Immune microenvironment plays an important role in the tumor progression, immune escape and immunotherapy resistance[10]. Hepatocellular carcinoma (HCC) is a typical inflammatory related malignancy, in which contains a large number of macrophages, innate and adaptive immune cells, thus forming a complex immune tolerance microenvironment[25]. In recent years, with the rapid development of bioinformatics on tumor immunity microenvironment, immunotherapy has gained a lot of interest from researchers due to its great potential for effectively treating HCC. Nevertheless, there remain many unanswered questions, such as low objective response rate, high adverse reaction, and high resistance rates, which obstruct the generalization of immunotherapy. Thus, deeper understanding of the role of the tumor microenvironment will improve the response rates of the current treatment approaches, and provide great clinical significance for precise treatment of HCC patients.
In this research, according to gene expression analysis of TCGA database, the ESTIMATE algorithm was applied to calculated the immune/stromal/Estimate scores of TME in HCC. Then, we explored the impact of the immune/stromal/Estimate scores on OS, PFS, and DFS, and further analyzed the correlation of immune/stromal/Estimate with clinicopathologic parameters (i.e. gender, age, tumor grade, and clinical stages). The results indicated the higher score of stromal/immune/Estimate were significantly associated with longer survival, including OS, PFS, and DFS in patients with HCC. Moreover, immune/stromal scores were inversely correlated with tumor grade and clinical stage, which manifested that stromal/immune/Estimate scores might predicate the survival of HCC patients and the malignancy of tumors. Given that lncRNAs could influence tumor microenvironment and finally affect tumor behaviors, we opted to focus on searching DElncRNAs basing on high- or low- stromal and immune scores, which were identified as TME-related lncRNAs, and then exploring the prognosis values of those DElncRNAs. Eventually, we identified six EMT-related lncRNAs (LINC01150, LINC02273, LINC00426, AP002954.1, AC007277.1, and AC008549.1) to construct the prognostic signature via Cox regression analysis. Then the risk scores were calculated based on the expression profiles and coefficients of the six lncRNAs. Kaplan–Meier analysis, ROC curve analysis, univariate and multivariate Cox regression analysis were carried out to explored the prognostic values of the risk signature, the results indicated that the survival in high-risk group were worse than those of patients in low-risk group, and the risk score model could correctly predict the OS as independent indicator of HCC patients compared with other clinical parameters. In addition, in order to better apply the risk model in clinical practice, we established a nomogram based on risk score and other clinical feathers (i.e. grade and clinical stage) and verified the nomogram with calibration curve and ROC curve, the identification results showed that the nomogram had a satisfactory uniformity with actual survival and provided a better clinical practicality than the traditional tumor grade or clinical stage system.
Some lncRANs in this risk model have been illuminated to be involved in the progression of HCC or other cancers, including LINC00426 and LINC02273. Previous study have demonstrated that the expression of LINC00426 was significantly down-regulated in the tumor tissue of including 72 NSCLC patients compared to normal lung tissue and the expression level was also remarkably correlated with the clinical stage[26].In contrast, another study showed an opposite result, the LINC00426 was significantly up-regulated in lung adenocarcinoma(LUAD) tissues and cell lines and play a notable role in accelerating tumor proliferation, invasion, metastasis, and epithelial–mesenchymal transition (EMT) in vitro and in vivo, which via regulating the miR-455-5p/ UBE2V1[27]. Besides, it has been reported that LINC00426 is also a key regulator in doxorubicin resistance of osteosarcoma [28]. Significantly, LINC00426 might reshape the tumor immune microenvironment, which positively associated with T-helper cell differentiation, cytokine signaling pathways, and multiple immune markers, including cytotoxic markers, coinhibitory and costimulatory molecules (i.e. PDCD1, CTLA4, HAVCR2, TIGIT, FOXP3, ICOS), and chemokine receptors and ligands (i.e. CXCR3/6, CXCL9/13, CCL4/5/7/19, CCR7)[29]. Another study revealed that LINC00426 significantly correlated with immune cell fraction in clear cell renal cell carcinoma based on bioinformatic analysis[30]. In this study, we found LINC 00426 was significantly upregulated in both high immune scores groups and high stromal scores group compared to low scores groups, which might play a key role in altering tumor microenvironment, this result were consistent with those previously reported. Another prognostic lncRNA LINC002273 was reported correlation with tumor proliferation, migration, and invasion by epigenetically increasing AGR2 transcription in breast cancer[31]. Nevertheless, the other four prognostic TME-related lncRNAs (LINC01150, AP002954.1, AC007277.1, and AC008549.1) were rarely reported. Thus, our research may provide a new perspective for their potential functions.
