In individual mammalian cells, DNA double-strand breaks trigger the tumor suppressor p53 and its antagonist Mdm2 to oscillate in discrete pulses of roughly 5.5-hour period with a roughly 2-hour delay between p53 and Mdm2 peaks. The pulses have nearly fixed amplitude and duration, yet the number of pulses varies from cell to cell, amplitudes fluctuate from peak to peak by about 70%, and periods by about 20% - variability that is functional, since it shapes cell-fate decisions and fractional killing (the death of only part of a genetically identical cell population) after chemotherapy. Later work largely established how the pulses are generated (repeated re-initiation by the upstream kinase ATM in an excitable network), but the quantitative question posed by the modeling literature of 2000-2015 remains open: which identifiable sources - intrinsic biochemical noise, extrinsic slowly varying (colored) noise, cell-to-cell heterogeneity in regulators such as Wip1, and stochasticity of damage and repair - produce which features of the observed variability, and how does the length of the feedback expression delay map quantitatively onto oscillation period, damping, and amplitude? The problem asks for a mechanistic and predictive account of these relationships.

Background

p53 is a transcription factor that coordinates the response to genomic stress: it activates genes for DNA repair, cell-cycle arrest, senescence (irreversible growth arrest), and apoptosis. Under resting conditions p53 protein is kept scarce because Mdm2, an E3 ubiquitin ligase (an enzyme that attaches the degradation tag ubiquitin to target proteins), binds p53 and sends it to the proteasome for degradation. p53 in turn drives transcription of the MDM2 gene, so p53 and Mdm2 form a negative feedback loop: p53 promotes its own destroyer, with a delay of roughly two hours set by Mdm2 messenger-RNA transcription and protein translation. Ionizing radiation creates DNA double-strand breaks that activate ATM, a protein kinase that phosphorylates both p53 (stabilizing it) and Mdm2 (destabilizing it); p53 also induces Wip1, a phosphatase that deactivates ATM, adding a second, slower negative feedback loop through the damage sensor.

Measurements averaged over cell populations showed damped p53 oscillations after irradiation (Lev Bar-Or et al. 2000), and single-cell time-lapse fluorescence microscopy then revealed what population averages had hidden: p53 appears as a train of discrete pulses whose amplitude and duration are nearly fixed while their number varies from cell to cell and grows on average with radiation dose (Lahav et al. 2004). Longer imaging showed sustained oscillations for up to three days, with peak-to-peak amplitude fluctuations of about 70%, period fluctuations of about 20%, and a sizeable fraction of non-oscillating cells in a clonal population - genetically identical descendants of a single cell (Geva-Zatorsky et al. 2006); population-level damping is largely an artifact of cells drifting out of phase. The variability is not meaningless noise: the number of p53 pulses influences whether cells recover, arrest, or senesce (Purvis et al. 2012), and cell-to-cell variation in p53 dynamics contributes to fractional killing - a genetically identical population being only partly killed by a given drug dose - (Paek et al. 2016).

Two broad classes of explanation were proposed for the pulse trains and their variability. Deterministic models, built from ordinary or delay differential equations (equations in which the rate of change depends on the system state at an earlier time, representing the transcriptional-translational lag), treat the p53-Mdm2 loop as a damped or self-sustained oscillator whose oscillations require a sufficiently long delay, with the delay acting as a bifurcation parameter (a parameter whose change transforms a stable steady state into a limit cycle, a closed repeating trajectory, via a Hopf bifurcation); variability is then attributed to cell-to-cell differences in parameters or to noise in upstream damage signaling (Ma et al. 2005; Wagner et al. 2005). Stochastic models, simulated with Gillespie's Monte Carlo algorithm for chemical reaction networks and analyzed with the linear noise approximation (a systematic expansion of the master equation governing molecule-number fluctuations), attribute sustained oscillations and their irregularity to intrinsic noise - the random birth and death of individual molecules - possibly amplified by the feedback loop (Proctor and Gray 2008; Cruz-Rodríguez, Figueroa-Morales and Mulet 2015). The stochastic picture is motivated by an apparent conflict: deterministic versions of standard p53-Mdm2 models predict decay to a steady state, yet single cells oscillate indefinitely. A middle position attributes variability mainly to extrinsic colored noise - slowly varying, temporally correlated fluctuations in production rates or regulator abundances acting cell by cell - rather than to fast intrinsic molecular noise.

