View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1828 (2013) 157–166

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamem

Review Connexins in ☆

Anna Pfenniger a,b, Marc Chanson c, Brenda R. Kwak a,b,⁎

a Department of Pathology and Immunology, University of Geneva, Switzerland b Cardiology, Department of Internal Medicine, University of Geneva, Switzerland c Department of Pediatrics, Geneva University Hospitals and University of Geneva, Switzerland

article info abstract

Article history: Atherosclerosis, a chronic inflammatory disease of the vessel wall, involves multiple cell types of different or- Received 8 February 2012 igins, and complex interactions and signaling pathways between them. Autocrine and paracrine communica- Received in revised form 26 April 2012 tion pathways provided by cytokines, chemokines, growth factors and lipid mediators are central to Accepted 4 May 2012 atherogenesis. However, it is becoming increasingly recognized that a more direct communication through Available online 15 May 2012 both hemichannels and gap junction channels formed by connexins also plays an important role in athero- sclerosis development. Three main connexins are expressed in cells involved in atherosclerosis: Cx37, Cx40 Keywords: Gap junction and Cx43. Cx37 is found in endothelial cells, monocytes/ and , Cx40 is predominantly Connexin an endothelial connexin, and Cx43 is found in a large variety of cells such as smooth muscle cells, resident Hemichannel and circulating leukocytes (neutrophils, dendritic cells, lymphocytes, activated macrophages, mast cells) Atherosclerosis and some endothelial cells. Here, we will systematically review the expression and function of connexins in cells and processes underlying atherosclerosis. This article is part of a Special Issue entitled: The Commu- nicating junctions, roles and dysfunctions. © 2012 Elsevier B.V. All rights reserved.

Contents

1. Introduction ...... 157 2. Connexins in the pathogenesis of atherosclerosis ...... 159 2.1. Endothelial cell dysfunction ...... 159 2.2. Inflammatory cell recruitment ...... 160 2.2.1. Monocytes/macrophages ...... 160 2.2.2. Lymphocytes ...... 160 2.2.3. Dendritic cells ...... 161 2.2.4. Neutrophils ...... 161 2.2.5. Platelets ...... 162 2.2.6. Mast cells ...... 162 2.3. Smooth muscle cell recruitment ...... 162 2.4. Other repair-related mechanisms ...... 163 2.4.1. Angiogenesis ...... 163 2.4.2. Lymphangiogenesis ...... 164 3. Conclusions and perspectives ...... 164 Acknowledgements ...... 164 References ...... 164

1. Introduction ☆ This article is part of a Special Issue entitled: The Communicating junctions, roles and dysfunctions. Atherosclerosis is a multifactorial inflammatory disease of the vessel ⁎ Corresponding author at: Dept. of Pathology and Immunology, University of Geneva, wall of medium to large-sized . It is the leading cause of mortal- FRM, 64 de la Roseraie, CH-1211 Geneva, Switzerland. Tel.: +41 22 3827237; fax: +41 22 3827245. ity in industrialized countries, and its incidence is rapidly increasing in E-mail address: [email protected] (B.R. Kwak). the developing nations [1]. Atherosclerosis is characterized by specific

0005-2736/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2012.05.011 158 A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 lesions of the intimal layer of arteries, called , where the tra- after the 4th decade of life. Atherosclerosis can thus be viewed as a ditional components of a chronic inflammatory response such as the ex- model of a slowly progressing inflammatory disorder. travasation and accumulation of leukocytes, the presence of necrotic Atherosclerosis starts with a dysfunction of the endothelial cells debris, the deposition of extracellular matrix, and the apparition of an- (ECs) lining an , which start to express increased levels of adhe- giogenesis, can all be found [2] (Fig. 1). This disease progresses over sion molecules [2,3]. At the same time, lipid particles, such as low decades. In fact, the earliest atheromatous plaques can be found in density (LDL)-, accumulate within the vessel teenagers, and most of the consequences are generally apparent wall and become oxidized. The traditional cardiovascular risk factors

