This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rao, R. M. Articles by Luscinskas, F. W. Search for Related Content PubMed PubMed Citation Articles by Rao, R. M. Articles by Luscinskas, F. W. Related Collections Acute myocardial infarction Peripheral vascular disease Mechanism of atherosclerosis/growth factors

(Circulation Research. 2007;101:234.)
© 2007 American Heart Association, Inc.

Review

Endothelial-Dependent Mechanisms of Leukocyte Recruitment to the Vascular Wall

Ravi M. Rao, Lin Yang, Guillermo Garcia-Cardena, Francis W. Luscinskas

From Vascular Science (R.M.R.), National Heart and Lung Institute, Imperial College School of Medicine, Hammersmith Hospital, London, UK; and Center for Excellence in Vascular Biology (L.Y., G.G.-C., F.W.L.), Department of Pathology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Mass.

Correspondence to Francis W. Luscinskas, Brigham and Women’s Hospital, 77 Ave Louis Pasteur, NRB 7 Rm 752, Boston, MA 02115. E-mail fluscinskas{at}rics.bwh.harvard.edu

This Review is part of a thematic series on Migration of Vascular Cells, which includes the following articles:

Mechanisms of Vascular Smooth Muscle Cell Migration

Endothelial Cell Migration During Angiogenesis

Endothelial Precursor Cell Migration During Vasculogenesis

Endothelial-Dependent Mechanisms of Leukocyte Recruitment to the Vascular Wall

Molecular Mechanisms of Endothelial Cell Migration
Kathy K. Griendling Editor

	Abstract

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
Inflammation is a fundamental process that protects organisms by removing or neutralizing injurious agents. A key event in the inflammatory response is the localized recruitment of various leukocyte subsets. Here we address the cellular and regulatory mechanisms of leukocyte recruitment to the vessel wall in cardiovascular disease and discuss our evolving understanding of the role of the vascular endothelium in this process. The vascular endothelium is the continuous single-cell lining of the cardiovascular system that forms a critical interface between the blood and its components on one side and the tissues and organs on the other. It is heterogeneous and has many synthetic and metabolic functions including secretion of platelet-derived growth factor, von Willebrand factor, prostacyclin, NO, endothelin-1, and chemokines and the expression of adhesion molecules. It also acts as a nonthrombogenic and selective permeable barrier. Endothelial cells also interact closely with the extracellular matrix and with adjacent cells including pericytes and smooth muscle cells within the vessel wall. A central question in vascular biology is the role of the endothelium in the initiation of inflammatory response, the extent of its "molecular conversations" with recruited leukocytes, and its influence on the extent and/or outcome of this response.

Key Words: inflammation • monocyte • lymphocytes • atherogenesis • adhesion molecules

	Introduction

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
There is now compelling evidence that the vascular endothelium is central to the cellular and molecular events that initiate the response of the body to infection, immune reactions, and tissue injury. It is the continuous single-cell lining of the cardiovascular system that forms a critical interface between the blood and its components on one side and the tissues and organs on the other.¹ Endothelial cell activation by a variety of stimuli including proinflammatory cytokines (eg, interleukin [IL]-1ß, tumor necrosis factor [TNF]-, interferon [IFN]-), certain bacterial endotoxins, hemodynamic factors, viruses, and thrombin can predispose to local thrombosis, loss of vessel barrier function, and rapid and robust leukocyte recruitment.² If unchecked, these alterations can contribute to cardiovascular diseases including atherosclerosis, ischemia/reperfusion injury, rheumatoid arthritis, and allograft rejection. A fundamental event in the inflammatory response is the localized and restricted recruitment of blood leukocyte subsets to tissues and organs through endothelial-dependent mechanisms. Over the past 2 decades, investigators have applied intravital microscopy of the microcirculation and complementary in vitro flow models, together with murine genetic knockout/-in and transgenic models and adoptive transfer protocols, to make dramatic strides toward a more complete understanding of leukocyte trafficking. Recruitment is a sequential, multistep adhesion cascade involving leukocyte and endothelial cell (EC) adhesion molecules that support leukocyte tethering and rolling (step 1), firm adhesion (step 2), and transmigration (step 3) (reviewed elsewhere³) under defined laminar shear stress (see Figure 1). Although the mechanisms of tethering/rolling and arrest steps are fairly well understood, the last step, transendothelial migration (TEM) (transmigration, diapedesis, emigration), has been more difficult to elucidate at the molecular level. Nonetheless, a current overview reveals significant advances in our understanding of which molecules and factors regulate transmigration, providing the focus of this review. This review is composed of 4 sections. We begin with a discussion of the cells and known adhesion pathways implicated in leukocyte recruitment and transmigration, focusing on the molecules relevant to cardiovascular diseases (eg, atherogenesis), followed by evaluation of factors that influence the location of leukocyte recruitment. The role of regulatory factors that integrate the effect of various stimuli on EC function is presented next. The last section reviews the concept that leukocyte engagement of endothelial adhesion molecules triggers important signaling in endothelium and discusses the role of such signals in leukocyte recruitment.

View larger version (43K): [in this window] [in a new window]   Figure 1. The multistep process of leukocyte recruitment.3 Initial attachment and rolling, arrest, and migration to cell–cell borders and transmigration across the vascular endothelium, shown here for monocytes. The leukocytes initially attach via selectin-mediated mechanisms along with contributions from the 4 and ß2 integrins interacting with their ligands VCAM-1 and ICAM-1, respectively. The next step is stable arrest; ß2 integrins become activated by arrest chemokines and trigger cell arrest at or near cell–cell junctions. Leukocytes then migrate to junctions and transmigrate across the vascular endothelium at both junctional and nonjunctional locations. The symbols used to represent adhesion molecules in endothelial cells are identified below each component of the figure.

	Adhesion Pathways That Regulate Mononuclear Leukocyte Transmigration: Implications for Atherosclerosis

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
The initiation of atherogenesis is multifactorial and complex and is caused by a collection of risk factors referred to as the metabolic syndrome. However, a significant body of evidence suggests that the progress of an atherosclerotic lesion strongly resembles that of a chronic inflammatory disease (reviewed elsewhere⁴). Investigators have identified several leukocyte types in early lesions, including monocytes/macrophages, T cells, B cells, dendritic cells (DCs), natural killer (NK) T cells, NK cells, mast cells, activated smooth muscle cells, as well as the vast number of products secreted by these cells (reviewed elsewhere^4,5). Mouse models of atherogenesis using LDL receptor (LDLR)- or apolipoprotein (Apo)E-null mice alone, or in combination with novel genetically modified mice, have been useful and hold promise to elucidate critical steps in atherogenesis. Most evidence points to the early recruitment of blood monocytes into the aortic intima or other lesion-prone areas of the cardiovascular system, their ingestion of lipids to form lipid-rich macrophages known as foam cells,⁶ secretion of inflammatory mediators that promote recruitment of smooth muscle and multiple types of immune cells, and cell proliferation, all of which ultimately lead to development of a complex lesion or plaque. Rupture of a complex lesion causes myocardial infarction and other serious complications. Thus, understanding the cellular pathways that regulate monocytes and other cells that accumulate in or traffic into lesions is of paramount importance to developing successful strategies to prevent, treat, or reverse chronic inflammatory processes that underlie many cardiovascular diseases.

