Mosaic Regulation of Stress Pathways Underlies Senescent Cell Heterogeneity
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Cellular senescence (CS) and quiescence (CQ) are stress responses characterised by persistent and reversible cell cycle arrest, respectively. These phenotypes are heterogeneous, dependent on the cell type arrested and the insult inciting arrest. Because a universal biomarker for CS has yet to be identified, combinations of senescence-associated biomarkers linked to various biological stress responses including lysosomal activity (β-galactosidase staining), inflammation (senescence-associated secretory phenotypes, SASPs), and apoptosis (senescent cell anti-apoptotic pathways) are used to identify senescent cells.
Using in vitro human bulk RNA-seq datasets, we find that senescent states enrich for various stress responses in a cell-type, temporal, and insult-dependent manner. We further demonstrate that various gene signatures used to identify senescent cells in the literature also enrich for stress responses, and are inadequate for universally and exclusively identifying senescent samples.
Genes regulating stress responses – including transcription factors and genes controlling chromatin accessibility – are contextually differentially expressed, along with key enzymes involved in metabolism across arrest phenotypes. Additionally, significant numbers of SASP proteins can be predicted from senescent cell transcriptomes and also heterogeneously enrich for various stress responses in a context-dependent manner.
We propose that ‘senescence’ cannot be meaningfully defined due to the lack of underlying preserved biology across senescent states, and CS is instead a mosaic of stress-induced phenotypes regulated by various factors, including metabolism, TFs, and chromatin accessibility. We introduce the concept of Stress Response Modules, clusters of genes modulating stress responses, and present a new model of CS and CQ induction conceptualised as the differential activation of these clusters.