Hepatocellular carcinoma (HCC) accounts for the majority type of primary liver cancer and is the sixth most common and fourth deadly malignancy globally (Bray et al., 2018). Liver resection or transplantation is still the first choice for radical cure, while only a small portion of patients fit the bill. As a matter of fact, more than 70% of the patients are already at a middle-to-late stage when first diagnosed. Despite the improvement of several new treatments such as immunotherapy and target therapy, the 5 years survival rate has remained less than 20% (Allemani et al., 2015; Katsanos et al., 2017). Thus, it is of great importance to develop novel therapeutic targets and reliable prognostic models for medical decision making.

Pyroptosis, a novel defined programmed cell death, is characterized by the cleavage of gasdermins and the subsequent cell membrane ruptures and necrotic cell death (Rogers et al., 2017; Shi et al., 2017). Pyroptosis was involved in innate immunity and initially found to play a crucial part in fighting against infection (Shao et al., 2021). However, increasing evidences have suggested that pyroptosis also takes part in cancer development (Fang et al., 2020; Tan et al., 2020). Deletion of NLRP3 (one of the key components of pyroptosis) in transgenic mice was found to be correlated with colon cancer development (Dupaul-Chicoine et al., 2010). NLRP3 inflammasome may also drive liver injury and fibrosis and may contribute to the development of liver cancer (Wree et al., 2018). More recently, the pyroptosis-related gene signatures for ovarian cancer and lung cancer has been proved to be prognostic (Lin et al., 2021; Ye et al., 2021). Interestingly, pyroptosis-induced inflammation was found to synergize with anti-PD1 therapy through triggering antitumour immunity (Wang et al., 2020). In HCC, pyroptosis was proved to be restrained in HCC tissues and cells, small molecule drugs such as berberine, euxanthone and miltirone may induce HCC cell death through triggering pyroptosis (Chu et al., 2016; Chen et al., 2018; Zhang et al., 2020a). However, the role of different pytoptosis-related genes in HCC has not been reported. Here, we explored the diverse expression, prognosis, associations with immune status and response to immunotherapy of pytoptosis-related genes in HCC.

Pyroptosis is a novel kind of cellular necrosis, characterized by releasing huge amount of inflammatory factors and the formation of bubble-like morphology. It has been suggested to have both tumor protecting and promoting roles. More recently, researchers have focused on the induction of pyroptosis in tumor cells, which may be a novel therapeutic target (Xia et al., 2019). The pyroptosis-related genes have been proved to be prognostic in both ovarian cancer and lung cancer. However, the diverse roles of pyroptosis-related genes in HCC have yet to be analyzed.

For the first time, we explored the expression and prognostic roles of pyroptosis-related genes in HCC. Our results indicated that 24 genes were differentially expressed between the HCC tissues and normal tissues. A gene signature containing five pyroptosis-related genes (GSDME, CASP8, SCAF11, NOD2, CASP6) could well predict OS in HCC patients. The risk score generated from the gene signature was shown to be an independent risk factor for poor prognosis. Similar to our results, prior studies have also proved the prognostic roles of pyroptosis-related genes in lung cancer and ovarian cancer (Lin et al., 2021; Ye et al., 2021).

GSDME, also known as deafness autosomal dominant 5 (DFNA5), one of the genes from our gene signature, was first identified for its involvement in inherited hearing impairment (Van Laer et al., 1998). GSDME was one of the core executors in cell pyroptosis. GSDME was cleaved by caspase-3 into pore-formation of GSDME-N domain, which triggered pyroptosis in cells (Rogers et al., 2017). GSDME was originally identified as a tumor suppressor because its expression was downregulated in breast cancer and reduced GSDME was associated with poor survival of breast cancer patients (Kim et al., 2008). However, GSDME was also found to be overexpressed in head and neck squamous cell carcinoma, lung squamous cell carcinoma and cholangiocarcinoma. Overexpression of GSDME was associated with poor survival of head and neck squamous cell carcinoma (Liu et al., 2021; Zhang et al., 2021). Hence, GSDME may play different roles in different types of cancer. In this study, our result showed that GSDME may play a tumor promoting role in HCC. Though, small molecule drugs such as cannabidiol and miltirone may induce HCC cell death through GSDME-mediated pyroptosis (Zhang et al., 2020b; Shangguan et al., 2021), non-cleavable or pore-defective GSDME was also proved to be not tumor suppressing (Zhang et al., 2020a). The detailed role of GSDME in HCC are worthy of further exploring.

