Mesothelioma is an unusual malignancy that originates from the mesothelial cells of the pleural or other regions. About 81% of the tumors originate from the pleura. The prevalence of malignant mesothelioma is increasing, but the mortality remains unchanged. In China, the incidence rate of malignant mesothelioma was only 1.50/10⁶ whereas the fatality rate was 1.22/10⁶ (1). MPM is mainly seen in older men exposed to asbestos. Compared with European and American countries, the onset age of mesothelioma is younger in China. The prevalence and fatality of malignant mesothelioma in China increase rapidly after the age of 35 or 40, reaching a peak at the age of 80 or 85 (1). Malignant pleural mesothelioma (MPM) is difficult to treat and has a dismal prognosis because most patients are at advanced stages when first diagnosed and are with early onset of evident clinical manifestations. Due to its resistance to conventional therapies and the absence of effective alternative regimens, MPM presents a highly difficult challenge. Despite the prompt advancement of immunotherapy and the fairly encouraging outcomes of ICIs in treating MPM, it remains high mortality on a global scale. The 5-year survival rate is around 10%, and the median overall survival is roughly one year.

Based on multiple studies conducted in MPM with immunotherapy alone or combined applied, the median PFS of 4~7 months does not seem particularly impressive. However, the increased median OS is mainly driven by a small portion of patients with long-lasting responses and deserves more explorations (2). CheckMate-743, a phase 3 randomized controlled trial, recently showed that MPM could benefit from PD-1 inhibitors combined with CTLA-4 inhibitors (3). Subgroup analyses revealed that the response rate to ICIs in MPM is somewhat but not entirely related to histology. Coupled with the fact that ICIs are more expensive and not covered by health insurance, it will result in a low cost-benefit ratio if the treatment is not effective. Therefore, there is an urgent need for identifying the subtypes of MPM patients who would potentially benefit from immunotherapy (4).

Immunogenic cell death (ICD) is a form of regulated cell death (RCD), acting as a major initiator of adaptive immune response in the context of malignant neoplasms (5). The promotion of ICD sensitizes MPM to ICIs treatments, as demonstrated by in-vitro experiments, preclinical models, and preliminary trials (6–11). which raises the possibility that ICD-associated biomarkers could serve as prospective predictive indicators for immunotherapy. An increased amount of work has discovered that induction of adaptive immune responses by cancer cells undergoing ICD is dependent on the emission and detection of a particular panel of DAMPs, including cell surface-exposed calreticulin (CALR), high mobility group box 1 (HMGB1) and extracellular adenosine triphosphate (ATP) (12, 13). In addition, previous studies also have demonstrated that ICD-associated DAMPs produced by chemotherapy or radiotherapy activate the cytotoxic CD8+ T cell and alleviate the immunosuppressive tumor microenvironment (TME), thus suggesting an essential role of DAMPs in immunotherapy (10). Pattern recognition receptors (PRRs) bound by DAMPs present adjuvanticity by activating transcription factors, eliciting APC cell activation, differentiation, and maturation, promoting the release of type 1 interferons and chemokines, resulting in the recruitment of APCs and T cells, and ultimately modulating intrinsic and adaptive immunity (14). Whether there is a pre-existing anti-tumor immune response is essential for effective immune checkpoint blockade. Effector T cells release interferon-γ (IFN-γ) by recognizing tumor neoantigens, which activates the Janus kinase (JAK)– signal transducer and activator of transcription (STAT) signaling pathway. The expression of programmed cell death ligand 1 (PD-L1) on the surface of tumor cells is mediated by the subsequent stimulation of the transcription factor interferon regulatory factor 1 (IRF1), which negatively regulates the effector T cell response in turn (15, 16). Immune checkpoint inhibitors (ICIs) disrupt this negative feedback loop as one of the primary mechanisms to restore anti-tumor immunity and exert anti-tumor efficiency. Hence, sensitizing tumors to ICIs by maneuvering ICD-associated DAMPs hinges on the inflammatory tumor immune microenvironment of MPM. Thus, we plan to assess the distinctive TiME to analyze the immune profiles of distinct MPM subtypes, which is crucial to interpret varied prognoses and efficacy of immunotherapy.

