Connexin37 Is the Major Connexin Expressed in the Media of Caudal Artery
Objective— To determine the connexins (Cxs) involved in intercellular coupling within vascular muscle, the present study has quantified mRNA and protein expression for Cx37, Cx40, Cx43, and Cx45 in the caudal artery (CA) and thoracic aorta (ThA) of the rat.
Methods and Results— Real-time polymerase chain reaction and immunohistochemistry identified Cx37 as the most abundantly expressed Cx in the CA, with fine punctate staining observed in the media. Conversely, mRNA for Cx43 was 40-fold greater in the ThA than in the CA, with punctate staining in the endothelium and media of the ThA but confined to the endothelium in the CA. Western blotting confirmed the differences in the relative amounts of Cx43 between the 2 vessels. For both arteries, Cx45 was expressed to a lesser degree in the media but not in the endothelium, whereas Cx40 was found only in the endothelium. Cx37, Cx40, and Cx43 were expressed in the endothelium of both vessels, although the density of Cx40 plaques was significantly greater in the CA.
Conclusions— The demonstration of Cx37 as the dominant Cx in the media of the CA highlights the potential heterogeneity in Cx involvement in vascular smooth muscle.
Gap junctions are intercellular channels composed of membrane proteins known as connexins (Cxs), of which 4 (Cx37, Cx40, Cx43, and Cx45) have been identified in vascular tissue.1–3⇓⇓ In blood vessels, gap junctions are found connecting adjacent endothelial cells,4 connecting adjacent smooth muscle cells, 5 and connecting endothelial and smooth muscle cells.6 Although 3 of the 4 vascular Cxs have been shown to be expressed by endothelial cells of most vessels,7 the identity of the Cxs connecting adjacent smooth muscle cells in arteries is less clear.
In large elastic arteries, such as the aorta, Cx43 is thought to be the major gap junctional protein expressed in the smooth muscle.8–11⇓⇓⇓ More recently, some studies have shown expression of other Cxs in addition to Cx43 in the media of elastic arteries, although some of these differences can be attributed to heterogeneity among animal species.7 For example, Cx40 has been shown in the aorta of the cow and pig but not the rat,11 whereas Cx37 has been reported in the aorta and pulmonary artery of the rat in some studies10 but not in others.11
In muscular arteries, identification of the major Cx isoform has been more difficult. Cx43 has not been found in the media of a number of large muscular arteries, including the caudal, basilar, mesenteric, and coronary arteries,8,12–15⇓⇓⇓⇓ although it has been reported in pial and cremaster arterioles in rats and cheek pouch arterioles in hamsters.16 On the other hand, Cx40 appears to be a potential candidate in the media, having been identified in the coronary artery in a number of species,11 in the basilar artery,17 and in a number of different arterioles in rats.16,18,19⇓⇓ Cx45 is also receiving some attention20,21⇓ because of its identification in arterial smooth muscle in embryonic and adult mice.2,3,22⇓⇓ On the other hand, Cx37 is generally considered to be an endothelial Cx, although it has been described in the media of the larger coronary arteries of the rat11 and in the media of collateral vessels during coronary arteriogenesis in dogs.23
Considering the absence of a clearly identifiable Cx in the media of muscular arteries and the potential for some confusion due to nonspecificity of antibodies,24 we chose to compare Cx expression at mRNA and protein levels in a large muscular artery with expression in an elastic artery. By using real-time polymerase chain reaction (PCR), we have been able to quantify mRNA expression in the 2 vessels and relate this to protein expression by using immunohistochemistry and Western blotting with the use of different sources of antibodies raised against different epitopes for which results might be contentious. Together, the mRNA and protein data are consistent with the predominant expression of Cx43 in the media of the thoracic aorta (ThA) but with the expression of Cx37 in the media of the caudal artery (CA). Cx45 was expressed to a lesser extent in the media of both arteries. Our results also demonstrate variation in the relative appearance of the phosphorylated and nonphosphorylated forms of Cx43 in different tissues.
An expanded Methods section can be found in the online supplement at http://atvb.ahajournals.org.
