Introduction
PAAD is a cancerous tumor characterized by rapid spread and poor prognosis, resulting in 466,003 new deaths in 2020 and 495,773 new cases of pancreatic cancer worldwide [
1]. Studies showed that pancreatic cancer incidence and mortality rate are nearly equivalent and the 5-year survival rate is estimated at only 1% [
2]. The majority of PAAD patients are already suffering from advanced diseases when they are diagnosed. However, up to 80% of pancreatic cancers cannot be resected because of their highly malignant and early metastasis [
3]. Patients with pancreatic cancer after resection still have a poor prognosis [
4]. Hence, it is urgent to develop a prognostic model and identify biomarkers in diagnosis of PAAD.
Pyroptosis is a type of regulated necrotic cell death induced by inflammatory caspases. It mainly relies on the inflammation to activate a part of the caspase family of proteins, so that inflammation cleaves the Gasdermin protein, activates the Gasdermin protein, and the activated Gasdermin protein translocates to the cell membrane to form holes, then the cell swells, the cell membrane ruptures, and finally leads to the efflux of the cytoplasm and the formation of pyroptosis. Caspase-1, 4, 5, 11 are pro-inflammatory cysteine proteases, all belong to the cysteine aspartate proteolytic enzyme family, and this class of proteases is critically involved in the body's generation of inflammatory responses and innate immune responses [
5]. Caspase-3 is a master regulator of apoptosis, while those associated with inflammation include caspase-1, 4, 5, 11, 12, 13, 14, among which caspase-1, 4, 5, 11 mainly mediates pyroptosis [
6]. Endogenous and exogenous stimulatory signals act on the inflammasome through different pathways to activate caspase-1, which mediates the osmotic swelling cleft of the cell, the formation of small pores in the cell membrane, the efflux of intracellular substances (such as lactate dehydrogenase, etc.). IL-1 β and IL-18 precursor cleaves and induces the synthesis and release of other inflammatory factors, adhesion molecules. Amplifying the local and systemic inflammatory response is the main mechanism by which pyroptosis occurs [
7]. It has been shown that pyroptosis plays a dual role in promoting and inhibiting the development of cervical cancer. Studies have shown that the NLRP3 inflammasome is involved in the innate immune response to cervical cancer, and its expression is widely present in tumor cells [
8]. NLRP3 inflammatory activation can be achieved through humans, lysosomal rupture, and reactive oxygen species. In cervical cancer, the NLRP3 inflammasome is mainly activated by reactive oxygen species to induce pyroptosis [
9]. In HPV infected cervical cancer cells, aim2 can play a tumor suppressive role by stimulating pyroptosis [
10]. However, several studies have found that removal of pro-inflammatory factors produced by pyroptosis can inhibit cervical cancer cell growth while impairing the body’s immune effect on tumor cells [
11,
12].
An increasing number of studies have shown that pyroptosis plays an important role in cancer progression. Therefore, in-depth study of the role of pyroptosis in pancreatic carcinogenesis and progression, as well as the establishment of a relevant prognostic model of pyroptosis, is of great importance for the treatment of PAAD. To the best of our knowledge, there is no pyroptosis-related prognostic model in PAAD has been established to predict the prognosis of patients with PAAD. Therefore, a novel prognostic model based on pyroptosis-related genes for predicting survival of patients with PAAD is highly needed. In this present study, we aimed to establish a prognostic model on the basis of pyroptosis-related genes to predict the prognosis of patients with PAAD. In addition, we verified the function of GSDMC in PANC-1 and CFPAC-1 cells, which might be a promising therapeutic target for the treatment of PAAD. Our study systematically explored the prognostic value of pyroptosis-related genes and their correlations with clinical characteristics, thus shedding light on the promising roles of pyroptosis-related genes as potential prognostic biomarkers and novel therapeutic targets for PAAD patients.
Materials and methods
Data acquisition and preprocessing
Data normalization
In order to integrate the expression data from TCGA and GEO database (GSE71729 and GSE57495), we performed the batch normalization to remove batch effects. Firstly, we downloaded mRNA-seq FPKM data of TCGA PAAD patients. Secondly, we downloaded the microarray expression data from GEO database (GSE71729 and GSE57495). Thirdly, these expression data from TCGA and GEO database were log2-transformed. Finally, batch normalization was performed across abovementioned data using the combat function in "sva" package in R software (version 4.1.2) [
13]. This method was widely utilized to combine different datasets in previous studies [
14‐
16].
