BMC Medicine (Feb 2021)

Glycolysis-associated lncRNAs identify a subgroup of cancer patients with poor prognoses and a high-infiltration immune microenvironment

  • Kuo-Hao Ho,
  • Tzu-Wen Huang,
  • Chwen-Ming Shih,
  • Yi-Ting Lee,
  • Ann-Jeng Liu,
  • Peng-Hsu Chen,
  • Ku-Chung Chen

DOI
https://doi.org/10.1186/s12916-021-01925-6
Journal volume & issue
Vol. 19, no. 1
pp. 1 – 17

Abstract

Read online

Abstract Background Long noncoding (lnc)RNAs and glycolysis are both recognized as key regulators of cancers. Some lncRNAs are also reportedly involved in regulating glycolysis metabolism. However, glycolysis-associated lncRNA signatures and their clinical relevance in cancers remain unclear. We investigated the roles of glycolysis-associated lncRNAs in cancers. Methods Glycolysis scores and glycolysis-associated lncRNA signatures were established using a single-sample gene set enrichment analysis (GSEA) of The Cancer Genome Atlas pan-cancer data. Consensus clustering assays and genomic classifiers were used to stratify patient subtypes and for validation. Fisher’s exact test was performed to investigate genomic mutations and molecular subtypes. A differentially expressed gene analysis, with GSEA, transcription factor (TF) activity scoring, cellular distributions, and immune cell infiltration, was conducted to explore the functions of glycolysis-associated lncRNAs. Results Glycolysis-associated lncRNA signatures across 33 cancer types were generated and used to stratify patients into distinct clusters. Patients in cluster 3 had high glycolysis scores and poor survival, especially in bladder carcinoma, low-grade gliomas, mesotheliomas, pancreatic adenocarcinomas, and uveal melanomas. The clinical significance of lncRNA-defined groups was validated using external datasets and genomic classifiers. Gene mutations, molecular subtypes associated with poor prognoses, TFs, oncogenic signaling such as the epithelial-to-mesenchymal transition (EMT), and high immune cell infiltration demonstrated significant associations with cluster 3 patients. Furthermore, five lncRNAs, namely MIR4435-2HG, AC078846.1, AL157392.3, AP001273.1, and RAD51-AS1, exhibited significant correlations with glycolysis across the five cancers. Except MIR4435-2HG, the lncRNAs were distributed in nuclei. MIR4435-2HG was connected to glycolysis, EMT, and immune infiltrations in cancers. Conclusions We identified a subgroup of cancer patients stratified by glycolysis-associated lncRNAs with poor prognoses, high immune infiltration, and EMT activation, thus providing new directions for cancer therapy.

Keywords