Monocyte adhesion aggregation platelet ATP

mφ

cAMP cGMP

Immunological synapse Angiogenesis

M

G2 G1

DC T cell S activation

? EC

Fig. 1. Key aspects of the function of connexins in atherosclerosis development. Top left: Cx37 inhibits monocyte (mΦ) adhesion to the by the release of ATP through hemichannels. ATP then acts in an autocrine fashion by binding to its receptors, which ultimately leads to a reduced adhesiveness of the monocyte. Bottom left: Cx43 is involved in the immunological synapse. Cx43 is expressed by DCs and T lymphocytes and gap junctions are present at the outer aspect of the immunological synapse. Activation of T lympho- cytes is favored by Cx43-dependent communication. Top right: Cx37 reduces platelet aggregation. Platelets form Cx37 gap junction channels. GJIC between platelets reduces aggre- gation, presumably through the diffusion of anti-aggregating metabolites (cAMP, cGMP) from the bulk of the thrombus to newly recruited platelets. Bottom right: Cx37 reduces angiogenesis. Cx37-deficient animals have increased angiogenesis and EC proliferation. The presence of Cx37 gap junction channels may reduce proliferation by blocking the tran- sition from G1 to S phase of the cell cycle. A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 159 such as , dyslipidemia, diabetes and smoking all pro- pathway for transmembrane signaling, whereas gap junction chan- mote these early steps of atherogenesis. The chemotactic activity of nels will enable a direct communication pathway between the cyto- oxidized lipids (oxLDL) combined with the enhanced expression of plasm of adjacent cells. Both communication pathways are believed adhesion molecules by ECs leads to the adhesion and diapedesis of to be essential for proper cell and tissue homeostasis, and have leukocytes such as monocytes, lymphocytes, neutrophils and baso- been shown to play a major role in a wide variety of biological and phils [4]. OxLDLs become phagocytosed by monocytes differentiated pathological processes. into macrophages, and the inflammation is then perpetuated by the Atherosclerosis is a chronic inflammatory disease with multiple recruited leukocytes. Lipids and inflammatory cells slowly accumu- cellular protagonists, and thus with complex interactions between late and eventually form a necrotic core. Simultaneously, chemotactic several cell types. Most suggested and investigated signaling path- factors secreted within the neointima also drive the de-differentiation, ways involve autocrine and paracrine communication through cyto- migration and proliferation of smooth muscle cells (SMCs) from the un- kines, chemokines, growth factors and their respective receptors. derlying media into the neointima. The cells ultimately form a fibrous However, it is now increasingly recognized that a more direct path- cap sealing off the necrotic core from the lumen. Further progression in- way, such as gap junction intercellular communication (GJIC), also volves the growth of newly formed vessels within the plaque, which plays a role in inflammatory diseases in general, and atherosclerosis can promote the arrival of new inflammatory cells. The growth of in particular. Moreover, signaling through connexin hemichannels is such a plaque will impinge on the vessel lumen (i.e. cause a ) also increasingly described in inflammatory disorders. In this review, which can lead to symptoms if the restriction of blood flow is too severe, we will describe the expression and function of connexins in each cell and especially in case of increased demand of blood supply (stable an- type involved in atherosclerosis and, if possible, their implication in gina pectoris, for instance, when stenosed coronary arteries impair the development of atherosclerosis. proper perfusion of the myocardium during exercise). However when the stability of such a plaque is compromised, i.e. when the balance be- 2. Connexins in the pathogenesis of atherosclerosis tween the fibrous elements and inflammatory components is lost, the plaque can rupture. This exposes a highly thrombogenic material to 2.1. Endothelial cell dysfunction the blood flow, which then leads to the activation and aggregation of platelets and to the activation of the coagulation cascade. If the throm- Endothelial cell dysfunction is regarded as the first event in ath- bus formation is not limited, it can lead to the rapid occlusion of the erosclerosis development. Following exposure to disturbed blood vessel, which in turn causes an ischemia of perfused organs (for in- flow patterns and pro-inflammatory or toxic circulating factors such stance, a myocardial infarction (MI), when a coronary artery is as encountered in the metabolic syndrome, dyslipidemia, in smokers abruptly occluded). or in systemic pro-inflammatory states, ECs start to express adhesion It is well recognized that atherosclerosis, like most cardiovascular molecules. Simultaneously, the expression and function of eNOS be- diseases, arises from interactions between environmental factors and comes altered [18]. Its product, NO, has several vasoprotective effects: multiple genes. The heritable component of cardiovascular diseases is it reduces inflammation, leukocyte adhesion, platelet aggregation, estimated between 40% and 60%, depending on the populations [5].In and SMC proliferation and induces endothelium-dependent vasodila- fact, the role of genetic polymorphisms in this disease has been tion [19]. The hallmark of dysfunctional ECs is a markedly reduced syn- highlighted by numerous large genome wide studies [6]. Every path- thesis of NO. Moreover, deleterious reactive oxygen species (ROS) can ophysiological aspect of atherosclerosis could be linked to several ge- be produced by uncoupling of eNOS itself, by the nicotinamide adenine netic polymorphisms. For instance, polymorphisms in genes involved dinucleotide phosphate (NAPDH) oxidase and by xanthine oxidase, or in lipid metabolism, such as the apolipoproteins (apoB, apoE), have by neighboring inflammatory cells. These ROS further reduce nitric been frequently associated with coronary artery disease (CAD) and/ oxide (NO) bioavailability by converting it to peroxynitrite and thus or MI [7,8]. Endothelial dysfunction is assumed to be predisposed by contribute to the phenomenon of endothelial dysfunction. Additionally, polymorphisms in the endothelial nitric oxide synthase (eNOS), ROS can oxidize eNOS itself, further reducing NO production. which were correlated with carotid intima/media thickness and In arteries of a healthy individual, ECs mainly express Cx37 and CAD [9,10]. Similarly, polymorphisms in genes involved in oxidative Cx40. However, both connexins are lost in the endothelium overlay- stress, inflammation, hemostasis, and vascular remodeling could all ing atherosclerotic plaques [20]. In addition, Cx43 can be detected be correlated with atherosclerosis-related diseases [6]. Interestingly, in ECs at specific regions of arteries [21], such as at branching sites, the 1019C>T polymorphism of the human Cx37 gene, which results which are highly susceptible to atherosclerosis development, and in a single amino acid substitution in the regulatory C-terminal re- also at the shoulder regions of plaques [20]. The mechanisms involved gion of this protein (P319S) has repetitively been associated with in the modification of connexin expression pattern in atherosclerosis atherosclerosis-related diseases such as and have not been elucidated yet. However, it was suggested that inflam- its consequence, stroke, as well as CAD and MI [11–15]. These obser- matory mediators may be involved. In fact, cultured ECs were shown vational studies pointed toward a possible implication of connexins to lose Cx37 and Cx40 expression after treatment with tumor necrosis in atherosclerosis development. factor (TNF)-α, whereas Cx43 expression was only modestly modi- Connexins are a homogeneous family of proteins expressed in a fied [22]. It is thus likely that the abundance of pro-inflammatory cy- large variety of tissues and known for their assembly into intercellular tokines in an may cause a reduction in endothelial Cx37 channels, called gap junctions [16]. Twenty-one different human con- and Cx40 expression. In addition, some cytokines, like interleukin nexin genes have been reported so far, each coding for a protein with (IL)-1β, were shown to regulate the function of gap junction channels 4 transmembrane domains, 2 extracellular and one cytoplasmic and hemichannels in other inflammatory systems [23,24]. It is thus loops, a short cytoplasmic N-terminal domain and a longer regulato- conceivable that connexin function may also be affected by cytokines ry C-terminal domain [17]. Connexin proteins are named after their in atherosclerosis. specific molecular weight in kDa (for instance, Cx43 has a mobility A functional coupling of arterial ECs seems to be important for a of 43 kDa, see http://www.genenames.org/genefamily/gj.php). Six synchronized response of ECs to various agonists. For instance, it connexins can oligomerize and form a hemichannel or connexon. was shown that individual responses to acetylcholine or histamine When two cells are closely apposed, connexons from one cell can are variable in ECs of rat aortic segments [25]. As the histamine recep- dock with their counterparts in the neighboring cell and form a gap tor H1 is only focally expressed in the endothelium, GJIC is necessary junction channel. Depending on the cell type and the connexin for a homogeneous increase in cytosolic calcium and subsequent acti- expressed, connexons can function as hemichannels, providing a vation of eNOS following exposure to histamine [26]. An additional 160 A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 role of connexins as regulators of eNOS expression and function was treated with TNF-α and IFN-γ [33]. In atherosclerotic plaques, macro- recently demonstrated. Cx37 is a direct protein partner for this pro- phages and foam cells were strongly immunoreactive for Cx37 and in tein, and a mutual functional regulation between eNOS and Cx37 some instances, Cx43 could also be found [20]. was demonstrated in ECs [27]. Downregulation of Cx37 caused an in- The function of Cx43 expression in monocytes/macrophages in ath- crease in NO release, without affecting eNOS expression. Conversely, erosclerosis has not been investigated so far. However, Cx43 seems to short eNOS-mimetic peptides modified the regulation properties of be important for phagocytosis by regulating Ras homolog Cx37 gap junction channels. Moreover, Cx40-deficient mice have a re- gene family, member A (RhoA)-dependent rearrangement of the actin duced expression of eNOS in ECs and a decreased endothelium- cytoskeleton [34]. This role was demonstrated for Fc receptor (FcR)- dependent vasorelaxation [28]. In addition to modified eNOS function, driven phagocytosis, but one could hypothesize a similar role in phago- dysfunctional ECs express increased amounts of selectins (E-selectin, cytosis mediated by other receptors, such as the scavenger receptors. P-selectin) and integrins (intercellular adhesion molecule (ICAM)-1, This could add a novel potential role for Cx43 in atherosclerosis devel- vascular cell adhesion molecule (VCAM)-1), which promote leukocyte opment, by favoring uptake of oxLDL by macrophages. adhesion to the endothelium and subsequent transmigration into the The role for Cx37 in monocytes/macrophages in atherosclerosis is intimal space. This increase in adhesion molecules is tightly regulated more evident. As shown in Fig. 1 (top left), hemichannels formed by in physiological conditions and for instance, VCAM-1 expression is this connexin confer an anti-adhesive property to monocytes via the re- downregulated by the ecto-nucleotidase CD73 on ECs. Here again, inter- lease of ATP in the extracellular space, which can then have an auto- cellular communication seems to be involved in the regulation mecha- crine/paracrine effect on monocytes [32]. In fact, monocytes deficient nisms. In fact, endothelial-specific deletion of Cx40 in atherosusceptible for Cx37 were shown to adhere more strongly to non-biological surfaces mice (on an ApoE−/− background) was associated with a reduction in as well as to a monolayer of ECs. Using adoptive transfers of murine expression and activity of CD73, as well as a massive increase in blood cells, it could be concluded that Cx37-deficient monocytes were VCAM-1 expression [29]. These mice subsequently developed earlier more likely to enter atherosclerotic plaques than Cx37-expressing and larger atherosclerotic plaques than their control littermates. cells. These results provided a likely mechanistic explanation to the A role for endothelial Cx43 in atherosclerosis development is less atherosusceptible state of Cx37-deficient mice. Interestingly, the afore- clear. Its expression is generally very low or absent in the arterial en- mentioned genetic polymorphism in the human Cx37 gene, which re- dothelium of healthy individuals. Cx43 expression was shown to be sults in either a proline (major allele) or a serine (minor allele) at increased in the aorta of platelet endothelial cell adhesion molecule position 319 of the Cx37 protein, appeared to affect cell adhesion. In (PECAM)-1 deficient mice, which, when bred in an ApoE−/− back- fact, cells transfected with the 319S isoform of Cx37 were more adhesive ground, developed less atherosclerosis than PECAM-1-expressing than when the 319P isoform was expressed. ApoE−/− mice [30]. However, Cx43 may also be increased in regions In conclusion, Cx37 expression in monocytes and macrophages subjected to disturbed flow, where increased atherosclerosis devel- has an anti-atherosclerotic effect by reducing cell adhesion, and the opment is present. Interestingly, inhibition of Cx43-mediated GJIC be- role of Cx43 in these cells in the context of atherosclerosis is not yet tween ECs by mimetic peptides decreased adhesion of inflammatory defined. cells [31]. These discrepancies underline the need for more detailed stud- ies using cell-specific deletion of Cx43 in the context of atherosclerosis. 2.2.2. Lymphocytes It therefore appears that endothelial Cx37 and Cx40 are important Although innate immunity incontestably plays a major role in in the maintenance of endothelial quiescence, by favoring a proper atherosclerosis, it became evident over the last decade that adaptive expression and regulation of eNOS, and by inhibiting leukocyte adhe- immunity also contributes to the pathogenesis of plaque develop- sion, whereas the function of endothelial Cx43 is less clear. Thus an ment. T cells are recruited to the intima by similar processes than alteration of Cx37 or Cx40 expression could potentially lead to EC monocytes [4]. They are much less numerous than macrophages dysfunction, and promote atherosclerosis development. within lesions (about 4 to 10 times less in human plaques). Once in the lesions, they are activated, secrete proatherogenic mediators 2.2. Inflammatory cell recruitment and participate in plaque growth. B cells have also been described in plaques, but they are more abundant in the adventitial layer of ar- 2.2.1. Monocytes/macrophages teries. Finally, tertiary lymphoid structures are commonly associated The combination of adhesion molecules on ECs and release of che- with advanced atherosclerotic plaques. Through the study of mice motactic factors from the vessel wall facilitates monocyte adhesion deficient for B or T cells, it has become evident that antigen-specific and transmigration from the artery lumen to the intima [2]. Mono- adaptive immune responses contribute to atherosclerosis develop- cytes then differentiate into macrophages and phagocyte oxLDL due ment [35]. B cells may play a protective role, for instance by produc- to its binding to their scavenger receptors. Lipids progressively accu- ing antibodies directed against oxLDL. However, it was suggested that mulate within macrophages, which become foam cells, the hallmark specific subsets of B cells could also favor atherosclerosis development. of atheroma [4]. OxLDL increases the activation of macrophages and Thus, the role of B cells does not appear straightforward. In contrast, T their synthesis of cytokines (such as TNF-α) which further promotes cells are central protagonists in atherogenesis. The majority are positive the recruitment of leukocytes. In addition, macrophages release toxic for the CD4 antigen, though some CD8 + cells can also be found. The