Subsequent experimental and combined experimental-modeling work substantially clarified the pulse-generation mechanism: the pulses are not a free-running p53-Mdm2 limit cycle but are driven by repeated re-initiation - persistent DNA damage reactivates ATM, whose activity itself pulses, and the p53-Wip1-ATM feedback keeps individual p53 pulses uniform in shape (Batchelor et al. 2008); the circuit behaves as an excitable system (one that rests at a steady state but emits an all-or-none pulse when an input crosses a threshold), with the ATM switch supplying positive feedback and cell-to-cell differences in Wip1 setting cell-specific activation thresholds (Loewer et al. 2010; Batchelor et al. 2011; Mönke et al. 2017). Recent work recast the core loop as a damped, resonant oscillator - a single p53-stabilizing drug pulse yields damped p53 oscillations, and periodic stimulation produces frequency-dependent amplitude responses maximal near the natural frequency (Heltberg et al. 2026) - and measured its entrainment - the locking of its period to that of an applied periodic input - including phase response curves (the phase shift caused by a brief perturbation as a function of its timing) and Arnold tongues (regions of locking between forcing period and oscillator period; Jiménez et al. 2024). What has not been settled is the quantitative question the modeling literature originally posed: a validated decomposition of the observed variability into mechanistic sources, and a validated quantitative relationship between feedback-delay length and oscillation features.

Problem Statement

Provide a quantitative, mechanistic account of the variability and timing of the p53-Mdm2 oscillatory response to DNA damage in single cells. Specifically: (1) decompose the observed variability - peak-to-peak amplitude fluctuations of roughly 70%, period fluctuations of roughly 20%, and cell-to-cell variation in pulse number, including the fraction of non-oscillating cells in a clonal population - into identifiable mechanistic sources: intrinsic biochemical noise (finite-molecule fluctuations in p53 and Mdm2 synthesis and degradation), extrinsic slowly varying (colored) noise such as low-frequency production-rate fluctuations, cell-to-cell heterogeneity in regulator abundances (for example Wip1, ATM, Mdm2, or Mdmx, a p53-binding protein related to Mdm2), stochasticity of damage generation and repair, and cell-cycle state; and (2) establish the quantitative relationship between the length of the feedback expression delay (the p53-to-Mdm2 lag of roughly 2 hours, reflecting Mdm2 transcription and translation) and the features of the oscillation it produces - period, damping rate, and amplitude - in the biological circuit itself. The account must attribute the observed amplitude, period, and pulse-number statistics to variability sources that are separately identifiable or manipulable, and predict how those statistics change under manipulation of the identified sources. Both sub-questions (1) and (2) are jointly posed. The problem includes deciding among the competing published attributions of variability (intrinsic noise, extrinsic colored noise, cellular heterogeneity, and stochastic re-initiation timing), which remain unsettled. The mechanism of pulse generation itself (repeated ATM-driven initiation in an excitable network) is established background, not part of what must be solved; whether stochastic re-initiation timing contributes to the observed variability remains in scope.

