Involvement of endothelial ephrin-B2 in adhesion and transmigration of EphB-receptor-expressing monocytes

The vascular endothelium forms an interface between the circulation and the different body compartments. As such, it controls the trafficking of leukocytes. The recruitment of circulating cells is orchestrated by the coordinated activities of adhesive and chemoattractive signals. The sequential activities of selectins, integrins and IgG-superfamily molecules in concert with chemokines control leukocyte tethering, rolling, firm adhesion and subsequent transmigration through the vascular endothelium (Vestweber, 2007). The specific repertoire of adhesion molecules and chemotactic signals establishes the site and cell-type-selective vascular `area code' necessary for spatiotemporal control of leukocyte trafficking (Vestweber, 2007). Whereas the major players of leukocyte trafficking have been well established, additional adhesive signals may be involved in the fine tuning of the vascular address code. Most notably, the steps leading from firm leukocyte adhesion to transmigration are still poorly understood.

Ephrin (Eph) receptors and their corresponding ephrin ligands are cell surface molecules that elicit positional guidance cues on corresponding receptor and ligand expressing cells that get into juxtapositional contact (Arvanitis and Davy, 2008; Kullander and Klein, 2002; Wilkinson, 2001). Originally identified as neuronal pathfinding molecules, Eph receptors and ephrin ligands are now recognized as being widely expressed controlling cell-cell interactions of multiple cell types (Kullander and Klein, 2002; Pfaff et al., 2006). Vascular functions of Eph receptors and their ligands were identified a decade ago when genetic loss-of-function experiments revealed surprising roles of the Eph-receptor ligand ephrin-B2 (also known as EFNB2) and its receptors EphB2, EphB3 and EphB4 in controlling arteriovenous assembly and differentiation (Adams et al., 1999; Gerety et al., 1999; Wang et al., 1998). Ephrin-B2 is selectively expressed by arterial and angiogenically activated endothelial cells (ECs). By contrast, the EphB4 receptor is preferentially expressed by venous ECs. As such, ephrin-B2 and the EphB4 receptor were established as the first markers to discriminate between arteries and veins, which has stimulated research into further defining the molecular repertoire of arterial vs venous ECs (Augustin and Reiss, 2003; Torres-Vazquez et al., 2003). EphB receptors (EphBRs) are classical receptor tyrosine kinases (RTKs) and interact with their corresponding ephrinB ligands that themselves act as transmembrane tyrosine kinase signaling molecules. EphBR-ephrinB interactions are capable of mediating bi-directional signaling events upon cell-cell contact, either into the receptor-expressing cell as `forward signaling' or into the ligand-expressing cell as `reverse signaling'.

Vascular morphogenic functions of EphBRs and their corresponding ephrin-B2 ligand in development are well established. Yet, much less is known about their roles in the adult vasculature. The arteriovenous asymmetrical expression patterns of ephrin-B2 and EphB4 are maintained in the adult (Diehl et al., 2005; Korff et al., 2006). Likewise, ephrin-B2 expression is prominently upregulated during adult neovascularization (Gale et al., 2001; Hayashi et al., 2005; Shin et al., 2001) and arteriogenic vessel remodeling (Korff et al., 2008). EphBR–ephrin-B2 interactions have recently been implicated in regulating inflammation (Ivanov et al., 2005; Yuan et al., 2004). Expression of EphBRs and their ligand ephrin-B2 has also been described in different circulating cell populations, including platelets (Prevost et al., 2002), T-cells (Yu et al., 2003), dendritic cells (de Saint-Vis et al., 2003), bone marrow hematopoietic cells (Okubo et al., 2006) and endothelial progenitor cells (Foubert et al., 2007).

We have shown that activation of endothelial ephrin-B2 reverse signaling exerts pro-adhesive and pro-migratory effects on ECs, whereas forward EphB4 signaling interferes negatively with the angiogenic cascade (Füller et al., 2003). Expression-profiling experiments in vitro and in vivo have established the microenvironmental control of EC-expressed ephrin-B2 by cytokines, contact with smooth muscle cells and biomechanical forces (Goettsch et al., 2004; Hayashi et al., 2005; Korff et al., 2006; Korff et al., 2008). Importantly, ephrin-B2 was found to be apically expressed in adult ECs (Korff et al., 2006). The apical expression of ephrin-B2 in ECs, in combination with the detection of EphBR-positive cells in the circulation prompted us to hypothesize that EphBR–ephrin-B2 interactions have a role in controlling leukocyte recruitment to the endothelium. The rationale for these experiments was also stimulated by our recent observation that arteriogenic vessel remodeling is associated with the induction of ephrin-B2 expression and concomitant monocyte recruitment (Korff et al., 2008). The present study was consequently aimed at systematically exploring this hypothesis at the cellular level. The experiments have solidly established a role of the EphBR–ephrin-B2 system in controlling monocytic cell adhesion to ECs and their subsequent transmigration.

