Identification of a Interferon-Stimulated gene Signature for Predicting Prognosis, Tumor Microenvironment, and Drug Candidates in Hepatocellular Carcinoma

doi:10.21203/rs.3.rs-3014889/v1

Download PDF

Research Article

Identification of a Interferon-Stimulated gene Signature for Predicting Prognosis, Tumor Microenvironment, and Drug Candidates in Hepatocellular Carcinoma

https://doi.org/10.21203/rs.3.rs-3014889/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Interferon-stimulated genes (ISGs) play critical roles not only in antiviral defense and adaptive immunity but also in the progression of cancer and the immune response. However, there is limited research delineating the relationship between ISGs and HCC prognosis, tumor microenvironment (TME), and response to immunotherapy.

Methods

The transcriptional and relevant clinical data of HCC were downloaded from The Cancer Genome Atlas (TCGA) and The International Cancer Genome Consortium (ICGC) databases, which were used for internal and external validation, respectively. First, ISGs that were differentially expressed in HCC tissues compared to adjacent non-tumor tissues and were also associated with prognosis were screened. Second, the prognostic model based on ISGs was constructed using the least absolute shrinkage and selection operator (LASSO) Cox regression analysis. Next, we analyzed the relationship between the prognosis model and clinical outcomes, clinical and pathological features, immune microenvironment, and response to immunotherapy. Finally, the expression levels of three ISGs were validated by real-time PCR in normal and HCC cell lines.

Results

Three ISGs (BUB1, NDC80, and SOCS2) were selected to establish the prognostic model. The model has good predictive power for clinical outcomes, clinical and pathological features, gene mutations, tumor microenvironment, and response to immunotherapy. The ROC curve analysis confirmed the predictive efficacy of the model. Furthermore, the results of real-time PCR showed that BUB1 and NDC80 were highly expressed in tumor cell lines, and SOCS2 was highly expressed in normal liver cell lines.

Conclusion

The prognostic model based on three ISGs can accurately predict the clinical outcomes, clinical and pathological features, gene mutations, tumor microenvironment, and response to immunotherapy in HCC patients.

Hepatocellular carcinoma

Interferon-Stimulated Genes

Prognosis

Signature

Tumor environment

Immunotherapy

Hepatocellular carcinoma (HCC) is the sixth most common cancer worldwide and the third leading cause of cancer-related death, with almost 85% of cases occurring in low- or middle-resource countries, particularly in Eastern Asia and sub-Saharan Africa [1]. In the past decade, although the use of targeted therapies and immune checkpoint inhibitors has led to significant improvements in the treatment of liver cancer, the overall survival of patients remains unsatisfactory [2]. Therefore, it is particularly important to explore new biomarkers for predicting patient survival outcomes and drug sensitivity to improve the prognosis of HCC.

Interferon-stimulated genes (ISGs) are genes induced by interferons (IFNs) [3]. When IFNs bind to their homologous cell surface receptor, they signal through the JAK/STAT pathway to activate the heterotrimeric transcription factor complex ISGF3 and IFNγ-activated factor (GAF), which in turn activate the expression of ISGs. Many of the protein products encoded by ISGs achieve one or more cellular outcomes, including antiviral defense, anti-proliferative activity, and stimulation of adaptive immunity [4]. Recently, it has been found that ISGs also play a crucial role in regulating tumor immunity. Studies have shown that the expression of ISGs in breast cancer cells increases after treatment with tamoxifen and irradiation, leading to the occurrence of acquired resistance and being closely related to the prognosis of patients [5]. In melanoma, the expression of ISGs in tumor cells can predict resistance to immune checkpoint blocker (ICB) treatment, and the expression of ISGs in immune cells can predict the response of the body to treatment [6]. In addition, ISGs are also closely related to the efficacy of ICB treatment and the prognosis of ovarian cancer [7]. Given the important role played by ISGs in the occurrence and development of tumors, exploring the relationship between ISGs and the prognosis of HCC patients is of great value.

The development of tumors is closely related to the mutation of oncogenes. Genome sequence analysis reveals the widespread occurrence of mutations in epigenetic regulatory genes in tumors [8]. The molecular mechanism of HCC is very complex, including genetic and epigenetic changes, chromosomal aberrations, mutations, and altered molecular pathways [9]. Therefore, determining the genomics of HCC can help us further improve the classification of HCC according to molecular features and develop more effective treatments to improve the survival rate of HCC patients [10]. Up to now, several studies have elucidated the prognostic impact of different genes and established relevant prognostic models by analyzing genomic data of HCC patients [11–13]. However, There is no establishment of a HCC prognosis model based on ISGs.