ROS that maintain the oxidation of LDL particles. Finally, macro- main pathogenic effects of T cells are mediated by a T helper (TH)1 re- phages secrete several proteases that will weaken the newly synthe- sponse, with high secretion of IFN-γ. However, a role for the TH17 subset sized fibrous cap, and thus promote plaque rupture. Macrophages can has also been proposed. Regulatory T (Treg) cells may also be involved, as also act as antigen presenting cells (APCs) and contribute to the im- a transfer of forkhead box P3 (Foxp3)+ T cells confers protective effects mune component of atherogenesis, but this will be discussed in against atherosclerosis development in mice. more detail in the section on dendritic cells (DCs), as they are proto- As with all leukocytes in general, the study of connexins in lym- typical APCs. phocytes is complex, due to the large panel of different lymphocyte Circulating monocytes express low levels of Cx37 [32],butother subsets, the rapid evolution of knowledge in that respect, and the dif- connexins could not be detected. However upon stimulation with a ficulty to properly characterize lymphocytes. Thus, all results should combination of cytokines (TNF-α and interferon (IFN)-γ)human be interpreted in view of the origin and specific characteristics of blood monocytes were shown to express Cx43. It could also be demon- the cells used in each study. strated that whereas clusters of native monocytes were not functionally T lymphocytes express Cx43. Cx40 has also been reported, but in- coupled, Lucifer Yellow could diffuse between cells once they were consistently, and only in T cells derived from human tonsils [36].To A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 161 determine whether subsets of T cells differentially express Cx43, myeloid DCs (mDCs) and plasmacytoid DCs (pDCs). Though much mouse spleen-derived TH0 cells were differentiated into TH1orTH2. less abundant than macrophages, both have been detected in athero- It appeared that the spleen-derived TH1 cells expressed more Cx43 sclerotic plaques, but more consistently in the associated adventitia than TH0orTH2 [37]. and lymphoid tissues. In healthy arteries, mDCs are thought to have Regarding T cells, two distinct aspects need to be considered: ini- a tolerogenic role by silencing T cells [4]. In the context of atheroscle- tial development and further differentiation/activation. A function for rosis, mDCs can accumulate lipids like macrophages, or become acti- Cx43 in T cell development has been investigated using a mouse vated by danger signals (oxLDL and its degradation products) and model with T cell-specific deletion of Cx43, as well as with chimeric thus become activators of T cells [43]. This will lead to increased pro- mice lacking Cx43 in all bone marrow-derived cells [38,39]. In both duction of proatherogenic cytokines such as IFN-γ and TNF-α. pDCs, models, T cell development appeared physiological despite Cx43 de- being one of the main sources of class I IFN, were first hypothesized letion. Thus, Cx43 expression in T cells does not appear to play a to be proatherogenic as well. However, a recent study demonstrated prominent role in their development. that pDCs protect against atherosclerosis by dampening the T cell re- Regarding T cell activation, a prominent role for Cx43 was recently sponse [44]. The secretion of the tolerogenic molecule indoleamine described in the T cell–DC antigen-dependent interaction, known as 2,3-dioxygenase by pDCs was able to suppress T cell proliferation. immunological synapse [40] (Fig. 1 bottom left). Cx43 was shown to Knowledge on connexins in DCs is currently restricted to mDCs, accumulate at the outer aspect of the immunological synapse both which also express Cx43. As above-mentioned, a role for Cx43 in in CD4 + cells and DCs, but only when the T cell receptor (TCR) was the immunological synapse between T cells and (m)DCs was recently engaged with a specific major histocompatibility complex class II demonstrated. In addition, GJIC between DCs, or between DCs and (MHC-II): antigen couple. Moreover, this concentration of Cx43 led to target cells, has also been demonstrated. First, fluorescent markers a bidirectional transfer of fluorescent probes, indicating functional gap can diffuse within clusters of DCs activated by lipopolysaccharide junction channels. Absence or blockade of Cx43 by a mimetic peptide (LPS) or TNF-α in addition to IFN-γ. Incubation with gap junction or 18-β-glycyrrhetinic acid reduced the intracellular calcium oscilla- blockers seems to inhibit the induction of the costimulatory mole- tions in T cells following antigen recognition, and reduced IFN-γ secre- cules CD80 and CD86 [45]. Thus, communication between DCs could tion. Thus, the presence of Cx43 appears to be important for T cell participate in their proper activation and subsequent antigen presen- activation. As a TH1 antigen-specific response is involved in atherogen- tation. Second, a role for connexins in cross-presentation of antigens esis, it is likely that Cx43 may participate in the development of this dis- has been extensively demonstrated [46–48]. In fact, it was shown ease by contributing to the activation of lymphocytes. that peptidic antigens can diffuse through gap junctions in DCs. The

Tregs also express Cx43, and several roles for this protein have been receiver DC is able to present this newly acquired peptide on MHC-I proposed. First, T cell-specificdeletionofCx43inmice(CD4-Cre+ molecules, and thus activate cytotoxic CD8 + T cells. CD8 + cells fl/fl Cx43 ) markedly reduced the amount of Treg cells, without affecting have not been extensively investigated in atherosclerosis develop- other T cell subsets, as mentioned above [38]. This indicates that Cx43 ment. However, it was suggested that they may be involved in disease is essential for the development of Treg cells. The overexpression of progression and plaque destabilization [49]. Cx43 in a lymphoblast cell line caused the overexpression of the cru- Here again, the function of connexins in DCs in the context of athero- cial Treg transcription factor Foxp3, and this effect was lost when the sclerosis remains hypothetical. Increase of DC activation by gap junc- C-terminal domain of Cx43 was truncated. Thus, it seems that Cx43 is tions, and participation in the immunological synapse and in antigen necessary for the differentiation of Treg cells, which may have conse- cross-presentation, could promote atherosclerosis development. quences in the context of atherosclerosis. Second, the formation of gap junctions with target T cells may participate in the suppressor func- 2.2.4. Neutrophils tion of Tregs, and as cyclic adenosine monophosphate (cAMP) is known Neutrophils are important actors of innate immunity and are rap- to mediate that suppressor effect in target cells, it was hypothesized idly recruited from the microcirculation to sites of inflammation. Less that its transfer through gap junctions could be involved [41]. is known about their contribution to chronic inflammation. Recently, B cells at different stages of differentiation were also shown to ex- however, a role for neutrophils in macrovascular inflammation and in press Cx43 [36]. Here again, chimeric mice lacking Cx43 in bone atherosclerosis has been reported [50–54]. Although neutrophils are marrow-derived cells exhibited normal B cell development [39].Using not easily detectable in human and murine atherosclerotic lesions, a various B cell lines, Machtaler and colleagues [42] suggest that Cx43 is correlation between blood neutrophil counts and size of atheroscle- involved in B cell activation and adhesion following the involvement rotic plaques has been observed [50,52]. Hypercholesterolemia was of the BCR immunological synapse. Splenic B cells from Cx43+/− mice shown to stimulate granulopoiesis and increase neutrophil mobiliza- also appeared to have a slightly reduced adhesion on anti-IgM-coated tion from the bone marrow. Simultaneously, high cholesterol in- surfaces than cells isolated from wild-type mice. Unlike as described creases the expression of adhesion molecules by ECs, which when in T lymphocytes, the effect of Cx43 on B cell activation seemed to de- combined with chemotactic cytokines can then recruit circulating pend on the presence of the C-terminal domain of Cx43, and not on neutrophils to atherosclerotic lesions [55]. Neutrophil secretory gran- intercellular transfer of molecules. As B cells, and mostly their produc- ules are reservoirs of antimicrobial proteins, proteases, components tion of antibodies derived against plaque antigens, are thought to be of the respiratory burst, as well as of a wide variety of ligands for en- atheroprotective, Cx43-dependent activation of these cells may be ben- dothelial membrane receptors, extracellular matrix proteins and sol- eficial against plaque growth. uble mediators of inflammation [56,57]. Thus, in addition to their In conclusion, the overall effect of Cx43 on lymphocyte participa- participation in the initiation of atherosclerosis, the release of pro- tion in atherosclerosis development is far from being straightforward. atherosclerotic factors by activated neutrophils may compromise Overall, it participates in the activation and function of most lympho- the stability of fibrous caps [58]. cyte subtypes, and could thus both have pro- and anti-atherogenic