Glioma is the most common malignancy in the brain, representing more than 70% of all central nervous system (CNS) malignancies (1). Glioblastoma multiforme (GBM), the most aggressive and malignant form of glioma, has a median survival of fewer than 21 months (2) despite the progress of neurosurgical resection, chemotherapy, radiation therapy, and novel approaches such as immunotherapy. Intratumoral heterogeneity widely exists in GBM (3) and has become an obstacle for molecular targeted therapy (4). To counteract the heterogeneity, therapies targeting the cytosolic innate immune receptors retinoic-acid inducible gene I (RIG-I) have been employed to gain a good response (5).

Acting as the upstream of RIG-I in the innate immune response, DDX60 is a novel DEAD-box RNA helicase and first identified through microarray research of genes induced by measles’ virus infection in dendritic cells (DCs) (6). Through the helicase domain and ATP-binding site, DDX60 can detect abnormal intracellular nucleic acids and then induce RIG-I-dependent type I interferons (type I IFNs) and other inflammatory cytokines (6–8). Besides, DDX60 induced RIG-I-independent antiviral responses have also been demonstrated (8). Involved in RIG-I-dependent and independent innate immune responses, DDX60 has been proven to be associated with the development of tumors (9–11). It was upregulated in oral squamous cell carcinoma and correlated with poor disease-free survival (10), while downregulated in colorectal cancer and related with the initiation and progression of the disease (11). Therefore, DDX60 represents a potential target for tumor therapy. Immunotherapy and particularly immune checkpoint inhibitors, such as programmed death-ligand 1 (PD-L1) inhibitors, have revolutionized the treatment landscape of glioma (12). However, because of the heterogeneity and immunosuppression of glioma, some of the checkpoint inhibitor therapies fail to get a positive effect (1, 13), and new biomarkers for immune therapies are urgently needed. Herein, we assume that DDX60 is a novel immune therapeutic target for glioma and explore its prognostic value and biological function in glioma.

This study demonstrated that DDX60 is highly expressed in GBM and predicts poor prognosis of glioma by the Cancer Genome Atlas (TCGA), Chinese Glioma Genome Atlas (CGGA), Repository for Molecular Brain Neoplasia Data (REMBRANDT), and Gravendeel databases. Then, the correlation between DDX60 expression and inflammatory responses, immune-related molecules, infiltrating immune cells as well as checkpoint protein in glioma was also established.

Although novel therapies such as immunotherapy have been used, less progress has been made in overall survival (OS) in GBM patients (26). Within-tumor heterogeneity is a major driver of progression, recurrence, and therapeutic resistance of GBM (27). Therefore, more biomarkers are needed to be identified to accurately access the prognosis of GBM patients and individualize treatment strategies.

DDX60 is a DEAD-box RNA helicase and has been proved upregulated in melanoma (28) and oral squamous cell carcinoma (OSCC) (10) while deregulating in colorectal cancer (11). Herein, we demonstrated that higher expression of DDX60 was correlated with high-grade glioma. Moreover, we observed that DDX60 was significantly overexpressed in glioma with IDH wild-type, ATRX wild-type, MGMT unmethylated as well as TERT promoter mutated. These findings suggest that gliomas with high DDX60 expression have increased malignancy and insensitive to chemotherapy (29).

A previous study has shown that high expression of DDX60 was significantly associated with poor survival in lip squamous cell carcinoma (10). As expected, this study demonstrated that high DDX60 expression was associated with poor prognosis both in LGG and GBM. As the nomogram could show better performance than conventional staging systems and more precise prognostic forecast in some tumors (30, 31), we identified DDX60 as a prognostic marker of glioma and built a nomogram with a risk classification system. The four parameters included in the nomogram complied with clinical relevance and Cox analysis (32). Studies have shown that gender, age, WHO grade, and IDH status are related to the prognosis of glioma (33–35). However, our univariate and multivariate Cox analyses did not find a correlation between sex and prognosis of glioma. Thus, gender was excluded from the visualization of the nomogram. The calibration plots of the two external validation cohorts were highly fitted, illustrating that the nomogram performed well in predicting 3- or 5-year survival for glioma patients.

GO and GSEA of DDX60 in this research showed that immune and inflammatory responses were the most enriched terms. It has been indicated that inflammation regulates various stages of the tumor process, such as promotion and invasion (36). Different proinflammatory mediators induced by inflammation promote tumor progression by regulating chemokines, cascades of cytokines, adhesion, and pro-angiogenic activities (37). DDX60 is an IFN-inducible gene, and its ectopic expression can promote RIG-I RNA-binding activity, causing RIG-I-mediated type I IFN expression (6). Type I IFNs (IFN-alpha and IFN-beta) are a family of cytokines with a diverse cellular processes such as regulation of inflammatory and immune responses (38). Through mediating type I IFNs, DDX60 can also activate STAT1and upregulate MHC-I (39); these results are compatible with our findings. Furthermore, we also found that DDX60 was positively associated with LCK and HCK while negatively associated with IgG response; these results refined the mechanism of DDX60 in the inflammatory response of glioma.