Despite the fact that an increasing amount of predictive models related to immunotherapy have been constructed to elaborate subtypes of MPM, ICD-associated DAMPs and their receptors were barely based upon to construct a predictive classification model. In this research, we performed consensus clustering analysis based on the ICD-associated DAMPs gene set and investigated the impact of DAMPs and their sensing receptors on the immune status from a variety of perspectives and on the survival expectancy of MPM patients. Additionally, the DAMPs-based classification we established was assessed for its predictive value of immune checkpoint inhibitors (ICIs) applied to mesothelioma (the flow chart of analysis is demonstrated in Figure 1 ). Our research offers novel information to discover the potential molecular mechanisms in different subtypes of MPM, which may fulfill the demand for precision immunotherapy of MPM.

Our study categorized MPM patients into two subgroups based on DAMPs and PRRs. In the nuclear DAMPs subtype, nuclear-associated HSP90AA1, HSPA4, CALR and HMGN1 were strongly expressed in the nuclear DAMPs subtype, whereas PRRs modulating activities of inflammasome or immune cells, such as TLRs, AIM2 and NLRP3, were primarily expressed in the inflammatory DAMPs subtype.

HMGN1(also known as alarmin), as a member of the high-mobility group protein family, is activated in undifferentiated cells which proliferate continuously. Extracellular HMGN1 functions as an innate danger-associated inflammatory mediator directly inducing the generation of cytokine and DC maturation. Upon translocation to the cytoplasm, it binds to PRRs to initiate proinflammatory signaling. Increasing expression of HMGN1 might provoke chronic inflammation contributing to carcinogenesis and indicating a poorer prognosis (55). The intracellular functions of CALR as a crucial regulator of Ca2+ homeostasis and the integrin-dependent signaling is probably required for tumor progression (56) which therefore implies that CALR expression is robustly related to prompt tumor progression and poor prognosis (57, 58) in the nuclear DAMPs subtype. HSPs in cells are essential in protein folding. The enriched Notch signaling positively regulate the activity of the mTOR pathway (59), and mTOR complex 1 (mTORC1) activates MYC-induced protein synthesis (60). Extreme endoplasmic reticulum (ER) stress and UPR are triggered by the uncontrolled buildup of misfolded proteins in the ER, which induce biological effects via upregulation of molecular chaperones such as HSP (e.g., HSP90AA1 and HSPA4). Elevated expression of HSPA4 and HSP90AA1 is related with poor clinical outcomes in cancer patients (61).

However, DAMPs alone are insufficient to elicit an ICD, and a corresponding receptor is required to generate biological effects. Toll-like receptors (TLRs), an evolutionarily conserved transmembrane protein expressed in epithelial cells and immune cells, serve as an important receptor to identify the DAMPs. TLRs bind with ligands such as HMGN1, CALR, HSPA4 and HSP90 to generate the affiliated biological effects via activation of MyD88-dependent and independent pathways. AIM2 and NLRP3 are sensors of DNA released from necrotic cells or increased Ca2+ release that initiate the inflammasome assembly (62, 63). Interleukin (IL)-1β and IL-18 are the inflammasome effector cytokines released as a result of signaling pathways that are regulated by inflammasomes. These cytokines assure an optimum inflammatory immune response against cancer cells (64). Therefore, MPM belonging to nuclear DAMPs subtype with higher expression of nuclear-associated DAMPs genes and lower corresponding PRRs expression are more likely to exhibit more aggressive biological behaviors. Nuclear-associated DAMPs are needed to regulate the progression of nuclear DAMPs subtype of MPM, which in consequence leads to poor outcomes. Under the circumstances of PRRs deficiency, the downstream signaling pathway cannot be activated, even with the presence or overexpression of DAMPs, causing effector cells suppressed and adapted immunity muted, which probably accounts for the suppressed immune microenvironment of the nuclear DAMPs subtype and primary resistance to immunotherapy, consistent with the predicted results of our survival analysis and exploration of immune status and ICIs efficacy.

