doi: 10.1161/CIRCRESAHA.107.165670
© 2007 American Heart Association, Inc.
Editorials |
Which Connexins Connect?
From the Department of Diagnostic Radiology, Wales Heart Research Institute, School of Medicine, Cardiff University, UK.
Correspondence to Prof Tudor Griffith, Department of Diagnostic Radiology, Wales Heart Research Institute, School of Medicine, Cardiff University, Health Park, Cardiff, CF14 4XN UK. E-mail griffith{at}cardiff.ac.uk
See related article, pages 1292–1299
Key Words: gap junction • electrotonic • EDHF • endothelium • smooth muscle
Electrical signaling via gap junctions modulates vascular function via 2 physiologically interrelated mechanisms, namely dilations/constrictions that propagate longitudinally along the vessel wall,1–4 and the EDHF phenomenon, in which relaxation is facilitated by the transmission of endothelial hyperpolarization to smooth muscle via a radial myoendothelial pathway.5 Gap junctions are composed from 2 hemichannels, each constructed from 6 connexin (Cx) protein subunits, whose alignment across the extracellular space permits intercellular diffusion of ions and small molecules <1 kDa in size and confers electrical continuity. Coupling is enhanced by the clustering of up to several hundred individual gap junction channels to form plaques that can be visualized by immunostaining at points of cell–cell contact. Vascular cells variably express 4 connexin subtypes (Cxs 37, 40, 43, and 45) with protein expression generally being more abundant in the endothelium than in smooth muscle cells. Most commonly, endothelial plaques contain Cx37 and Cx40 and medial plaques contain Cx43. This distribution is seen, for example, in the rabbit iliac artery in which synthetic peptides selectively targeted against Cx37 and Cx40 attenuate endothelium-dependent subintimal smooth muscle hyperpolarization, whereas peptides targeted against Cx43 attenuate the spread of medial hyperpolarization.6 In muscular arteries electrical signaling pathways thus appear to correlate closely with immunostaining findings.
In this issue of Circulation Research Wolfle and colleagues1 have analyzed longitudinal conduction in mice with replacement of Cx40 by "knock in" of Cx45 (denoted as Cx40KI45 mice). Responses were initiated by localized application of the endothelium-dependent dilators acetylcholine (ACh) and bradykinin (BK) to cremasteric arterioles in wild-type (WT), Cx40 knockout (Cx40KO), and Cx40KI45 animals, and diameter changes were monitored dynamically by intravital microscopy. Responses were found to propagate for more than 1 mm within 1 sec in WT mice without significant decay, whereas remote dilation was reduced by 50% in Cx40KO and Cx40KI45 animals. Because focal endothelial disruption abolishes conducted dilations to ACh in mouse cremasteric arterioles,2 and immunostaining confirmed endothelial KO of Cx40 and KI of Cx45, these data suggest that Cx45 is unable to substitute functionally for Cx40 in the endothelium of these vessels. As the authors note, Cx40 channels exhibit a large conductance of 200 pS, whereas Cx45 channels possess much lower conductivity in the range 20 to 40 pS7 and may therefore be unable to rescue intercellular signaling after loss of Cx40. It seems likely that the residual conducted dilations evoked by ACh and BK in Cx40KO and Cx40KI45 mice involve gap junctions constructed from Cx37 as this connexin is also expressed in the endothelium of cremasteric arterioles2 whereas Cx43 is not detectable.1 Indeed, Cx37 channels exhibit a high conductance of 300 pS,7 and Cx37 and Cx40 both contribute to endothelial communication in mouse aortic endothelium with dye transfer being reduced by 60% in Cx40KO mice and by 25% in Cx37KO mice.8 Surprisingly, expression of these connexins may be mutually interdependent because there is evidence that deletion of Cx40 causes a 
20-fold reduction in aortic endothelial Cx37 levels, and deletion of Cx37 a 4-fold reduction in Cx40, probably via posttranscriptional mechanisms.8 Whether Cx37 expression is similarly downregulated in arterioles from Cx40KO and Cx40KI45 mice remains to be investigated.