A number of adhesion molecules have been implicated in leukocyte transmigration based on in vivo and in vitro models (Figure 1 and Table 1). Given the number of possible pathways available to leukocytes, the specificity of recruitment of the appropriate leukocyte types is likely attributable to combinations of adhesion molecules and inducible and locally expressed chemoattractants, chemokines, or cytokines and, in addition, the ability of leukocyte to transiently disrupt EC junctions including components of the adherens junctions (VE-cadherin complex^7,8) and tight junctions (occludin, claudins, junctional adhesion molecules [JAMs]) that may act as a barrier to prevent transmigration. As mentioned above, a key step in leukocyte recruitment is firm adhesion of leukocyte on the surface of the endothelium, which positions the leukocyte to migrate into the vessel wall. This step occurs through at least 2 mechanisms: chemokine-triggered activation of leukocyte ß1 and ß2 integrins, which bind to their counterreceptors on the endothelium (vascular EC adhesion molecule [VCAM]-1 and intercellular adhesion molecule [ICAM]-1/ICAM-2, and JAMs [see Figure 1 and 2; Tables I and II]); and, secondly, by selectin-dependent activation of ß2 integrins on rolling leukocytes (primarily neutrophils),^9,10 although the importance of this mechanisms for monocyte recruitment has yet to be studied.¹¹ The chemokines implicated in leukocyte recruitment in animal models of atherosclerosis include monocyte chemoattractant protein (MCP)-1 (CCL2), MCP-2 (CCL8), KC (CXCL1), MIP-1 (CCL3), MIP-1ß (CCL4), RANTES (CCL5), and fractalkine (CX3CL1), although this list is likely to expand^4,5 (Table 2). Certain chemokines have been shown to bind to endothelium and can be presented on the cell surface by glycosaminoglycans or syndecan-1 and -2 and trigger leukocyte arrest.^12,13 The ability to present chemokines remains an important topic and more in vivo work is necessary. Recent in vitro studies suggest that chemokine-triggered arrest via ß1 and ß2 integrin activation requires close localization among apical chemokines, their cell receptors, and integrin adhesion molecules.¹⁴

View this table: [in this window] [in a new window]   Table 1. Endothelial Cell Adhesion Molecules in Leukocyte Recruitment

View this table: [in this window] [in a new window]   Table 2. Chemokines Implicated in Leukocite Recruitment in Experimental Models of Artherosclerosis

View larger version (38K): [in this window] [in a new window]   Figure 2. Leukocyte recruitment to the vessel wall. A, Atherosclerotic lesions develop reproducibly in distinct regions of the cardiovascular system. Regions of high probability63 for lesion development are the lesser curvature of the aortic arch and at arterial bifurcations (highlighted in yellow). B, Atherogenesis in the ApoE–/– model on a western diet was associated with a gradual blood monocytosis and, in particular, a selective elevation in the Ly-6Chi subset.68,69 C, The Ly-6Chi monocyte subset appears to preferentially traffic to lesions as compared with the Ly-6Clo subset. The recruited Ly-6Chi monocytes become macrophages (Ly-6Clo) and Ly-6Clo become DCs (not shown for clarity). D, The vasa vasorum in the adventitia also maybe an important portal for leukocyte egress, primarily B and T cells, as well as DCs. It remains to be determined whether these 2 pools (intimal and adventitia leukocytes) remain segregated or comingle as lesion evolves.

Chemoattractants, chemokines, and other as yet unidentified factors emanating from injured, infected, apoptotic, or activated cells and tissues elicit leukocyte recruitment. However, detailed coverage of this topic is beyond the scope of the current work (see previous reviews^5,15). It is worthwhile to note that we still do not know whether a single chemokine or chemokine networks are the real agents that promote correct spatial and temporal recruitment and functional effects on leukocytes. Recent work by Tacke¹⁶ lends support to the latter hypothesis as discussed next. Also under intense investigation are the mechanisms that control chemokine- and chemoattractant-mediated leukocyte arrest and directed migration. This is likely to involve complex regulatory pathways that coordinate the actin cytoskeleton and the location and affinity of adhesion molecules and chemokine receptors (see elsewhere¹⁷ for general concepts in cell migration).

	Endothelial Cell Adhesion Molecules Involved in Mononuclear Leukocyte Recruitment

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
Several members of the immunoglobulin superfamily involved in recruitment to the vessel wall are discussed below. We highlight the well-described molecules and present other molecules that have been less well-characterized in leukocyte adhesion and transmigration (Figure 1 and Table 1).

VCAM-1 (CD102) and ICAM-1 (CD54)
VCAM-1 was initially identified as a CD11/CD18-independent endothelial ligand for mononuclear leukocytes.¹⁸ VCAM-1 recognizes leukocyte 4ß1 and 4ß7 integrins, with 4ß1 integrin considered the preferential counter-receptor. The expression of VCAM-1 is under transcriptional regulation^19,20 and is one of the earliest markers of lesions in animal models of atherogenesis.²¹ Over the past 15 years, VCAM-1 has emerged as a key inducible EC-expressed adhesion molecule that mediates monocyte recruitment to early lesions in experimental animal models of atherogenesis.^21,22

ICAM-1 is expressed on the surface endothelium and immune cells. In addition to ICAM-1, 4 other family members also bind leukocyte integrins (ICAM-2, -3, -4, -5). ICAM-1 expression is regulated primarily by gene transcription and induced to high levels by a variety of inflammatory mediators. ICAM-1 binds to leukocyte ß2 integrins (CD11/CD18); its binding to lymphocyte function–associated antigen-1 and Mac-1 are particularly well-characterized (reviewed elsewhere²³). ICAM-1 has been reported on the surface of endothelium in a dimeric form,²⁴ is constitutively expressed in the peripheral vasculature, and is upregulated at lesions. ICAM-1 has been implicated in "mature" atherosclerotic lesions²⁵; however, its role in initiation and formation of lesions is overshadowed by the contribution of VCAM-1.^22,26 One could postulate that the prominence of VCAM-1 in atherogenesis is a result of its restricted expression pattern in early lesions and in regions of vasculature predisposed to lesions, its capacity to support leukocyte arrest and firm adhesion, as well as its ability to mediate transmigration of monocytes and lymphocytes (see below for further discussion of VCAM-1 expression pattern in the aorta).

