Artwork: Evolution of phenotypic plasticity results in genetic and phenotypic tumor heterogeneity, with a dense tumor core of fast-growing cells (teal) surrounded by a diffuse rim of invasive cells (brown).
Tumors can be viewed as ecosystems consisting of cancer cells that differ in genotype and phenotypic traits such as metabolism, motility, proliferation, and treatment resistance [1] . These heterogeneous populations of cancer cells interact with one another as well as with their surrounding microenvironment, which includes stromal cells, immune cells, extracellular matrix components, and signaling molecules. As a consequence, no tumor is alike, even for patients with the same diagnosis! This tumor heterogeneity is a major obstacle to cancer treatment since it enables treatment resistance and leads to heterogeneous treatment outcomes.
Two crucial factors that lead to tumor heterogeneity are irreversible genetic changes and reversible phenotypic changes. Notably, the latter was recently proposed as a novel hallmark of cancer [2] . However, most studies focus only on one of these processes.
But how can we model the interplay between genetic changes and phenotypic plasticity and what's their effect on cancer treatment?
In our latest paper, we shed light on this question by focusing on the phenotypic switch between migration and proliferation (also known as the "go-or-grow" dichotomy), which is essential in glioblastoma, the most deadly brain tumor.
Figure 1:
(A) The cell’s genotype dictates the regulation of the phenotypic switch. This switch determines the cell’s reaction to its microenvironment. Subsequently, the cell interacts and shapes its microenvironment according to its phenotype. (B) In the mathematical model and the context of the go-or-grow dichotomy, the genotype is represented by a parameter which controls the phenotypic switch between the migratory and proliferating phenotypes dependent on the local cell density. After assuming either phenotype the cell influences the local microenvironment by either reducing the cell density (by migration) or increasing it (by proliferation).
Our core idea is that mutations alter the regulation of phenotypic plasticity, specifically the reaction to the microenvironment, rather than just phenotypic traits like the proliferation rate or migration speed.
We study this hypothesis using a novel, spatially explicit model that tracks individual cells’ phenotypic and genetic states. This model is an extension of the BIO-LGCA framework, a cellular automaton class that is especially useful for modeling cell migration [3] . When a new cancer cell is born, a mutation can change its genotype and thereby its regulation of the phenotypic switch. Each cell has its unique phenotypic switch function (or "go-or-grow strategy") which determines its phenotype in dependence on the cell density, a proxy for the microenvironment. This approach contrasts with traditional models by explicitly incorporating both genetic and reversible phenotypic changes, allowing for a more nuanced understanding of tumor evolution.
We observe that cells at the tumor edge evolve to favor migration over proliferation and vice versa in the tumor bulk. This leads to a dense tumor core of fast-growing cells surrounded by a diffuse tumor rim of invasive cells. Notably, different genetic configurations, i.e., different regulations of the phenotypic switch, can result in this pattern of phenotypic heterogeneity.
Figure 2:
Snapshots of an example simulation on a hexagonal lattice. (A) Total cell density, (B) migratory cells, (C) proliferating cells and (D) average local switch parameter. Spatial heterogeneity of genotypes, phenotypes, and density emerges, with a high-density tumor core of proliferating cells surrounded by a low-density tumor rim of migratory cells.
Finally, we investigate implications for cancer therapy by simulating a simplified treatment scenario using the synthetic tumors as initial conditions. Since model parameters are usually not obtainable from patient data we instead test the predictive power of genetic and phenotypic heterogeneity measures. We find that it is phenotypic, rather than genetic, heterogeneity that predicts tumor recurrence after therapy. This offers new insights into the significant variability in glioblastoma recurrence times post-treatment.
Figure 3:
Genetic and phenotypic heterogeneity predict treatment success. Shown is the time to recurrence of synthetic tumors versus the genetic entropy (A) and phenotypic entropy (B) as measures for tumor heterogeneity. Higher phenotypic entropy is associated with lower recurrence times, while higher genetic entropy is associated with higher recurrence time, corresponding to better prognosis. Black dashed lines are linear regressions.
In the future we plan on building on the foundation laid in this paper by exploring additional aspects of tumor evolution and therapy response. For example, explicitly incorporating dependencies on multiple environmental factors like oxygen and growth factors would lead to a complex, non-monotonic phenotypic switch function, potentially resulting in even richer dynamics.
Summing up, in this paper we demonstrated how the interplay of phenotypic plasticity and Darwinian evolution leads to multi-scale heterogeneity, i.e., heterogeneous spatial distributions of cancer subclones and phenotypes. This heterogeneity, in turn, affects treatment outcomes and leads to considerable variability between patients.
References
Merlo LMF, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and ecological process. Nat Rev Cancer. 2006;6: 924–935. doi:10/bhqtbz
Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov. 2022;12: 31–46. doi:10/gn3vvz
Deutsch A, Nava-Sedeño JM, Syga S, Hatzikirou H. BIO-LGCA: A cellular automaton modelling class for analysing collective cell migration. PLOS Comput Biol. 2021;17: e1009066. doi:10/gqpwd3
