Single-cell live imaging shows that low-frequency features of ERK (extracellular signal-regulated kinase) and Akt (protein kinase B) kinase activity time courses predict whether an individual cell will divide, and model interpretation favors a time-integrator picture in which the cumulative time a pathway is active sets division probability. However, this relationship has only been established correlationally, in a narrow set of contexts (EGF (epidermal growth factor) plus insulin stimulation of serum-starved MCF10A human mammary epithelial cells, and asynchronously cycling RPE (retinal pigment epithelium) cells observed for 48 hours). The open problem is to determine whether ERK and/or Akt dynamics causally control stochastic division probability; whether the predictive relationship and the relative importance of ERK versus Akt depend on the upstream growth factor and receptor pathway that activates them; and how the relationship generalizes across cell types, growth-factor doses, chronic stimulation conditions, and longer observation windows in which cells may divide repeatedly.

Background

ERK (extracellular signal-regulated kinase, the canonical effector of the MAPK (mitogen-activated protein kinase) pathway) and Akt (protein kinase B, the canonical effector of the PI3K (phosphoinositide 3-kinase) pathway) are the two central kinase relays transducing growth-factor signals from receptor tyrosine kinases such as the EGF (epidermal growth factor) receptor. Both can be monitored in single living cells with kinase translocation reporters (KTR): fluorescent fusion proteins that shuttle between nucleus and cytoplasm in proportion to kinase activity, so the cytoplasmic-to-nuclear fluorescence ratio gives a time-resolved readout of pathway activity in the same cell over tens of hours. Because genetically identical cells exposed to the same stimulus still show heterogeneous pathway dynamics and make apparently stochastic division decisions, the field asks which features of the dynamics, if any, encode the fate decision. Classical results established that ERK activity is pulsatile and that pulse frequency and duration correlate with proliferative versus differentiative outcomes (Aoki et al. 2013 showed noise-driven stochastic ERK pulses regulating density-dependent proliferation; Goglia et al. 2020 used a live-cell screen of altered ERK dynamics to derive principles of proliferative control). Min et al. 2020 demonstrated that mother cells temporally integrate their mitogen (ERK-pulse) history and that this integrated history controls whether daughter cells proliferate, establishing time integration rather than instantaneous activity as a plausible decoding variable. Arekatla et al. 2023 used optogenetic (light-activated) manipulation of ERK and Akt to show that their dynamics control exit from pluripotency in mouse embryonic stem cells, demonstrating that causal tests of dynamic encoding are experimentally feasible. On the data side, Stern et al. 2022 generated single-cell ERK and Akt activity time courses linked to individually annotated division events in MCF10A cells (a non-tumorigenic human mammary epithelial line). Bennett et al. 2024 then showed, using discrete wavelet transforms (mathematical decompositions that separate a time course into distinct frequency bands) to extract low-frequency features of these time courses followed by ensemble machine-learning classification, that an individual cell's division within a 48-hour window can be predicted from its ERK and Akt dynamics (F-measure 0.524 and AUC 0.726 in MCF10A, where the F-measure is the harmonic mean of precision and recall and the AUC is the area under the receiver operating characteristic curve, a threshold-free measure of classifier discrimination), that ERK dynamics are more predictive than Akt dynamics with the combination best, and that the ERK model transfers to asynchronously cycling RPE (retinal pigment epithelial) cells in full growth medium. Model interpretation found no single time point uniquely informative, consistent with a time-integrator model in which the amount of time the pathway is on dictates division probability. The authors are explicit, however, that these results are predictive and not causal, that the contexts examined were narrow (acute EGF plus insulin stimulation after serum and growth-factor starvation in MCF10A; ERK inhibition experiments in RPE), and that causality, ligand dependence, and broader generality remain to be established.

Problem Statement

Determine the causal status and contextual scope of the relationship between single-cell ERK and Akt activity dynamics and stochastic cell division, as posed by Bennett et al. (2024). Concretely: (1) Causality - determine whether the duration or integral of ERK and/or Akt pathway activity causally determines the probability that an individual cell divides, as the time-integrator interpretation suggests, or whether the observed predictive relationship is only correlative (for example because cell-cycle progression feeds back onto pathway activity). (2) Ligand and pathway dependence - determine whether the predictive relationship, including the relative importance of ERK versus Akt dynamics, changes depending on which upstream growth factor or receptor pathway activates the signaling, beyond the EGF-plus-insulin stimulation of MCF10A cells and ERK-inhibition perturbations in RPE cells examined so far. (3) Generality - establish how the relationship behaves across cell types, growth-factor doses, chronic (as opposed to starvation-acute) growth-factor conditions, and observation windows long enough to include repeated divisions and additional fate events, which requires generating and annotating the corresponding single-cell ERK/Akt time-course and division datasets. The axes listed in component (3) delimit the class of contexts over which generality is to be mapped rather than a checklist that must be covered exhaustively; resolution of this component nonetheless requires extending the relationship beyond the EGF-plus-insulin MCF10A and RPE contexts examined so far. These three components jointly serve one objective: replacing the correlational, narrow-context evidence with a causal and mapped account of how ERK/Akt dynamics control stochastic cell division.

