Skip to main content
Download PDF

A novel super-enhancer-related risk model for predicting prognosis and guiding personalized treatment in hepatocellular carcinoma

BMC Cancer volume 24, Article number: 1087 (2024) Cite this article

Abstract

Background

Our research endeavored to develop a robust predictive signature grounded in super-enhancer-related genes (SERGs), with the dual objectives of forecasting survival outcomes and evaluating the tumor immune microenvironment (TiME) in hepatocellular carcinoma (HCC).

Methods

HCC RNA-sequencing data were retrieved from The Cancer Genome Atlas (TCGA), and 365 patients were randomly assigned to training or testing sets in 1:1 ratio. SERGs of HCC were downloaded from Super-Enhancer Database (SEdb). On the basis of training set, a SERGs signature was identified, and its prognostic value was confirmed by internal and external validation (GSE14520) sets. We subsequently examined the model for potential functional enrichment and the degree of tumor immune infiltration. Additionally, we carried out in vitro experiments to delve into the biological functions of CBX2 gene.

Results

An SE-related prognostic model including CBX2, TPX2, EFNA3, DNASE1L3 and SOCS2 was established and validated. According to this risk model, patients in the high-risk group had a significantly worse prognosis, and their immune cell infiltration was significantly different from that of low-risk group. Moreover, the high-risk group exhibited a significant enrichment of tumor-associated pathological pathways. The SERGs signature can generally be utilized to screen HCC patients who are likely to respond to immunotherapy, as there is a positive correlation between the risk score and the Tumor Immune Dysfunction and Exclusion (TIDE) score. Furthermore, the downregulation of the CBX2 gene expression was found to inhibit HCC cell viability, migration, and cell cycle progression, while simultaneously promoting apoptosis.

Conclusions

We developed a novel HCC prognostic model utilizing SERGs, indicating that patients with high-risk score not only face a poorer prognosis but also may exhibit a diminished therapeutic response to immune checkpoint inhibitors (ICIs). This model is designed to tailor personalized treatment strategies to the individual needs of each patient, thereby improving the overall clinical outcomes for HCC patients. Furthermore, CBX2 is a promising candidate for therapeutic intervention in HCC.

Peer Review reports

Introduction

Hepatocellular carcinoma (HCC) has been the most common primary malignancy of liver, with a high rate of fatalities [1, 2]. The treatment of HCC has evolved over the past years. After the SHARP trial in 2007 [3], Sorbafenib was approved as the only evidence-based treatment in advanced HCC for almost a decade. Subsequently, targeted drugs such as Lenvatinib, Regorafenib, and Cabozantinib successively showed promising therapeutic effects in advanced HCC [4,5,6]. Additionally, an increasing body of evidence has suggested that the advent of immune checkpoint inhibitors (ICIs) is profoundly transforming the therapeutic landscape and prognostic outcomes for HCC [7,8,9]. It should not be ignored that the current treatment methods still have some limitations, such as drug resistance, “ceiling effect”, adverse reactions, etc. In addition, due to the heterogeneity of HCC pathogenesis, development, and metastatic processes, we still lack biomarkers that can accurately predict the prognosis of HCC patients [10, 11]. Studies have revealed that serum alpha-fetoprotein (AFP), chronic hepatitis B virus infection, des-gamma carboxyprothrombin (DCP), circulating tumor cell DNA (ctDNA), extracellular vesicles (EV), centrosome and spindle pole-associated protein (CSPP1) and other biomarkers were related to diagnosis and prognosis of HCC, but without specificity and stability [12,13,14,15]. Thus, it is crucial to identify appropriate biomarkers for prognosis and prediction of HCC.

Super-enhancer (SE) was a class of cis-regulatory elements with super transcriptional activation characteristics, which was firstly proposed by Richard A. Young and his colleagues [16, 17], with a wide range of potential applications, such as the discovery of pathogenic drivers, analyzing variation sites associated with disease susceptibility, and developing accurate diagnoses and therapies for complex diseases [18, 19]. Researchers have been exploring whether SE could be used to identify effective biomarkers in cancer and develop therapeutics that target transcriptional addictions in recent years [20,21,22]. SE-related genes (SERGs) are therefore expected to be biomarkers for prognosis and antitumor therapy [23, 24]. In spite of this, additional studies are required to elucidate the strong connection between SE and HCC [25, 26]. Hence, we are committed to harnessing the power of bioinformatics to identify and develop potential HCC biomarkers, followed by a rigorous validation process [27,28,29]. The aim of this study is to develop a new risk assessment model to accurately predict the prognosis for HCC patients by analyzing specific SERGs, thereby enhancing clinical decision-making and improving patient outcomes.

Based on The Cancer Genome Atlas (TCGA), we collected RNA sequencing and clinicopathological information from patients with HCC. A thorough bioinformatics analysis was subsequently undertaken, culminating in the development of five prognostic SERGs, which were then validated using both internal and external validation sets. To uncover potential mechanisms and pathways, enrichment analyses were performed between low-risk and high-risk groups. Additionally, the relationship between SERGs signature and tumor immune microenvironment (TiME) was also investigated using analyses of tumor infiltration immune levels, tumor-infiltrating immune cells, immune checkpoints, and chemotactic factors. Furthermore, a suite of techniques including reverse transcription quantitative polymerase chain reaction (RT-qPCR), Western Blots, immunohistochemistry (IHC), and a range of functional assays, such as the cell counting kit-8 (CCK-8) assay, cellular migration assay, cell cycle analysis, and apoptosis assay, were employed to ascertain the expression levels and clinical relevance of CBX2, a risk gene within our prognostic model, in HCC cell lines and tissues.

Materials and methods

Acquisition of data

In accordance with TCGA policy, gene expression profiles (RNA-sequencing) and clinical information for HCC and normal samples were obtained from the publicly available TCGA database (https://portal.gdc.cancer.gov/) [30]. To develop and validate a novel predictive model, we randomly allocated HCC patients from the TCGA-LIHC dataset into training and internal validation sets in a balanced 1:1 ratio. The inclusion criteria were stringent, necessitating the presence of both survival time and status data, along with comprehensive transcriptome profiles. In addition, the transcriptomic data matrix and clinical parameters for the external validation set (GSE14520) were downloaded from the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.go/geo/).

Identification of HCC SERGs

From SEdb 2.0 (www.licpathway.net/sedb/), 2073 SERGs of HCC were identified after removing duplicates using the following screening method: data-browsing, human, NCBI GEO/SRA, tissue, liver (Sample_02_1299) (the details of 2073 SERGs were shown in Additional File 1).

Filtering of SERGs and clustering analyses

Differentially expressed genes (DEGs) between tumor and normal samples were screened using cut-offs of |log2 (fold change) |>2 and false discovery rate (FDR) control (adjusted P < 0.05). An analysis of Venn curves was used to select overlapping genes, and Volcano plots were generated using the “ggplot2” package (https://ggplot2.tidyverse.org). In addition, we conducted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis for HCC patients based on SERGs using the “cluster Profiler” package [31]. We also analyzed the degree of protein-protein interaction (PPI) network of significantly positive DEGs using the Search Tool for the Retrieval of Interacting Genes (STRING) database (http://string-db.org) [32].

Construction and validation of the novel SERGs signature

For SERGs with prognostic value, a univariate Cox analysis of overall survival (OS) was conducted and forest plots were presented (P < 0.05) [33]. To develop and verify a novel signature, we randomly allocated 365 HCC patients from the TCGA-LIHC database, each with detailed survival data, status information, and complete transcriptome profiles, into training and testing cohorts in a balanced 1:1 ratio. Utilizing the training dataset, we conducted Least Absolute Shrinkage and Selection Operator (LASSO) regression, complemented by tenfold cross-validation [34]. Subsequently, we identified five SERGs and incorporated them into a prognostic risk signature. For each HCC patient, the individual risk score was computed by applying the normalized expression levels of these SERGs to their respective regression coefficients, employing the following formula:

$$Risk\,score=\:\sum\:_{i=1}^{n}{expression}_{gene\_i}\:\times\:\:{lasso\_coeffieicent}_{gene\_i}$$

With the median risk score serving as the cutoff, patients in the training set were stratified into low- and high-risk groups. Kaplan-Meier (KM) analysis was then applied to evaluate the OS differences between these groups. For validation, we replicated the risk score calculation for each HCC patient in the internal and external validation (GSE14520) sets, following the formula from the training phase, and subsequently categorized them into low- and high-risk groups based on the median risk score.

