AIBP augments LEC fate commitment.

a Representative confocal images of Prox1⁺ cells in the fli1a:egfp and apoa1bp2^−/−; fli1a:egfp at 4 dpf following immunostaining using GFP and Prox1 antibodies. Arrows show the Prox1⁺ LECs in TD. Blood vessels are labeled in green, and Prox1 staining is shown in red. Data are pooled from 3 independent experiments. b Quantitative data of Prox1⁺ LEC in TD of 7 somites. Mean ± SE, n = 23 (fli1a:egfp) and n = 28 (apoa1bp2^−/−; fli1a:egfp) embryos; unpaired two-sided t-test with Welch’s correction. c Scheme illustration of mESC to LEC differentiation. The embryoid bodies (EBs) of mESCs were prepared and cultured in the EC differentiation medium containing BMP4 and bFGF for 3 days. Recombinant VEGFA and VEGFC in combination or AIBP were supplemented at day 3 and remained in culture for additional 4 days. d qPCR analysis of the expression of LEC-associated Lyve1 and Prox1 and endothelial cell marker Pecam1 at the indicated time points of mESC to LEC differentiation. Mean ± SD, n = 3 independent repeats; two-way ANOVA with Tukey’s post-hoc test. The P-values comparing control and AIBP-treated samples are shown. e qPCR analysis of the expression of Vegfc and Adamts3. Mean ± SD, n = 3-4; two-way ANOVA with Tukey’s post-hoc test. f Effect of VEGFR3 and NOTCH inhibition on mESC to LEC differentiation. The murine ESCs were subjected to LEC differentiation as described in panel c in the presence or absence VEGFR3 inhibitor SAR (SAR13165) or NOTCH inhibitor DAPT from day 3 to day 7. qPCR analysis of Prox1, Lyve1, and Pecam1 expression was performed. Mean ± SD, n = 3 independent repeats; two-way ANOVA with Tukey’s post-hoc test. ****p < 0.0001. Scale: 50 µm. Source data are provided as a Source Data file.

AIBP regulates LEC caveolae.

a TEM analysis of control and recombinant AIBP-treated hLECs. Arrows depict caveolae. b Quantification of caveolae in (a). n = 35 (control) and n = 36 (AIBP) cells. Data are Mean ± SE; unpaired two-sided t-test with Welch’s correction. c TEM analysis of caveolae in ECs of the cardinal vein from control and apoa1bp2^−/− zebrafish. Arrows indicate closed caveolae and arrowheads show open caveolae. d Quantification of caveolae in cardinal vein ECs. n = 27 (control) and n = 29 (apo1bp2^−/−) cells. Data are Mean ± SE; unpaired two-sided t-test with Welch’s correction. e Scheme illustration of 4 hydroxy-tamoxifen (4OHT)-induced ubiquitous mEos2-APOA1 expression. f Representative images of control and mEos2-APOA1 expressed embryos after 4OHT treatment. Embryos were imaged at 2 dpf. The white dashed line demarcates the control animals. The results are representative of 3 independent repeats. g Photoconversion and vascular circulation of mEos2-APOA1. The head regions of embryos were exposed to UV light for 1 min to induce photoconversion. h Analysis of mEos2-APOA1 secretion. The specified tail region was imaged, and the maximum RFP-to-GFP signal ratio within the ISV lumen was quantified. n = 8 intersegmental vessels from 2 embryos. Data are presented as mean ± SEM, analyzed using one-way ANOVA with Tukey’s post-hoc test. i Maxi-projection confocal images of Prox1⁺ cells in the apoa1bp^−/−; fli1a:egfp zebrafish with control (mEos2-APOA1) and mEos2-APOA1 overexpression (ubi:Gal4-ERT2+mEos2-APOA1) at 4 dpf and immunostained with GFP (green) and Prox1 (red) antibodies. Arrowheads show the Prox1⁺ LECs in TD. j Quantitative data of Prox1⁺ LEC in TD (7 somites). n = 10 (mEos-APOA1) and n = 11 (ubi:Ert2-Gal4; mEos-APOA1) embryos. Data are Mean ± SE; unpaired two-sided t-test with Welch’s correction. CV cardinal vein. Scale bar: 400 nm in a and 500 nm in c; 100 µm in (f, g, i). Source data are provided as a Source Data file.

