI network and visualized them by operating Cytoscape software.

2.4 Nomogram Establishment and calibration

To research the relationship between CRFGs-based signature and clinical variables, we launched univariate and multivariate Cox regression analysis to explore whether risk scores were provided with an independent prognostic value in pancreatic patients. We established a nomogram associated with outcome for assessing the probability of 1-, 2-, and 3-year Overall Survival (OS) for pancreatic cancer patients. We performed the concordance index (C-index) and calibration curve to evaluate the predictive utility of the accomplished nomogram.

2.5 Tumor Mutational Burden and Tumor Immune Analysis

To understand the tumor mutational burden (TMB) of patients with HCC, the R package of “maftools” was utilized to examine and integrate the TCGA data. The original mutation annotation format (MAF) was divided into two groups according to the risk score. Then, on the basis of the somatic mutation data for each patient in the two groups, TMB score was calculated. We utilized the CIBERSORT(Newman et al., 2015) to extract the relative proportions of 22 types of human-infiltrated immune cells correlating between the risk signature and immune-cell characteristics. Additionally, the enrichment levels of 29 immune-related functions between the two groups were evaluated by single sample Gene Set Enrichment Analysis (ssGSEA) scores. Furthermore, with ESTIMATE R package(Yoshihara et al., 2013), we ciphered out the tumor microenvironment (TME) score, including stromal score, immune score, and estimate score between high- and low-risk groups. By the tumor immune dysfunction and exclusion (TIDE) algorithm(Jiang et al., 2018), the potential immune checkpoint blockade (ICB) response was assessed.

2.6 Profiles of CRFGs Identified Three Distinct Clusters of HCC

By employing hierarchical agglomerative clustering based on Euclidean distance and Ward’s linkage(Wilkerson and Hayes, 2010), HCC patients with qualitatively varying CRFGs expressions were clustered. The K-means method was used to classify patients for further research by applying the “ConsensusClusterPlus” package.

2.7 Prediction of Potential Compounds in the Treatment of Hepatocellular Carcinoma

To explore the sensitivity difference of drugs between high- and low-risk groups, the Genomics of Drug Sensitivity in Cancer (GDSC, https://www.cancerrxgene.org/) database to analyze the half-maximal inhibitory concentration (IC50) of drugs for predicting the drug sensitivity by using the R package. p < 0.05 was considered statistically significant.

2.8 Quantitative Real-Time Polymerase Chain Reaction

Human hepatocellular carcinoma cell line MHCC-97H, Hep-G2, HCC-LM3, PLC-PRF-5, and human normal hepatic cell line HL-7702 were incubated in 37℃ and 5% CO2 circumstance. The total RNA of these cells was extracted and purified by an RNA-easy Isolation Reagent (Vazyme, Nanjing, PRC). The reverse transcription was completed through T100 Thermal Cycler (BIO-RAD, United States) and qRT-PCR was finished via PerfectStat Green qPCR SuperMix (TransGen, Beijing, PRC) as well as Quantagene q225 (Kubotechnology, Beijing, PRC). All experimental operations followed the steps of the product manual. The primer sequences of PCR amplification were as follows:

ATG13, forward: 5’-AGACAGTTCGTGTTGGGACAG-3’

reverse: 5’-CTCAAATTGCCTGGTAGACATGA-3’

BAP1, forward: 5’-GATACGTCCGTGATTGATGATGA-3’

reverse: 5’-TGAGTTGCACAAGAGTTGGGTA-3’

SLC38A1, forward: 5’-AACCTCCTTAGGCATGTCTGT-3’

reverse: 5’-GCAAAGGCGAGTCCCAAAAT-3’

ELAVL1, forward: 5’-AACTACGTGACCGCGAAGG-3’

reverse: 5’-CGCCCAAACCGAGAGAACA-3’

YY1AP1, forward: 5’-CTTCCCTGTACTCTCATCGCT-3’

reverse: 5’-CCTTCGCTATCTGATGTTGTTCA-3’

β-catenin, forward: 5’-TGATGGTGGGAATGGGTCAG-3’

reverse: 5’-GGTGTGGTGCCAGATCTTCT-3’

β-catenin was designed as the internal control and each sample was replicated three times to ensure the reliability of experimental results. The relative expression level was determined by the 2^-ΔΔCt method. The differences in ATG13, BAP1, SLC38A1, ELAVL1, and YY1AP1 expression were tested by t-test. Subsequently, the data visualization was accomplished via GraphPad Prism (version 8.02).