The underlying biological pathways were then explored with functional analyses. Based on the DEGs, GSEA and GSVA identified inflammatory DAMPs subtype considerably enriched in the pro-inflammatory pathways that enhance adaptive immune responses. Apart from the toll-like receptor signaling pathway in line with higher expression of TLRs, TNFR binding and regulation of type I interferon-mediated signaling generates the antiproliferative effects of type I IFNs and TNFs (65) via activating the proinflammatory NF-κb pathway (66). Cytokine/chemokines pathways (such as IL-2 family, IL-12, IFN-γ, TNF superfamily, chemokine signaling, and IL6-JAK-STAT signaling) and effector immune cell pathways (such as NKT, DC, and T cell activation) were also significantly overexpressed. Upregulated IFN-γ signaling amplifies the antitumor response by mediating induced effects of IL-12 (67) and produces chemokines that attract immune effector cells, effectively changing the TME (68). The release of IFN-γ also leads to increasing antigen presentation of cancer and noncancer cells. Inflammatory chemokines induce recruitment of monocytes and help to support and regulate activated T cells (69). IL-2 mainly produced by CD4+ T cells activates effector T cells and innate lymphoid cells (ILCs) (70). The cytokines aforementioned regulate pro-inflammatory immunity by linking intrinsic and adaptive immune responses. Furthermore, ATM pathway, integrin A4B1 pathway and FGF pathway are enriched in this subtype. ATM (Ataxia-telangiectasia mutated proteins) is a key regulator of the DNA damage response (DDR) and contributes to cell cycle checkpoint maintenance, DNA damage repair and telomere maintenance in DNA double-strand breaks (DSB). The ATM pathway also regulates the suppression of anti-tumor immune cancer-associated fibroblasts (CAFs) differentiation (71). Advanced desmoplasia and stromal changes caused by CAFs have been identified as substantial factors in the progression of MPM (72). Accordingly, inhibition of ATM has the potential to overcome immune resistance in combination with ICIs in a minor subset of MPM patients of inflammatory DAMPs subtype. Fibroblast growth factors (FGFs) and integrin-α4β1 pathways are involved in oncogenic behaviors such as metastasis, angiogenesis, and activation of CAFs (73–75). Thereby, inhibitors targeting FGFs or integrin A4B1, or administrating anti-angiogenic agents may introduce promising directions for the management of this subtype. For the nuclear DAMPs subtype, GSEA revealed that expression of cancer hallmarks of MYC targets, NOTCH signaling, TGF-β signaling, unfolded protein response (UPR), MTORC1 signaling, and IL-8 production was higher in the nuclear DAMPs subtype. MM cells are dependent on Notch signaling, leading to activation of the phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) signaling pathway (76). NOTCH signaling facilitates immune escape by up-regulating PD-L1 and is associated with the expansion of exhausted CD8+ T cells (76). Expression profiles of malignant mesotheliomas revealed that 46% displayed altered expression of RPTOR (mTORC1 component) that activates the mechanistic target of rapamycin complex 1 (mTORC1) to enhance MM cell growth (77) suggesting the worse outcome of nuclear DAMPs subtype. On the other hand, studies have demonstrated the potential role of IL-8 as a driver of resistance to ICIs and that IL-8 has an essential role in reinforcing the immunosuppressive microenvironment and triggering EMT by determining the types and quantity of myeloid cells infiltrating tumors (78, 79). TGF-β is correlated with suppressing T cell proliferation and activation, impairing DC and NK cell function, encouraging Treg cell differentiation, and boosting CAF activities, ultimately gives rise to resistance to the immunotherapy (80) in the nuclear DAMPs subtype. Thus, targeting the aforementioned pathways is a plausible way to modify the suppressive immune microenvironment and provides new therapeutic options for this subtype.

