Search

Objective: Vascular endothelial (VE)-cadherin is of dominant importance for the formation and stability of endothelial junctions, yet induced gene inactivation enhances vascular permeability in the lung but does not cause junction rupture. This study aims at identifying the junctional adhesion molecule, which is responsible for preventing endothelial junction rupture in the pulmonary vasculature in the absence of VE-cadherin. Approach and Results: We have compared the relevance of ESAM (endothelial cell-selective adhesion molecule), JAM (junctional adhesion molecule)-A, PECAM (platelet endothelial cell adhesion molecule)-1, and VE-cadherin for vascular barrier integrity in various mouse tissues. Gene inactivation of ESAM enhanced vascular permeability in the lung but not in the heart, skin, and brain. In contrast, deletion of JAM-A or PECAM-1 did not affect barrier integrity in any of these organs. Blocking VE-cadherin with antibodies caused lethality in ESAM -/- mice within 30 minutes but had no such effect in JAM-A -/- , PECAM-1 -/- or wild-type mice. Likewise, induced gene inactivation of VE-cadherin caused rapid lethality only in the absence of ESAM. Ultrastructural analysis revealed that only combined interference with VE-cadherin and ESAM disrupted endothelial junctions and caused massive blood coagulation in the lung. Mechanistically, we could exclude a role of platelet ESAM in coagulation, changes in the expression of other junctional proteins or a contribution of cytoplasmic signaling domains of ESAM., Conclusions: Despite well-documented roles of JAM-A and PECAM-1 for the regulation of endothelial junctions, only for ESAM, we detected an essential role for endothelial barrier integrity in a tissue-specific way. In addition, we found that it is ESAM which prevents endothelial junction rupture in the lung when VE-cadherin is absent.

Endothelial cells line blood and lymphatic vessels and form intercellular junctions, which preserve vessel structure and integrity. The vascular endothelial cadherin, VE-cadherin, mediates endothelial adhesion and is indispensible for blood vessel development and permeability regulation. However, its requirement for lymphatic vessels has not been addressed. During development, VE-cadherin deletion in lymphatic endothelial cells resulted in abortive lymphangiogenesis, edema, and prenatal death. Unexpectedly, inducible postnatal or adult deletion elicited vessel bed-specific responses. Mature dermal lymph vessels resisted VE-cadherin loss and maintained button junctions, which was associated with an upregulation of junctional molecules. Very different, mesenteric lymphatic collectors deteriorated and formed a strongly hyperplastic layer of lymphatic endothelial cells on the mesothelium. This massive hyperproliferation may have been favored by high mesenteric VEGF-C expression and was associated with VEGFR-3 phosphorylation and upregulation of the transcriptional activator TAZ Finally, intestinal lacteals fragmented into cysts or became highly distended possibly as a consequence of the mesenteric defects. Taken together, we demonstrate here the importance of VE-cadherin for lymphatic vessel development and maintenance, which is however remarkably vessel bed-specific., (© 2018 The Authors.)

CD99-like 2 (CD99L2) is a membrane protein with moderate sequence homology to CD99, which initiates cell aggregation of transfected cells and that is strongly expressed on endothelial cells, neutrophils, and lymphocytes. We showed recently that Abs against CD99L2 inhibit neutrophil, but not T lymphocyte, recruitment into inflamed tissues. In this study, we have generated conditional gene-deficient mice for CD99L2 and show by analyzing them in various inflammation models several results. First, gene ablation of CD99L2 impairs neutrophil recruitment into inflamed cremaster and peritoneum. Second, despite the strong expression of CD99L2 on peripheral neutrophils, only gene ablation on endothelial cells but not on myeloid cells affects neutrophil extravasation. Third, in contrast to our previous Ab-based results, recruitment of activated T cells into inflamed skin was impaired in mice lacking CD99L2 on endothelial cells. We conclude that CD99L2 is an essential endothelial Ag for leukocyte extravasation, which does not require homophilic interactions with CD99L2 on leukocytes.

Leukocyte extravasation depends on various adhesion receptors at endothelial cell contacts. Here we have analyzed how mouse CD99 and CD99L2 cooperate with PECAM-1. We found that antibodies against mouse CD99 and PECAM-1 trap neutrophils between endothelial cells in in vitro transmigration assays. A sequential function, as has been suggested for human PECAM-1 and CD99, could not be demonstrated. In contrast to these in vitro results, blocking CD99 or CD99L2 or gene disruption of PECAM-1 trapped neutrophils in vivo between endothelial cells and the underlying basement membrane as revealed by electron microscopy and by 3-dimensional confocal fluorescence microscopy in the inflamed cremaster tissue. Leukocyte extravasation was inhibited in interleukin-1beta-inflamed peritoneum and in the cremaster by PECAM-1 gene disruption and was further attenuated by blocking antibodies against CD99 and CD99L2. In addition, CD99 and CD99L2 were required for leukocyte extravasation in the cremaster after stimulation with tumor necrosis factor-alpha, where the need for PECAM-1 is known to be bypassed. We conclude that CD99 and CD99L2 act independently of PECAM-1 in leukocyte extravasation and cooperate in an independent way to help neutrophils overcome the endothelial basement membrane.

