Reduced Expression of Endothelial Connexin37 and Connexin40 in Hyperlipidemic Mice
Recovery of Connexin37 After 7-Day Simvastatin Treatment
Objective— We sought to clarify the response of endothelial connexins to hyperlipidemia and lipid-lowering therapy.
Methods and Results— Aortic endothelial gap junctions were analyzed by en face immunoconfocal microscopy and electron microscopy in C57BL/6 mice subjected to the following regimens: (1) normal chow (NC) for 3 months (3 mo), (2) NC for 9 mo, (3) NC for 3 mo, followed by a cholesterol-enriched diet (CED) for 6 mo, (4) NC for 3 mo and CED for 6 mo, with simvastatin in the final week, and (5) (in apoprotein E [apoE]-deficient mice) NC and examined at 3 mo and 7 to 9 mo. In wild-type mice, connexin37 (Cx37) and Cx40 were markedly downregulated in the CED-fed animals compared with those fed NC (CED vs 9-mo NC, 77% reduction in Cx37 and 65% reduction in Cx40; both P<0.01). After simvastatin treatment, Cx40 remained depressed, but Cx37 recovered to 94% of the level found in non–cholesterol-fed animals (P<0.01). Electron microscopy demonstrated that gap junctions were smaller in animals fed the CED compared with those given simvastatin and with controls fed NC (P<0.01). Endothelial connexins were rare in the atherosclerotic plaques of apoE-deficient mice.
Conclusions— Mouse aortic endothelial gap junctions and connexins are downregulated during long-term hyperlipidemia. Short-term treatment with simvastatin leads to recovery of Cx37 expression but not Cx40 expression.
Atherosclerosis originates from the abnormal accumulation of lipid in the vascular wall,1,2⇓ which is accelerated in the presence of hyperlipidemia.3,4⇓ Studies in animal models have demonstrated that hyperlipidemia is associated with alterations to the junctions between neighboring endothelial cells, including gap junctions.5,6⇓ Gap junctions are cell-membrane protein channels made of connexin molecules, which belong to a multigene family. In mammals, connexin37 (Cx37), Cx40, and Cx43 are known to be variously expressed in endothelial cells of different sites within the vascular tree.7,8⇓ In vitro studies indicate that the properties of gap-junctional channels are determined by their connexin makeup,9 and gene-knockout studies show that mice deficient in specific type(s) of endothelial connexins are associated with vascular functional defects or morphological changes.10–12⇓⇓ Furthermore, distinct differences in the distribution and expression between connexin types in the vascular wall have been demonstrated in processes related to atherogenesis, such as hypertension,13,14⇓ postinjury regeneration,15,16⇓ and aging,17 raising the possibility that each connexin type might play a distinctive role. Although recent investigations of hyperlipidemia induced by different experimental approaches in animal models of different species have reported that endothelial gap junctions and connexins change in the hyperlipidemic state,5,6⇓ whether the change is specific to the model remained unclear. In addition, in those studies, the extent of endothelium examined was limited, and knowledge regarding the response of each endothelial connexin to lipid-lowering therapy was lacking.
In this study, we examined the expression patterns of endothelial gap junctions and connexins in C57BL/6 mice fed a cholesterol-enriched diet (CED) and in the same strain of mice deprived of the apoprotein E (apoE) gene. In addition, the effect of statins, a class of potent, lipid-lowering agents reported to decrease progression and/or lead to regression of atherosclerosis by its pleiotropic effects in the vascular wall,18 was investigated. Previous studies have shown that female C57BL/6 mice fed the CED developed more aortic fatty lesions than their male counterparts.19
Animals, Diets, and Tissue Processing
Thirty female C57BL/6 mice, purchased from the National Animal Center of Taiwan (Taipei), were fed normal chow (NC) until 3 months of age, when 6 animals (group 1) were examined. The remaining 24 animals were divided equally into 3 groups. Group 2 received NC for another 6 months; group 3 received NC enriched with 1.25% cholesterol, 15% lard, and 0.5% cholic acid (CED; TestDiet Co) for 6 months; and group 4 received the CED for 6 months plus oral feeding of simvastatin (10 mg · kg−1 · d−1; kindly donated by Merck & Co, Whitehouse Station, NJ) during the final week. In parallel, 12 male, homozygous apoE-deficient mice of C57BL/6 background, purchased from the Jackson Laboratory (Bar Harbor, Me) and fed NC were also examined at 3 (n=5, group 5) and 7 to 9 (n=7, group 6) months of age. The serum cholesterol level was determined before perfusion-fixation.
