The Matrisome Is Associated with Metabolic Reprograming in Stem-like Phenotypes of Gastric Cancer

Abstract

The extracellular matrix (ECM) is an important regulator of all cellular functions, and the matrisome represents a major component of the tumor microenvironment. The matrisome is an essential component comprising genes encoding ECM glycoproteins, collagens, and proteoglycans; however, its role in cancer progression and the development of stem-like molecular subtypes in gastric cancer is unknown. We analyzed gastric cancer data from five molecular subtypes (n = 497) and found that metabolic reprograming differs based on the state of the matrisome. Approximately 95% of stem-like cancer type samples of gastric cancer were in the high-matrisome category, and energy metabolism was considerably increased in the high-matrisome group. Particularly, high glycosaminoglycan biosynthesis-chondroitin sulfate metabolic reprograming was associated with an unfavorable prognosis. Glycosaminoglycan biosynthesis-chondroitin sulfate metabolic reprograming may occur according to the matrisome status and contribute to the development of stem-like phenotypes. Our analysis suggests the possibility of precision medicine for anticancer therapies.

Keywords: epithelial-mesenchymal transition; extracellular matrix; glycosaminoglycan biosynthesis-chondroitin sulfate; matrisome; stem-like gastric cancer.

Figure 1

The landscape of metabolic reprograming between high-matrisome and low-matrisome groups. (a) Box plot showing seven metabolic signatures between high-matrisome and low-matrisome in the Y497 cohort. (b) Bar graph showing matrisome state across five molecular subtypes. (c) Heat map showing activity of 83 metabolic pathways in high-matrisome and low-matrisome groups. (d) Heat map showing 22 cell types in high-matrisome and low-matrisome groups. (e) Heat map showing Pearson correlation between top enriched metabolic pathways and top ranked cell types. (f) Bar graph of Pearson correlation between matrisome values and metabolic pathway gene expression levels. (g) Bar graph of Pearson correlation between matrisome values enriched in various cell types.

Figure 2

The extracellular matrix modulates the hallmarks of cancer. (a) Heat map of 50 cancer hallmarks in the high-matrisome group and low-matrisome group in 4 gastric cancer cohorts. (b) Heat map of telomere maintenance mechanism in the high- and low-matrisome groups in 4 gastric cohorts. (c) Boxplot of mutation burden in high- and low-matrisome groups in STAD (left); boxplot of copy alternation number in high- and low-matrisome groups in the STAD cohort (right). (d) Network of biological pathways of integrated energy metabolism with elevated gene expressions enriched in high-matrisome samples (4 cohorts). (e) Heat map of similarity for Pearson correlation between glucagon-like peptide 1-regulated insulin secretion genes and glycosaminoglycan biosynthesis-chondroitin sulfate genes in cohort Y497. (f) Boxplot of overexpressed immune checkpoint target genes in high-matrisome samples in cohort Y497. (g) Heat map of Pearson correlation between highly expressed immune checkpoint target genes and highly expressed glycosaminoglycan biosynthesis-chondroitin sulfate genes in cohort Y497. (h) Boxplot of overexpressed oncogenes/tumor suppressor genes in high- and low-matrisome groups. (i) Heat map of Pearson correlation between highly expressed glycosaminoglycan biosynthesis-chondroitin sulfate genes and highly expressed oncogenes/tumor suppressor genes in cohort Y497. STAD, The Cancer Genome Atlas stomach adenocarcinoma.

Figure 3

High-matrisome type is associated with poor prognosis in gastric cancer. (a) Kaplan–Meier plots show the overall survival rates for the high- and low-matrisome groups. The p-values were analyzed using the log-rank test and adjusted by Bonferroni correction. (b) Bar graph showing the tumor stages for the high- and low-matrisome groups. (c) Network of gene ontology analysis of high-matrisome groups of 4 cohorts (FDR < 0.001). (d) Network of transcription factor target genes and transcription factors in the high-matrisome group in cohort Y497. (e) Network of transcription factor target genes and transcription factors in the high-matrisome group in STAD cohort. (f) Network of transcription factor target genes and transcription factors in the high-matrisome group in cohort GSE15459. (g) Network of transcription factor target genes and transcription factors in the high-matrisome group in cohort GSE62254. STAD, The Cancer Genome Atlas stomach adenocarcinoma.

Figure 4

Therapeutic targets in the high-matrisome group in gastric cancer. (a) Heat map of Pearson correlation between high expression of glycosaminoglycan biosynthesis-chondroitin sulfate genes and that of 22q11.2 copy number variation pathway genes in cohort Y497. (b) Heat map of Pearson correlation between high expression of glycosaminoglycan biosynthesis-chondroitin sulfate genes and that of angiogenesis development genes in cohort STAD. (c) Network of protein–protein interaction in glycosaminoglycan biosynthesis-chondroitin sulfate in cohort Y497. (d) Predicted drugs from Genomics of Drug Sensitivity in Cancer database for CHST7, CHSY3, and DSE. (e) Kaplan–Meier plots for glycosaminoglycan biosynthesis-chondroitin sulfate gene signatures (left) and CHST7 expression in Y497. STAD, The Cancer Genome Atlas stomach adenocarcinoma.