Cholesterol regulates VEGFR3 signaling.

a−d Effect of MβCD on VEGFR3 phosphorylation. a hLECs were growth factor-starved, and treated with 10 mM MβCD for 5, 15, and 30 min, and the resulting cells were further stimulated with 100 ng/mL VEGFC. The resulting cells were lysed and blotted using CAV-1, VEGFR3, GAPDH antibodies. b Quantitative analysis of panel a. Mean ± SD, n = 3 repeats; two-way ANOVA with Dunnett’s post-hoc test. c, hLECs were treated as in panel a. and cell lysates were immunoprecipitated using VEGFR3 antibody. Immunoblotting was performed using anti-phosphotyrosine (4G10) and VEGFR3 antibodies. d Quantitative analysis of (c). n = 3 repeats. Data are presented as mean ± SD and were analyzed using one-way ANOVA with Tukey’s post-hoc test. e, f Effect of AIBP treatment on VEGFR3 distribution in caveolar fractions. e hLECs were incubated with either recombinant AIBP or vehicle control in EBM2 supplemented with 10% FBS for 2 h, and the cells were subjected to sucrose gradient ultracentrifugation. n = 3 repeats. The resulting fractions were collected for Western blot analysis as indicated. Tx treatment; cav: caveolar fraction; n.c non-caveolar fraction. f Quantitative data of (e). Mean ± SD; two-way ANOVA with Sidak’s post-hoc test. n = 3 repeats. g, h Effect of AIBP and HDL co-treatment on VEGFR3 signaling. g hLECs were growth factor-starved and treated with HDL, AIBP, or HDL and AIBP in combination, and further stimulated with VEGFC. The resulting cells were lysed and immunoblotted as indicated. h Quantitative data of ERK and AKT activation. Mean ± SD; two-way ANOVA with Tukey’s post-hoc test. n = 3 repeats. Ctrl: control. i Maxi-projection confocal images of Prox1⁺ and pErk1/2⁺ cells in the apoa1bp^−/−; fli1a:egfp zebrafish at 36 hpf following immunostaining using GFP, pErk1/2, and Prox1 antibodies. Dorsal (DA) aorta and cardinal vein (CV) were imaged. Arrows show the Prox1⁺ LECs with pErk1/2 expression. j Quantitative data of pErk1/2 intensity in Prox1⁺ LECs. Data are Mean ± SE; unpaired two-sided t-test with Welch’s correction. n = 146 (control) and n = 166 (apoa1bp^−/−) cells. Scale bar: 50 µm. Source data are provided as a Source Data file.

Effect of CAV-1 on VEGFR3 signaling.

a Conserved CAV-1 binding site on VEGFR3 in human (Hu), mouse (Ms), and zebrafish (Zf). The conserved amino acids are shown in blue. b Co-immunoprecipitation of endogenous VEGFR3 and CAV-1 in hLECs. Lysates from two 10 cm confluent plates of hLECs were combined, then equally divided for immunoprecipitation using VEGFR3 antibody or control protein A beads. The samples were subsequently immunoblotted for CAV-1 and VEGFR3. c, d VEGFR3^AAA loses its binding to CAV-1. c hLECs were transfected with control EGFP, VEGFR3-EGFP (R3), or VEGFR3^AAA-EGFP (R3^AAA) using lentivirus-mediated gene transduction. After 72 hours, the resulting cells were lysed and immunoprecipitated with GFP antibody conjugated to agarose beads and immunoblotted using GFP and CAV-1 antibodies. d The input lysates were immunoblotted using GFP, CAV1, or GAPDH antibody as indicated. e Localization of VEGFR3 and VEGFR3^AAA in caveolae. hLECs were transduced with VEGFR3-APEX2 or VEGFR3^AAA-APEX2 Lenti-viral particles, and after 72 h, cells were fixed with 2.5% glutaraldehyde, stained using DAB substrate kit, and pelleted for TEM analysis. An enlarged view of a single caveola, highlighted with a white box, is shown in the top left corner of each image. f–h hLECs were transduced using lentivirus, and the resulting cells were growth factor starved and treated with 100 ng/mL VEGFC for 20 min, cells were then lysed and immunoblotted as indicated. R3/R3^AAA-EGFP denotes detection using GFP antibody. Quantitative data of VEGFR3 activation (g), ERK activation (i), and AKT activation ( j) were shown. Mean ± SD; two-way ANOVA with Tukey’s post-hoc test. n = 3 independent repeats in g, i, j. Endg: endogenous. Scale bar: 400 nm. Source data are provided as a Source Data file.

