ARTICLE ABSTRACTOne of the main consequences of inhibition of neovessel growth and vessel pruning produced by angiogenesis inhibitors is increased intratumor hypoxia. Growing evidence indicates that tumor cells escape from this hypoxic environment to better nourished locations, presenting hypoxia as a positive stimulus for invasion. In particular, anti-VEGF/R therapies produce hypoxia-induced invasion and metastasis in a spontaneous mouse model of pancreatic neuroendocrine cancer (PanNET), RIP1-Tag2. Here, a novel vascular-targeting agent targeting semaphorin 4D (Sema4D) demonstrated impaired tumor growth and extended survival in the RIP1-Tag2 model. Surprisingly, although there was no induction of intratumor hypoxia by anti-Sema4D therapy, the increase in local invasion and distant metastases was comparable with the one produced by VEGFR inhibition. Mechanistically, the antitumor effect was due to an alteration in vascular function by modification of pericyte coverage involving platelet-derived growth factor B. On the other hand, the aggressive phenotype involved a macrophage-derived Sema4D signaling engagement, which induced their recruitment to the tumor invasive fronts and secretion of stromal cell–derived factor 1 (SDF1) that triggered tumor cell invasive behavior via CXCR4. A comprehensive clinical validation of the targets in different stages of PanNETs demonstrated the implication of both Sema4D and CXCR4 in tumor progression. Taken together, we demonstrate beneficial antitumor and prosurvival effects of anti-Sema4D antibody but also unravel a novel mechanism of tumor aggressivity. This mechanism implicates recruitment of Sema4D-positive macrophages to invasive fronts and their secretion of proinvasive molecules that ultimately induce local tumor invasion and distant metastasis in PanNETs.
An anti-semaphorin-4D vascular targeting agent demonstrates antitumor and prosurvival effects but also unravels a novel promalignant effect involving macrophage-derived SDF1 that promotes tumor invasion and metastasis, both in animal models and patients.See related commentary by Tamagnone and Franzolin, p. 5146