NOD2, another gene from our gene signature, is reported to promote inflammation and apoptosis. It has been indicated that NOD2 can contribute to myocardial ischemia/reperfusion injury through inducing cardiomyocyte apoptosis and inflammation (Liu et al., 2016). Emerging evidence has been found between NOD2 and human cancers. NOD2 was found to be prognostic in kidney cancer (Xu et al., 2017). NOD2 was also proved to be upregulated and activated in HCC samples, overexpression of NOD2 correlated with poor survival. Mechanically, NOD2 increased liver inflammation via inducing the nuclear autophagy pathway (Zhou et al., 2021). Similar to their results, our results showed that NOD2 was elevated in HCC tissues and correlated with poor survival.

By GSEA analysis, our results indicated that endocytosis, ubiquitin mediated proteolysis and regulation of autophagy were enriched in the high-risk group, which indicated stronger protein metabolism in the high-risk group. We speculated the hypermetabolic behavior in the high-risk group may lead to high malignancy and poor prognosis, which needed to be elucidated in further studies. In order to explore the correlations between risk score and immune cell infiltration in tumor, ssGSEA was adopted. Generally speaking, the high abundance of Treg contributes to immunosuppression and poor survival (Zhou et al., 2016; Zhang et al., 2019). What’s more, macrophages, especially the tumor associated macrophages are mainly M2 macrophages, which are shown to be tumor promoting (Deng et al., 2021). Here, in this study, our results showed that macrophages and Treg were enriched in the high-risk group and patients in the high-risk group also had poor prognosis. Although, our ssGSEA results seemed not always to be consistent, it was acceptable from an overall perspective.

Immunotherapy of ICIs has become an inspiring treatment for various of advanced cancers including HCC. Recently, results from the study of KEYNOTE-240 also proved the value of pembrolizumab as second-line therapy in patients with advanced HCC(Finn et al., 2020). However, not all HCC patients can benefit from the ICIs therapy. It is worthy selecting patients who may benefit from immunotherapy. In this study, we explored the relationship between IPS and risk score. The results suggested that the low-risk group had higher IPS scores, IPS-PD1 blocker scores, IPS-PD1/CTLA4 blocker scores and IPS-CTLA4 blocker scores, indicating the possibility of applying the risk score to predict response of immunotherapy. Besides, we also investigated the relationship between risk score and common immune checkpoints including PD1, CTLA-4, LAG-3 and TIGHT. Our results showed that the expressions of PD1, CTLA-4, LAG-3 and TIGHT were elevated in the high-risk group which also indicated that the low-risk patients may have a better response to ICIs. Thus, our predictive model may be helpful for selecting patients in clinical practice.

Nevertheless, our study had some limitations. As all the data included for analysis were retrieved from online databases, further in vivo/vitro studies are still needed to verify these findings. Besides, we did not explore the post-translational modifications of these genes, as post-translational modifications were of great importance for controlling intracellular signal transduction and cell functions. What’s more, the detailed signal pathways for these genes affecting clinical survival deserved to be investigated in future studies. Last but not least, with the development of single-cell RNA-seq (scRNA-seq) technique, we are now able to focusing on the specific types of cells which mainly express the pyroptosis-related genes (Zhao et al., 2020; Song et al., 2021). This would be certainly more meaningful for us to understand the HCC microenvironment. Till now, few single-cell studies have explored the role of pyroptosis in HCC, it is a promising research direction for our future study.

In conclusion, we systemically explored the expression, predictive signaling pathways, prognosis and associations with immune infiltration of pyroptosis-related genes in patients with HCC. Our research gives new insights regarding the roles of pyroptosis-related genes to HCC and may be helpful for selecting patients in clinical practice.