All animals were anesthetized by ether inhalation and were perfusion-fixed by direct intracardiac injection, initially with heparinized phosphate-buffered saline (10 U/mL), followed by phosphate-buffered 2% paraformaldehyde (pH 7.4) for 10 minutes. In all animals, the descending thoracic aortas were dissected and cut into transverse rings for rapid freezing in isopentane at −160°C. The samples were then stored under LN2 before being immunolabeled. For thin-section electron microscopy, selected arterial samples were prepared by standard procedures.20 The work was conducted in accordance with the Republic of China Animal Protection Law (Scientific Application of Animals), 1998.
Immunofluorescence Labeling of Connexins and Nucleus Identification
Three antibodies were used for immunofluorescence detection of Cx37, Cx40, and Cx43. The polyclonal antisera against Cx37 or Cx43 were produced in rabbits [designated Cx37(R382) and Cx43(R530), respectively] against the synthetic peptides corresponding to residues 266 to 281 (for Cx37) or 314 to 322 (for Cx43) of the cytoplasmic C-terminal tail of rat connexins. The Cx40 antiserum was produced in guinea pigs against a synthetic peptide corresponding to residues 256 to 270 of the cytoplasmic C-terminal tail of rat Cx40 [designated Cx40(GP8)]. These polyclonal sera were affinity-purified and have previously been confirmed to be isotype-specific.17,21⇓
Secondary Antibody/Detection Systems
Donkey anti-rabbit and anti–guinea pig immunoglobulins conjugated to either CY3 or CY5 (Chemicon) were used to visualize immunolabeled connexins. For single labeling of individual connexins, CY3-conjugated antibodies were used. For double labeling of 2 connexins, 1 CY3-conjugated antibody and 1 CY5-conjugated antibody were used in combination.
Immunolabeling of Connexins
The perfusion-fixed aortic rings were rinsed in phosphate-buffered saline for 5 minutes, blocked in 0.5% bovine serum albumin for 15 minutes, and incubated with anti-Cx37 (1:200), anti-Cx40 (1:100), or anti-Cx43 (1:500) at 37°C for 2 hours. The samples were then treated with CY3-conjugated secondary antibody (1:500, at room temperature for 1 hour). In double-labeling experiments, incubation was conducted with a mixture of anti-Cx37 (1:200) and anti-Cx40 (1:100), followed by incubation with a mixture of the 2 corresponding species-specific secondary antibodies (CY3 and CY5, 1:500). At this stage, in selected experiments, samples were incubated with bisbenzamide (1 μg/mL, Sigma) for 15 minutes to visualize the cell nuclei.
Confocal Laser Scanning Microscopy and Image Analysis
Please see http://atvb.ahajournals.org.
Serum cholesterol level differed between the groups. Compared with the wild-type animals fed NC (group 1, 54±4 mg/dL; group 2, 60±7, mg/dL), the cholesterol level was markedly elevated in apoE-deficient mice of both age groups (group 5, 698±88 mg/dL; group 6, 507±86 mg/dL), as well as in the wild-type animals fed the CED (group 3, 665±157 mg/dL; group 1 or 2 vs group 3, 5, or 6; all P<0.01). After administration of simvastatin for 1 week, the cholesterol level dropped rapidly (group 4, 188±24 mg/dL; group 4 vs group 3, P<0.01), although it remained higher than that of the wild-type animals of the same age fed NC (group 4 vs group 2, P<0.01).
Histologic examination of the aortic wall revealed the presence of atherosclerotic plaques in the older age group of apoE-deficient mice (Figure 1A). In wild-type animals maintained on the CED, fatty streaks and intracytoplasmic vacuoles in both endothelial cells and smooth muscle cells were apparent (Figure 1B and 1C), and these features remained after the administration of simvastatin.