Guided by our previous finding that ephrin-B2 is expressed on the luminal side of ECs and is upregulated during angiogenesis (Korff et al., 2006), as well as the arteriogenesis-associated upregulation of ephrin-B2 with enhanced monocyte transmigration (Korff et al., 2008), we set out to systematically analyze the cellular mechanisms of ephrin-B2- and EphBR-mediated adhesive interactions between ECs and leukocytes. Fluorescence-activated cell sorting (FACS) analysis of leukocyte EphBR expression identified several EphBR-positive leukocyte populations (Fig. 1). Mouse bone-marrow-derived hematopoietic cells, mouse peripheral blood leukocytes, mouse splenocytes and mouse macrophages (F4/80+) were all found to express EphBRs (Fig. 1A). EphBRs were also expressed by human CD14+ cells, the human leukemic monocyte cell line U937 and the murine macrophage cell line J774 (Fig. 1B). The detailed analysis of EphBR expression by FACS experiments identified the high-affinity ephrin-B2 receptors EphB2 and EphB4 (Chrencik et al., 2006) to be expressed on the cell surface of CD14+ cells (Fig. 1C).

Previous research has established attractive functions of EC-expressed ephrin-B2 (Füller et al., 2003; Hamada et al., 2003), supporting the hypothesis that ephrin-B2-positive ECs exert pro-adhesive effects on EphBR-positive circulating mononuclear cells. We consequently analyzed the adhesive properties of ECs that were retrovirally transduced to express different forms of ephrin-B2, i.e. full-length, cytoplasmically truncated (ΔC-ephrin-B2) ephrin-B2. All transduced cells also expressed GFP via an IRES site (see Materials and Methods). Transduced ECs were probed for morphological changes (live imaging by using the GFP signal), and differences in their ephrin-B2 and ΔC-ephrin-B2 expression patterns were assessed by immunoprecipitation using EphB4 linked to the Fc portion of human IgG (EphB4-Fc) followed by western blot analysis using an anti-ephrin-B2 antibody (Fig. 2A,B). Ephrin-B2 and EphB4 expressions in transduced human umbilical vein endothelial cells (HUVECs) (mock-transduced, ephrin-B2- or ΔC-ephrin-B2-transduced) were traced by reverse transcription (RT)-PCR and western blotting (supplementary material Fig. S1A,B). No changes in EphB4 phosphorylation was detected in the transduced cells (supplementary material Fig. S1B). Both, full-length ephrin-B2 as well as ΔC-ephrin-B2 were expressed on the cell surface and no obvious morphological changes were detectable in the transduced cells expressing the different constructs (supplementary material Fig. S1C). Protein levels in cells expressing transduced ephrin-B2 were compared with those in freshly isolated lung ECs expressing endogenous ephrin-B2. We found that ephrin-B2-transduced ECs expressed ephrin-B2 at levels that were comparable with those found in ECs in vivo (supplementary material Fig. S2). Likewise, no changes in the proliferative behavior of the transduced cells were observed (Fig. 2C). Corresponding to in vivo ephrin-B2 expression on the luminal side of ECs (Korff et al., 2006), cultured ECs that overexpressed ephrin-B2 presented ephrin-B2 uniformly on their apical cell surface (Fig. 2D). Similarly, ΔC-ephrin-B2 was apically expressed on the cell surface of cultured ECs (data not shown).

U937 and J774 cells preferentially adhered to ephrin-B2-overexpressing HUVECs compared with mock-transduced or ΔC-ephrin-B2-overexpressing ECs (Fig. 3A,B), suggesting a requirement of the cytoplasmic domain of ephrin-B2 for this adhesion process. By contrast, A375 melanoma cells that expressed EphB1 and EphB2 receptors did not show this pro-adhesive phenotype (data not shown). Adhesive interactions of EphBR–ephrin-B2 controlled early steps of the leukocyte adhesion cascade, as evidenced by the observation that long-term (16 hours) adhesion of J774 cells led to adhesion similar to that in mock-transduced, ephrin-B2- and ΔC-ephrin-B2-overexpressing HUVECs (Fig. 3C). Similarly, adhesion experiments under laminar flow revealed the same pro-adhesive phenotype of J774 cells adhering to ephrin-B2-overexpressing HUVECs than mock-transduced or ΔC-ephrin-B2-overexpressing cells (Fig. 3D; supplementary material Movies 1-3).

