Pan-cancer analysis identifies telomerase-associated signatures and cancer subtypes

Cancer cells become immortalized through telomere maintenance mechanisms, such as telomerase reverse transcriptase (TERT) activation. In addition to maintaining telomere length, TERT activates manifold cell survival signaling pathways. However, telomerase-associated gene signatures in cancer remain elusive.

We performed a systematic analysis of TERT high (TERT^high) and low (TERT^low) cancers using multidimensional data from The Cancer Genome Atlas (TCGA). Multidimensional data were analyzed by propensity score matching weight algorithm. Coexpression networks were constructed by weight gene coexpression network analysis (WGCNA). Random forest classifiers were generated to identify cancer subtypes.

The TERT^high-specific mRNA expression signature is associated with cell cycle-related coexpression modules across cancer types. Experimental screening of hub genes in the cell cycle module suggested TPX2 and EXO1 as potential regulators of telomerase activity and cell survival. MiRNA analysis revealed that the TERT^high-specific miR-17-92 cluster can target biological processes enriched in TERT^low cancer and that its expression is negatively correlated with the tumor/normal telomere length ratio. Intriguingly, TERT^high cancers tend to have mutations in extracellular matrix organization genes and amplify MAPK signaling. By mining the clinical actionable gene database, we uncovered a number of TERT^high-specific somatic mutations, amplifications and high expression genes containing therapeutic targets. Finally, a random forest classifier integrating telomerase-associated multi-omics signatures identifies two cancer subtypes showed profound differences in telomerase activity and patient survival.

In summary, our results depict a telomerase-associated molecular landscape in cancers and provide therapeutic opportunities for cancer treatment.