In this study, we developed the first risk score model based on ISGs by analyzing high-throughput sequencing data of HCC patients. Subsequently, we analyzed the relationship between the risk score model and some factors, including clinical features, immune infiltration and treatment sensitivity of patients. Finally, we further validated the differential expression of score-related ISGs in human normal liver cell lines and tumor cell lines. This study systematically elucidated the impact of ISGs on the prognosis of HCC patients, providing potential targets for the treatment and prognosis of HCC.

2.1 Data Collection

The mRNA and clinical data of 370 HCC samples and 50 normal samples were downloaded from the TCGA database (https://portal.gdc.cancer.gov/). And mRNA and clinical information data from other samples were obtained from the ICGC for external validation. The data were in fragments per kilobase of exon model per million mapped fragments (FPKM) format, normalized by reads per kilobase per million mapped reads (RPKM). The study strictly follows the TCGA and ICGC database access policies and publication guidelines. The list of ISGs was collected from previous studies and a total of 312 ISGs were included in the analysis.

2.2 Construction and validation of the ISGs prognostic model

ISGs showing differential expression between tumor and non-tumor tissues were screened using the "limma" package in R language. Genes with a |log2FC| > 1 and FDR < 0.05 were selected as differentially expressed genes. Univariate Cox regression analysis was performed on overall survival (OS) to determine the ISGs that have prognostic value. Next, the differentially expressed genes were intersected with prognosis-associated genes to obtain overlapping genes. To minimize the risk of overfitting, Lasso regression analysis was performed with the "glmnet" package in R language. The risk score of each patient was calculated as follows:

Risk score = Exp1β1 + Exp2β2 + Exp3β3 + … + Expxβx, where Exp represents the corresponding gene expression level for prognosis and β is the Lasso regression coefficient.

Patients were divided into high-risk (HR) and low-risk (LR) groups based on the median value of the risk score. The "survivalROC" package in R language was used to plot the survival curve, and the "timeROC" package was used for ROC curve analysis to evaluate the predictive ability of the ISGs prognostic model for OS, PFS, DSS, and DFS. To validate the efficacy of the model, patients from TCGA were used for internal validation, and subgroup analysis was performed based on clinical and pathological features, including AFP, vascular invasion, poor tumor differentiation, advanced clinical stage, and so on. Patients from ICGC were used for external validation.

2.3 Functional analysis based on model grouping

To explore the universality of the risk score among groups with different clinical and pathological features, subgroup analysis was performed. The TCGA patients were divided into different subgroups based on clinical and pathological features, and Kaplan-Meier survival analysis was performed in each subgroup. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted to investigate the cellular composition, molecular function, biological processes, and signaling pathways of differentially expressed genes. Then, mutation data (MAF format) and corresponding clinical information for HCC samples were obtained from The Cancer Genome Atlas (TCGA) dataset (https://portal.gdc.cancer.gov/), and the differences in mutation between the HR and LR groups were analyzed. The tumor mutation burden (TMB) scores of HCC patients were computed by somatic mutation analysis. The tumor immune dysfunction and exclusion (TIDE) algorithm was used to predict the clinical response to immune checkpoint inhibitors between the two groups. To assess the degree of HCC stemness, the logistic regression machine learning algorithm (OCLR) developed by Malta et al. was used to calculate the stemness indexes [14]. Next, immunocyte infiltration levels and immune activation were evaluated by single-sample gene set enrichment analysis (ssGSEA). The expression levels of immune checkpoint-related genes were analyzed between the HR and LR groups. Finally, the sensitivity of the HR and LR groups to different anticancer drugs was predicted based on the data from the largest publicly available drug genomics database, the Genomicsof Drug Sensitivity in Cancer (GDSC) (https://www.cancerrxgene.org/).

2.4 Cell culture

Human hepatocellular carcinoma cell Hep3B, Huh-7, HCCLM3 and normal human liver cell QSG7701 were obtained from the American Type Culture Collection (ATCC, USA), and cultured in DMEM supplemented with 10% fetal bovine serum, 100 units/mL penicillin and 100 ug/mL streptomycin. All cells were maintained in a humidified incubator with an atmosphere containing 5% CO₂ at 37°C.

2.5 RNA extraction and quantitative real-time PCR

The total RNA of cells was extracted by TRIzol (Cat# R401-01, Vazyme, Nanjing, China), and cDNA was synthesized using HiScript III RT SuperMix for qPCR (+ gDNA wiper) (Cat# R323-01, Vazyme, Nanjing, China) according to the manufacturer’s instructions. Quantitative real-time PCR was performed using ChamQ SYBR qPCR Master Mix (High ROX Premixed) (Cat#Q341-02, Vazyme, Nanjing, China) on the real-time PCR system (Applied QuantStudio™ 5 Real-Time PCR System, Thermo Fisher Scientific, USA). The relative expression was normalized to GAPDH by the 2-ΔΔCt method. The primer sequences are listed in Table 1.

