A Novel Defined Pyroptosis-Related Gene Signature for Predicting the Prognosis of Endometrial Cancer.

Dis Markers 2022 AI 6 Explanations View Original
Original Paper (PDF)

Unable to display PDF. Download it here or view on PMC.

Plain-English Explanations
Pages 1-2
Pyroptosis and Its Role in Cancer Prognosis

Pyroptosis overview: Pyroptosis is an inflammatory form of programmed cell death, distinct from apoptosis and necrosis. First identified in macrophages responding to pathogen infection, it is now recognized as playing critical roles in inflammatory and immune defenses, as well as in tumor biology.

Pyroptosis and tumors: Pyroptosis can promote inflammatory death of cancer cells and inhibit proliferation and migration of tumor cells, thereby affecting tumor progression and prognosis. For example, decreased expression of the pyroptotic inflammasome GSDMD in gastric cancer cells promotes tumor cell proliferation.

PRG signatures in other cancers: Pyroptosis-related gene (PRG) signatures have been validated as effective prognostic markers in ovarian cancer, lung cancer, bladder cancer, and head and neck squamous cell carcinoma. However, the prognostic value of PRGs in endometrial cancer had not been established prior to this study.

Study objective: This bioinformatics study used TCGA data from 515 EC tumor and 35 normal endometrial tissue samples to construct a novel PRG prognostic signature for EC, then assess its relationship to tumor immunity and immunotherapy response prediction.

TL;DR: Pyroptosis-related gene signatures show prognostic value in multiple cancers, and this study applied TCGA bioinformatics analysis to construct the first PRG-based prognostic signature specifically for endometrial cancer.
Pages 1-2
Bioinformatics Pipeline and Signature Construction

Data source and preprocessing: Gene expression and clinical data were obtained from TCGA (515 tumor, 35 normal samples). After excluding FFPE samples, duplicate data, and patients with survival time under 30 days, the remaining EC samples were divided 3:1 into a training set (n = 387) and testing set (n = 128).

PRG identification: 133 PRGs were extracted from GeneCards and compared between tumor and normal tissues using the DESeq2 package in R. 97 differentially expressed PRGs were identified (all p less than 0.05), including 64 upregulated and 33 downregulated genes. Protein-protein interaction networks were constructed via the STRING database.

Signature development steps: (1) Univariate Cox regression identified 19 prognostic-associated PRGs; (2) LASSO regression reduced this to 14 candidates; (3) Multivariate Cox regression produced the final 7-gene signature: NFKB1, EEF2K, CTSV, MDM2, GZMB, PANX1, and PTEN.

Risk score formula: RS = (-0.623 x NFKB1) + (-0.231 x CTSV) + (0.335 x PANX1) + (-0.322 x PTEN) + (-0.359 x MDM2) + (-0.219 x GZMB) + (1.031 x EEF2K). Patients were classified as high-risk or low-risk based on median risk score.

TL;DR: Using TCGA data and a multi-step bioinformatics pipeline (DESeq2, LASSO, Cox regression), researchers identified a 7-gene pyroptosis-related signature (NFKB1, EEF2K, CTSV, MDM2, GZMB, PANX1, PTEN) for EC prognosis prediction.
Pages 5-7
Prognostic Performance of the 7-Gene Signature

Training set performance: Among 387 training set patients divided by median risk score, the high-risk group had significantly more deaths and shorter survival (p less than 0.0001). ROC AUC values for predicting survival at 1, 3, and 5 years were 0.732, 0.763, and 0.793, respectively.

Testing set validation: Among 128 testing set patients, the high-risk group again showed significantly worse survival (p = 0.0015). Test set AUC values were 0.92 (1-year), 0.789 (3-year), and 0.739 (5-year), demonstrating strong predictive performance. Principal component analysis confirmed clear separation between risk groups.

Clinical correlation: Risk scores were significantly associated with age (p = 0.0015), histology (p less than 2.22e-16), and disease stage (p = 0.00024), but not tumor grade. High-risk status predicted worse OS across multiple clinical subgroups including age groups, FIGO stages I-II and III-IV, and grades 1, 2, and 3.

Independent prognostic value: Multivariate Cox analysis confirmed that PRG risk score, age, and stage independently predict EC prognosis (all p less than 0.001). The risk score AUC (0.716) outperformed age (0.602), histology (0.556), weight (0.556), grade (0.537), and approached stage (0.708).

