Cells within the same tumor can exhibit different
EMT-associated phenotypes—an epithelial phenotype, a mesenchymal phenotype, and one or more hybrid E/M phenotypes.
This is a canonical example of nongenetic intra-tumoral heterogeneity observed across cancer types including in breast cancer [46],
melanoma [47], colorectal cancer [48], and in prostate cancer
[49]. Different EMT-associated phenotypes exhibit varying
tumor-initiating capabilities [6, 7] and sensitivities to anticancer
drugs [50, 51]. How does such epithelial-mesenchymal heterogeneity emerge in a population of cancer cells? How is this heterogeneity maintained and propagated across generations and passages?
These are key questions that must be answered if we are to be able
to attenuate the role of epithelial-mesenchymal heterogeneity in
driving the failure of anticancer therapies.
Multiple nongenetic mechanisms can contribute towards the
emergence of phenotypic heterogeneity. The regulatory circuits
that govern the phenotypes of different cells often respond differently to the same external cues leading to a phenotypically heterogeneous population. Phenotypes of cells in a population can change
stochastically due to the noisy transcription of genes [52] or due to
the random partitioning of the parent cell molecules among the
daughter cells during cell division [53, 54]. Finally, cell–cell communication can cause cells in a population to acquire distinct phenotypes in a non-cell autonomous manner. Each of these three
mechanisms has been implicated in the emergence and maintenance of epithelial-mesenchymal heterogeneity. Mathematical and
computational modeling approaches have played a key role in
determining how these mechanisms can drive epithelialmesenchymal heterogeneity in populations of cancer cells. Here,
we describe mathematical modeling approaches corresponding to
each of the three mechanisms.
4 Heterogeneity from Cell-to-Cell Variation in Regulatory Kinetics
Large and complex gene regulatory networks underlie different
cellular functions such as stem cell differentiation [55, 56] and
circadian rhythm [57, 58]. The dynamical behavior of such large
networks can be understood as being driven by a core regulatory
circuit with the remaining genes in the circuit being peripheral to
circuit dynamics, acting only to alter the signaling status of the core
regulatory circuit [59]. The effects of peripheral genes and exogenous signaling can then be modeled as perturbations to the kinetic
parameters governing the dynamics of the core regulatory module.
This is the approach underlying the framework known as random
circuit perturbation or RACIPE [60]. Here, we describe how to use
RACIPE for modeling epithelial-mesenchymal heterogeneity.
392
Shubham Tripathi et al.
EMT-associated phenotypes—an epithelial phenotype, a mesenchymal phenotype, and one or more hybrid E/M phenotypes.
This is a canonical example of nongenetic intra-tumoral heterogeneity observed across cancer types including in breast cancer [46],
melanoma [47], colorectal cancer [48], and in prostate cancer
[49]. Different EMT-associated phenotypes exhibit varying
tumor-initiating capabilities [6, 7] and sensitivities to anticancer
drugs [50, 51]. How does such epithelial-mesenchymal heterogeneity emerge in a population of cancer cells? How is this heterogeneity maintained and propagated across generations and passages?
These are key questions that must be answered if we are to be able
to attenuate the role of epithelial-mesenchymal heterogeneity in
driving the failure of anticancer therapies.
Multiple nongenetic mechanisms can contribute towards the
emergence of phenotypic heterogeneity. The regulatory circuits
that govern the phenotypes of different cells often respond differently to the same external cues leading to a phenotypically heterogeneous population. Phenotypes of cells in a population can change
stochastically due to the noisy transcription of genes [52] or due to
the random partitioning of the parent cell molecules among the
daughter cells during cell division [53, 54]. Finally, cell–cell communication can cause cells in a population to acquire distinct phenotypes in a non-cell autonomous manner. Each of these three
mechanisms has been implicated in the emergence and maintenance of epithelial-mesenchymal heterogeneity. Mathematical and
computational modeling approaches have played a key role in
determining how these mechanisms can drive epithelialmesenchymal heterogeneity in populations of cancer cells. Here,
we describe mathematical modeling approaches corresponding to
each of the three mechanisms.
4 Heterogeneity from Cell-to-Cell Variation in Regulatory Kinetics
Large and complex gene regulatory networks underlie different
cellular functions such as stem cell differentiation [55, 56] and
circadian rhythm [57, 58]. The dynamical behavior of such large
networks can be understood as being driven by a core regulatory
circuit with the remaining genes in the circuit being peripheral to
circuit dynamics, acting only to alter the signaling status of the core
regulatory circuit [59]. The effects of peripheral genes and exogenous signaling can then be modeled as perturbations to the kinetic
parameters governing the dynamics of the core regulatory module.
This is the approach underlying the framework known as random
circuit perturbation or RACIPE [60]. Here, we describe how to use
RACIPE for modeling epithelial-mesenchymal heterogeneity.
392
Shubham Tripathi et al.