Platelet Endothelial Cell Adhesion Molecule-1 (CD31) and CD99
CD99 is a highly glycosylated transmembrane protein concentrated at the junctions of confluent ECs and is also expressed by most types of leukocytes. CD99 is not an immunoglobulin family member and does not belong to any known protein family. Platelet EC adhesion molecule (PECAM)-1 is an immunoglobulin gene superfamily member that is abundant at endothelial cell–cell junctions. Both CD99 and PECAM-1 have been demonstrated to regulate paracellular transmigration of monocytes and neutrophils, hence will be considered together (reviewed in²⁷). Interestingly, the role of CD99 or PECAM-1 in models of atherosclerosis have yet to be studied in detail (reviewed elsewhere¹⁵). In function-blocking monoclonal antibody (mAb) studies, both CD99 and PECAM-1 contribute to monocyte migration through endothelial cell–cell junctions. Both molecules work by forming homophilic binding interactions.²⁸ For example, CD99 on leukocytes interacts with CD99 on the endothelium, and PECAM-1 on the leukocyte and EC behaves in an analogous fashion.²⁹ In vitro studies with blocking mAb showed that the effect of blockade of both PECAM-1 and CD99 molecules was greater than that of either alone, and further analysis revealed that the molecules function sequentially during monocyte transmigration. In vivo studies of PECAM-1 function in inflammation have demonstrated a role in leukocyte recruitment, although the effect varies and depends on the mode of inhibition (knockout mice versus blocking antibody), the mouse strain, and the stimuli tested (see³⁰ for review). The role of CD99 in in vivo models of inflammation has been reported for lymphocyte trafficking to skin in a murine delayed-type hypersensitivity model.³¹

Junctional Adhesion Molecules-A, -B, and -C
JAMs comprise a family of six immunoglobulin-like proteins named CAR, ESAM, JAM-4, JAM-A, JAM-B, and JAM-C. The JAMs contain two extracellular immunoglobulin domains, a transmembrane domain, and a short cytoplasmic domain containing signaling motifs.³⁰ These molecules have been placed in two subfamilies: type I (CAR, ESAM, JAM-4) or type II (JAM-A, -B and –C) PDZ domain binding motifs. JAM-A, -B, and C- are concentrated at endothelial cell–cell junctions in vitro and have been implicated in leukocyte transmigration of endothelium and epithelium (reviewed elsewhere³⁰).

Function-blocking mAb to EC JAM-A inhibited chemokine-induced monocyte transmigration in an in vivo subcutaneous air pouch model, and mAbs blocked chemokine-mediated monocyte TEM in vitro.³² Thus, existing data do support a role for JAM-A in leukocyte recruitment, although the mechanisms remain incompletely understood, because blocking or ablation of JAM-A is not effective in all models, as recently reviewed.³⁰ JAM-A can dimerize in solution³³ and has been shown to bind leukocyte lymphocyte function–associated antigen-1 integrin.³⁴ However, it contributed to T-cell recruitment in vitro only when the VCAM-1 and very-late antigen (VLA)-4 pathway was blocked,³⁴ suggesting a secondary role. More recently JAM-A expression in leukocytes, rather than endothelium, has been suggested to play a significant role in the movement of leukocyte to sites of inflammation.³⁰ JAM-A has also been implicated in regulating EC migration, formation of intact endothelial monolayer,³⁵ and restitution of epithelial barrier function in vitro.³⁶ However, JAM-A–knockout mice do not appear to exhibit abnormalities in vascular development; however, its role in epithelial barrier function remains to be thoroughly investigated.

The function of JAM-B in leukocyte recruitment in models of inflammation is less clear, perhaps because of its multiple binding partners and because its expression in endothelium is restricted to high-endothelial venules within the lymph nodes and Peyer patches in mice.³⁷ JAM-B is reported to bind leukocyte VLA-4³⁸ and also interacts with JAM-C in a heterophilic manner in endothelium and in overexpression systems. One study has suggested that JAM-C participates in T-cell TEM in vitro.³⁹ Both JAM-B and JAM-C formed homodimers in solution and in a cell overexpression system (JAM-B/JAM-B and JAM-C/JAM-C) and both bind to leukocyte integrins (JAM-C binds to Mac-1, and JAM-B binds to VLA-4 integrin; reviewed elsewhere⁴⁰). A recent proposal suggests that disrupting constitutive JAM-B–JAM-C interactions can "free" surface-expressed JAM-C to interact with cells expressing Mac-1.⁴¹ This raises the question of whether the disruption also frees JAM-B to interact with VLA-4 and support adhesion of both leukocyte integrins Mac-1 and VLA-4.

JAM-C binds leukocyte Mac-1.⁴² JAM-C has been implicated in neutrophil emigration in vitro and in vivo using soluble JAM-C as a function-blocking reagent.⁴³ Other studies using mAbs that interfere with Mac-1–JAM-C binding did not inhibit neutrophil transmigration under shear flow in vitro; however, a small inhibitory effect was detected when JAM-C mAb was combined with blocking mAbs directed to PECAM-1 and CD99,⁴⁴ suggesting an ancillary role for neutrophil transmigration. This study also found that JAM-C did not cluster around adherent or transmigrating neutrophils, although ICAM-1 clustering was readily observed. JAM-C has been reported to regulate EC barrier function through interactions with VE-cadherin and Rap1⁴⁵ and has also been implicated in recruitment of monocytes across oxidized LDL-treated endothelium.⁴⁶

Poliovirus Receptor and DNAX
Additional work has suggested that EC express poliovirus receptor (CD155), an immunoglobulin superfamily member that interacts with monocyte-expressed DNAX accessory molecule-1 (DNAM-1, CD226) in MCP-1–driven TEM in a transwell assay.⁴⁷ Immunofluorescence microscopy showed that the DNAM-1-Fc chimera preferentially localizes to endothelial cell–cell junctions and specifically recognizes endothelial-expressed poliovirus receptor. Studies revealed that monocyte migration to MCP-1 was blocked by mAbs directed to DNAM-1 and poliovirus receptor. These data are of interest, but future studies are necessary to characterize this pathway in monocyte (and T-cell) adhesion and transmigration of endothelium in vivo and in other in vitro models of inflammation.

CD47 or Integrin-Associated Protein
CD47 is broadly expressed by most cell types, including endothelium, epithelium, platelets, and most leukocytes. Two ligands for CD47 have been reported, thrombospondin-1 and leukocyte-expressed signal-regulatory protein (SIRP)- (reviewed elsewhere⁴⁸). A role has been reported for the CD47–SIRP- pathway in monocyte transmigration of rat cerebral ECs.⁴⁹ This finding is consistent with a significant inhibitory effect of anti-CD47 mAbs on neutrophil trans–endothelial cell migration to IL-8 or neutrophil TEM of TNF-–activated human umbilical vein ECs (HUVECs) in vitro⁵⁰ and a delay in neutrophil emigration in CD47-null mice subjected to a model of bacterial peritonitis.⁵¹ Crosslinking CD47 in B-cell lines triggered polarization with lamellipodia formation.⁵² In addition, CD47 has been reported to participate in Jurkat T-cell arrest on cytokine-activated endothelium under shear flow⁵³ via CD47-dependent modulation of 4ß1 integrins. There are no reported studies on the role of CD47 in T-cell or monocyte trafficking in vivo. Future studies are necessary to gain a better understanding of the interactions between CD47 and matrix proteins and integrins and how these interaction impact leukocyte recruitment.

In summary, there are multiple leukocyte adhesion pathways and a myriad of proinflammatory chemokines, chemoattractants, and lipids that in various combinations mediate recruitment of several leukocyte types. Moreover, the EC in different vascular beds can exhibit very different phenotypes, including specialized morphologies, functions, and different set points,⁵⁴ adding additional complexity. Thus, a major challenge remains to determine whether there is a hierarchy among these pathways in various vascular beds during the inflammatory response.

