Background: Cellular plasticity, the capacity of cells to transition between states, is pivotal in cancer evolution. This fundamental process is observed not only in cancer but also during development, repair, and maintenance of equilibrium. This study draws inspiration from the exceptional cellular plasticity exhibited during pre-implantation (PIE) embryo development, marking the pinnacle of plasticity during totipotency establishment. In early developmental stages, transcriptional control depends on maternal factors, but at the four-to-eight-cell stage, the zygotic genome starts actively transcribing. To detect genome reactivation akin to preimplantation states, we present PIECanceR, a machine-learning model to identify plasticity reactivation using transcriptomic data.
Aim: To explore the significance of transcriptional signatures associated with early embryonic plasticity in cancer.
Methods: A single-cell RNA sequencing atlas of early human embryos was assembled from pre-existing datasets. We selected genes with the highest variability in pan-cancer datasets and delineated their expression pattern during early embryonic stages as training features. A k-Top Scoring Pairs (kTSP) model was trained to classify cellular states along the developmental trajectory. Multiple datasets, including Genomic Data Commons (GDC), Genotype-Tissue Expression (GTEx), and the Cancer Cell Line Encyclopedia (CCLE), were then classified using PIECanceR to identify relevant patient samples, tissues, and cell lines enriched for PIE plasticity.
Results: We have built a machine-learning model, PIECanceR, that successfully classifies early embryonic stages. When applied to cancer samples, PIECanceR predicted elevated plasticity in a subset of patients, with significant differences in overall survival benefits.
Conclusions: PIECanceR translates developmental plasticity into cancer plasticity. This tool enables the discovery of reactivation of preimplantation plasticity from transcriptomic data in oncogenic settings. Investigation into the mechanisms and significance of ZGA in the context of cancer cells is ongoing.