Table 1

Primer list
Gene	Primer sequence
BUB1 (forward)	ACAATCAACGGAGAAAGCATGA
BUB1 (reverse)	CTCCACCACCTGATGCAACT
NDC80 (forward)	TCAAGGACCCGAGACCACTTA
NDC80 (reverse)	GGGAGCTTGTAGAGATTTCATGG
SOCS2 (forward)	TTAAAAGAGGCACCAGAAGGAAC
SOCS2 (reverse)	AGTCGATCAGATGAACCACACT
GAPDH (forward)	CATGCCTTCTTGCCTCTTGTCTCTTAGAT
GAPDH (reverse)	CATGCCTTCTTGCCTCTTGTCTCTTAGAT

2.5 Statistical analysis

The correlation between the risk score and clinical and pathological features was analyzed using Student's t-test. Fisher's exact test or Pearson's chi-square test was used to analyze qualitative variables. Significant differences between two groups were compared using the Wilcoxon test, and those of three or more groups were compared using the Kruskal-Wallis test. The Mann-Whitney test with p values adjusted by Benjamini and Hochberg (BH) correction was used to measure the ssGSEA scores between the groups. Kaplan-Meier curves were plotted to analyze OS, PFS, DFS, and DSS, and a log-rank test was used to compare survival among different groups. Univariate and multivariate Cox regression analyses were performed to analyze independent prognostic factors for OS. A p-value < 0.05 was considered statistically significant. Data analysis and image plotting were performed using R language (4.1.2) and GraphPad (8.0.1).

3.1 Construction of the Prognostic Model based on three ISGs

There were 313 interferon-stimulated genes (ISGs) in the TCGA database, among which 15 genes were differentially expressed in HCC tissues compared to adjacent non-tumor tissues (Fig. 1A, B). Prognostic analysis showed that 93 genes were associated with tumor prognosis. The intersection of 15 differentially expressed genes and 93 prognosis-associated genes was taken, resulting in 4 overlapping differentially expressed genes, including BLVRA, BUB1, NDC80, and SOCS2. Then univariate Cox regression analysis was used to retain genes that were prognostic for OS, and a three-gene prognostic model was constructed through LASSO regression analysis (Fig. 1C). The risk score formula of the prognostic model is as follows:

Risk score = (0.1092)*BUB1 + (0.2351)*NDC80 + (-0.2909)*SOCS2.

BUB1 and NDC80 were identified as high-risk genes that were highly expressed in tumor tissues and associated with poor prognosis, while SOCS2 was identified as a protective gene that was down-regulated in tumor tissues and indicative of a better prognosis (Fig. 1D). Then the risk score of each patient in TCGA was calculated using this model. Univariate and multivariate Cox regression analysis showed that the risk score was an independent prognostic indicator for OS, PFS, DFS, and DSS (Fig. 1E-H).

3.2 The model predicts the Overall survival of HCC patients

The risk score for each patient was calculated based on the risk score formula. According to the median value of the risk scores, patients in the training set were divided into HR and LR groups. The scatter plot shows the patient survival rates based on the risk score, and the heatmap further illustrates the expression differences of the three genes between the two groups (Fig. 2A). In the TCGA training cohort, the HR group had a better OS than the LR group (p < 0.001) (Fig. 2B). The AUC values of the risk score predicting the 1-year, 2-year, and 3-year OS of patients were 0.772, 0.744, and 0.713, respectively (Fig. 2B). Using the ICGC dataset as a validation cohort, the high risk score also represented a poorer prognosis. The AUC values of the risk score predicting the 1-year, 2-year, and 3-year OS of patients were 0.683, 0.713, and 0.692, respectively (Fig. 2C). The results revealed that the risk score was the independent prognostic factor of OS.

3.3 The model significantly correlated with the clinical and pathological features

To elucidate the relationship between risk scores and clinical and pathological features, we performed group analysis based on different clinical and pathological features (Fig. 3A-F). The results showed that higher risk scores were significantly associated with the presence of Asian ethnicity, AFP ≥ 25ng/ml, vascular invasion, Edmondson G3-4, T stage 3–4, and TNM stage Ⅲ-Ⅳ (P < 0.05).

Subsequently, patients were divided into 16 subgroups based on age, sex, ethnicity, AFP, tumor T stage, TNM stage, Edmondson grade, and presence of vascular invasion. As revealed by survival analyses, patients in the HR group had a shorter OS than those in the LR group, and this outcome was achieved in each group. In addition, the AUC values of 1-year, 2-year, and 3-year OS prediction using the risk score were not less than 0.6 (Fig. 4). These results suggested that the model was a reliable index for predicting clinical and pathological features.