DNA methylation is a pivotal factor in cancer prognosis and holds promise as a biomarker [35]. In this study, we leveraged MethSurv (https://biit.cs.ut.ee/methsurv/), a bioinformatics tool for analyzing the impact of DNA methylation on survival [36], to explore the prognostic significance and expression patterns of single CpG sites within SERGs in HCC. The analysis was based on beta values, which quantify methylation levels on a scale from 0 (unmethylated) to 1 (fully methylated).

Clinicopathological characteristics and SERGs signature

In the training cohort of HCC patients, subgroups were delineated based on a spectrum of clinicopathological features, encompassing age, gender, Eastern Cooperative Oncology Group performance status (ECOG PS), pathological T stage, tumor grade, Child-Pugh score, presence of vascular invasion, alpha-fetoprotein (AFP) levels, and body mass index (BMI). Subsequently, we scrutinized the correlation between the risk score and these clinical attributes, as well as the prognostic efficacy of the SERGs signature across subgroups stratified by distinct clinical parameters.

Development and validation of a nomogram

Clinicopathological indicators, encompassing age, gender, ECOG PS, pathological T stage, tumor grade, Child-Pugh grade, presence of vascular invasion, AFP levels, and BMI, were subjected to both univariate and multivariate Cox regression analyses. A comprehensive multivariate regression model was then formulated, integrating the risk score with other independent prognostic factors, to predict the 1-, 3-, and 5-year OS rates. The predictive accuracy of the nomogram was evaluated through calibration plots, time-dependent receiver operating characteristic (ROC) curve analysis, concordance index (C-index), and decision curve analysis (DCA).

The SERGs signature and TiME

Employing the ESTIMATE algorithm [37] and the Immune Cell Abundance Identifier (ImmuCellAI) (http://bioinfo.life.hust.edu.cn/web/ImmuCellAI/) [38], we quantified immune infiltration scores and characterized tumor-infiltrating immune cells (TIICs) within HCC samples. Utilizing the “TIMER” (http://timer.cistrome.org/) analysis tool [39], we further established the correlation between the identified SERGs and the immune cell profiles. By examining the relationship between the risk score and Tumor Immune Dysfunction and Exclusion (TIDE) score (http://tide.dfci.harvard.edu/), we have preliminarily assessed the predictive capacity of the risk score system for patient responses to ICIs. Additionally, we conducted an in-depth analysis to explore the relationships between SERGs signature, immune checkpoints, and chemotactic factors, aiming to prognosticate the TiME of HCC.

DEGs of high-risk group compared with low-risk group

DEG analysis was conducted to distinguish between high-risk and low-risk groups, employing the ‘edgeR’ R package with a FDR threshold of less than 0.05 and |log2FC|≥2. Statistical significance for GO and KEGG pathway analyses was determined by P values below 0.05. Subsequently, a ridge plot was employed to graphically represent the distinct gene expression patterns characteristic of the high-risk group, utilizing the GSEA (http://software.broadinstitute.org/gsea/index.jsp) (version 3.14.3) [40].

Preliminary experimental validation of the CBX2 gene

The mRNA expression levels of the CBX2 gene across various HCC cell lines were corroborated using data from the Cancer Cell Line Encyclopedia (CCLE) (https://sites.broadinstitute.org/ccle) and verified through RT-qPCR assays. Concurrently, the protein expression levels of CBX2 were ascertained in both cell lines and tissues through Western Blot and IHC analyses. Furthermore, the presence and localization of CBX2 protein were elucidated in human A-431, osteosarcoma U2OS, and U-251MG cell lines via immunocytochemistry (ICC), utilizing image data sourced from the Human Protein Atlas (HPA) Protein Atlas (http://www.proteinatlas.org/). Additionally, the functional impact of CBX2 was investigated by siRNA-mediated gene knockdown, focusing on its influence on cellular viability, cellular migration, cell cycle progression, and apoptosis in HepG2 cells.

Cell lines and cell culture

The normal human liver cell line MIHA was sourced from Fenghui Biological Co., Ltd. (Hunan, China) and maintained in RPMI-1640 medium (Meilunbio, China) enriched with 10% fetal bovine serum (FBS; Meilunbio, China) and 100 U/ml penicillin-streptomycin. The culture was conducted in a humidified incubator with an atmosphere of 5% CO2 at a temperature of 37 °C. The human HCC cell lines HepG2 and C3A were procured from Procell Life Science and Technology Co., Ltd. (Wuhan, China) and ATCC® (Manassas, VA), respectively. Both were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM; Meilunbio, China), supplemented similarly with 10% FBS (Meilunbio, China) and 100 U/ml penicillin-streptomycin, under the same controlled environmental conditions.

RT-qPCR

Complementary DNA (cDNA) was synthesized from total cellular RNA using the HiScript III All-in-one RT SuperMix Perfect for qPCR (Vazyme, China). Subsequently, RT-qPCR was conducted on an ABI 7500 thermocycler (Applied Biosystems, Foster City, CA, USA) with the SYBR Green PCR Master Mix (Vazyme, China). Primer sequences are detailed in Additional File 2. Relative fold changes in mRNA expression were determined by comparing sample Ct values to control groups using the comparative ΔΔCt method.

Clinical tissue collection

HCC tissue samples were procured in 2022 from patients diagnosed at the First Affiliated Hospital of Fujian Medical University. The study excluded patients who had undergone any pre-surgical treatment and those with a history of other malignancies. Prior to sample collection, all participants provided their written informed consent for the clinical application and research use of their tumor tissues.

IHC

IHC was conducted using the standard avidin-biotin complex method. The staining intensity was graded on a scale of 0 to 3, and the extent of positively stained cells was scored from 1 to 4. The IHC score for each sample was calculated as the product of the staining intensity and the positive cell range scores. The CBX2-stained tissue sections were categorized into high and low expression groups based on the comparative IHC scores between the HCC tissues and their corresponding normal adjacent tissues. This assessment was performed by two independent pathologists in a blinded fashion to ensure objective evaluation of all specimens.

Transfection with siRNA

For the knockdown of CBX2, HepG2 cells were seeded in 6-well plates at a density of 3 × 10^5 cells per well. Once the cells achieved 30–40% confluence, transfection was initiated using siRNA and the GP-transfect-Mate reagent (Shanghai GenePharma Co., Ltd.) following the manufacturer’s protocol. The siRNA duplexes, designed with the sense and antisense sequences GUCUUGGACUUGCAGAGUGTT and CACUCUGCAAGUCCAAGACTT, respectively, were synthesized by Shanghai GenePharma Co., Ltd. The transfection process was carried out for 48 h to ensure efficient depletion of CBX2.

CCK-8 assay

Cells were cultured at 5% CO2 and 37 °C on 96-well plates (100 µL/well), and ten microliters of CCK-8 reagent (Beyotime Bio Inc., China) were added to each well after 24 h, 48 h, and 72 h, respectively, with OD450 values determined by a microplate reader.

Cellular migration assay

The cellular migration assay was performed in vitro using 24-well transwell chambers. Cells (8 × 10^4) were seeded in the top chambers, and the bottom chambers were filled with DMEM medium (600 ml) containing 10% FBS to stimulate migration. After 60 h of incubation, the cells were stained with 0.1% crystal violet. The cells that had migrated through the ostioles to the reverse side were counted under a microscope in five predetermined fields at magnification of 200. Each assay was performed in triplicate.

Cell cycle assay

After harvesting and fixing cells in 70% cold ethanol overnight at 4 °C, the cells were centrifuged at 2500 rpm for 5 min the next day. In flow cytometry (Beckman Coulter), the cell cycle was assessed after pellets were washed with PBS, treated with RNase A, and stained with PI at 37 °C for 30 min in the dark.

Apoptosis assay

Apoptosis ratios were determined using the Annexin V-FITC/PI apoptosis kit (MultiSciences (Lianke) Biotech Co., Ltd.). The cells were harvested 48 h after transfection and resuspended in binding buffer containing Annexin V-FITC and PI. Flow cytometry (Beckman Coulter) was used to analyze the samples. In three repeated experiments, viable cells, necrotic cells, and apoptotic cells were separated, and the percentages of apoptotic cells from each group were compared.