Single-cell force spectroscopy (SCFS) was used to study the initial binding properties of mock-transduced, ephrin-B2- and ΔC-ephrin-B2-overexpressing HUVECs to EphBR-expressing monocytes. GFP-positive ECs were selected and used for SCFS. The principle of SCFS is represented in Fig. 4A,B. The maximum unbinding force and unbinding work of EphBR-expressing J774 cells from ephrin-B2- and ΔC-ephrin-B2-overexpressing HUVECs following an adhesion period of 1 minute was significantly higher than the force required to remove rapidly adhering cells from mock transduced cells (Fig. 4C). Yet, the adhesive strength of ephrin-B2 and ΔC-ephrin-B2 was identical (Fig. 4C).

Pre-treatment of J774 cells with an anti-EphB2 antibody or an anti-EphB4 antibody led to a significant decrease in adhesion to ephrin-B2-overexpressing HUVECs (Fig. 5A). Pre-treatment of J774 cells with both antibodies further blocked adhesion (Fig. 5A). Accordingly, pre-treatment of HUVEC monolayers with an anti-ephrin-B2 antibody led to a significant decrease in U937 cell adhesion to ephrin-B2-overexpressing HUVECs (Fig. 5B). Baseline adhesion of mock-transduced HUVECs was unaffected (Fig. 5A,B). Taken together, endogenously expressed EphBRs of monocytes were involved in the adhesion to ephrin-B2-overexpressing ECs.

Next, we hypothesized that the pro-adhesive phenotype of endogenous EphBR-expressing monocytes to ephrin-B2 does not just depend on the reverse signaling of ephrin-B2, but also involves mechanisms that depend on the forward signaling of EphBRs. To address this question, J774 cells overexpressing EphB4-GFP or ΔC-EphB4-GFP were generated and compared with mock-transduced cells. Protein expression was assessed by western blot analysis (Fig. 6A). EphB4-GFP or ΔC-EphB4-GFP were expressed on the cell surface of J774 cells as confirmed by surface trypsinization followed by western blot analysis (data not shown). The ΔC-EphB4-GFP mutant lacked the cytoplasmic domain of EphB4 and was therefore signaling incompetent. J774 cells that expressed full-length or truncated EphB4, as well as mock-transduced J774 control cells were allowed to adhere to ephrin-B2-overexpressing HUVEC monolayers and to mock-transduced HUVEC (Fig. 6B). Overexpression of EphB4 in J774 cells led to a significant increase (1.6-fold) in adhesion to ephrin-B2-overexpressing HUVECs compared with mock-transduced HUVECs. Adhesion of J774 EphB4 cells was dependent on the cytoplasmic domain of EphB4, as evidenced by similar adhesion of ΔC-EphB4 J774 cells and mock-transduced J774 cells.

EphBR-positive monocytic cells adhered preferentially to ECs that overexpressed full-length ephrin-B2, but not to those expressing truncated ΔC-ephrin-B2 (Fig. 3A,B). We therefore compared the fate of ephrin-B2 and ΔC-ephrin-B2 expressed on the EC surface following their binding to the receptor. EC monolayers were stimulated with the soluble EphB4 (EphB4-Fc) and the cellular location of the resulting complex was traced over time. Monolayers were at different time points surface-trypsinized to discriminate between cell-surface-presented and internalized EphB4–ephrin-B2 complexes. Non-stimulated and non-trypsinized cells expressed ephrin-B2 and ΔC-ephrin-B2 on the respective cells (Fig. 7A,B, first lane). Trypsin treatment of unstimulated cells removed ephrin-B2 and ΔC-ephrin-B2 almost completely, indicating that both molecules were primarily expressed on the cell surface (Fig. 7A,B, second lane). Stimulation of ephrin-B2-overexpressing cells with EphB4-Fc induced the internalization of the resulting EphB4–ephrin-B2 complex, as evidenced by the presence of ephrin-B2 in surface-trypsinized cells. Internalized ephrin-B2 was first detected after 15 minutes and maximum internalization could be observed between 30 and 60 minutes (Fig. 7A, lanes 3-6). By contrast, ΔC-ephrin-B2, albeit being correctly expressed at the cell surface (Fig. 7B, lane 2 compared with lane 1), was not internalized following stimulation with EphB4-Fc (Fig. 7B, lanes 3-6).

