Computational investigation of sphingosine kinase 1 (SphK1) and calcium dependent ERK1/2 activation downstream of VEGFR2 in endothelial cells

Abstract

Vascular endothelial growth factor (VEGF) is a powerful regulator of neovascularization. VEGF binding to its cognate receptor, VEGFR2, activates a number of signaling pathways including ERK1/2. Activation of ERK1/2 is experimentally shown to involve sphingosine kinase 1 (SphK1) activation and its calcium-dependent translocation downstream of ERK1/2. Here we construct a rule-based computational model of signaling downstream of VEGFR2, by including SphK1 and calcium positive feedback mechanisms, and investigate their consequences on ERK1/2 activation. The model predicts the existence of VEGF threshold in ERK1/2 activation that can be continuously tuned by cellular concentrations of SphK1 and sphingosine 1 phosphate (S1P). The computer model also predicts powerful effects of perturbations in plasma and ER calcium pump rates and the current through the CRAC channels on ERK1/2 activation dynamics, highlighting the critical role of intracellular calcium in shaping the pERK1/2 signal. The model is then utilized to simulate anti-angiogenic therapeutic interventions targeting VEGFR2-ERK1/2 axis. Simulations indicate that monotherapies that exclusively target VEGFR2 phosphorylation, VEGF, or VEGFR2 are ineffective in shutting down signaling to ERK1/2. By simulating therapeutic strategies that target multiple nodes of the pathway such as Raf and SphK1, we conclude that combination therapy should be much more effective in blocking VEGF signaling to EKR1/2. The model has important implications for interventions that target signaling pathways in angiogenesis relevant to cancer, vascular diseases, and wound healing.

Fig 1. Rules for constructing the SphK-1 dependent ERK1/2 pathway.

A. Species with domain structures labeled. VEGF with three binding sites, two for binding to the receptors and a C-terminal binding domain for binding to NRP1. VEGFR2, with ligand binding site (L), receptor coupling site (C), and the tyrosine in the position 1175 (Y1175), VEGFR1 with ligand binding domain (L), receptor coupling site (C), and the NRP1 binding site (NRP1bd). NRP1 with the ligand binding site and the VEGFR2 binding site labeled (R1bd/L), B. The rule for the interaction of VEGF and VEGFR and subsequent dimerization included, C. The rule for ligand-independent interaction of VEGFR1 and NRP1, D. Ligand-independent dimerization of the receptors through the coupling sites (top). Ligand-dependent phosphorylation of the Y1175 on VEGFR2 (bottom), E. Receptor-complexes undergo internalization and degradation from the plasma membrane, F. Signaling pathway for the activation of ERK1/2 downstream of phosphorylated Y1175 (pY1175), G. Elements of the calcium-cycling module of the model. Activated PLCγ results in the generation of IP3 that then diffuses and binds to and activates IP3 sensitive calcium release channels on the ER membrane. The decline in ER calcium concentration results in the opening of the calcium release activated calcium (CRAC) channels.

Fig 2. Model fit to the experimental data.

A. Total VEGFR2 fraction from the model (blue, solid) fitted to the experimental data in [38] (red circles) and [39] (black circles), B. Total VEGFR2 level in the absence of NRP1 from the model (blue) is fitted to the experimental data in [39] (black circles). The control curve from the model (in the presence of NRP1) is shown (black), C. Surface VEGFR2 fraction from the model (blue) is fitted to the experimental data from [40] (red circles) and [38] (black circles), D. Normalized total phosphorylated VEGFR2 from the model is fitted to the data of Chabot et al. [41], E. Predicted fractional surface level of pVEGFR2 from the model, F. Predicted endosomal pVEGFR2, G. pPLCγ from the model is fitted to the normalized experimental data from [41] (red circles), H. Normalized calcium transient from the model (blue) fitted to the data in [36] (red circles), I. Raw calcium transient from the model, J. Phosphorylated pERK1/2 from the model is fitted to the experimental data from [41] (red circles), K. Inhibiting SphK1 in the model (red line) blocks ERK1/2 activation. Experimental data point from endothelial cells at t = 10 min is also shown [11] (red circle). Control pERK1/2 curve is also shown (blue), L. Inhibiting PKC blocks ERK1/2 activation in the model (red line) with the experimental data point at t = 10 min shown [11] (red circle). Control pERK12 is shown (blue), M. Bound VEGF from the model (blue) is fitted to the experimental data from [42] (red circles), N. Dose response curve for pVEGFR2 from the model (blue) fitted to the data in [43] (red circle), O. Maximum value of pERK1/2 from the model (blue) versus VEGF concentrations fitted to the experimental data in [44] (black circles) and [45] (red circles), P. Activated SphK1(blue) and S1P (red).