Extraction of proteins and western blot analysis

In order to lyse the cells, Radio Immunoprecipitation Assay (RIPA) lysis buffer (Beyotime, China) was used. We determined the protein concentration with a BCA protein concentration kit (Beyotime, China) and separated the proteins using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Following electroblotting, the proteins were transferred to polyvinylidene difluoride membranes (Millipore, 150 Billerica, MA, USA). In a solution of 5% nonfat dry milk in Tris Buffer Solution Tween (TBST), the membranes were blocked for 2 h. The primary antibodies CBX2 (Chromobox 2; proteintech, 1: 5000) and GAPDH (HRP-60004, proteintech, 1:5000) were then incubated at 4 °C overnight. Every 10 min, the membrane was washed with TBST, for three times. With goat anti-rabbit IgG secondary antibodies (HUABIO, 1: 50,000), membranes were washed and stained for 1 h at room temperature. We then washed the membranes three times with TBST every 10 min, cut the blots, and hybridized them with antibodies. Using an ECL kit from Thermo Scientific (Rockford, USA), the bands were developed. Image Lab Software (Bio-Rad) was used to analyze the images taken with GelDoc XR equipment. The original, uncropped images of Western Blot were in Additional File 3.

Statistical analysis

In this study, statistical analyses were performed using R software (version 4.3.1). The KM was analyzed using the log-rank test. We compared the proportion of immune infiltration scores, TIICs, TIDE scores, immune checkpoints, and chemotactic factors between low-risk and high-risk groups using T or Wilcoxon tests. To compare the means of two independent groups, the independent t-test was used when the data were normally distributed. For statistical analysis, Graphpad Prism 8.0 (GraphPad Software Inc., USA) was used (P value of 0.05 was considered statistically significant). * P < 0.05, ** P < 0.01, *** P < 0.001.

Results

Identification of SERGs in HCC

Figure 1 illustrated the study’s flow. With raw data preprocessing, 1,630 DEGs of TCGA-LIHC (Fig. 2A) were obtained (see Additional File 4 for details on 1630 DEGs). After removing duplicates from SEdb 2.0, a total of 2,073 SERGs of HCC were identified. As a result of the overlap between the two algorithms mentioned above, 104 SERGs were identified for further analysis (Fig. 2B) (the details of 104 SERGs are shown in Additional File 5).

Fig. 1
figure 1

The Flow Diagram of the Study

Full size image
Fig. 2
figure 2

Differentially Expressed Super-Enhancer-Related Genes (SERGs) in Hepatocellular Carcinoma (HCC). A The Volcano plot illustrates the differentially expressed genes (DEGs) in the TCGA-LIHC dataset, applying stringent thresholds for |log2 (fold change)| of greater than 2 and employing false discovery rate (FDR) adjustments with a significance level of P < 0.05; B utilizing a Venn diagram, 104 SERGs were identified through an intersection of DEGs from the TCGA-LIHC dataset and SERGs cataloged in the SEdb database; C Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (www.kegg.jp/kegg/kegg1.html) pathway analyses were conducted on the 104 differentially expressed SERGs to explore their biological functions and pathways; D a protein-protein interaction (PPI) network was constructed from the 104 differentially expressed SERGs using the STRING database to map their interactive relationships

Full size image

Functional clustering analysis of SERGs in HCC

The functions of the 104 SERGs were determined using GO and KEGG analyses (see Additional File 6). In terms of ion gated channels, cell division processes, and ligand interaction, these genes were significantly enriched (Fig. 2C). Furthermore, the protein-protein interaction (PPI) network among these genes revealed that they were interacted (Fig. 2D).

Creating and validating SERGs signature

The SERGs with prognostic value in HCC were screened using univariate Cox regression analysis. We then identified 33 SERGs as potential prognostic genes by using a cut-off threshold of P < 0.05 (Fig. 3A, B). By calculating regression coefficients in the training set, LASSO regression algorithm was used to refine the model (Fig. 3C, D). As a result, the SE-related risk model, comprising CBX2, TPX2, EFNA3, DNASE1L3 and SOCS2, was obtained and the risk score system was established (Table 1) (the relationships between genes and clinical outcomes were detailed in Additional File 7).

Fig. 3
figure 3

Identification and Validation of the 5-SERGs Signature. A A Venn diagram facilitated the identification of 33 SERGs from the TCGA-LIHC dataset that were associated with SE-related DEGs and genes linked to overall survival (OS); B forest plots were employed to graphically represent the prognostic significance of the 33 SE-related DEGs; C the LASSO model incorporated tenfold cross-validation for the selection of variables that are most predictive of survival outcomes; D the profiles of the LASSO coefficients for the five differentially expressed SERGs were depicted to illustrate their contribution to the predictive model; in the E training set, F internal validation set and G external validation set (GSE14520), risk plots were distributed alongside the survival status of patients, and heat maps were used to visualize the expression patterns of the five SERGs; Kaplan-Meier (KM) survival curves were compared between low-risk and high-risk groups in the H training set, I internal validation set and J external validation set (GSE14520) to demonstrate the predictive power of the SERGs signature for OS

Full size image
Table 1 Differentially expressed SERGs and their coefficients in LASSO regression model
Full size table

Cell growth is suppressed or promoted by DNA methyltransferases on CpG islands, which are reversible transcription factors [33]. DNA methylation clustering of the expression levels of the 5 SERGs in HCC is presented in this study, along with its prognostic value (Additional File 8). Among DNA methylation expression levels, cg03523533, cg00790461, cg16896344 from CBX2; cg10143807, cg17582777 from EFNA3 had the highest values and significant prognostic significance (likelihood ratio (LR) test P < 0.05) in HCC.

We compared the distribution of risk scores, survival status, survival time, and expression levels of five SERGs for HCC in training set (Fig. 3E), internal validation set (Fig. 3F), and external validation set (GSE14520) (Fig. 3G) using the risk score formula. There was a statistically significant difference between the high-risk group and the low-risk group in terms of the OS. KM analysis showed that the high-risk group had a significantly worse OS than the low-risk group in training set (Fig. 3H), internal validation set (Fig. 3I), and external validation set (GSE14520) (Fig. 3J).

Clinicopathological characteristics and risk score

In Fig. 4A, the association between risk score and clinicopathological characteristics was analyzed. There was a lower risk score in HCC patients with ECOG PS = 0, pathologic T1&T2 stage, G1&G2, without vascular invasion, and AFP ≤ 400ng/ml. We also validated the prediction efficiency of risk groups in several subgroups. It was found that patients in the high-risk group had a worse OS in subgroups of age ≤ 60 years, age > 60 years, male, ECOG PS = 0, ECOG PS ≥ 2, pathologic T1&T2 stage, N0, M0, G1&G2, G3&G4, Child-Pugh grade A, without vascular invasion, AFP ≤ 400ng/ml, BMI ≤ 25, and BMI > 25 (Fig. 4B).

Fig. 4
figure 4

Predictive Power of the Risk Score System. A The relationship between the risk score and various clinical characteristics was assessed, including age, gender, Eastern Cooperative Oncology Group (ECOG) performance status, pathologic T stage, tumor grade, Child-Pugh grade, presence of vascular invasion, alpha-fetoprotein (AFP) levels, and body mass index (BMI); B Kaplan-Meier (KM) curves were constructed to predict overall survival (OS) in different subgroups defined by these clinical characteristics: age ≤ 60 years and age > 60 years, male, ECOG performance status of 0 and ECOG performance status ≥ 2, pathologic T1&T2 stages, pathologic N0 stage, pathologic M0 stage, tumor grades G1&G2 and G3&G4, Child-Pugh grade A, absence of vascular invasion, AFP levels ≤ 400 ng/ml, and BMI categories of > 25 and ≤ 25

Full size image

Nomogram and validation

Clinicopathological factors associated with prognosis, including age, gender, ECOG PS, pathological T stage, grade, Child-Pugh grade, vascular invasion, AFP, and BMI, have been analyzed in training set using univariate and multivariate Cox regression (Table 2). The independent prognostic parameters (ECOG PS and pathologic T stage) and risk score were then calculated by univariate and multivariate Cox regression (Fig. 5A, B). In training set, high-risk group had significantly worse OS [hazard ratio (HR) = 5.132, 95% confidence interval (CI): 2.733–9.636, P < 0.001]. The model was verified in testing set (HR = 1.492, 95% CI: 1.032–2.155, P = 0.033) (Fig. 5C, D). After integrating ECOG PS, pathologic T stage, and risk score, the nomogram model was established, with C-index 0.790 (95% CI: 0.757–0.823) (Fig. 5E). Compared to the ideal model and the nomogram model, the calibration plots demonstrated excellent agreement among 1-, 3-, and 5-year OS rates (Fig. 5F). Additionally, DCA analyses of the model in training set (Fig. 5G, H, I) and time-dependent ROC analyses (Fig. 5J, K) of the model in both training and testing sets indicated robust discriminatory ability of the nomogram.

