Connexins: New genes in atherosclerosis

Abstract

Atherosclerosis, the main cause of death and disability in adult populations of industrialized societies, is a multifactorial progressive process involving a variety of pathogenic mechanisms. Our current view on the pathogenesis of the disease implies complex patterns of interactions between a dysfunctional endothelium, leukocytes, and activated smooth muscle cells in which cytokines and growth factors are known to play a crucial role. Apart from paracrine cell‐to‐cell signalling, a role for gap junction‐mediated intercellular communication in the development of the disease has been recently suggested. Gap junction channels result from the docking of two hemichannels or connexons, formed by the hexameric assembly of connexins, and directly connect the cytoplasm of adjacent cells. In this review, we summarize existing evidence implicating connexins in atherosclerosis. Indeed, the expression pattern of vascular connexins is altered during atherosclerotic plaque formation. In addition, changes in connexin expression or gap junctional communication have been observed in vascular cells in vitro by disturbances in blood flow, cholesterol, inflammatory cytokines, and growth factors. Furthermore, genetically modifying connexin expression affects the course of the atherosclerotic process in mouse models of the disease. Finally, the involvement of connexins in treatment of atherosclerotic disease will be discussed.

• Atherosclerosis
• connexin
• endothelium
• gap junction
• leukocytes
• smooth muscle cells
• vascular wall

Abbreviations
ADP	=	adenosine diphosphate
apoB	=	apolipoprotein B
apoE	=	apolipoprotein E
ATP	=	adenosine triphosphate
Cx	=	connexion
ECs	=	endothelial cells
ECM	=	extra‐cellular matrix
GJIC	=	gap junctional intercellular communication
HMG‐CoA	=	3‐hydroxy‐3‐methylglutaryl‐CoA
IL	=	interleukin
LDL	=	low‐density lipoprotein
LDLR	=	LOW‐DENSITY LIPOPROTEIN RECEPTOR
LPS	=	lipopolysaccharide
SMCs	=	smooth muscle cells
TNF	=	tumour necrosis factor

Pathogenesis of atherosclerosis

Atherosclerosis leads as cause of death and disability in the Western world. This progressive disease is characterized by the accumulation of lipids, T lymphocytes and macrophages as well as fibrous elements in large and medium‐sized arteries. Numerous observations in human and animals have led to the formulation of the ‘response‐to‐injury’ hypothesis of atherosclerosis 1, 2. This hypothesis emphasizes endothelium dysfunction as the initiating step of the disease. Possible causes of endothelium dysfunction leading to atherosclerosis include elevated low‐density lipoprotein (LDL), obesity, free radicals (caused by cigarette smoking, for example), hypertension, diabetes, and infectious micro‐organisms 3–5. As a response to these harmful agents, the injured endothelium displays increased expression of a variety of cell adhesion molecules and secretes chemoattractants to recruit specific leukocytes 6. However, despite a given patient's risk factor profile or type of animal model, the early lipid‐rich lesions of atherosclerosis show a markedly non‐random pattern of distribution within the arterial vasculature. Atherosclerotic plaques often form at branch points of arteries 7, which are regions of disturbed blood flow, suggesting that local fluid dynamic factors may contribute to the focal distribution of lesions. Indeed, various in vitro and in vivo experiments have shown flow‐induced changes in the endothelial expression of pro‐inflammatory chemokines and adhesion molecules 8–10.

Leukocyte recruitment in the early phases of atherosclerosis involves mostly monocytes. However, T lymphocytes are also implicated in the early development of the disease 11. After adhering to the dysfunctional endothelium, monocytes transmigrate between intact endothelial cells (ECs) to infiltrate into the arterial intima. Once in the intima, monocytes propagate and mature under the influence of cytokines, chemokines, and growth factors secreted by themselves and other atheroma‐associated cells. In addition, the induced expression of scavenger receptors allows macrophages to accumulate lipids within their cytoplasm and eventually progress to the arterial foam cells, a hall‐mark of the atherosclerotic lesion. These foam cells along with the T cells represent the fatty streak known as the earliest form of atherosclerotic plaque.