Multiple non-neoplastic cells exist in the GBM microenvironment, such as infiltrating immune cells (40). The immune surveillance of these immune cells would be converted to detrimental function when the immune system is overwhelmed by cancer burden during tumor development (40). The high correlation between infiltrating immune cells and DDX60 expression has been demonstrated in this study, such as type 17 T helper cell (Th17) and macrophages. Previous research had verified that Th17 cells in the GBM microenvironment may participate in immune suppression via TGF-β1-induced IL-10 secretion (41). Tumor-associated macrophages (TAMs) in GBM have been proven to be the dominant infiltrating immune cell population and engaged in interactions with tumor cells to aid tumor infiltration and proliferation (42). These results suggest that DDX60 might involve in immunosuppression by mediating immune cells in glioma.

CNS used to be considered as “immunologically privileged” in the past decades. However, as more and more researchers have verified that the leukocyte lymphatics not only be present in CNS but also have the ability to transport antigens to cervical lymph nodes (43–46), researchers agree that CNS is more likely “immunologically unique” rather than “immunologically privileged”. These provide a basis for glioma immunotherapy. Undoubtedly, immunotherapy holds a bright future for the treatment of glioma. However, it seems difficult to achieve stable and better outcomes for immunotherapy in clinical. This can be mostly attributed to the tumor heterogeneity of glioma (47, 48). New biomarkers which can predict and monitor immunotherapy response have become urgently needed. In this study, DDX60 was not only a prognosis prediction for glioma patients but also an indicator of the immune microenvironment of glioma and might become a novel biomarker and potential therapeutic target.

Checkpoint inhibitors, advancing rapidly in recent years, have been the immunotherapy most advanced in clinical use. Among them, death protein 1 (PD-1) and PD-L1 are the most broadly studied (49). PD-L1 is widely expressed on the GBM infiltrating T cells and is a negative prognosticator for GBM outcome (50). PD-1 combining with PD-L1, negatively modulates T cell receptor-induced signaling transduction, blocks the activation of cytotoxic T cell, and inhibits the producing of inflammatory factors, causing T cell inability (49). Many clinical trials in GBM are evaluating anti-PD-L1 agents such as Durvalumab, Atezolizumab, Avelumab, alone or combined with other therapies (51). However, not all the research studies on anti-PD-L1 agents come out with a meaningful benefit (51). Thus, new biomarkers that can accurately predict the efficacy of PD-L1 inhibitor therapy are urgently needed. In this study, we demonstrated that the correlation coefficients (R) between PD-L1 and DDX60 were 0.54 in the TCGA glioma dataset and 0.61 in the CGGA glioma dataset. We further performed western blot both in patient tissues and glioma cell lines to verify the strong correlation between PD-L1 and DDX60 (P < 0.0001, R = 0.86). The mechanism of the positive correlation between DDX60 and PD-L1 might lie in the IFN/PD-L1 axis. As an IFN-inducible gene, the ectopic expression of DDX60 can improve RIG-I RNA-binding activity, causing RIG-I-mediated IFN expression (6). A previous study has shown that IFN was a crucial factor of PD-L1 expression in the glioma model (52). Thus, the upregulation of DDX60 might lead to a higher expression level of PD-L1. These results implied that glioma patients with higher DDX60 expression might benefit more from PD-L1 blocker therapy.

Besides, the correlation between DDX60 and some other immune checkpoint genes such as cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4), epidermal growth factor (EGF), CD226, and CD96 was shown ( Figure 7 , Table 2 ). As the correlation coefficients (R) between EGF and DDX60 were >0.5 in both CGGA glioma and CGGA GBM database, western blot was then employed to demonstrate the strong association between EGF and DDX60 (P = 0.002, R = 0.56). Thus, these results illustrate the predictive significance and potential synergistic responses of DDX60 to immune checkpoint treatments.

There have been reports that DDX60 was overexpressed in other types of cancers (10, 11), but most of them did not further explore the intrinsic mechanisms. Thus, the novelties of this paper lie not merely in the findings of prognostically significant of DDX60 in glioma, but also in the mechanism of DDX60 on glioma. Our research verified the strong association between DDX60 and glioma immune microenvironment, clarified the mechanism of DDX60, and proposed that DDX60 might become a novel biomarker for immunotherapy.

In conclusion, these results would widen our knowledge of the expression and prognostic value of DDX60 in gliomas. Furthermore, as a potential therapeutic target, DDX60 is positively correlated with PD-L1 and other checkpoints. Thus, these findings will help to optimize the immunotherapy in glioma.

Publicly available datasets were analyzed in this study. This data can be found here: http://gliovis.bioinfo.cnio.es/.

The studies involving human participants were reviewed and approved by the Human Ethics Committee of Huashan Hospital. The patients/participants provided their written informed consent to participate in this study.

JingZ designed and conducted the study and drafted the original manuscript. MF, JinZ, and MZ helped collect the databases. ZD helped conduct the IHC and IRS. WH, HZ, and YM supervised the study. WH revised the manuscript. All authors contributed to the article and approved the submitted version.

The study was funded by Shanghai Science and Technology Commission (17430750200), the National Natural Science Foundation of China (82072785, 82072784), Join Breakthrough Project for New Frontier Technologies of Shanghai Hospital Development Center (SHDC12016120).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.