Table 2 Univariate and multivariate Cox analysis of OS in training set
Full size table
Fig. 5
figure 5

Prognostic Model Nomogram and Verification. In A, B training set and C, D testing set, univariate and multivariate Cox analyses of clinical factors and risk score with OS were performed; E nomogram predicting 1-, 3- and 5-years survival rate of HCC patients; F the calibration curves for 1-, 3-, 5-year OS in training set; G 1-, H 3-, I 5-year decision curve analysis (DCA) for the model in training set; time-dependent ROC curves for predictive performance of the model in J training set and K testing set

Full size image

The signature of SERGs and TiME

A negative correlation was found between the risk score and the Stromal score (Fig. 6A, B); furthermore, the risk score was negatively correlated with immune response-related cells (such as MAIT and Th17 cells) (Fig. 6C); additionally, the risk score correlated positively with the TIDE score, with higher scores indicating a poorer prognosis and decreased sensitivity to ICIs (Fig. 6D). The expression levels of each SERG were correlated with tumor purity and markers of different immune cells (Additional File 9). A correlation between immune checkpoints and risk score was also evaluated (Fig. 6E). A higher expression level of immune checkpoints was observed in the high-risk group (Fig. 6F), including PD-1, CTLA4, PDCD1, LAG3, ICOS, LMTK3, and TIGIT. Additionally, we examined the relationship between risk score and chemotactic factors (Fig. 6G). Chemotactic factors (including CCL26, CCL28, CCR3, CCR10, CXCL5, XCL1) were more abundant in the high-risk group (Fig. 6H).

Fig. 6
figure 6

Prognostic Model and Tumor Immune Microenvironment (TiME). A The correlation between the risk score and immune infiltration scores was examined, encompassing the Stromal score, Immune score, and ESTIMATE score to gauge the level of immune cell presence within the tumor microenvironment; B a comparison of immune infiltration scores was made between the low-risk and high-risk groups to identify any significant differences that might influence patient prognosis; C the distribution of tumor-infiltrating immune cells (TIICs) was compared between the low-risk and high-risk groups to assess whether the risk score could predict the level of immune cell infiltration; D the association between the risk groups and Tumor Immune Dysfunction and Exclusion (TIDE) scores; E the relationship between the risk score and immune checkpoints was investigated; F a comparative analysis of six key immune checkpoints was conducted between the low-risk and high-risk groups to evaluate the potential impact of these checkpoints on survival outcomes; G the connection between the risk score and the expression of chemotactic factors was examined; H the expression levels of six chemotactic factors were compared between the low-risk and high-risk groups to identify any disparities that could be attributed to the risk score and their implications for immune cell trafficking

Full size image

DEGs of high-risk group compared with low-risk group

GO and KEGG analyses of DEGs between high-risk and low-risk groups (Fig. 7A) revealed these genes were significantly enriched in transcriptional and post-transcriptional regulation, and cell cycle (Fig. 7B). Additionally, GSEA analysis was conducted in order to clarify the potential regulatory mechanisms responsible for the differences between high-risk and low-risk groups. Some cancer-related hallmarks, including transcriptional regulation, were significantly associated with high-risk groups (Fig. 7C). Cancer-related metabolism was associated with low-risk group (Fig. 7D).

Fig. 7
figure 7

Differentially Expressed Genes (DEGs) in the High-Risk Group Relative to the Low-Risk Group. A A variance ranking chart is presented to illustrate the variability of genes within the high-risk group when contrasted with the low-risk group; B Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses are conducted on the DEGs to explore their biological functions and associated pathways, with reference to the KEGG resource at www.kegg.jp/kegg/kegg1.html; C, D Gene Set Enrichment Analysis (GSEA) ridge maps are utilized to compare the enrichment patterns of DEGs between the high-risk and low-risk groups, providing a visual representation of the gene sets that are significantly associated with the risk status

Full size image

Preliminary experimental verification for CBX2 gene

In order to further confirm the prognostic significance of CBX2 in HCC, we focused on its expression, clinical significance and prognostic value. The expression of CBX2 at the transcriptome level in HCC cell lines was displayed using the database of CCLE (Fig. 8A) and RT-qPCR (Fig. 8B). As predicted by transcriptional results, CBX2 was highly expressed in HepG2 and C3A cell lines. The HPA database was used to reveal the subcellular localization of CBX2 protein in A-431, U2OS, and U-251MG cells by ICC, which indicated that CBX2 was primarily expressed in the nucleoplasm (Fig. 8C). Then, CBX2 expression was further confirmed at the protein level in cell lines using Western Blot (Fig. 8D). To confirm the dysregulation of CBX2, IHC analysis was also performed on 20 pairs of HCC and peritumoral tissues (Fig. 8E). Moreover, we knocked down the CBX2 by siRNA (Fig. 8F, G). Subsequently, the outcomes of the CCK-8 assay (Fig. 8H), cellular migration assay (Fig. 8I), cell cycle assay (Fig. 8J) and apoptosis assay (Fig. 8K) demonstrated that the suppression of CBX2 gene expression hindered the cell viability, cellular migration and cell cycle progression in HCC, while promoted apoptosis. A higher expression level of CBX2 was associated with pathologic T3&T4 stage, TNM stage III & stage IV, male, G3&G4, AFP > 400 ng/ml, and vascular invasion (Fig. 8L). Patients with high levels of CBX2 had a shorter OS (HR = 1.64, 95% CI: 1.16–2.32, P = 0.005) (Fig. 8M). Additionally, we also conducted an in-depth investigation into the influence of the CBX2 gene on a variety of immune molecules within HCC, utilizing a range of immune infiltration algorithms (Additional File 10).

Fig. 8
figure 8

Preliminary Experimental Validation of CBX2 Characteristics. A CBX2 mRNA expression levels in HCC cell lines from CCLE; B the mRNA relative expression levels of CBX2 in a normal liver cell line (MIHA) and HCC cell lines (HepG2 and C3A) determined by RT-qPCR; C ICC for determining the subcellular location of CBX2 in A-431, U2OS, and U-251MG cell lines by HPA. CBX2 localized to the nucleoplasm (green). Microtubules are stained in red and the nucleus in blue (DAPI); D the protein expression levels of CBX2 in a normal liver cell line (MIHA) and HCC cell lines (HepG2 and C3A) determined by cropped Western Blot; E the representative images of H&E and IHC for CBX2 expressed on HCC and peritumoral tissues, and the protein expression levels of CBX2 in 20 pairs of tumoral and peritumoral tissues by IHC; the mRNA and protein expression levels of CBX2 in HepG2 cells treated with siNC or siCBX2 via F RT-qPCR and G cropped Western Blot; H the cell viability of HepG2 cells treated with siNC or siCBX2 by CCK-8 assay; I cellular migration assay, J cell cycle assay, and K apoptosis assay of HepG2 cells treated with siNC or siZIC2; L the association between CBX2 expression levels and clinicopathological features in HCC; M KM survival curves of OS for HCC patients according to the expression level of CBX2. (* P < 0.05, ** P < 0.01, *** P < 0.001)

Full size image

Discussion

The SE plays a vital role in determining cell identity and fate, which are frequently associated with up-regulated expression of coding genes [16, 18]. Researchers have found that SE contribute to a variety of diseases as important regulatory regions [41, 42]. By activating various oncogenes and other cancer-related genes, SE plays an important regulatory role in cancer occurrence, cell differentiation, immune response, and other important biological processes [18, 43, 44].

Current research is intensely focused on the role and prognostic implications of SE in HCC development [25, 45, 46]. Despite the obscure nature of the underlying mechanisms, we have utilized SERGs to delineate the molecular profiles of HCC patients, thereby constructing prognostic models tailored to these profiles. Furthermore, we have investigated the interplay between the TiME and the prognostic biomarkers to guide the individualized treatment of HCC patients.