The continued inflammatory response and accumulation of lipids collaborate with other events to promote atherosclerotic plaque growth and eventually rupture 4, 5. During the growing phase, medial smooth muscle cells (SMCs) migrate to the top of the intima where they proliferate and generate extra‐cellular matrix components. The SMCs and matrix molecules coalesce to structure a strong fibrous cap that covers the original atherosclerotic lesion. Although this adds to the size of the plaque, it also seals the plaque off safely from the blood and reduces the possibility of rupture. As this cap matures some of the foam cells underneath die and lipids are released, thus forming the necrotic or lipid core of the advanced atherosclerotic plaque. Finally, the fibrous cap of a plaque might rupture, triggering thrombus formation at the site of the lesion.

Rupture of an atherosclerotic plaque is the primary cause of sudden cardiac death, accounting for 60% of sudden deaths with thrombosis 12. In human, plaques that are most likely to break possess a thinned cap, a large lipid core, and many inflammatory cells 5, 6, 11, 12. This plaque phenotype is partially dependent on the macrophage activity. Macrophage foam cells produce reactive oxygen species and pro‐inflammatory cytokines that intensify the local inflammatory response and further induce macrophage proliferation and lipid uptake. In addition, the activated macrophages produce matrix metalloproteinases that can degrade the extra‐cellular matrix (ECM) components thus deteriorating the plaque's fibrous cap and increasing its likelihood of rupture 13.

Key messages

• The development of atherosclerosis, a chronic immuno‐inflammatory disease, critically depends on co‐ordinated interactions between circulating blood cells and cells that reside within the arterial wall.

• Gap junction channels, formed by connexins, allow the direct exchange of ions and small molecules between cells in contact thus co‐ordinating physiological processes such as cell growth and differentiation.

• Significant changes have been observed in the pattern of vascular connexin expression during atherosclerotic plaque formation. Recent studies on connexin‐deficient mouse models for atherosclerosis have revealed key roles for these proteins in the development of the disease.

Gap junctions in the vascular wall

Atherosclerotic lesion formation involves complex patterns of interaction between inflammatory cells and normal cellular elements of the arterial wall, in which cytokines, chemokines, and growth factors are known to play a critical role. Aside from these paracrine mechanisms, another form of intercellular communication involves gap junctions.

Gap junctions are composed of intercellular channels that allow the direct exchange of ions, small metabolites, and other second messenger molecules between adjacent cells, this way synchronizing responses in multicellular organisms 14. This type of intercellular communication permits rapidly co‐ordinated activities, such as contraction of cardiac and smooth muscle, and transmission of neuronal signals at electrical synapses. In addition, gap junctional intercellular communication (GJIC) plays a role in slower physiological processes, such as cell growth and development. Molecular cloning studies have demonstrated that gap junction channels are formed by members of a family of related proteins called connexins (Cx) in vertebrates. There are more than 20 different Cx types in the human and mouse genomes 15. The commonly used nomenclature distinguishes Cx by their molecular mass deduced from their respective sequences. Full gap junction channels are complex in that they span two plasma membranes. One gap junction channel results from the docking of two hemichannels or connexons. Each connexon is assembled from six Cx proteins. Recent findings suggest that hemichannels themselves can also open under both physiological and pathological conditions, and this activity may participate in a number of cellular processes 16, 17.

As a consequence of the molecular diversity of the Cx family, connexons and gap junction channels may be assembled from different Cx. Each type of connexin‐made channel has unique inherent gating properties or permeabilities to various molecules and ions. Historically, gap junctions have been described as relatively non‐selective and permeable to a wide range of molecules up to ∼1 kD. However, experiments carefully examining the electrophysiological properties of gap junction channels between cells expressing different Cx have revealed that channels made of different Cx exhibit unique conductances and permeabilities to ions and fluorescent dyes 18, 19. Also endogenous metabolites, such as adenosine diphosphate (ADP), adenosine triphosphate (ATP), glutathione, and glutamate, have been shown to permeate gap junction channels between living cells 20. Moreover, the Cx composition of gap junction channels appeared to determine selectivity among second messengers 21.

