Abstract: The focus of pancreatic cancer research has been shifted from pancreatic cancer cells towards their microenvironment, involving pancreatic stellate cells that interact with cancer cells and influence tumor progression. To quantitatively understand the pancreatic cancer microenvironment, we construct a computational model for intracellular signaling networks of cancer cells and stellate cells as well as their intercellular communication. We extend the rule-based BioNetGen language to depict intra- and inter-cellular dynamics using discrete and continuous variables respectively. Our framework also enables a statistical model checking procedure for analyzing the system behavior in response to various perturbations. The results demonstrate the predictive power of our model by identifying important system properties that are consistent with existing experimental observations. We also obtain interesting insights into the development of novel therapeutic strategies for pancreatic cancer.
Introduction
Pancreatic cancer (PC) remains one of the most aggressive and treatment-resistant malignancies, with a prognosis that has seen little improvement over several decades. Traditionally, research has focused almost exclusively on the intrinsic genetic mutations of pancreatic cancer cells (PCCs). However, recent evidence highlights the tumor microenvironment (TME)—the complex ecosystem of cells and molecules surrounding the tumor—as a primary driver of treatment failure.
A central player in this environment is the Pancreatic Stellate Cell (PSC). In a healthy pancreas, these cells are quiescent. In the presence of cancer, they become activated, migrating toward the tumor to form a dense, protective "cocoon" that shields cancer cells from both chemotherapy and immune surveillance. Understanding the signaling "conversation" between PCCs and PSCs is critical for developing effective therapies. This paper presents a hybrid computational framework designed to model these multi-scale interactions and identify potential new strategies for multi-drug intervention.
Methodology and Model Construction
To address the complexity of the TME without the need for thousands of precise chemical reaction rates (which are often unknown), the authors employed a multi-scale formal modeling approach:
- Hybrid Modeling Language: The researchers extended the BioNetGen language to handle different types of dynamics. Intracellular signaling (inside the cell) is modeled using Boolean Networks (BNs), where proteins are treated as binary states ("on" or "off"). This captures the logic of signaling pathways without requiring exact kinetic data. Intercellular communication (between cells) is modeled using continuous variables to represent the concentration of signaling molecules like growth factors in the extracellular space.
- Statistical Model Checking (StatMC): The model's behavior was analyzed using Bounded Linear Temporal Logic (BLTL). This allows researchers to ask complex probabilistic questions, such as the likelihood of a cancer population reaching a stable high-growth state under specific conditions.
- Simulation: The NFsim stochastic simulator was used to manage the massive state space inherent in multicellular systems.
The resulting model, illustrated in Figure 1, integrates the signaling pathways of both PSCs and PCCs. It maps how external stimuli like VEGF, EGF, and TGFβ1 trigger internal cascades involving key proteins such as RAS, PI3K, AKT, and p53, ultimately influencing cell behaviors like proliferation, apoptosis, and autophagy.
Key Insights and Results
The Role of the Microenvironment
The model confirms the critical role of PSCs in tumor stability. Simulations showed that the probability of a cancer population reaching a high, stable level jumps from 40.5% (without stellate cells) to 99.6% when stellate cells are present. This aligns with clinical observations of PSC-driven tumor progression.
Mechanisms of Drug Resistance
The model provides a clear explanation for the failure of monotherapies, such as drugs targeting only the EGF receptor (e.g., Erlotinib). PSCs provide redundant pro-survival signals that allow cancer cells to bypass single-point blockades. Furthermore, the model captures the transition from apoptosis (programmed cell death) to autophagy (self-eating for survival), a switch that helps cancer cells endure nutrient-poor conditions and therapeutic stress.
Identifying New Therapeutic Targets
By perturbing different nodes in the network, the researchers identified promising new targets:
- Targeting PSC Migration: Inhibiting ERK specifically in stellate cells was found to be more effective than many current PCC-targeted drugs because it prevents PSCs from migrating and forming their protective barrier.
- Polypharmacological Strategies: The model suggests that simultaneously inhibiting RAS in cancer cells and ERK in stellate cells significantly reduces the concentration of growth factors (like bFGF) in the environment, effectively "starving" the tumor and overcoming the redundancy of the signaling network.
Discussion and Future Work
While the current model successfully reproduces known clinical outcomes and offers novel therapeutic insights, it remains a simplified representation of the pancreatic microenvironment. The authors identify several areas for future development:
- Expanding the TME: Future versions of the model will aim to include other critical components, such as cancer-associated macrophages and T-cells, to better understand immune evasion.
- Experimental Validation: The novel drug combinations identified, such as dual RAS/ERK inhibition, require validation in wet-lab animal models or clinical trials.
- Spatial Dynamics: Incorporating 3D spatial architecture will be necessary to account for how signal diffusion is affected by the physical structure of the tumor and its surrounding stroma.
Summary updated