There is emerging evidence that the mechanisms that underpin local and conducted dilations are distinct (Figure) because agents that block hyperpolarizing KCa and Kir channels attenuate the initiating response to ACh when applied locally, but are far less effective (or exert or no effect) against conducted dilations when applied remotely.9–12 More controversially, in mouse cremasteric arterioles the local endothelial hyperpolarizing response to ACh involves activation of endothelial SKCa channels, whereas it has been proposed that local release of a diffusible EDHF activates hyperpolarizing smooth muscle BKCachannels directly in this vessel.9 By contrast, in hamster feed arteries endothelial SKCa and IKCa channels underpin the initiating local response to ACh, whereas BKCa channel blockade is without effect.11 There is nevertheless a growing consensus that conducted dilation involves the spread of hyperpolarization from the endothelium to subjacent smooth muscle cells; localized endothelial disruption often blocks the conduction of ACh-evoked responses,2,12,13 although in some arterioles (eg, hamster cheek pouch) a parallel smooth muscle pathway may also contribute to signal propagation.3 Whether the reduction in conducted ACh- and BK-induced dilations observed by Wolfle and colleagues in cremasteric arterioles in Cx40KO mice in part reflects loss of Cx40 from myoendothelial gap junctions (as well as the endothelium) remains to be determined, because the composition of myoendothelial channels in these vessels is unknown. Cx37 appears to be specifically excluded from myoendothelial gap junctions in mouse endothelial/vascular smooth muscle cocultures, despite expression in both cell types, leading to the conclusion that channels constructed from Cx40 or Cx43 in the endothelium and Cx43 in smooth muscle cells underpin coupling.14 By contrast, Cx37 and Cx40, but not Cx43, have been located in myoendothelial channels in rat mesenteric arteries, even though all 3 connexins are present in homocellular endothelial gap junction plaques.15
|
Wolfle and colleagues also report that local dilations induced by adenosine and constrictions induced by K+ conducted to an equivalent extent in arterioles from WT, Cx40KO, and Cx40KI45 mice, but that the amplitude of these responses decayed over distance, thus contrasting with the nondecremental effects of ACh and BK.1 Because the conduction mechanism was therefore Cx40-independent, they postulate that adenosine and K+ evoke changes in smooth muscle membrane potential that are conducted via the media rather than the endothelium. This conclusion gains support from observations that focal smooth muscle disruption attenuates conducted K+-evoked constriction without affecting conducted dilations in hamster cheek pouch arterioles.3 It is conceivable that plaques that are too small to be visualized by immunostaining contribute to smooth muscle signaling in mouse cremasteric arterioles as Cx37, Cx40, or Cx43 were not detected in the media of these vessels and expression of Cx45 was very limited.1
Many of the experimental features of conducted responses can be reproduced by a resistive/capacitative model of coupled endothelial and smooth muscle cells in the vascular wall.16 Simulations illustrate how localized electrical events conduct robustly along the endothelium while eliciting a parallel response in underlying smooth muscle because loss of charge is minimized by low endothelial coupling resistance (3 M
) and high myoendothelial coupling resistance (1800 M). Small movements of charge from the endothelium are able to effect substantial changes in the membrane potential of smooth muscle cells because the input resistance of the media is high. By contrast, localized electrical events in smooth muscle conduct poorly to neighboring cells because their high coupling resistance (90 M) promotes signal dissipation, and when charge does conduct to the endothelium, its effects on membrane potential are minimized by the low endothelial input resistance unless a large number of smooth muscle cells are stimulated simultaneously. The model also predicts that endothelial disruption will attenuate conducted responses far more effectively than disruption of the medial layer.