The components in the tumor immune microenvironment (TiME) provide clues to predict MPM patient outcomes and ICIs responses (81). Activated memory CD4+ T cells, M1-like macrophages, and activated DCs were more prevalent in the inflammatory DAMPs subtype according to the CIBERSORT, while the ESTIMATE revealed that a significantly higher immune score, which have positive relations with better outcomes and more benefits from ICIs. Comparatively, M2-like macrophages of the nuclear DAMPs subtype were massively recruited, which have recently been discovered to participate in promoting resistance to ICIs therapy and predicting a poor outcome (82, 83). Furthermore, the improved OS and response rate to ICIs in the inflammatory DAMPs subtype may be associated with the dramatically elevated expression of CD8A and PD-1. These findings support the previous report that MPM patients with increased expression of CD8A and PD-1 enjoy a favorable prognosis and potentially benefit from ICIs in preceding clinical trials (84) (85). The expression of MHC class I and class II protein, particularly human leukocyte antigen class I (HLA-I) alleles, and proinflammatory chemokines and immunomodulators, especially CXCL10 and CXCR3, is more robust in the inflammatory DAMPs subtypes than the other. MHC expression on tumor cells from treatment-naive patients positively correlates with the clinical outcome and response to anti-CTLA-4, anti-PD-1, or their combination by recognizing tumor-specific antigens (86). In respect of CXCL10 and CXCR3 as downstream adjuvant effectors of type I IFNs, their signaling boosts the efficacy of immunotherapy by increasing immune infiltration of cytotoxic lymphocytes (CTLs), natural killer cells (NKs) and DCs (87). Therefore, upregulated CXCL10 and CXCR3 are positively linked with the efficacy of immunotherapy (88, 89). The above results laid further foundations for our reasonably predicting the better prognosis and response rate of ICIs treatment in the inflammatory DAMPs subtype than the other.

In the TCGA cohort, mutations of TTN were found more frequently in the inflammatory DAMPs subtypes and mutations of TP53 in the nuclear DAMPs subtypes were clinical outcome and efficacy on the trend. Previous studies showed that patients with mutated TTN are associated with longer progression-free survival or overall survival (90). TTN expression was favorably associated with the infiltration levels of effector T cells owning an inflammatory TiME and TMB in numerous tumor types, and therefore is linked with susceptibility to immune checkpoint inhibitors (91–93). We can infer that TTN may have a connection with clinical outcomes and efficacy of immunotherapy as our model projected (94). As for TP53 mutation implying a more malignant nature, the nuclear DAMPs subtype is indicated to require more intense management (95).

Copy number deletion in MPM is a characteristic genetic alteration that may result from altered methylation caused by external factors such as asbestos. Copy number deletions of cancer suppressor genes (including CDKN2A/B, FOCAD, NF2, etc.) are always accompanied by adjacent functional genes (including MTAP, MLLT3, etc.) deletion that synergistically contributes to oncogenesis. Studies have shown that copy number deletion is associated with loss of tumor neoantigens and reduced gene expression of immune-related pathways (96), which predicts dismal immune efficacy in this subtype as our subtyping model does (97). Copy number loss of Cyclin-dependent kinase (CDK) inhibitor 2A (CDKN2A) occurs at a significantly higher frequency in the nuclear DAMPs subtype. CDKN2A copy number loss suggests dismal outcomes and predicts immunotherapy resistance (98). CDKN2A encodes p16INK4a which regulates cell-cycle by inhibition of CDK4/6. Emerging clinical data demonstrates selective CDK4/6 inhibitors widely used in clinical practice contribute to PD-L1 up-regulation and immune surveillance enhancement. Hence the combination of ICIs and CDK4/6 inhibitors is a worthwhile strategy for improving outcomes in the immunotherapy-tolerant nuclear DAMPs subtype of MPM. Fifty-seven percent of patients with CDKN2A copy number loss had methylthioadenosine phosphorylase (MTAP) co-deletion in the TCGA validation cohort, since both genes reside in the same cluster of the 9p21 region (99). Co-deletion occurs more frequently indicating poorer prognosis in nuclear DAMPs subtype than the other (72% vs 48%, P=0.032). Studies show that methionine adenosyltransferase 2A (MAT2A) inhibitors induce synthetic lethality of MTAP-deleted cancer, especially in combination with taxanes and gemcitabine (100), which demonstrate a potential to treat MPM of the nuclear DAMPs subtype. Accordingly, distinctive copy number deletion also offers novel approaches to the management of the immune-resistant subtype of MPM.