2.9 Statistical Analysis

R software (version 4.1.1) was applied to conduct the data analysis and visualization. The Wilcoxon test was utilized for the comparison of data between two groups, while the ANOVA test for the comparison of data among three groups. The prognostic prediction genes for constructing the prognostic model were screened through multivariate COX regression analysis. The Kaplan-Meier method with a two-sided log-rank test was applied for survival analysis. A p-value < 0.05 was considered statistically significant (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001). The flow chart of this study is shown in Fig. 1. The detailed clinical characteristics of a total of 374 HCC from the TCGA-LIHC dataset are summarized in Table 1.

3.1. Subsection

3.1.1. Construction and validation of CRFG-based signature

A total of 52 cuproptosis-related ferroptosis genes were sought out from the ferroptosis genes following the criteria of the absolute value of correlation coefficient more than 0.5 and P value less than 0.05. Compared with the normal haptic tissues, a total of 17 CRFGs were identified as differently expressed CRFGs in the TCGA-LIHC dataset (Fig. 2A). We gained 11 genes of them provided with the prognostic value by utilizing univariate Cox regression analysis based on these differentially expressed genes (Fig. 2B). Then, LASSO Cox regression analysis was applied to establish a prognostic model with the expression profile of the 11 genes previously obtained. A 5-gene signature was identified based on the optimal value of λ (Fig. 2C-D). A risk model was successfully constructed with 5 genes (ATG13, BAP1, ELAVL1, SLC38A1, and YY1AP1) and the risk score was calculated with relevant coefficients of 5 CRFGs as follows (Table): risk score= (0.011* expression level of ATG13 + 0.003* expression level of BAP1 + 0.006* expression level of ELAVL1 + 0.005* expression level of SLC38A1 + 0.008* expression level of YY1AP1). The patients were secerned into a high-risk group(n = 182) and a low-risk group (n = 183) in line with the median cut-off value (Fig. 2E-F.). The predictive performance of the risk score for OS was evaluated through time-dependent ROC curves, and the area under the curve (AUC) reached 0.713 at 1 year, 0.630 at 3 years, and 0.593 at 5 years (Fig. 2G-H). As depicted in Fig. 2I, those patients in the high-risk group bore a higher probability of death earlier than their counterparts in the low-risk group. Conformably, patients with high risk possessed a remarkably worse OS than those with low risk indicated by the Kaplan-Meier curve (Fig. 2I).

3.1.2. The CRFG-based signature was an independent indicator of HCC prognosis

Subsequently, we extracted the main clinical characteristics of HCC patients from TCGA database, including age, gender, grade, stage, and T. Both univariate and multivariate Cox analyses were actualized to testify whether this signature could be an independent prognostic indicator. Univariate analysis represented that risk score, TNM stage, and T stage were significantly relevant to the survival of HCC patients (p < 0.001) (Fig. 3A), whereas multivariate analysis revealed that risk score remained dramatically related to prognosis (p < 0.05) (Fig. 3B).

3.1.3. Clinicopathological manifestations & CRFG prognostic signature

We managed to figure out the association between this signature and clinicopathological manifestations via the Chi-square test. As the heatmap delineated, there were pronounced differences between high- and low risk groups in pathological stage, T stage, N stage, and tumor grade (Fig. 4A). The superior performance of CRFG-based signature in predicting the outcome of HCC patients in age < = 65, female, stage I-II, T1-T2 stage through conduction of a further stratification analysis (Fig. 4B-E). Moreover, we constructed the hybrid nomogram incorporating clinicopathological characteristics and the newly excavated CRFG prognostic signature which was stable and authentic and thus may be applied in the clinical management of HCC patients (Fig. 5A). The calibration curve illustrated that the 1-year, 3-years, and 5-year actual survival of the patients was in line with the predicted value (Fig. 5B). The concordance index (Riskscore = 0.68) approved the plausive predictive ability of the nomogram (Fig. 5C).