3.4 The independent prognostic value of the model for PFS, DFS, and DSS

To clarify the role of the risk score in predicting PFS, DFS, and DSS, we conducted further survival analysis. The results showed that patients in the HR group had significantly shorter PFS, DFS, and DSS compared with those in the LR group (P < 0.0001). In the TCGA dataset, the AUC values of the risk score for predicting 1-year, 2-year, and 3-year PFS were 0.721, 0.678, and 0.603, respectively (Fig. 5A). The AUC values for predicting 1-year, 2-year, and 3-year DFS were 0.733, 0.682, and 0.605, respectively (Fig. 5B). The AUC values for predicting 1-year, 2-year, and 3-year DSS were 0.849, 0.795, and 0.743, respectively (Fig. 5C). These results indicate that the risk score not only has excellent predictive ability for OS but also has excellent predictive ability for PFS, DFS, and DSS in HCC patients.

3.5 The model predicts gene mutations

To determine the differences in the mutational profile between the two groups, we analyzed the occurrence of gene mutations in each group. The results showed that the somatic mutation rate in the HR group was 90.5% (162 of 179 samples), with missense mutations mainly found in the TP53 (37%), TTN, CTNNB1, MUC16, and CSMD3 genes. In contrast, the somatic mutation rate in the LR group was 84.83% (151 of 178 samples), with missense mutations mainly found in the TTN, CTNNB1, MUC16, PCLO, and MUC5B genes. The mutation frequency of TP53 was lower in the LR group, at only 11% (Fig. 6A). Compared with the LR group, the oxidative phosphorylation and adipogenesis signaling pathways were significantly activated in the HR group, while the E2F targets and G2M checkpoint signaling pathways were significantly suppressed (Fig. 6B, C). The TMB in the HR group was significantly higher than that in the LR group, and high TMB was associated with poor prognosis (Fig. 6D). Next, we calculated the stemness score of the patients using the OCLR algorithm. Compared with the LR group and normal tissues, the stemness score was significantly higher in the HR group, which often indicated a poorer prognosis (Fig. 6E).

3.6 The model predicts immune microenvironment

To further investigate the instructive role of ISGs in the immune microenvironment, it is necessary to study the types of infiltrating immune cells in HCC patients. By analyzing the immune cell infiltration of HCC patients in the TCGA database, we found that the HR group had higher levels of common lymphoid progenitor, memory CD4 + T-cell, and CD4 + T-cell (Th2) infiltration, while the levels of endothelial cell and granulocyte-monocyte progenitor infiltration were lower (Fig. 7A). Consistently, the enrichment of common lymphoid progenitors, memory CD4 + T-cell, and CD4 + T-cell (Th2)often indicated poor prognosis (Fig. 7B-D), while the enrichment of endothelial cells and granulocyte-monocyte progenitors was associated with better prognosis (Fig. 7E, F). In addition, we found that the immune score of the LR group was lower than that of the HR group, but the stroma score and microenvironment score were higher than those of the HR group (Fig. 7G). Similarly, higher stroma and microenvironment scores often indicated better prognosis (Fig. 7H and I), while there was no significant correlation between the immune score and the prognosis (Fig. 7J).

3.7 The model predicts the response to immunotherapy

We found that the expression levels of PD1, CTLA4, TIM-3, LAG3, and TIGIT were significantly higher in the HR group than in the LR group (Fig. 8A-H). Then, we used the TIDE algorithm to predict that the HR group might have a better response to immune checkpoint inhibitors (Fig. 8I). We also compared the difference in sensitivity to anticancer drugs between the HR group and the LR group by calculating the IC₅₀ values of the drugs. We found that the IC₅₀ values of targeted drugs (sorafenib, gefitinib, and imatinib) and chemotherapy drugs (cisplatin, docetaxel, gemcitabine, and methotrexate) were significantly lower in the HR group than in the LR group (Fig. 8J-P). These findings suggested that the HR group may benefit more from targeted therapy, immunotherapy, and chemotherapy.

3.8 The expression of BUB1, NDC80, and SOCS2 in normal and tumor cells.

As shown in Fig. 1E, BUB1 and NDC80 were highly expressed in tumor tissues, while SOCS2 was lowly expressed. However, the results were only based on the analysis of public databases. Therefore, we further validated the mRNA expression of these genes in the human normal liver cell line QSG-7701 and liver cancer cell lines Hep3B, Huh7, and LM3. Consistent with the analysis results of the public database, BUB1 and NDC80 were highly expressed in tumor cell lines, while SOCS2 was highly expressed in normal liver cell lines (Fig. 9A-C).

