Fig 1: RAB11B controls the recycling of endocytosed active a5ß1 integrins and polarized ED-A FN secretion in ECs.(a) Time-course analysis of recycled active a5ß1 integrin in siCTL ECs versus siRAB11A ECs or siRAB11B ECs. RAB11B, but not RAB11A, silencing significantly impairs active a5ß1 integrin recycling. Data are mean±s.e.m., n=3 independent experiments (two technical replicates for each experiment). (b) Confocal microscopy analysis of anti-active a5ß1 integrin mAb SNAKA51 localization (red) in living confluent ECs (20 min of incubation). SNAKA51+ active a5ß1 integrin localizes in fibrillar adhesion in siCTL and in siRAB11A, but not in siRAB11B ECs in which it accumulates in perinuclear punctae. The number of active a5ß1 integrin-containing adhesions per 100 µm2 of cell area was quantified in siCTL, siRAB11A and siRAB11B ECs. Data are mean values±s.e.m., n=20 cells per condition pooled from two independent experiments. (c) Confocal xz sectioning microscopy analysis of anti-active a5ß1 integrin mAb SNAKA51 localization (red) in living confluent ECs (20 min of incubation). Quantitative analysis of apico-basal mean intensity ratio reveals that SNAKA51+ active a5ß1 integrin localizes on the basolateral surface of siCTL and siRAB11A, but not of siRAB11B ECs, where it redistributes all around the cell surface. Data are mean values±s.e.m., n=20 cells per condition pooled from two independent experiments. (d) Confocal microscopy analysis of IST9 mAb+ endogenous cellular ED-A FN (green) in confluent ECs. ED-A FN polymerizes into a fibrillar network in siCTL and siRAB11A, but not in siRAB11B ECs, where it accumulates in a perinuclear compartment. Relative amount of fibrillary ED-A FN area was calculated in siCTL, siRAB11A and siRAB11B ECs. Data are mean values±s.e.m., n=20 cells per condition pooled from 2 independent experiments. (e) Western blot analysis of soluble ED-A FN released by confluent ECs seeded on Transwell inserts. An equal percentage of apical and basolateral volumes of medium were collected after 72 h of culture from different wells of siCTL or siRAB11B ECs. Equal amounts of rabbit IgG were exogenously added to samples (spike normalization) for loading control purposes. Quantification of the ratio between apical or basolateral amount of ED-A FN released by siCTL over siRAB11B ECs. RAB11B silencing much more severely impairs basolateral than apical ED-A FN secretion. Data are mean±s.e.m., n=6 wells per condition pooled from three independent experiments. Scale bar, 50 µm (d), 20 µm (b), 5µm (c). **P<0.01; ***P<0.001; Student's t-test.