Identification of differentially expressed pyroptosis-related genes
The “DESeq2” package was used to identify differentially expressed pyroptosis-related genes in 178 PAAD samples and 167 normal samples. P < 0.05 and |log2 fold change (FC)|> 1.2 were set as as cut-off values. The volcano of pyroptosis-related genes and heatmap of differentially expressed pyroptosis-related genes were drawn using the OmicStudio tools (
https://www.omicstudio.cn/tool). Protein protein interactions (PPIs) were plotted by using string database (
https://string-db.org/) and boxplots were drawn with the R package “ggpubr”. The minimum interaction score required for PPI analysis was set at 0.4 (medium confidence).
Enrichment analysis of differentially expressed pyroptosis-related genes
The biological process enrichment of 25 differentially expressed genes were analyzed with Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) through R statistical software including “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, “ggplot2”, and “GOplot” packages. In addition, gene set enrichment analysis (GSEA) was performed in order to identify the biological process and signaling pathways that differ between high-risk groups and low-risk groups in PAAD. Our reference gene sets were derived from the C2 subcollection (c2.cp.kegg.v7.5.1.symbols.gmt). The significance thresholds were determined by 1000 permutation analyses, and we considered significant results when the p value was less than 0.05.
Identification of prognostic genes
Our training set consisted of 178 PAAD samples and 167 normal pancreatic samples from the TCGA and GTEx databases. To investigate the relationship between the expression levels of pyroptosis-related genes and overall survival (OS) of PAAD patients, we conducted a univariate Cox regression analysis using the "survival" package. A significant filtering criterion was set at p < 0.05 for further analysis. We next eliminated gene collinearity and reduced the number of genes using LASSO Cox regression. Finally, we conducted multivariate Cox regression analysis on the basis of univariate Cox regression.
Construction and validation of a prognostic model based on pyroptosis-related genes
The risk score was calculated according to the centralized and standardized PAAD mRNA expression data in the train set.
$${\text{Risk}}\, {\text{score}} = \sum\nolimits_i^n {{x_i}{y_i}}$$
X represents the coefficient of pyroptosis-related genes in LASSO Cox regression analysis, Y represents the gene expression of pyroptosis-related genes. PAAD patients were divided into high-risk and low-risk groups based on the median risk score, and the overall survival (OS) between these two groups was analyzed. Receiver operating characteristic (ROC) curves were produced by the timeROC package to evaluate the prognostic efficiency of the model. To make the model more convincing, we utilized the PAAD cohort in the GEO database for validation. The expression of each pyroptosis-related genes was also normalized, and the risk score was then calculated by the above formula. PAAD patients in the GEO cohort were also grouped into high-risk and low-risk groups according to the median risk score, and the OS between the two groups was also compared. Next, to determine if risk score was an independent prognostic factor for OS in PAAD patients in the train set, univariate and multivariate Cox regression analyses were conducted. Covariates included age, gender, grade, stage, T, and N.
Construction of nomogram and calibration curves
The nomogram was built using the “RMS” package of R software to predict individual survival probability, and calibration curves for the prediction of 1 -, 2-, and 3-year survival rate of PAAD patients were plotted.
Drug sensitivity analysis
Using the pRRophetic package in R software, the sensitivity score of each small moleular compound was calculated for each patient in the high-risk group and low-risk group. Then, we used PubChem (
https://pubchem.ncbi.nlm.nih.gov/) website to visualize the conformations of drugs in 3D.
Cell culture
CFPAC-1 and PANC-1 cells were purchased from Procell Life Science and Technology Co., Ltd. (Wuhan, China). CFPAC-1 and PANC-1 cells were cultivated in RPMI-1640 (Hyclone) supplemented containing 2 mM L-glutamine and 10% FBS (Life Technologies).
siRNA sequence
The siRNA sequences were as follows: si-GSDMC-1:5′-GGAUCCAGAGCC AUCAUUU-3′. si-GSDMC-2: 5′- CCUAGA AACUGUUGUGACA-3′.
CCK8 assay
Cell viability was determined by CCK8 kit. In short, CFPAC-1 and PANC-1 cells transfected with si-GSDMC-1 or si-GSDMC-2 were inoculated in 96 well plates × 1000 cells/well). Add CCK8 and use the multimode microplate reader at 0, 24, 48, 72 h respectively. The optical density of each well was measured at 450 nm. Each experiment was repeated three times.