HCC is a highly malignant disease with a poor prognosis. Currently, scholars have developed some scoring systems to predict the prognosis of HCC patients and guide the selection of treatment, such as the Barcelona Clinic Liver Cancer (BCLC), Child-Pugh score, and Cancer of the Liver Italian Program (CLIP). However, the utility of these scoring systems in predicting the survival rates of late-stage HCC patients is usually limited [15]. Therefore, establishing an accurate and effective HCC prognosis evaluation system is essential. Several studies have revealed that ISGs are closely related to the prognosis of various tumors and the effect of treatments such as radiotherapy and immunotherapy [5–7, 16]. However, the role of ISGs in HCC remains unclear. In this study, we constructed a novel prognostic model based on three ISGs, revealing the value of ISGs in predicting the survival outcomes of HCC patients for the first time.

As is well known, the accumulation of genomic alterations can alter the molecular and cellular biology processes of cells, driving the occurrence and development of HCC. At the same time, these changes can be used as biomarkers for predicting treatment response and evaluating tumor prognosis [17]. In our study, we identified three important genes (BUB1, NDC80, and SOCS2) that are differentially expressed in tumor tissue and adjacent normal tissue and are also correlated with the survival outcome of HCC patients. These genes play important roles not only in the interferon response but also in tumorigenesis and tumor progression. BUB1, a mitotic serine/threonine kinase, plays an important role in maintaining correct ploidy through mitosis [18] and it drives the occurrence and development of bladder cancer by mediating the STAT3 signaling pathway [19]. In addition, the abnormal expression of BUB1 is also associated with the prognosis of various tumors, such as sarcoma [20] and breast cancer [21]. SOCS2, a kind of adaptor protein that acts as the substrate recognition subunit of Cullin/Ring ubiquitin ligases [22], is closely related to somatic cell growth by regulating the GH/IGF-1 signal pathway [23]. Previous research has shown that overexpression of SOCS2 inhibits the growth of ovarian cancer cells and breast cancer cells [24], as well as the proliferation and migration of hepatocellular carcinoma cells [25]. Therefore, it is reasonable to consider SOCS2 as a tumor suppressor factor. NDC80 (also called Hec1) is a core component of the kinetochore and plays a critical role in chromosome segregation during mitosis[26]. It has been shown to promote the proliferation and migration of liver cancer [27], gastric cancer [28], and pancreatic cancer cells [29], but its exact mechanism of action in tumorigenesis is not clear.

Then, a risk score model based on these three genes was constructed, and the risk score was calculated by using the risk score formula. Patients were divided into the HR group and LR group according to the median risk score value. The results showed that patients with higher risk scores had poorer overall survival (OS). A high risk score was closely related to clinical pathological parameters representing poor prognoses, such as high AFP, vascular infiltration, low tumor differentiation, and advanced tumor stage. Univariate and multivariate analyses showed that this score was an independent risk factor affecting the prognosis of HCC patients. Furthermore, to verify its universal applicability, patients were stratified into several subgroups based on clinical and pathological features such as age, gender, race, AFP level, tumor T stage, TNM stage, Edmondson grade, and vascular invasion. Internal validation results showed that the model had good predictive ability in several subgroups of patients with different clinical characteristics. Moreover, in the TCGA database, this score demonstrated excellent predictive efficacy for patients’ OS, PFS, and DSS. External validation using the ICGC database confirmed the robustness of the score’s predictive ability. In summary, this risk score exhibits strong predictive power in the prognosis of HCC patients in both the training set and testing set.

It has been reported that the occurrence and development of tumors are associated with the accumulation of genetic mutations. In our study, we analyzed the differences in the genetic mutation profiles between the HR group and LR group and found that missense mutations were the most commonly observed form of mutation in both groups. The HR group had a higher proportion of gene mutations, such as TP53, TTN, CTNNB1, MUC16, and CSMD3. In particular, the proportion of TP53 mutations in the HR group was significantly higher than that in the LR group. Notably, previous studies have reported that TP53 mutations are closely associated with higher Edmondson grade, microvascular invasion [30], and a poorer prognosis [31]. Then, the GSEA results showed that the oxidative phosphorylation and adipogenesis signaling pathways were significantly activated in the HR group. Changes in mitochondrial oxidative phosphorylation function contribute to the proliferation, metastasis, and invasion of HCC [32]. Continuous de novo lipogenesis provides membrane building blocks, lipid signaling molecules, protein translation post-translational modifications, and energy supply to support the rapid proliferation of cancer cells [33]. In addition, the HR group also showed higher TMB levels and stemness scores, both of which are associated with a poorer prognosis. Consistent with our study, patients with high TMB in liver cancer have a significantly poorer overall prognosis [34] and high stemness scores are associated with malignant characteristics in many cancers [35]. In summary, these results explained the underlying reasons for the poorer prognosis represented by the HR group from the perspective of genetic alterations.