Four connexins, Cx37, Cx40, Cx43, and Cx45, have been described in the vascular wall, a tissue that contains EC‐EC, SMC‐SMC and EC‐SMC gap junctions 22–24. Although Cx expression profiles have not yet been completely described for all parts of the vascular tree, it has become apparent that it is not uniform in all blood vessels 22. In addition, differences in Cx expression have been reported in some vessels, like coronary arteries, when comparing different species 25. In general, ECs express Cx37 and Cx40, whereas SMCs express Cx43 and Cx45. Cx43 has also been found in a subset of ECs near branch points of arteries and in capillaries 26, 27. Additional expression of Cx37 or Cx40 in SMCs has been observed in small elastic or resistance arteries or during development 28, 29.

The recent availability of connexin‐specific tools, such as dominant negative connexins or connexin‐mimetic blocking peptides, has boosted our knowledge about the physiological function of vascular gap junctions. Thus, GJIC plays a significant role in the conduction of vasomotor responses along arterioles 23, 24, 30. In addition, ECs are induced to migrate during the process of new capillary sprout formation and during repair of the endothelial lining after injury in large vessels, a process that seems dependent on temporary switches in Cx expression 31. More recently, Cx43, Cx40, Cx37, and Cx45 gene‐targeted mice have been created, each displaying a different vascular phenotype. From these knock‐out mice, it appeared that Cx40 is required for normal transmission of endothelium‐dependent vasodilator responses 32. In addition, Cx40‐deficient mice appeared hypertensive, pointing to a possible role of this Cx in the regulation of blood pressure 32, 33. Interestingly, EC‐specific deletion of Cx43 seems to cause hypotension in mice 34. However, this observation remains to be confirmed as similar mice that were developed by another laboratory do not display a vascular phenotype 27. The complete deletion of Cx45 causes abnormalities in vascular development and mouse embryos die early between days 9.5 and 10.5 35. Although the deletion of Cx37 leads to female infertility because the mice do not ovulate, these animals do not show an obvious vascular phenotype 36, 37. In contrast to the single knock‐out animals, mice lacking both Cx37 and Cx40 display severe vascular abnormalities and die perinatally 38. An overview of the different Cx‐deficient mice and their vascular phenotype is illustrated in Table I.

Table I. Cx‐deficient mice and their associated vascular phenotype.

Cx‐deficient mice	Vascular phenotype	References
Cx37^−/−	normal	36, 37
Cx40^−/−	irregular arteriolar vasomotion, hypertension	32, 33
EC‐specific deletion of Cx40	normal	86
Cx40^−/−Cx37^−/−	lethal (P0–P1), severe vascular abnormalities	38
Cx43^−/−	lethal (P0), swelling and blockage of right ventricular outflow tract	72
EC‐specific deletion of Cx43	Normal/hypotension	27/34
Cx45^−/−	lethal (E9.5–10.5), defective vascular development	35

Gap junctions and atherosclerosis

Cx are dynamic proteins with half‐lives ranging from 1 to 5 hours, indicating that gap junction channels are fully exchanged several times per day 14. This may provide a mechanism to regulate direct cytoplasmic cross‐talk between cells under normal or pathological conditions. Over the last 15 years, there has been increasing support that Cx might participate in the process of atherogenesis. Thus, significant changes in the pattern of vascular Cx expression have been observed during atherosclerotic plaque formation. Furthermore, atherosclerotic risk factors acting on the endothelium, inflammation, and SMC activation/proliferation have been shown to affect Cx expression or GJIC. It suggests that Cx may play a role in the pathogenesis, but it does not establish causality between these proteins and atherosclerosis. More extensive research on the vascular Cx has been conducted in recent years, to further investigate this association. For this purpose, mouse models have been created by interbreeding Cx‐deficient mice and atherosclerosis‐prone mice (LDLR^−/− mice or apolipoprotein E‐deficient (ApoE^−/−) mice).

