Modeling of mesenchymal hybrid epithelial state and phenotypic transitions in EMT and MET processes of cancer cells

Peng He,Ya Jia,Kang Qiu

doi:10.1038/s41598-018-32737-z

Abstract

Based on the transcriptional regulatory mechanisms between microRNA-200 and transcription factor ZEB in an individual cancer cell, a minimal dynamic model is proposed to study the epithelial-mesenchymal transition (EMT) and mesenchymal-epithelial transition (MET) processes of cancer cells. It is shown that each cancer cell can exit in any of three phenotypic states: the epithelial (E) state, the mesenchymal (M) state, and the epithelial/mesenchymal (E/M) hybrid state, and the state of cancer cell can interconvert between different states. The phase diagram shows that there are monostable, bistable, and tristable phenotypic states regions in a parameters plane. It is found that different pathway in the phase diagram can correspond to the EMT or the MET process of cancer cells, and there are two possible EMT processes. It is important that the experimental phenomenon of E/M hybrid state appearing in the EMT process but rather in the MET process can be understood through different pathways in the phase diagram. Our numerical simulations show that the effects of noise are opposite to these of time delay on the expression of transcription factor ZEB, and there is competition between noise and time delay in phenotypic transitions process of cancer cells.

Highlights

Modeling of microscopic regulatory mechanisms in single cancer cell is a crucial step towards understanding macroscopic physiological phenomena of cancer cells
In this paper, based on the core regulatory network proposed by Lu et al.[19], a general minimal dynamic model with two variables is proposed to understand the epithelialmesenchymal transition (EMT) and mesenchymal-epithelial transition (MET) processes and the effects of both noise and time delay
It was reported that cancer cells in the E/M hybrid phenotype exhibit stemness characters during EMT in developmental, regenerative, as well as pathological contexts[48,49,50,51], and related to drug resistance[52,53,54]

Summary

Model and Method

The four components model of the miR-34/SNAIL and the miR-200/ZEB mutually inhibiting loops[19,37] of cancer cells is shown in Fig. 1(a), it can be found that the interactions between transcription factors and microRNAs have symmetrical properties. The four components model of the miR-34/SNAIL and the miR-200/ZEB mutually inhibiting loops[19,37] of cancer cells is shown, it can be found that the interactions between transcription factors and microRNAs have symmetrical properties. One can simplify core regulatory network by a general minimal regulatory model with two variables as shown by Fig. 1(b), in this paper, X1 represents the transcription factor ZEB family, and X2 represents the miR-200, and the SNAIL signal activates the expression of ZEB and inhibits the expression of miR-200 simultaneously. A is the activation strength of ZEB induced by both SNAIL signal and itself, b is the inhibited strength of miR-200 induced by both. The stationary probability Pst(x1, x2) obtained from the Fokker-Planck Eq (5) can be used to indicate the phenotypes proportion distribution of cancer cells.

Results

Conclusions and Discussions

Additional Information