Padala et al., (2017). Cancerous perturbations within the ERK, PI3K/Akt, and Wnt/β-catenin signaling network constitutively activate inter-pathway positive feedback loops.

September 2017, model of the month by Emma Louise Fairbanks
Original models: BIOMD0000000648, BIOMD0000000652, BIOMD0000000653, BIOMD0000000654, BIOMD0000000655 and BIOMD0000000656.

Background

Cellular networks are usually robust and can maintain normal dynamics when subjected to many perturbations and noises. Nevertheless they show extreme sensitivity to certain perturbation of components involved in crosstalk between different signaling pathways. Such perturbations are often cancerous and result in constitutive activation of inter-pathway feedback loops, thereby magnifying the effect on cancer growth and proliferation. Mathematical models of signalling crosstalk in cancer can identify hidden mechanisms that can be experimentally tested and offer insight to further findings. In this study [1], Padala et al. (2017) developed a model to study crosstalk between the ERK, Akt and Wnt signalling pathways in cancerous conditions.

Mutations in intracellular signalling cascade proteins are found in most human tumours. The ERK pathway is reported to be dysregulated in approximately 30% of all human cancers [2]. Therefore, this pathway has been the focus of many studies. Epidermal growth factor (EGF) binds to its receptor (EGFR) starting a chain of reactions which result in Ras activation, initiating the activation of the MAPK cascade (Raf->MEK->ERK). Once ERK is double phosphorylated (ppERK) it can translocate into the nucleus to phosphorylate other signalling molecules and transcription factors that participate in the regulation of proliferation.

EGFR is also responsible for the activation of the Akt signalling. Upon binding to its ligand the receptor recruits PI3K which phosphorylates PIP2 resulting in PIP3; this reaction is negatively regulated by PTEN. Akt is then translocated to the plasma membrane by PIP3, where it is phosphorylated. Akt inhibits GS3Kβ and promotes cell cycle progression. GS3Kβ phosphorylates β-catenin which causes β-catenin to be degraded; otherwise, β-catenin can bind to free T-cell factor (TCF) resulting in activation of Wnt target genes which drives proliferation. Mutations can affect these signalling pathways and induce uncontrolled cell proliferation leading to cancer. For example Raf mutation can introduce positive feedback loop in ERK and Wnt/β-Catenin pathway (Figure 1a). Similaryly, PTEN mutation results in constitutive activation of Akt signalling which in turns induces sustained activation of Wnt/β-Catenin pathways (Figure 1b). Padala et al [1],used mathematical model to investigate the effect of the common cancerous mutations (Table 1) including bRaf and PTEN mutations.

The model

Crosstalk models are often derived from combinations of single pathway models or smaller crosstalk models. Padala et al. (2017) [1] recently proposed a model comprising of 64 reactions, 54 nodes and 103 parameters, describing the crosstalk in three signalling networks; viz. ERK, Akt and Wnt. This model was derived from Kim et al. (2007), Orton et al. (2009), and Brown et al. (2004) [2, 3, 4]. Curated models for these pathways can be found on the biomodels database with identifiers BIOMD0000000149, BIOMD0000000623 and BIOMD0000000033, respectively. Cancerous perturbations were introduced into the mathematical model to study the signalling dynamics in response to mutations and over-expression, thereby investigating how these perturbations influence other signalling pathways in the network.

There are many known crosstalks between the ERK, Akt and Wnt pathways. The key interactions include the following:

  • ppERK and pAkt phosphorylate and inhibit GSK3β
  • Akt is activated by EGFR.
  • GSK3β binds to RKIP.
  • Raf1 is phosphorylated by Wnt and β-Catenin/TCF through unknown mechanisms that lead to ERK activation.
  • ERK signalling is reduced by a negative feed-forward loop from pAkt to Raf1 and bRaf.
  • GSK3β inhibits the formation of PKC and therefore reduces activation of SOS.
Taking these crosstalks into consideration, the mathematical model was constructed using mass action and Michaelis-Menten kinetics in MatLab Simbiology toolbox. Padala et al. (2017) [1] then performed simulations on the model for common cancerous perturbations (Table 1).
Table 1: List of models from Padala et al. (2017) and the corresponding cancerous perturbation.

Model ID

Perturbation

BIOMD0000000648

normal condition

BIOMD0000000653

bRaf mutation

BIOMD0000000656

EGFR over expression

BIOMD0000000654

Ras mutation

BIOMD0000000653

PTEN mutation

BIOMD0000000652

PI3K mutation

Figure 1

Figure 1 A diagram summarising the network for a) mutated bRaf and b) mutated PTEN. This figure is the reproduction of figure 3H and 5H from [1]. The part of the pathways that are affected by the mutation are shown in red color.

