Malignant mesothelioma is one of the most devastating and aggressive solid tumors, with less than a 10% survival rate at 5 years (Huang et al., 2019). Worldwide, there were 30,870 and 26,278 estimated new mesothelioma cases and deaths, respectively, in 2020 (Sung et al., 2021). Although it is not a common tumor type, the majority of mesothelioma patients are diagnosed at an advanced stage or metastatic phase due to the absence of early symptoms (Fels Elliott and Jones, 2020). Accompanied by progress in medical technology, several therapeutic approaches, such as surgery, radiation, chemotherapy and combination treatment extend the survival of mesothelioma patients. However, the current outcome of treatment remains unsatisfactory (Taioli et al., 2015; Fels Elliott and Jones, 2020; Viscardi et al., 2020). Therefore, novel treatment strategies are urgently needed to improve the outcome of patients with mesothelioma.

Innate and adaptive immunity play key roles in the initiation, progression, and prognosis of tumors. In the three phases of tumor immunoediting, the host immune system can eliminate tumor cells or condition tumor immunogenicity to promote disease progression. Immunotherapy based on immune checkpoint (ICP) inhibitors (e.g., PD-1/PD-L1 and CTLA-4) has achieved success in combating some malignant tumors, however, only a minority of patients with mesothelioma benefit from ICP inhibitor therapy (Chatwal and Tanvetyanon, 2018; Mutti et al., 2018; Forde et al., 2019). After targeting ICP proteins, tumor vaccines are considered to be another effective tool for immunotherapy (Chiang et al., 2010; Wang et al., 2018). This has become attractive to oncologists because vaccines can promote anti-tumor-specific T cell responses by inducing active immunity. Recently, a CTGF/MSLN tumor vaccine targeting MSLN-expressing malignant mesothelioma generated anti-tumor effects by decreasing tumor volumes and prolonging survival (Chen et al., 2018a). Also, an induced pluripotent stem cells (iPSC) vaccine suppressed mesothelioma progression by promoting tumor antigen-specific T cell responses (Kooreman et al., 2018). Although the therapeutic effects of this clinical application remain unclear, evidence is sufficient to warrant further studies on the potential of mesothelioma-associated vaccines.

Tumor vaccines can trigger and enhance anti-tumor-specific immune responses by reprogramming the immune system to identify and eliminate malignant tumor cells (Chiang et al., 2010; Wang et al., 2018). They exhibit a broad therapeutic window, limited non-specific action, few adverse effects, and produce a long-term immunological memory response. Thus, tumor vaccine therapies have several advantages (e.g., less vulnerable to develop drug resistance and adverse reactions, and durable therapeutic effects) compared with standard chemotherapy and immunotherapy. Based on the antigen form, tumor vaccines can be classified into tumor cell, dendritic cell (DC), peptide, DNA, and RNA types (Mockey et al., 2007; Bouzid et al., 2020). However, the first four vaccine types have shortcomings limiting their potential for clinical application. For example, tumor cell vaccines have shown promising results in laboratory experiments and preclinical models, but have failed in clinical trials because of poor tumor cell immunogenicity (Emens, 2006; Chiang et al., 2010). Although DC vaccines loaded with tumor antigens can induce cytotoxic T cell and T helper cell responses, the effects are not satisfactory (Emens, 2006). Peptide-based tumor vaccines in clinical testing exhibit significantly enhanced immunogenicity and clinical anti-tumor activity. Unfortunately, malignant mesothelioma is not a candidate for surgical therapy, so it is difficult to obtain tumor genomic data for developing personalized peptide vaccines (Sayour et al., 2018; Viscardi et al., 2020). Moreover, the integration of DNA-encoded antigen sequences into the genome of tumor cells has mutation risks, leading to suboptimal treatment (Myhr, 2017; Pardi et al., 2018). Conversely, mRNA vaccines do not integrate into the genome or interact with DNA, thus they do not result in a mutation risk to the host. The half-life of mRNA in vivo may be modulated through various modifications or delivery systems. Also, the inherent immunogenicity of mRNA can be downregulated to ensure an acceptable safety profile (Pardi et al., 2018). Additionally, various modifications and carrier strategies can enhance mRNA stability, cellular uptake, and expression (McNamara et al., 2015; Sayour et al., 2015). Through in vitro transcription reactions, rapid development and production of mRNA vaccines may be achieved, which reduces the time needed to reach the clinic. It is easy to encode any pathological antigen by designing and modifying the mRNA sequence, which is important for constructing individualized mRNA vaccines (Luo et al., 2020). Therefore, in the context of clinical applications, mRNA vaccines may be the best choice for targeting tumor-specific antigens (TSAs). Preclinical models and clinical trials have indicated that mRNA vaccines encoding TSAs exhibit potent anti-tumor activity in multiple tumors, such as melanoma, gastrointestinal cancer, colorectal cancer, pancreatic adenocarcinoma, and hepatocellular carcinoma (Pardi et al., 2018; Pardi et al., 2020; Xu et al., 2020).