3.1.4. Enrichment analyses, Protein-Protein interaction (PPI) and TMB

Gene ontology (GO) enrichment and Kyoto encyclopedia of genes and genomes (KEGG) pathway analyses were carried out to elucidate the biological functions and pathways associated with the CRFGs. The biological process (BP) participated in the cellular response to chemical stress, response to oxidative stress, and cellular response to oxidative stress. The cellular component (CC) mainly regulated the production of the autophagosome, the phagophore assembly site. Molecular function (MF) was remarkably up-regulated in protein serine/threonine kinase activity and ubiquitin protein ligase binding. In the KEGG pathways, the results revealed that these genes were mainly involved in the FoxO signaling pathway, apoptosis, and Prolactin signaling pathway (Fig. 6A-B). Gene set enrichment analysis (GSEA) was performed to further elucidate the molecular mechanisms of CRFG-based signature. The results of GSEA showed that cell cycle, mTOR signaling pathway, Notch signaling pathway, p53 signaling pathway, pathways in cancer, and ErbB signaling pathway were mainly enriched in the high-risk group, whereas drug metabolism cytochrome P450, PPAR signaling pathway, complement, and coagulation cascades, primary bile acid biosynthesis, retinol metabolism, and tryptophan metabolism were mainly enriched in low-risk group (Fig. 6C). The PPI network of the differentially expressed CRFGs was established on the STRING database, and the top 10 nodes were ranked by the topology method of Maximal Clique Centrality (MCC) with the cytoHubba plug-in (Fig. 6D). The TMB of the high-risk group was much higher than that of the low-risk group based on the TMB score (1139 vs. 351) generated from the somatic mutation data of TCGA. As the waterfall plot shows, TP53 (36%) ranked at the top of the mutation rate in the high-risk group, while CTNNB1 (28%) peaked in the low-risk group (Fig. 6E-F).

3.1.5. Tumor Immunity analysis

The heatmap exhibited the association between the signature and immune infiltration according to the analyses of TIMER, CIBERSORT, CIBERSORT-ABS, XCELL, QUANTISEQ, EPIC, and MCP-counter (Fig. 7A). Given the significance of checkpoint inhibitor-based immunotherapies, we further commenced with exploring the difference in the expression of immune checkpoint between the high and low-risk groups. A substantial difference in the expression of CTAL4, TNFSF18, HAVCR2, and PDCD-1 among others between these two groups of patients (Fig. 7B). The box plot was used to show that aDCs, B cells, DCs, Neutrophils, NK cells, Th2 cells, and Treg had obvious correlations with the risk score via “limma”, “reshape2”, “scales”, “tidyverse”, and “ggpubr” (Fig. 7C). In addition, correlations analysis between immune cell subpopulations and related functions based on ssGSEA of TCGA-LIHC data uncovered that APC co-inhibition, cytolytic activity, MHC class I, type I IFN response, and type II IFN response were observably different between the low-risk and high-risk groups (Fig. 7D). We evaluated the difference in the immune score, stromal score, and estimate score between these two groups by ESTIMATE. The results of ESTIMATE exhibited that HCC patients in high-risk groups possessed lower immune and estimate scores than those in low-risk groups with p < 0.05 (Fig. 7E-F). In line with the result of tumor immune dysfunction and exclusion analysis, there was a significant difference in TIDE score between these two groups (Fig. 7G-J).

3.1.6. Consensus Cluster Analysis of Prognosis CRFGs

Based on the expression profiles of the 5-gene model, we performed consensus cluster analysis. k = 3 was identified with optimal clustering stability from k = 2 to 9, which harvested the greatest correlation within the group and the comparatively low correlation among groups (Fig. 8A-B), indicating the practicability of dividing the patients into three clusters. A consensus cumulative distribution function (CDF) diagram demonstrated that CDF reached an approximate maximum while k = 4, and classification was applicable (Fig. 8C). As Sankey portrayed, most of the Cluster 1 (C1) shunted to low-risk groups, while all of the Cluster 2 (C2) and most of the Cluster 3 (C3) swarmed into high-risk groups (Fig. 8D). The K-M curve revealed that patients in C1, approximating in the low-risk groups, had the best OS, whereas those in C2 had the worst OS (Fig. 8E). We evaluated the difference in the immune score, stromal score, and estimate score between these clusters by ESTIMATE. The results of ESTIMATE exhibited that HCC patients in C1 and C2 possessed higher immune, estimate, and stromal scores than those in C3 with a significant difference, which kept the consistency with the result before clustering (Fig. 8F-H). Compared with the difference between high- and low-risk subsets, we employed PCA to examine the difference among clusters (Fig. 8I-J) Next, based on the clustering pattern, we conducted a series of immune analyses previously performed to further testify to the correlation between CRFGs and tumor immunity in HCC, suggesting that CRFGs may provide a new reference marker for immunotherapy of HCC (Fig. 8K-L).