Altered connexin expression during atherogenesis

In human coronary atherosclerosis, Cx43 expression in intimal SMCs is increased at early stages of the disease but reduced in advanced atheroma 39. Hypercholesterolaemia‐induced atherosclerosis in the rabbit arterial wall resulted in Cx43 expression associated with macrophage foam cells and, comparable with the human situation, reduced levels of Cx43 between intimal SMCs were observed in advanced lesions 40. A similar temporal pattern of Cx43 expression in intimal SMCs and macrophage foam cells was observed during atherogenesis in LDLR^−/− mice fed a high‐fat diet for 0, 6, 10, or 14 weeks 41. Although Cx43 is mostly absent in aortic endothelium of healthy LDLR^−/− mice, this protein appeared in ECs at the shoulder region of advanced atherosclerotic plaques, a localization known to experience turbulent blood flow. However, the endothelium covering advanced atherosclerotic plaques no longer expressed Cx37 or Cx40. A similar observation has been reported in a study in C57BL/6 mice 42. Cx37 and Cx40 were dramatically reduced when these mice were fed a high‐cholesterol diet for several months. In addition, Cx37 expression has been reported in macrophages in early and late atheroma 41. This protein also appeared in medial SMCs beneath advanced atherosclerotic lesions 41. Of note, similar Cx37 expression patterns were observed in advanced atherosclerotic plaques in human carotid artery.

Initiation of atherosclerosis

Endothelium dysfunction, the initiating step of atherosclerosis, is frequently observed with hypercholesterolaemia and mostly in arterial regions experiencing disturbed flow patterns. As described above, gap junctions in arterial endothelium in situ are known to consist mainly of Cx40 and Cx37. An increasing number of reports reveal, however, that Cx43 is moderately expressed or absent in quiescent endothelia, but is induced under conditions associated with endothelium dysfunction. Indeed, Cx43 was found to be abundant in endothelia localized at the downstream edge of the ostia of branching vessels and at flow dividers, regions that experience turbulent shear stress 26. In addition, various in vitro studies have shown a positive correlation between Cx43 expression and mechanical load or disturbed flow patterns 43–45. In a recent study, we have systematically investigated the effects of pulsed unidirectional and oscillatory flows in combination with different levels of hydrostatic pressure and circumferential stretch on the expression of Cx43 using an in vitro perfusion system 46. We observed that endothelial Cx43 expression significantly increased in response to oscillatory shear stress within 4 hours, with the highest levels (up to 3‐fold) reached after 24 hours. In contrast, increasing hydrostatic pressure from 100 to 150 mmHg did not affect Cx43 expression levels. Interestingly, 4% cyclic circumferential stretch combined with high unidirectional shear induced only a 1.5‐fold increase in Cx43 expression, whereas the same cyclic circumferential stretch combined with oscillatory shear stress increased Cx43 by almost 7‐fold. Taken together, these studies suggest a causal relation between haemodynamic conditions and endothelial Cx43 expression, which may be significant to focal vulnerability to atherosclerosis. Interestingly, it has been demonstrated that gap junction assembly between cultured Novikoff hepatoma cells can be increased by cholesterol supplementation 47, LDL, and apolipoprotein B (apoB) treatments 48. Such enhancement of gap junction assembly seems dependent on stimulation of Cx trafficking from intercellular stores 49. In addition to its effects on assembly, increasing cellular cholesterol content affects the biophysical properties of Cx43 gap junction channels. Indeed, neonatal cardiomyocytes were protected against heptanol‐induced closure of Cx43 gap junction channels upon increased cellular cholesterol content 50. Whether cholesterol/LDL exerts similar effects on Cx43 in ECs remains to be investigated. In this context, it is worth mentioning that lipoprotein‐derived phospholipid oxidation products upregulate Cx43, downregulate Cx37, and do not affect Cx40 expression in the endothelium of murine carotid arteries in vivo and EC cultures 51.

Progression of the disease

Leukocyte recruitment in the early phases of atherosclerosis involves monocytes and T lymphocytes 11. There is evidence in the literature, although mainly circumstantial, for a role of GJIC in the inflammatory response. Thus, altered Cx expression and/or GJIC has been described in a number of inflammatory conditions in vivo, such as the ischaemic heart, and after injection of endotoxin or lipopolysaccharide (LPS) in liver or heart (see for reviews: 52, 53). More specifically, reduced GJIC has been observed with the pro‐inflammatory mediators LPS, tumour necrosis factor α (TNF‐α), interleukin 1α (IL‐1α) or Il‐1β in in vitro studies on ECs 54, 55 and SMCs 56. Moreover, rapid closure of myoendothelial gap junctions has been reported in a co‐culture of human ECs and SMCs that were exposed to TNF‐α 57. Most of these rapid changes in GJIC might be explained by actions of the pro‐inflammatory mediators on gating and permeability properties of the gap junction channels involved.