Figure 2

Figure 2 shows the dynamics of ppERK, pAkt and β-Catenin/TCF under normal condition (BIOMD0000000648). ppERK and pAkt have transient responses, whereas β-Catenin/TCF depends on the presence of Wnt. Blue and black lines represent simulations performed in the absence (W=0) and presence (W=1) of Wnt, respectively. This figure is a reproduction of Figure 2 of [1].

Figure 3

Figure 3 Normalised protein concentration with bRaf mutation compared to normal condition. The plot was obtained by simulating BIOMD0000000653. This figure is a reproduction of figure 3 (d-f) of [1]. The blue lines represent the level of activation in the absence of a Wnt signal, black line represents the concentration of β-catenin/TCF in the presence of Wnt and the red line present the concentrations when bRaf is mutated.

Figure 4

Figure 4 Normalised protein concentrations with PTEN mutation compared to normal condition. The plot was obtained by simulating BIOMD0000000655. This figure is a reproduction of figure 5 (e-f) of [1]. The blue lines represent the level of activation in the absence of a Wnt signal (too small to be visible for Akt) and the red lines represent the concentrations when PTEN is mutated.

Results and Conclusions

In this study, Padala et al. (2017) [1] used mathematical modelling to understand the molecular mechanisms underlying signalling crosstalk and its implication in cancerous conditions. The model reproduced previously published dynamics of ERK, PI3K/AKT and Wnt/ β-Catenin signalling pathways (Figure 2) for the normal condition. Furthermore, Padala et al. (2017) also modified the models to induce common cancerous conditions, described in Table 1, and investigated the dysregulation in signalling pathways which facilitate uncontrolled cell proliferation.

bRaf is found to be mutated in 7% of cancers. In normal cells a negative feedback loop from Akt to Raf1 and bRaf regulates concurrent over-activation of the two pathways. It was found that when bRaf is mutated the activation of the Akt and ERK signalling pathways become disjoint events. Due to this mutation, bRaf is not naturally deactivated and the negative feed-forward loop from pAkt to bRaf is also inactivated. Therefore, bRaf remains active which sequentially results in the activation of MEK and ERK. Simulations in Figure 3 show this perturbation to have an undesirable effect on the concentrations of ppMEK, ppERK and β-Catenin/TCF which promotes excessive cell proliferation. The negative feedback loop, which would usually terminates the signal in the MAPK cascade, no longer exists and ppERKs constant activation causes β-Catenin/TCF to accumulate through GSK3β. Similar behaviour is observed when Ras is mutated since Ras activates bRaf.

When PTEN is mutated it can no longer catalyse the dephosphorylation of PIP3 to PIP2 and results in increased activation of Akt by PIP3. These mutations also increase the rate of β-catenin/TCF formation since active Akt inhibits GSK3β through phosphorylation. Model simulation also revealed that PTEN mutation affects the ERK pathway by reducing ERK activation. Unlike perturbations in the ERK pathway, the feed-forward loops from Akt to Raf1 and bRaf allow hyperactivation of Akt to reduce signalling of Raf1 and bRaf proteins and their downstream proteins MEK and ERK. Similar to PTEN mutation, when PI3K is mutated it is hyper-activated and leads to the accumulation of PIP3, therefore, similar behaviours are observed. In summary, this study revealed the common cancerous mutations facilitate uncontrolled proliferation, by switching the dynamics of ERK and Akt from transient to sustained activation and increasing the level of β-Catenin/TCF.

In silico studies are a useful tool to analyse and increase knowledge of the dynamics of signaling pathways and the complex interactions within them. Using computational simulation Padala et al. (2017) showed that a perturbation in a component involved in feedback loops can cause permanent activation of multiple pathways without the presence of the external stimuli. Activation of these pathways leads to increased proliferation and growth of cells resulting in cancer. Similar to the simulation of cancerous mutations, these models can also be used to investigate the effect of drug targeting these pathways and therefore will also be a useful tool for drug development research.

 

References

  1. Padala, R. R., Karnawat, R., Viswanathan, S. B., Thakkar, A. V., & Das, A. B. (2017). Cancerous perturbations within the ERK, PI3K/Akt, and Wnt/β-catenin signaling network constitutively activate inter-pathway positive feedback loops. NMolecular BioSystems, 13(5), –830-840.

  2. Kim, D., Rath, O., Kolch, W., & Cho, K. H. (2007). A hidden oncogenic positive feedback loop caused by crosstalk between Wnt and ERK pathways . Oncogene, 26(31), , 4571.

  3. Orton, R. J., Adriaens, M. E., Gormand, A., Sturm, O. E., Kolch, W., & Gilbert, D. R. (2009). Computational modelling of cancerous mutations in the EGFR/ERK signalling pathway. BMC Systems Biology, 3(1), 100.

  4. Brown, K. S., Hill, C. C., Calero, G. A., Myers, C. R., Lee, K. H., Sethna, J. P., & Cerione, R. A. (2004). The statistical mechanics of complex signaling networks: nerve growth factor signaling. Physical biology, 1(3), , 184.