Novel pyroptosis-associated genes signature for predicting the prognosis of sarcoma and validation

Abstract

Background: Sarcoma is a rare mesenchymal malignant tumor. Recently, pyroptosis has been reported to be a mode of programmed cell death. Nonetheless, levels of pyroptosis-associated genes in sarcoma and its relevance to prognostic outcomes are yet to be elucidated.

Results: Sarcoma cases were classified into two subtypes with regards to differentially expressed genes. We established a profile composed of seven genes and classified the sarcoma patients into low- and high-risk groups through least absolute shrinkage and selection operator Cox regression. Survival rate of low-risk sarcoma patients was markedly higher, relative to high-risk group (P<0.001). In combination with clinical features, the risk score was established to be an independent predictive factor for OS of sarcoma patients. Chemotherapeutic drug sensitivity response analysis found 65 drugs with higher drug sensitivity in low-risk, than in high-risk group and 14 drugs with higher drug sensitivity in the high-risk patient group, compared with low-risk patient group. In addition, functional enrichment, pathway and gene mutation of the two modules were analyzed. Finally, we used qRT-PCR to detect the expression of seven pyroptosis-related genes in tumor cells, and human skeletal muscle cells, compared with human skeletal muscle cells, PODXL2, LRRC17, GABRA3, SCUBE3 and RFLNB genes show high expression levels in tumor cells, while IGHG2 and hepatic leukemia factor show low expression levels in tumor cells.

Conclusions: Our research suggest that pyroptosis is closely associated with sarcoma, and these findings confirm that pyroptosis-associated seven genes have a critical role in sarcoma and are potential prognostic factors for sarcoma.

Keywords: biomarkers; pyroptosis; sacoma.

Figure 1. Tumor classification based on the pyroptosis-related DEGs

(A) Pairwise correlations among the levels of 33 pyroptosis-associated genes in sarcoma. (B) About 262 Sarcoma patients assigned into two clusters based on consensus clustering matrix (k = 2). (C) When the clustering variable (k) was increased from 2 to 10, (k = 2) is the best clustering. (D) The PCA plot for sarcoma patients, which was based on the pyroptosis-related genes. (E) Heatmap and clinic-pathologic features of the two clusters, as classified by the DEGs. (F) Kaplan–Meier OS curves of the two clusters.

Figure 2. The screening and functional enrichment analyses of the DEGs

(A) Volcano map of DEGs. (B) GO and pathway analyses of DEGs. (C) Sixty-three were established by univariate Cox Regression analyses for dimensionality reduction. (D) Error rate for the data as a function of the classification tree. (E) The importance values for 12 the predictors.

Figure 3. Creation of the risk signature in TCGA cohort

(A) Cross-validation for tuning parameter selections in LASSO regression. (B) LASSO regression of seven genes. (C–E) Risk score distribution, survival status, and expressions of seven prognostic pyroptosis-related genes in sarcoma. (F) Heatmap of connections between clinic-pathologic characteristics and seven genes. (G,H) Overall survival curves for high-/low-risk group sarcoma patients and ROC curve for evaluating the predictive value. (I) Circos plot shows the chromosomal distribution of the seven genes.

Figure 4. Univariate as well as multivariate Cox regression analyses of the risk score

(A) Univariate and Multivariate analysis for the TCGA cohort. (B) Heatmap of connections between clinic-pathologic characteristics and risk groups.

Figure 5. Chemotherapeutic response for risk models

High-risk group patients were sensitive to chemotherapy with cisplatin, docetaxel, doxorubicin and gemcitabine.

Figure 6. Functional analysis of risk models

(A) GO enrichment of the risk models. (B) KEGG pathway enrichment of the risk models. (C) Immune response analysis of risk models. (D) Mutation analysis of risk models.

Figure 7. Validation of seven genes expression in RH30, SW982 and HSKMC by qRT-PCR

Bars represent mean ± SEM (n=3), *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.