Extracellular Signal-Regulated Kinases: One Pathway, Multiple Fates

Abstract

This comprehensive review delves into the multifaceted aspects of ERK signaling and the intricate mechanisms underlying distinct cellular fates. ERK1 and ERK2 (ERK) govern proliferation, transformation, epithelial-mesenchymal transition, differentiation, senescence, or cell death, contingent upon activation strength, duration, and context. The biochemical mechanisms underlying these outcomes are inadequately understood, shaped by signaling feedback and the spatial localization of ERK activation. Generally, ERK activation aligns with the Goldilocks principle in cell fate determination. Inadequate or excessive ERK activity hinders cell proliferation, while balanced activation promotes both cell proliferation and survival. Unraveling the intricacies of how the degree of ERK activation dictates cell fate requires deciphering mechanisms encompassing protein stability, transcription factors downstream of ERK, and the chromatin landscape.

Keywords: EMT; ERK; apoptosis; cell fate; cell proliferation; cell signaling; pluripotency; senescence.

Figure 1

Simplified schematic of the activation of the ERK pathway by ligands binding to membrane receptors. Receptors activate the GTPase RAS via Grb2 (Growth Factor Receptor–Bound Protein 2) and Sos (Son of Sevenless). The signaling module downstream of RAS includes a cascade of kinases formed by RAF, MEK, and ERK. The outcomes of the pathway depend largely on the phosphorylation of ERK targets that modulate protein activity and gene expression but also attenuate (negative feedback) or reinforce the pathway (positive feedback). (Adapted from BioRender).

Figure 2

Decoding ERK-signaling duration by integrating ERK-dependent immediate early genes (IEGs) transcription (early effect) with ERK-dependent IEG products stabilization (late effect). (a) IEG are induced by transcription factors phosphorylated by ERK, but they are subsequently degraded. (b) Sustained ERK activity stabilizes IEG by further phosphorylation events. These IEGs code for transcription factors that control cell cycle genes. This model was proposed by Blenis and colleagues [19]. Long-lasting ERK signals can also stabilize mRNA of late response genes [20], although the direct targets of ERK involved were not identified (adapted from BioRender).

Figure 3

Decoding ERK-signaling duration and intensity by multisite phosphorylation. In this model, originally described for Elk1 phosphorylation, prolonged ERK signals are required to phosphorylate low-affinity inhibitory sites that turn off early-acting transcription factors [21] (adapted from BioRender).

Figure 4

Decoding ERK oscillations by pulses of transcription factor activation. (a) Pulses of ERK activities detected by live-cell reporters [4,29,42] could be transmitted to effector molecules such as transcription factors [22]. (b) Modulating the frequency (# oscillation/time), amplitude and duration (the time it takes for a single oscillation) of ERK activity pulses can control different outcomes [4].

Figure 5

Decoding ERK-signaling intensity by cumulative ERK activation. The red line represents ERK activity. Fate 2 can be attained by a lasting ERK induction or the accumulation of the same signal after several pulses of activation (based on Johnson and Toettcher [52]). This model implies that ERK signaling is somehow remembered in cells receiving short pulses of ERK induction (green line). The nature of the memory signal may depend on the context. In the yeast pheromone pathway, the levels of the CDK inhibitor Far1 reflect an integral of the duration and concentration of past pheromone exposure [53]. Albeck and colleagues earlier proposed this decoding mechanism while studying the induction of the ERK target gene Fra-1. Using a single-cell approach based on fluorescent ERK reporters, they found that both the amplitude and the duration of ERK activity contributed to Fra-1 induction. This graded response apparently contrasts with the Blenis model, where lasting ERK activity was required to induce late-response genes in a switch-like fashion [54]. However, mathematical modeling of the network regulating Fra-1 expression shows that the basal activity of Fra-1 transcription determines whether the network will only respond to long-lasting stimulation as proposed by Blenis or as a signal integrator capable of remembering previous stimulation (adapted from BioRender).

Figure 6

Decoding ERK signal intensity by protein degradation. In this model, high-intensity ERK signals trigger the degradation of many ERK targets by coupling protein phosphorylation to ubiquitin-mediated protein degradation. Prolonged or intense ERK signals are required to increase the stoichiometry of phosphorylation, so subsequent degradation effectively reduces protein levels. This was described in senescent cells as SAPD (senescence-associated protein degradation) [70,75,76] but also in mouse ES cells that lose naïve pluripotency upon ERK activation (adapted from BioRender).

Figure 7

Non-monotonic relationship between ERK activity and cell proliferation. The Goldilocks effect. Proliferation requires moderate ERK activity [10], while differentiation [10], senescence [1,70,71] and apoptosis [79] require stronger signals. Aberrantly high ERK activation triggers protein degradation (SAPD) [70,75,76], changes in gene expression (GE) [77], and nucleolar stress (NoS) [78]. These processes regulate senescence and perhaps apoptosis. EMT reprograms the response to high-intensity ERK signals, acting downstream of ERK to inhibit senescence and apoptosis (Adapted from BioRender).