In this study, we identified a total of 1,630 DEGs from TCGA-LIHC and 2,073 SERGs from SEdb 2.0. Based on overlapping genes between the two algorithms mentioned above, we screened 104 SERGs and derived a 5-SERGs signature with prognostic value using univariate Cox regression analysis and LASSO regression analysis. The risk score system was validated under different circumstances. HCC patients were divided into low-risk and high-risk groups based on the risk score system. By KM curve, the effectiveness and stability of the risk score system were validated in training set (including several subgroups), internal and external validation sets. In the training set, ECOG PS, pathologic T stage, and risk score were combined into a nomogram model, and calibration plots demonstrated robust predictive ability with C-index 0.790. Additionally, time-dependent ROC analysis and DCA for training and testing sets showed that the nomogram was highly discriminatory. Consequently, we are well-equipped to forecast the clinical outcomes for HCC patients with confidence, utilizing our prognostic model that is grounded in the analysis of 5-SERGs.

The 5-SERG prognostic model comprised of CBX2, TPX2, EFNA3, DNASE1L3 and SOCS2, was easy to understand and operate in practical application. In previous studies, the 5 SERGs have been shown to play a role in invasion, malignant phenotypes, and drug resistance of HCC [47,48,49,50,51]. It has been found that TPX2 accelerates sorafenib metabolism or clearance in HCC cells, resulting in resistance to sorafenib [47]; EFNA3 was frequently upregulated in HCC and its overexpression was associated with more aggressive tumor behaviors. EFNA3 functionally contributed to self-renewal, proliferation and migration in HCC cells [48]; as a novel HCC cell cycle regulator and tumor suppressor, DNASE1L3 impairs HCC tumorigenesis by delaying cell cycle progression possibly through disrupting the positive E2F1-CDK2 regulatory loop [49]; it has been demonstrated that high-expressed SOCS2 was one of the biomarkers predicting radiosensitivity of HCC by advancing the ubiquitination degradation of SLC7A11 and promoting ferroptosis [50]. In various cancers, such as breast cancer, esophageal cancer, and gastric cancer, CBX2 was overexpressed and acted as an oncogene [51,52,53,54]. Currently, limited is known about the role of CBX2 in HCC.

According to our study, CBX2 has shown significantly high expression in both cell lines and tissues of HCC. Notably, the abnormally high expression of CBX2 in HCC may be caused by a variety of factors, including genome copy number variation, promoter hypomethylation, chromatin openness, and increased activity markers such as H3K4me3, H3K4me1, and H3K27ac [55]. Furthermore, the expression level of CBX2 is closely related to malignant phenotype of tumors. In vitro experiments have demonstrated that the suppression of CBX2 expression in HCC cells resulted in reduced cell viability, hindered cellular migration and disrupted cell cycle progression, along with an induction of apoptosis. These findings indicate the potential of CBX2 as a novel target for therapeutic intervention in the treatment of HCC. On the mechanism, CBX2 may regulate the proliferation and apoptosis of HCC cells by influencing the phosphorylation of YAP protein in Hippo signaling pathway [51]. Considering the complex and diverse, more investigations are required to further explore the specific mechanisms by which CBX2 mediates the malignant function in HCC, as well as its potential and application prospects as a therapeutic target.

To evaluate the risk score system, we performed GO and KEGG analyses based on DEGs between low-risk and high-risk groups. The results displayed that DEGs were significantly enriched in terms of mRNA processing, receptor ligand activity, cadherin binding, and cell cycle, which were closely related to the occurrence and development of tumors [56,57,58,59]. We then performed GSEA between high-risk and low-risk groups. The results indicated that pathways associated with signaling pathway transduction and epigenetic change were enriched in the high-risk group, while cancer-related metabolism was enriched in the low-risk group. According to several studies, TiME was associated with oncogenesis and prognosis of HCC [60, 61]. For a more comprehensive evaluation of the relationship between TiME and 5-SERG prognostic model, immune analyses were employed.

There was a difference in TiME between the high-risk group and low-risk group. The results revealed that the risk score was negatively correlated with Stromal score and immune response-related cells (such as MAIT and Th17 cells), which are often associated with tumor prognosis and response to treatment [62,63,64]. To achieve a more holistic evaluation of the tumor microenvironment, it might be essential to integrate additional scoring systems and biomarkers into the analysis. Our risk score was positively correlated with TIDE score, which mirror the extent of T cell dysfunction and exclusion within the tumor microenvironment, with elevated scores typically correlating with a less favorable treatment outcome for ICIs and a more adverse prognosis, suggesting that a low-risk population potentially rendering more responsive to ICIs [65, 66].

The study was innovative in some ways: it explored SE as a potential biomarker for predicting the prognosis and ICIs treatment effect of HCC, and it developed a prognostic model based on SERGs; as well as validating the model through internal and external cohorts, preliminary experiments were conducted in vitro to explore the biological function of CBX2. There were also two limitations to this study: firstly, only one representative risk gene (CBX2) in the prognostic model was validated in true bench work, but without in vivo experimentation; secondly, most patients from TCGA-LIHC were pathologic T1 or T2 stage, which might be a cause to the bias.

In conclusion, the risk model, grounded in the analysis of 5-SERGs, provides superior predictive insights into the prognosis and TiME of HCC patients. This advancement significantly aids in the formulation of tailored therapeutic approaches. Looking ahead, research efforts may concentrate on pinpointing high-risk patient populations and elucidating the mechanisms by which ‘cold’ tumors transition into ‘hot’ tumors [67,68,69].

Conclusions

We developed a novel HCC prognostic model utilizing SERGs, indicating that patients with high-risk score not only face a poorer prognosis but also may exhibit a diminished therapeutic response to ICIs. This model is designed to tailor personalized treatment strategies to the individual needs of each patient, thereby enhancing the overall clinical outcomes for HCC patients. Furthermore, CBX2 is a promising candidate for therapeutic intervention in HCC.

Availability of data and materials

The datasets used and/or analyzed in this study are available from the cancer genome database (TCGA-LIHC) (https://portal.gdc.cancer.gov), SEdb 2.0 (https://www.licpathway.net/sedb/) (Sample_02_1299), NCBI Gene Expression Omnibus (GEO: dataset of GSE14520) (https://www.ncbi.nlm.nih.gov/geo/), Human Protein Atlas (HPA) database (http://www.proteinatlas.org/), KEGG pathway database (https://www.kegg.jp/kegg/kegg1.html), STRING (http://string-db.org) (version 11.5), MethSurv database (https://biit.cs.ut.ee/methsurv/), ImmuCellAI database (http://bioinfo.life.hust.edu.cn/ImmuCellAI), Tumor Immune Dysfunction and Exclusion (TIDE) (http://tide.dfci.harvard.edu/), GSEA (http://software.broadinstitute.org/gsea/index.jsp), “ggplot2” package (https://ggplot2.tidyverse.org), TIMER (http://timer.cistrome.org/), Cancer Cell Line Encyclopedia (CCLE) (https://sites.broadinstitute.org/ccle).

References

  1. Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17–48. https://doi.org/10.3322/caac.21763.

    Article  PubMed  Google Scholar 

  2. Haber PK, Puigvehí M, Castet F, Lourdusamy V, Montal R, Tabrizian P, Buckstein M, Kim E, Villanueva A, Schwartz M, Llovet JM. Evidence-based management of Hepatocellular Carcinoma: systematic review and Meta-analysis of Randomized controlled trials (2002–2020). Gastroenterology. 2021;161(3):879–98. https://doi.org/10.1053/j.gastro.2021.06.008.

    Article  CAS  PubMed  Google Scholar 

  3. Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF, de Oliveira AC, Santoro A, Raoul JL, Forner A, Schwartz M, Porta C, Zeuzem S, Bolondi L, Greten TF, Galle PR, Seitz JF, Borbath I, Häussinger D, Giannaris T, Shan M, Moscovici M, Voliotis D, Bruix J, SHARP Investigators Study Group. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med. 2008;359(4):378–90. https://doi.org/10.1056/NEJMoa0708857.

    Article  CAS  PubMed  Google Scholar 

  4. Bruix J, Qin S, Merle P, Granito A, Huang YH, Bodoky G, Pracht M, Yokosuka O, Rosmorduc O, Breder V, Gerolami R, Masi G, Ross PJ, Song T, Bronowicki JP, Ollivier-Hourmand I, Kudo M, Cheng AL, Llovet JM, Finn RS, LeBerre MA, Baumhauer A, Meinhardt G, Han G. RESORCE investigators. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2017;389(10064):56–66. https://doi.org/10.1016/S0140-6736(16)32453-9.

