Stochastic modeling of p53-regulated apoptosis upon radiation damage

We develop and study the evolution of a model of radiation induced apoptosis in cells using stochastic simulations, and identified key protein targets for effective mitigation of radiation damage. We identified several key proteins associated with cellular apoptosis using an extensive literature survey. In particular, we focus on the p53 transcription dependent and p53 transcription independent pathways for mitochondrial apoptosis. Our model reproduces known p53 oscillations following radiation damage. The key, experimentally testable hypotheses that we generate are - inhibition of PUMA is an effective strategy for mitigation of radiation damage if the treatment is administered immediately, at later stages following radiation damage, inhibition of tBid is more effective.

💡 Research Summary

This paper presents a stochastic systems‑biology model of radiation‑induced apoptosis, focusing on the dual roles of p53 in transcription‑dependent and transcription‑independent pathways that converge on mitochondrial death signaling. Using the Gillespie direct algorithm, the authors construct a reaction network that captures p53 synthesis, nuclear translocation, oligomerization, Mdm2‑mediated ubiquitination, and degradation, as well as the upstream “A→B→p53C” activation cascade that mimics the gradual rise of p53 after ionizing radiation. The model reproduces the experimentally observed p53 oscillations (periods of 2–3 h) and the subsequent induction of pro‑apoptotic genes PUMA and Bax.

In the mitochondrial arm, the network includes competitive binding of PUMA and p53 to the anti‑apoptotic protein Bcl‑2, the recruitment of tBid to mitochondria, activation of Bax/Bak, Bax oligomerization (dimers → tetramers → octamers), mitochondrial outer‑membrane permeabilization (MOMP), release of cytochrome c and Smac, formation of the apoptosome (Apaf‑1·Cyto‑c·Caspase‑9), activation of Caspase‑3, and regulation by XIAP and Smac. All species are subject to synthesis and degradation reactions, ensuring realistic steady‑state concentrations.

Four pharmacological interventions are modeled: inhibition of Mdm2, PUMA, Bid, and active Caspase‑3 (C3*). The authors simulate drug addition at different post‑irradiation times and evaluate the impact on the probability of cell death. The key findings are:

1. Early inhibition (≤ 1 h) of PUMA dramatically reduces downstream Bax/Bak activation and lowers the overall death probability by more than 60 %. This reflects the fact that PUMA binds Bcl‑2 with higher affinity than p53, displacing p53 and freeing Bax for activation.

2. Later inhibition (4–6 h after exposure) of Bid is more effective than early PUMA inhibition because, at this stage, tBid is the primary driver of Bax activation and oligomerization. Blocking Bid at this window suppresses the amplification loop that would otherwise sustain MOMP.

3. Mdm2 inhibition paradoxically enhances p53 oscillations and transcription‑dependent apoptosis, making it unsuitable for radioprotection.

4. Direct Caspase‑3 inhibition reduces the final execution phase but has limited benefit if upstream mitochondrial permeabilization has already occurred.

Parameter sensitivity analysis identifies the Mdm2‑p53 feedback loop and the PUMA‑Bcl‑2 binding constants as the most influential determinants of cell‑fate outcomes. The stochastic framework naturally captures cell‑to‑cell variability: even under identical radiation doses, simulated cells display a distribution of death times, consistent with experimental observations of heterogeneous responses.

The authors argue that deterministic ODE models, which assume large molecule numbers and ignore stochastic fluctuations, cannot reproduce these heterogeneities or the timing‑dependent efficacy of inhibitors. By integrating spatial compartmentalization (cytoplasm vs. mitochondria) and low‑copy‑number stochasticity, the model provides mechanistic insight into why the same drug can be protective at one time point and ineffective at another.

The paper concludes with several actionable recommendations: (i) develop or repurpose high‑affinity PUMA inhibitors for immediate post‑irradiation administration; (ii) design tBid‑targeting agents (e.g., BH3‑mimetic peptides) for delayed treatment; (iii) avoid Mdm2 antagonists in radioprotection contexts; and (iv) consider combination regimens that address both early (PUMA) and late (Bid) phases. Future extensions could incorporate DNA damage response kinases (ATM/ATR), p53 mutant variants, and tissue‑specific expression profiles to enable personalized radioprotective strategies.