Additional complexity may, however, result from a wave of elevated endothelial [Ca2+] that accompanies conducted dilation and can propagate at a velocity much slower than electrotonic signaling for distances more than 1 mm from the initiating site.11 In hamster cheek pouch and cremasteric arterioles this phenomenon may promote activation of eNOS with a wave of secondary NO release prolonging the duration of ACh- and BK-evoked conducted dilations without affecting their amplitude.10,11 Wolfle and colleagues found that both local and remote dilator responses evoked by BK in mouse cremasteric arterioles were attenuated by inhibition of eNOS.1 Surprisingly, however, responses evoked by ACh were unaffected,1 consistent with the finding that conducted responses to ACh are similar in WT and eNOS–/– mice.4 This agonist-specific difference remains to be fully explained. It should be noted, however, that nondecremental conducted responses do not have an obligatory requirement for a Ca2+ wave that spreads along the endothelium. In rat mesenteric arteries, for example, ACh stimulates a localized rise in endothelial [Ca2+], but no increase is observed at distances >0.5 mm upstream of the stimulating pipette whereas dilation is conducted without decay for 1 to 2 mm.13 Furthermore, in such vessels localized activation of smooth muscle KATP channels by levcromakalim evokes a conducted dilation and hyperpolarization, which propagates via an endothelial rather than smooth muscle pathway, but is not accompanied by parallel elevations in endothelial [Ca2+]i.13 The ubiquitous role of electrotonic mechanisms is also emphasized by the finding that local ACh-induced endothelial hyperpolarizations can be transformed to depolarizations by KCa channel blockade in rat mesenteric arteries, and these may also propagate via the endothelium with minimal decay.12
There is evidence that NO may also attenuate conducted responses in mouse cremasteric arterioles because the length constant for the decay of the constrictor response initiated by focal stimulation with K+ increases by 25% after eNOS inhibition.4 Furthermore, the characteristics of the conducted response to K+ in eNOS–/– mice closely match those in WT mice after eNOS inhibition and, in addition, exogenous NO depresses propagation.4 Because the effects of NO on K+-induced constriction in WT mice are comparable to those in Cx40KO mice, Cx40 does not appear to be a specific target for NO.4 However, there is evidence from a mouse model of sepsis in which there is enhanced NO production that the connexin subtype targeted by NO might be Cx37.17 Conduction of K+-induced constriction in cremasteric arterioles is depressed in these septic WT mice, partially restored by local eNOS inhibition, and similar to that in nonseptic Cx37KO mice.17
Cx40KO mice are severely hypertensive, and Wolfle and colleagues demonstrated elevations in blood pressure of ca. 45 mm Hg using a telemetric technique.1 This compared with a more modest elevation of 20 mm Hg in Cx40KICx45 mice, suggesting that Cx45 can partially restore normal blood pressure control in Cx40-deficient animals. Initially, hypertension in Cx40KO mice was attributed to the localized microcirculatory vasospasm observed in such animals, but it has emerged that impaired pressure control of renin synthesis may also be a key factor.18,19 Cx40 is the dominant connexin expressed by renin-producing cells and in adult Cx40KO mice such cells are displaced from their normal location in the afferent arteriole to the extra/periglomerular interstitium, indicating that gap junctional communication is essential for the correct juxtaglomerular positioning of renin-producing cells and signaling with the macula densa.19 Whether this abnormality is corrected in Cx40KICx45 mice remains to be established, although as the authors note, Cx45 is able to replace Cx40 in the atrioventricular node and right atrium of the heart in Cx40-deficient mice, whereas Cx40 function cannot be restored in the left atrium.
Although the complexities of gap junctional signaling in the vascular system are beginning to unravel, many important issues remain and will undoubtedly be illuminated by KO/KI approaches such as those adopted by Wolfle and colleagues. For example, the role of direct cell–cell communication in vascular development is poorly understood. Mice with dual ablation of Cx37 and Cx40 initially exhibit normal vasculogenesis, but die perinatally with pronounced vascular abnormalities including vessel dilatation, hemangioma-like defects, and localized hemorrhages, and in Cx45KO embryos vasculogenesis is normal, but transformation into mature vessels is interrupted by apoptosis leading to death in utero.7,8
 |
Acknowledgments |
|---|
Source of Funding
Work performed in the author’s laboratory is funded by the British Heart Foundation.
Disclosures
None.
|
Footnotes |
|---|
The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association.
|
References |
|---|
 Top References |
|---|
1. Wolfle SE, Schmidt VJ, Hoepfl B, Gebert A, Alcolea S, Gros D, de Wit C. Connexin45 cannot replace the function of connexin40 in conducting endothelium-dependent dilations along arterioles. Circ Res. 2007; 101: 1292–1299.
2. Looft-Wilson RC, Payne GW, Segal SS. Connexin expression and conducted vasodilation along arteriolar endothelium in mouse skeletal muscle. J Appl Physiol. 2004; 97: 1152–1158.
3. Budel S, Bartlett IS, Segal SS. Homocellular conduction along endothelium and smooth muscle of arterioles in hamster cheek pouch: unmasking an NO wave. Circ Res. 2003; 93: 61–68.