There is circumstantial evidence that neutrophils do express con- functions, depending on the degree of activation of each cell type. nexins and that gap junctions may be established between inflammato- However, specific studies on lymphocyte connexins in the context ry cells during the course of inflammation [59]. Hence, it was proposed of atherosclerosis remain to be performed. that unstimulated human neutrophils express the major vascular con- nexins, namely Cx37, Cx40, and Cx43 [60]. In contrast, other studies 2.2.3. Dendritic cells failed to detect mRNAs for Cx26, Cx32, Cx37, Cx40, Cx43 and Cx45, DCs are professional APCs that can present exogenous or endoge- suggesting that human circulating neutrophils do not express con- nous antigens to T cells. There are 2 subtypes of DCs: conventional or nexins [31,61,62]. However, positive immunodetection of Cx43 was 162 A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 observed after LPS activation of peripheral hamster blood leukocytes the extent of platelet aggregation was reproduced when wild-type [61], suggesting that Toll-like receptor (TLR) activation of neutrophils mouse or human platelets were treated with pharmacological gap may induce the transcription of Cx43. Moreover, in humans, LPS in- junction blockers or with a Cx37-mimetic peptide. Thus, it appears duced the translocation of Cx43 to the membrane of neutrophils and that functional Cx37 gap junction channels moderate platelet promoted their aggregation. Aggregated neutrophils exhibited GJIC activation. only when treated with conditioned medium from activated ECs, Although the signals that move through Cx37 gap junctions have not suggesting that soluble factors secreted by these cells may be involved been discovered, the authors further showed that the 1019C>T poly- in the induction of gap junctions in neutrophils during tissue injury morphism of the human Cx37 gene, coding for Cx37-319P or -319S, af- [63]. In contrast to these studies, neutrophils recovered from inflamed fected the gap junctional permeability to neurobiotin. In fact, enhanced lungs after bronchoalveolar lavages did not show Cx43 expression or intercellular spread of the tracer was observed in cell lines expressing cell coupling [62]. Controversies between laboratories regarding GJIC the Cx37-319S isoform as compared to cells expressing Cx37-319P in neutrophils therefore leave many hypotheses unresolved. channels. Interestingly, platelets from individuals with the latter poly- Neutrophil extravasation seems to involve GJIC, since transendothelial morphism (decreased Cx37 channel permeability) showed enhanced migration of these cells was increased in the presence of connexin- reactivity in in vitro aggregation assays, consistent with the idea that ef- mimetic peptides or gap junction channel blockers [60,64]. Recently, it ficient platelet-to-platelet coupling limits thrombus growth. As men- has been proposed that Cx43 in neutrophils may exert anti- tioned above, the Cx37 1019C>T polymorphism strongly associates adhesion properties by modulating the endothelium barrier func- with carotid and coronary artery stenosis [11–13,15], and is an indepen- tions [65]. The mechanism underlying such a role appeared to in- dent risk factor for myocardial infarction [12,14]. Together with the pre- volve hemichannels rather than GJIC. Indeed, this anti-adhesive vious finding that Cx37 in monocytes regulates the leukocyte adhesion effect of Cx43 was explained by the extracellular release of ATP to the endothelium [32], the data suggest that impaired Cx37 channel from these cells via hemichannels, which is similar to the anti- function may accelerate the development of atherosclerosis. adhesive effect of Cx37 in monocytes described above [65]. However in this case, ATP does not act directly on the inflammatory cell. Rath- 2.2.6. Mast cells er, ATP undergoes hydrolysis to adenosine by the action of ecto- Mast cells are primarily known as essential effectors in the resolu- apyrase (CD39) and CD73 at the surface of ECs. Adenosine then acti- tion of allergic responses. Recent investigations however have now vates its receptors on the EC surface that trigger a cAMP-dependent also implicated mast cells in the initiation and progression of athero- intracellular signaling, resulting in the decreased expression of ad- sclerosis. Mast cells derive from bone marrow cells and circulate in hesion molecules [65]. Although the latter study did not take into ac- the peripheral blood as mast cell precursors before being recruited count changes in endothelial GJIC in response to the ATP/adenosine to specific tissues and organs where they further mature. Mast cells ratio, a dysfunctional endothelium may interfere with these protec- also reside in the vessel wall, particularly in perivascular tissue, and tive mechanisms, which may promote neutrophil recruitment to accumulate in atherosclerotic lesions [70,71]. Upon activation by atherosclerotic lesions. IgE, complement components or viral or bacterial pathogens, mast cells acutely release a fraction of their cytoplasmic granules that con- 2.2.5. Platelets tain a large panel of mediators including neutral proteases, cathep- Platelets are anucleated fragments of bone marrow megakaryo- sins, histamine, heparin and several cytokines and growth factors. cytes that play a major role in inflammation, homeostasis and throm- The activities of these various mediators have been found to contribute bus formation after endothelial injury. Upon activation, platelets to the initial fatty streak formation as well as to the destabilization of undergo morphological changes and express molecules that enhance plaques, thereby making it prone to rupture [72–74]. In vitro studies their adhesion to the endothelium as well as to other inflammatory have revealed that mast cells regulate the behavior of atheroma- cells, including neutrophils [66]. The platelet–neutrophil interaction associated cell types through their secreted mediators, although oppo- plays a pivotal role in acute inflammation, and aggregates of both site results have been obtained depending on the mediator involved. are increased in patients with CAD [66,67]. Although the causal role In fact, histamine and basic fibroblast growth factor (bFGF) released of platelet–neutrophil interaction in atherosclerosis remains to be from mast cells activate SMC surface receptors and accelerate their mi- elucidated, it may help to explain some of the pathophysiological gration and proliferation [75,76], whereas mast cell-derived heparin events associated with different clinical states [68]. Based on the ra- proteoglycans inhibit proliferation of aortic SMCs in culture [77].In tionale of reducing thrombus formation after plaque rupture, anti- this respect, the recent work of Ehrlich and colleagues describing a aggregating drugs are widely prescribed, and have shown clear bene- role for Cx43-mediated GJIC between mast cells and fibroblasts in fits over the last decades. But in addition, therapeutic inhibition of wound repair and fibrosis is of particular interest. Using a sophisticated neutrophil–platelet and platelet–platelet aggregation may prove use- co-culture model, they showed that heterocellular GJIC between mast ful in the stage of atherosclerosis development as well, by moderating cells and fibroblasts increases fibroblast proliferation and fibroblast- inflammation in the vessel wall. induced collagen lattice contraction, two important hypertrophic scar Recently, connexins and GJIC have been shown to contribute to fibroblast activities [78]. Moreover, heterocellular GJIC between a platelet aggregation by providing an intercellular pathway that nega- mast cell line (RMC-1) and human dermal fibroblasts transformed tively regulate thrombus propensity [69] (Fig. 1 top right). In this these cells into myofibroblasts, expressing α-SM actin within cytoplas- study, Angellilo-Scherrer and colleagues report the presence of Cx37 mic stress fibers [79]. Whether GJIC between mast cells and other cell in mouse and human platelets and the intercellular transfer of the types within the atherosclerotic lesion also directs the activities of Cx37-permeant gap junction tracer neurobiotin. Neurobiotin transfer these cells remains to be established. was reduced by gap junction blockers and was abrogated in platelets isolated from Cx37-deficient mice. Interestingly, bleeding time after 2.3. Smooth muscle cell recruitment tail transsection was reduced by half in Cx37-deficient mice as com- pared to wild-type mice. Moreover, thrombus propensity was in- Lesion progression involves the migration of SMCs from the media creased in Cx37-deficient mice both after application of FeCl3 to the into the intima where they proliferate in response to mediators such mesenteric arteries and after intravenous injection of collagen and as platelet-derived growth factor (PDGF). In the intima, SMCs produce epinephrine in anesthetized mice. In keeping with these in vivo ob- extracellular matrix macromolecules, including collagen and elastin, servations, in vitro aggregation assays showed increased reactivity and form a fibrous cap that covers the plaque [2,80]. Progenitor cells to low-dose agonists of mouse platelets lacking Cx37. Increase in from the blood might also settle as SMCs in the atherosclerotic lesion A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 163

