Quantifying replication stress in cancer without proliferation confounding

Abstract

Replication stress (RS) is a major driver of genomic instability and cancer development, caused by impaired DNA replication that can lead to chromosomal instability (CIN). Although RS is mechanistically linked to CIN, its relationship with cellular proliferation is complex. Depending on the context, RS can either promote or suppress cell growth. Existing RS gene expression signatures overlook this complexity, relying on the overexpression of oncogenes such as MYC, which introduces a proliferation bias. To disentangle genuine RS from confounding cell cycle and proliferation transcription profiles, we developed and validated a novel gene expression signature that accurately predicts RS independently of oncogene activity. This signature captures RS-related transcriptional changes across diverse cellular contexts, enabling a more robust and proliferation-independent measure of RS in both experimental and clinical samples. Applying our signature to patient data, we discovered a link between RS and the non-homologous end-joining (NHEJ) DNA repair pathway. Specifically, we observed that MSH2 and MSH6--core components of mismatch repair--are associated with elevated RS and may indicate a shift toward NHEJ-mediated repair under stress conditions. Our study provides a refined approach to quantify RS and sheds light on its broader impact on DNA repair network dynamics. Author summaryCancer cells often display defects in faithful DNA replication, which can result in permanent chromosomal alterations during cell division. One major cause of such damage is underreplicated DNA resulting from replication stress (RS). RS describes the slowing or stalling of DNA replication due to factors like a premature start of replication, before necessary replication factors are assembled. This makes RS a key driver in cancer development. Current methods quantifying RS rely heavily on overexpression of oncogenes that deregulate the cell cycle leading to an increase in cell growth and confounding the identification of RS. To address this, we carefully selected samples displaying RS from diverse sources and identified genes biologically and statistically linked to RS, independent of cell growth signals. Using this new signature, we found a link between RS and a low-fidelity DNA repair mechanism. By improving RS quantification, our work could provide new insights into how cells respond to replication stress and therefore lead to improved methods for diagnosing and treating cancers linked to RS.