(A) Schematic of drug treatment protocol. (B) Time course images of microvessel perfusion with 70 kDa dextran-rhod B (scale bar 500 μm). (C) Quantification of microvessel 70 kDa dextran-rhod B intensity normalized to t = 1 of 25% KRAS4AG12V microvasculatures treated with different drugs starting at day 5. Fluorescent intensity was measured as a ratio of interstitial to lumen intensity normalized to t = 1. MEKi-treated AVM-like microvasculatures displayed significantly lower dextran extravasation than DMSO or PI3Ki. (One-way ANOVA with Tukey’s multiple comparison test; N = 3, ***p < 0.001; mean ± SD) (D) Permeability coefficient of 25% KRAS4AG12V microvasculatures treated with DMSO (gray), MEKi (orange) or PI3Ki (purple). The permeability of MEKi treatment microvasculatures was significantly lower than that of DMSO or PI3Ki treated AVM-like vessels. (One-way ANOVA with Tukey’s multiple comparison test; N = 3, ***p < 0.001; mean ± SD) (E) Immunofluorescence images at day 7 of day 5 treated AVM-on-a-chip. MEKi-treated AVM-like microvasculatures have an increase in VE-cadherin-rich (purple) adherens junction staining in KRAS4AG12V ECs (white, arrows indicate areas of increased VE-cadherin) compared to DMSO or PI3Ki treatment (arrows indicate areas of VE-cadherin disconnect) (scale bar, 100 μm).