The clinical success of neutralizing vascular endothelial growth factor (VEGF) has unequivocally identified VEGF as a driver of retinal edema that underlies a variety of blinding conditions. VEGF is not the only input that is received and integrated by the endothelium. For instance, the permeability of blood vessels is also regulated by the large and ubiquitously expressed transforming growth factor beta (TGF-β) family. In this project, we tested the hypothesis that members of the TGF-β family influence the VEGF-mediated control of the endothelial cell barrier. To this end, we compared the effect of bone morphogenetic protein-9 (BMP-9), TGF-β1, and activin A on the VEGF-driven permeability of primary human retinal endothelial cells. While BMP-9 and TGF-β1 had no effect on VEGF-induced permeability, activin A limited the extent to which VEGF relaxed the barrier. This activin A effect was associated with the reduced activation of VEGFR2 and its downstream effectors and an increased expression of vascular endothelial tyrosine phosphatase (VE-PTP). Attenuating the expression or activity of VE-PTP overcame the effect of activin A. Taken together, these observations indicate that the TGF-β superfamily governed VEGF-mediated responsiveness in a ligand-specific manner. Furthermore, activin A suppressed the responsiveness of cells to VEGF, and the underlying mechanism involved the VE-PTP-mediated dephosphorylation of VEGFR2.
The therapeutic benefit provided by anti-vascular endothelial growth factor (VEGF) for patients with vision-threatening conditions such as diabetic retinopathy (DR) demonstrates the important role of VEGF in this affliction. Cytokines, which can be elevated in the vitreous of patients with DR, promote leakage of retinal blood vessels, and may also contribute to pathology, especially in those patients for whom anti-VEGF does not provide adequate benefit. In this in vitro study using primary human retinal endothelial cells, we compared anti-VEGF with the (transforming growth factor beta) TGFβ receptor inhibitor RepSox (RS) for their ability to enforce barrier function in the face of VEGF, cytokines, and the combination of both. RS was superior to anti-VEGF because it prevented permeability in response to VEGF, cytokines, and their combination, whereas anti-VEGF was effective against VEGF alone. The inhibitory effect of RS was associated with suppression of both agonist-induced pore formation and disorganization of adherens junctions. RS-mediated inhibition of the TGFβ pathway and increased expression of claudin-5 did not adequately explain how RS stabilized the endothelial cell barrier. Finally, RS not only prevented barrier relaxation, but also completely or partially reclosed a barrier relaxed with tumor necrosis factor α (TNF α) or VEGF, respectively. These studies demonstrate that RS stabilized the endothelial barrier in the face of both cytokines and VEGF, and thereby identify RS as a therapeutic that has the potential to overcome permeability driven by multiple agonists that play a role in the pathology of DR.