The tumor microenvironment is a unique environment that appears due to the interaction between the tumor and host during the progression of the disease, consisting of proliferating tumor cells, tumor stroma, blood vessels, infiltrating inflammatory cells, and various related tissue cells [36]. In our study, we found that the infiltration of common lymphoid progenitors, memory CD4 + T-cell, and CD4 + T-cell (Th2) was higher in the HR group, and the enrichment of these cells often led to a poor prognosis. Similarly, Yang et al.'s study showed that the infiltration of common lymphoid progenitor and CD4 + T-cell (Th2) cells is negatively correlated with the survival rate of pancreatic cancer patients, which also confirms our point [37]. In addition, our findings also revealed that the HR group had lower stroma scores and microenvironment scores compared to the LR group, which were strongly associated with a poorer prognosis. These results are consistent with the findings from Huo et al.'s study [38]. The immune microenvironment of HCC patients with a higher risk score is poorer, which may be one of the reasons for their poorer prognosis.

In order to better guide the selection of clinical treatment methods, we evaluated the sensitivity of patients in different risk groups to immunotherapy, targeted therapy, and chemotherapy. TIDE score is a novel biomarker for predicting response to immunotherapy. We found that the HR group had a lower TIDE score, which indicates a lower likelihood of tumor immune evasion [39]. Moreover, the HR group had significantly higher expression levels of immune checkpoint genes and higher TMB levels [40]. This suggests that patients in the HR group may have a better response to immunotherapy. Additionally, we also discussed the potential sensitivity of different risk groups to various drugs, and the results showed that the HR group had a higher sensitivity to multiple chemotherapeutic and targeted drugs. Therefore, although a higher risk score represents a poor prognosis, it also indicates a higher sensitivity to chemotherapy, targeted therapy, and immunotherapy. However, the ultimate efficacy of these drugs still needs to be verified in clinical trials.

To determine the differences in the mRNA expression of BUB1, NDC80, and SOCS2 genes in tumor cell lines and normal liver cell lines, we detected the mRNA levels of these genes using real-time PCR. Consistent with the results obtained in tumor tissues, we found that BUB1 and NDC80 were significantly more highly expressed in tumor cell lines, while SOCS2 was significantly more highly expressed in tumor cell lines compared with normal liver cell lines.

To sum up, we have developed an excellent stability and universality risk score model for HCC based on the analysis of high-throughput sequencing data, which consists of only three genes. This prognostic model is not only beneficial for clinical applications but also helps to reduce the sequencing cost. However, there are limitations to our study. On the one hand, this study is a retrospective analysis based on public data, and further validation in a large multi-center prospective study is needed. On the other hand, the selected genes have important prognostic value, and our study has not fully revealed their underlying mechanisms, which requires more in vitro and in vivo experiments for further verification.

We have developed the first risk score model for HCC based on ISGs, which can predict the OS, PFS, DSS, and DFS of HCC patients. A high risk score was correlated with clinical and pathological features, such as tumor gene mutations, tumor microenvironment infiltration, immune checkpoint expression, and drug sensitivity. Our study provides new insights into the molecular mechanisms of HCC and offers new targets and options for the clinical treatment of HCC.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The datasets analysed for this study were obtained from The Cancer Genome Atlas (TCGA)(https://portal.gdc.cancer.gov/), International Cancer Genome Consortium (https://dcc.icgc.org/) and the Genomicsof Drug Sensitivity in Cancer (GDSC) (https://www.cancerrxgene.org/).

Competing interests

The authors have no conficts of interest to declare

Funding statement

This work was supported by the National Natural Science Foundation of China (82272819,81972888); the Research Project of Jinan Microecological Biomedicine Shandong Laboratory (JNL202219B, JNL202204A); the Primary Research & Development Plan of Jiangsu Province (BE2022840); and the Open Project of Chinese Materia Medica First-Class Discipline of Nanjing University of Chinese Medicine (2020YLXK007).

Authors’ contributions

Shuo Wang and Chunping Jiang designed this study, Shuo Wang and Zhonghan Zhou analyzed the data in this study and interpreted the findings and drafted the manuscript. Weimin Du and Zhongxia Wang carried out data management and revised the manuscript. All authors reviewed the final version of the manuscript.

