Eph-B4 prevents venous adaptive remodeling in the adult arterial environment

Abstract

Eph-B4 determines mammalian venous differentiation in the embryo but is thought to be a quiescent marker of adult veins. We have previously shown that surgical transposition of a vein into the arterial environment is characterized by loss of venous identity, as indicated by the loss of Eph-B4, and intimal thickening. We used a mouse model of vein graft implantation to test the hypothesis that Eph-B4 is a critical determinant of venous wall thickness during postsurgical adaptation to the arterial environment. We show that stimulation of Eph-B4 signaling, either via ligand stimulation or expression of a constitutively active Eph-B4, inhibits venous wall thickening and preserves venous identity; conversely, reduction of Eph-B4 signaling is associated with increased venous wall thickness. Stimulated Eph-B4 associates with caveolin-1 (Cav-1); loss of Cav-1 or Eph-B4 kinase function abolishes inhibition of vein graft thickening. These results show that Eph-B4 is active in adult veins and regulates venous remodeling. Eph-B4-Cav-1-mediated vessel remodeling may be a venous-specific adaptive mechanism. Controlled stimulation of embryonic signaling pathways such as Eph-B4 may be a novel strategy to manipulate venous wall remodeling in adults.

Figure 1.

Venous expression of Eph-B4 during vein graft adaptation in both a mouse model and in human specimens. (A–C) Expression of ephb4 (A), efnb2 (B), and opn mRNA (C). n = 4. IVC, inferior vena cava; VG, vein graft; Ao, aorta. Error bars denote SEM. (D) Eph-B4 and Ephrin-B2 immunoreactivity at 3 wk of vein graft adaptation in mice. Photomicrographs show representative samples from n = 4 independent samples for each group. Bars, 100 µm. Black arrows show the vein graft wall. Red arrowheads indicate Eph-B4– and Ephrin-B2–positive signals. (E) Eph-B4 immunoreactivity during human vein graft adaptation. Representative photomicrographs from two independent samples are shown. Bar, 250 µm. Red arrowheads point to the Eph-B4–positive signal. (F) Distribution of Eph-B4 in mouse veins (n = 4) as determined by immunofluorescence. The density of cells immunoreactive for both Eph-B4 and either CD31 or α-actin was analyzed by ImageJ (National Institutes of Health). **, P < 0.01; *, P < 0.05. Error bars denote SEM.

Figure 2.

Recombinant Ephrin-B2/human IgG Fc chimeric protein activates the Eph-B4 receptor in vitro. (A) EC expression of Eph-B4 (green), phosphotyrosine (red), or both (merge) after 5 min of treatment with 2 µg/ml of control CD6/Fc or 2 µg/ml Ephrin-B2/Fc. Bar, 50 µm. (B) Immunoprecipitation analysis of tyrosine-phosphorylated Eph-B4 (arrowhead) in EC treated with Ephrin-B2/Fc. *, IgG heavy chain; **, IgG light chain. Bottom shows total cell lysate. (C) Expression of VCAM and ICAM on EC treated with Ephrin-B2/Fc. n = 3. (D) Migration of WT EC in response to Ephrin-B2/Fc, control CD6/Fc, or VEGF. n = 3. Error bars denote SEM. (E) Migration of WT or Eph-B4^+/− EC in response to Ephrin-B2/Fc or FBS. n = 3. **, P < 0.01; *, P < 0.05. Error bars denote SEM.

Figure 3.

Ephrin-B2/Fc activates Eph-B4 and decreases vein graft wall thickness in vivo. (A) Immunofluorescence of anti–human IgG-FITC in whole mount inferior vena cava (IVC) isolated from mice 24 h after injection with control CD6/Fc or Ephrin-B2/Fc. Representative photomicrographs from three independent samples are shown. Bar, 20 µm. (B) Expression of Eph-B4 (green), phosphotyrosine (red), or both (merge) in pulmonary vein (PV), inferior vena cava, or vein graft (VG) of mice after 3 wk of treatment with control CD6/Fc or Ephrin-B2/Fc. Representative photomicrographs from three independent samples are shown. Star indicates vessel lumen. Arrows show vein graft wall. Bars, 20 µm. pTyr, phosphotyrosine. (C) Mouse body weight during treatment with control CD6/Fc or Ephrin-B2/Fc. Error bars denote SEM. (D, G, and H) Mouse vein grafts after 3 wk of treatment with control CD6/Fc or Ephrin-B2/Fc, examined with H&E staining (D), immunoreactivity against α-actin (SMA; G), or immunofluorescence against Eph-B4 (H). Arrowheads show vein graft wall. *, lumen. Bars, 20 µm. (E and F) Vein graft wall thickness and lumen/total vessel area after control CD6/Fc or Ephrin-B2/Fc treatment. Error bars denote SEM. (I and J) Percentage of cells in the walls of the inferior vena cava or vein grafts, treated with control CD6/Fc or Ephrin-B2/Fc and positively staining for proliferation (I) or apoptosis (J). Error bars denote SEM. All experiments analyzed five (CD6/Fc) and seven (Ephrin-B2/Fc) mice. *, P < 0.05; **, P < 0.01.

