Poster H2 - Mapping Clonal Diversity: Spatiotemporal Evolution of Pancreatic Tumors via Agent-Based Modeling

Friday, October 25, 2024

3:30 PM - 4:30 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Siqi "Kiki" Zhang (she/her/hers)

Undergraduate
Biomedical Engineering, Johns Hopkins University
Baltimore, Maryland, United States

Primary Investigator(s)

Rachel Karchin, MS PhD (she/her/hers)

Professor of Biomedical Engineering
Johns Hopkins University, Maryland, United States

Introduction: Cancer is characterized by the continuous evolution of diverse subclonal populations, a process known as clonal evolution, driving tumor progression and heterogeneity from early lesions to metastasis. This evolution enables tumors to adapt and survive, impacting their response to treatments. Metastasis, the leading cause of cancer death, involves complex interactions between primary tumors and distant sites, where tumor cells acquire genetic and epigenetic changes to spread and colonize new areas. Understanding these dynamics is crucial for developing targeted therapies.

In this study, we developed an agent-based model (ABM) to simulate the spatial and temporal evolution of pancreatic tumor cells, focusing on mutation dynamics, subclone formation, and growth rates. Our ABM extends a previous logistic stochastic model by incorporating passenger and driver mutations, allowing for finer granularity of cell-cell interactions. The model starts with an initiating population of type-0 cells, which have not yet evolved the potential to disseminate. Type-0 cells can mutate into type-1 cells with metastatic potential, and type-1 cells can further evolve into type-2 cells capable of establishing new colonies.

This provides a detailed depiction of clonal dynamics and metastasis in pancreatic tumors, offering new insights into tumor heterogeneity and evolution. By tracking the spatial distribution and temporal evolution of different cell types and subclones, our model captures the complexity of tumor growth and metastatic spread. This approach highlights the importance of clonal interactions and the role of mutations in driving tumor heterogeneity, providing a robust framework for exploring therapeutic strategies targeting specific subclones or mutation pathways.

Materials and

Methods: We developed an agent-based model (ABM) using the Mesa framework in Python to simulate the progression of pancreatic tumors, incorporating three cell types: Type-0 cells (initial population with no metastatic potential), Type-1 cells (evolved with metastatic potential), and Type-2 cells (capable of establishing new metastatic colonies). Cells are positioned on a two-dimensional lattice with a Moore neighborhood of 8 adjacent sites. Replication occurs if adjacent sites are unoccupied, introducing unique mutations following a Poisson distribution and the infinite alleles model. Driver mutations provide a selective advantage of 0.004 to the birth rate, while passenger mutations do not confer fitness benefits.

Population growth is constrained by carrying capacities: McarryP for Type-0 and Type-1 cells in the primary site, and McarryM for Type-2 cells in metastatic sites. Birth rates decrease as populations approach these limits, simulating density-dependent growth. Model parameters, including initial birth and death rates for the three cell types, McarryP, McarryM, and metastasis rate, were derived from previous publications that analyzed clinical data of 101 pancreatic cancer patients from Johns Hopkins University. The primary tumor and metastatic site sizes provided the basis for setting these carrying capacities.

Simulations were conducted over 100 steps, exploring various mutation and driver mutation rates to study their impact on tumor growth, mutation dynamics, subclone formation, and metastasis. We collected model and agent-specific data every 10 steps. The model tracked the spatial and temporal evolution of tumor cells, providing detailed analyses of clonal dynamics and heterogeneity.

Results, Conclusions, and Discussions: Our simulations provided valuable insights into the dynamics of pancreatic tumor progression under varying mutation and driver mutation rates. Figures 1 illustrates changes in tumor cell birth and death rates, with significant differences observed post-step 30, indicating a critical phase in tumor evolution. Higher mutation rates (0.1) and increased driver mutation rates (0.001 and 0.002) led to substantial rises in both birth and death rates, suggesting heightened cellular activity. This aligns with the increased number of Type-1 and Type-2 cells (Figure 2), indicating enhanced metastatic potential. Notably, Type-0 cell populations (Figure 2) grew more rapidly at lower mutation rates, highlighting their influence on early tumor growth. Figure 3 shows the genetic similarity among tumor cells at step 30, represented by the Jaccard index. Most cell pairs exhibit low genetic similarity, indicating significant genetic diversity. Proximal cells have similar Jaccard indices, suggesting localized genetic similarity within a diverse landscape, crucial for tumor adaptability. Figure 4 depicts the spatial distribution of cell types at step 100. Type-1 cells concentrate around the central tumor region, while Type-2 cells, with metastatic potential, are located at the periphery, suggesting active migration.

Our agent-based model elucidates the spatial and temporal dynamics of pancreatic tumor cells under varying mutation and driver mutation rates. The results highlight the significant impact of these rates on tumor cell proliferation, differentiation, and genetic diversity. The spatial patterns of cell types and genetic similarity underscore the complexity of tumor evolution and the critical role of genetic heterogeneity in tumor progression and metastasis.

This study emphasizes the importance of clonal evolution and genetic diversity within tumors. Our ongoing research will investigate subclonal interactions, focusing on cooperation or competition among subclones. Understanding these interactions is crucial for developing targeted therapies to disrupt tumor dynamics and inhibit growth and metastasis. Future work will refine the model to incorporate detailed cellular behaviors and explore the effects of different therapeutic strategies on pancreatic tumor evolution.

Acknowledgements (Optional):