References

  1. Lev Bar-Or R, Maya R, Segel LA, Alon U, Levine AJ, Oren M (2000). Generation of oscillations by the p53-Mdm2 feedback loop: a theoretical and experimental study. Proceedings of the National Academy of Sciences USA. doi:10.1073/pnas.210171597, https://doi.org/10.1073/pnas.210171597
  2. Lahav G., Rosenfeld N, Sigal A, Geva-Zatorsky N, Levine AJ, Elowitz MB, Alon U (2004). Dynamics of the p53-Mdm2 feedback loop in individual cells. Nature Genetics. doi:10.1038/ng1293, https://doi.org/10.1038/ng1293
  3. Ma L, Wagner J, Rice JJ, Hu W, Levine AJ, Stolovitzky GA (2005). A plausible model for the digital response of p53 to DNA damage. Proceedings of the National Academy of Sciences USA. doi:10.1073/pnas.0501352102, https://doi.org/10.1073/pnas.0501352102
  4. Wagner J, Ma L, Rice JJ, Hu W, Levine AJ, Stolovitzky GA (2005). p53-Mdm2 loop controlled by a balance of its feedback strength and effective dampening using ATM and delayed feedback. IEE Proceedings - Systems Biology. doi:10.1049/ip-syb:20050025, https://doi.org/10.1049/ip-syb:20050025
  5. Geva-Zatorsky N, Rosenfeld N, Itzkovitz S, Milo R, Sigal A, Dekel E, Yarnitzky T, Liron Y, Polak P, Lahav G, Alon U (2006). Oscillations and variability in the p53 system. Molecular Systems Biology. doi:10.1038/msb4100068, https://doi.org/10.1038/msb4100068
  6. Proctor CJ, Gray DA (2008). Explaining oscillations and variability in the p53-Mdm2 system. BMC Systems Biology 2:75. doi:10.1186/1752-0509-2-75, https://doi.org/10.1186/1752-0509-2-75
  7. Batchelor E, Mock CS, Bhan I, Loewer A, Lahav G (2008). Recurrent initiation: a mechanism for triggering p53 pulses in response to DNA damage. Molecular Cell. doi:10.1016/j.molcel.2008.03.016, https://doi.org/10.1016/j.molcel.2008.03.016
  8. Loewer A, Batchelor E, Gaglia G, Lahav G (2010). Basal dynamics of p53 reveal transcriptionally attenuated pulses in cycling cells. Cell. doi:10.1016/j.cell.2010.05.031, https://doi.org/10.1016/j.cell.2010.05.031
  9. Cruz-Rodríguez L, Figueroa-Morales N, Mulet R (2015). On the role of intrinsic noise on the response of the p53-Mdm2 module. Journal of Statistical Mechanics: Theory and Experiment P09015 (preprint arXiv:1211.5073). doi:10.1088/1742-5468/2015/09/p09015, https://doi.org/10.1088/1742-5468/2015/09/p09015
  10. Batchelor E, Loewer A, Mock C, Lahav G (2011). Stimulus-dependent dynamics of p53 in single cells. Molecular Systems Biology. doi:10.1038/msb.2011.20, https://doi.org/10.1038/msb.2011.20
  11. Purvis JE, Karhohs KW, Mock C, Batchelor E, Loewer A, Lahav G (2012). p53 dynamics control cell fate. Science. doi:10.1126/science.1218351, https://doi.org/10.1126/science.1218351
  12. Paek AL, Liu JC, Loewer A, Forrester WC, Lahav G (2016). Cell-to-cell variation in p53 dynamics leads to fractional killing. Cell. doi:10.1016/j.cell.2016.03.025, https://doi.org/10.1016/j.cell.2016.03.025
  13. Mönke G, Cristiano E, Finzel A, Friedrich D, Herzel H, Falcke M, Loewer A (2017). Excitability in the p53 network mediates robust signaling with tunable activation thresholds in single cells. Scientific Reports 7:46571. doi:10.1038/srep46571, https://doi.org/10.1038/srep46571
  14. Wang C, Liu H, Zhou J (2019). Contribution of time delays to p53 oscillation in DNA damage response. IET Systems Biology. doi:10.1049/iet-syb.2019.0006, https://doi.org/10.1049/iet-syb.2019.0006
  15. Wang DG, Wang S, Huang B, Liu F (2019). Roles of cellular heterogeneity, intrinsic and extrinsic noise in variability of p53 oscillation. Scientific Reports 9:5883. doi:10.1038/s41598-019-41904-9, https://doi.org/10.1038/s41598-019-41904-9
  16. Eliaš J, Macnamara CK (2021). Mathematical modelling of p53 signalling during DNA damage response: a survey. International Journal of Molecular Sciences 22:10590. doi:10.3390/ijms221910590, https://doi.org/10.3390/ijms221910590
  17. Yang Z, Hanson RL, Batchelor E (2021). Progress and challenges in understanding the regulation and function of p53 dynamics. Biochemical Society Transactions 49:2123-2131. doi:10.1042/bst20210148, https://doi.org/10.1042/bst20210148
  18. Jiménez A, Lu D, Kalocsay M, Berberich MJ, Balbi P, Jambhekar A, Lahav G (2022). Time-series transcriptomics and proteomics reveal alternative modes to decode p53 oscillations. Molecular Systems Biology. doi:10.15252/msb.202110588, https://doi.org/10.15252/msb.202110588
  19. Gutu N, Binish N, Keilholz U, Herzel H, Granada AE (2023). p53 and p21 dynamics encode single-cell DNA damage levels, fine-tuning proliferation and shaping population heterogeneity. Communications Biology. doi:10.1038/s42003-023-05585-5, https://doi.org/10.1038/s42003-023-05585-5
  20. Wang P, Wang HY, Gao XJ, Zhu HX, Zhang XP, Liu F, Wang W (2023). Encoding and decoding of p53 dynamics in cellular response to stresses. Cells 12:490. doi:10.3390/cells12030490, https://doi.org/10.3390/cells12030490
  21. Hsieh FS, Nguyen DPM, Heltberg MS, Wu CC, Lee YC, Jensen MH, Chen SH (2024). Plausible, robust biological oscillations through allelic buffering. Cell Systems 15:1018-1032. doi:10.1016/j.cels.2024.10.002, https://doi.org/10.1016/j.cels.2024.10.002
  22. Jiménez A, Lucchetti A, Heltberg MS, Moretto L, Sanchez C, Jambhekar A, Jensen MH, Lahav G (2024). Entrainment and multi-stability of the p53 oscillator in human cells. Cell Systems. doi:10.1016/j.cels.2024.09.001, https://doi.org/10.1016/j.cels.2024.09.001
  23. Venkatachalapathy H, Dallon S, Yang Z, Azarin SM, Sarkar CA, Batchelor E (2025). Pulsed stimuli enable p53 phase resetting to synchronize single cells and modulate cell fate. Molecular Systems Biology 21:4. doi:10.1038/s44320-025-00091-8, https://doi.org/10.1038/s44320-025-00091-8
  24. Heltberg MS, Jimenez A, Lahav G, Jensen MH (2026). Genetic resonance in the p53 signaling network. Cell Systems 17:101514. doi:10.1016/j.cels.2025.101514, https://doi.org/10.1016/j.cels.2025.101514
  25. François J, Jambhekar A, Lahav G (2026). Decoding the central dogma: quantitative insights into transcription and translation dynamics in the p53-mediated DNA damage response. Journal of Molecular Biology 438:169436. doi:10.1016/j.jmb.2025.169436, https://doi.org/10.1016/j.jmb.2025.169436