Gap-Junction Distribution and Connexin Expression
En face confocal views of the luminal surface of the aortas from wild-type animals fed NC after single labeling clearly displayed a punctate signal for Cx37 and Cx40, typical of gap junctions and delineating the borders of endothelial cells (Figure 2). In general, the 2 connexins shared a similar expression pattern, including (1) the number of gap-junctional spots per unit area (Figure 3, upper histogram); (2) spot size (Figure 3, middle); and (3) the percentage area of connexin signal (Figure 3, lower). In addition, there was a tendency for the total gap-junction area for each connexin to be larger in the 9-month-old, wild-type animals fed NC compared with the younger animals (Figure 2A, 2B, 2G, and 2H and Figure 3).
Connexin expression was dramatically reduced in the wild-type animals fed the CED (Figure 2C and 2I). Image analysis showed that the total gap-junction number as well as gap-junction spot size decreased, such that the total gap-junction area was reduced by 77% for Cx37 and by 65% for Cx40 compared with wild-type animals fed NC (both P<0.01, Figure 3). After treatment with simvastatin, Cx37 expression was recovered (Figures 2D and 3⇑; group 4 vs group 3, P<0.01), whereas Cx40 remained depressed (Figures 2J and 3⇑).
To exclude the possibility that the reduction in connexins was attributable to the detachment of endothelial cells during sample processing, the nuclei of the arterial wall were rendered visible by bisbenzamide staining, and the images containing the nuclei of the whole layer were superimposed onto the corresponding images of the endothelial connexins (Figure 4). The results demonstrated that, although both endothelial cells and smooth muscle cells had ovoid nuclei, nuclei of the endothelium were wider than those of smooth muscle. In addition, the orientation of the nuclei between these 2 cell types was roughly at right angles (Figure 4A and 4B). With these characteristics plus the outlining of the endothelial cell borders by the connexin labels, nuclei of the endothelial cells were unequivocally distinguished from those of underlying smooth muscle cells. Application of this approach, as illustrated in Figure 4C through 4E, confirmed the presence of endothelial nuclei, thereby reinforcing the conclusion that the reduction in connexins of the wild-type animals fed the CED could indeed be ascribed to the endothelium and that the effect of simvastatin was connexin type–specific. In addition, bisbenzamide staining allowed identification of plaques, in which the cell nuclei were found to have lost their orderly orientation and become unevenly distributed (Figure 4G and 4H).
The expression and distribution of Cx37 and Cx40 in the apoE-deficient mice also differed from those of the wild-type animals. Although in the 3-month-old apoE-deficient animals both connexins appeared more or less evenly distributed, the total gap-junction area for each of the 2 connexins tended to be lower than that in the corresponding age groups of the wild-type animals fed NC (Figure 2E and 2K, Figure 3, and Figure 4F). Similarly, a tendency for a lower level of expression was seen in the older age group of apoE-deficient mice compared with the corresponding age group of wild-type animals fed NC. However, the expression and distribution of endothelial connexins varied markedly from region to region in the older age group of apoE-deficient animals. In general, Cx37 and Cx40 were scanty and heterogeneously distributed in the endothelium covering the plaques (Figure 4G and 4H). In areas outside the plaques, the distribution of connexins was more or less homogeneous (Figure 4I).
Double labeling enabled us to examine the spatial relation between Cx37 and Cx40. Colocalization of the 2 connexins within the same spot was detected as yellow fluorescence, due to direct superimposition of red and green (please see online Figure I at http://atvb.ahajournals.org). For the 9-month-old, wild-type animals, 93±1% of the spots positive for Cx37 were also positive for Cx40 in those fed NC (online Figure IA). The extent of colocalization was reduced to 45±3% in those fed the CED and to 58±3% in those fed the CED plus simvastatin (group 2 vs either group 3 or group 4, both P<0.01; online Figure IB and IC). Cx43 was undetectable in the endothelium of all animals.