We have shown recently that vascular endothelial protein tyrosine phosphatase (VE-PTP), an endothelial-specific membrane protein, associates with vascular endothelial (VE)-cadherin and enhances VE-cadherin function in transfected cells (Nawroth, R., G. Poell, A. Ranft, U. Samulowitz, G. Fachinger, M. Golding, D.T. Shima, U. Deutsch, and D. Vestweber. 2002. EMBO J. 21:4885-4895). We show that VE-PTP is indeed required for endothelial cell contact integrity, because down-regulation of its expression enhanced endothelial cell permeability, augmented leukocyte transmigration, and inhibited VE-cadherin-mediated adhesion. Binding of neutrophils as well as lymphocytes to endothelial cells triggered rapid (5 min) dissociation of VE-PTP from VE-cadherin. This dissociation was only seen with tumor necrosis factor alpha-activated, but not resting, endothelial cells. Besides leukocytes, vascular endothelial growth factor also rapidly dissociated VE-PTP from VE-cadherin, indicative of a more general role of VE-PTP in the regulation of endothelial cell contacts. Dissociation of VE-PTP and VE-cadherin in endothelial cells was accompanied by tyrosine phoshorylation of VE-cadherin, beta-catenin, and plakoglobin. Surprisingly, only plakoglobin but not beta-catenin was necessary for VE-PTP to support VE-cadherin adhesion in endothelial cells. In addition, inhibiting the expression of VE-PTP preferentially increased tyrosine phosphorylation of plakoglobin but not beta-catenin. In conclusion, leukocytes interacting with endothelial cells rapidly dissociate VE-PTP from VE-cadherin, weakening endothelial cell contacts via a mechanism that requires plakoglobin but not beta-catenin.

Immature dendritic cells (DC) reside in tissues where they initiate immune responses by taking up foreign antigens. Since DC have a limited tissue half-life, the DC pool in tissues has to be replenished constantly. This implies that precursor/immature DC must be able to cross non-activated endothelium using as yet unknown mechanisms. Here we show that immature, but not mature bone marrow-derived murine DC migrate across resting endothelial monolayers in vitro. We find that endothelial intercellular adhesion molecule-2 (ICAM-2) is a major player in transendothelial migration (TEM) of immature DC, accounting for at least 41% of TEM. Surprisingly, the ICAM-2-mediated TEM was independent of beta2-integrins, the known ICAM-2 ligands, since neither blocking of beta2-integrins with antibodies nor the use of CD18-deficient DC affected the ICAM-2-specific TEM. In humans, the C-type lectin DC-specific ICAM-3-grabbing nonintegrin (DC-SIGN) was shown to interact with ICAM-2, suggesting a similar role in mice. However, we find that none of the murine DC-SIGN homologues mDC-SIGN, murine DC-SIGN-related molecule-1 (mSIGN-R1) and mSIGN-R3 is expressed on the surface of bone marrow-derived mouse DC. Taken together, this study shows that ICAM-2 strongly supports transmigration of immature DC across resting endothelium by interacting with ligands that are distinct from beta2-integrins and DC-SIGN homologues.

Human CD99 is a small highly O-glycosylated cell-surface protein expressed on most leukocytes. It was recently found to be expressed at endothelial cell contacts and to participate in the transendothelial migration (TEM) of monocytes in vitro. In order to analyze the physiologic relevance of CD99 in vivo we searched for the mouse homolog. We cloned a mouse cDNA coding for a protein 45% identical in its sequence with human CD99. Based on the cDNA, we generated antibodies against this mouse homolog of CD99, which detected the antigen on most leukocytes, on endothelia of various tissues, and at cell contacts of cultured endothelial cells. Cell aggregation of CD99-transfected Chinese hamster ovary (CHO) cells was completely blocked by anti-CD99 antibodies. The same antibodies inhibited TEM of lymphocytes in vitro, independent of whether T cells or endothelial cells were preincubated with antibodies. In a cutaneous delayed-type hypersensitivity (DTH) reaction, anti-CD99 antibodies inhibited the recruitment of in vivo-activated T cells into inflamed skin as well as edema formation. We conclude that mouse CD99 participates in the TEM of lymphocytes and in their recruitment to inflamed skin in vivo. This establishes CD99 as a valid target for interference with cutaneous inflammatory processes.