Although mRNA tumor vaccines have more advantages compared with other conventional vaccine platforms, it remains challenging to develop specific potent antigens from numerous mutated candidates. For patients with mesothelioma, no effective mRNA vaccine has been reported and the subpopulation that may benefit from an mRNA vaccination remains undefined. Tumor-infiltrating immune cells and immune-related genes (IRGs) play important roles in regulating anti-tumor immune responses and determining patient survival (Oehl et al., 2018; Fusco et al., 2020). Therefore, IRG-based immuno-subtyping can be exploited to stratify mesothelioma patients and select suitable candidates for mRNA vaccination.

In this study, five putative tumor antigens were identified for mesothelioma mRNA vaccine development. These tumor-associated antigens were associated with prognosis, genomic alterations, and antigen-presenting cell (APC) infiltration. Based on the expression of IRGs, we constructed a consistent cluster for identifying suitable mRNA vaccine candidates for mesothelioma patients. Two powerful immune subtypes (IS1 and IS2) were defined and validated, which exhibited different molecular, cellular and clinical characteristics. In addition, the immune landscape of individual patients was determined by graph learning-based dimensional reduction. Our findings provide valuable insight for the development and administration of mRNA vaccines for mesothelioma patients.

Malignant mesothelioma is a rare and aggressive solid tumor with no known cure (Li et al., 2019). Current immunotherapies have completely changed oncology from a therapeutic perspective, while its effectiveness in mesothelioma remains unclear (Ribas and Wolchok, 2018). Vectors are one of the flexible means of vaccine delivery. Although viral vector-based cancer vaccines have the advantage that the entire tumor antigen gene or a portion of the gene can be inserted and several genes can be inserted into the vector, there are several challenges to wide clinical application of these vaccines due to the complex manufacturing and the immune responses against components of the viral vector backbone (Larocca and Schlom, 2011). Notably, comparing with viral vector-based cancer vaccines, the delivery vehicles of mRNA vaccines transfect immune cells without inducing toxicity or unwanted immunogenicity, such as lipid-based nanoparticles, and polyplexes and polymeric nanoparticles (Sahin and Türeci, 2018). Compared to mRNA vaccines, peptide vaccines have several disadvantages that limit their clinical potential. For example, high variability in physicochemical properties of individual peptides complicates fabrication. The second challenge is unrelated immune responses against artificial epitopes generated by peptide degradation in the extracellular space (Larocca and Schlom, 2011; Sahin and Türeci, 2018). The mRNA-based vaccine is well-suited for targeting tumor-specific antigens (Huang et al., 2021a; Huang et al., 2021b). It has achieved significant clinical effects and may represent a new avenue for cancer immunotherapy, however, the effects of mRNA vaccines on patients with mesothelioma have not been determined.

In this study, we identified five tumor antigens (AUNIP, FANCI, LASP1, PSMD8, and XPO5) that were considered promising candidates as mRNA-based vaccines. Their overexpressions were associated with adverse prognosis and RFS together with elevated APCs and B cell infiltration, which may play important functional roles for the occurrence and development of mesothelioma. Thus, targeting these antigens may promote APC and B cell infiltration and induce a host immune response. As expected, previous reports have also indicated that they have the potential of becoming targets for the development of mRNA vaccines. AUNIP, also known as AIBp and C1orf135, plays an important role in DNA damage repair and the cell cycle. AUNIP expression is associated with the tumor microenvironment and immune infiltration in oral squamous cell carcinoma, hepatocellular carcinoma, and lung adenocarcinoma (Yang et al., 2019; Ma et al., 2020). Thus, it is considered a candidate diagnostic and prognostic biomarker for these cancers. FANCI is a key protein in DNA repair and ribosome biogenesis and cooperates with IMPDH2 to promote cell proliferation of (lung adenocarcinoma) LUAD (Zheng and Li, 2020). LASP1 is an actin-binding protein and can enhance tumor growth and metastasis in multiple cancers, including nasopharyngeal carcinoma, colorectal cancer, triple-negative breast cancer, and pancreatic ductal adenocarcinoma (Hu et al., 2017; Butt and Raman, 2018). PSMD8 was found to be overexpressed or mutated in breast cancer and diffuse large B cell lymphoma (Yuan et al., 2021). XPO5 is responsible for the nuclear export of pre-miRNA into the cytoplasm and is correlated with tumorigenesis and prognosis in various cancers (Patrão et al., 2018; Clancy et al., 2019).

Because the benefits of survival and therapeutic responses of patients subjected to mRNA vaccine-based cancer immunotherapy remain limited to a small population, stratifying patients to obtain better insight for the optimal use of tumor immunotherapy is an effective strategy. According to the expression profiles of IRGs, patients with mesothelioma may be divided into two immune subtypes. They exhibited significantly different cellular, molecular and clinical characteristics. A previous study identified six immune subtypes by performing an immunogenomics analyses (Thorsson et al., 2018). Patients with C1, C2, and C3 were correlated with excellent prognosis, C4 and C6 were associated with poor prognosis. In the present study, mesothelioma patients were classified into five subtypes (C1, C2, C3, C4, and C6). C2 and C3 were mainly clustered into the IS1 group, and C6 was associated with the IS2 group. These results were consistent with a better survival rate for IS1, and a worse survival for IS2. Moreover, IS1 eigengenes were significantly higher in greenyellow, brown, and magenta modules characterized by an unfavorable prognosis, which also confirmed the prognostic classification (IS1 and IS2). Taken together, these findings demonstrate the reliability of immunotyping and augment the previous classification.