Antibodies were raised in sheep against amino acids 266 to 281 of the C-terminus of rat Cx37 (Cx37/266), amino acids 254 to 270 of the C-terminus of rat Cx40 (Cx40/254), and amino acids 354 to 367 of human Cx45 (Cx45/354). The latter peptide was 93% homologous to the mouse sequence, whereas the rat sequence is unknown. The specificity of these antigenic sequences has been demonstrated previously by use of Western blotting and immunolabeling of transfected cells.9,25,26⇓⇓ Antibodies against rat Cx43 were commercially raised in rabbits (Zymed), and reactivity is independent of phosphorylation status. Commercial antibodies were also raised in rabbits against amino acids 318 to 333 of mouse Cx37, amino acids 340 to 358 of mouse Cx40, and amino acids 285 to 298 of mouse Cx45 (Alpha Diagnostic International).
Morphometric measurements of endothelial cell parameters, after staining with anti-Cx37, and Cx expression within the smooth muscle and endothelial cell layers were made by using the AIS imaging system (version 3.0, Imaging Research). Cx expression was quantified by measuring the average size of the smallest labeled structures in each preparation and defining these as individual gap junction plaques. The size of individual plaques was used in the grain-counting function to determine the number of connexin plaques per square micrometer of smooth muscle or endothelium. By use of the endothelial cell parameters, counts were adjusted to give the number of plaques per endothelial cell and the number of plaques per 100 μm of endothelial cell perimeter. In the case of the smooth muscle cells, expression was defined as the number of plaques per square micrometer of transverse or longitudinal medial area, because identification of individual cells in the multilayered media was not possible.
Expression of mRNA for Vascular Cxs
Quantification of mRNA expression of the vascular Cxs with the use of real-time PCR was achieved by constructing standard curves from the plasmid clones of the 4 Cxs (see online Figure IA and IB, available at http://atvb.ahajournals.org, for Cx40). Expression of mRNA in the arterial samples was expressed as Cx copy number and was normalized to the number of copies of 18S RNA. Data showed that in the ThA, expression of mRNA for Cx43 was significantly greater than expression of mRNA for Cx37, Cx40, and Cx45 (P<0.05, Figure 1). In the CA, expression of mRNA for Cx37 was significantly greater than expression of mRNA for Cx40, Cx43, and Cx45 (P<0.05, Figure 1). Expression of Cx37 mRNA was also significantly greater in the CA than in the ThA (P<0.05, Figure 1), whereas mRNA for Cx43 was significantly greater in the ThA than in the CA (P<0.05, Figure 1). Expression of mRNA for Cx37 in the CA was not significantly different from mRNA expression for Cx43 in the ThA. Expressions of mRNA for Cx40 and Cx45 were not significantly different between the 2 arteries (Figure 1).
Expression of Protein for Vascular Cxs
The distribution of Cx37 immunolabeling, typical for each of the 2 vessels, was the same regardless of whether the commercial Cx37 antibody or the affinity-purified Cx37/266 antibody was used. When tested on sections of rat lung, Cx37/266 showed abundant expression of Cx37 within the endothelium of blood vessels throughout the lung and within the smooth muscle layers of the bronchioles, consistent with studies of Nakamura et al.10 Staining was completely blocked by peptide (online Figure II, available at http://atvb.ahajournals.org). On the other hand, the commercial Cx40 antibody cross-reacted with Cx43, as evidenced by punctate staining in the media of the ThA and by the appearance of bands in Western blots of brain tissue, equivalent in size to those stained with antibodies against Cx43. Consequently, only results obtained with Cx40/254 are discussed. Staining with the 2 antibodies against Cx45 also revealed differences in sections of the ventricle but not in the 2 arteries. In the ventricle, the commercial anti-Cx45 stained cardiac myocytes throughout the myocardium, whereas Cx45/354 stained only the myocytes along the endocardial border, consistent with the restricted distribution found by Coppen et al,26 who used an antibody against the same sequence (see Figure 2E and 2F).
For each of the Cx antisera used in the present study, no staining at all was observed in the absence of the primary antibody or when the primary antibody was preincubated with the appropriate antigenic peptide.
Smooth Muscle Cells
Longitudinal and transverse sections of the rat ThA and CA were analyzed for expression of Cx37, Cx40, Cx43, and Cx45. No significant difference was found in the data for either orientation for any of the 4 Cxs in either artery. In the media of the CA, Cx37 was highly expressed (Figures 2A and 3⇓A), whereas Cx37 expression in the ThA was very sparse (P<0.05, Figure 3A and online Figure IIIA, available at http://atvb.ahajournals.org). Expression of Cx40 was absent from the media of both arteries (Figures 2B and 3⇓A and online Figure IIIB).