Fig 3. Global sensitivity analysis and VEGF threshold response of pERK1/2.

A. Top 15 parameters with positive correlation according to the PRCC method. The top parameter is the total concentration of Raf. B. Top 15 parameters with negative correlation according to PRCC. The Michaelis-Menten-type parameter for the activation of MEK1/2 by Raf exhibits the highest negative correlation. C. Response of maximum fractional pERK1/2 relative to VEGF, with different SphK1 concentrations. The threshold value of VEGF is lowered in response to increasing SphK1 levels. D. VEGF threshold value is a decreasing function of the SphK1 concentrations, monotonically decreasing relative to SphK1. E. Sample curves for the time course of pERK1/2 for different values of SphK1 concentration. F. SphK1 catalytic rate modulates the VEGF threshold for ERK1/2 activation. G. There is a monotonically decreasing relation between the threshold value of VEGF and kcat_SK1. H. Sample pERK1/2 versus time curves for different values of kcat_SK1. I. The effect of the strength of S1P dephosphorylation on the threshold value of VEGF. J. VEGF threshold increases in response to increasing dephosphorylation rate of S1P, indicating an approximately linear response. K. Sample curves of pERK1/2 versus time in response to different values of kdpS1P.

Fig 4. The effects of perturbations in calcium cycling on pERK1/2 dynamics.

A. PMCA inhibition results in sustained pERK1/2 curve (blue), while the maximum value of pERK1/2 is unchanged. The plateau phase of pERK1/2 is progressively prolonged as the pump rate is reduced from the base value (pink, multiple of one), B. pERK1/2 versus time curves for five different values of the SERCA pump rate. C. The effect of varying CRAC channel current amplitude on the shape of the pERK1/2 versus time curve, illustrating the prolongation of the pERK1/2 signal, D. The changes in maximum pERK1/2 as a function of variations in total concentration of CIB1. There is rapid decline of pERK1/2 signal for CIB1 concentrations below ~80 nM. E. pERK1/2 versus time in response to CIB1 concentration, F. The dissociation constant for the binding of Ca²⁺/CIB1 to SphK1 starkly influences the maximum value of pERK1/2, G. pERK1/2 for different values of the dissociation constant, H. The maximum pERK1/2 as a function of the translocation rate of CIB1/SphK1 to the membrane (kt_SK1), I. pERK1/2 versus time in response to changes in kt_SK1.

Fig 5. Relation between pERK1/2 and pVEGFR2 and its modulation by SphK1 pathway.

A. pERK1/2 versus the phosphorylation rate of the receptor, B. Maximum pERK1/2 versus pVEGFR2 demonstrates the existence of a threshold value of pVEGFR2 below which there is no ERK1/2 activation. The critical value is 0.046, C. pVEGFR2 threshold is decreased in response to increases in SphK1 concentration. Increasing SphK1 to 1 μM from the baseline value of 0.1 μM reduces the threshold from 0.046 to 0.008, D. pVEGFR2 threshold value is sharply increased by increasing S1P dephosphorylation rate, from 0.008 to 0.16, E. Maximum active SphK1 versus max pVEGFR2 curve with similar threshold behavior, F. Maximum S1P shows monotonically increasing behavior relative to pVEGFR2 for pVEGFR2 values above the threshold value of 0.046, G. Maximum pERK1/2 versus percent inhibition of VEGF. The inset shows the threshold region: 99% sequestration of VEGF is necessary before ERK1/2 activation is blocked, H. Maximum pERK1/2 versus percent inhibition of the receptor phosphorylation. Inset shows the threshold region, I. Simulating the effect of VEGFR2 depletion with a mAb VEGFR2 inhibitor (e.g., ramucirumab) on pERK1/2, J. Simulating the effect of a generic multi-target tyrosine kinase inhibitor (e.g., sorafenib) on pERK1/2 simultaneously blocking VEGFR2 phosphorylation and Raf activation, K. Combined effect of sorafenib and ramucirumab demonstrating enhanced ERK1/2 inhibition, L. Combined effect of sorafenib, ramucirumab, and SphK1 inhibition on pERK1/2.