Primary liver cancer remains one of the most common human cancers around the world. Although the treatment strategies for primary liver cancer are diverse, the prognosis of the patients are dismal, with 5-year survival rate is 18.1% in America [12]. HCC accounts for the largest proportion of primary liver cancer, whose overall survival is significantly difference in the world. On the other hand, it’s difficult to estimate the overall survival of HCC exactly. Previous studies have shown that the ERS is of great significance in the biogenesis and development of HCC[8], but the relationship between ERS and the overall survival of HCC is not clear. Hence the study is urgently needed to explore the relationship between ERS and HCC. Our study aimed to establish a new gene signature according to ERS and to provide assistance for the treatment and diagnosis of HCC. In this study, we established a new 9-genes signature related to ERS. The expression level of EXTL3, CXCL8, MAGEA3, TMX1, TPP1, BRSK2, CREB3L1 were negative while PIK3R1, ATP2A3 were positive with the prognosis of HCC patients. The products of EXTL3 gene are glycosyltransferase which plays important role in biosynthesis of heparan sulfate[13]. Lai et al. found that the interaction of RAG3A and EXTL3 can regulate keratinocyte differentiation via PI3K-AKT signaling pathway[14]. In another study, Zhang et al. found that REG3A/REG3B binding to EXTL3 can promote acinar to ductal metaplasia and stimulate the RAS-RAF-MEK-ERK signaling pathway[15]. CXCL8 is a crucial proinflammatory chemokine and serves as a vital regulator in the tumor microenvironment[16]. It was reported that ERS can promote the expression of CXCL8 in breast carcinomas[17]. In another study, CXCL8 can promote HCC metastasis and progression[18, 19], which is consistent with our study. However, the relationship between ERS, CXCL8 and HCC need to be explored in the future. MAGEA3 is one of the members of cancer-testis antigen. The aberrant expression of MAGEA3 has been identified in various cancers[20]. In recent study, MAGEA3 has been verified as a driver of tumor progression in HCC[21], and some long non-coding RNAs can influence proliferation, chemoresistance and the progression of HCC via MAGEA3[22, 23]. The products of TMX1 is the disulfide isomerase (PDI) which plays a very important role in ERS[24]. Raturi et al. found that TMX1 can regulate the transport of Ca2+ and cancer cell metabolism[25]. Transcription product of TPP1 is a lysosomal protease and can cleave substrate N-terminal tripeptides. It was reported that the high expression level of TPP1 can lead the poor prognosis of HCC[26]. BRSK2 is a serine/threonine kinase, Wang et al. found that ERS can down-regulate the protein level of BRSK2 and reduce the apoptosis induced by ERS[27]. CREB3L1 is a cellular transcription factor in response to ERS[28]. Besides, research on CREB3L1 are rare in HCC, but one study has shown that CREB3L1 is associated with the prognosis of glioma cases[29]. PIK3R1 encodes p85α subunit of PI3K enzymes and PIK3R1 is considered as a tumor-suppressor gene traditionally[30], but another study showed that overexpression of PIK3R1 accelerates hepatocellular carcinoma progression in cell lines[31]. Moreover, PIK3R1 can promote the transportation of XBP1 isoform 2 to nuclear, which can regulate the cellular response to ERS and improve the glucose tolerance of the liver[32]. The proteins encoded by ATP2A3 gene is Ca2+-ATPase3 from the sarco/endoplasmic reticulum (SERCA3) which plays a very important role in maintaining intracellular Ca2+ homeostasis[33]. Some studies showed that the overexpression of ATP2A3 can induce apoptosis, suppress cell cycle progression and trigger ERS[34]. In HCC cases, the expression of ATP2A3 is downregulated and is related to the poor prognosis of HCC patients[35]. To conclude, among the 9 genes included in our gene signature, EXTL3, TMX1, BRSK2, CREB3L1 are related to the overall survival of HCC patients, but the research on them is rare and more effort need to take to investigate the function of those genes in HCC. The roles of residual genes in the process of ERS in HCC need to be identified in the future.
In our study, the Riskscore of HCC patients were calculated and the immune status were explored in different groups with different methods in TCGA cohort. The high-risk group had a lower immune infiltration state, and a higher tumor purity. As our results shown, the high-risk group had a higher infiltration of M2 macrophages, regulatory T cells (Tregs) and a lower infiltration of CD8+ T cells. In results of MCP counter, the infiltration of NK cells is lower in the high-risk group. Besides, most of the immune response are reduced in high-risk group. The major functions of M2 macrophages are anti-inflammatory and immunosuppressive effects[36]. Those cells can lead to immune suppression and promote tumor cell proliferation and metastasis[37]. Tregs are highly immune suppressive cells, they can suppress effective antitumor immunity, which can contribute to the poor prognosis of various types of human cancers[38]. It’s reported that CD8+ T cells can interact with other immune cells, and it is of great importance in anti-tumor process[39]. Clinical studies have illustrated that the high frequency of CD8+ T cells is good for cancer-free survival in patients of different cancers[40–42]. In humans, NK cells play a very import role in tumor immunosurveillance and reduced NK cells function has been proved with the worse outcomes in cancer patients[43]. According to our study, one reason for poor prognosis in high-risk group is the immunosuppressive microenvironment, these subgroup patients might get more benefit from immunotherapy by targeting CD8+ T cells or NK cells.
Besides, we also explored the mutation status in the different groups. The results suggested that the mutation rate of TP53, MUC16 are higher in high-risk group. TP53 gene is a widely studied tumor suppressor gene and the mutation of TP53 is common in human malignancies[44]. The TP53 protein plays a critical role in the cellular response to different stresses and can maintain genomic integrity[45]. MUC16 is also known as CA125, which has been confirmed as a biomarker in ovarian cancer. Besides, it has been proved to overexpress in human malignancies. The overexpression of MUC16 can promote disease progression and metastasis in some cancers, thus it is an ideal target for diagnosis and therapy[46]. In the analysis of somatic copy number variation, we observed that amplified genomic peaks of oncogenic drivers were discovered in high-risk group, including MYC, MET, MTDH and RPS6KB1. Besides, tumor suppressor genes, such as WRN, CDKN2A and BRCA2 were detected a deletion peak. Those genes are critical for the process of the HCC[47], the amplification of oncogenes and deletion of suppressor genes may lead to the poor prognosis in high-risk group. More importantly, the alternations and heterogeneity of genomic have an impact on tumor microenvironment transformation, progression, and treatment resistance[48], thus those mutation can also serve as the therapeutic targets for high-risk group.
In the following analysis, we identified Riskscore and ECOG performance status as independent predictors and constructed a nomogram based on them. ECOG performance status is a widely applied for comprehensive scale of symptoms and mobility, which has been existed for a long time[49]. Our nomogram shows a robust accuracy and efficiency practicality compared to other models[50, 51]. Besides, the model can provide a practical tool for clinic which can assess the prognosis of the HCC patients. In addition, there are also some shortcomings in our study. First, our study was retrospective and the data was obtained from the public database, the performance of the model need to be confirmed by a mass of clinical cases. Moreover, the immune states and somatic mutations related to the gene signature need to be explored by basic experiment. Finally, the construction of nomogram was based on the RNA-seq technology, the high cost may limit the clinical application of the model.