    Article  CAS  PubMed  Google Scholar 

  5. Abou-Alfa GK, Meyer T, Cheng AL, El-Khoueiry AB, Rimassa L, Ryoo BY, Cicin I, Merle P, Chen Y, Park JW, Blanc JF, Bolondi L, Klümpen HJ, Chan SL, Zagonel V, Pressiani T, Ryu MH, Venook AP, Hessel C, Borgman-Hagey AE, Schwab G, Kelley RK. Cabozantinib in patients with Advanced and Progressing Hepatocellular Carcinoma. N Engl J Med. 2018;379(1):54–63. https://doi.org/10.1056/NEJMoa1717002.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Zhu AX, Kang YK, Yen CJ, Finn RS, Galle PR, Llovet JM, Assenat E, Brandi G, Pracht M, Lim HY, Rau KM, Motomura K, Ohno I, Merle P, Daniele B, Shin DB, Gerken G, Borg C, Hiriart JB, Okusaka T, Morimoto M, Hsu Y, Abada PB, Kudo M. REACH-2 study investigators. Ramucirumab after Sorafenib in patients with advanced hepatocellular carcinoma and increased α-fetoprotein concentrations (REACH-2): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2019;20(2):282–96. https://doi.org/10.1016/S1470-2045(18)30937-9.

    Article  CAS  PubMed  Google Scholar 

  7. Wang R, Lin N, Mao B, Wu Q. The efficacy of immune checkpoint inhibitors in advanced hepatocellular carcinoma: a meta-analysis based on 40 cohorts incorporating 3697 individuals. J Cancer Res Clin Oncol. 2022;148(5):1195–210.

    Article  CAS  PubMed  Google Scholar 

  8. Finn RS, Qin S, Ikeda M, Galle PR, Ducreux M, Kim TY, Kudo M, Breder V, Merle P, Kaseb AO, Li D, Verret W, Xu DZ, Hernandez S, Liu J, Huang C, Mulla S, Wang Y, Lim HY, Zhu AX, Cheng AL. IMbrave150 investigators. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. N Engl J Med. 2020;382(20):1894–905.

    Article  CAS  PubMed  Google Scholar 

  9. Casadei-Gardini A, Rimini M, Tada T, Suda G, Shimose S, Kudo M, Cheon J, Finkelmeier F, Lim HY, Rimassa L, Presa J, Masi G, Yoo C, Lonardi S, Tovoli F, Kumada T, Sakamoto N, Iwamoto H, Aoki T, Chon HJ, Himmelsbach V, Pressiani T, Montes M, Vivaldi C, Soldà C, Piscaglia F, Hiraoka A, Sho T, Niizeki T, Nishida N, Steup C, Iavarone M, Di Costanzo G, Marra F, Scartozzi M, Tamburini E, Cabibbo G, Foschi FG, Silletta M, Hirooka M, Kariyama K, Tani J, Atsukawa M, Takaguchi K, Itobayashi E, Fukunishi S, Tsuji K, Ishikawa T, Tajiri K, Ochi H, Yasuda S, Toyoda H, Ogawa C, Nishimura T, Hatanaka T, Kakizaki S, Shimada N, Kawata K, Tada F, Ohama H, Nouso K, Morishita A, Tsutsui A, Nagano T, Itokawa N, Okubo T, Arai T, Imai M, Kosaka H, Naganuma A, Koizumi Y, Nakamura S, Kaibori M, Iijima H, Hiasa Y, Burgio V, Persano M, Della Corte A, Ratti F, De Cobelli F, Aldrighetti L, Cascinu S, Cucchetti A. Atezolizumab plus Bevacizumab versus Lenvatinib for unresectable hepatocellular carcinoma: a large real-life worldwide population. Eur J Cancer. 2023;180:9–20. https://doi.org/10.1016/j.ejca.2022.11.017.

    Article  CAS  PubMed  Google Scholar 

  10. Wang Y, Deng B. Hepatocellular carcinoma: molecular mechanism, targeted therapy, and biomarkers. Cancer Metastasis Rev. 2023;42(3):629–52.

    Article  PubMed  Google Scholar 

  11. Guo DZ, Zhang X, Zhang SQ, Zhang SY, Zhang XY, Yan JY, Dong SY, Zhu K, Yang XR, Fan J, Zhou J, Huang A. Single-cell tumor heterogeneity landscape of hepatocellular carcinoma: unraveling the pro-metastatic subtype and its interaction loop with fibroblasts. Mol Cancer. 2024;23(1):157.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Wang Y, Deng B. Hepatocellular carcinoma: molecular mechanism, targeted therapy, and biomarkers. Cancer Metastasis Rev. 2023;42(3):629–52. https://doi.org/10.1007/s10555-023-10084-4.

    Article  PubMed  Google Scholar 

  13. Parikh ND, Tayob N, Singal AG. Blood-based biomarkers for hepatocellular carcinoma screening: approaching the end of the ultrasound era? J Hepatol. 2023;78(1):207–16. https://doi.org/10.1016/j.jhep.2022.08.036.

    Article  CAS  PubMed  Google Scholar 

  14. Fernández-Palanca P, Payo-Serafín T, Méndez-Blanco C, San-Miguel B, Tuñón MJ, González-Gallego J, Mauriz JL. Neuropilins as potential biomarkers in hepatocellular carcinoma: a systematic review of basic and clinical implications. Clin Mol Hepatol. 2023;29(2):293–319. https://doi.org/10.3350/cmh.2022.0425.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Chen VL, Sharma P. Role of biomarkers and Biopsy in Hepatocellular Carcinoma. Clin Liver Dis. 2020;24(4):577–90. https://doi.org/10.1016/j.cld.2020.07.001.

    Article  PubMed  Google Scholar 

  16. Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-André V, Sigova AA, Hoke HA, Young RA. Super-enhancers in the control of cell identity and disease. Cell. 2013;155(4):934–47. https://doi.org/10.1016/j.cell.2013.09.053.

    Article  CAS  PubMed  Google Scholar 

  17. Whyte WA, Orlando DA, Hnisz D, Abraham BJ, Lin CY, Kagey MH, Rahl PB, Lee TI, Young RA. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell. 2013;153(2):307–19. https://doi.org/10.1016/j.cell.2013.03.035.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Thandapani P. Super-enhancers in cancer. Pharmacol Ther. 2019;199:129–38. https://doi.org/10.1016/j.pharmthera.2019.02.014.

    Article  CAS  PubMed  Google Scholar 

  19. Cao C, Cai Z, Ye R, Su R, Hu N, Zhao H, Xue Y. Global in situ profiling of RNA-RNA spatial interactions with RIC-seq. Nat Protoc. 2021;16(6):2916–46. https://doi.org/10.1038/s41596-021-00524-2.

    Article  CAS  PubMed  Google Scholar 

  20. Zheng C, Liu M, Fan H. Targeting complexes of super-enhancers is a promising strategy for cancer therapy. Oncol Lett. 2020;20(3):2557–66. https://doi.org/10.3892/ol.2020.11855.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Zhang J, Liu W, Zou C, Zhao Z, Lai Y, Shi Z, Xie X, Huang G, Wang Y, Zhang X, Fan Z, Su Q, Yin J, Shen J. Targeting super-enhancer-associated oncogenes in Osteosarcoma with THZ2, a covalent CDK7 inhibitor. Clin Cancer Res. 2020;26(11):2681–92. https://doi.org/10.1158/1078-0432.CCR-19-1418.

    Article  CAS  PubMed  Google Scholar 

  22. Li XP, Qu J, Teng XQ, Zhuang HH, Dai YH, Yang Z, Qu Q. The emerging role of super-enhancers as therapeutic targets in the Digestive System tumors. Int J Biol Sci. 2023;19(4):1036–48. https://doi.org/10.7150/ijbs.78535.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Liu Q, Guo L, Lou Z, Xiang X, Shao J. Super-enhancers and novel therapeutic targets in colorectal cancer. Cell Death Dis. 2022;13(3):228. https://doi.org/10.1038/s41419-022-04673-4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Qi SH, Wang QL, Zhang JY, Liu Q, Li CY. The regulatory mechanisms by enhancers during cancer initiation and progression. Yi Chuan. 2022;44(4):275–88. https://doi.org/10.16288/j.yczz.21-440.

    Article  PubMed  Google Scholar 

  25. Li J, Wang J, Wang Y, Zhao X, Su T. E2F1 combined with LINC01004 super-enhancer to promote hepatocellular carcinoma cell proliferation and metastasis. Clin Epigenetics. 2023;15(1):17. https://doi.org/10.1186/s13148-023-01428-6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Tsang FH, Law CT, Tang TC, Cheng CL, Chin DW, Tam WV, Wei L, Wong CC, Ng IO, Wong CM. Aberrant Super-enhancer Landscape in Human Hepatocellular Carcinoma. Hepatology. 2019;69(6):2502–17. https://doi.org/10.1002/hep.30544.