The possibility that GJIC may play a role in the modulation of the inflammatory response has been subject of a number of in vitro studies in recent years. Thus, various laboratories have tested the possibility that gap junctions are involved in leukocyte transmigration in vitro with somewhat conflicting outcomes. In general, transmigration of leukocytes across endothelial or epithelial monolayers was assessed in these studies in the presence of connexin‐mimetic peptides or gap junction channel‐blockers. The results revealed that transmigration of neutrophils was increased or not affected under those conditions 53, 58, 59, transmigration of monocytes was decreased 60, but that these treatments had only modest effects on lymphocytes 61. The differences in the in vitro cell systems used are most likely responsible for the variability in results between laboratories. Beyond doubt, in vivo studies on Cx‐deficient mice will provide a better defined and more physiological setting to address whether Cx are involved in the inflammatory response.

As described before, the expression profile of Cx37 is altered in mouse and human atherosclerotic plaques. Notably, Cx37 is no longer found in the endothelium covering advanced plaques but detected in macrophages recruited to the lesions 41, 62. Subsequently, we investigated the effect of Cx37 deletion (Cx37^−/−) on atherogenesis in ApoE^−/− mice, an animal model for the disease 63. We induced atherosclerosis in Cx37^+/+ApoE^−/− and Cx37^−/−ApoE^−/− mice by feeding them a high‐cholesterol diet and demonstrated that atherosclerotic lesion development was accelerated in mice lacking Cx37, both in the descending aortas and aortic sinuses. Knowing that recruitment of leukocytes is a prerequisite for atherosclerotic lesion formation, we tested whether the increased atherosclerosis in Cx37^−/−ApoE^−/− mice was caused by accelerated monocyte migration across the endothelium. Thus, criss‐cross in vivo adoptive transfer assays revealed that the presence of Cx37 in monocyte/macrophages but not in the endothelium contributed to the recruitment of these leukocytes to atherosclerotic lesions. To explore whether Cx37 could control the earliest step of the leukocyte recruitment paradigm, we then compared adhesion of Cx37^+/+ApoE^−/− and Cx37^−/−ApoE^−/− monocyte/macrophages to an activated mouse EC monolayer. Indeed, an increased number of Cx37‐deficient monocytes or macrophages adhered. Because individual monocytes and macrophages cannot form gap junctions, we speculated that cell adhesion may be regulated by the activity of Cx37 hemichannels. We further demonstrated in primary monocytes, macrophages, and a macrophage cell line transfected with Cx37 cDNA that Cx37 hemichannel activity inhibited leukocyte adhesion. This anti‐adhesive effect was mediated by ATP release into the extra‐cellular space. Thus, Cx37 hemichannels control the initiation of atherosclerotic plaque development by modulating the autocrine ATP‐dependent regulation of monocyte adhesion.

A genetic polymorphism in the human Cx37 gene (C1019T) has been reported as a potential prognostic marker for atherosclerotic plaque development and myocardial infarction 64–68. This specific polymorphism results in a non‐conservative amino acid change (serine to proline) in the intracellular C‐terminus of the Cx37 protein. In this context, the observation that expression of Cx37‐S or Cx37‐P by transfection of a macrophage cell line revealed differential adhesiveness to substrates is of particular importance 63. These observations suggests that the C1019T polymorphism of Cx37 generates hemichannels with different properties in terms of ATP transport and provide a potential mechanism by which Cx37 may function as a prognostic marker for atherosclerosis.