Cx43 was abundantly expressed in the media of the ThA (Figure 3A and online Figure IIIC) but was absent from the media of the CA (Figures 2C and 3⇑A). Cx45 could be detected sparsely in the media of CA and ThA (Figures 2D and 3⇑A and online Figure IIID).
Plaques of Cx37 and Cx45 in the media of the CA and Cx45 in the ThA were significantly smaller than Cx43 plaques in the ThA (P<0.001; for CA, 0.08±0.003 μm [Cx37] and 0.06±0.001 μm [Cx45]; for ThA, 0.11±0.01 μm [Cx43] and 0.06±0.004 μm [Cx45]).
En face views of the luminal surface showed punctate staining for Cx37, Cx40, and Cx43 along the periphery of endothelial cells in ThA and CA (online Figures IVA, IVB, and IVC, respectively [available at http://atvb.ahajournals. org], and Figure 4A, 4B, and 4C, respectively). Cx45 was not found in the endothelium of either artery (online Figure IVD and Figure 4D). The area, length, and width of endothelial cells did not differ between the CA and the ThA, although the perimeter of endothelial cells was significantly greater in the CA than in the ThA (P<0.05; online Figure V, available at http://atvb.ahajournals.org). Expression of Cx37 and Cx43 per square micrometer of luminal surface was similar between the ThA and CA, whereas expression of Cx40 was significantly less in the ThA than in the CA (P<0.05, Figure 3B). Because Cx expression in the endothelium was seen exclusively around the cell periphery (online Figure IV and Figure 4), the density of Cx plaques in the cell membrane was also calculated. The density of Cx40 plaques was significantly greater in the CA than in the ThA, whereas no such difference was seen for Cx37 and Cx43 (P<0.05). Plaque sizes for each Cx did not vary between endothelial cells in either artery. However, in the CA, Cx37 plaques between endothelial cells were significantly larger than plaques between smooth muscle cells (0.150±0.01 μm for endothelial cells and 0.075±0.003 μm for smooth muscle cells, P<0.0001), whereas in the ThA, plaques for Cx43 were greater in the endothelium than in the smooth muscle (0.143±0.01 μm for endothelial cells and 0.106±0.01 μm for smooth muscle cells, P<0.001).
Rabbit Cx43 antibodies revealed the presence of multiple isoforms of Cx43 in extracts of brain, heart, lung, and ThA (Figure 5), with the higher molecular weight isoforms appearing to predominate in the heart and ThA (Figure 5, −control peptide). All isoforms were blocked by preincubation of the Cx43 antibody with immunogenic peptide (Figure 5, +control peptide). Cx43 was detectable only in extracts of CA when 10 μg of protein was loaded on the gel (online Figure VI, available at http://atvb.ahajournals.org) but not when 5 μg was used (Figure 5). Quantification of Cx43 by phosphoimaging revealed that compared with CA, ThA contained at least 9 times more Cx43 (data not shown). When samples of heart, ThA, and CA were treated with phosphatase, a shift in mobility toward the lower molecular weight form was found (online Figure VI), confirming that the high molecular weight forms were phosphorylated Cx43, as previously demonstrated.27,28⇓
The Cx40/254, Cx45/354, and commercial Cx45 antibodies all stained numerous bands in Western blots, although only a small number of bands disappeared when the antibody was preincubated with immunogenic peptide (online Figure VII, available at http://atvb.ahajournals.org). Thus, Cx40/254 antibody specifically recognized a band of 40 kDa from lung, CA, and ThA but not liver (online Figure VIIA, −/+peptide). In the lung, however, a band at 45 kDa also appeared to be reduced. The Cx45/354 antibody revealed the presence of a specific 45-kDa band in all tissues tested, although it was very weak in the arteries. A higher molecular weight band, which was blocked by peptide, was also seen in the brain (online Figure VIIB, −/+peptide). On the other hand, the commercial Cx45 antibody labeled a 45-kDa band and also a lower molecular weight species in brain, heart, and CA (online Figure VIIC, asterisk). Although the Cx37/266 antibody and the commercial Cx37 antibody worked well in immunohistochemistry, numerous bands were stained in Western blots, but none disappeared after peptide incubation.