    Article  CAS  PubMed  Google Scholar 

  27. Bian G, Cao J, Li W, Huang D, Ding X, Zang X, Ye Y, Li P. Identification and validation of a Cancer-Testis Antigen-related signature to predict the prognosis in stomach adenocarcinoma. J Cancer. 2024;15(11):3596–611.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Gui Z, Ye Y, Li Y, Ren Z, Wei N, Liu L, Wang H, Zhang M. Construction of a novel cancer-associated fibroblast-related signature to predict clinical outcome and immune response in cervical cancer. Transl Oncol. 2024;46:102001.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Wei X, Zhou Z, Long M, Lin Q, Qiu M, Chen P, Huang Q, Qiu J, Jiang Y, Wen Q, Liu Y, Li R, Nong C, Guo Q, Yu H, Zhou X. A novel signature constructed by super-enhancer-related genes for the prediction of prognosis in hepatocellular carcinoma and associated with immune infiltration. Front Oncol. 2023;13:1043203.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Love MI, Huber W, Anders S. Moderated estimation of Fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. https://doi.org/10.1186/s13059-014-0550-8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16(5):284–7. https://doi.org/10.1089/omi.2011.0118.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Franceschini A, Szklarczyk D, Frankild S, Kuhn M, Simonovic M, Roth A, Lin J, Minguez P, Bork P, von Mering C, Jensen LJ. STRING v9.1: protein-protein interaction networks, with increased coverage and integration. Nucleic Acids Res. 2013;41(Database issue):D808–15. https://doi.org/10.1093/nar/gks1094.

    Article  CAS  PubMed  Google Scholar 

  33. Liu J, Lichtenberg T, Hoadley KA, Poisson LM, Lazar AJ, Cherniack AD, Kovatich AJ, Benz CC, Levine DA, Lee AV, Omberg L, Wolf DM, Shriver CD, Thorsson V, Cancer Genome Atlas Research Network, Hu H. An Integrated TCGA Pan-cancer Clinical Data Resource to Drive High-Quality Survival Outcome Analytics. Cell. 2018;173(2):400–e41611. https://doi.org/10.1016/j.cell.2018.02.052.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Friedman J, Hastie T, Tibshirani R. Regularization paths for generalized Linear models via Coordinate Descent. J Stat Softw. 2010;33(1):1–22.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Martens JW, Margossian AL, Schmitt M, Foekens J, Harbeck N. DNA methylation as a biomarker in breast cancer. Future Oncol. 2009;5(8):1245–56. https://doi.org/10.2217/fon.09.89.

    Article  CAS  PubMed  Google Scholar 

  36. Modhukur V, Iljasenko T, Metsalu T, Lokk K, Laisk-Podar T, Vilo J. MethSurv: a web tool to perform multivariable survival analysis using DNA methylation data. Epigenomics. 2018;10(3):277–88. https://doi.org/10.2217/epi-2017-0118.

    Article  CAS  PubMed  Google Scholar 

  37. Yoshihara K, Shahmoradgoli M, Martínez E, Vegesna R, Kim H, Torres-Garcia W, Treviño V, Shen H, Laird PW, Levine DA, Carter SL, Getz G, Stemke-Hale K, Mills GB, Verhaak RG. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013;4:2612. https://doi.org/10.1038/ncomms3612.

    Article  CAS  PubMed  Google Scholar 

  38. Miao YR, Zhang Q, Lei Q, Luo M, Xie GY, Wang H, Guo AY. ImmuCellAI: a Unique Method for Comprehensive T-Cell subsets abundance prediction and its application in Cancer Immunotherapy. Adv Sci (Weinh). 2020;7(7):1902880. https://doi.org/10.1002/advs.201902880.

    Article  CAS  PubMed  Google Scholar 

  39. Li T, Fu J, Zeng Z, Cohen D, Li J, Chen Q, Li B, Liu XS. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 2020;48(W1):W509–14. https://doi.org/10.1093/nar/gkaa407.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50. https://doi.org/10.1073/pnas.0506580102.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Koutsi MA, Pouliou M, Champezou L, Vatsellas G, Giannopoulou AI, Piperi C, Agelopoulos M. Typical Enhancers, super-enhancers, and cancers. Cancers (Basel). 2022;14(18):4375.

    Article  CAS  PubMed  Google Scholar 

  42. Yang Z, Liu Y, Cheng Q, Chen T. Targeting super enhancers for liver disease: a review. PeerJ. 2023;11:e14780. https://doi.org/10.7717/peerj.14780.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Bacabac M, Xu W. Oncogenic super-enhancers in cancer: mechanisms and therapeutic targets. Cancer Metastasis Rev. 2023;42(2):471–80. https://doi.org/10.1007/s10555-023-10103-4.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Wang M, Chen Q, Wang S, Xie H, Liu J, Huang R, Xiang Y, Jiang Y, Tian D, Bian E. Super-enhancers complexes zoom in transcription in cancer. J Exp Clin Cancer Res. 2023;42(1):183. https://doi.org/10.1186/s13046-023-02763-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Yuan XQ, Zhou N, Wang JP, Yang XZ, Wang S, Zhang CY, Li GC, Peng L. Anchoring super-enhancer-driven oncogenic lncRNAs for anti-tumor therapy in hepatocellular carcinoma. Mol Ther. 2023;31(6):1756–74. https://doi.org/10.1016/j.ymthe.2022.11.013.

    Article  CAS  PubMed  Google Scholar 

  46. Su T, Zhang N, Wang T, Zeng J, Li W, Han L, Yang M. Super enhancer-regulated LncRNA LINC01089 induces alternative splicing of DIAPH3 to Drive Hepatocellular Carcinoma Metastasis. Cancer Res. 2023;83(24):4080–94. https://doi.org/10.1158/0008-5472.CAN-23-0544.

    Article  PubMed  Google Scholar 

  47. Wang H, Chu F, Zhang XF, Zhang P, Li LX, Zhuang YL, Niu XF, He X, Li ZJ, Bai Y, Mao D, Liu ZW, Zhang DL, Li BA. TPX2 enhances the transcription factor activation of PXR and enhances the resistance of hepatocellular carcinoma cells to antitumor drugs. Cell Death Dis. 2023;14(1):64. https://doi.org/10.1038/s41419-022-05537-7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Husain A, Chiu YT, Sze KM, Ho DW, Tsui YM, Suarez EMS, Zhang VX, Chan LK, Lee E, Lee JM, Cheung TT, Wong CC, Chung CY, Ng IO. Ephrin-A3/EphA2 axis regulates cellular metabolic plasticity to enhance cancer stemness in hypoxic hepatocellular carcinoma. J Hepatol. 2022;77(2):383–96. https://doi.org/10.1016/j.jhep.2022.02.018.

    Article  CAS  PubMed  Google Scholar 

  49. Sun J, Wang X, Shen Q, Wang M, Chen S, Zhang X, Huang Y, Zhang Z, Li W, Yuan Y, Huang Z. DNASE1L3 inhibits hepatocellular carcinoma by delaying cell cycle progression through CDK2. Cell Oncol (Dordr). 2022;45(6):1187–202. https://doi.org/10.1007/s13402-022-00709-1.

    Article  CAS  PubMed  Google Scholar 

  50. Chen Q, Zheng W, Guan J, Liu H, Dan Y, Zhu L, Song Y, Zhou Y, Zhao X, Zhang Y, Bai Y, Pan Y, Zhang J, Shao C. SOCS2-enhanced ubiquitination of SLC7A11 promotes ferroptosis and radiosensitization in hepatocellular carcinoma. Cell Death Differ. 2023;30(1):137–51. https://doi.org/10.1038/s41418-022-01051-7.

    Article  CAS  PubMed  Google Scholar 

  51. Mao J, Tian Y, Wang C, Jiang K, Li R, Yao Y, Zhang R, Sun D, Liang R, Gao Z, Wang Q, Wang L. CBX2 regulates proliferation and apoptosis via the phosphorylation of YAP in Hepatocellular Carcinoma. J Cancer. 2019;10(12):2706–19. https://doi.org/10.7150/jca.31845.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Zheng S, Lv P, Su J, Miao K, Xu H, Li M. Overexpression of CBX2 in breast cancer promotes tumor progression through the PI3K/AKT signaling pathway. Am J Transl Res. 2019;11(3):1668–82.