Following the biochemical analysis of EphB4–ephrin-B2 complex trafficking, we performed staining experiments to analyze in detail the spatiotemporal pattern of EphB4–ephrin-B2-complex formation and internalization. Cytochemical analysis confirmed the biochemical findings, in that both ephrin-B2 and ΔC-ephrin-B2 were expressed uniformly on the cell surface of cultured cells (Fig. 7C). Stimulation of cells with EphB4-Fc for 20 minutes induced the junctional translocation of cell-surface-expressed ephrin-B2 and ΔC-ephrin-B2 (Fig. 7C). A 60-minute EphB4-Fc stimulation of cells that overexpressed full-length ephrin-B2 resulted in a granular cytoplasmic staining pattern of ephrin-B2, reflecting the internalization of the EphB4–ephrin-B2 complex (Fig. 7C). By contrast, ΔC-ephrin-B2 was still found at inter-endothelial junctions following prolonged stimulation with EphB4-Fc, reflecting the inability to internalize the resulting EphB4–ΔC-ephrin-B2 complex. Confocal microscopic analysis of ephrin-B2- and ΔC-ephrin-B2-overexpressing ECs revealed an unaltered location of the adherens junction molecule VE-cadherin (Fig. 7D), confirming an unchanged vascular integrity. VE-cadherin did not colocalize with ephrin-B2 and ΔC-ephrin-B2 (Fig. 7D). Collectively, our data established that cell-surface-expressed ephrin-B2 translocated into cell junctions upon receptor engagement prior to the cytoplasmic domain-dependent internalization of the resulting complex.

On the basis of the observed translocation of ephrin-B2 to inter-endothelial junctions, we hypothesized that ephrin-B2 expressed in ECs is functionally involved in the transmigration of EphBR-positive cells. Long-term adhesion experiments have shown that maximum adhesion of J774 cells resulted in similar numbers of J774 cells adhering to ECs that overexpressed full-length ephrin-B2 and ΔC-ephrin-B2 or had been mock-transduced (Fig. 3C). Next, HUVECs overexpressing ephrin-B2 or ΔC-ephrin-B2, or those that had been mock-transduced, were grown to confluence on Transwell filters prior to seeding of J774 cells. Confluence of the HUVEC monolayers was controlled by calcein staining (Fig. 8, top). J774 cells were allowed to transmigrate for 24 hours and were counted on the lower side of the Transwell membrane. The number of J774 cells migrating through ephrin-B2-overexpressing HUVEC monolayers was significantly higher than those migrating through ΔC-ephrin-B2-overexpressing cells and mock-transduced HUVEC monolayers (Fig. 8, center). The ephrin-B2 antibody, which blocked short-term adhesion, did not block transmigration (data not shown). Collectively, these results point towards a role of ephrin-B2 – when expressed in ECs – in the transmigration of EphBR-positive cells.

Eph receptors form the largest family of receptor tyrosine kinases, consisting of ten A- and six B-class receptors that bind their corresponding A- or B-class ephrin ligands (Pfaff et al., 2006). EphrinA ligands are GPI-anchored molecules, whereas ephrinB ligands are transmembrane molecules that are capable of transducing signals through their cytoplasmic tail. The EphBRs and ephrinB-class ligands have multiple phoshorylation sites in their cytoplasmic domain and a PDZ-binding domain at their cytoplasmic C-terminal end that interacts with molecules important for downstream signaling (Ran and Song, 2005; Su et al., 2004). Eph receptors and ephrin ligands have originally been identified as neuronal pathfinding molecules that transduce attractive and repulsive signals upon contact of cells that express corresponding receptors and ligands (Kullander and Klein, 2002); they thereby act as crucial regulators of neuronal connectivity.

Eph receptors and ephrin ligands have been identified as being widely expressed and as controlling cell-cell interactions of numerous cell types, including cells of the vascular system (Adams, 2002). The research focus of vascular Eph receptors and ephrin ligands has primarily been on developmental issues, unravelling the role of B-class Eph receptors and ephrin ligands in arteriovenous differentiation (Adams et al., 1999; Gerety et al., 1999; Wang et al., 1998), interactions between EC and smooth-muscle cells (Foo et al., 2006) and lymphangiogenesis (Makinen et al., 2005). Studies of interactive mechanisms between Eph receptors and ephrin ligands in the adult vascular system primarily concentrated on their role in tumor angiogenesis and tumor progression (Erber et al., 2006; Kertesz et al., 2006; Martiny-Baron et al., 2004; Noren et al., 2004).