Significant differences have been found in the Cx makeup of the media, but not the endothelium, of an elastic and a muscular rat artery by using real-time PCR, immunohistochemistry, and Western blotting. In the CA, mRNA for Cx37 was the most prevalent of the 4 Cxs, and immunohistochemically demonstrable staining was found in the media of that artery, but it was not found to any great extent in the ThA. In the present study, which used high-resolution confocal microscopy, a single fluorescent spot was defined as a gap junctional plaque, as validated previously.29,30⇓ On the other hand, mRNA for Cx43 was the most prevalent Cx in the ThA and greatly exceeded the expression of Cx43 in the CA. In similar fashion, the specific activity of Cx43 protein on Western blots was significantly greater in the aorta than in the CA, in agreement with the extensive punctate staining in the aortic media and an absence of staining in the CA, as described previously.8,31⇓ In the endothelium, protein for Cx37, Cx40, and Cx43 was detected in both vessels, as was Cx45 in the media of the 2 vessels. These data strongly suggest that the predominant Cx connecting smooth muscle cells in the CA is Cx37, in contrast to the aorta, where the predominant Cx is confirmed to be Cx43, with Cx45 playing a minor role in both arteries.
Although the mRNA extracts of both arteries included smooth muscle cells and endothelial cells, the predominant cell type was the smooth muscle cell, inasmuch as our ultrastructural studies have demonstrated that there are ≥7 layers of smooth muscle cells surrounding the single layer of endothelial cells in the 2 vessels (S. Sandow, unpublished data, 2002). Thus, when protein expression for each of the Cx subtypes in the media of the 2 arteries was compared with the mRNA expression in the vessels, a good correlation was found for all 4 Cxs in both arteries. In contrast, in cultured smooth muscle cells from preglomerular arterioles of the rat, only Cx40 protein was detected in spite of mRNA expression for Cx37, Cx40, and Cx43.18 Although Cx40 protein was also found in sections of preglomerular arterioles in vivo, no mRNA analyses of this Cx or of other Cxs were performed.18 In preliminary experiments for the present study, attempts were made to rub off the endothelium to attribute mRNA for specific Cx isoforms to specific cellular layers. Unfortunately, these experiments were not entirely successful, inasmuch as real-time PCR showed that expression of mRNA for the endothelial cell marker von Willebrand factor could still be detected in these “endothelium-denuded” preparations.
Although Cx37 was extensively expressed in the media of the CA, it was effectively absent in the media of the ThA. These results are in agreement with the study of van Kempen and Jongsma,11 who failed to find any Cx37 labeling in the aortas of several species, but in contrast to the study of Nakamura et al,10 who described the presence of Cx37 in the smooth muscle of the rat aorta and pulmonary artery. However, the authors did not define which region of the aorta was used, and because expression of other Cxs has been shown to vary along the length of the aorta,8,20⇓ it may be difficult to directly compare the results. Expression of mRNA for Cx37 was 10-fold greater in the CA than in the ThA, in line with the observed staining within the muscle layers. In the same species, Cx37 has also been identified in the media of large coronary arteries,11 perhaps suggesting a more widespread role for Cx37 in cell coupling in the media of large muscular arteries in the rat.
In the present study, Cx40 was not detected in the media of either the ThA or CAs. These results are in contrast to several previous studies that have described Cx40 in the media of blood vessels from several species, including preglomerular and pial arterioles of the rat,16,18⇓ hamster cremaster muscle arterioles,16 and coronary arteries from the cow, pig, and rat.11 Taken together, these results may suggest a greater role for Cx40 in smaller vessels, supporting the idea that heterogeneity exists in the expression of Cxs within different parts of the vascular tree.
In the ThA, Cx45 was relatively sparsely expressed in contrast to the dense expression of Cx43. A reciprocal relationship between these 2 Cxs was found throughout the aortic vessels to the iliac artery.20 In the CA, double labeling with antibodies against smooth muscle myosin showed Cx45 between smooth muscle cells, and with the use of real-time PCR, expression was similar to that in the ThA. The greater expression of Cx37 than of Cx45 in the CA suggests that Cx37 may share a similar inverse relation to Cx45 in the media of the CA. Alternatively, expression of the 2 Cxs may have implications for radial versus longitudinal coupling of smooth muscle cells. Recent dye-coupling studies in the CA from our laboratory (N. Bramich, unpublished data, 2001) have shown selective spread of Lucifer yellow dye between smooth muscle cells in the radial, but not the longitudinal, direction. However, in the present study, in the media of both arteries, protein expression of Cx37 and Cx45 did not differ between longitudinal and transverse orientations.