[81]. SMCs may die in advancing plaques, often by apoptosis, and con- proliferation [92]. These intriguing results unmasked that post- tribute to the extension of the necrotic lipid core. As mentioned translational modifications may add an additional level of complexi- above, the atherosclerotic plaque generally grows silently for years ty to the implication of Cx43 in atherosclerosis. without producing any clinical symptoms. Large atheroma may cause clinical manifestations by inducing flow-limiting stenoses. Alternative- 2.4. Other repair-related mechanisms ly, rupture of the plaque's fibrous cap may provoke thrombus formation that can create an acute ischemia. Paradoxically, plaque rupture does not 2.4.1. Angiogenesis often occur at the sites with the most severe narrowing. Rather, patho- The formation of microvessels (angiogenesis) has recently re- logical observations have shown that ruptured plaques typically have a ceived increasing attention as a possible contributor to the risk of thin, collagen-poor fibrous cap with few SMCs but abundant macro- plaque rupture. Vasa vasorum, a network of microvasculature origi- phages and a large necrotic core [82]. Macrophages in the lesion general- nating primarily in the adventitia of large arteries, become activated ly produce enzymes that degrade collagen (matrix metalloproteinases; during atherosclerosis in human [93] and mice [94]. Initial observa- (MMPs))aswellasmediatorsthatinducefurtherapoptoticcelldeath tional studies as early as 1936 already pointed toward the existence of SMCs. Hence, the balance and interaction between macrophages and of microvessels in atherosclerotic lesions and their absence from the SMCs is a critical determinant for the progression and stability of ad- intima of healthy arteries [95]. More recent studies have revealed vanced atheroma. that microvessel content increases with plaque progression, a pro- Migration and proliferation of vascular SMCs as well as synthesis cess that is likely stimulated by plaque hypoxia, ROS and hypoxia- of extracellular matrix by these cells commonly involve phenotypic inducible factor signaling (see for a review: [96]). As such, plaque an- transformation of SMCs from the differentiated contractile state to giogenesis seems to be a physiological response to the diseased state the activated synthetic state. Connexins appeared to play a crucial of the arterial wall rather than a requirement for atherogenesis. role in phenotypic switch of vascular SMCs. An early study from the However, plaque microvessels are immature and fragile. This dis- group of Nicolas Severs showed that cytokine-induced modulation torted integrity of the microvessel endothelium likely leads to of contractile SMCs to synthetic SMCs coincided with more numerous intraplaque hemorrhage. The associated accumulation of red blood and larger gap junctions as well as increased Cx43 expression [83]. cell (RBC)-derived cholesterol and rapid expansion of the necrotic Transforming growth factor (TGF)-β-treated human aortic SMCs core as well as the influx of macrophages and other cells involved also upregulate Cx43 expression which correlates with increased syn- in RBC and iron phagocytosis places the plaque at increased risk for thetic activity and, paradoxically, enhanced contractile differentiation rupture [96–98]. Molecular therapies aimed at reducing plaque an- [84]. Coronary SMCs show a heterogeneous phenotype, with both giogenesis might thus be an attractive novel target to alleviate the spindle-shaped SMCs (S-SMCs) and rhomboid SMCs (R-SMCs) [85]. risk of plaque rupture. R-SMCs display an increased proliferative, migratory, proteolytic phe- The formation of new blood vessels in adults (angiogenesis) in- notype and express more Cx43 than S-SMCs. Furthermore, R-SMCs do volves branching and extension of pre-existing vessels. It may however not express Cx40, an additional gap junction protein found in S-SMCs. also occur via recruitment of endothelial progenitor cells (EPCs) from When S-SMCs are treated with PDGF they acquire a rhomboid pheno- thebonemarrow[2]. Vascular endothelial growth factor (VEGF) is the type, with a concomitant upregulation of Cx43 and a loss of Cx40 and major growth factor involved in these processes. It stimulates the sur- α-SM actin [86]. Increased Cx43 expression between intimal SMC was vival of ECs and increases their proliferation and motility in angiogene- also observed in vivo in early atherosclerotic lesions in human and sis originating from pre-existing vessels. VEGF rapidly and transiently mice [20,87]. Interestingly, these increased Cx43 expression levels be- disrupts GJIC in ECs. This effect has been correlated with changes in tween intimal SMCs declined with progression of the lesion [87].The Cx43 phosphorylation [99], thus may likely involve a change in kinetics level of Cx43 expression positively correlates with nuclear factor kappa of Cx43 gap junction channels. Connexins have been frequently associ- B(NF-κB) activation in human radial artery, suggesting the involve- ated with migration and proliferation of ECs. Microvascular ECs migrat- ment of this transcription factor in the regulation of Cx43 expression ing from the edges of a mechanically induced wound in culture have in vascular SMCs [88]. Binding of NF-κB to the Cx43 gene promoter increased Cx43 levels and enhanced Lucifer Yellow dye diffusion com- was recently demonstrated in the context of renin-dependent hyper- pared with cells at distance from the wound [100,101].Interestingly, tension in murine arteries [89]. The contribution of Cx43 to lesion for- Cx37 expression levels and permeability to Propidium Iodide, a Cx37- mation and progression was studied in atherosusceptible LDL receptor permeant molecule, showed an opposite response from Cx43 upon knockout (LDLR−/−) mice in which Cx43 expression was reduced by wounding of a monolayer bEnd3 cells, a microvascular EC line express- half (Cx43+/−) [90]. Cx43+/−LDLR−/− mice displayed reduced plaque ing all three endothelial connexins [101]. The requirement for proper formation but, more impressively, lesions in these mice had thicker fi- Cx43/Cx37 expression for coordinated migration of ECs has been dem- brous caps containing more SMCs and interstitial collagen. In addition, onstrated using a chimeric connexin (3243H7) or a fusion protein smaller lipid-cores and fewer macrophages were observed in lesions (Cx43-βGal) with dominant negative properties [101]. Specific GJIC of Cx43+/−LDLR−/− mice, whereas leukocyte counts in peripheral also regulates the formation and complexity of branches in blood were similar to the control group. Thus, targeting Cx43 may an in vitro Matrigel angiogenesis assay [102]. Many connexins have favor potential plaque stabilizing processes rather than affecting plaque been implicated in the proliferation of various cancer cell types [103], formation alone. The scenario by which reduced Cx43 ultimately leads but Cx37 seems to be the most important growth regulator for ECs. It to plaque stabilization remains to be identified, but likely involves addi- suppresses cell proliferation by increasing cell cycle time through an ex- tional functions of Cx43 in macrophages as well [91]. tension of all phases of the cell cycle and accumulating cells at the G1/S Oxidized phospholipids, the active components of oxLDL, are be- checkpoint [104] (Fig. 1 bottom right). In agreement with this notion, lieved to promote key changes in vascular SMC phenotype during ath- Cx37 deletion enhances vascular growth, including angiogenesis, and erogenesis. The effects on Cx43 expression and phosphorylation of facilitates hindlimb recovery after an ischemic insult in mice [105]. two oxidized phospholipid derivatives (1-palmitoyl-2-oxovaleroyl-sn- Whether Cx37 also affects the formation of new microvessels in athero- glycero-3-phosphorylcholine (POVPC) and 1-palmitoyl-2-glutaroyl- sclerotic lesions is at present not known. It is however of interest that sn-glycero-3-phosphocholine (PGPC)) were therefore compared in the 1019C>T polymorphism in the Cx37 gene that has been associated the mouse carotid artery after direct application of the compounds. Ap- with CAD and MI differentially affects the proliferation of the endothe- plication of POVPC decreased Cx43 levels, enhanced its phosphorylation lial SK-HEP-1 cell line [106]. Moreover, this polymorphism has also at serine 279/282 and increased SMC proliferation, whereas PGPC en- been associated with an altered incidence of hemangiosarcoma, a hanced serine 368 phosphorylation with no associated change in human endothelium-derived cancer [107]. 164 A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166