The antigenicity and adjuvanticity determine the immunogenicity of cell death. Despite MPM is characterized by genetic alterations in tumor suppressor genes (101, 102) suggesting a lack of tumor-associated antigens (TAAs) and a low level of tumor mutation burden (TMB) (103), studies showed that the lowest antigenicity associated with tumorigenesis might be minor but is sufficient to support immunogenicity (12). In this regard, the adjuvanticity of DAMPs and their PRRs in immunogenic cell death may constitute a promising target for activating the immune response, since it is in a superior position to preserve homeostasis of immune microenvironment. Counteracting the inhibition of DAMPs and PRRs (e.g., TLRs stimulators) may suppress tumor growth as well as balance the production of cytokines within the TiME, and further, suppress the immunosuppressive cells while activating the immunostimulatory or effector cells (12), and thus the efficacy of immunotherapy can be enhanced by intervening with the immunomodulatory effects of DAMPs and its downstream signaling. According to the perspectives above, pathways and genomic alterations more peculiar in the nuclear DAMPs subtype of MPM which is primarily resistant to ICIs can be manipulated to modulate the immune status to some extent.

However, there are drawbacks to our study. Firstly, a few genes in the established gene list were excluded due to limitations of expression sequencing by microarray, yet the modified gene set was representative of the concerned genes in the process of ICD. Secondly, there is insufficient sequencing data on patients treated with ICIs since mesothelioma is a rare tumor. Models constructed of mice of the same strain (BALB/c) with identical genetic backgrounds inoculated subcutaneously with the same cell lines (AB1-HA cells) and then treated with anti-CTLA-4 and anti-PD-L1 were selected for additional validation. Moreover, the ORR of PD-1 inhibitors applied to MPM patients of two separate subtypes in the validation cohort did not reach statistical significance, although there was a trend that implied patients belonging to the inflammatory subtype benefit more from ICIs and the DCR, probably on account of small sample size or limited efficacy of ICIs as a single agent in MPM. Additional research is expected to confirm the validity and clinical practicability in larger cohorts or elaborately designed clinical trials.

In conclusion, we identified two molecular subtypes via K-means analysis based on the expression of ICD-associated DAMPs and their corresponding receptors in MPM. Characteristic signaling pathways and different immune statuses in these two subtypes result in disparate prognoses and efficacy of immune checkpoint inhibitors.

Our research offers a novel method to predict the prognoses and identify the MPM patients with a potential to benefit from ICIs and provides a new perspective to enhance the efficacy of immunotherapy for MPM patients with primary resistance to ICIs. Our work has made a step forward in the process of development of precision therapy in MPM.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found within the article/supplementary materials.

ZL: conceptualization, methodology, software, data curation, investigation, validation, visualization, writing- original draft preparation, writing- reviewing and editing. HB: supervision, funding acquisition. RW: writing- reviewing and editing. JW: conceptualization, supervision, writing- reviewing and editing, funding acquisition. All authors contributed to the article and approved the submitted version.