As previously demonstrated, protein for Cx37, Cx40, and Cx43 was expressed in the endothelium of the ThA and CA,15 whereas the expression of Cx45 was not detected in the endothelium of either artery.2,20⇓ Staining was essentially confined to the intercellular borders. The expression of Cx40 was significantly greater in the endothelium of the CA than in the ThA, whereas the expression of Cx37 and Cx43 was not significantly different between the 2 arteries. Morphological measurements demonstrated some differences in the shape of endothelial cells between the 2 arteries, resulting in an increase in cell perimeter in the CA relative to that in the ThA. In spite of this increase in cell perimeter, the density of Cx40 plaques was also significantly greater in the CA than in the ThA.
The validity of protein data ultimately depends on the specificity of the antibodies used. The commercial Cx43 antibody stained a number of molecular weight bands that were confirmed in the present study to represent unphosphorylated and phosphorylated Cx43.32 Using a similar antibody, Hossain et al27 found no differences in immunohistochemical staining of brain tissues in which different relative amounts of these phosphorylated forms existed. For each of Cx37, Cx40, and Cx45, 2 antibodies directed against well-separated epitopes were used. Both Cx37 antibodies yielded similar immunohistochemical results, specific for the 2 arteries. Furthermore, in the lung, there was extensive Cx37 staining of blood vessels and bronchioles as previously described,10 confirming specificity. A similar antibody to Cx37/266 has been previously characterized with the use of Western blotting of Cx-transfected cells.9,25⇓ The Cx40/254 antibody detected a 40-kDa protein in Western blots of tissue samples (present study), and a similar antibody demonstrated specificity in Cx-transfected cells.9 However, the commercial antibody to Cx40 appeared to cross-react with Cx43, and use was discontinued. The Cx45/354 antibody produced restricted labeling of the ventricular endocardium consistent with that previously demonstrated by Coppen et al26 for an antibody raised against the same epitope and fully characterized in transfected cells. However, the commercial Cx45 antibody cross-reacted with Cx43, in a manner similar to that described by Coppen et al,26 for a similarly directed commercial Cx45 antibody. Interestingly, the commercial Cx45 antibody did not stain the muscle of the aorta or the endothelium of either vessel, which are sites of extensive Cx43 expression. Coppen et al26 have elegantly identified the cross-reacting peptide as residues 283 to 286 of Cx43. We found that this site lies in an area of the carboxy terminus rich in predicted phosphorylation sites, immediately adjacent to a serine and containing a tyrosine. Thus, we suggest that tissue-specific phosphorylation of Cx43 occurs in the vessels but not in the heart, thus reducing accessibility of this antibody in the former but not in the latter.
Tissue-specific distribution of the unphosphorylated and phosphorylated forms of Cx43 has been previously reported.28 In the present study, the higher molecular weight phosphorylated forms predominated in the heart and ThA, whereas the lower molecular weight forms were more prevalent in the brain and CA. Hossain et al27 demonstrated that the predominance of the lower molecular weight forms in the brain was due to rapid dephosphorylation of Cx43. Because Cx43 in the CA is restricted to the endothelium, whereas Cx43 in the ThA is predominantly in the smooth muscle, differences in specific kinases or phosphatases may exist between these 2 tissues.
The present study has demonstrated that Cx37 is the major Cx expressed in the media of the CA, whereas this role is played by Cx43 in the ThA. Reduced expression of Cx45 suggests that it plays a minor role in the media of both arteries, whereas Cx40 is not expressed in the muscle of either vessel. In contrast, 3 Cxs are expressed in the endothelium of both vessels, and of these, Cx40 is more prevalent in the CA than in the ThA
We thank Dr Klaus Matthaei, Dr Hilton Grayson, Dr Kim Powell, and Matthew Newhouse for advice and assistance and the National Heart Foundation of Australia for financial support.
Received June 4, 2002; revision accepted June 19, 2002.
- ↵Severs NJ. Cardiovascular disease.In: Novartis Foundation Symposium: Gap Junction-Mediated Intercellular Signaling in Health and Disease. New York, NY: Wiley; 1999.
- ↵Krüger O, Plum A, Kim J-S, Winterhager E, Maxeiner S, Hallas G, Kirchhoff S, Traub O, Lamers WH, Willecke K. Defective vascular development in connexin 45-deficient mice. Development. 2000; 127: 4179–4193.
- ↵Kumai M, Nishii K, Nakamura K, Takeda N, Suzuki M, Shibata Y. Loss of connexin45 causes a cushion defect in early cardiogenesis. Development. 2000; 127: 3501–3512.