2.4.2. Lymphangiogenesis Acknowledgements Lymphatic vessels have also been detected in human coronary and carotid atherosclerotic lesions [96]. Similar to blood microvessels, This work was supported by grants from the Swiss National Sci- lymphatic microvessel content increases with plaque progression ence Foundation (310030‐127551 to BRK and 310000‐119739 to [108]. It is presently unknown whether intraplaque lymphatics are MC), and a joint grant from the Swiss National Science Foundation, protective or deleterious for plaque progression and rupture. On the the Swiss Academy of Medical Sciences and the Velux Foundation one hand, inadequate lymphangiogenesis in atherosclerotic plaques (323630‐123735 to AP). may result in inability to drain extravasated RBC, leukocytes and lipo- proteins, thus sustaining local inflammation. On the other hand, re- duction in lymphatic vessels would result in less efficient draining References plaque leukocytes and cytokines to the lymphatic nodes and lym- [1] V.L. Roger, A.S. Go, D.M. Lloyd-Jones, R.J. Adams, J.D. Berry, T.M. Brown, M.R. phoid tissues where immune cells may be sensitized and activated. Carnethon, S. Dai, G. de Simone, E.S. Ford, C.S. Fox, H.J. Fullerton, C. Gillespie, A role for gap junctions in lymphatic vessels has already been K.J. Greenlund, S.M. Hailpern, J.A. Heit, P.M. Ho, V.J. Howard, B.M. Kissela, S.J. Kittner, D.T. Lackland, J.H. Lichtman, L.D. Lisabeth, D.M. Makuc, G.M. Marcus, A. suggested in 1978 [109], but in spite of numerous studies on GJIC in Marelli, D.B. Matchar, M.M. McDermott, J.B. Meigs, C.S. Moy, D. Mozaffarian, cardiovascular system, there has only been limited characterization M.E. Mussolino, G. Nichol, N.P. Paynter, W.D. Rosamond, P.D. Sorlie, R.S. in lymphatic vessels or lymphatic ECs [110,111]. Survey microarray Stafford, T.N. Turan, M.B. Turner, N.D. Wong, J. Wylie-Rosett, Heart disease and — fi stroke statistics 2011 update: a report from the American Heart Association, studies comparing lymphatic and blood ECs identi ed the expression Circulation 123 (2011) e18–e209. of Cx43 and Cx47 in lymphatic ECs [112]. Recent in situ studies re- [2] R.N. Mitchell, F.J. Schoen, Atherosclerosis, in: V. Kumar, A.K. Abbas, N. Fausto, J.C. vealed the expression of Cx37, Cx43 and Cx47 during mouse lym- Aster (Eds.), Robbins and Cotran Pathologic Basis of Disease, Saunders, Philadelphia, 2009, pp. 496–506. phatic development [113,114]. Interestingly, differential Cx37/Cx43 [3] R. Ross, Atherosclerosis—an inflammatory disease, N. Engl. J. Med. 340 (1999) expression patterns were observed between (future) valve-forming 115–126. regions and the lymphangion [114] as well as on upstream and down- [4] G.K. Hansson, A. Hermansson, The immune system in atherosclerosis, Nat. Immunol. 12 (2011) 204–212. stream sides of lymphatic valves [113]. Lymphatic valve develop- [5] T.T. Tuomisto, B.R. Binder, S. Yla-Herttuala, Genetics, genomics and proteomics ment and function are profoundly disturbed in Cx37-deficient mice in atherosclerosis research, Ann. Med. 37 (2005) 323–332. [113,114]. Finally, Cx47 expression was only weakly detected in early [6] H. Roy, S. Bhardwaj, S. Yla-Herttuala, Molecular genetics of atherosclerosis, – lymphatic vessels but became later on highly enriched in a subset of Hum. Genet. 125 (2009) 467 491. [7] B.D. Chiodini, S. Barlera, M.G. Franzosi, V.L. Beceiro, M. Introna, G. Tognoni, APO ECs in lymphatic valves [113]. Intriguing mutations in Cx47 gene have B gene polymorphisms and coronary artery disease: a meta-analysis, Athero- been recently linked to a subset of human hereditary lymphedemas, sclerosis 167 (2003) 355–366. although the nature of lymphatic vascular defect or molecular mecha- [8] P.W. Wilson, E.J. Schaefer, M.G. Larson, J.M. Ordovas, Apolipoprotein E alleles and risk of coronary disease. A meta-analysis, Arterioscler. Thromb. Vasc. Biol. nism is at present unknown [115,116]. Although these recent studies 16 (1996) 1250–1255. point toward a role for connexins in physiological development of lym- [9] U. Paradossi, E. Ciofini, A. Clerico, N. Botto, A. Biagini, M.G. Colombo, Endothelial phatics, no studies have been published so far that evaluate the role of function and carotid intima-media thickness in young healthy subjects among endothelial nitric oxide synthase Glu298–>Asp and T-786-->C polymorphisms, connexins in lymphatic vascular function in disease. Stroke 35 (2004) 1305–1309. [10] J.P. Casas, L.E. Bautista, S.E. Humphries, A.D. Hingorani, Endothelial nitric oxide 3. Conclusions and perspectives synthase genotype and ischemic heart disease: meta-analysis of 26 studies in- volving 23028 subjects, Circulation 109 (2004) 1359–1365. [11] M. Boerma, L. Forsberg, L. Van Zeijl, R. Morgenstern, U. De Faire, C. Lemne, D. In summary, three connexins in several subtypes have been di- Erlinge, T. Thulin, Y. Hong, I.A. Cotgreave, A genetic polymorphism in connexin rectly implicated in the development of atherosclerosis. These are 37 as a prognostic marker for atherosclerotic plaque development, J. Intern. Med. 246 (1999) 211–218. Cx37 in monocytes and platelets, Cx40 in the endothelium, and [12] C.W. Wong, T. Christen, A. Pfenniger, R.W. James, B.R. Kwak, Do allelic variants of Cx43 in SMCs. However numerous other cell types that play a role the connexin37 1019 gene polymorphism differentially predict for coronary ar- in atherosclerosis also express one (or more) of these connexins. As tery disease and myocardial infarction? Atherosclerosis 191 (2007) 355–361. described in this review, a function could generally be ascribed to a [13] H.I. Yeh, Y. Chou, H.F. Liu, S.C. Chang, C.H. Tsai, Connexin37 gene polymorphism and coronary artery disease in Taiwan, Int. J. Cardiol. 81 (2001) 251–255. connexin expressed by each cell type. It thus seems very likely that [14] F. Listi, G. Candore, D. Lio, M. Russo, G. Colonna-Romano, M. Caruso, E. future studies will be able to find other implications for connexins Hoffmann, C. Caruso, Association between C1019T polymorphism of connexin37 in atherosclerosis development. and acute myocardial infarction: a study in patients from Sicily, Int. J. Cardiol. 102 (2005) 269–271. It is increasingly evident that connexins are generally not expressed [15] H.B. Leu, C.M. Chung, S.Y. Chuang, C.H. Bai, J.R. Chen, J.W. Chen, W.H. Pan, Genetic by a single cell type, and that they may have distinct or even opposite variants of connexin37 are associated with carotid intima-medial thickness and fu- roles depending on the cell type studied. Thus, the association between ture onset of ischemic stroke, Atherosclerosis 214 (2011) 101–106. [16] J.C. Saez, V.M. Berthoud, M.C. Branes, A.D. Martinez, E.C. Beyer, Plasma mem- one connexin and a complex disease like atherosclerosis is not straight- brane channels formed by connexins: their regulation and functions, Physiol. forward. A genetic polymorphism in a connexin gene may have multiple Rev. 83 (2003) 1359–1400. additive or counteracting effects on disease development. This high- [17] W.H. Evans, P.E. Martin, Gap junctions: structure and function (Review), Mol. – fi Membr. Biol. 19 (2002) 121 136. lights the need for cell-speci c connexin knock-out models to detangle [18] T. Munzel, C. Sinning, F. Post, A. Warnholtz, E. Schulz, Pathophysiology, diagno- the complicated network of connexin implications in atherosclerosis. sis and prognostic implications of endothelial dysfunction, Ann. Med. 40 (2008) Another recurrent question is whether the results obtained in 180–196. [19] Z. Yang, X.F. Ming, Recent advances in understanding endothelial dysfunction in mouse models can be applied to humans. Although mouse models atherosclerosis, Clin. Med. Res. 4 (2006) 53–65. closely mimic human pathology, there are important differences. [20] B.R. Kwak, F. Mulhaupt, N. Veillard, D.B. Gros, F. Mach, Altered pattern of vascu- For instance, plaque rupture with superimposed , the lar connexin expression in atherosclerotic plaques, Arterioscler. Thromb. Vasc. – most common complication of human atherosclerosis, is rarely ob- Biol. 22 (2002) 225 230. [21] J.E. Gabriels, D.L. Paul, Connexin43 is highly localized to sites of disturbed flow in served in mice. Consequently, clinical events such as myocardial in- rat aortic endothelium but connexin37 and connexin40 are more uniformly dis- farction or ischemic stroke are almost never seen in these models. tributed, Circ. Res. 83 (1998) 636–643. Moreover, the small size of mice excludes research in interventional [22] H.V. van Rijen, M.J. van Kempen, S. Postma, H.J. Jongsma, Tumour necrosis factor alpha alters the expression of connexin43, connexin40, and connexin37 in cardiology and diagnostic imaging. Although these problems general- human umbilical endothelial cells, Cytokine 10 (1998) 258–264. ly apply to atherosclerosis research, they will hinder further develop- [23] W. Meme, P. Ezan, L. Venance, J. Glowinski, C. Giaume, ATP-induced inhibition of ments toward connexin-based therapies and await the development gap junctional communication is enhanced by interleukin-1 beta treatment in cultured astrocytes, Neuroscience 126 (2004) 95–104. of new animal models as well as of minimally-invasive methods [24] M.A. Retamal, N. Froger, N. Palacios-Prado, P. Ezan, P.J. Saez, J.C. Saez, C. Giaume, that would enable to confirm the results in humans. Cx43 hemichannels and gap junction channels in astrocytes are regulated A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166 165