EdU assay
EdU kit was used for EdU determination (Ribobio, # C10310-2). The EdU test solution was inoculated into CFPAC-1 and PANC-1 cells transferred with si-gsdm-1 or si-gsdm-2, respectively. Continue to culture in the incubator for 2 h, and then fix with 4% paraformaldehyde for 30 min. The dye is then dyed according to the manufacturer's scheme and the image is taken using EVOS FL automatic microscope. Finally, using Image J software to count the number of cells. Each experiment was repeated three times.
Wound healing assay
The ability of cell migration was evaluated by a wound healing experiment. CFPAC-1 and PANC-1 cells transfected with si-GSDMC-1 or si-GSDMC-2 were inoculated in 6-well plates. When the cells reach reaching a confluence of 100%. Use a 10 µl pipette to form a wound in the center of the cell monolayer, and then continue to culture in the incubator. At a specific time, using Image J software to count the wound area. Each experiment was repeated three times.
Western blotting
Two small inferring RNAs (siRNAs) were employed to knock down GSDMC. After 48 h of transfection, cells were lysed in RIPA buffer with a phosphatase inhibitor cocktail (biomake, #B14001, #B15001). Proteins were loaded and separated by electrophoresis on SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a nitrocellulose membrane. The signals were visualized using the ECL Kit (Meilunbio, #MA0186). Antibodies used were anti-GSDMC (diluted 1:1000, proteintech, #27,630-1AP) and anti-ACTIN antibody (diluted 1:1000, Abclonal, #AC026).
Lentiviral production, infection, and construction of stable cell lines
We constructed shRNA sequences (sh-1:5′-GGAUCCAGAGCC AUCAUUU-3′. sh-2:5′-CCUAGA AACUGUUGUGACA-3′) using the plko.1-puro-gfp vector, and a scramble sequence (5′-TTCTCCGAACGTGTCACGT-3′) was designed as a negative control. We also constructed GSDMC-overexpressing lentivirus using the pCDH-CMV-Puro vector (GeneChem, China). Next, we followed the manufacturer's instructions for lentiviral transfection. Briefly, shRNA or pCDH plasmid was cotransfected with the packaging plasmids psPAX2 and MD2G in 293 T cells using lipofectamine 3000. Supernatants were collected 48 h after transfection and filtered through 0.22 μm low protein binding filters. 1 ml of supernatant were used to transduce PANC-1 or CFPAC-1 cells, and the medium was changed 2 days after infection. The infected cells were screened by puromycin for 10 days.
Cell cycle analysis
Propidium iodide (PI) staining was utilized to analyze cell cycle. In short, for the cell cycle analysis, cells (1 × 107) were washed using PBS and fixed using 70% ethanol for 30 min at room temperature. We then stained the cells with PI containing RNase A (Thermo, # F10797) after washing them three times with PBS. A FACScan (Millipore) was used to measure the red signal and ModFIT LT v3.1 software was used to analyze the FSC data.
Chemical treatment of cells
In six-well dishes at a density of 1000 cells per well, the PANC-1 cells were trypsinized and plated. Cells were allowed to attach overnight and then exposed to corresponding concentrations of chemical treatment. Approximately 48 h after chemical treatment, we replaced the media with fresh media, and incubated the plates at 37 °C.
Statistical analysis
Statistical analysis was performed by two-tailed unpaired t-test, one-way ANOVA and two-way ANOVA in GraphPad Prism software (version 8.0.2). When the p value was less than 0.05, the results were considered statistically significant.
Discussion
Pyroptosis is a caspase-1-or caspase-11-dependent programmed cell death [
17,
18]. It has become increasingly evident that pyroptosis plays an important role in the progression of cancer in recent years. Research has shown that pyroptosis-related genes play a different role in different types of cancer. Cell death releases inflammatory factors to provide tumor cells with a suitable environment for survival [
19]. Pyroptosis can promote tumor cell death, making pyroptosis a potential prognostic and therapeutic target in cancer [
20]. In HPV infected cervical cancer cells, AIM2 plays a tumor suppressive role by stimulating pyroptosis [
10]. However, the role of pyroptosis genes in PAAD is unclear. This study aimed to construct a prognostic model regarding pyroptosis-related genes for diagnosis and prediction of prognosis in patients with PAAD.
