NAFLD encompasses a spectrum of liver lesions, from steatosis to cirrhosis, via NASH. The multiple-hit hypothesis involving plenty of factors has been posited in pathological mechanism of NASH14,15. Pyroptosis is a type of programmed cell death that is closely related to the inflammatory response16,17. While historically much emphasis was placed on apoptosis (i.e., programmed cell death) and necrosis (i.e., non-programmed cell death) in NAFLD17–19, however, the role of pyroptosis in NASH remains unclear. Our present study first constructed a diagnosis model for NASH based on PRGs and then explored the role of pyroptosis in NASH.
We firstly studied the mRNA levels of 57 currently known pyroptosis-related genes (PRGs) in NASH and normal tissues and found that a total of 20 PRGs were differentially expressed. Through analysis of PRGs between normal and NASH samples, we found that most of differentially expressed PRGs were caspase family members (CASP6, CASP9, CASP3, CASP4) and inflammasome sensors (NLRP7, NLRP9, NLRP2, NLRP1, NLRP6). To further assess the diagnostic value of PRGs, we constructed a seven-gene signature via logistic univariate analysis, LASSO regression analysis and logistic multivariate regression analysis and finally screened feature genes and developed a diagnostic model based on seven genes (CASP3, ELANE, GZMA, CASP4, CASP9, IL6 and TP63). Our present diagnostic model showed the AUC was 0.978 for this model, indicating the high performance for differentiating NASH cases from healthy controls. CASP3, known as a classic marker of apoptosis, produces active executors that cleave structural and regulatory proteins in the nucleus and cytoplasm of cells upon activation, thereby regulating cell death. Recent study by Shao et al. showed that activation of CASP3 can trigger pyroptosis through cleaving GSDME and generating a GSDME-N fragment that perforated membranes20. ELANE is one of the major serine proteases secreted by neutrophils, and it activates proinflammatory cytokines such as TNF-α, IL-1β, and IL-18, which are known to be pyroptosis promoters21,22. Kambara et al. proved that GSDMD could be cleaved and activated by ELANE and induce neutrophils to undergo pyroptosis23. In our study, the expression of ELANE was significantly higher while the neutrophil infiltration score was lower in NASH group than in control group, which indicated that ELANE might activate pyroptosis in neutrophils. Cytotoxic lymphocyte-mediated immunity relies on granzymes. Shao et al. has reported that GZMA could cleave and unleash GSDMD's pore-forming activity, which enables natural killer T cells and cytotoxic T lymphocytes to kill gasdermin B (GSDMB)-positive cells through pyroptosis24. In our study, the expression of GZMA was higher in NASH group with higher cytotoxic lymphocytes including natural killer T cells. In summary, 3 genes (CASP3, GZMA and ELANE) in the prognostic model were proven to be pyroptosis promoters. However, we also identified down-regulation of CASP4, CASP9 and IL6 in NASH patients and related to reduced risk of NASH, while these genes should also be involved in pyroptosis. Till now, pyroptosis has not been fully understood, though shared similarities and crossovers in mechanisms to apoptosis have been found17. As disease progresses, multiple ways of cell death may coexist and interact with each other17. So, how these genes interact with each other during pyroptosis still remains unclear and requires further investigation.
It is becoming clear that innate immune signaling could modulate the development of certain metabolic diseases, including nonalcoholic fatty liver disease, characterized by chronic, low-grade inflammation due to a dysregulated innate immune signaling25. Through further analysis of the immune, we found that most of infiltrated immune cells, including macrophages, activated dendritic cells, immature dendritic cells, activated CD8 T-cells, effector memory CD4 T cells, nature killer T cells, activated B cells, and immature B cells were higher in NASH patients compared with healthy controls. Thus, our results also proved that NASH patients present immune imbalance. As pyroptosis is closely related to the immune response, we further applied correlation analysis between PRGs with immune response and discovered that PRGs were closely associated with the changes of HLA-related genes, immune cells and immune-related pathways in the dataset.
Since we were also interested to explore the immune regulation and search effective anti-inflammatory therapy based on PRGs to target a specific group of NASH patients, we further divided NASH patients into two subtypes (cluster 1 and cluster 2) by using consensus clustering analysis and found different immune status in two subtypes. Patients in cluster 2 presented lower profiles of expression of immune cells, including active dendritic cell, central memory CD4 cell, memory B cell, type-1 T helper cell, indicating lower inflammatory condition. In addition, HLA genes play an important role in immune response and immune therapy. Attractively, most HLA genes were also up-regulated in cluster 1, indicating that patients in cluster 1 might be more sensitive to immune therapy. Five immune pathways showed higher activity in cluster1, including Antigen processing and presentation, BCR signaling pathway, interferon Receptor, TCR signaling pathway, and nature killer cell cytotoxicity, while other four immune pathways showed higher activity in cluster 2. Then, we identified potential drugs of NASH disease by using DGIdb database, and discovered that Oltipraz, Metaproterenol, Fingolimod might target certain genes in NASH progress. However, unfortunately, the clinical information of NASH patients is limited, the two subtypes did not show any significant differences in clinical characteristics (Data not shown). So, we only conducted a partial study to observe the phenomenon, and further study is needed.
Since very few studies have comprehensively assessed PRGs in NASH, our study preliminarily studied the diagnostic value of these PRGs and systematically analyzed the relationships between PRGs and the immune state in NASH, providing theoretical support for future research. However, our study has several limitations. Firstly, the total sample size was still relatively small (NASH: n = 257; Normal: n = 142) due to a lack of data and we cannot determine the specific role of these PRGs in NASH, which warrants further study. Besides, our study didn't include samples from different NAFLD and NASH stages, which could not reflect the transition from NAFLD to NASH stage. Secondly, this study was based on bioinformatics analysis, and thus many of the results were theoretical. Thirdly, RNA -seq data in GEO database used in our study are all derived from the lysates of whole livers. Though hepatocytes are the major cell type in the liver tissue, other cell types, including Kupffer cells, immune cells and hepatic stellate cells, are also involved in the pathological process of NASH development. Therefore, further in-depth studies are required to deepen our understandings.
In summary, our study provided a molecular model based on PRGs for NASH diagnosis. We also revealed that pyroptosis was related with immune imbalance in NASH.