Overall, this strongly suggests that a paracrine loop could be involved in B. henselae VEGF action. However, up to now, the majority of these studies have been carried out on macrovascular ECs. Interestingly, using our species- and organ-specific ECs model, we showed that B. henselae increased VEGF production by Human Skin Microvascular ECs but not by iHUVEC or feline macro- or micro-vascular ECs. These results, which might explain the decreased pathogenic potential of B. henselae infection for cat as compared to human, strongly suggests the involvement of an autocrine secretion of VEGF by skin ECs. This finding points out that the mechanism of B. henselae-mediated angiogenesis induction implies autocrine VEGF production and stimulation of the infected endothelium. In parallel to VEGF production, the phosphorylation of VEGFR-2 was observed in ECs upon infection, mostly by homologous strains. This phenomenon is of high interest in the context of epidemiological human infection. As B. henselae isolates infecting a human have always a feline origin, it is tempting to speculate that an adaptive switch is taking place upon cat scratch in humans. Altogether, our results raise the hypothesis of hostdependent signaling in ECs upon bacterial infection as VEGFR-2 activation is more prominent in human ECs exposed to homologous strains. The same holds true for feline ECs. Because bacillary angiomatosis is not detected in cats infected by homologous strains, alternate mechanisms are likely taking place, Our results support the fact that VEGF signaling might be increased in human ECs through an autocrine production, which does not occur in feline ECs. Paracrine VEGF activation cannot be excluded; therefore, it will be interesting to further Vismodegib Hedgehog inhibitor investigate whether VEGF can be released in the perivascular microenvironment by other cells, such as macrophages and polymorphonuclear cells. Our results on VEGF and VEGFR-2 are in BI-D1870 apparent contradiction with those obtained by Scheidegger et al. However, the concentration ranges and source of VEGF were far for being comparable. Our study highlights the ECs organo- and species- specificity in their interactions with B. henselae. Indeed, our model further demonstrates the difference of reactivity displayed by human vs feline ECs on the one hand and by the macro- vs the microvasculature on the other hand. Additionally, local physiological microenvironment is most likely to play an important role. In this context, our cellular systems will allow studying the process of specific ECs infection by B. henselae. They will offer new possibilities to investigate bacterial and cellular factors which determine human vs cat reactivity, as well as the mechanisms involved in anti-apoptotic and/or pro-angiogenic effects, ultimately resulting to bacillary angiomatosis and peliosis only in humans. In conclusion, according to the validated hypothesis stating the organospecificity of the endothelium, our work extends this concept to the species-specific endothelial cell phenotypes.