We have previously observed that ephrin-B2 expression in ECs is under microenvironmental control and that ECs express ephrin-B2 on their luminal side (Korff et al., 2006). Other studies have established shear stress and hypoxic stress as major determinants of expression of ephrin-B2 in ECs (Goettsch et al., 2004; le Noble et al., 2004; Suenobu et al., 2002; Vihanto et al., 2005). The luminal expression of ephrin-B2 has prompted us to hypothesize that ephrin-B2 presented in ECs has a role in the recruitment of leukocytes by interacting with EphBRs expressed by circulating cells. This hypothesis was also supported by our recent observation that arteriogenic vascular remodeling is associated with the induction of ephrin-B2 expression and concomitant monocyte recruitment (Korff et al., 2008). We also investigated the expression of EphBRs and found them to be widely expressed by circulating cells as well as by bone-marrow-derived cells. Further expression-profiling experiments identified EphB2 and EphB4 as the preferentially and most-consistently expressed monocyte EphBRs. These findings confirm and extend previous studies demonstrating EphBR expression in platelets (Prevost et al., 2005), T-cells (Yu et al., 2003), dendritic cells (de Saint-Vis et al., 2003), bone-marrow hematopoietic cells (Okubo et al., 2006) and endothelial progenitor cells (Foubert et al., 2007).

Luminal ephrin-B2 expression and expression of EphBRs by monocytes provided the rationale for our analysis of the contribution of EphB4–ephrin-B2 adhesive interactions during monocyte recruitment. This hypothesis was also supported by recent observations associating EphB4 and ephrin-B2 with inflammatory processes (Yuan et al., 2004; Zamora et al., 2006). The subsequent experiments showed conclusively that EphB-expressing monocytes adhered preferentially to and transmigrate through ephrin-B2-expressing endothelium. This adhesion was dependent on the cytoplasmic domain of both EphB4 and ephrin-B2, which indicates distinct signaling mechanisms in ECs (ephrin-B2 reverse signaling) and in monocytes (EphB4 forward signaling). The mechanistic analysis of the bi-directional signaling processes during monocyte-to-EC adhesion is subject of future work. Surprisingly, preliminary data suggest that the PDZ domain of ephrin-B2 is not required for the preferential adhesion of monocytes to ephrin-B2-expressing ECs (data not shown). In contrast to the 30-minute cellular adhesion experiments, single cell atomic force spectroscopy (SCFS) detected adhesive forces resulting from a direct molecular interaction of the extracellular domains of ephrin-B2 with corresponding EphBRs in 1-minute short-term experiments. In these experiments, both ephrin-B2 and truncated ephrin-B2 had a similar binding strength that was significantly higher than the binding strength of mock-transduced cells. The differences between the 1-minute force-measurement experiments and the 30-minute functional experiments most probably reflect the involvement of the cytoplasmic domains of receptor and ligand and might also affect adhesion strength by affecting intracellular signal-transduction pathways.

EphBR–ephrin-B2 interactions mediate transient and interactive cell-cell processes (Zimmer et al., 2003), and are therefore perfectly positioned to contribute to the dynamic adhesive mechanisms that occur during leukocyte recruitment. Future work will have to address in more mechanistic detail at which stage of the leukocyte recruitment cascade EphBR–ephrin-B2 interactions need to be positioned, and how they interact with the classical selectin, integrin and Ig-superfamily adhesion events. For example, EphB4 activation has recently been shown to upregulate PSGL-1 in endothelial progenitor cells, thereby leading to an increased homing to lesions and incorporation into newly formed microvessels (Foubert et al., 2007). The results described in this study established a role for EphBR–ephrin-B2 interactions during monocyte adhesion as well as during transmigration. Intriguingly, stimulation experiments with EphB4-Fc revealed that activated endothelial ephrin-B2 translocates from the cell surface to inter-endothelial junctions, where it is internalized. Previous experiments have shown that this translocation is associated with CD31 (Korff et al., 2006). Conceptualizing these findings, it is tempting to hypothesize that EphBR–ephrin-B2 interactions have a role in gating adherent monocytes from the site of adhesion to inter-endothelial junctions to prepare them for transmigration. Our findings also stress the notion that heterotypic EphBR–ephrin-B2 interactions may primarily be pro-adhesive, whereas homotypic EphBR–ephrin-B2 interactions may lead to repulsive functions (Marston et al., 2003; Zimmer et al., 2003).

In summary, the interaction of EphBR-expressing hematopoietic cells with ephrin-B2-expressing ECs elicits pro-adhesive functions that lead to the translocation of EphBR–ephrin-B2 complexes to inter-endothelial junctions and also supports monocyte transmigration. Collectively, these findings suggest important roles of the EphBR–ephrinB system in controlling the recruitment of hematopoietic cells to the vascular endothelium. Further study of complementary EphBR–ephrinB expression patterns as well as the underlying signaling mechanisms will shed light into the leukocyte-adhesion cascade and will significantly expand our understanding of cell trafficking in the vasculature.

