Sample-based modeling reveals bidirectional interplay between cell cycle progression and extrinsic apoptosis

Dirke Imig, Nadine Pollak, Frank Allgöwer, Markus RehmPLOS Computational Biology 2020 · OpenReview

Abstract

TRAIL (TNF-related apoptosis-inducing ligand) induces cell death preferentially in cancer cells. Whether cell cycle progression notably affects extrinsic apoptosis has remained unclear, since systematic experimental and mathematical studies to quantitatively understand such interdependencies remained challenging. Here, we applied statistical and mechanistic modeling, linked with experimental analyses, to determine if cell cycle and apoptosis progression are interconnected. Using sample-based modeling, we demonstrate that times required to commit apoptotic cell death depend on the cell cycle position at the time of TRAIL exposure, with delays manifesting during S phase, around a time that can be defined as a point of apoptosis deceleration (PAD). Overall, cells receiving TRAIL in the G1 phase were more likely to die, providing scope to optimize TRAIL-based treatment strategies with respect to cell cycle dynamics.

Introduction

Programmed cell death (apoptosis) can be triggered from within the cell (intrinsic) or via external "death ligands" like TRAIL (extrinsic). While the relationship between the cell cycle and intrinsic apoptosis—often triggered by DNA damage—is well-documented, the interplay between the cell cycle and the extrinsic pathway has remained poorly understood. Biological populations exhibit significant heterogeneity; even genetically identical cells do not die at the same time or rate. This study investigates whether a cell's phase ($G_1$ vs. $S/G_2/M$) influences its sensitivity to death signals and, conversely, if death signaling interferes with the cell cycle "clock."

Conceptual overview of the bidirectional interplay between cell cycle and apoptosis.

Methods

The researchers employed a systems biology approach combining high-resolution live-cell imaging with mathematical modeling. Cancer cell lines (NCI-H460 and HCT-116) expressing Fucci (Fluorescent Ubiquitination-based Cell Cycle Indicator) reporters were used to track individual cell cycle phases in real-time following TRAIL exposure. A "sample-based modeling" framework was developed to simulate thousands of individual virtual cells, moving beyond population averages. Phenomenological models were then used to test if death signaling kinetics varied by cell cycle position, with Bayesian Information Criteria (BIC) used for model selection and parameter estimation.

Integrated workflow: from experimental data and phase length modeling to the analysis of cell fate and death timing.

Results

The study revealed a bidirectional "conversation" between the two processes. TRAIL exposure was found to prolong the cell cycle, adding an average of 2.4 hours to the $G_1$ phase and 1.3 hours to $S/G_2/M$. Simultaneously, the cell cycle regulates the speed of apoptosis. Cells in $G_1$ died significantly faster (median 2.9 hours) than those in $S/G_2/M$ (median 3.5 hours). A "Point of Apoptosis Deceleration" (PAD) was identified in mid-S phase (at 49-55% of the cycle duration), where the biochemical processes leading to death significantly slow down. This suggests that cells able to complete division before the death signal reaches a critical threshold may survive, explaining some instances of treatment resistance.

Discussion

These findings highlight $G_1$ as a period of heightened vulnerability to TRAIL-induced apoptosis. As a proof-of-concept, synchronizing cells in $G_1$ using a CDK4/6 inhibitor (Abemaciclib) significantly accelerated cell death. While the study provides a robust phenomenological framework, the specific molecular mechanisms—such as fluctuations in caspase or Bcl-2 levels—driving the deceleration at the PAD remain to be identified. Future work integrating these findings into mechanistic signaling network models will be the next step toward "precision timing" in cancer chronotherapy.

Summary updated

Abstract

Author summary TRAIL (TNF-related apoptosis-inducing ligand) induces cell death preferentially in cancer cells. Whether cell cycle progression notably affects extrinsic apoptosis has remained unclear, since systematic experimental and mathematical studies to quantitatively understand such interdependencies remained challenging. Here, we applied statistical and mechanistic modeling, linked with experimental analyses, to determine if cell cycle and apoptosis progression are interconnected. Using sample-based modeling, we demonstrate that times required to commit apoptotic cell death depend on the cell cycle position at the time of TRAIL exposure, with delays manifesting during S phase, around a time that can be defined as a point of apoptosis deceleration (PAD). Overall, cells receiving TRAIL in the G1 phase were more likely to die, providing scope to optimize TRAIL-based treatment strategies with respect to cell cycle dynamics.
BibTeX
@article{imig2020sample,
  author = {Dirke Imig and Nadine Pollak and Frank Allgöwer and Markus Rehm},
  title = {{Sample-based modeling reveals bidirectional interplay between cell cycle progression and extrinsic apoptosis}},
  journal = {PLOS Computational Biology},
  year = {2020},
  url = {https://openreview.net/forum?id=TAcNA-ObsJ},
}
Figure