Considering that the tumor immune status is a key factor influencing the effectiveness of mRNA vaccines, the tumor immune microenvironment characteristics of two immune subtypes were evaluated. IS1 exhibited an immune “cold” phenotype characterized by a lack of immune cell infiltration, indicative of a non-inflammatory TME (Huang et al., 2021a; Huang et al., 2021b). In contrast, IS2 displayed TME characteristics that were opposite to those of the immune “cold” phenotype, and were characterized by an enhanced immune cell infiltration and elevated scores for TGF-β response, indicative of an immune “hot” and immunosuppressive phenotype (Huang et al., 2021a; Huang et al., 2021b). These two immune subtypes could represent the distinct potential mechanisms that modulate tumor immune escape, corresponding to different treatment strategies. Recent studies suggest that patients with mesothelioma undergoing immunotherapy (e.g., T cell therapies) or chemotherapy (e.g., gemcitabine) showed enhanced tumor immune cell infiltration (e.g., CD4 + T helper cells and CD8 + T cells) and prolonged survival (de Gooijer et al., 2020; Gray and Mutti, 2020). Therefore, immunotherapy using mRNA vaccines targeting previously recognized tumor antigens (AUNIP, FANCI, LASP1, PSMD8, and XPO5) could induce immune cell infiltration to reinvigorate the ISI patient’s immune system. Furthermore, the molecular signatures of IS1 (e.g., low scores of TGF-β response, low expression of ICPs, and overexpression of ICD genes) indicate that these patients are suitable for mRNA vaccination therapies. Suzuki et al. revealed that blocking TGF-β signal transduction significantly enhanced the anti-tumor response of CD8⁺ T cells, and ultimately inhibited the growth of malignant mesothelioma (Suzuki et al., 2007). Combaz-Lair et al. found that TLR3 could restore the anti-tumor immune response in patients with mesothelioma (Combaz-Lair et al., 2016). EIF2AK2 is an interferon-inducible, double-stranded RNA protein kinase and its activation can suppress the metastatic capabilities of several cancer cell types (Garcia-Ortega et al., 2017). The MET gene encodes a receptor tyrosine kinase, c-MET, that binds to hepatocyte growth factor and dysregulation of MET signaling is implicated in cell migration, proliferation, and angiogenesis (Guo et al., 2020). The hot tumor (IS2) corresponded to a more complex TME. Inflammatory cytokines, such as, interleukin, tumor necrosis factor and TGF-β, play key roles in tumor progression and promotion of early metastasis (Han et al., 2020). Although IS2 exhibited increased immune cell infiltration, the survival of IS2 patients was worse than that of IS1. Thus, the key factor in determining the prognosis may be the immunosuppressive environment or the stimulation environment dominating. In the immune-suppressive environment, tumors exhibited the high expressions of TGF-β and ICPs, as well as poorly immunogenic of cancer cell death. Previous studies demonstrated that the activated TGF-β signaling pathways may suppress immune function by preventing lymphocytes from entering the tumor parenchyma, whereas specific molecular inhibitors that target TGF-β restored anti-tumor immunity by remodeling the tumor immune microenvironment (Travis and Sheppard, 2014; Larson et al., 2020). Thus, patients exhibiting the IS2 subtype may benefit from combination therapy of ICP inhibitor drugs and TGF-β blockers.

The complex tumor immune microenvironment implies that there is heterogeneity within the same immune subtype and between individual patients. Compared with IS1A, patients with IS1B showed a relative increase in tumor-infiltrating T cells. This suggests that combination therapy with mRNA vaccine and ICP inhibitors may alter the tumor microenvironment in these patients to a state that is more conducive to successful treatment by promoting immune cell infiltration and enhancing an anti-tumor T cell response. We also identified that IS2 patients may be unsuitable for RNA vaccines compared with IS1, however, mRNA vaccines could be relatively feasible for IS2C patients among the IS2A, IS2B, and IS2C subtypes. Of note, it is important to integrate the results of the two immune subtypes and the immune landscape for mesothelioma. Furthermore, to shrink the immune components used to develop personalized therapies based on mRNA vaccines, the co-expression modules of immune gene were constructed. The two co-expression modules with hub IRGs (TIE1 and CLEC1A) were negatively correlated with component 2 of the immune landscape, implying that patients with low expression of the two genes could be suitable for mRNA vaccines. This paper has some limitations that further experimental verification are needed to validate these findings, including mRNA vaccine synthesis and purification, and murine models with mesothelioma.

AUNIP, FANCI, LASP1, PSMD8, and XPO5 are putative tumor antigens for mRNA vaccine development. Patients with the IS1 subtype could be more suitable for mRNA vaccination. This study provides a theoretical foundation for developing effective tumor vaccine therapy for mesothelioma patients.