Cancer Medicine (Jul 2021)

A five‐gene signature is a prognostic biomarker in pan‐cancer and related with immunologically associated extracellular matrix

  • Chunlai Yu,
  • Mingliang You,
  • Peizhen Zhang,
  • Sheng Zhang,
  • Yuzhu Yin,
  • Xiao Zhang

DOI
https://doi.org/10.1002/cam4.3986
Journal volume & issue
Vol. 10, no. 13
pp. 4629 – 4643

Abstract

Read online

Abstract The tumor microenvironment (TME) is related to extracellular matrix (ECM) dynamics and has a broad fundamental and mechanistic role in tumorigenesis and cancer progression. We hypothesized that ECM regulators might play an essential role in pan‐cancer attribution by causing a generic effect through its regulation of the dynamics of ECM alteration. By analyzing data from TCGA using GSEA and univariate Cox regression analysis, we found that ECM regulator genes were significantly enriched and contributed to mortality in various cancer types. Notably, UMAP analysis revealed that ECM regulator genes dominated the differences between tumor and adjacent normal tissues based on 59 or 31 pan‐survival‐related ECM gene sets. Subsequently, a five‐gene signature consisting of the predominant ECM regulators ADAM12, MMP1, SERPINE1, PLOD3, and P4HA3 was identified. We found that this five‐gene signature was pro‐mortality in 18 types of cancer in TCGA, and validated eleven other cancer types in TCGA and seven types in the TARGET and CoMMpass databases using overall survival analysis. KEGG pathway enrichment and Pearson correlation analysis indicated that these five component genes that were correlated with specific ECM proteins involved in tumorigenesis from the ECM receptor interaction gene set. Additionally, the fitted results of a linear model were applied to strengthen the discovery, demonstrating that the five genes were correlated with immune infiltration score and especially associated with typically immunologically “cold” tumors. We thus conclude that the ADAM12, MMP1, SERPINE1, PLOD3, and P4HA3 signature showed a close association with a pan‐cancer effect on prognosis and is related to ECM proteins in the TME which corresponding with immunologically “cold” cancer types.

Keywords