3.1.7. Drug sensitivity analysis

The prediction of potential therapeutic drugs indicated that the sensitivity to eight chemicals in the high- and low-risk groups dramatically differed. Compared with the low-risk group, the high-risk group obtained more sensitivity with four compounds (AT13148, AZD1332, SBZ16763), while patients in the low-risk group were more susceptive to four compounds (BMS-536924, BPD-00008900, PD-173074, GDC0810) than those in high-risk group (Fig. 9).

3.1.8. Analysis of Quantitative Real-Time Polymerase Chain Reaction

The mRNA relative expression of five cuproptosis-related ferroptosis genes (ATG13, BAP1, SLC38A1, ELAVL1, and YY1AP1) we screened out previously were identified in MHCC-97H, Hep-G2, HCC-LM3, PLC-PRF-5 and HL-7702 cells by the approach of qRT-PCR. As depicted in Figs. 9I-M, the expression level of these selected genes represented significantly higher in hepatocellular carcinoma cells compared with normal hepatic cells, which indicated that these factors are differentially expressed between tumor and normal tissues. In a nutshell, the experimental results confirmed the reliability of the risk signature as well. (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

3.2. Figures, and Tables

Table 1

The clinical characteristic of patients in the TCGA database.
Variable	Number of samples
Gender Male/Female	254/122
Age <=65/>65	235/141
Grade G1/G2/G3/G4/NA	55/180/123/13/5
Stage I/II/III/IV/NA	175/86/86/5/24
T T1/T2/T3/T4/NA	185/94/81/13/3
M M0/M1/NA	272/4/100
N N0/N1/NA	257/4/115

Cuproptosis, a novel type of regulated cell death, reveals that the balanced concentration of copper in an organism contributes to organismal homeostasis, and the imbalance in copper homeostasis may affect the organism by triggering cuproptosis (Zhao et al., 2022). The liver is considered a biochemical factory of metabolism for an organism. Research showed an accumulation of copper occurred in the cirrhotic liver compared with the healthy liver (Poznański et al., 2021). As the body’s metabolic hub for nutrients like glucose, lipids, and amino acids, disharmony of these metabolites can lead to oxidative stress, liver disease, and even mortality(Masoodi et al., 2021). It has been a decade since ferroptosis was coined, and there has been no lack of research on the ferroptotic role in hepatocellular carcinoma (HCC) (Tang et al., 2020). Ferroptosis, characterized by a massive iron overload and lethal peroxidation of polyunsaturated fatty acids (PUFAs), breeds the potential of hampering the carcinogenesis in HCC and attenuating the increasing resistance to current therapy of sorafenib (Bekric et al., 2022). Ferroptosis may act as the principal mechanism undergirding the anticancer effect of sorafenib by suppressing the cystine/glutamate antiporter SLC7A11(Lachaier et al., 2014). Interestingly, it should be pointed out that sorafenib was not a simon-pure inducer of ferroptosis based on screening results using a panel of cancer lines (Zheng et al., 2021). Moreover, IFNγ was used to sensitize HCC cells to ferroptosis via inhibiting system xc- by triggering the JAK/STAT signaling pathway, which served new insights into the feasibility of utilizing IFNγ to induce ferroptosis in treating HCC. Conversely, HCC cells with increased extracellular lactate levels are more resistant to ferroptosis induced by RSL3 and erastin(Zhao et al., 2020, Chen et al., 2022).