Figure 4.

Increased intimal thickening in vein grafts derived from Eph-B4^+/− mice (4 wk). (A) Histology of veins derived from WT (WT) or Eph-B4 heterozygous KO (Eph-B4^+/−) mice stained with H&E. Bar, 50 µm. (B) Morphology of vein graft adaptation of WT or Eph-B4^+/− vein grafts placed into WT or Eph-B4^+/− hosts, examined with H&E staining or smooth muscle α-actin (SMA) immunoreactivity. Arrows show intimal thickening. Bar, 100 µm. (C–E) Vein and vein graft wall thickness (C), vein graft lumen area (D), and total vein graft area (E). Error bars denote SEM. (F) Numbers of CD3 immunoreactive cells infiltrated into veins and vein grafts. Error bars denote SEM. n = 5–7 in each group. **, P < 0.01.

Figure 5.

Plasmids with mutated Eph-B4 kinase domain have defective Eph-B4 phosphorylation. (A) Time course of Eph-B4 phosphorylation, stimulated with 2 µg/ml Ephrin-B2/Fc, in HEK293 cells transfected with HA-tagged mEph-B4 plasmid. (B) Time course of Ras-GAP association with Eph-B4, stimulated with 2 µg/ml Ephrin-B2/Fc, in HEK293 cells transfected with HA-tagged mEph-B4 plasmid. (C) Immunofluorescence for HA-tagged Eph-B4 in HEK293 cells transfected with either mEph-B4 wt plasmid or mutant (w804a and w811f) mEphB4 plasmids, counterstained with DAPI. Bar, 50 µm. (D) Tyrosine phosphorylation in response to Ephrin-B2/Fc in HEK293 cells transfected with WT or mutant (w804a and w811f) mEph-B4 plasmids. (E) Immunofluorescence of HA-tagged Eph-B4 (green), cav-1 (red), or merge (yellow) in COS cells transfected with WT or mutant Eph-B4 plasmids, counterstained with DAPI, and without or with stimulation with Ephrin-B2/Fc. White arrowheads show Eph-B4 in the intracellular vesicles. Red arrowheads show colocalization of Eph-B4 with cav-1 at the cell membrane, after treatment with Ephrin-B2/Fc, in WT but not mutant Eph-B4 plasmids. Bar, 100 µm. (F) Tyrosine phosphorylation in response to Ephrin-B2/Fc in COS cells transfected with WT or mutant (w804a, w811f) mEph-B4 plasmids. (G) Ras-GAP or Ephexin association with Eph-B4, stimulated with 2 µg/ml Ephrin-B2/Fc, in COS cells transfected with HA-tagged mEph-B4 plasmid. All panels show representative samples from three independent experiments. WCL, whole cell lysate.

Figure 6.

Eph-B4 inhibition of vein graft thickening depends on a functional Eph-B4 kinase domain. (A) Tyrosine phosphorylation on Eph-B4 in EC transfected with Ad-Eph-B4 wt or Ad-Eph-B4 w804a. n = 3. (B and G) Vein grafts, transfected with Ad-GFP, Ad-Eph-B4 wt, or Ad-Eph-B4 w804a, were implanted into WT mice for 3 wk and then examined for GFP immunofluorescence (B) or H&E staining (G). Arrows show vein graft wall. Bars, 50 µm. (C–E) Immunofluorescence of CD31 (red) and GFP (green; C), SM α-actin (red) and GFP (green; D), and F4/80 (red) and GFP (green; E) in Ad-Eph-B4 wt-transfected vein grafts. Bars, 20 µm. Arrowheads show colocalized cells. *, lumen. (F) Vein graft wall thickness as determined by serial ultrasounds in mice. Vein grafts were control grafts (filled circle) or treated with Ad-Eph-B4 wt (filled red square) or Ad-Eph-B4 w804a (empty red square). Error bars denote SEM. (H and I) Vein graft wall thickness, and lumen and total vessel area, after 3 wk, in vein grafts transfected with Ad-GFP, Ad-Eph-B4 wt, or Ad-Eph-B4 w804a. **, P < 0.01. n.s., not significant. Error bars denote SEM. B-I show representative data from n = 6 independent samples.