    CAS  PubMed  PubMed Central  Google Scholar 

  53. Peng L, Huang X, Qing D, Lu H, Liu X, Chen J, Long X, Pang Q. MiR-30a-5p inhibits cell behaviors in esophageal cancer via modulating CBX2. Mutat Res. 2023;826:111818. https://doi.org/10.1016/j.mrfmmm.2023.111818.

    Article  CAS  PubMed  Google Scholar 

  54. Zeng M, Li B, Yang L, Guan Q. CBX2 depletion inhibits the proliferation, invasion and migration of gastric cancer cells by inactivating the YAP/β-catenin pathway. Mol Med Rep. 2021;23(2):137. https://doi.org/10.3892/mmr.2020.11776.

    Article  CAS  PubMed  Google Scholar 

  55. Meng Q, Zhou Q, Chen X, Chen J. Prognostic hub gene CBX2 drives a cancer stem cell-like phenotype in HCC revealed by multi-omics and multi-cohorts. Aging. 2023;15(22):12817–51.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Zaccara S, Ries RJ, Jaffrey SR. Reading, writing and erasing mRNA methylation. Nat Rev Mol Cell Biol. 2019;20(10):608–24. https://doi.org/10.1038/s41580-019-0168-5.

    Article  CAS  PubMed  Google Scholar 

  57. Mendonsa AM, Na TY, Gumbiner BM. E-cadherin in contact inhibition and cancer. Oncogene. 2018;37(35):4769–80. https://doi.org/10.1038/s41388-018-0304-2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Liu J, Peng Y, Wei W. Cell cycle on the crossroad of tumorigenesis and cancer therapy. Trends Cell Biol. 2022;32(1):30–44. https://doi.org/10.1016/j.tcb.2021.07.001.

    Article  CAS  PubMed  Google Scholar 

  59. Murray IA, Patterson AD, Perdew GH. Aryl hydrocarbon receptor ligands in cancer: friend and foe. Nat Rev Cancer. 2014;14(12):801–14. https://doi.org/10.1038/nrc3846.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Donne R, Lujambio A. The liver cancer immune microenvironment: therapeutic implications for hepatocellular carcinoma. Hepatology. 2023;77(5):1773–96. https://doi.org/10.1002/hep.32740.

    Article  PubMed  Google Scholar 

  61. Chen C, Wang Z, Ding Y, Qin Y. Tumor microenvironment-mediated immune evasion in hepatocellular carcinoma. Front Immunol. 2023;14:1133308. https://doi.org/10.3389/fimmu.2023.1133308.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Mao M, Yu Q, Huang R, Lu Y, Wang Z, Liao L. Stromal score as a prognostic factor in primary gastric cancer and close association with tumor immune microenvironment. Cancer Med. 2020;9(14):4980–90.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  63. Duan M, Goswami S, Shi JY, Wu LJ, Wang XY, Ma JQ, Zhang Z, Shi Y, Ma LJ, Zhang S, Xi RB, Cao Y, Zhou J, Fan J, Zhang XM, Gao Q. Activated and exhausted MAIT cells Foster Disease Progression and Indicate Poor Outcome in Hepatocellular Carcinoma. Clin Cancer Res. 2019;25(11):3304–16.

    Article  PubMed  Google Scholar 

  64. Li S, Na R, Li X, Zhang Y, Zheng T. Targeting interleukin-17 enhances tumor response to immune checkpoint inhibitors in colorectal cancer. Biochim Biophys Acta Rev Cancer. 2022;1877(4):188758.

    Article  CAS  PubMed  Google Scholar 

  65. Raskov H, Orhan A, Christensen JP, Gögenur I. Cytotoxic CD8 + T cells in cancer and cancer immunotherapy. Br J Cancer. 2021;124(2):359–67. https://doi.org/10.1038/s41416-020-01048-4.

    Article  CAS  PubMed  Google Scholar 

  66. Donne R, Lujambio A. The liver cancer immune microenvironment: therapeutic implications for hepatocellular carcinoma. Hepatology. 2023;77(5):1773–96.

    PubMed  Google Scholar 

  67. Hapuarachi B, Danson S, Wadsley J, Muthana M. Exercise to transform tumours from cold to hot and improve immunotherapy responsiveness. Front Immunol. 2023;14:1335256. https://doi.org/10.3389/fimmu.2023.1335256.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. Xuan C, Hu R. Chemical Biology perspectives on STING agonists as Tumor Immunotherapy. ChemMedChem. 2023;18(23):e202300405. https://doi.org/10.1002/cmdc.202300405.

    Article  CAS  PubMed  Google Scholar 

  69. Yamauchi M, Ono A, Amioka K, Fujii Y, Nakahara H, Teraoka Y, Uchikawa S, Fujino H, Nakahara T, Murakami E, Okamoto W, Miki D, Kawaoka T, Tsuge M, Imamura M, Hayes CN, Ohishi W, Kishi T, Kimura M, Suzuki N, Arihiro K, Aikata H, Chayama K, Oka S. Lenvatinib activates anti-tumor immunity by suppressing immunoinhibitory infiltrates in the tumor microenvironment of advanced hepatocellular carcinoma. Commun Med (Lond). 2023;3(1):152. https://doi.org/10.1038/s43856-023-00390-x.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

This work was supported by Natural Science Foundation of Fujian Province (No. 2022J01211).

Author information

Author notes

  1. Qing Wu, Ping Li and Xuan Tao contributed equally to this work.

Authors and Affiliations

  1. Department of Oncology, The First Affiliated Hospital of Fujian Medical University, No. 20 Chazhong Road, Fuzhou, 350005, China

    Qing Wu, Ping Li & Xianhe Xie

  2. Department of Oncology, National Regional Medical Center, Binhai Campus of The First Affiliated Hospital, Fujian Medical University, Fuzhou, 350212, China

    Qing Wu & Xianhe Xie

  3. Molecular Oncology Research Institute, The First Affiliated Hospital of Fujian Medical University, Fuzhou, 350005, China

    Qing Wu & Xianhe Xie

  4. Department of Pathology, The First Affiliated Hospital of Fujian Medical University, Fuzhou, China

    Xuan Tao

  5. Fuzong Clinical Medical College of Fujian Medical University, Fuzhou, Fujian, China

    Nan Lin

  6. Department of Gastrointestinal Surgery, The 900th Hospital of Joint Logistics Support Forces of Chinese PLA, Fuzhou, Fujian, China

    Nan Lin

  7. Department of Interventional Radiology, National Regional Medical Center, Binhai Campus of The First Affiliated Hospital, Fujian Medical University, Fuzhou, 350212, China

    BinBin Mao

  8. Fujian Key Laboratory of Precision Medicine for Cancer, The First Affiliated Hospital, Fujian Medical University, Fuzhou, 350005, China

    Xianhe Xie

Authors
  1. Qing Wu
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Ping Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Xuan Tao
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Nan Lin
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. BinBin Mao
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Xianhe Xie
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

Qing Wu: Conceptualization, Methodology, Software, RT-qPCR, Western blot, Data curation, Formal analysis, Writing-Original Draft; Ping Li: Software, RT-qPCR, Western blot; CCK-8 assay, cellular migration assay, cell cycle assay and apoptosis assay; Xuan Tao: Software, IHC; Nan Lin: Methodology, Visualization; BinBin Mao: Visualization; Xianhe Xie: Conceptualization, Validation, Writing-Review & Editing. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Xianhe Xie.

Ethics declarations

Ethics approval and consent to participate

Our study was performed in accordance with the Declaration of Helsinki and was approved by Branch for Medical Research and Clinical Technology Application, Ethics Committee of the First Affiliated Hospital of Fujian Medical University (Approval No.: MRCTA, ECFAH of FMU [2021]477). Patients were informed that the resected specimens were stored by the hospital and potentially used for scientific research, and that their privacy would be maintained. All patients provided informed consent prior to undergoing surgery.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wu, Q., Li, P., Tao, X. et al. A novel super-enhancer-related risk model for predicting prognosis and guiding personalized treatment in hepatocellular carcinoma. BMC Cancer 24, 1087 (2024). https://doi.org/10.1186/s12885-024-12874-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12885-024-12874-7

Keywords

BMC Cancer

ISSN: 1471-2407

Contact us