In our present study, we systematically carried out an innovative investigation that construct cuproptosis-related ferroptosis genes (CRFGs) by conjoined the cuproptosis-related genes and ferroptosis-related genes in accord with the correlation coefficient | r | > 0.5 and p < 0.05. Subsequently, we screened out 11 CRFGs with prognostic value by using univariate Cox regression, and further ruled out 5 CRFGs (ATG13, BAP1, ELAVL1, SLC38A1, YY1AP1) through the LASSO regression. The relative expression of the five-CRFGs in MHCC-97H, Hep-G2, HCC-LM3, PLC-PRF-5 HCC cells, and HL-7702 cells was validated via carrying out qPCR and exhibited in the bar graph. ATG13/ULK1 complex, the upstream of autophagy, can be activated by Licochalcone A(LicA), a chemotherapy drug inducing apoptosis as Bcl-2 inhibitor for HCC patients (Niu et al., 2018). The BAP1 gene has emerged as a major tumor suppressor mutated with various frequencies in numerous human malignancies, including hepatocellular carcinoma, intrahepatic cholangiocarcinoma, and clear cell renal cell carcinoma. BAP1 belongs to the deubiquitinase superfamily of enzymes, which are responsible for the maturation and regulation of ubiquitin signaling (Masclef et al., 2021). Still, recent research demonstrated that BAP1 mutations could define a homogeneous subgroup of hepatocellular carcinoma with fibrolamellar-like features and activated PKA (Hirsch et al., 2020). YAP1 and TAZ can regulate the expression of the Na+-dependent amino acid transporter SLC38A1 and the bidirectional transporter SLC7A5 to regulate the uptake of glutamine and leucine, thereby activating mTORC1, which are critical for tumor formation, growth, and progression in HCC and significantly associated with HCC patients’ survival. These results suggested that SLC38A1, SLC7A5, and mTORC1 are potential therapeutic targets in HCC (Park et al., 2016). In addition, the level of SLC38A1 could accurately predict the overall survival and relapse-free survival in patients with alcoholic steatohepatitis (ASH), and HCC (Cai et al., 2022). Interestingly, upregulated expression of SLC38A1 was dramatically related to an unfavorable prognosis and defective immune infiltration in HCC by utilizing immunohistochemical analysis of tissue samples (Liu et al., 2021). The interaction of the RNA-binding protein ELAVL1 serves as a key molecular event that triggers autophagy activation, promotes autophagic ferritin degradation, and in turn, leads to ferroptosis, which may provide promising diagnostic and therapeutic approaches to regulate hepatic stellate cells (HSCs) survival and death in liver fibrosis (Zhang et al., 2018). Moreover, ELAVL1 was demonstrated to contribute to not only HBV replication but also HCC cell growth, which might be a potent therapeutic target for the treatment of HBV-related HCC (Kanzaki et al., 2022). Previous studies have discovered that YY1-associated protein1 (YY1AP1) is a critical oncoprotein specifically activated in EpCAM + AFP + HCC. EpCAM is transcriptionally regulated by YY1AP1, which is dependent on YY1 to bind to EpCAM promoter and activate its transcription (Zhao et al., 2015).

We constructed a risk model and divided the patients of HCC into high-risk and low-risk subsets in accordance with a risk score. Survival analysis indicated that low-risk patients performed better than high-risk (p = 0.002). ROC curves showed that the CRFGs-based signature was of high accuracy and reliability, and the AUC value in the training group was 0.713 R 1 year. Clinicopathological analysis, enrichment analysis, PPI, and TMB implied that this 5-gene-based model offers remarkable sensitivity for survival prediction. Furthermore, through the level of tumor immunity (immune cells, immune function, immunological checkpoint, immune score), the results proved that the screened-out 5 CRFGs are related to tumor immune infiltration which might provide fresh insights into immunotherapy for patients with HCC. In addition, we performed the consensus clustering analysis and shunted patients into three clusters, most of the Cluster 1 (C1) shunted to low-risk groups, while all of the Cluster 2 (C2) and most of the Cluster 3 (C3) swarmed into high-risk groups. Based on the clustering, we conducted survival analysis, clinicopathological analysis, and tumor immunity analysis similar to previous steps. Those results kept consistent with the risk model, which testified to the stringency of the prognostic signature. The drug sensitivity we employed might be the predictive potential chemical compounds for a novel treatment strategy.

Our work in this paper still needs to be further improved. Although we accomplished the construction of the 5-CRGs signature with the prediction of prognosis and immune microenvironment, retrospective studies based on biological data still need further experimental and clinical data to confirm.

In this article, we systematically evaluated the value of cuproptosis-related ferroptosis genes in predicting survival, the role of the tumor microenvironment and immune cell infiltration, the underlying regulatory mechanism of genes associated with ferroptosis, and the prediction in potential chemical compounds for the treatment of HCC. Five genes’ features associated with cuproptosis-related ferroptosis might possess the predictive potency in the survival of patients with HCC and instruct patients to individualized treatment in the future.