4. Rodenwaldt B, Pohl U, de Wit C. Endogenous and exogenous NO attenuates conduction of vasoconstrictions along arterioles in the microcirculation. Am J Physiol. 2007; 292: H2341–H2348.
5. Griffith TM. Endothelium-dependent smooth muscle hyperpolarization: do gap junctions provide a unifying hypothesis? Br J Pharmacol. 2004; 141: 881–903.[CrossRef][Medline] [Order article via Infotrieve]
6. Chaytor AT, Bakker LM, Edwards DH, Griffith TM. Connexin-mimetic peptides dissociate electrotonic EDHF-type signalling via myoendothelial and smooth muscle gap junctions in the rabbit iliac artery. Br J Pharmacol. 2005; 144: 108–114.[CrossRef][Medline] [Order article via Infotrieve]
7. Sohl G, Willecke K. Gap junctions and the connexin protein family. Cardiovasc Res. 2004; 62: 228–232.[CrossRef][Medline] [Order article via Infotrieve]
8. Simon AM, McWhorter AR. Role of connexin37 and connexin40 in vascular development. Cell Commun Adhes. 2003; 10: 379–385.[CrossRef][Medline] [Order article via Infotrieve]
9. Siegl D, Koeppen M, Wolfle SE, Pohl U, de Wit C. Myoendothelial coupling is not prominent in arterioles within the mouse cremaster microcirculation in vivo. Circ Res. 2005 14; 97: 781–788.[CrossRef]
10. Hoepfl B, Rodenwaldt B, Pohl U, De Wit C. EDHF, but not NO or prostaglandins, is critical to evoke a conducted dilation upon ACh in hamster arterioles. Am J Physiol. 2002; 283: H996–H1004.
11. Domeier TL, Segal SS. Electromechanical and pharmacomechanical signalling pathways for conducted vasodilatation along endothelium of hamster feed arteries. J Physiol. 2007; 579: 175–186.
12. Goto K, Rummery NM, Grayson TH, Hill CE. Attenuation of conducted vasodilatation in rat mesenteric arteries during hypertension: role of inwardly rectifying potassium channels. J Physiol. 2004; 561: 215–231.
13. Takano H, Dora KA, Spitaler MM, Garland CJ. Spreading dilatation in rat mesenteric arteries associated with calcium-independent endothelial cell hyperpolarization. J Physiol. 2004; 556: 887–903.
14. Isakson BE, Duling BR. Heterocellular contact at the myoendothelial junction influences gap junction organization. Circ Res. 2005; 97: 44–51.
15. Sandow SL, Neylon CB, Chen MX, Garland CJ. Spatial separation of endothelial small- and intermediate-conductance calcium-activated potassium channels (KCa) and connexins: possible relationship to vasodilator function? J Anat. 2006; 209: 689–698.[CrossRef][Medline] [Order article via Infotrieve]
16. Diep HK, Vigmond EJ, Segal SS, Welsh DG. Defining electrical communication in skeletal muscle resistance arteries: a computational approach. J Physiol. 2005; 568: 267–281.
17. McKinnon RL, Lidington D, Bolon M, Ouellette Y, Kidder GM, Tyml K. Reduced arteriolar conducted vasoconstriction in septic mouse cremaster muscle is mediated by nNOS-derived NO. Cardiovasc Res. 2006; 69: 236–244.
18. Krattinger N, Capponi A, Mazzolai L, Aubert JF, Caille D, Nicod P, Waeber G, Meda P, Haefliger JA. Connexin40 regulates renin production and blood pressure. Kidney Int. 2007; 72: 814–822.[CrossRef][Medline] [Order article via Infotrieve]
19. Kurtz L, Schweda F, de Wit C, Kriz W, Witzgall R, Warth R, Sauter A, Kurtz A, Wagner C. Lack of connexin 40 causes displacement of renin-producing cells from afferent arterioles to the extraglomerular mesangium. J Am Soc Nephrol. 2007; 18: 1103–1111.
Related Article:
Circ. Res. 2007 101: 1292-1299.
This article has been cited by other articles:
 |
|
 |
|
  D. A. Goodenough and D. L. Paul Gap Junctions Cold Spring Harb Perspect Biol, July 1, 2009; 1(1): a002576 - a002576. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||