Abstract
Chromosomal replication and cell division are inherently genotoxic processes, especially in rapidly proliferating cells like ES cell lines and expanding hematopoietic stem/progenitors. Metazoan animals have a very efficient system to ensure genome fidelity. It is comprised of two interdependent branches. First are cell cycle checkpoints that detect genetic damage and delay cell cycle progression to allow time for repair. Second is apoptosis, an orderly elimination of cells damaged beyond the capability to be repaired. Defects in either of these branches contribute to spontaneous tumorigenesis and genetic instability. Evidence exists in various species that checkpoint signals are transient. Down-regulation of these cell cycle delaying signals occurs after repair, called “recovery”, or without repair, called “adaptation”. However, there is no clear evidence of checkpoint adaptation reported in mammals. Mouse ES cell lines (mES) are highly unstable, genetically and epigenetically, but the mechanism of this genetic infidelity is unknown. It is controversial whether human ES cell lines are likewise unstable, but an understanding of the mechanisms of instability in mouse model mES could be useful. Here, for the first time, we provide evidence of checkpoint adaptation in genomically stressed mES. There is failure of the checkpoint adaptation branch of genome surveillance to activate the apoptosis branch, resulting in aberrant cell cycle progression and polyploidy/aneuploidy. Treatment of E14, R1, CCE, or JSR mES with nocodazole, taxol, or etoposide induced polyploidization (8N DNA content) in as much as 50% of cells measured by bivariate flow analysis of phosphorylation of histone H3 and DNA content. This was accompanied by a low level (12–15%) of apoptosis measured by intracellular activated caspase-3. Moreover, caspase-3 was not activated in polyploid cells. This situation is reversed during embryoid body formation and differentiation to various lineages including primitive hematopoietic cells. Etoposide treatment resulted in nearly total cell death. Importantly, no polyploidization occurred, and the cells apoptose from 4N (not 8N). A screen of activation state (site-specific phosphorylation) of 10 DNA-damage checkpoint-relevant signaling intermediates (including pCHK1, pCDC25c, pCDC2, p53) suggested that all were phosphorylated by etoposide, indicating that damage was detected and the cell cycle was suspended as in somatic cells. Thus the checkpoint branch may not be responsible for apoptosis failure. This pattern of checkpoint adaptation, failure to initiate apoptosis, and polyploidization is mimicked in highly differentiated, pre-B lymphocyte cell line, Ba/F3, by suppressing apoptosis via Survivin or Anamorsin overexpression. Overexpression of Survivin resulted in 4-fold decreased etoposide-induced apoptosis concomitant with 4-fold increased polyploidy. Taxol caused similar results. We suggest that undifferentiated mES are insensitive to genotoxic-stress-induced cell death because of checkpoint adaptation without apoptosis as the endpoint. This could be due to uncoupling of the two branches of the genome-surveillance system. Uncoupling could be a mechanism for spontaneous genetic instability in mES and may have implications for human ES cell lines, cancer-linked genetic instability, and ex-vivo expanded HSCs.
Generation of mouse ES cell lines engineered for the forced induction of transcription factors