References

  1. Bennett, J. J. R., Stern, A. D., Zhang, X., Birtwistle, M. R., Pandey, G. 'Low-frequency ERK and Akt activity dynamics are predictive of stochastic cell division events.' npj Systems Biology and Applications 10, 65 (2024). DOI 10.1038/s41540-024-00389-7, https://doi.org/10.1038/s41540-024-00389-7 (source paper)
  2. Stern, A. D., Smith, G. R., Santos, L. C., Sarmah, D., Zhang, X., Lu, X., Iuricich, F., Pandey, G., Iyengar, R., Birtwistle, M. R. 'Relating individual cell division events to single-cell ERK and Akt activity time courses.' Scientific Reports 12, 18077 (2022). DOI 10.1038/s41598-022-23071-6, https://doi.org/10.1038/s41598-022-23071-6
  3. Min, M., Rong, Y., Tian, C., Spencer, S. L. 'Temporal integration of mitogen history in mother cells controls proliferation of daughter cells.' Science 368, 1261-1265 (2020). DOI 10.1126/science.aay8241, https://doi.org/10.1126/science.aay8241
  4. Aoki, K., Kumagai, Y., Sakurai, A., Komatsu, N., Fujita, Y., Shionyu, C., Matsuda, M. 'Stochastic ERK activation induced by noise and cell-to-cell propagation regulates cell density-dependent proliferation.' Molecular Cell 52, 529-540 (2013). DOI 10.1016/j.molcel.2013.09.015, https://doi.org/10.1016/j.molcel.2013.09.015
  5. Goglia, A. G., Wilson, M. Z., Jena, S. G., Silbert, J., Basta, L. P., Devenport, D., Toettcher, J. E. 'A live-cell screen for altered Erk dynamics reveals principles of proliferative control.' Cell Systems 10, 240-253 (2020). DOI 10.1016/j.cels.2020.02.005, https://doi.org/10.1016/j.cels.2020.02.005
  6. Arekatla, G., Trenzinger, C., Reimann, A., Loeffler, D., Kull, T., Schroeder, T. 'Optogenetic manipulation identifies the roles of ERK and AKT dynamics in controlling mouse embryonic stem cell exit from pluripotency.' Developmental Cell 58, 1022-1036 (2023). DOI 10.1016/j.devcel.2023.04.013, https://doi.org/10.1016/j.devcel.2023.04.013
  7. Bonsignore, F., Pozzi, S., Aloi, E., Mazza, D., Zambrano, S. 'Linking signaling dynamics and cell fate decisions through single-cell imaging: evidence and challenges.' Frontiers in Cell and Developmental Biology 13, 1656051 (2025). DOI 10.3389/fcell.2025.1656051, https://doi.org/10.3389/fcell.2025.1656051
  8. Zuniga, A. G., Aikin, T. J., McKenney, C., Lendner, Y., Phung, A., Hook, P. W., Meltzer, A., Timp, W., Regot, S. 'Sustained ERK signaling promotes G2 cell cycle exit and primes cells for whole-genome duplication.' Developmental Cell 59, 1724-1736 (2024). DOI 10.1016/j.devcel.2024.03.032, https://doi.org/10.1016/j.devcel.2024.03.032
  9. Hoffman, T. E., Tian, C., Nangia, V., Yang, C., Regot, S., Gerosa, L., Spencer, S. L. 'CDK2 activity crosstalk on the ERK kinase translocation reporter can be resolved computationally.' Cell Systems 16, 101162 (2025). DOI 10.1016/j.cels.2024.12.003, https://doi.org/10.1016/j.cels.2024.12.003
  10. Kohane, F. V., Johnstone, C., Neumann, D. P., Gunawan, I., Huang, T., Vafaee, F., Chaffer, C. L., Lock, J. G. 'Epithelial-mesenchymal cell state heterogeneity predetermines differential phospho-signaling responses to EGF stimulation.' Molecular Biology of the Cell 37 (2026; published online December 2025). DOI 10.1091/mbc.e25-07-0349, https://doi.org/10.1091/mbc.e25-07-0349
  11. Nussinov, R., Zhang, W., Liu, Y., Jang, H. 'Mitogen signaling strength and duration can control cell cycle decisions.' Science Advances 10 (2024). DOI 10.1126/sciadv.adm9211, https://doi.org/10.1126/sciadv.adm9211
  12. Riede, L., Borowiec, A., Mallick, S., Mallick, S., Chakrabarti, J., Thorne, C. A., Pond, K. W. 'Preservation of human colonic stem cells requires an ERK dynamics checkpoint mediated by AKT.' bioRxiv preprint (2026). DOI 10.64898/2026.04.02.715982, https://doi.org/10.64898/2026.04.02.715982