Author Contributions: This research was conducted in collaboration with all authors. S.Z., J.W., J.X., and B.S. performed the data collation and analysis and analyzed and interpreted the results. S.Z. and J.J. drafted and reviewed the manuscript. All authors have read and approved the published version of the manuscript.

Funding: This study was supported by the National Natural Science Foundation of China (No. 81871965).

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Acknowledgments: We genuinely express our sincere gratitude to TCGA for sharing a large amount of data and the convenience provided by several online database tools.

Conflict of interests: The authors declare no conflict of interests.

Availability of Data and Materias: The public mRNA expression and corresponding clinical information used in this study were obtained from the public database the Cancer Genome Atlas, TCGA, (https://portal.gdc.cancer.gov/), Ferroptosis-related genes (FRGs) were downloaded from the public database (FerrDb, http://www.zhounan.org/ferrdb/legacy/index.html)

ANWANWAN, D., SINGH, S. K., SINGH, S., SAIKAM, V. & SINGH, R. (2020), "Challenges in liver cancer and possible treatment approaches", Biochim Biophys Acta Rev Cancer, Vol. 1873 No. 1, pp. 188314.
BEKRIC, D., OCKER, M., MAYR, C., STINTZING, S., RITTER, M., KIESSLICH, T. & NEUREITER, D. (2022), "Ferroptosis in Hepatocellular Carcinoma: Mechanisms, Drug Targets and Approaches to Clinical Translation", Cancers (Basel), Vol. 14 No. 7, pp.
CAI, C., LIN, J., LI, J., WANG, X. D., XU, L. M., CHEN, D. Z. & CHEN, Y. P. (2022), "miRNA-432 and SLC38A1 as Predictors of Hepatocellular Carcinoma Complicated with Alcoholic Steatohepatitis", Oxid Med Cell Longev, Vol. 20224832611.
CENTER, M. M. & JEMAL, A. (2011), "International trends in liver cancer incidence rates", Cancer Epidemiol Biomarkers Prev, Vol. 20 No. 11, pp. 2362-8.
CHEN, J., LI, X., GE, C., MIN, J. & WANG, F. (2022), "The multifaceted role of ferroptosis in liver disease", Cell Death Differ, Vol. 29 No. 3, pp. 467-480.
CHEN, X., KANG, R., KROEMER, G. & TANG, D. (2021), "Broadening horizons: the role of ferroptosis in cancer", Nat Rev Clin Oncol, Vol. 18 No. 5, pp. 280-296.
CRAIG, A. J., von FELDEN, J., GARCIA-LEZANA, T., SARCOGNATO, S. & VILLANUEVA, A. (2020), "Tumour evolution in hepatocellular carcinoma", Nat Rev Gastroenterol Hepatol, Vol. 17 No. 3, pp. 139-152.
HADIAN, K. & STOCKWELL, B. R. (2020), "SnapShot: Ferroptosis", Cell, Vol. 181 No. 5, pp. 1188-1188.e1.
HIRSCH, T. Z., NEGULESCU, A., GUPTA, B., CARUSO, S., NOBLET, B., COUCHY, G., BAYARD, Q., MEUNIER, L., MORCRETTE, G., SCOAZEC, J. Y., BLANC, J. F., AMADDEO, G., NAULT, J. C., BIOULAC-SAGE, P., ZIOL, M., BEAUFRÈRE, A., PARADIS, V., CALDERARO, J., IMBEAUD, S. & ZUCMAN-ROSSI, J. (2020), "BAP1 mutations define a homogeneous subgroup of hepatocellular carcinoma with fibrolamellar-like features and activated PKA", J Hepatol, Vol. 72 No. 5, pp. 924-936.
JIANG, P., GU, S., PAN, D., FU, J., SAHU, A., HU, X., LI, Z., TRAUGH, N., BU, X., LI, B., LIU, J., FREEMAN, G. J., BROWN, M. A., WUCHERPFENNIG, K. W. & LIU, X. S. (2018), "Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response", Nat Med, Vol. 24 No. 10, pp. 1550-1558.
KANZAKI, H., CHIBA, T., KANEKO, T., AO, J., KAN, M., MUROYAMA, R., NAKAMOTO, S., KANDA, T., MARUYAMA, H., KATO, J., ZEN, Y., KOTANI, A., SEKIBA, K., OTSUKA, M., OHTSUKA, M. & KATO, N. (2022), "The RNA-Binding Protein ELAVL1 Regulates Hepatitis B Virus Replication and Growth of Hepatocellular Carcinoma Cells", Int J Mol Sci, Vol. 23 No. 14, pp.
LACHAIER, E., LOUANDRE, C., GODIN, C., SAIDAK, Z., BAERT, M., DIOUF, M., CHAUFFERT, B. & GALMICHE, A. (2014), "Sorafenib induces ferroptosis in human cancer cell lines originating from different solid tumors", Anticancer Res, Vol. 34 No. 11, pp. 6417-22.
LIU, Y., YANG, Y., JIANG, L., XU, H. & WEI, J. (2021), "High Expression Levels of SLC38A1 Are Correlated with Poor Prognosis and Defective Immune Infiltration in Hepatocellular Carcinoma", J Oncol, Vol. 20215680968.
MASCLEF, L., AHMED, O., ESTAVOYER, B., LARRIVÉE, B., LABRECQUE, N., NIJNIK, A. & AFFAR, E. B. (2021), "Roles and mechanisms of BAP1 deubiquitinase in tumor suppression", Cell Death Differ, Vol. 28 No. 2, pp. 606-625.