In recent years, the search for PAAD biomarkers, prognostic markers, and prognostic models has been gaining increasing attention [
21‐
25]. Patients with PAAD may benefit from these models since they have great ability to predict prognosis. Consistent with previous studies, the prognostic model we constructed still has good performances for predicting the prognosis of patients with PAAD.
We constructed a prognostic risk model using 5 genes (IL18, CASP4, NLRP1, NLRP2, GSDMC) through univariate Cox and Lasso Cox regression analysis. IL18 is a proinflammatory cytokine that promotes IFN- γ Of secretion [
26]. The levels of IL18 are significantly elevated in patients with gouty arthritis [
27] and rheumatoid arthritis [
28]. CASP4 is a gene involved in encoding a protein involved in immune response and inflammation [
29], and studies have shown that decreased expression of CASP4 is associated with poor prognosis in esophageal squamous cell carcinoma [
30], but low expression of CASP4 in our prognostic model is more favorable for the survival of PAAD patients. NLRP1 is an innate immune receptor that assembles into an inflammasome to induce pyroptosis in human corneal epithelial cells [
31], which is consistent with the manifestations in our constructed model. NLRP2 is highly expressed in renal tubular epithelial cells and plays a role in promoting inflammation [
32]. GSDMC is also one of the most important model genes in our study. Pyroptosis is primarily a programmed cell death mediated by Gasdermins (GSDM) [
33]. In which GSDM contains molecules such as Gasdermin C (GSDMC), Gasdermin D (GSDMD) [
34]. Studies have pointed out that high expression of GSDMC promotes melanoma metastasis [
35]. In gastric cancer, GSDMC inhibits tumor cell growth [
36]. Overexpression of GSDMC causes poor prognosis in lung adenocarcinoma [
37]. An increasing number of studies have shown that the application of GSDMC is extensive. However, studies of GSDMC in PAAD are quite limited. Therefore, we performed cell experiments with GSDMC alone to verify the specific role of GSDMC in PAAD. Our results showed that the growth, proliferation, and migration were inhibited in PAAD cells with silencing of GSDMC, which is consistent with previous studies. Therefore, we speculated that GSDMC contributed to the poor prognosis of pancreatic cancer mainly by promoting tumor growth and migration.
Previous studies showed that PAAD patients who benefit from immunotherapy is limited [
38,
39]. Studies have pointed out that the future treatment of pancreatic cancer should be through active combination and adoptive immunotherapy [
40]. The combinatorial approach of immunotherapy in conjunction with other modalities is believed to be a promising treatment strategy. Increasing studies have shown that T cells play a key role in immunotherapy [
41]. The higher CD8 expression on T cells confers a better prognosis in esophageal, colorectal, and non-small cell lung cancer [
42,
43]. The expression of T cells CD8 in low-risk subgroup was higher than that in high-risk subgroup in our study, which is consistent with previous study [
44]. In addition, high expression of mast cells resting leads to poor prognosis in hepatocellular carcinoma [
45,
46], which is consistent with our study.
In addition, we screened out four potential small molecular compounds, including A.443654, PD.173074, Epothilone.B, and Lapatinib. A.443654 is a well-known Akt serine/threonine kinase inhibitor [
47], which is equally potent against Akt1, Akt2, and Akt3 within cells (Ki = 160 pM) [
48]. Studies showed that A.443654 could induce apoptosis in chronic lymphocytic leukemia cells in a dose-dependent manner [
49]. In addition, A.443654 plays a key role in cells transition to the G2/M phase [
50]. Therefore, A.443654 may inhibit PAAD by mediating the cell proliferation, which is consistent with the results of our validation experiments. PD.173074, a small-molecule tyrosine kinase inhibitor, which could interfere with the relevant signaling of fibroblast growth factor [
51]. The growth and invasion of Epithelial-mesenchymal transition-induced tumor cells could be inhibited by PD.173074 through EGFR pathway [
52]. It has been demonstrated that the FDA approved antitumor drug Epothilone B could improve microtubule stability and promote α-Ability of tubulin polymerization [
53]. As an EGFR and HER2 tyrosine kinase inhibitor, Lapatinib is approved by FDA to treat patients with HER2-positive breast cancer [
54]. These results indicated that these potential drugs might provide novel insights into the treatment of patients with PAAD.
However, our study has some limitations. First, GSDMC could promote the proliferation, migration, and invasion of PAAD, but its molecular mechanism is still unknown. Second, in vivo function of GSDMC in PAAD still need to be explored in the future.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.