Plaque rupture

The continued inflammatory response and accumulation of lipids work together with other events to promote atherosclerotic plaque growth and eventually rupture 4, 5. During the growing phase, medial SMCs migrate to the top of the intima where they multiply and produce components of the ECM. The SMCs and matrix molecules coalesce to form a strong fibrous cap that covers the original atherosclerotic site. Migration and proliferation of SMCs as well as synthesis of ECM by these cells commonly involve phenotypic transformation of SMCs from the differentiated contractile state to the activated synthetic state. The major gap junction protein expressed in vascular SMCs is Cx43. Interestingly, the synthetic SMC phenotype is known to express considerably higher levels of this protein in vitro as compared to contractile SMCs 69, 70, suggesting that the level of Cx43 expression critically depends on SMC phenotype. Furthermore, increased Cx43 expression between intimal SMC is observed in vivo in early atherosclerotic lesions in human and mice 39, 41. These increased Cx43 expression levels between intimal SMCs declined with progression of the lesion 39. Besides the changes in Cx43 expression in SMCs, the expression of this protein is increased in other cell types during atherogenesis. Indeed, relatively high expression levels of Cx43 are observed in macrophage foam cells 41, 71 and in ECs covering the shoulder region of atherosclerotic lesions 41. To elucidate whether these changes in Cx43 expression participate in the development of atherosclerotic plaques in vivo, we have intercrossed atherosclerosis‐susceptible LDLR^−/− mice with mice heterozygous for a Cx43 null mutation (Cx43^+/− mice). Indeed, the Cx43 knock‐out mutation (Cx43^−/−) in mice is lethal because of cardiac malformations 72, and the amount of Cx43 is reduced by half in Cx43^+/− mice 73. Male LDLR^−/− mice, 10 weeks old, with normal (Cx43^+/+LDLR^−/−) or reduced (Cx43^+/−LDLR^−/) Cx43 expression were fed a cholesterol‐rich diet for 14 weeks. The progression of atherosclerosis was reduced by about 50% both at the level of the aortic roots and in the thoraco‐abdominal aortas of Cx43^+/−LDLR^−/− mice in comparison to control littermates 74. Of note, both groups of mice showed similar increases in serum cholesterol and triglycerides as well as body weight. Interestingly, our results also revealed that the composition of atherosclerotic plaques in Cx43^+/−LDLR^−/− mice was strikingly different from control mice. Lesions of Cx43^+/−LDLR^−/− mice had smaller lipid‐cores and fewer macrophages, whereas leukocyte counts in peripheral blood were similar for both groups. In addition, the lesions had thicker fibrous caps containing more SMCs and interstitial collagen. The content of SMCs versus macrophages, the extent of collagen within the lesion, and the size of the lipid core are related to the vulnerability of human atherosclerotic lesions to rupture, an event which might lead to acute myocardial infarction. Thus, targeting Cx43 may favour potential plaque stabilizing processes rather than affecting plaque size alone. The mechanism by which reduced Cx43 ultimately leads to this dual benefit remains to be identified 75.

Possible role for connexins in treatment of atherosclerotic disease

Statins

Inhibitors of 3‐hydroxy‐3‐methylglutaryl‐CoA (HMG‐CoA) reductase (called statins) lower plasma cholesterol in human and reduce atherosclerosis‐related morbidity and mortality 76, 77. Because several in vitro studies have identified numerous pleiotropic effects of statins on vascular cells that could modulate atherogenesis and plaque rupture via mechanisms other than lowering cholesterol 78, we evaluated the effects of statins on gap junctions in primary human vascular cells 75. We observed that various types of statins dose‐dependently inhibited Cx43 expression. Of note, the lower concentrations of simvastatin (0.4–0.08 µM) used in our study are within the range of expected tissue levels derived from prescribed pharmacological dosages. The effect of the statins on Cx43 expression was abolished in the presence of L‐mevalonate, indicating that inhibition of HMG‐CoA reductase is responsible for the reduction in Cx43. The statin‐induced reduction in Cx43 expression was associated with a marked reduction in GJIC. The clinical benefits of lipid lowering with statins are often attributed to changes in atherosclerotic plaque composition leading to plaque stability 79. Interestingly, statins do not reduce plasma lipid levels in mice due to compensatory upregulation of HMG‐CoA reductase. Consequently, possible beneficial effects of statin treatment on the composition of mouse atherosclerotic plaques can be interpreted without this confounding variable. We observed that statin treatment reduced Cx43 expression in atherosclerotic lesions in LDLR^−/− mice. Interestingly, statin‐treated mice displayed beneficial changes in plaque morphology that were similar to the ones observed in Cx43^+/−LDLR^−/− mice 74. In another study, it has been shown that Cx37 and Cx40 expression was decreased in mouse aorta during long‐term hyperlipidaemia, a major risk factor of atherosclerosis. Interestingly, simvastatin has been shown to reverse the hyperlipidaemia‐induced decrease of endothelial Cx37 in mice, thus possibly yielding an additional connexin‐mediated benefice to this treatment 42.