oppositely by proinflammatory cytokines released from activated microglia, [50] M. Drechsler, R.T. Megens, M. van Zandvoort, C. Weber, O. Soehnlein, Hyperlipidemia- J. Neurosci. 27 (2007) 13781–13792. triggered neutrophilia promotes early atherosclerosis, Circulation 122 (2010) [25] T.Y. Huang, T.F. Chu, H.I. Chen, C.J. Jen, Heterogeneity of [Ca(2+)](i) signaling in 1837–1845. intact rat aortic endothelium, FASEB J. 14 (2000) 797–804. [51] P. Rotzius, S. Thams, O. Soehnlein, E. Kenne, C.N. Tseng, N.K. Bjorkstrom, K.J. [26] P. Kameritsch, K. Pogoda, A. Ritter, S. Munzing, U. Pohl, Gap junctional commu- Malmberg, L. Lindbom, E.E. Eriksson, Distinct infiltration of neutrophils in lesion nication controls the overall endothelial calcium response to vasoactive ago- shoulders in ApoE−/− mice, Am. J. Pathol. 177 (2010) 493–500. nists, Cardiovasc. Res. 93 (2012) 508–515. [52] A. Zernecke, I. Bot, Y. Djalali-Talab, E. Shagdarsuren, K. Bidzhekov, S. Meiler, R. [27] A. Pfenniger, J.P. Derouette, V. Verma, X. Lin, B. Foglia, W. Coombs, I. Roth, N. Krohn, A. Schober, M. Sperandio, O. Soehnlein, J. Bornemann, F. Tacke, E.A. Satta, S. Dunoyer-Geindre, P. Sorgen, S. Taffet, B.R. Kwak, M. Delmar, Gap junc- Biessen, C. Weber, Protective role of CXC receptor 4/CXC ligand 12 unveils tion protein Cx37 interacts with endothelial nitric oxide synthase in endothelial the importance of neutrophils in atherosclerosis, Circ. Res. 102 (2008) cells, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 827–834. 209–217. [28] F. Alonso, F.X. Boittin, J.L. Beny, J.A. Haefliger, Loss of connexin40 is associated [53] M.G. Ionita, P. van den Borne, L.M. Catanzariti, F.L. Moll, J.P. de Vries, G. with decreased endothelium-dependent relaxations and eNOS levels in the Pasterkamp, A. Vink, D.P. de Kleijn, High neutrophil numbers in human carotid mouse aorta, Am. J. Physiol. Heart Circ. Physiol. 299 (2010) H1365–H1373. atherosclerotic plaques are associated with characteristics of rupture-prone le- [29] C.E. Chadjichristos, K.E. Scheckenbach, T.A. van Veen, M.Z. Richani Sarieddine, C. sions, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 1842–1848. de Wit, Z. Yang, I. Roth, M. Bacchetta, H. Viswambharan, B. Foglia, T. Dudez, M.J. [54] M.M. Averill, S. Barnhart, L. Becker, X. Li, J.W. Heinecke, R.C. Leboeuf, J.A. van Kempen, F.E. Coenjaerts, L. Miquerol, U. Deutsch, H.J. Jongsma, M. Chanson, Hamerman, C. Sorg, C. Kerkhoff, K.E. Bornfeldt, S100A9 differentially modifies B.R. Kwak, Endothelial-specific deletion of connexin40 promotes atherosclerosis phenotypic states of neutrophils, macrophages, and dendritic cells: implications by increasing CD73-dependent leukocyte adhesion, Circulation 121 (2010) for atherosclerosis and adipose tissue inflammation, Circulation 123 (2011) 123–131. 1216–1226. [30] H.Y. Stevens, B. Melchior, K.S. Bell, S. Yun, J.C. Yeh, J.A. Frangos, PECAM-1 is a critical [55] M. Drechsler, Y. Doring, R.T. Megens, O. Soehnlein, Neutrophilic granulocytes — mediator of atherosclerosis, Dis. Model Mech. 1 (2008) 175–181 (discussion 179). promiscuous accelerators of atherosclerosis, Thromb. Haemost. 106 (2011) [31] M.Z. Sarieddine, K.E. Scheckenbach, B. Foglia, K. Maass, I. Garcia, B.R. Kwak, M. 839–848. Chanson, Connexin43 modulates neutrophil recruitment to the lung, J. Cell. [56] O. Soehnlein, C. Weber, L. Lindbom, Neutrophil granule proteins tune monocytic Mol. Med. 13 (2009) 4560–4570. cell function, Trends Immunol. 30 (2009) 538–546. [32] C.W. Wong, T. Christen, I. Roth, C.E. Chadjichristos, J.P. Derouette, B.F. Foglia, M. [57] N. Borregaard, O.E. Sorensen, K. Theilgaard-Monch, Neutrophil granules: a li- Chanson, D.A. Goodenough, B.R. Kwak, Connexin37 protects against atheroscle- brary of innate immunity proteins, Trends Immunol. 28 (2007) 340–345. rosis by regulating monocyte adhesion, Nat. Med. 12 (2006) 950–954. [58] F.R. Tavora, M. Ripple, L. Li, A.P. Burke, Monocytes and neutrophils expressing [33] E.A. Eugenin, M.C. Branes, J.W. Berman, J.C. Saez, TNF-alpha plus IFN-gamma in- myeloperoxidase occur in fibrous caps and thrombi in unstable coronary duce connexin43 expression and formation of gap junctions between human plaques, BMC Cardiovasc. Disord. 9 (2009) 27. monocytes/macrophages that enhance physiological responses, J. Immunol. [59] J.C. Saez, M.C. Branes, L.A. Corvalan, E.A. Eugenin, H. Gonzalez, A.D. 170 (2003) 1320–1328. Martinez, F. Palisson, Gap junctions in cells of the immune system: struc- [34] R.J. Anand, S. Dai, S.C. Gribar, W. Richardson, J.W. Kohler, R.A. Hoffman, M.F. Branca, ture, regulation and possible functional roles, Braz. J. Med. Biol. Res. 33 J. Li, X.H. Shi, C.P. Sodhi, D.J. Hackam, A role for connexin43 in macrophage phago- (2000) 447–455. cytosis and host survival after bacterial peritoneal infection, J. Immunol. 181 (2008) [60] S. Zahler, A. Hoffmann, T. Gloe, U. Pohl, Gap-junctional coupling between neu- 8534–8543. trophils and endothelial cells: a novel modulator of transendothelial migration, [35] J. Andersson, P. Libby, G.K. Hansson, Adaptive immunity and atherosclerosis, J. Leukoc. Biol. 73 (2003) 118–126. Clin. Immunol. 134 (2010) 33–46. [61] P.I. Jara, M.P. Boric, J.C. Saez, Leukocytes express connexin 43 after activation [36] E. Oviedo-Orta, T. Hoy, W.H. Evans, Intercellular communication in the immune with lipopolysaccharide and appear to form gap junctions with endothelial system: differential expression of connexin40 and 43, and perturbation of gap cells after ischemia–reperfusion, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) junction channel functions in peripheral blood and tonsil human lymphocyte 7011–7015. subpopulations, Immunology 99 (2000) 578–590. [62] I. Scerri, O. Tabary, T. Dudez, J. Jacquot, B. Foglia, S. Suter, M. Chanson, Gap junc- [37] A. Bermudez-Fajardo, M. Yliharsila, W.H. Evans, A.C. Newby, E. Oviedo-Orta, CD4+ tional communication does not contribute to the interaction between neutro- T lymphocyte subsets express connexin 43 and establish gap junction chan- phils and airway epithelial cells, Cell Commun. Adhes. 13 (2006) 1–12. nel communication with macrophages in vitro,J.Leukoc.Biol.82(2007) [63] M.C. Branes, J.E. Contreras, J.C. Saez, Activation of human polymorphonuclear 608–612. cells induces formation of functional gap junctions and expression of connexins, [38] M. Kuczma, J.R. Lee, P. Kraj, Connexin 43 signaling enhances the generation of Med. Sci. Monit. 8 (2002) BR313–BR323. Foxp3+ regulatory T cells, J. Immunol. 187 (2011) 248–257. [64] E. Oviedo-Orta, W. Howard Evans, Gap junctions and connexin-mediated com- [39] T.D. Nguyen, S.M. Taffet, A model system to study Connexin 43 in the immune munication in the immune system, Biochim. Biophys. Acta 1662 (2004) system, Mol. Immunol. 46 (2009) 2938–2946. 102–112. [40] A. Mendoza-Naranjo, G. Bouma, C. Pereda, M. Ramirez, K.F. Webb, A. Tittarelli, [65] H.K. Eltzschig, T. Eckle, A. Mager, N. Kuper, C. Karcher, T. Weissmuller, K. M.N. Lopez, A.M. Kalergis, A.J. Thrasher, D.L. Becker, F. Salazar-Onfray, Functional Boengler, R. Schulz, S.C. Robson, S.P. Colgan, ATP release from activated neutro- gap junctions accumulate at the immunological synapse and contribute to T cell phils occurs via connexin 43 and modulates adenosine-dependent endothelial activation, J. Immunol. 187 (2011) 3121–3132. cell function, Circ. Res. 99 (2006) 1100–1108. [41] T. Bopp, C. Becker, M. Klein, S. Klein-Hessling, A. Palmetshofer, E. Serfling, V. [66] A. Zarbock, R.K. Polanowska-Grabowska, K. Ley, Platelet–neutrophil-interac- Heib, M. Becker, J. Kubach, S. Schmitt, S. Stoll, H. Schild, M.S. Staege, M. tions: linking hemostasis and inflammation, Blood Rev. 21 (2007) 99–111. Stassen, H. Jonuleit, E. Schmitt, Cyclic adenosine monophosphate is a key com- [67] K.N. Kornerup, G.P. Salmon, S.C. Pitchford, W.L. Liu, C.P. Page, Circulating platelet– ponent of regulatory T cell-mediated suppression, J. Exp. Med. 204 (2007) neutrophil complexes are important for subsequent neutrophil activation and 1303–1310. migration, J. Appl. Physiol. 109 (2010) 758–767. [42] S. Machtaler, M. Dang-Lawson, K. Choi, C. Jang, C.C. Naus, L. Matsuuchi, The gap [68] R.R. Koenen, C. Weber, Platelet-derived chemokines in vascular remodeling and junction protein Cx43 regulates B-lymphocyte spreading and adhesion, J. Cell atherosclerosis, Semin. Thromb. Hemost. 36 (2010) 163–169. Sci. 124 (2011) 2611–2621. [69] A. Angelillo-Scherrer, P. Fontana, L. Burnier, I. Roth, R. Sugamele, A. Brisset, S. [43] E.K. Koltsova, K. Ley, How dendritic cells shape atherosclerosis, Trends Immunol. Morel, S. Nolli, E. Sutter, A. Chassot, C. Capron, D. Borgel, F. Saller, M. Chanson, 32 (2011) 540–547. B.R. Kwak, Connexin 37 limits thrombus propensity by downregulating platelet [44] I.T. Daissormont, A. Christ, L. Temmerman, S. Sampedro Millares, T. Seijkens, M. reactivity, Circulation 124 (2011) 930–939. Manca, M. Rousch, M. Poggi, L. Boon, C. van der Loos, M. Daemen, E. Lutgens, B. [70] M. Kaartinen, A. Penttila, P.T. Kovanen, Extracellular mast cell granules carry Halvorsen, P. Aukrust, E. Janssen, E.A. Biessen, Plasmacytoid dendritic cells pro- apolipoprotein B-100-containing into phagocytes in human arterial tect against atherosclerosis by tuning T-cell proliferation and activity, Circ. Res. intima. Functional coupling of exocytosis and phagodytosis in neighboring cells, 109 (2011) 1387–1395. Arterioscler. Thromb. Vasc. Biol. 15 (1995) 2047–2054. [45] H. Matsue, J. Yao, K. Matsue, A. Nagasaka, H. Sugiyama, R. Aoki, M. Kitamura, S. [71] M. Kaartinen, A. Penttila, P.T. Kovanen, Accumulation of activated mast cells in Shimada, Gap junction-mediated intercellular communication between dendrit- the shoulder region of human coronary atheroma, the predilection site of ather- ic cells (DCs) is required for effective activation of DCs, J. Immunol. 176 (2006) omatous rupture, Circulation 90 (1994) 1669–1678. 181–190. [72] P.T. Kovanen, Mast cells and degradation of pericellular and extracellular matrices: [46] J. Neijssen, C. Herberts, J.W. Drijfhout, E. Reits, L. Janssen, J. Neefjes, Cross- potential contributions to erosion, rupture and intraplaque haemorrhage of athero- presentation by intercellular peptide transfer through gap junctions, Nature sclerotic plaques, Biochem. Soc. Trans. 35 (2007) 857–861. 434 (2005) 83–88. [73] J. Sun, G.K. Sukhova, P.J. Wolters, M. Yang, S. Kitamoto, P. Libby, L.A. MacFarlane, [47] A. Mendoza-Naranjo, P.J. Saez, C.C. Johansson, M. Ramirez, D. Mandakovic, C. J. Mallen-St Clair, G.P. Shi, Mast cells promote atherosclerosis by releasing Pereda, M.N. Lopez, R. Kiessling, J.C. Saez, F. Salazar-Onfray, Functional gap junc- proinflammatory cytokines, Nat. Med. 13 (2007) 719–724. tions facilitate melanoma antigen transfer and cross-presentation between [74] I. Bot, S.C. de Jager, A. Zernecke, K.A. Lindstedt, T.J. van Berkel, C. Weber, E.A. Biessen, human dendritic cells, J. Immunol. 178 (2007) 6949–6957. Perivascular mast cells promote atherogenesis and induce plaque destabilization in [48] B. Pang, J. Neijssen, X. Qiao, L. Janssen, H. Janssen, C. Lippuner, J. Neefjes, Direct apolipoprotein E-deficient mice, Circulation 115 (2007) 2516–2525. antigen presentation and gap junction mediated cross-presentation during apo- [75] Y. Inoue, T.E. King Jr., S.S. Tinkle, K. Dockstader, L.S. Newman, Human mast cell ptosis, J. Immunol. 183 (2009) 1083–1090. basic fibroblast growth factor in pulmonary fibrotic disorders, Am. J. Pathol. [49] J. Gewaltig, M. Kummer, C. Koella, G. Cathomas, B.C. Biedermann, Requirements 149 (1996) 2037–2054. for CD8 T-cell migration into the human arterial wall, Hum. Pathol. 39 (2008) [76] N. Toda, Mechanism of histamine actions in human coronary arteries, Circ. Res. 1756–1762. 61 (1987) 280–286. 166 A. Pfenniger et al. / Biochimica et Biophysica Acta 1828 (2013) 157–166

