The landscape of enhancer RNA identify prognosis-related molecular subtypes in gastric cancer

Abstract

Background: Enhancer RNAs (eRNAs), the transcriptional products of active enhancers, are of great significance in the initial progression of cancers. However, the biological function and bioinformatics profiles of eRNA in gastric cancer remains largely enigmatic.

Methods: Firstly, STAD were clustered into three subtypes with the data of eRNA expression from TCeA. Then we explored the difference of the tumor immune microenvironment, transcription levels, and transcription regulation among the three clusters. Finally, samples collected from 12 patients diagnosed with STAD were used to conduct qRT-PCR, verifying the conclusion based on network database.

Results: The three clusters were detected to have different tumor microenvironments: Cluster A has an immune "cold" microenvironment. While cluster B features as more infiltration of immune cells, accompanied with higher expression of immune checkpoints such as PDCD1, LAG3, and TIGIT. Besides, Cluster C shows a higher stromal feature with B lineage, neutrophils, and fibroblasts. Further analyses indicated that CpG island methylation level of Cluster B is different from the other two clusters. Meanwhile, Cluster A and B showed significant enrichment of TP53 and KRAS mutation respectively while Cluster C has higher tumor mutation burden (TMB) and microsatellite instability (MSI). With the elaboration of transcriptional regulation of epigenetic clustering, we detected that Cluster A enriched in epithelial phenotype pathways. Cluster B enriched in cell-cell adhesion. Cluster C enriched in fibroblast proliferation. The clinical cohort show that Cluster B patients have lower interstitial cell characteristics and CAF infiltration.

Conclusion: We identified three unique epigenetic clusters of STAD through the differential activation of super-enhancers, and identified Cluster B with a higher immune infiltrating and a better prognosis, which provides a novel understanding of eRNAs and potential clinical applicability of eRNA-based molecular subtypes in gastric cancer.

Keywords: eRNAs; gastric cancer; immune microenvironment; methylation; mutation.

FIGURE 1

(A) The Delta area of several clustering numbers via un‐consensus clustering method. (B) The consensus matrix of gene expression when clustering number is 3. (C) The Kaplan–Meier survival array of three epigenetic clusters. (D) The differently expression analysis of eRNA in three epigenetic cluster.

FIGURE 2

(A–C) The immune score, stromal score, and tumor purity of three epigenetic clusters via ESTIMATE algorithm. (D) The tumor environment cells infiltration of three epigenetic clusters via CIBERSORTS algorithm. (E) The expression of immune checkpoints of three epigenetic clusters. (F) The landscape of 450 K DNA methylation panel array of three epigenetic clusters via Champ algorithm.

FIGURE 3

(A) The landscape of genetic mutation events of three epigenetic clusters. (B) The mutation classification and variant type of all TCGA‐STAD patients. (C) The percentage of TP53 and KRAS mutation patients in three epigenetic clusters. (D, E) The tumor mutation burden (TMB) and Microsatellite instability (MSI) of three epigenetic clusters.

FIGURE 4

(A) The Schematic of epigenetic regulation of distant RNA by enhancers. The heatmap of differentially expressed genes among the three clusters. (B, C) The IGV screenshot of H3K27ac and EP300 ChIP‐seq in gastric tissue from ENCODE dataset.

FIGURE 5

(A–C) The pathway enrichment analysis of three epigenetic clusters via Metascape.

FIGURE 6

(A) The LASSO algorithm to constructed 18‐gene epigenetic clusters signature. (B) 18‐gene epigenetic clusters signature predictive ability to distinguish clusters. (C–D) The survival analysis of patients according to 18‐gene epigenetic clusters signature from GSE51105 and GSE62254. (E) Representative picture of immunohistochemistry of E‐cad, N‐cad, and α‐SMA. (F) The Schematic of STAD patients' phenotypic differences due to epigenetic heterogeneity.