Percutaneous coronary intervention

A consequence of coronary atherosclerosis is ischemic heart disease. In addition to coronary bypass surgery, percutaneous coronary interventions are commonly used to treat critically narrowed atherosclerotic coronary arteries. However, their long‐term efficacy is hampered by restenosis or re‐narrowing of the arteries at the site of the intervention 80, 81. In fact, stretching a diseased artery often induces an exaggerated response to injury that involves recruitment and infiltration of leukocytes to the damaged site as well as a surge in cytokines and growth factors. Subsequently, vascular SMCs undergo a phenotypic modulation from a contractile to a synthetic phenotype, proliferate within the media, and migrate towards the intima. These changes are associated with modulation of the ECM. The sum of all these events leads to neointima formation.

As described above, increased Cx43 expression is intimately linked to intimal hyperplasia in the context of chronic injury that promotes atherosclerosis. The correlation between Cx43 and neointima formation in the setting of acute vascular injury has first been investigated by Severs and collaborators 82. Balloon injury in the rat carotid artery induced the expression of Cx43 in the media and intima in parallel with SMC activation and phenotype. Enhanced Cx43 expression in vascular SMCs as well as in neointimal macrophages was reported later in restenotic lesions of injured vessels in other species including mice 40, 83–85. To investigate whether reducing Cx43 expression would affect neointima formation in vivo, we subjected hypercholesterolaemic Cx43^+/−LDLR^−/− mice and Cx43^+/+LDLR^−/− control littermates to carotid balloon distension injury, which induced marked endothelial denudation and activation of medial SMCs 85. We observed restricted neointima formation after balloon injury in Cx43^+/−LDLR^−/− mice as compared to their Cx43^+/+LDLR^−/− littermates. This phenotype was associated with decreased macrophage infiltration and reduced SMC migration and proliferation, as well as accelerated endothelial repair in Cx43^+/−LDLR^−/− mice. Consequently, decreasing Cx43 expression may offer a novel therapeutic strategy for reducing restenosis after percutaneous coronary interventions.

Concluding remarks

In this review, we have provided an overview about the current knowledge on pathophysiology of vascular Cx and possible therapeutical implications. It is now well recognized from studies on small animals that remodelling of vascular gap junctions does occur during atherogenesis and in the restenotic process following balloon angioplasty. However, these studies have only been confirmed for certain stages of the disease process in human. Studies on knock‐out mouse models have definitely revealed the importance of Cx in vascular disease, but the precise role of these proteins remains often puzzling as Cx are often expressed in multiple cell types. Further investigations on mice in which Cx are deleted or over‐expressed in a temporal or spatial fashion will provide more detailed insight in their role in disease. In addition, better‐defined primary cell culture models will have to be developed to reproduce more realistically crucial steps in disease development. If the mechanisms behind alterations in Cx expression and the changes in cell behaviour resulting from such changes can be elucidated using down‐to‐earth models with reduced levels of complexity, this will certainly contribute to our further understanding in the pathophysiology of atherosclerosis and restenosis. Furthermore, such models might facilitate the development of Cx‐based intervention strategies in the future.

Acknowledgements

This work was supported by grants from the Swiss National Science Foundation (#PPOOA‐68883), Fondation Simone and Gustave PREVOT, Fondation Carlos et Elsie DE REUTER, and the Swiss University Conference Program ‘Heart Remodeling in Health and Disease’.