Figure 7.

Eph-B4 signaling is linked to Cav-1 in vitro. (A) Eph-B4 expression in ECs derived from WT mice, Cav-1 KO mice, or Cav-1 KO mice with EC-specific reconstitution of Cav-1 (Cav-1 RC). n = 3. Bar, 50 µm. (B) Sucrose gradient analysis of WT EC or Cav-1 KO EC. n = 3. (C) Immunofluorescence of Eph-B4 (green), cav-1 (red), or both (merge) in EC stimulated with control CD6/Fc or Ephrin-B2/Fc. Arrowheads show colocalization. n = 3. (D) Cav-1 association with Eph-B4, stimulated with 2 µg/ml Ephrin-B2/Fc, in EC. n = 3. (E) Migration of WT, Cav-1 KO, or Cav-1 RC EC in response to CD6/Fc, Ephrin-B2/Fc, or FBS. n = 3. *, P < 0.05; **, P < 0.01. Error bars denote SEM. (F) Nitric oxide production by WT, Cav-1 KO, or Cav-1 RC EC in response to control or Ephrin-B2/Fc. n = 3. *, P < 0.05. Error bars denote SEM. NCT, negative control. n.s., not significant.

Figure 8.

Eph-B4 activation stimulates Cav-1 phosphorylation in vitro and in vivo. (A) FACS analysis of venous EC derived from WT or Eph-B4^+/− mice. Ephrin-B2/Fc was allowed to bind to EC and then detected with anti–Fc-FITC. Control groups did not include the anti–Fc-FITC antibody. n = 3. (B) Phosphorylation of Cav-1 on tyrosine-14 (Y14) in WT or Eph-B4^+/− EC stimulated with control CD6/Fc or Ephrin-B2/Fc. n = 3. (C) WT or Eph-B4^+/− EC expression of phosphorylated-Cav-1 (red), stimulated with control CD6/Fc or Ephrin-B2/Fc. n = 3. Bar, 50 µm. (D) Expression in inferior vena cava (IVC) of Cav-1 WT or Cav-1 KO mice of phosphorylated Cav-1 (red). n = 3. Bar, 50 µm. (E) Expression in inferior vena cava, vein graft (VG), or aorta (Ao) of WT mice, of Cav-1 (green, top two rows), Eph-B4 (green, bottom three rows), phosphorylated Cav-1 (red), and both (merge). Bar, 50 µm. n = 3. Arrowheads denote loss of pCav-1 signal. (F) Expression in vein grafts derived from WT or Eph-B4^+/− mice of CD31 (green), phosphorylated Cav-1 (red), or both (merge). Counterstain is with DAPI (blue). Bar, 20 µm. n = 3. (G) Expression in vein grafts derived from control or Ephrin-B2/Fc–treated mice, of phosphorylated Cav-1 (red), both (merge) phosphorylated Cav-1 (red) and total Cav-1 (green), and both (merge) phosphorylated Cav-1 (red) and Eph-B4 (green). Counterstain is with DAPI (blue signal). n = 3. Bar, 50 µm. Arrowheads denote intimal-medial thickness. *, lumen.

Figure 9.

Ephrin-B2/Fc inhibition of vein wall thickening depends on Cav-1. (A) Vein grafts derived from WT, homozygous Cav-1 KO mice, or Cav-1 EC-reconstituted (Cav-1 RC) mice were implanted into littermate mice, examined after 4 wk, and stained with H&E. Arrows show vein graft wall. n = 2–7. Bar, 50 µm. (B) Vein graft wall thickness as determined by serial ultrasounds in mice. Vein grafts were derived from WT (squares) or Cav-1 KO mice (circles) and treated with control CD6/Fc (black) or Ephrin-B2/Fc (red). n = 5–7. Error bars denote SEM. (C) Vein grafts were derived from WT or Cav-1 KO mice, implanted into littermate WT mice, treated with control CD6/Fc or Ephrin-B2/Fc, examined after 4 wk, and stained with H&E. Arrows show vein graft wall. n = 5–7. Bar, 50 µm. (D) Vein graft wall thickness in WT and Cav-1 KO vein grafts, with control CD6/Fc or Ephrin-B2/Fc treatment. n = 5–7. Error bars denote SEM. (E) Numbers of CD3 immunoreactive cells infiltrated into veins and vein grafts. Error bars denote SEM. n = 5–7. **, P < 0.01. n.s., not significant.