Acknowledgements

None

Sung H, Ferlay J, Siegel RL: Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. 2021, 71(3):209-249.
Chen Z, Xie H, Hu M, Huang T, Hu Y, Sang N, Zhao Y: Recent progress in treatment of hepatocellular carcinoma. American journal of cancer research 2020, 10(9):2993-3036.
Schoggins JW: Interferon-Stimulated Genes: What Do They All Do? Annual review of virology 2019, 6(1):567-584.
Au-Yeung N, Horvath CM: Transcriptional and chromatin regulation in interferon and innate antiviral gene expression. Cytokine & growth factor reviews 2018, 44:11-17.
Post AEM, Smid M, Nagelkerke A, Martens JWM, Bussink J, Sweep F, Span PN: Interferon-Stimulated Genes Are Involved in Cross-resistance to Radiotherapy in Tamoxifen-Resistant Breast Cancer. 2018, 24(14):3397-3408.
Benci JL, Johnson LR, Choa R, Xu Y, Qiu J, Zhou Z, Xu B, Ye D, Nathanson KL, June CH et al: Opposing Functions of Interferon Coordinate Adaptive and Innate Immune Responses to Cancer Immune Checkpoint Blockade. Cell 2019, 178(4):933-948.e914.
Sprooten J, Vankerckhoven A, Vanmeerbeek I: Peripherally-driven myeloid NFkB and IFN/ISG responses predict malignancy risk, survival, and immunotherapy regime in ovarian cancer. 2021, 9(11).
You JS, Jones PA: Cancer genetics and epigenetics: two sides of the same coin? Cancer cell 2012, 22(1):9-20.
Podlaha O, Riester M, De S, Michor F: Evolution of the cancer genome. Trends in genetics : TIG 2012, 28(4):155-163.
Dhanasekaran R, Nault JC, Roberts LR, Zucman-Rossi J: Genomic Medicine and Implications for Hepatocellular Carcinoma Prevention and Therapy. Gastroenterology 2019, 156(2):492-509.
Fu XW, Song CQ: Identification and Validation of Pyroptosis-Related Gene Signature to Predict Prognosis and Reveal Immune Infiltration in Hepatocellular Carcinoma. Frontiers in cell and developmental biology 2021, 9:748039.
Dai Y, Qiang W, Lin K, Gui Y, Lan X, Wang D: An immune-related gene signature for predicting survival and immunotherapy efficacy in hepatocellular carcinoma. 2021, 70(4):967-979.
Liang JY, Wang DS, Lin HC, Chen XX, Yang H, Zheng Y, Li YH: A Novel Ferroptosis-related Gene Signature for Overall Survival Prediction in Patients with Hepatocellular Carcinoma. International journal of biological sciences 2020, 16(13):2430-2441.
Malta TM, Sokolov A, Gentles AJ, Burzykowski T, Poisson L, Weinstein JN, Kamińska B, Huelsken J, Omberg L, Gevaert O et al: Machine Learning Identifies Stemness Features Associated with Oncogenic Dedifferentiation. Cell 2018, 173(2):338-354.e315.
Huitzil-Melendez FD, Capanu M, O'Reilly EM, Duffy A, Gansukh B, Saltz LL, Abou-Alfa GK: Advanced hepatocellular carcinoma: which staging systems best predict prognosis? Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2010, 28(17):2889-2895.
Goedegebuure RSA, Kleibeuker EA, Buffa FM, Castricum KCM, Haider S, Schulkens IA, Ten Kroode L, van den Berg J, Jacobs M, van Berkel AM et al: Interferon- and STING-independent induction of type I interferon stimulated genes during fractionated irradiation. Journal of experimental & clinical cancer research : CR 2021, 40(1):161.
Grady WM: Epigenetic alterations in the gastrointestinal tract: Current and emerging use for biomarkers of cancer. Advances in cancer research 2021, 151:425-468.
Bernard P, Hardwick K, Javerzat JP: Fission yeast bub1 is a mitotic centromere protein essential for the spindle checkpoint and the preservation of correct ploidy through mitosis. The Journal of cell biology 1998, 143(7):1775-1787.
Jiang N, Liao Y, Wang M, Wang Y, Wang K, Guo J, Wu P, Zhong B, Guo T, Wu C: BUB1 drives the occurrence and development of bladder cancer by mediating the STAT3 signaling pathway. Journal of experimental & clinical cancer research : CR 2021, 40(1):378.
Long Z, Wu T, Tian Q, Carlson LA, Wang W, Wu G: Expression and prognosis analyses of BUB1, BUB1B and BUB3 in human sarcoma. Aging 2021, 13(9):12395-12409.
Ocaña A, Pérez-Peña J, Díez-González L, Sánchez-Corrales V, Templeton A, Seruga B, Amir E, Pandiella A: Transcriptomic analyses identify association between mitotic kinases, PDZ-binding kinase and BUB1, and clinical outcome in breast cancer. Breast cancer research and treatment 2016, 156(1):1-8.
Kung WW, Ramachandran S, Makukhin N, Bruno E, Ciulli A: Structural insights into substrate recognition by the SOCS2 E3 ubiquitin ligase. 2019, 10(1):2534.
