Coupled Source-Sink Habitats Produce Spatial and Temporal Variation of Cancer Cell Molecular Properties as an Alternative to Branched Clonal Evolution and Stem Cell Paradigms

Abstract

Intratumoral molecular cancer cell heterogeneity is conventionally ascribed to the accumulation of random mutations that occasionally generate fitter phenotypes. This model is built upon the “mutation-selection” paradigm in which mutations drive ever-fitter cancer cells independent of environmental circumstances. An alternative model posits spatio-temporal variation (e.g., blood flow heterogeneity) drives speciation by selecting for cancer cells adapted to each different environment. Here, spatial genetic variation is the consequence rather than the cause of intratumoral evolution. In nature, spatially heterogenous environments are frequently coupled through migration. Drawing from ecological models, we investigate adjacent well-perfused and poorly-perfused tumor regions as “source” and “sink” habitats, respectively. The source habitat has a high carrying capacity resulting in more emigration than immigration. Sink habitats may support a small (“soft-sink”) or no (“hard-sink”) local population. Ecologically, sink habitats can reduce the population size of the source habitat so that, for example, the density of cancer cells directly around blood vessels may be lower than expected. Evolutionarily, sink habitats can exert a selective pressure favoring traits different from those in the source habitat so that, for example, cancer cells adjacent to blood vessels may be suboptimally adapted for that habitat. Soft sinks favor a generalist cancer cell type that moves between the environment but can, under some circumstances, produce speciation events forming source and sink habitat specialists resulting in significant molecular variation in cancer cells separated by small distances. Finally, sink habitats, with limited blood supply, may receive reduced concentrations of systemic drug treatments; and local hypoxia and acidosis may further decrease drug efficacy allowing cells to survive treatment and evolve resistance. In such cases, the sink transforms into the source habitat for resistant cancer cells, leading to treatment failure and tumor progression. We note these dynamics will result in spatial variations in molecular properties as an alternative to the conventional branched evolution model and will result in cellular migration as well as variation in cancer cell phenotype and proliferation currently described by the stem cell paradigm.

Modeling cancer’s ecological and evolutionary dynamics

“…Although we have presented a simplified version of the evolution of drug resistance in cancer, there are many ways this model can be extended, depending on the question of interest. For example, one could consider the influence of other cell types such as immune cells and fibroblasts, spatiotemporal variation and heterogeneity in the microenvironment [ 27 , 61 , 85 ], side effects of drugs [ 86 ], drug scheduling [ 109 ], evolutionarily informed therapies [ 18 , 19 ], and the effects of plasticity and cell states [ 20 , 21 ].…”

“…This modeling framework, inspired by traditional quantitative genetics approaches, has traditionally been used to investigate problems in evolutionary ecology such as predator–prey coevolution [ 15 ] or consumer-resource games [ 87 ]. Recently, it has been adapted to explore problems in cancer [ 19 – 21 , 27 , 28 , 86 ]. The hallmarks of cancer provide a useful perspective to form the basis of a fitness generating G -function framework for cancer [ 52 ].…”

A life history model of the ecological and evolutionary dynamics of polyaneuploid cancer cells

“…Instead of having a separate G function for each state, since cells in both 2N+ and PACC states are part of the same species, we must derive a unifying G function that incorporates both in a life history-enlightened manner. To do this, we follow an approach introduced in our recent paper 41 where we represent our model as a population projection matrix (PPM) 42 – 44 and use the spectral bound of the matrix as a measure of fitness, as it controls the long-term (asymptotic) growth rate of the population. Namely, our PPM takes the following form: …”

Digest: Natural selection fluctuates at an extremely fine spatial scale inside a wild population of snapdragon plants *

“…This requires the development of new theoretical techniques that incorporate state structured populations across heterogeneous environments within a single population as well as experimental studies to supplement the theory (Cunningham et al. 2021). In cancer, microenvironmental habitats close to blood vessels are significantly different than those closer to the tumor core; cancer cells in the former have plentiful access to nutrients but are also more subject to “predatory” immune cells and therapy.…”

“…Theoretical studies show how these microenvironmental differences generate frequency and density‐dependent selection both within and among microenvironments, generating local cancer phenotypes (Cunningham et al. 2021). Understanding how these different cancer phenotypes evolve is essential to develop evolutionarily informed strategies for cancer treatment.…”