Gene Expression Monotonicity across Bladder Cancer Stages Informs on the Molecular Pathogenesis and Identifies a Prognostic Eight-Gene Signature.

Abstract

Simple SummaryGene expression monotonicity is an important feature in the evolution and progression of cancer, yet has not been investigated in bladder cancer (BLCA). Most of the transcriptomic stage investigations of BLCA are limited either to a subset of stages, or to a small number of samples. Here, we leverage publicly available data to create a meta-dataset of 1135 primary BLCA transcriptomes, to identify genes and processes with a monotonal change related to higher clinical or pathologic stages. Our analysis aims to deepen the current understanding of the disease’s molecular pathogenesis, as well as to propose a prognostic gene signature based on the trait of monotonicity. Results demonstrate tumor dependencies on specific cell-cycle and metabolic microprocesses, and highlight an eight-gene signature capable of prognosing 5-year outcomes in both the discovery and validation sets.Despite advancements in molecular classification, tumor stage and grade still remain the most relevant prognosticators used by clinicians to decide on patient management. Here, we leverage publicly available data to characterize bladder cancer (BLCA)’s stage biology based on increased sample sizes, identify potential therapeutic targets, and extract putative biomarkers. A total of 1135 primary BLCA transcriptomes from 12 microarray studies were compiled in a meta-cohort and analyzed for monotonal alterations in pathway activities, gene expression, and co-expression patterns with increasing stage (Ta–T1–T2–T3–T4), starting from the non-malignant tumor-adjacent urothelium. The TCGA-2017 and IMvigor-210 RNA-Seq data were used to validate our findings. Wnt, MTORC1 signaling, and MYC activity were monotonically increased with increasing stage, while an opposite trend was detected for the catabolism of fatty acids, circadian clock genes, and the metabolism of heme. Co-expression network analysis highlighted stage- and cell-type-specific genes of potentially synergistic therapeutic value. An eight-gene signature, consisting of the genes AKAP7, ANLN, CBX7, CDC14B, ENO1, GTPBP4, MED19, and ZFP2, had independent prognostic value in both the discovery and validation sets. This novel eight-gene signature may increase the granularity of current risk-to-progression estimators.