BM cells were obtained by flushing femurs of C57Bl/6 mice with HBSS, filtered through a 40 μm cells strainer (Becton Dickinson) and washed twice in cold phosphate-buffered saline. Mouse peripheral blood was collected and lymphocytes were obtained after sedimentation of erythrocytes in PBS, 5 mM EDTA, 2% dextran for 30 minutes at 37°C and further depleted of red blood cells by lysis, using a wholeblood erythrocyte lysing kit from R&D Systems as previously described (Reiss et al., 2001). Splenocytes were freshly isolated from disaggregated mouse spleens and cleared of red blood cells using a whole blood erythrocyte lysing kit (R&D Systems). Cells were incubated for 1 hour at 37°C in RPMI, then washed and resuspended to 1.5×10⁷ cells/ml as described (Warnock et al., 2000). F4/80+ cell were isolated by rat anti-mouse F4/80 antibodies coupled to anti-rat magnetic beads and sorted according to the manufacturer's instructions (MACS system, Miltenyi Biotek). Purity was determined by F4/80 FITC staining. PBMNC (peripheral blood mononuclear cells) from human blood were positively selected for monocytes using CD14+ magnetic beads according to the manufacturer's protocol (MACS system, Miltenyi Biotek). Human umbilical vein endothelial cells (HUVECs) were purchased from Promocell (Heidelberg, Germany). HUVECs cultured from passages 4-8 were used for the experiments. EC growth medium (ECGM), EC basal medium (ECBM) and the corresponding medium supplements were purchased from Promocell (Heidelberg, Germany). The human monocytic cell line U937 was cultured in RPMI 1640 using Glutamax-1 (Invitrogen, Karlsruhe, Germany). The mouse macrophage cell line J774 and the human melanoma cell line A375 were grown in DMEM L-glutamine supplemented with 4.5 g/l glucose (Invitrogen, Karlsruhe, Germany). All cells were cultured at 37°C, 5% CO₂, 100% humidity in the corresponding medium, containing 10% fetal calf serum (FCS) and 1% penicillin-streptomycin (PAA, Pasching, Austria).

Recombinant mouse ephrin-B2-Fc and mouse EphB4-Fc, goat anti-ephrin-B2 antibody, goat anti-EphB2 antibody and goat anti-EphB4 antibody were obtained from R&D Systems (Wiesbaden, Germany). Goat IgG, human IgG (Fc-fragment) and mouse IgG were purchased from Jackson ImmunoResearch (Suffolk, UK). Goat anti-human actin antibody was obtained from Santa Cruz (Heidelberg, Germany). Mouse anti-human VE-cadherin antibody was purchased from Santa Cruz (Heidelberg, Germany). Rat anti-mouse F4/80 antibody was obtained from Serotec (Dusseldorf, Germany). As secondary antibody, goat anti-human IgG-Cy3 was obtained from Sigma-Aldrich (Taufkirchen, Germany), rabbit anti-goat HRP was obtained from Dako (Hamburg, Germany), rabbit anti-mouse IgG–Alexa-Fluor-488 and donkey anti-goat IgG–Alexa-Fluor-488 were obtained from Invitrogen (Karlsruhe, Germany).

Full-length ephrin-B2 and EphB4 were cloned from freshly isolated HUVEC cDNA by using reverse transcriptase (RT)-PCR amplification, using specific primers. Ephrin-B2 constructs that lack the cytoplasmic domain were generated using primers that included the first 266 amino acid residues and lacked the terminal 67 amino acids (designated ΔC-ephrin-B2). EphB4 constructs that lack the cytoplasmic domain were generated using primers that only included the first 607 amino acid residues and lacked the terminal 380 amino acids (designated ΔC-EphB4). Sequence-controlled cDNAs were subcloned into a plib-IRES-GFP-neo expression vector for retroviral transduction. HUVECs or J774 cells were retrovirally transduced following the instructions of a retroviral expression system (Clontech, Mountain View, CA). Individual clones resistant for G418 (PAA, Pasching, Austria) were isolated and expanded. The GFP-tag was used to monitor the efficacy of transduction via live imaging using a GFP filter on the fluorescent microscope (Olympus). Protein expression was confirmed by western blotting and corresponding staining procedures.