TL;DR: The 7-gene PRG signature demonstrated strong and consistent prognostic performance across training and testing sets, independently predicting EC survival with AUCs reaching 0.92 in the test cohort.
Page 8
Nomogram Integration for Clinical Use

Nomogram construction: A comprehensive prognostic nomogram was built incorporating six parameters: age, weight, histology, grade, stage, and PRG risk score. This tool enables patient-specific survival probability estimates for 1-, 3-, and 5-year time points.

Calibration accuracy: Calibration plots showed excellent agreement between nomogram-predicted and actual observed 1-year, 3-year, and 5-year survival rates in the entire TCGA cohort, confirming the practical utility of integrating the PRG signature into a composite prediction tool.

TL;DR: The PRG risk score was integrated with clinical variables into a nomogram that provides well-calibrated patient-level survival probability estimates across 1-, 3-, and 5-year time horizons.
Pages 9, 13, 14
Tumor Immunity and Immunotherapy Implications

Immune cell infiltration differences: CIBERSORT analysis revealed that the high-risk group had significantly higher proportions of naive B cells, resting CD4 memory T cells, and activated NK and dendritic cells, while showing lower proportions of plasma cells, CD8+ T cells, regulatory T cells (Tregs), and resting NK and dendritic cells.

Immune pathway enrichment: ssGSEA analysis showed significantly higher enrichment of APC co-inhibition, checkpoint, HLA, and T cell co-inhibition pathways in the high-risk group. Higher proportions of dendritic cells and neutrophils were also observed, suggesting an immunosuppressive microenvironment associated with poor prognosis.

Immune checkpoint gene expression: CTLA4, PD1, and PDL1 were all significantly decreased in the high-risk group (CTLA4: p = 1.4e-14; PD1: p = 1.4e-14; PDL1: p = 0.00075). Risk score was negatively correlated with CTLA4 (r = -0.4), PD1 (r = -0.27), and PDL1 (r = -0.21), suggesting the PRG signature may guide immunotherapy response prediction in EC.

TL;DR: High-risk PRG patients show reduced immune checkpoint gene expression and altered immune cell composition, suggesting impaired anti-tumor immunity and potential utility of the signature for predicting immunotherapy response.
Pages 12-13
Biological Roles of the Seven Signature Genes

NFKB1 and CTSV: NFKB1, a subunit of the NF-kB pathway, acts as a pathway-specific suppressor of inflammation and tumor development; upregulation promotes breast cancer cell invasiveness. CTSV (cathepsin V) is expressed in activated macrophages and associated with poor prognosis and increased invasion in breast and colorectal cancer.

PANX1 and PTEN: PANX1 (pannexin 1) is an ATP-releasing channel that regulates tumor immune microenvironment through interactions with cancer-associated fibroblasts, macrophages, and lymphocytes; its overexpression correlates with EMT and poor outcomes in breast and pancreatic cancers. PTEN is a well-established tumor suppressor operating through the PI3K/AKT pathway, and its inactivation is a key event in early EC tumorigenesis.

MDM2, GZMB, and EEF2K: MDM2 is a negative regulator of tumor suppressor p53 that is abnormally upregulated in multiple tumor types; in EC, characteristic low MDM2 with high p53 marks low-differentiated tumors. GZMB (granzyme B) is a cytotoxic T lymphocyte and NK cell component associated with poor outcomes in lung and colorectal cancers. EEF2K (eukaryotic elongation factor 2 kinase) is an atypical kinase overexpressed in multiple cancers that promotes cancer cell survival under nutrient deprivation.

Clinical significance: Together, these seven genes represent an interconnected network of inflammatory, immune, and tumor suppressor pathways whose combined expression profile captures the molecular complexity of EC prognosis more effectively than any single biomarker or conventional clinical variable alone.

TL;DR: The seven PRG signature genes collectively represent NF-kB inflammatory signaling, tumor immune microenvironment regulation (PANX1, GZMB), classical tumor suppression (PTEN), p53 regulation (MDM2), and translational stress response (EEF2K) - providing broad molecular coverage of EC disease biology.
Citation: Open Access, 2022. Available at: PMC9806687.