MASOODI, M., GASTALDELLI, A., HYÖTYLÄINEN, T., ARRETXE, E., ALONSO, C., GAGGINI, M., BROSNAN, J., ANSTEE, Q. M., MILLET, O., ORTIZ, P., MATO, J. M., DUFOUR, J. F. & OREŠIČ, M. (2021), "Metabolomics and lipidomics in NAFLD: biomarkers and non-invasive diagnostic tests", Nat Rev Gastroenterol Hepatol, Vol. 18 No. 12, pp. 835-856.
NEWMAN, A. M., LIU, C. L., GREEN, M. R., GENTLES, A. J., FENG, W., XU, Y., HOANG, C. D., DIEHN, M. & ALIZADEH, A. A. (2015), "Robust enumeration of cell subsets from tissue expression profiles", Nat Methods, Vol. 12 No. 5, pp. 453-7.
NIU, Q., ZHAO, W., WANG, J., LI, C., YAN, T., LV, W., WANG, G., DUAN, W., ZHANG, T., WANG, K. & ZHOU, D. (2018), "LicA induces autophagy through ULK1/Atg13 and ROS pathway in human hepatocellular carcinoma cells", Int J Mol Med, Vol. 41 No. 5, pp. 2601-2608.
PARK, Y. Y., SOHN, B. H., JOHNSON, R. L., KANG, M. H., KIM, S. B., SHIM, J. J., MANGALA, L. S., KIM, J. H., YOO, J. E., RODRIGUEZ-AGUAYO, C., PRADEEP, S., HWANG, J. E., JANG, H. J., LEE, H. S., RUPAIMOOLE, R., LOPEZ-BERESTEIN, G., JEONG, W., PARK, I. S., PARK, Y. N., SOOD, A. K., MILLS, G. B. & LEE, J. S. (2016), "Yes-associated protein 1 and transcriptional coactivator with PDZ-binding motif activate the mammalian target of rapamycin complex 1 pathway by regulating amino acid transporters in hepatocellular carcinoma", Hepatology, Vol. 63 No. 1, pp. 159-72.
POZNAŃSKI, J., SOŁDACKI, D., CZARKOWSKA-PĄCZEK, B., BONNA, A., KORNASIEWICZ, O., KRAWCZYK, M., BAL, W. & PĄCZEK, L. (2021), "Cirrhotic Liver of Liver Transplant Recipients Accumulate Silver and Co-Accumulate Copper", Int J Mol Sci, Vol. 22 No. 4, pp.
TANG, B., ZHU, J., LI, J., FAN, K., GAO, Y., CHENG, S., KONG, C., ZHENG, L., WU, F., WENG, Q., LU, C. & JI, J. (2020), "The ferroptosis and iron-metabolism signature robustly predicts clinical diagnosis, prognosis and immune microenvironment for hepatocellular carcinoma", Cell Commun Signal, Vol. 18 No. 1, pp. 174.
TSVETKOV, P., COY, S., PETROVA, B., DREISHPOON, M., VERMA, A., ABDUSAMAD, M., ROSSEN, J., JOESCH-COHEN, L., HUMEIDI, R., SPANGLER, R. D., EATON, J. K., FRENKEL, E., KOCAK, M., CORSELLO, S. M., LUTSENKO, S., KANAREK, N., SANTAGATA, S. & GOLUB, T. R. (2022), "Copper induces cell death by targeting lipoylated TCA cycle proteins", Science, Vol. 375 No. 6586, pp. 1254-1261.
VILLANUEVA, A. (2019), "Hepatocellular Carcinoma", N Engl J Med, Vol. 380 No. 15, pp. 1450-1462.
WILKERSON, M. D. & HAYES, D. N. (2010), "ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking", Bioinformatics, Vol. 26 No. 12, pp. 1572-3.
YOSHIHARA, K., SHAHMORADGOLI, M., MARTÍNEZ, E., VEGESNA, R., KIM, H., TORRES-GARCIA, W., TREVIÑO, V., SHEN, H., LAIRD, P. W., LEVINE, D. A., CARTER, S. L., GETZ, G., STEMKE-HALE, K., MILLS, G. B. & VERHAAK, R. G. (2013), "Inferring tumour purity and stromal and immune cell admixture from expression data", Nat Commun, Vol. 42612.
ZHANG, Z., YAO, Z., WANG, L., DING, H., SHAO, J., CHEN, A., ZHANG, F. & ZHENG, S. (2018), "Activation of ferritinophagy is required for the RNA-binding protein ELAVL1/HuR to regulate ferroptosis in hepatic stellate cells", Autophagy, Vol. 14 No. 12, pp. 2083-2103.
ZHAO, J., GUO, S., SCHRODI, S. J. & HE, D. (2022), "Cuproptosis and cuproptosis-related genes in rheumatoid arthritis: Implication, prospects, and perspectives", Front Immunol, Vol. 13930278.
ZHAO, X., PARPART, S., TAKAI, A., ROESSLER, S., BUDHU, A., YU, Z., BLANK, M., ZHANG, Y. E., JIA, H. L., YE, Q. H., QIN, L. X., TANG, Z. Y., THORGEIRSSON, S. S. & WANG, X. W. (2015), "Integrative genomics identifies YY1AP1 as an oncogenic driver in EpCAM(+) AFP(+) hepatocellular carcinoma", Oncogene, Vol. 34 No. 39, pp. 5095-104.
ZHAO, Y., LI, M., YAO, X., FEI, Y., LIN, Z., LI, Z., CAI, K., ZHAO, Y. & LUO, Z. (2020), "HCAR1/MCT1 Regulates Tumor Ferroptosis through the Lactate-Mediated AMPK-SCD1 Activity and Its Therapeutic Implications", Cell Rep, Vol. 33 No. 10, pp. 108487.
ZHENG, J., SATO, M., MISHIMA, E., SATO, H., PRONETH, B. & CONRAD, M. (2021), "Sorafenib fails to trigger ferroptosis across a wide range of cancer cell lines", Cell Death Dis, Vol. 12 No. 7, pp. 698.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

