BioModels Database logo

BioModels Database

spacer

Schilling et al., (2009). Theoretical and experimental analysis links isoform-specific ERK signalling to cell fate decisions.

February 2011, model of the month by Benedetta Frida Baldi
Original model: BIOMD0000000270

The ERK signaling pathway is one of the most studied network, due to its major implications in cell proliferation, differentiation and survival; deregulation of this pathway has being linked to several diseases and cancer types.

In the MAP kinase cascade there are two pairs of highly related isoforms: MEK1 and 2 (dual-specificity mitogen activated protein kinases) that share 80% of sequence identity, and ERK1 and 2 (mitogen activated protein kinases) that share 85% of sequence identity. While knockout mice for MEK1 and ERK2 isoforms are lethal, they are available for MEK2 and ERK1. This suggest that these isoforms has different activities even if most experimental results suggest that they have redundant functions.

External signals that are able to activate the ERK pathway are various growth factors , mitogen , cytokines and G protein-coupled receptors (GPCRs). One of these factors is Erythropoietin, a 193 amino-acid long hormone involved in the regulation of erythrocyte differentiation, and the maintenance of a physiological level of circulating erythrocyte mass. Erythropoietin (Epo) interacts with its receptor EpoR , a homodimeric cytokine receptor on the plasma membrane. Epo induces conformational changes in the receptor that subsequently allow the activation of Janus protein tyrosine kinase 2 (JAK2) , which is coupled to the proximal part of the receptor in the cytoplasmic region.

JAK2 after being activated by autophosphorylation, phosphorylates eight tyrosine residues in the distal part of cytoplasmic region of EpoR. The phospho-tyrosines on EpoR act as a docking sites for intracellular proteins such as SHC2. Epo also activate the JAK-STAT , PI3K-AKT and PKC pathways.

After Epo induced stimulation, the signal is silenced by two major mechanisms: dephosphorylation of JAK2 and EpoR mediated by the phosphatase SH1, and by down-regulation of the receptor via ubiquitination. The pathway is further regulated by negative feedback from the double phosphorylated form of ERK1/2 to SOS. Schematic representation of the Epo pathway is shown in Figure 1.

Figure 1

Figure 1: Schematic representation of the Epo pathways. Figure taken from [1].



Figure 2

Figure 2: Mathematical model. A) Process diagram of the model. B) Trajectories of the model variables. Experimental data depicted as open circles; for ERKs and MEKS processive (dashed line) or distributive ( solid line) are shown. Figure taken from [2].

In this work, Shilling and colleagues used primary cell data instead of tumor-derived cell lines. This should reduce the risk of artefacts due to the prevalent deregulation of ERK signaling pathway in cancer cells, and provides real physiological data.

In order to estimate kinetic parameters for the model, they acquired time-resolved data from primary murine erythriod progenitor (CFU-E) cells under continuous stimulation with Erythropoietin. Quantitative immunoblotting were performed for 29 different time points up to 70 minutes post-stimulation, with 50 U/ml of Epo to quantify the expression of phosphorylated JAK2, phosphorylated EpoR, double phosphorylated MEK1 and 2, double phosphorylated ERK1 and 2, phosphorylated SOS and the sum of mSOS and SOS (Figure 2B).

The number of Epo Receptor expressed on the membrane was calculated with saturation-binding assays using radio-labelled Epo. Mass spectrometry analysis was performed to determine the fraction of mono and double phosphorylated ERK isoforms, using a label free approach on a nano-electrospray mass spectrometer coupled with an Ultra-Performance Liquid Chromatography (UPLC) column (Figure 3A).

The model consists of 40 irreversible reactions following the mass action law. As shown in figure 2, the Epo Receptor (EpoR) is in a stable complex with the Janus protein kinase JAK2, that is activated by EpoR phosphorylation due to the binding of its ligand, Epo. Activation of SHC2, Grb2 and then SOS is simplified in a single step of SOS activation. This results in direct activation of SOS by JAK2.



Once SOS is active, it promotes the phosphorylation of Ras and then Raf. Again this has been merged into a single step leading from SOS to Raf. The active form of Raf activates the phosphorylation of MEK1/2, which is subsequently able to activate ERK1/2.