	Mechanisms Regulating Leukocyte Transmigration to Vessel Wall

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
The mechanisms underlying the process of transmigration continue to be fruitfully studied. An area of emerging interest, notably in atherosclerosis, is the factors that determine the location of leukocyte recruitment. This can be broadly divided into (1) regional, the effect of flow patterns on leukocyte recruitment; (2) lesional, where recruitment occurs in a developing lesion; and (3) cellular, at a cellular level, where transmigration of endothelium occurs.

Regional Recruitment
An important consideration in the pathogenesis of arterial lesions is the effect of disturbed flow on EC function and, specifically, on leukocyte recruitment. Whereas flow in smaller vessels and straight arteries is laminar, patterns of flow in large arteries are complex. The location of atherosclerotic plaques correlates well with areas of disturbed flow, notably bifurcations and branch points. A summary of all the changes seen in ECs under different conditions of flow is beyond the scope of this review and has been reviewed elsewhere.^55,56 A number of different observations pertinent to cell recruitment have been made and, depending on the assay and the cell type, a number of different phenotypes have been identified (reviewed by Matharu et al⁵⁶). The overriding message is that alteration of blood flow from laminar to disturbed shear stress causes a change in expression of adhesion molecules (notably ICAM-1, VCAM-1, and E- and P-selectin) and the synthesis of critical chemokines (MCP-1, IL-8, RANTES, and Gro-) and hence can influence leukocyte recruitment.

Only a few in vitro studies have examined the effects of flow on leukocyte recruitment. Early studies examining binding of leukocytes to endothelium grown in either static or fluid shear flow conditions found increased binding to HUVECs grown under shear conditions.^57,58 These increases were attributed to upregulation of ICAM-1, although other mechanisms are likely to be involved.⁵⁹ Other studies have reported enhanced binding of a monocytic cell line to ECs grown under oscillatory flow conditions compared with cells grown without shear.⁶⁰ A genuine concern regarding these earlier reports is that all used either a leukocytic cell line or a static binding assay (or both). In contrast, a more recent (and robust) study examining ECs grown under different conditions of shear found very little adhesion of neutrophils.⁶¹ In the same study, the effects of shear preconditioning on cytokine-induced recruitment were examined. In response to TNF-, neutrophil recruitment (adherence and transmigration) was reduced in a shear-dependent manner, with a concomitant reduction in cytokine-induced expression of E-selectin, IL-8, and Gro-, whereas no such difference was observed for IL-1ß.⁶² The mechanism underlying these observations remains unidentified, but the emergence of flow-induced transcription factors such as Kruppel-like factor 2 (KLF)2 might provide an explanation (see below for detailed discussion on KLF2).

A series of elegant studies has examined the effects of disturbed flow on adhesion molecule expression and leukocyte recruitment in en face sections of the aortic arch of wild-type and atherosclerotic mice. Firstly, areas predisposed to atherosclerotic plaque formation were mapped in sections from LDLR^–/– mice, using oil red O as a marker.⁶³ High expression of VCAM-1 was found in these high-probability areas on the lesser curve of the aortic arch, compared with reduced expression in low-probability areas, namely the greater curve. Next, assessment in wild-type mice of EC in the high-probability areas demonstrated a large variation in cell shape and orientation. In addition, high levels of p65, IB, and IBß were found in these areas, and robust nuclear translocation of p65 was detected either following lipopolysaccharide stimulation or, in the case of the LDLR^–/–, after cholesterol feeding.⁶⁴ Translocation of p65 correlated with upregulation of its target adhesion molecules, VCAM-1, and E-selectin. More recently, this work has been extended to demonstrate inflammatory gene expression in high-probability areas as well as abundant CD68⁺ cells and significant numbers of CD3⁺ cells in the intima in wild-type mice.⁶⁵ Cells were also observed in the adventitia, but accumulation was the same in high-probability or low-probability areas. CD68⁺ cells in the intima had a DC phenotype and expressed DC markers. These cells were derived from the bone marrow and accumulated in the intima, but not the adventitia, in a VCAM-1–dependent manner. The role of adventitial recruitment is discussed below.

A recently published study has examined the effect of either low shear stress (LSS) or oscillatory shear stress (OSS) on the development of atherosclerotic lesions in the carotid artery of ApoE^–/– mice.⁶⁶ Particular attention was paid to the differential expression of the proatherogenic chemokines at 3 time points (early, 1 week; developing, 3 weeks; and late lesions, 9 weeks) in areas of the artery artificially exposed to LSS or OSS by placement of a cast. Both patterns of shear stress result in the development of atherosclerotic lesions. In areas exposed to LSS, these lesions appeared earlier, were larger, and appeared more "unstable" (larger necrotic core, thin fibrous cap) than their OSS counterparts, where a smaller, more stable lesion developed later. Both LSS and OSS stimulated expression of MCP-1 (CCL2) at early time points, whereas IP-10 and KC appeared to be preferentially expressed in LSS areas. However, at later time points, only CX3CL1 (fractalkine) was expressed at significant levels, largely restricted to LSS areas. This resulted in accumulation of CX3CR1 monocyte-derived macrophages. Antibody inhibition of CX3CL1 lowered cellular recruitment (both CD68⁺ and adventitial mast cells) and lesion size and increased cap thickness in LSS areas, although no reduction in CX3CR1 cellular recruitment was detected. This study may therefore add to our understanding of the mechanisms underlying the emerging complexities of monocyte recruitment in atherosclerosis.

Recent studies have revealed new insight into the heterogeneity of monocyte subsets and their trafficking in lesions in murine models of inflammation and atherogenesis. As mentioned above, there had been limited characterization of the actual dynamics of leukocyte trafficking in lesions, and the aggregate information resolved only a momentary snapshot of a dynamic process. Moreover, the tools to define heterogeneous populations of monocytes have only recently become available. The advent of highly sensitive cell imaging techniques and the development of cell surface markers that identify monocyte subsets have enabled high-resolution in situ monitoring of monocyte trafficking during various stages of atherogenesis in mouse models.^16,67–69 The current paradigm proposes at least 2 (probably 3) monocyte subsets in mice, which fortuitously show good correlation to human subsets, although the Ly6C antigen is not yet defined in humans (see a previous review⁷⁰ and references therein), making the murine findings valuable (see Figure 2). Thus, in humans the "inflammatory" monocyte is identified by surface markers as CD14^hi/CD16^–/CCR2⁺/CX3CR1^lo, and the mouse counterpart is identified as Ly6C^hi/CD62L⁺/CX3CR1^lo/CCR2⁺. The second subset is considered the "resident" monocyte subset and is identified in humans as CD14^lo/CD16⁺/CX3CR1^hi, whereas the murine characteristics are Ly6C^lo/CD62L^–/CX3CR1^hi/CCR2^–. Of interest is the recent work by 2 groups independently reporting preferential recruitment of the Ly6C^hi monocyte subset in lesions in ApoE^–/– mouse models of atherogenesis. Swirski et al used a variety of approaches to demonstrate that Ly6C^hi blood monocytes accumulate in circulating blood of ApoE^–/– mice fed a Western diet and that this population of cells was preferentially recruited to lesions and differentiated into macrophage-like "foam" cells⁶⁸ (refer to Figure 2). Adoptive transfer studies with Ly6C^hi monocytes revealed a requirement for CCR2 to enter lesions; once in lesions, the transferred Ly-6C^hi cells rapidly differentiated into macrophages expressing Ly-6C^lo. Interestingly, ApoE^–/– mice on a Western diet with atorvastatin for 25 weeks showed reduced serum cholesterol and blood monocytosis, as well as lower Ly-6C^hi cell burden in various organs. These findings shed light on monocyte heterogeneity in atherogenesis and identify a monocyte subset with a high propensity to traffic to lesions and develop into lesion macrophages. In the second study, Tacke et al also used the ApoE-null mouse on a Western diet to address the role of chemokine receptors CCR2, CCR5, and CX3CR1 in monocyte subset trafficking to inflammatory sites and lesions.¹⁶ These authors also reported preferential recruitment of Ly-6C^hi cells (identified by a labeling protocol different from the work by Swirski et al⁶⁸) as compared with Ly-6C^lo cells and that Ly-6C^lo cells recruited into lesions can give rise to CD11c⁺ DC-like cells. These studies also revealed that CCR2, CX3L1, and CCR5 regulated the migration of these 2 monocyte subsets into lesions, although a caveat is that the experimental approach required invasive surgical and tissue transplant protocols. These studies confirm the value of murine models in the study of human cardiovascular disease and provide clues as to which subset(s) are dominant in atherogenesis. A logical next step would be to identify the human counterpart of murine Ly6C^hi monocytes and to study their relevance in human monocyte subsets and, if appropriate, in human disease.