- ↵Ko Y-S, Yeh H-I, Rothery S, Dupont E, Coppen SR, Severs NJ. Connexin make-up of endothelial gap junctions in the rat pulmonary artery as revealed by immunoconfocal microscopy and triple-label immunogold electron microscopy. J Histochem Cytochem. 1999; 47: 683–691.
- ↵Beny J-L, Connat JL. An electron-microscopic study of smooth muscle cell dye coupling in the pig coronary arteries: role of gap junctions. Circ Res. 1992; 70: 49–55.
- ↵Sandow SL, Hill CE. The incidence of myoendothelial gap junctions in the proximal and distal mesenteric arteries of the rat is suggestive of a role in EDHF-mediated responses. Circ Res. 2000; 86: 341–346.
- ↵Yeh HI, Rothery S, Dupont E, Coppen S, Severs NJ. Individual gap junction plaques contain multiple connexins in arterial endothelium. Circ Res. 1998; 83: 1248–1263.
- ↵Bastide B, Neyses L, Ganten D, Paul M, Willecke K, Traub O. Gap junction protein connexin40 is preferentially expressed in vascular endothelium and conductive bundles of rat myocardium and is increased under hypertensive conditions. Circ Res. 1993; 73: 1138–1149.
- ↵Bruzzone R, Haefliger JA, Gimlich RL, Paul DL. Connexin40, a component of gap junctions in vascular endothelium, is restricted in its ability to interact with other connexins. Mol Biol Cell. 1993; 4: 7–20.
- ↵Gros D, Jarry-Guichard T, Ten Velde I, de Maziere A, van Kempen MJA, Davoust J, Briand JP, Moorman AFM, Jongsma HJ. Restricted distribution of connexin40, a gap junctional protein, in mammalian heart. Circ Res. 1994; 74: 839–851.
- ↵Yeh HI, Dupont E, Coppen S, Rothery S, Severs NJ. Gap junction localization and connexin expression in cytochemically identified endothelial cells of arterial tissue. J Histochem Cytochem. 1997; 45: 539–550.
- ↵Little TL, Beyer EC, Duling BR. Connexin 43 and connexin 40 gap junctional proteins are present in arteriolar smooth muscle and endothelium in vivo. Am J Physiol. 1995; 268: H729–H739.
- ↵Li X, Simard M. Multiple connexins form gap junction channels in rat basilar artery smooth muscle cells. Circ Res. 1999; 84: 1277–1284.
- ↵Ko Y-S, Coppen SR, Dupont E, Rothery S, Severs NJ. Regional differentiation of desmin, connexin43, and connexin45 expression patterns in rat aortic smooth muscle. Arterioscler Thromb Vasc Biol. 2001; 21: 355–364.
- ↵Li X, Simard M. Connexin45 gap junction channels in rat cerebral vascular smooth muscle cells. Am J Physiol. 2001; 281: H1890–H1898.
- ↵Alcoléa S, Theveniau RM, Jarry GT, Marics I, Tzouanacou E, Chauvin JP, Briand JP, Moorman AF, Lamers WH, Gros DB. Downregulation of connexin 45 gene products during mouse heart development. Circ Res. 1999; 84: 1365–1379.
- ↵Yeh H-I, Chang H-M, Lu W-W, Lee Y-N, Ko Y-S, Severs NJ, Tsai C-H. Age-related alteration of gap junction distribution and connexin expression in rat aortic endothelium. J Histochem Cytochem. 2000; 48: 1377–1389.
- ↵Coppen SR, Dupont E, Rothery S, Severs NJ. Connexin45 expression is preferentially associated with the ventricular conduction system in mouse and rat heart. Circ Res. 1998; 82: 232–243.
- ↵Kadle R, Zhang JT, Nicholson J. Tissue-specific distribution of differentially phosphorylated forms of Cx43. Mol Cell Biol. 1991; 11: 363–369.
- ↵Green CR, Peters NS, Gourdie RG, Rothery S, Severs NJ. Validation of immunohistochemical quantification in confocal scanning laser microscopy: a comparative assessment of gap junction size with confocal and ultrastructural techniques. J Histochem Cytochem. 1993; 41: 1339–1349.
- ↵Yeh HI, Lupu F, Dupont E, Severs NJ. Upregulation of connexin43 gap junctions between smooth muscle cells after balloon catheter injury in the rat carotid artery. Arterioscler Thromb Vasc Biol. 1997; 17: 3174–3184.
- ↵van Veen TAB, Van Rijen H, Opthof T. Cardiac gap junction channels: modulation of expression and channel properties. Cardiovasc Res. 2001; 51: 217–229.