The double phosphorylated form of ERK1/2 inhibits SOS activity with a negative feedback loop. Ligand dependent down-regulation of the receptor and dephoshorylation are summarised with a SHP1 mediated dephosphorylation reaction of JAK2, which has been associated a delay.

Figure 3

Figure 3: Mass spectrometry validation of the distributive model of ERK1/2 activation. A) Model prediction. B) UPLC-ESI-MS analysis. Figure taken from [2].

Figure 4

Figure 4: Bow-tie architecture of the signaling network. Activated molecule abundance shown in grey, in dashes lines is shown the standard deviation of the model prediction. Figure taken from [2].



Figure 5

Figure 5: Effect of over-expression of ERK1 or ERK2 on the signalling pathway. Model trajectories are shown as solid lines and experimentally quantified phosphorylation level of ERK1/2 are shown as open symbols. Figure taken from [2].

Two different models for the activation of MEK and ERK have been proposed in this paper. One is a single-step processive model, where the double phosphorylation on the protein kinases is treated as a single step. The second one is a two-step distributive model where the two phosphorylations are treated in two separate reactions (Figure 2A).

The processive model failed to fit the experimental data and so has been discarded. Nevertheless, MEK activation can still be described as a processive model, since in the distributive form the first phosphorylation result is three-fold slower for MEK1, and two-fold slower for MEK2 than the second one. As mass-spec data shows, ERK's isoforms data fits well with the distributive model (Figure 3), where single and double phosphorylated forms of the kinase are detected.

Analysing the maximum number of activated molecule and signal amplifications, the model shows a bow tie architecture. EpoR activates the signal cascade through SOS and Raf only weakly, whereas Raf and MEKs then massively amplify the signal (Figure 4). This type of structure remarkably points out the fragility point of the cascade, the knot in the bow tie, that in this case is the activation of Ras and subsequently Raf by SOS. This is an evidence which agrees with the well-established role of Ras and Raf in carcinogenesis and could help provide a more target based approach in drug discovery.

To understand the dynamics of the system, and the role of the two different isoforms of ERK, overexpression experiments have being performed both in vitro and in silico to analyse ERK1/2 phosphorylation patterns. Due to the strong negative feedback loop from ERK to SOS when ERK1 levels are elevated, Raf and MEK activation is reduced and the signal amplification of this cascade ultimately reduces the fully activated ERK2 levels. In other words, when ERK1 is overexpressed, the expression of double phosphorylated ERK1 also increases but this reduces the amplitude of double phosphorylated ERK2 (Figure 5).



Other in vitro and in silico experiments have been carried out to link ERK response to different initial condition, such as protein concentration and cell proliferation. Interestingly, both approaches show that increasing the level of ERK1/2 reduced proliferation and increased differentiation. Moreover, extrapolating the contribution of the two different isoforms to proliferation suggests that ERK2 promotes proliferation, whereas ERK1 inhibits it in case of ERK2 hyperactivity (Figure 6).

In conclusion this model provide new insights into Epo regulation on cell fate decisions and elucidate the different role of the two ERK’s isoform. This work is also a perfect example of the system biology cycle, where experimental data and mathematical modeling interacts constantly in refinements and improvements cycles.

Figure 6

Figure 6: Positive and negative contribution to average proliferation of double-phosphorylated ERK1/2 (solid lines) with confident intervals (dashed lines). Figure taken from [2].

Bibliographic References

  1. Szenajch, J., Wcislo, G., Jeong, J.Y., Szczylik, C., Feldman, L. The role of erythropoietin and its receptor in growth, survival and therapeutic response of human tumor cells From clinic to bench - a critical review. Biochim Biophys Acta. , 1806(1):82-95, 2010. [CiteXplore] .
  2. Schilling, M., Maiwald, T., Hengl, S., Winter, D., Kreutz, C., Kolch, W., Lehmann, W.D., Timmer, J., Klingmüller, U. Theoretical and experimental analysis links isoform-specific ERK signalling to cell fate decisions. Mol Syst Biol. , 5:334, 2009. [CiteXplore]
spacer
spacer