Greenhalgh CJ, Rico-Bautista E, Lorentzon M, Thaus AL, Morgan PO, Willson TA, Zervoudakis P, Metcalf D, Street I, Nicola NA et al: SOCS2 negatively regulates growth hormone action in vitro and in vivo. The Journal of clinical investigation 2005, 115(2):397-406.
Sutherland KD, Lindeman GJ, Choong DY, Wittlin S, Brentzell L, Phillips W, Campbell IG, Visvader JE: Differential hypermethylation of SOCS genes in ovarian and breast carcinomas. Oncogene 2004, 23(46):7726-7733.
Liu J, Liu Z, Li W, Zhang S: SOCS2 is a potential prognostic marker that suppresses the viability of hepatocellular carcinoma cells. Oncology letters 2021, 21(5):399.
Wimbish RT, DeLuca JG: Hec1/Ndc80 Tail Domain Function at the Kinetochore-Microtubule Interface. Frontiers in cell and developmental biology 2020, 8:43.
Ju LL, Chen L, Li JH, Wang YF, Lu RJ, Bian ZL, Shao JG: Effect of NDC80 in human hepatocellular carcinoma. World journal of gastroenterology 2017, 23(20):3675-3683.
Qu Y, Li J, Cai Q, Liu B: Hec1/Ndc80 is overexpressed in human gastric cancer and regulates cell growth. Journal of gastroenterology 2014, 49(3):408-418.
Meng QC, Wang HC, Song ZL, Shan ZZ, Yuan Z, Zheng Q, Huang XY: Overexpression of NDC80 is correlated with prognosis of pancreatic cancer and regulates cell proliferation. American journal of cancer research 2015, 5(5):1730-1740.
Wang S, Shi H, Liu T, Li M, Zhou S, Qiu X, Wang Z, Hu W, Guo W, Chen X et al: Mutation profile and its correlation with clinicopathology in Chinese hepatocellular carcinoma patients. Hepatobiliary surgery and nutrition 2021, 10(2):172-179.
Liu J, Ma Q, Zhang M, Wang X, Zhang D, Li W, Wang F, Wu E: Alterations of TP53 are associated with a poor outcome for patients with hepatocellular carcinoma: evidence from a systematic review and meta-analysis. European journal of cancer (Oxford, England : 1990) 2012, 48(15):2328-2338.
Zhang Y, Li W, Bian Y, Li Y, Cong L: Multifaceted roles of aerobic glycolysis and oxidative phosphorylation in hepatocellular carcinoma. PeerJ 2023, 11:e14797.
Zhang F, Du G: Dysregulated lipid metabolism in cancer. World journal of biological chemistry 2012, 3(8):167-174.
Shrestha R, Prithviraj P, Anaka M, Bridle KR, Crawford DHG, Dhungel B, Steel JC, Jayachandran A: Monitoring Immune Checkpoint Regulators as Predictive Biomarkers in Hepatocellular Carcinoma. Frontiers in oncology 2018, 8:269.
Bai KH, He SY, Shu LL, Wang WD, Lin SY, Zhang QY, Li L, Cheng L, Dai YJ: Identification of cancer stem cell characteristics in liver hepatocellular carcinoma by WGCNA analysis of transcriptome stemness index. Cancer medicine 2020, 9(12):4290-4298.
Whiteside TL: The tumor microenvironment and its role in promoting tumor growth. Oncogene 2008, 27(45):5904-5912.
Yang B, Zhou M, Wu Y, Ma Y, Tan Q, Yuan W, Ma J: The Impact of Immune Microenvironment on the Prognosis of Pancreatic Ductal Adenocarcinoma Based on Multi-Omics Analysis. Frontiers in immunology 2021, 12:769047.
Huo J, Wu L, Zang Y: A Prognostic Model of 15 Immune-Related Gene Pairs Associated With Tumor Mutation Burden for Hepatocellular Carcinoma. Frontiers in molecular biosciences 2020, 7:581354.
Jiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, Li Z, Traugh N, Bu X, Li B et al: Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nature medicine 2018, 24(10):1550-1558.
Samstein RM, Lee CH, Shoushtari AN: Tumor mutational load predicts survival after immunotherapy across multiple cancer types. 2019, 51(2):202-206.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification of a Interferon-Stimulated gene Signature for Predicting Prognosis, Tumor Microenvironment, and Drug Candidates in Hepatocellular Carcinoma

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1 Introduction

2 Method

2.1 Data Collection

2.2 Construction and validation of the ISGs prognostic model

2.3 Functional analysis based on model grouping

2.4 Cell culture

2.5 RNA extraction and quantitative real-time PCR

2.5 Statistical analysis

3 Results

3.1 Construction of the Prognostic Model based on three ISGs

3.2 The model predicts the Overall survival of HCC patients

3.3 The model significantly correlated with the clinical and pathological features

3.4 The independent prognostic value of the model for PFS, DFS, and DSS

3.5 The model predicts gene mutations

3.6 The model predicts immune microenvironment

3.7 The model predicts the response to immunotherapy

3.8 The expression of BUB1, NDC80, and SOCS2 in normal and tumor cells.

4 Discussion

5 Conclusion

Declarations

References

Additional Declarations

Status:

Version 1