Mechanisms of vascular morphogenesis and stabilization by VEGF dose

VEGF is the master regulator of angiogenesis and the most investigated factor in therapeutic angiogenesis approaches to induce the growth of new vessels that could restore oxygen and nutrients supply in tissues affected by ischemia. Several VEGF-based therapies have been tested in clinical trials, but, despite a good safety profile, they could not prove therapeutic efficacy because of several issues with the VEGF gene delivery approaches applied, among which the dose and duration of expression are key parameters that define the narrow therapeutic window of VEGF. Taking advantage of a cell-based ex-vivo approach to gene delivery, we previously found that VEGF can induce either normal capillaries or aberrant angioma-like structures depending strictly on its dose in the microenvironment around each producing cell in vivo, and not on the total amount. However, stimulation of pericyte recruitment by co-expression of PDGF-BB could increase the number of pericytes enrolled to the area of neo-angiogenesis and prevent the aberrant angiogenic phenotype induced by high VEGF levels alone. Moreover, VEGF-sustained delivery of 4 weeks was required to maintain newly induced normal capillaries upon VEGF-signaling abrogation, whereas aberrant vascular structures never became VEGF-independent.
Here we sought to further investigate the cellular and molecular mechanisms regulating 1) the switch between normal and aberrant angiogenesis and 2) the achievement of vascular stabilization in the presence of increasing VEGF doses, in order to identify novel and potentially more specific molecular targets to improve both the safety and the efficacy of VEGF-based strategies for therapeutic angiogenesis.
The point 1) aimed to identify the role of the specific endothelium-pericyte signaling pathways such as TGF-?1/TGF?R, Angs/Tie2 and ephrinB2/EphB4 in the VEGF dose-dependent transition between normal and aberrant angiogenesis. A monoclonal population of transduced myoblasts, expressing a moderate VEGF dose that induces only normal angiogenesis, was further transduced to secrete soluble blockers of the TGF?-1/TGF?-R, Tie2/Angiopoietin or EphB4/EphrinB2 pathways (LAP, sTie2Fc and sEphB4, respectively). Two weeks after implantation into mouse skeletal muscles, neither TGF? nor Angiopoietin blockade altered the normal angiogenesis induced by low VEGF, whereas EphrinB2/EphB4 inhibition caused a switch to aberrant angioma-like structures, similar to the effects of blocking pericyte recruitment. Conversely, gain-of-function of EphB4 signaling by systemic treatment with recombinant EphrinB2-Fc completely prevented aberrant angiogenesis induced by high VEGF levels and yielded a physiological network of normal and mature capillaries. We recently found that VEGF over-expression in muscle induces angiogenesis without sprouting, but rather by circumferential enlargement followed by longitudinal splitting. Analysis of the initial stage of vascular induction (4 days) showed that EphB4 inhibition did not interfere with pericyte recruitment, contrary to high VEGF alone. However, it increased both endothelial proliferation and the diameter of initial enlargements induced by low VEGF, leading to a failure of splitting and progressive angioma growth. In conclusion, these results suggest that ephrinB2/EphB4 signaling plays a key role in controlling the switch between normal and aberrant angiogenesis by increasing VEGF doses, independently of pericytes recruitment.
The point 2) aimed to define whether the stabilization of newly formed vessels is regulated by the dose of VEGF over-expression in a therapeutic setting and to investigate the underlying cellular and molecular mechanisms. For this purpose, monoclonal populations of transduced myoblasts, homogeneously expressing increasing VEGF doses, were implanted in mouse muscles and VEGF signaling was abrogated after 10 and 17 days by VEGF-TrapR1R2 treatment. Stabilization of new vessels was fastest with low VEGF, but was delayed or prevented by progressively higher doses. Unexpectedly, this was not due to differences in pericyte coverage or functional perfusion. Rather, VEGF dose-dependently inhibited endothelial expression of Semaphorin3A, leading to impaired recruitment of CD11b+/Neuropilin-1+ monocytes (NEM), decreased tissue levels of Transforming Growth Factor-?1 (TGF-?1) and reduced endothelial SMAD2/3 activation. TGF-?1 itself stimulated further endothelial Sema3A expression, providing a positive feedback loop to amplify and maintain the stabilizing signals in low VEGF conditions. Therefore, VEGF impairs vascular stabilization in a dose-dependent manner without affecting pericyte recruitment, but rather by directly inhibiting the endothelial Sema3A/NEM/TGF-?1 paracrine loop. These findings suggest that co-delivery of Sema3A could accelerate vascular stabilization with therapeutic doses of VEGF, enabling short-term and safer therapies.