Subsequently, GSEA analysis was conducted to deeply probe the underlying mechanism and functional enrichment of the risk signature built by the six lncRNAs in HCC. The results indicated that some pathways associated with HCC tumorigenesis and progression might be activated in high-risk group, including ERBB signaling pathway, Notch signaling pathway, DNA replication and cell cycle. Besides, the GO analysis suggested that DEGs, which screened out between high- and low-risk group, were mainly involved in immune features, such as “humoral immune response”, “complement activation”, “adaptive immune response”, and “B cell mediated immunity”. Tumor cells interact with TME through various paracrine signaling pathways, among which Notch signaling pathway is considered to be one of the important pathways. The role of Notch signaling pathway in promoting or inhibiting cancer in tumor cells has been widely recognized[32]. Moreover, previous studies also have demonstrated that Notch signal is involved in regulating the differentiation and function of lymphocytes, DCs, Th cells and Treg cells[33–36]. Notch signaling is vital for the regulating activation of cytotoxic T cell (CTL), that Dll1 binging to Notch1 or Notch2 to express granzyme B and IFN- γ is necessities for naive CD8 + cells activation and differentiation, thereby enhancing antigen-specific cytotoxicity and inhibiting tumor growth[35, 37]. Nevertheless, a study in colon cancer found even if inhibits the Notch signal of CD8 + T cells, T cells can enhance their cytotoxic activity by reducing the expression of PD-1[38]. Therefore, the function of Notch signaling pathway in a specific tumor microenvironment needs to be further clarified.
Given the significant role of the tumor microenvironment in the process of tumorigenesis, we constructed TME-related lncRNAs risk signature. We then analyzed the correlation between the risk signature and specific immune cells or immune pathways with ssGSEA and CIBERSORT algorithm. We discovered that some immune cells, which were beneficial to enhance immune responses towards cancer, were remarkably upregulated in low-risk group, including B cells, CD8 + T cells, mast cells, Follicular helper T cells (Tfh), Th1 cells, and tumor infiltrating lymphocyte (TIL). CIBERSORT analysis indicated certain immune cells infiltration levels were negatively relevant to risk score, such naive B cells, activated memory CD4 + T cells, CD8 + T cells, γδ T cells, which can be understood as the higher the risk score, the lower infiltration level of immune response cells. Conversely, M2 macrophages were significant positive correlated with risk score, which contribute to tumor angiogenesis, metastasis, epithelial-mesenchymal transition (EMT), and immune suppression in HCC[39]. The differences of immune cells in tumor microenvironment between high-risk and low-risk group may partly account for why HCC patients in high-risk group showed worse prognosis than those low-risk group patients. In addition, we found the immune-related pathways ( i.e. immune checkpoint pathway, Cytolytic activity, inflammation-promoting, T cell co-inhibition/stimulation, INF-II response) were significantly activated in low-risk group compared with high-risk group. Among them, immune checkpoint pathway is one of the hotspots in recent years, immune checkpoint inhibitors (ICIs) can restore the immune recognition and attack ability of T cells through blocking the inhibition of checkpoint, so as to enhance the anti-tumor ability. Nevertheless, only a minority of HCC patients can benefit from the immune therapy and less than 30% of patients were observed for objective response to immune checkpoint inhibitors treatment. Thus, exploring more accurate biomarkers to forecast responsiveness of ICBs treatment and further screen the dominant population has become a major challenge in the treatment of HCC. In this study, we found 20 out of the 46 immune checkpoint blockade-related genes were significantly different between high- and low-risk group, and the expression level of PD-L2, and CTLA4 was significantly up-regulated in low-risk group. Furthermore, immunophenoscore (IPS) was compared to explore the correlation between risk score and ICIs immunotherapy response in HCC. Our results showed the HCC patients in the low-risk group exhibited higher IPS and tended to have better response after ICIs treatment. These findings manifested that the prognostic TME-related lncRNAs may possess the ability to predict clinical outcome of ICIs therapy in HCC samples.
In conclusion, we conducted overall analysis the impact of immune and stromal scores of the tumor microenvironment on the prognosis of patients with HCC. 100 DElncRNAs were filtered according to different expression both in high immune and stomal score group compared with low score groups, and eventually six lncRNAs was used to establish a novel TME-related prognostic risk signature basing on Cox regression analysis, which may improve prognostic predictive accuracy and guide individualized immunotherapy for HCC patients.