Local Recruitment
In vivo examination of leukocyte interactions in arteries has been difficult until recently. Roles have been established for a number of different adhesion receptors, notably VCAM-1 and the combination of E- and P-selectin, and chemokine receptors, notably CCR2, CXCR2, CX3CR1, and CCR5. This progress has been made possible by genetically modified mice or specific inhibitors and assessment of the histological composition of plaque rather than direct examination of the effects on leukocyte recruitment. A number of techniques have changed this approach. Initially, experiments on perfused ex vivo carotid segments demonstrated a critical role for PSGL-1/P-selectin and VLA-4/VCAM-1, as well as a minor role for VLA-4/CS-1 in mononuclear cell adhesion in arteries harvested from Western diet–fed atherosclerosis-prone mice.⁷¹ Interestingly, although blockage of KC (Gro-) did inhibit mononuclear cell arrest, this model revealed no effect of blocking JE (murine MCP-1).⁷²

The subsequent development of intravital microscopic examination of large arteries (carotid or descending aorta) has allowed the visualization of dynamic leukocyte–EC interactions on inflamed arteries or developing plaques.⁷³ An early study in 2000 demonstrated the role of endothelial selectins and 4ß1 (VLA4) integrin in mediating leukocyte rolling events on cytokine-activated inflamed aorta.⁷³ When this technique was extended to examine recruitment on developing plaques, 2 interesting findings emerged. The first was that under the higher shear conditions of the arterial circulation and on early atherosclerotic plaque, secondary capture of leukocytes was an important mode of recruitment. Secondary capture occurs when already-adherent leukocytes mediate a P-selectin glycoprotein ligand (PSGL)-1/L-selectin–dependent capture of other leukocytes from free flow, resulting in rolling strings of leukocytes, a phenomenon we and others have described recently in vitro.⁷⁴ The second important point was that established plaques were recruited on the periphery of the lesion, was P-selectin-dependent, and was modulated by E-selectin but not 4ß1 integrins.⁷⁵ These selectin-dependent interactions were transient and rarely progressed to firm adhesion, secondary capture was not seen, and (perhaps surprisingly) the leukocyte populations involved were composed largely of neutrophils. Intravital microscopy of carotid arteries also has led to confirmation that activated platelets, found in the serum of patients with atherosclerosis, can bind to monocytes via P-selectin/PSGL-1.⁷⁶ This results in activation of monocyte 4ß1 integrin and enhanced binding of monocyte-platelet complexes to endothelial VCAM-1. The same in vivo studies confirmed the established in vitro finding that activated platelets bind endothelium. This can result in enhanced recruitment of monocytes attributable to secretion by endothelial-bound platelets of RANTES, CCL5, and PAF/CXCL4 as well as stimuli (IL-ß) that induced expression of adhesion molecules and chemokines (reviewed elsewhere^77,78). Repeated injection of activated platelets into ApoE^–/– mice caused atherosclerotic lesions to increase by 40% over 12 weeks.

The use of intravital microscopic techniques has allowed visualization of tethering, rolling, and arrest of leukocytes to the aortic luminal endothelium, but transmigration of leukocytes remains difficult to study. Anesthetic and microscopic limitations have thus far prevented direct examination of transmigration of any of the leukocyte subsets that may be involved in atherosclerosis. However, as discussed above, VCAM-1 expression in lesion-prone areas has been convincingly demonstrated, as has its role in early atherosclerosis.²² In recent experiments, labeling of monocytes in vivo with 5-bromodeoxyuridine has demonstrated the accumulation of CD68⁺ cells in the intima of lesion-prone areas within 24 hours in normal mice. This accumulation is VCAM-1 dependent, as it was absent in hypomorphic VCAM-1–deficient (VCAM-1^D4D/D4D) animals and thus provides indirect evidence of luminal recruitment of cells.⁶⁵

Another area of interest is cell trafficking in the vasa vasorum of the luminal adventitia (refer to Figure 2). This region is also important in other aspects of atherogenesis including production of adipokines and progenitor cells. Resident DCs have been described at the adventitia–media border in other medium-large artery diseases, notably giant-cell arteritis (reviewed elsewhere^79,80). The precise role in atherosclerosis of immunocompetent cells in this region is not known, but their presence has been documented throughout the lifetime of a lesion. The presence of cellular infiltrates was first described some years ago⁸¹; more recent characterization has identified predominantly a B-cell infiltrate, but T cells, DCs, and CD68⁺ monocytes have also been identified. In advanced lesions, organized adventitial lymphoid aggregates have been detected, which are similar to those observed in autoimmune conditions such as Hahimoto thyroiditis and rheumatoid arthritis.⁸² Organized structures were observed largely in the presence of medial thinning of aortic sections and within them, von Willebrand factor–positive ECs express HECA452 reactivity akin to that seen in high-endothelial venules. A study in old, cholesterol-fed ApoE^–/– mice reiterated these findings and demonstrated a time-dependent accumulation of B and T cells as well as the presence of DC, NK, and plasma cells.⁸³ Again, this study identified breaches in the media coincident with these structures. These observations have thus led to the proposition that antigen presentation and B-cell maturation might occur in these adventitial structures.

