Fig 2: Absence of PLVAP does not affect fenestral formation or liver angiogenesis but alters immunocomplex binding to LSEC. (a) Scanning electron micrographs of sinusoidal vessels in 5 wk old wild-type and Plvap-/- mice. Bars, 1 µm. (b) Quantification (mean ± s.e.m., n = 9, 3 field/mouse, 3 mice/genotype) of the number of fenestral openings per sieve plate, the number of sieve plates per endothelial surface area, the number of fenestrations per endothelial cell area and the average surface area of a fenestral opening. (c) Flow cytometric analysis of liver EC (LYVE-1+CD144+ LSEC) and their quantification per CD45- parenchymal cells (left) and LYVE-1-CD144+ non-sinusoidal ECs per CD45- parenchymal cells (right)) in 5 wk old wild-type and Plvap-/- mice. Quantitative data are mean ± s.e.m. (n = 3–4 mice/genotype). (d) Maximal z-stack projections (48 µm) of whole-mount stainings of E14.5 liver from wild-type and Plvap-/- mice for LYVE-1 and PLVAP. Bars, 50 µm. (e), Quantification of sinusoidal vessel diameters in wild-type and Plvap-/- mice (mean ± s.e.m, 30 vessels/genotype, 2 mice/genotype). (f) Quantification of LYVE-1 and PLVAP signal areas from single optical sections (mean ± s.e.m, 10 vessels/genotype, 2 mice/genotype). (g–i) Flow cytometric analyses of LSEC (Live+CD45-LYVE-1+CD144+ cells) 2 h after an intravenous injection of Atto-488 conjugated OVA-IC in 5 wk old wild-type and Plvap-/- mice. (g) The percentage of LSEC (from all LSEC) that have taken up OVA-IC (mean ± s.e.m, 3–4 mice/genotype). (h) Representative histograms of the distribution of Atto-488-labeled OVA-IC in LSEC. Control, a mouse to which vehicle only (PBS) is administered. i Quantification of the mean fluorescence intensities (MFI) from (g). Representative images are from (n = 3 mice/genotype (a), and n = 2 mice/genotype (d) and 3–4 mice/genotype (h). *p < 0.05, ***p < 0.001.

Fig 3: Expression of PLVAP in Cav1-deficient liver. (a) Immunoblot analyses of adult liver lysates (5 µg protein/lane) from 5 wk old wild-type, Cav1-/- and Plvap-/- mice for PLVAP and GAPDH (a house keeping molecule) under non-reducing conditions. Quantitative data are mean ± s.e.m. (n = 3 mice/genotype). The same filter was sequentially probed for PLVAP and GAPDH. Full blots are shown in Suppl. Figure 10. (b) Confocal microscopy analyses of PLVAP and LYVE-1 expression in the midlobular area of 5 wk old wild-type and Cav1-/- mice and quantification of the PLVAP+ area (mean ± s.e.m, each dot represents one mouse). Bars, 10 µm. (c) Analyses of PLVAP expression in 24 wk old wild-type mice. Bars, 10 µm. (d) Quantitative PCR analyses of Lyve-1 and Plvap mRNA expression in sorted sinusoidal ECs (CD45-Podoplanin-LYVE-1+CD144+ cells) and non-sinusoidal (CD45-Podoplanin-LYVE-1-CD144+) EC and leukocytes (control cells) (n = 5 mice for both cell types). ß-actin was used as a control gene. Shown are representative images (numbers of mice/genotype: n = 3(a), n = 4–6 (b), n = 1 (c).

Fig 4: PLVAP expression in the diaphragmed and non-diaphragmed domains of endothelial cells in adult liver. (a) Transmission electron micrographs from a vein in liver of a 9 wk old wild-type mouse. VL, vascular lumen, BM, basement membrane (yellow dotted line). The two boxed areas are shown at a higher magnification in the insets. Red arrows, diaphragms in caveolae, fenestrae, and transendothelial channels. Bar, 40 nm. (b) Confocal microscopy analyses of PLVAP and LYVE-1 expression in 5 wk old wild-type mice. Blue is DAPI. CV, central venule; PV, portal venule. Bar, 100 µm. (c) Higher magnifications of the areas defined by the green boxes in (b) Bar, 10 µm. (d) Quantification of relative PLVAP staining intensities (mean ± s.e.m, n = 4 mice) from the portal vein, periportal, midlobular and pericentral sinusoids and from the central vein. Intensity value 1.0 was assigned to the portal vein in each animal. (e,f) Confocal microscopy analyses of PLVAP, LYVE-1 and CD144 (e) and PLVAP, LYVE-1 and CD31 (f) expression in 5 wk old wild-type mice. White arrowheads point to LYVE-1-PLVAP+CD144+ (e) and LYVE-1-PLVAP+CD31+ (f) sinusoids. Red arrowheads point to LYVE-1+PLVAP+CD144+ (e) and LYVE-1+PLVAP+CD31+ sinusoids (f). Bars, 50 µm. Shown are representative images (numbers of mice/genotype: n = 3 (a,e,f); n = 4 (b,c)). *p < 0.05.

Fig 5: PLVAP associates with LYVE-1, neuropilin-1 and VEGFR2 in fetal but not adult LSEC. (a–d) Proximal ligation analyses (PLA) of the associations between PLVAP and LYVE-1 (a), PLVAP and neuropilin-1 (Nrp1) (b), PLVAP and VEGFR2 (c) and VEGFR2 and Nrp1 (d) in fetal and adult livers. The white signals show associations between the indicated molecules, and the green stainings show expression of PLVAP (a–c) and VEGFR2 and Nrp1 (d). The boxed areas are shown at a higher magnification in the insets. Red arrows point to representative PLA signals. Bars, 10 µm. The quantifications (the area fraction of the white signal per PLVAP (a–c) or Nrp1 + VEGFR2 (d) positive sinusoids) are shown (mean ± s.e.m., n = 2–5 fields/mouse, 3–4 mice/genotype). *p < 0.05.