Inhibition of VEGFR2 Activation and Its Downstream Signaling to ERK1/2 and Calcium by Thrombospondin-1 (TSP1): In silico Investigation

Abstract

VEGF signaling through VEGFR2 is a central regulator of the angiogenic response. Inhibition of VEGF signaling by the stress-induced matricellular protein TSP1 plays a role in modulating the angiogenic response to VEGF in both health and disease. TSP1 binding to CD47 inhibits VEGFR2 activation. The full implications of this inhibitory interaction are unknown. We developed a detailed rule-based computational model to inquire if TSP1-CD47 signaling through VEGF had downstream effects upon ERK1/2 and calcium. Our Simulations suggest that enhanced degradation of VEGFR2 initiated by the binding of TSP1 to CD47 is sufficient to explain the inhibition of VEGFR2 phosphorylation, calcium elevation, and ERK1/2 activation downstream of VEGF. A complementary mechanism involving the recruitment of phosphatases to the VEGFR2 complex with consequent increase in the rate of receptor dephosphorylation may augment the inhibition of the VEGF signal. The model was then utilized to simulate the effect of inhibiting external TSP1 or the depletion of CD47 as potential therapeutic strategies in restoring VEGF signaling. Results suggest that depleting CD47 is a more efficient strategy in inhibiting the effects of TSP1/CD47 on VEGF signaling. Our results highlight the utility of in silico investigations in elucidating and clarifying molecular mechanisms at the intersection of TSP1 and VEGF biology and in differentiating between competing pro-angiogenic therapeutic strategies relevant to peripheral arterial disease (PAD) and wound healing.

Keywords: CD47; ERK1/2; TSP1; VEGF; VEGFR2; calcium; computational modeling.

Figure 1

Rules and pathways for VEGF signaling to calcium and ERK1/2. (A) Seed species used as input into BioNetGen including VEGF, VEGFR1, VEGFR2/CD47, NRP1, and TSP1. (B) Schematic representation of the rules for VEGF-mediated receptor dimerization, VEGF interaction with NRP1, (C). TSP1 binding to CD47 on the cell surface, (D). Ligand-independent VEGFR1 binding to NRP1 and multi-complex formation, VEGF-mediated heterodimer formation between VEGFR1 and VEGFR2. Binding of CD47 to VEGFR2 to form a preassociated complex, (E). Ligand-independent dimer formation resulting in VEGFR2/VEGFR2, VEGFR1/VEGFR1, and VEGFR1/VEGFR2 heterodimers. The rule for autophosphorylation is also shown, (F). Internalization of the complexes that contain VEGF/VEGFR2/VEGFR2 homodimers. (G) Signal transduction pathway from the phosphorylated tyrosine to ERK1/2 and calcium incorporating SphK1 feedback and translocation utilizing CIB1, (H). Calcium cycling module incorporating IP3-dependent release from the ER, CRAC, and calcium regulation with plasma membrane and SERCA pumps.

Figure 2

Model parameterization. (A) Total VEGFR2 (R2) computed as fraction total versus time fitted to two sets of experimental data from Bruns et al. (2010) and Ballmer-Hofer et al. (2011) (red and black circles), (B). R2 versus time with (blue) and without (black) NRP1 fitted to the data in Ballmer-Hofer et al. (2011) (red circles), (C). Surface level of VEGFR2 constrained with two sets of experimental data from Ewan et al. (2006) and Bruns et al. (2010) (red and black circles). (D) Fraction of phosphorylated receptors (pR2) is fitted to the experimental data in Chabot et al. (2009) (red circles). (E) Phosphorylated PLCγ from the model fitted to the data in Chabot et al. (2009) (red circles). (F) Normalized calcium transient from the model fitted to the normalized experimental data in Li et al. (2011) (red circles). (G) Raw calcium trace corresponding to the normalized transient in (F). (H) pERK1/2 fit to the data in Chabot et al. (2009) (red circles). (I) SphK1 inhibition abolishes pERK1/1 consistent with the data in Shu et al. (2002) (red versus blue curves). (J) Binding of VEGF to the binding sites on the cell surface versus VEGF concentration (Bikfalvi et al., 1991), (K). Dose response curve for pR2 is fitted to the experimental data in Whitaker et al. (2001) (red circles), (L). Dose response curve for the activation of ERK1/2 is fitted to two sets of data from Wijelath et al. (2006) and Yu et al. (2015) (red and black circles). The inset indicates sudden jump in pERK1/2 as VEGF in increased above threshold.

Figure 3

TSP1 inhibition of VEGFR2 signaling with enhanced VEGFR2 degradation. Two nanometer TSP1 is added for 10 min followed by 40 min of 50 ng/ml VEGF. (A–F) pR2 (A,B), calcium (C,D), and pERK1/2 (E,F) response as VEGFR2 degradation is increased. ERK1/2 exhibits threshold behavior in response to receptor degradation rate (E). (G) Maximum pERK1/2 as a function of TSP1 concentration and the fold-change in VEGFR2 degradation in the presence of TSP1. The dark blue region is the parameter regime where no ERK1/2 activation is observed, (H). Maximum ERK1/2 as a function of CD47 level [# (number) per surface area] and the fold-change in VEGFR2 degradation with TSP1 present.

Figure 4

Enhanced dephosphorylation of VEGFR2 via TSP1-mediated phosphatase recruitment. (A,B) Max pR2 and pR2 at 10 min vs. the fold-change in VEGFR2 dephosphorylation rate for receptors containing TSP1 with sample traces, (C,D) Max [Ca] as a function of the fold decrease in the rate of receptor dephosphorylation and sample traces, (E,F) Maximum pERK1/2 vs. fold-change in dephosphorylation rate and sample traces showing the threshold behavior, (G–I) Max pERK1/2 as a function of enhanced VEGFR2 degradation (shown as fold-change relative to control with no TSP1) and the fold-change in receptor dephosphorylation rate for 2, 0.6, and 0.2 nM TSP. This identifys the parameter regime for effective ERK1/2 inhibition (dark blue regions). there is no inhibition with [TSP1] = 0.2 nM (I). (J) Max pERK1/2 as a function degradation and dephosphoryation rates with 50% reduction in CD47 levels and 2 nM TSP1.

Figure 5

CRAC channel inhibition and VEGF signaling to ERK1/2 and calcium. (A) Max Ca versus CRAC channel amplitude, (B). Ca²⁺ traces for different CRAC channel current values. (C) Max pERK1/2 vs. current showing no appreciable effect. (D) pERK1/2 traces showing prolongation of pERK1/2 signal in response to increase in the current. (E) Max pERK1/2 and pERK (at 10 min) as a function of the time constant for the activation of the CRAC current. (F) pERK1/2 traces for five different values of the time constant.

Figure 6

TSP1 and CD47 inhibition and the restoration of VEGF signaling. (A–D) TSP1 inhibition and the recovery of the calcium (A,B) and pERK1/2 signal indicating 72% inhibition threshold for effective recovery (C,D). TSP1 inhibition involves 10-fold increase in VEGFR2 degradation and 51-fold increase in dephosphorylation relative to the case with no TSP1. (E–H) VEGF signal recovery by CD47 depletion. (E,F) intracellular calcium recovery as CD47 is depleted along with sample traces, (G,H) ERK1/2 activation is restored when 18% of CD47 is depleted. Sample traces are also shown.