An Innovative Ferroptosis-related Genes Signature for Prediction of Prognosis and Immune Microenvironment in Patients with Hepatocellular Carcinoma

Status:

Version 1

Abstract

Figures

Simple Summary

1. Introduction

2. Materials And Methods

2.1 Dataset Acquisition and Processing

2.2 Construction and Validation of Risk Score Based on CRFGs

2.3 Enrichment analyses and protein-protein interaction (PPI)

2.4 Nomogram Establishment and calibration

2.5 Tumor Mutational Burden and Tumor Immune Analysis

2.6 Profiles of CRFGs Identified Three Distinct Clusters of HCC

2.7 Prediction of Potential Compounds in the Treatment of Hepatocellular Carcinoma

2.8 Quantitative Real-Time Polymerase Chain Reaction

2.9 Statistical Analysis

3. Results

3.1. Subsection

3.1.1. Construction and validation of CRFG-based signature

3.1.2. The CRFG-based signature was an independent indicator of HCC prognosis

3.1.3. Clinicopathological manifestations & CRFG prognostic signature

3.1.4. Enrichment analyses, Protein-Protein interaction (PPI) and TMB

3.1.5. Tumor Immunity analysis

3.1.6. Consensus Cluster Analysis of Prognosis CRFGs

3.1.7. Drug sensitivity analysis

3.1.8. Analysis of Quantitative Real-Time Polymerase Chain Reaction

3.2. Figures, and Tables

4. Discussion

5. Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1