A recently published study examined B- and T-cell accumulation in the arterial wall in wild-type and in ApoE^–/– mice in early and late disease. Both B and T cells can be identified in the adventitia of wild-type mice, and they appear to constitutively traffic via the vasa vasorum (rather than via the lumen) in an L-selectin–dependent manner; cellular recruitment increased in ApoE^–/– animals (refer to Figure 2). The same study also reported the presence of lymphoid aggregates in advanced cholesterol-induced disease in the ApoE^–/– mouse, and, although cellular recruitment is once again partially L-selectin-dependent, no staining for its high-endothelial venule ligand, MECA79, was detected. However, to confirm functionality, in situ T-cell proliferation was shown to be induced by adoptive transfer of DC pulsed with antigen.⁸⁴

Taken together, these studies suggest the vasa vasorum may serve as a portal of entry for leukocytes in the early and late stages of atherosclerotic disease. Adoptive transfer and competitive homing studies suggest that the mechanisms of cellular recruitment involve L-selectin and immunostaining shows the presence of MIP-1. Our own studies suggest that adventitial CCR2⁺ monocyte recruitment might also occur very rapidly following imposition of a cholesterol diet (R.M.R., unpublished observations, 2006). In this case, it is possible that the source of the relevant chemokine (CCL2/MCP-1) might be adipocytes or fibroblasts. However, the relationship between intimal and adventitial recruitment remains unclear. A number of intriguing studies have identified breaches in the tunica media during lesion development, suggesting cell trafficking across the arterial wall in late disease, although this would be unlikely in earlier lesions, where the media is intact. Many experimental models of adventitial trafficking involve balloon dilatation or wire injury, resulting in disruption of the media; alternative strategies will be required to address this issue.

Cellular Recruitment
Early investigators of leukocyte transmigration appeared to provide proof of migration between rather than through ECs (eg,⁸⁵), although studies from the 1960s challenged this idea. However, based on data from serial electron microscopy studies, the long-held paradigm was that leukocyte TEM occurred in a paracellular fashion and only in certain specialized vascular beds (eg, the retina and the lung) was a transcellular route reported (well reviewed by Muller⁸⁶). Recent research has largely focused on the roles of leukocyte and endothelial junctional molecules in this process.

Over the last few years, there is renewed interest in the precise location of endothelial interactions. Prompted by a recent and seminal study⁸⁷ of neutrophil transmigration in guinea pig skin in response to the peptide FMLP, in vitro studies have attempted to examine the underlying mechanism for transcellular rather than paracellular migration. Marked cytoskeletal changes in endothelium during the migratory process have also been described. In certain in vitro systems, microvilli enriched with either ICAM-1 or VCAM-1 appear to engulf adherent leukocytes.⁸⁸ The precise functions of this docking structure are unclear: it is possible that resultant transmigration is, in fact, inhibited rather than augmented. Over the last 2 years, we and other investigators have reported that a variable but significant fraction of leukocytes undergo transcellular migration. Fluorescence microscopy was used to locate endothelial cell–cell junctions with antibodies to VE-cadherin. In one study, cells were incubated under static conditions with stimulated HUVECs and allowed to transmigrate for a defined time period. Transcellular migration, observed in 5% to 11% of episodes using neutrophils, monocytes, or T lymphoblasts,⁸⁹ was strongly associated with the formation of ICAM-1 projections (which the authors call "transmigratory cups") as well as with linear ß2 integrin clustering. The ICAM-1 projections were thought to be separate from the transmigratory pore but extended down the pore to the basal surface. Adherent (but not transmigratory) cells were associated with clustering of VCAM-1, and a weaker association was reported for ICAM-2 and PECAM-1. We adopted a different strategy: observing live-cell transmigration under shear conditions. We detected higher levels of transcellular migration following induction of HUVECs with TNF- to increase ICAM-1 expression as well as with ICAM–green fluorescent protein–transfected HUVECs.⁹⁰ This higher level of transcellular migration was also associated with nonjunctional adhesion of neutrophils. Absence of the cytoplasmic domain of ICAM-1 reduced transmigration generally and transcellular migration in particular. Interestingly, we could not detect transcellular migration of freshly isolated blood-derived CD3⁺ T cells under the same conditions. More recently, a study examined the effect of antibody cross-linking ICAM-1 on TNF-–stimulated HUVECs and demonstrated redistribution of ICAM-1 to caveolin-rich areas where F-actin stress fibers converged.⁹¹ Using a combination of time lapse and total internal reflection microscopy, transcytosis of ICAM-1 to the basolateral surface was seen. These results were reproduced using T lymphoblasts, with resultant transcellular migration of the cells in approximately 10% of cases (again using VE-cadherin staining to define junctions). A ring of caveolin and ICAM-1 formed around the transmigrating lymphoblast, with associated accumulation of F-actin. Knockdown of caveolin-1 using small interfering (si)RNA partially blocked transcellular but not paracellular migration. Of considerable interest, the extent of transcellular migration increased on HDMVEC, a microvascular cell type with some lymphatic properties. The role of intermediate filaments was assessed in another study, which showed transcellular migration of fresh peripheral blood mononuclear cells, but not neutrophils, on activated HUVECs, with associated staining for vimentin on microvilli from both leukocytes and ECs.⁹² Homing of vimentin (vim)^–/– lymphocytes to lymph node and spleen was impaired in vitro and in vivo in WT or vim^–/– mice. Transmigration in the inflamed cremaster of vim^–/– mice was, however, enhanced with an associated increase in endothelial permeability. In static assays, followed by fixation and staining, vimentin-rich "cups" were seen engulfing lymphocytes. In live cell time experiments under flow conditions, vimentin clustered around adhered and transmigrating cells. Finally, in ECs from vim^–/– mice, expression of ICAM-1 and VCAM-1 was markedly reduced, and lymphocytes exhibited low ß1 integrin expression and a loss of polarization.

A number of interesting questions arise from these studies. (1) Which cells are using these pathways? One study examined the effects of activated lymphoblasts, which appear to use a transcellular pathway. A criticism is that these cells are unlikely to exist in the bloodstream. Our own studies, however, have found no evidence for CD3⁺ T-cell transcellular migration of immortalized human umbilical vein endothelium, only for neutrophils, which is in direct conflict with the results obtained by Nieminen et al.⁹² Although certain technical differences exist in these studies, there is no viable explanation for these apparent contradictions. (2) What is the significance of the transmigratory cup? Leukocyte adhesion and transmigration prompts clustering and then reorganization of both ICAM-1 and VCAM-1, as well as cytoskeletal components F-actin, tubulin, and vimentin and several actin-binding proteins (ezrin, moesin, cortactin, and -actinin). Is it pertinent that these structures have been detected more frequently in assays under static conditions in cells that are subsequently fixed? There is little doubt, however, that significant cytoskeletal rearrangement is required of both leukocyte and ECs. (3) Do these pathways coexist? It seems from all the studies that only some cells use the transcellular pathway. It also appears that cells of microvascular origin support more transcellular migration than do HUVECs. One explanation offered is that in cells with tight junctional organization, such as the brain, leukocytes might prefer a transcellular route, whereas in post-capillary venules, where junctions are poorly organized, they might use a paracellular pathway. Another explanation could be that the strength of the triggering stimuli dictates what fraction of cells use the transcellular pathway. High levels of endothelial ICAM-1 do appear to favor transcellular TEM of neutrophils in certain models. (4) How might this be pertinent to atherosclerosis? Firstly, if microvascular trafficking is more susceptible to transcellular migration than large-vessel trafficking, it would be more relevant in the vasa vasorum, where significant mononuclear cell trafficking occurs in atherosclerosis (refer to Figure 2). On the other hand, the work from Millan et al⁹¹ suggests an alternative: that, following cross-linking or lymphoblast binding, ICAM-1 localizes to areas of stress fiber formation, which prompted transcellular migration. In areas of the vascular tree exposed to disturbed shear stress (as opposed to laminar shear), stress fiber formation is altered and the EC shape becomes more polygonal. We postulate that an altered cytoskeleton and density of adhesion molecules (eg, VCAM-1, ICAM-1, PECAM-1, CD47, CD99, and DNAM-1), and not just altered junctional components, may also contribute to the process of transcellular migration. Needless to say, exciting work lies ahead to answer these questions.

