We would like to thank Prof. Serini for his comment on our paper. Our model, like the model by Prof. Serini and coworkers (the “Torino group”), assumes that endothelial cells secrete an autocrine chemoattractant, i.e., a chemoattractant to which the ECs respond themselves.

Serini comments on a passage in the introduction of our paper where we review putative chemoattractants that endothelial cells might secrete to produce cellular aggregates and vascular networks. The Torino group argued that VEGF-A was the short-range autocrine chemoattractant, since ECs express receptors for VEGF (VEGFR-2) chemotax towards sources of VEGF under favorable conditions, and secrete VEGFs.
Based on a recent study by Lee et al. we argued that VEGF secreted by endothelial cells is not likely involved in endothelial cell aggregation and angiogenesis. Lee et al. studied angiogenesis in a genetically engineered line of mice in which the VEGF gene is specifically inactivated in endothelial cells (VEGF-ECKO, for VEGF endothelial cell knock-out). Of these mice 31.6 % died in utero, 11.5% died as young newborns a further 26.1% died between 4 weeks and 6 months. In the latter group and in mice surviving into adulthood, Lee et al. found severe vascular and cardiac defects.

The VEGF-ECKO transgenic mice had normal vascular density, endothelial fenestration and permeability, and interestingly, they showed normal angiogenic response in subcutaneously injected Matrigel containing VEGF, suggesting that angiogenesis is unaffected. Lee et al. found that fewer VEGF-ECKO cells than wild-type cells survived in cell cultures, and showed evidence that this is due to enhanced apoptosis rather than reduced proliferation. By studying the response of the VEGF-ECKO-ECs to hypoxia-simulating agents they concluded that autocrine VEGF signals promote survival-mechanisms, preventing apoptosis. Interestingly, this seems to be an entirely cell-autonomous effect, with VEGF activating VEGFR2 intracellularly (note that Lee et al. use the term autocrine signaling to indicate cell autonomous signaling, where both signal and receptors stay within the cell, while we used it to indicate signaling to the cell itself (and to neighboring cells) via a secreted signal and extracellular receptors, with the signal exiting the cell and binding to the same cell). Based on their findings, Lee et al. conclude that “autocrine VEGF does not contribute to the angiogenic response, as vascular density and patterning were virtually identical between control and VEGF-ECKO mice” and then “together these findings indicate that cell-autonomous signaling triggers a response that does not fully overlap with the events initiated by paracrine acitivation.” Thus, the endothelial VEGF required for endothelial survival does not leave the cells; the portion of VEGF that does leave the endothelial cells (if any: note that Lee et al. did not find a significant reduction in circulating VEGF in VEGF-ECKO mice) does not matter for the angiogenic cascade: angiogenesis is severely affected in full VEGF knock-outs, but progresses normally in VEGF-ECKO mice where only ECs do not produce VEGF.

What about those 31.6 % VEGF-ECKO mice dying in utero? The VEGF-gene is excised in only 95% of the VEGF-ECKO endothelial cells and indeed VEGF-ECKO still produce a little bit of VEGF under simulated hypoxia. Would it be possible that in those mice dying early in development EC-specific VEGF excision was more successful, so to crucially disrupt de novo EC assembly and angiogenic sprouting? I could not say for sure, but the Lee et al. findings suggest that early ruptures, cardiac failures, intestinal perforations and microinfarcts would be the most likely explanation for poor survival of VEGF-ECKO mice.

We would like to thank Prof. Serini again for taking the time to comment on our paper, and we would welcome further discussion on the role of VEGF secretion by endothelial cells during de novo and sprouting blood-vessel growth.