Effect of CAV-1 on Aibp2-regulated LEC specification.

a Representative TEM images of control and cav1^−/− zebrafish. Arrows show typical caveolae in cardinal vein ECs. b Quantification of caveolae in a. n = 27 (control) and n = 20 (cav1^−/−) cells. c Maxi-projection confocal images of Prox1⁺ cells in the 36 hpf apoa1bp^−/−; fli1a:egfp zebrafish with cav-1 crRNA alone or cav1-crRNA and Cas12a protein co-administration. Arrows show the Prox1⁺ LECs in cardinal vein. Green marks blood vessels, and red marks Prox1 staining. d Quantitative data of Prox1⁺ LECs in cardinal vein (7 somites) of (c). n = 34 embryos in both groups. e Representative confocal images of LEC specification in the cardinal veins of Vegfr3^AAA knock-in zebrafish at 36 hpf. Green marks blood vessels, and red marks Prox1+ cells. f Quantitative data of LEC numbers in (e). n = 13 (flt4^+/+) and n = 11 (flt4^+/AAA) embryos. g Confocal imaging of Prox1⁺ LECs in the indicated genetically modified fli1a:egfp zebrafish at 36 hpf. Green marks blood vessels, and red marks Prox1 staining. h Quantitative data of Prox1⁺ LECs of 7 somites in (g). n = 22 (apoa1bp^−/−; flt4^+/+) and n = 26 (apoa1bp^−/−; flt4^+/AAA) embryos. Mean ± SE in b, d, f, and h; unpaired two-sided t-test with Welch’s correction. CV cardinal vein. Scale bar: 500 nm in a; 50 µm in c, e, g. Source data are provided as a Source Data file.

AIBP administration promotes corneal lymphangiogenesis.

a Illustration of murine cornea lymphangiogenesis assay. b Representative images of murine corneal lymphangiogenesis by implantation of pellets containing the indicated recombinant proteins and immunostained using LYVE1 (green) and CD31 (red) antibodies. Enlarged images of the boxed regions are shown in the lower panels. c Quantification of LYVE1⁺ lymphatic vessel area per cornea. VC: VEGFC; AA: AIBP + APOA1; CSD: CAV-1 scaffolding domain peptide. Data are presented as Mean ± SE,and were analyzed using one-way ANOVA with Holm-Sidak’s post-hoc test. n = 8 (control), 9 (VC), 7 (VC + AA), 6 (VC + AIBP), and 9 (VC + CSD) corneas. Scale bar: 1000 μm in upper image and 500 μm in enlarged image. Source data are provided as a Source Data file.

AIBP facilitates VEGFC-mediated recovery from lymphedema in mice.

a Representative images of mouse tail lymphedema 14 days post-surgery. One-year-old mice were subjected to tail surgery to create lymphedema. Immediately following the surgery, recombinant human VEGFC, VEGFC in combination with AIBP, or saline control were delivered via peritoneal injection, twice a week for 4 weeks. The tails were imaged after the injury, and the recovery from swelling analyzed. b The quantification of tail volume in (a). Data are Mean ± SE; two-way ANOVA with Tukey’s post-hoc test. n = 3 (PBS), 3 (VEGFC), and 4 (AIBP + VEGFC) mice. P-values for VEGFC vs. VEGFC + AIBP-treated group are shown. Source data are provided as a Source Data file.