	Novel Endothelial Transcriptional Regulators of Inflammation and Leukocyte Recruitment

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
The functional phenotype of the vascular endothelium is dynamically regulated in space and time by several types of stimuli. In particular, the effects of proinflammatory cytokines and biomechanical forces have emerged as critical determinants for endothelial phenotypes in physiology and disease.⁹³ Furthermore, genomic studies on the interplay between these humoral and mechanical factors have begun to provide information about signaling pathways in the endothelium critical for the orchestration of an inflammatory response.

For example, using transcriptional profiling to assess changes in gene expression induced in cultured ECs by laminar shear stress, a well-characterized antiinflammatory biomechanical stimulus, and proinflammatory cytokines, 2 studies identified the transcription factor KLF2 as a gene upregulated by laminar shear stress and down regulated by TNF- or IL-1ß.^94,95 Based on these observations and the pattern of expression of KLF2 in regions protected from the development of atherosclerosis in human arteries, it was suggested that KLF2 protects against the development of atherosclerotic lesions.⁹⁴ KLFs are members of the zinc finger–containing family of transcription factors that can act as activators or repressors of gene expression and are important for the regulation of proliferation, migration, differentiation, and function of several cell types.⁹⁶ SenBanerjee et al showed that KLF2 overexpression in human umbilical vein ECs muted the IL-1ß–dependent induction of the proinflammatory adhesion molecules VCAM-1 and E-selectin (but not ICAM-1) while increasing the expression and activity of endothelial NO synthase.⁹⁵ Subsequent studies using complex models of arterial waveforms characteristic of the in vivo atherosclerosis-resistant (atheroprotective flow) and atherosclerosis-susceptible (atheroprone flow) regions of the human carotid artery demonstrated a selective upregulation of KLF2 by atheroprotective flow.⁹⁷ Collectively, these observations suggested that KLF2 is a transcriptional regulator of the atheroprotective flow-dependent modulation of inflammatory responses in endothelium. Indeed, recent studies by Parmar et al using a systems approach identified KLF2 as a critical integrator of the endothelial global transcriptional responses to atheroprotective flow and demonstrated that overexpression of KLF2 repressed the induction of 32 cytokines/chemokines (eg, IL-6, IL-8, RANTES, IP-10, MCP-1) in response to IL-1ß stimulation.⁹⁸ Moreover, expression of endothelial KLF2 promotes a robust antiinflammatory program, including the protective cytokines ELAFIN and IL-11. The mechanisms by which KLF2 suppresses the expression of proinflammatory stimuli in static and flow conditions are just beginning to be elucidated. Likely mechanisms include the inhibition of nuclear factor B activation by preventing recruitment of CBP/p300, an essential coactivator,⁹⁵ and the inhibition of the nuclear activity of the transcription factor ATF2.⁹⁹ Functionally, KLF2 overexpression in cultured ECs leads to a strong suppression of T-cell adhesion or rolling interactions with ECs under flow conditions.⁹⁵ Moreover, the flow-mediated inhibition of IL-1ß–mediated leukocyte adhesion to cultured ECs is potently reduced in the absence of KLF2.⁹⁸ In addition to biomechanical stimuli, KLF2 expression is also increased by 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, the "statins,"^100,101 a class of drugs originally designed to reduce cholesterol levels, implicating KLF2 in the well-characterized statin-mediated pleiotropic beneficial vascular effects.¹⁰⁰

In summary, KLF2 has emerged as a critical integrator of flow-mediated changes in the EC phenotype, including antiinflammatory and vasoprotective states, as well as a novel endothelial transcriptional regulator of leukocyte adhesion. Further exploration of this pathway may allow the identification of pharmacological modifiers of KLF2 expression capable of mimicking the KLF2-mediated endothelial vasoprotective effects. Furthermore, this example validates the use of a systems biology approach to identify novel molecular pathways responsible for complex cellular phenotypes and function in the vessel wall. Continued study using similar approaches should improve our understanding of unsolved questions in the field, including how leukocyte adhesion and transmigration modulate the endothelial phenotype and how the resolution of inflammation is orchestrated on a spatial and temporal level by the endothelium. The answers to these questions will undoubtedly increase our understanding of the inflammatory process and may lead to new strategies for the prevention and treatment of several vascular diseases.

	Signaling via Endothelial Cell Adhesion Molecules and Their Role in Transmigration

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
There is strong evidence that ligation of EC adhesion molecules initiates intracellular signaling cascades. However, although the integration of these signaling cascades and how they facilitate leukocyte transmigration and other functions have yet to be fully worked out, several recent studies have made important progress. Briefly, E-selectin ligation by leukocytes (HL-60) triggers its association with cytoskeleton and initiates outside-in signaling.¹⁰² Pathways activated include the formation of a Ras/Raf-1/phospho-MEK supramolecular complex, extracellular signal-regulated protein kinase (ERK)1/2 activation, and upregulation of c-fos. Similar signaling readouts have been reported for leukocyte binding or mAbs crosslinking of PECAM-1, P-selectin, and VCAM-1 (reviewed elsewhere¹⁰³). VCAM-1 clustering leads to activation of Rac1 and p38 mitogen-activated protein kinase (MAPK) and generation of reactive oxygen species (ROS) and inhibition of Rac signaling reduced transmigration of the leukocyte cell line U937.¹⁰⁴ The cytoplasmic tail of PECAM-1 contains 2 immunoreceptor tyrosine-based inhibition motifs that are tyrosine phosphorylated by src kinases.¹⁰⁵ These immunoreceptor tyrosine-based inhibition motifs act as scaffolds for docking of signaling molecules. PECAM-1 ligation has been shown to result in MAPK activation. Further work is necessary to determine whether PECAM-1–dependent signaling is involved in leukocyte transmigration.

