Multi range ERK responses shape the proliferative trajectory of single cells following oncogene induced senescence

Abstract

Oncogene-induced senescence (OIS) is a phenomenon in which aberrant oncogene expression causes non-transformed cells to enter a non-proliferative state. Cells undergoing OIS display phenotypic heterogeneity, with some cells senescing and others remaining proliferative. The causes of the heterogeneity remain poorly understood. We studied the sources of heterogeneity in the responses of human epithelial cells to oncogenic BRAFV600E expression. We found that a narrow expression range of BRAFV600E generated a wide range of activities of its downstream effector ERK. In population-level and single cell assays, ERK activity displayed a non-monotonic relationship to proliferation, with intermediate ERK activities leading to maximal proliferation. We profiled gene expression across a range of ERK activities over time and characterized four distinct ERK response classes, which we propose act in concert to generate the unique ERK-proliferation response. Altogether, our studies mapped the input-output relationships between ERK activity and proliferation providing important insights into how heterogeneity can be generated during OIS.

Low-frequency ERK and Akt activity dynamics are predictive of stochastic cell division events

“…We also found that the ERK-based model developed from the starved and growth-factor induced experimental setup was predictive of division fates from chronically growth factor treated, asynchronously cycling contexts in an independent dataset [77]. This independent dataset was from a different cell line, RPE cells, which although still of epithelial origin like MCF10A, provides further evidence of the potential generality of the models developed.…”

“…This raises the question of whether the models developed from this context apply to other cell types and experimental setups. To begin to answer this question, we analyzed data from a study on retinal pigment epithelium (RPE) cells containing the EKAREN5 FRET-based ERK reporter [53]. In this study, RPE cells were not serum-starved and then treated with growth factors, but rather were asynchronously cycling in full growth medium.…”

“…MCF10A datasets used in this study were the dual-reporter datasets presented in a previous paper [27], and are freely available at the repository mentioned therein. The RPE datasets were produced independently [53], and were obtained by contacting the authors.…”

“…The RPE experiments (part of a larger study [53]), in addition to using a different cell line, only measured ERK activity with the EKAREN5 reporter (and not Akt, or with a KTR reporter), had cells asynchronously cycling in full growth media (not initially serum-starved), and included measurements over a 4 day period with a 10 minute frequency and multiple division events along the time course (versus 2 days, 15 minute frequency, and the first division event in the MCF10A datasets). Key differences are summarized in Table 1.…”

Low-frequency ERK and Akt activity dynamics are predictive of stochastic cell division events

“…Moreover, the previous study only considered a single cell line (MCF10A), a single set of growth factors (EGF and insulin), and acute, synchronized response to growth factors (as opposed to chronic, asynchronous scenarios). In the current work, we analyzed the previously generated dataset by utilizing time series transformations to extract features [47][48][49] , and predicting cell divisions using effective machine learning techniques [50][51][52] for both MCF10A cell data with acute response to growth factors from a serum-starved state, and retinal pigment epithelial (RPE) cells asynchronously cycling in full growth media 53 . Models developed from only MCF10A data performed well on both datasets, and analyses suggested operational relationships between ERK and Akt dynamics and the probability of cell division.…”