[77] Y. Wang, P.T. Kovanen, Heparin proteoglycans released from rat serosal mast [97] F.D. Kolodgie, J. Narula, C. Yuan, A.P. Burke, A.V. Finn, R. Virmani, Elimination of cells inhibit proliferation of rat aortic smooth muscle cells in culture, Circ. Res. neoangiogenesis for plaque stabilization: is there a role for local drug therapy? 84 (1999) 74–83. J. Am. Coll. Cardiol. 49 (2007) 2093–2101. [78] T.T. Foley, G.C. Saggers, K.E. Moyer, H.P. Ehrlich, Rat mast cells enhance fibroblast [98] J.B. Michel, R. Virmani, E. Arbustini, G. Pasterkamp, Intraplaque haemorrhages as the proliferation and fibroblast-populated collagen lattice contraction through gap trigger of plaque vulnerability, Eur. Heart J. 32 (2011) 1977–1985 (1985a, 1985b). junctional intercellular communications, Plast. Reconstr. Surg. 127 (2011) [99] S. Suarez, K. Ballmer-Hofer, VEGF transiently disrupts gap junctional communi- 1478–1486. cation in endothelial cells, J. Cell Sci. 114 (2001) 1229–1235. [79] A.L. Pistorio, H.P. Ehrlich, Modulatory effects of connexin-43 expression on gap [100] M.S. Pepper, D.C. Spray, M. Chanson, R. Montesano, L. Orci, P. Meda, Junctional junction intercellular communications with mast cells and fibroblasts, J. Cell. communication is induced in migrating capillary endothelial cells, J. Cell Biol. Biochem. 112 (2011) 1441–1449. 109 (1989) 3027–3038. [80] G.K. Hansson, Inflammation, atherosclerosis, and coronary artery disease, N. [101] B.R. Kwak, M.S. Pepper, D.B. Gros, P. Meda, Inhibition of endothelial wound re- Engl. J. Med. 352 (2005) 1685–1695. pair by dominant negative connexin inhibitors, Mol. Biol. Cell 12 (2001) [81] A. Saiura, M. Sata, Y. Hirata, R. Nagai, M. Makuuchi, Circulating smooth muscle 831–845. progenitor cells contribute to atherosclerosis, Nat. Med. 7 (2001) 382–383. [102] C. Gartner, B. Ziegelhoffer, M. Kostelka, H. Stepan, F.W. Mohr, S. Dhein, Knock- [82] A.V. Finn, M. Nakano, J. Narula, F.D. Kolodgie, R. Virmani, Concept of vul- down of endothelial connexins impairs angiogenesis, Pharmacol. Res. 65 nerable/unstable plaque, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 1282–1292. (2012) 347–357. [83] R.E. Rennick, J.L. Connat, G. Burnstock, S. Rothery, N.J. Severs, C.R. Green, Expres- [103] L. Cronier, S. Crespin, P.O. Strale, N. Defamie, M. Mesnil, Gap junctions and can- sion of connexin43 gap junctions between cultured vascular smooth muscle cer: new functions for an old story, Antioxid. Redox Signal. 11 (2009) 323–338. cells is dependent upon phenotype, Cell Tissue Res. 271 (1993) 323–332. [104] J.M. Burt, T.K. Nelson, A.M. Simon, J.S. Fang, Connexin 37 profoundly slows cell [84] A. Rama, T. Matsushita, N. Charolidi, S. Rothery, E. Dupont, N.J. Severs, Up-regulation cycle progression in rat insulinoma cells, Am. J. Physiol. Cell Physiol. 295 of connexin43 correlates with increased synthetic activity and enhanced contractile (2008) C1103–C1112. differentiation in TGF-beta-treated human aortic smooth muscle cells, Eur. J. Cell [105] J.S. Fang, S.N. Angelov, A.M. Simon, J.M. Burt, Cx37 deletion enhances vascular Biol. 85 (2006) 375–386. growth and facilitates ischemic limb recovery, Am. J. Physiol. Heart Circ. Physiol. [85] H. Hao, P. Ropraz, V. Verin, E. Camenzind, A. Geinoz, M.S. Pepper, G. Gabbiani, 301 (2011) H1872–H1881. M.L. Bochaton-Piallat, Heterogeneity of smooth muscle cell populations cultured [106] S. Morel, L. Burnier, A. Roatti, A. Chassot, I. Roth, E. Sutter, K. Galan, A. Pfenniger, from pig coronary artery, Arterioscler. Thromb. Vasc. Biol. 22 (2002) 1093–1099. M. Chanson, B.R. Kwak, Unexpected role for the human Cx37 C1019T polymor- [86] C.E. Chadjichristos, S. Morel, J.P. Derouette, E. Sutter, I. Roth, A.C. Brisset, M.L. phism in tumour cell proliferation, Carcinogenesis 31 (2010) 1922–1931. Bochaton-Piallat, B.R. Kwak, Targeting connexin 43 prevents platelet-derived [107] T. Saito, V. Krutovskikh, M.J. Marion, K.G. Ishak, W.P. Bennett, H. Yamasaki, growth factor-BB-induced phenotypic change in porcine coronary artery Human hemangiosarcomas have a common polymorphism but no mutations smooth muscle cells, Circ. Res. 102 (2008) 653–660. in the connexin37 gene, Int. J. Cancer 86 (2000) 67–70. [87] J.P. Blackburn, N.S. Peters, H.I. Yeh, S. Rothery, C.R. Green, N.J. Severs, [108] I. Kholova, G. Dragneva, P. Cermakova, S. Laidinen, N. Kaskenpaa, T. Hazes, E. Upregulation of connexin43 gap junctions during early stages of human coro- Cermakova, I. Steiner, S. Yla-Herttuala, Lymphatic vasculature is increased in nary atherosclerosis, Arterioscler. Thromb. Vasc. Biol. 15 (1995) 1219–1228. heart valves, ischaemic and inflamed hearts and in cholesterol-rich and calcified [88] K. Arishiro, M. Hoshiga, T. Ishihara, K. Kondo, T. Hanafusa, Connexin 43 expres- atherosclerotic lesions, Eur. J. Clin. Invest. 41 (2011) 487–497. sion is associated with vascular activation in human radial artery, Int. J. Cardiol. [109] J.A. Rhodin, Microscopic anatomy of the pulmonary vascular bed in the cat lung, 145 (2010) 270–272. Microvasc. Res. 15 (1978) 169–193. [89] F. Alonso, N. Krattinger, L. Mazzolai, A. Simon, G. Waeber, P. Meda, J.A. Haefliger, [110] D.C. Zawieja, K.L. Davis, R. Schuster, W.M. Hinds, H.J. Granger, Distribution, prop- An angiotensin II- and NF-kappaB-dependent mechanism increases connexin 43 agation, and coordination of contractile activity in lymphatics, Am. J. Physiol. in murine arteries targeted by renin-dependent hypertension, Cardiovasc. Res. 264 (1993) H1283–H1291. 87 (2010) 166–176. [111] N.G. McHale, M.K. Meharg, Co-ordination of pumping in isolated bovine lym- [90] B.R. Kwak, N. Veillard, G. Pelli, F. Mulhaupt, R.W. James, M. Chanson, F. Mach, Re- phatic vessels, J. Physiol. 450 (1992) 503–512. duced connexin43 expression inhibits atherosclerotic lesion formation in low- [112] N. Wick, P. Saharinen, J. Saharinen, E. Gurnhofer, C.W. Steiner, I. Raab, D. Stokic, density lipoprotein receptor-deficient mice, Circulation 107 (2003) 1033–1039. P. Giovanoli, S. Buchsbaum, A. Burchard, S. Thurner, K. Alitalo, D. Kerjaschki, [91] C.W. Wong, F. Burger, G. Pelli, F. Mach, B.R. Kwak, Dual benefit of reduced Cx43 Transcriptomal comparison of human dermal lymphatic endothelial cells ex on atherosclerosis in LDL receptor-deficient mice, Cell Commun. Adhes. 10 vivo and in vitro, Physiol. Genomics 28 (2007) 179–192. (2003) 395–400. [113] J.D. Kanady, M.T. Dellinger, S.J. Munger, M.H. Witte, A.M. Simon, Connexin37 and [92] S.R. Johnstone, J. Ross, M.J. Rizzo, A.C. Straub, P.D. Lampe, N. Leitinger, B.E. Connexin43 deficiencies in mice disrupt lymphatic valve development and re- Isakson, Oxidized phospholipid species promote in vivo differential cx43 phos- sult in lymphatic disorders including lymphedema and , Dev. Biol. phorylation and vascular smooth muscle cell proliferation, Am. J. Pathol. 175 354 (2011) 253–266. (2009) 916–924. [114] A. Sabine, Y. Agalarov, E. Maby-El Hajjami, M. Jaquet, R. Hägerling, C. Pollmann, [93] R. Virmani, F.D. Kolodgie, A.P. Burke, A.V. Finn, H.K. Gold, T.N. Tulenko, S.P. D. Bebber, A. Pfenniger, N. Miura, O. Dormond, J.M. Calmes, R.H. Adams, T. Wrenn, J. Narula, Atherosclerotic plaque progression and vulnerability to rup- Mäkinen, F. Kiefer, B.R. Kwak, T.V. Petrova, Mechanotransduction, PROX1, and ture: angiogenesis as a source of intraplaque hemorrhage, Arterioscler. Thromb. FOXC2 cooperate to control connexin37 and calcineurin during lymphatic- Vasc. Biol. 25 (2005) 2054–2061. valve formation, Dev. Cell 22 (2012) 430–445. [94] K.S. Moulton, K. Vakili, D. Zurakowski, M. Soliman, C. Butterfield, E. Sylvin, K.M. [115] R.E. Ferrell, C.J. Baty, M.A. Kimak, J.M. Karlsson, E.C. Lawrence, M. Franke-Snyder, Lo, S. Gillies, K. Javaherian, J. Folkman, Inhibition of plaque neovascularization S.D. Meriney, E. Feingold, D.N. Finegold, GJC2 missense mutations cause human reduces macrophage accumulation and progression of advanced atherosclerosis, lymphedema, Am. J. Hum. Genet. 86 (2010) 943–948. Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 4736–4741. [116] P. Ostergaard, M.A. Simpson, G. Brice, S. Mansour, F.C. Connell, A. Onoufriadis, [95] J.C. Paterson, Vascularization and hemorrhage of the intima of atherosclerotic A.H. Child, J. Hwang, K. Kalidas, P.S. Mortimer, R. Trembath, S. Jeffery, Rapid coronary arteries, Arch. Pathol. 22 (1936) 313–324. identification of mutations in GJC2 in primary lymphoedema using whole [96] J.C. Sluimer, M.J. Daemen, Novel concepts in atherogenesis: angiogenesis and exome sequencing combined with linkage analysis with delineation of the phe- hypoxia in atherosclerosis, J. Pathol. 218 (2009) 7–29. notype, J. Med. Genet. 48 (2011) 251–255.