Signaling through ICAM-1 has been studied in more detail than other molecules, and multiple studies have found that the cytoplasmic tail is essential for leukocyte transmigration and cytoskeletal remodeling. Engagement of ICAM-1 expressed by cytokine-activated endothelium by adherent leukocytes triggers elevations in cytosolic Ca²⁺ concentration and myosin contractility,^106,107 as well as activation of small GTPases (Rho family¹⁰⁸), p38 MAPK,¹⁰⁹ and the tyrosine kinase p60^Src.¹¹⁰ A prominent substrate for p60^src and other src family kinases in endothelium is the actin-binding protein cortactin.¹¹¹ Most of these above events depend on signaling through the cytoplasmic tail.^{90,107,112,113} An attractive candidate to integrate these ICAM-1 signals, and possibly signals from E-selectin and VCAM-1, is the multifunctional scaffold protein cortactin.¹¹¹ Biochemical studies have shown that ligation of ICAM-1 triggers cortactin association with ICAM-1 and E-selectin in lipid rafts as well as cortactin phosphorylation by Src tyrosine family kinases.¹¹⁴ In addition, cortactin contains an actin-binding domain and a separate N-terminal domain that stimulates actin polymerization (in vitro) by the Arp2/3 complex.¹¹⁵ Arp2/3 is the dominant mechanism for the explosive actin polymerization characteristic of platelet shape change after stimulus-induced activation.¹¹⁶ Our laboratory has reported that cortactin coordinates ICAM-1 clustering and actin cytoskeleton remodeling during neutrophil adhesion and transmigration.^113,117 We found that cortactin–green fluorescent protein clustered around transmigrating neutrophils and colocalized with both ICAM-1 and F-actin in a ring-like cluster.¹¹³ Furthermore, inhibition of Src tyrosine kinases inhibitors PP2 or SU6656, or siRNA of cortactin, reduced neutrophil transmigration by 50%. The tyrosine phosphorylation of cortactin was critical for transmigration, because expression of a cortactin mutant (cortactin3F) in which 3 critical tyrosines were mutated to phenylalanine did not rescue transmigration in siRNA cortactin-treated endothelium, whereas a wild-type cortactin construct did. Subsequent studies showed that siRNA of cortactin also decreased both actin and ICAM-1 clustering around adherent or transmigrating neutrophils.¹¹⁷ Interestingly, even though siRNA of cortactin or Src kinase inhibitors have striking inhibitory effects on ICAM-1 clustering around leukocytes and largely prevented cytoskeletal remodeling, transmigration was reduced by only 50%, suggesting other scaffold proteins, signaling pathways, and/or adhesion molecules participate in neutrophil transmigration.

Others studies have implicated nonmuscle -actinins as important binding partners for ICAM-1 during transmigration in vitro. A recent study by Celli and colleagues,¹¹⁸ using a yeast 2-hybrid approach, found that -actinin-4 constitutively interacts with ICAM-1 cytoplasmic tail. They identified ICAM-1 resides Arg480, Lys481, and Arg486 as critical for association with -actinin. The authors did not examine whether these mutations alter ICAM-1 clustering during leukocyte transmigration or affect cytoskeletal remodeling but did show that knockdown of -actinin-4 by siRNA reduced neutrophil transmigration of TNF-–activated HUVECs. Future studies are necessary to identify other cytoskeletal molecules or networks that participate in adhesion molecule-mediated signal transduction in endothelium and to clarify the role of ezrin/moesin and Rho GTPases in cytoskeletal remodeling and leukocyte transmigration.

	Conclusions

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 
The study and elucidation of EC-dependent mechanisms that regulate leukocyte recruitment is crucial to enlarge the therapeutic window for cardiovascular diseases. Importantly, work reviewed in this article has added new molecules to the list of adhesion molecules mediating mononuclear leukocyte recruitment and transmigration and made several substantial changes in our understanding of mechanisms that regulate leukocyte recruitment. Many of these advances still point to the vascular endothelium as intimately associated with the initiation, retention, and presumably regression of inflammatory cell recruitment to essentially all tissues and organs.

	Acknowledgments

 
We thank Dr Michael Gimbrone (Brigham and Women’s Hospital) and Prof Dorian Haskard (Imperial College London) for helpful discussions and continued support, as well as members of the Center for Excellent in Vascular Biology, Brigham and Women’s Hospital, for helpful discussions and advice for this article. We apologize to the authors of articles that we have failed to mention because of space limitations.

Sources of Funding

Funding was provided by the NIH grants HL36028, HL53993, and HL 56985 to F.W.L.) and RO1-HL076686 and PO1-HL36028 (to G.G.-C.) and Arthritis Research Campaign UK grants R0600 and R0625 (to R.M.R.).

Disclosures

None.

	Footnotes

 
Original received March 12, 2007; revision received May 3, 2007; accepted May 31, 2007.

	References

Top Abstract Introduction Adhesion Pathways That Regulate... Endothelial Cell Adhesion... Mechanisms Regulating Leukocyte... Novel Endothelial... Signaling via Endothelial Cell... Conclusions References

 

1. Gimbrone MA Jr. Vascular endothelium: nature’s blood-compatible container. Ann N Y Acad Sci. 1987; 516: 5–11.[Medline] [Order article via Infotrieve]
2. Gimbrone MA. Endothelial dysfunction, hemodynamic forces, and atherosclerosis. Thromb Haemost. 1999; 82: 722–726.[Medline] [Order article via Infotrieve]
3. Butcher EC. Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity. Cell. 1991; 67: 1033–1036.[CrossRef][Medline] [Order article via Infotrieve]
4. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med. 2005; 352: 1685–1695.[Free Full Text]
5. Charo IF, Ransohoff RM. The many roles of chemokines and chemokine receptors in inflammation. N Engl J Med. 2006; 354: 610–621.[Free Full Text]
6. Glass CK, Witztum JL. Atherosclerosis. The road ahead. Cell. 2001; 104: 503–516.[CrossRef][Medline] [Order article via Infotrieve]
7. Shaw SK, Bamba PS, Perkins BN, Luscinskas FW. Real-time imaging of vascular endothelial-cadherin during leukocyte transmigration across endothelium. J Immunol. 2001; 167: 2323–2330.[Abstract/Free Full Text]
8. Gotsch U, Borges E, Bosse R, Boggemeyer E, Simon M, Mossmann H, Vestweber D. VE-cadherin antibody accelerates neutrophil recruitment in vivo. J Cell Sci. 1997; 110: 583–588.[Abstract]
9. Lo SK, Lee S, Ramos RA, Lobb R, Rosa M, Chi-Rosso G, Wright SD. Endothelial-leukocyte adhesion molecule 1 stimulates the adhesive activity of leukocyte integrin CD3 (CD11b/CD18,Mac-1, _mß₂) on human neutrophils. J Exp Med. 1991; 173: 1493–1500.[Abstract/Free Full Text]
10. Simon SI, Hu Y, Vestweber D, Smith CW. Neutrophil tethering on E-selectin activates beta 2 integrin binding to ICAM-1 through a mitogen-activated protein kinase signal transduction pathway. J Immunol. 2000; 164: 4348–4358.[Abstract/Free Full Text]
11. Woollard KJ, Kling D, Kulkarni S, Dart AM, Jackson S, Chin-Dusting J. Raised plasma soluble P-selectin in peripheral arterial occlusive disease enhances leukocyte adhesion. Circ Res. 2006; 98: 149–156.[Abstract/Free Full Text]
12. Tanaka Y, Adams DH, Shaw S. Proteoglycans on endothelial cells present adhesion-inducing cytokines to leukocytes. Immunol Today. 1993; 14: 111–115.[CrossRef][Medline] [Order article via Infotrieve]
13. Rot A, Hub E, Middleton J, Pons F, Rabeck C, Thierer K, Wintle J, Wolff B, Zsak M, Dukor P. Some aspects of IL-8 pathophysiology. III: Chemokine interaction with endothelial cells. J Leukoc Biol. 1996; 59: 39–44.[Abstract]
14. Shamri R, Grabovsky V, Gauguet JM, Feigelson S, Manevich E, Kolanus W, Robinson MK, Staunton DE, von Andrian UH, Alon R. Lymphocyte arrest requires instantaneous induction of an extended LFA-1 conformation mediated by endothelium-bound chemokines. Nat Immunol. 2005; 6: 497–506.[CrossRef][Medline]