Transduced HUVECs were seeded into 96-well tissue culture plates (800 cells per well); proliferation assays were performed 24 hours, 48 hours and 72 hours after seeding using the Cell Proliferation Kit I (Roche, Mannheim, Germany) according to the manufacturer's protocol. Quantification was performed in triplicates

Transduced HUVECs were grown to confluence in 24-well plates. U937 or J774 were pre-labelled using the PKH26 Red Fluorescent linker kit from Sigma-Aldrich (Taufkirchen, Germany) according to the manufacturer's protocol. Subsequently, 2.5×10⁵ cells were seeded on top of the confluent HUVECs. Adhesion of monocytes was allowed either for 30 minutes under agitation or for 16 hours without agitation. Correspondingly, HUVECs were pre-treated for 1 hour with anti-ephrin-B2 antibody (2 μg/ml) or corresponding IgG control (2 μg/ml) for blocking experiments. J774 cells were pre-treated for 1 hour using anti-EphB4 and anti-EphB2 antibodies (2 μg/ml) or corresponding IgG control (2 μg/ml) for blocking experiments. Non-adherent cells were washed off with PBS and the remaining adherent cells were fixed with 4% PFA for 10 minutes. Quantification was performed with the particle analysis Cell-F software (Olympus) by counting the labelled cells at 4× magnification per high-power field (HPF). Quantification was performed in three wells counting six to eight fields per well.

For flow experiments, flow chamber μ-Slides I (ibidi, Martinsried, Germany) were coated with 0.2% gelatine for 30 minutes at 37°C. Flow chambers were washed twice with PBS, and 10⁵ transduced HUVECs in 100 μl medium were plated into each chamber and grown to confluence. Subsequently, 2×10⁵ J774 cells were passed over the monolayers with a shear rate of 50 per second. Non-adherent cells were washed off with PBS and the adherent cells were fixed with 4% PFA for 10 minutes. Quantification was performed as described above. Quantification was performed by counting 20 fields per flow chamber slide.

Transduced HUVECs (5×10⁴) were grown to confluence on a transwell filter (8-μm pore size, Corning, Schiphol-Rijk, The Netherlands) in ECGM for 48 hours. Confluence was checked using calcein staining, by applying 1 μg/ml calcein-AM (Merck, Darmstadt, Germany) to a separate transwell. Pre-labelled J774 cells (1×10⁵) were added on top of the monolayers and allowed to transmigrate towards FCS (lower chamber). After 24 hours of incubation, cells were fixed with PFA and stained with Hoechst dye. Migrated cells were counted after removal of the cells on the upper side of the transwell membrane using a cotton swab. Quantification was performed with the particle analysis Cell-F software (Olympus) by counting the labelled cells at 4× magnification per high power field (HPF). Quantification was performed in three wells counting six fields per well.

To precipitate ephrin-B2 or EphB4, EphB4-Fc or ephrin-B2-Fc were coupled to proteinG-agarose as previously described (Korff et al., 2006). Cell extracts were separated by SDS-PAGE and transferred onto nitrocellulose membrane for 1 hour at 100 V. The membranes were blocked for 2 hours at room temperature with 3% BSA in TBS-T (140 mM NaCl, 10 mM Tris-HCl pH 7.4, 0.05% Tween-20) and incubated with primary antibody (2 μg/ml in 3% BSA supplemented with TBS-T) for 1 hour at room temperature. Following extensive washings, the membranes were incubated with peroxidase-coupled secondary antibody. Bound antibody was detected with ECL (Amersham Biosciences, Uppsala, Sweden) and exposure to Biomax MR films.

For fluorescence-activated cell sorting (FACS), cells were harvested by centrifugation at 800 g for 10 minutes. They were then washed once in binding buffer (20 mM HEPES pH 7.4, 2% fetal bovine serum, 2 μg/ml mouse IgG, and 0.02% sodium azide). Cells were resuspended in binding buffer at 10⁶ cells/ml and divided into 100 μl aliquots with 2 μg of primary antibody (anti-EphB2, anti-EphB4, ephrin-B2-Fc; goat IgG and human IgG as negative control) added to each aliquot. Cells were incubated at 4°C for 30 minutes in the dark, washed in binding buffer twice, and incubated with 0.5 μg/ml of corresponding secondary antibody for 30 minutes at 4°C in the dark. Cells were washed twice with binding buffer and thereafter assayed on the FACScan cytometer (Becton Dickinson).

Cells were fixed in acetone-methanol for 5 minutes, blocked for 1 hour with 3% BSA and incubated for 1 hour with 2 μg/ml mouse EphB4-Fc (R&D, Wiesbaden, Germany) and mouse anti-human VE-cadherin (Santa Cruz, Heidelberg, Germany). Following extensive washings, Cy3-conjugated anti-human Fc (Sigma-Aldrich, Taufkirchen, Germany) and rabbit anti-mouse Alexa-Fluor-488 (Invitrogen, Karlsruhe, Germany) were used as secondary antibodies. Samples were incubated for 30 minutes at room temperature. Nuclei were counterstained with Hoechst-33258 dye (Sigma-Aldrich) for 10 minutes.

Transduced HUVECs (with ephrin-B2, ΔC-ephrin-B2) were grown to confluence and stimulated with EphB4-Fc for various time points. Thereafter, cells were washed twice with PBS, fixed on ice for 7 minutes in 20 mM sodium acetate (ice-cold, pH 3.7). Cells were washed twice with ice-cold PBS and treated with trypsin (47 μM in PBS) on ice for 30 minutes to remove cell-surface-bound molecules. Pure FCS was added to stop the reaction. Ephrin-B2 was detected by western blotting as described above.

The technical details and background of single-cell force spectroscopy (SCFS) have been described elsewhere in detail (Franz et al., 2007; Ludwig et al., 2008). Cell-adhesion measurements were performed with a NanoWizard II atomic force microscope (AFM) in combination with a Cellhesion™ module (both from JPK, Berlin Germany). The AFM was operated on an inverted optical microscope (Axiovert 200, Zeiss; Göttingen, Germany) equipped with the specialized stage from the Cellhesion module. HUVECs that overexpressed either ephrin-B2, ΔC-ephrin-B2 or the empty vector (mock) were seeded on round coverslips of a diameter of 22 mm. GFP-positive cells on coverslips were selected by a GFP filter integrated in the optical microscope. Coverslips were mounted into the temperature-controlled perfusion chamber (BioCell, JPK; Berlin, Germany) of the AFM and kept at 37°C.

Concanavalin A (ConA) functionalization of cantilevers was performed as described previously (Zhang et al., 2006). In brief, tipless Arrow TL1 cantilevers (Nanoworld, Neuchâtel, Switzerland) with a nominal spring constant of 0.03 N/m were used. After plasma cleaning for 4 minutes (custom-built plasma cleaner), the cantilevers were coated with ConA in three consecutive steps. First, cantilevers were incubated overnight at 37°C in a humidified chamber in 100 mM NaHCO₃ buffer pH of 8.6 containing 0.5 mg/ml biotin-BSA. Second, cantilevers were washed three times in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 1.5 mM KH₂PO₄, 8.1 mM Na₂HPO₄, 0.5 mM EDTA pH 7.4) and incubated in 0.5 mg/ml streptavidin in PBS for 30 minutes at room temperature. Third, cantilevers were washed again three times in PBS and incubated in 0.4 mg/ml ConA-biotin solved in PBS for 30 minutes at room temperature. The cantilevers were stored in PBS at 4°C after washing.

ConA-coated cantilevers were calibrated individually with the thermal-noise method directly before the measurement with the built in AFM-Software routines. J774 cells were injected into the BioCell™ and attached to the ConA-coated cantilever using the AFM's cell capture utilities applying a gentle force of <600 pN for 3 seconds. Attached cells were allowed to establish contact to the coated cantilever for 5 minutes prior to the adhesion experiments. For force-distance measurements, the AFM was operated in constant force mode and a force of 800 pN was applied to J774 cells during contact with the underlying HUVECs. The cells were separated after 60 seconds of contact time. J774 cells were allowed to rest for 120 seconds between individual measurements. Measurements were performed in ECGM. Recorded force-distance plots were analyzed using the software supplied with the AFM (JPK-IP, version 3.1.6). The maximal unbinding force and the work required to detach the J774 completely from the HUVECs were determined.

All experiments are expressed as mean ± s.d.; differences between experimental groups were analyzed by impaired Student's t-test. Quantitative AFM experiments were analyzed by one way ANOVA analysis of variance. All groups were compared with each other using the all pairwise multiple comparison procedure (Holm-Sidak). P<0.05 was considered statistically significant

The authors acknowledge the Nikon Imaging Center at the University of Heidelberg for the use of the confocal microscope. This research was supported by the following grants from the Deutsche Forschungsgemeinschaft: SPP1190 `The tumor-vessel interface' (Au83/10-1) and SFB-TR23 `Vascular Differentiation and Remodeling', projects A3 to H.G.A., C1 to Y.R., and C5. to T.K. and M.H. H.G.A. is supported by an endowed chair from the Aventis Foundation.