Electron Microscopic Examination
In the 9-month-old, wild-type mice fed NC, gap junctions were frequently seen in the endothelial cells, which in general lie directly on the internal elastic lamina, leaving minimal subendothelial space (online Figure IIA; please see http://atvb.ahajournals.org). By contrast, in the wild-type mice fed the CED for 6 months, gap junctions were only infrequently observed in the endothelium, portions of which were observed to overlie a thickened subendothelial layer (online Figure IIB). In the wild-type mice treated with simvastatin, although the intima shared the same morphology as seen in mice fed the CED alone, gap junctions were not infrequently detected (online Figure IIC). Analysis of 50 randomly selected junctions from each group showed that gap junctions were smaller in the cholesterol-fed mice (147±12 nm) compared with those fed NC (437±40 nm) and cholesterol-fed mice that had received simvastatin (368±36 nm; both P<0.01). On the other hand, in the older age group of apoE-deficient mice, endothelial cells covering the plaques contained few gap junctions (online Figure IID). However, outside the plaques, gap junctions were common (online Figure IIE). Gap junctions were frequently seen in the 3-month-old animals of both genotypes.
The present study demonstrates that in aortic endothelium of C57BL/6 mice, gap junctions and their component connexins, Cx37 and Cx40, are downregulated in the hyperlipidemic state induced by long-term feeding of a CED. A dramatic upregulation of the junctions was apparent after short-term treatment with simvastatin, a 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor, but this was due to restoration of Cx37 expression rather than Cx40 expression. Similarly, in the hyperlipidemic state due to apoE gene deletion, there was a general tendency for the endothelial junctions to become less abundant compared with the wild-type animals fed NC. In particular, endothelial cells overlying atherosclerotic plaques expressed only few gap junctions. These findings substantially expand current knowledge of endothelial gap-junction biology in the hyperlipidemic state.
Structurally, gap-junction proteins assemble in the lipid milieu of the plasma membrane. Previous in vitro studies have shown that the lipid content of the plasma membrane, as well as the extracellular space, have diverse effects on gap junctions.22 In animal studies investigating the effect of hyperlipidemia on endothelial gap junctions, a recent ultrastructural study reported that fewer aortic endothelial gap junctions were present in rats fed a long-term, cholesterol-enriched diet compared with those fed normal chow.6 However, because immunohistochemical examination was not conducted in that study, the details of the component connexins remained unknown. Moreover, the cholesterol-fed, Sprague-Dawley rats used by those authors were reported only to develop fatty streaks6; no information was provided regarding the gap junctions in the endothelium overlying overt atherosclerotic plaques. A recent immunohistochemical study examining sections of aortas from LDL receptor–deficient mice fed a high-cholesterol diet shed light on this issue, revealing that in that model, connexins are downregulated in the endothelium covering the plaques.5 Although the finding is striking, the question remained as to whether the response of connexins in LDL receptor–deficient mice was unique or whether a similar response might characterize hyperlipidemia in general. In addition, whether hyperlipidemia affects murine endothelial gap junctions in areas free of atherosclerotic plaques remained unclear. These gaps in our knowledge are filled by the present study, which shows that regulation of endothelial gap junctions is similar but not identical between hyperlipidemic mice induced by different mechanisms. For example, a marked downregulation of endothelial gap junctions in the wild-type mice fed a CED was seen in areas free of fatty lesions, whereas in apoE-deficient mice, the downregulation was apparent only in the endothelial cells covering the atheromas. The discordant responses to hyperlipidemia between the wild-type and gene-knockout mice might be attributed to deletion of the apoE gene and/or lipid contents of the diets. However, the precise mechanisms underlying the differences require further studies. In addition, in aortic endothelium of LDL receptor–deficient mice, Cx43 expression was reported specifically at the plaque periphery5; however, the present study shows that no detectable signal for Cx43 was apparent at any site in the aortic endothelium. The findings are based on en face immunoconfocal views of the endothelium, which provide information from much larger expanses of endothelial areas compared with previous reports based on sectional views. In addition, the complementary electron microscopy performed in the present study verified the validity of and extended the immunoconfocal findings.
The present results potentially have important physiologic and pathophysiologic implications. Inhibition of gap-junctional communication in the vascular wall is reported to result in vasodilatation dysfunction through endothelium-dependent mechanisms,23,24⇓ which indicates a requirement for gap junctions in endothelium-dependent vasodilatation. In keeping with this finding, investigation of Cx40-knockout mice has demonstrated that the absence of endothelial Cx40 expression leads to vasodilatation dysfunction in the microcirculation.10 Previous studies have also shown that hyperlipidemia impairs endothelium-dependent vasodilatation.25,26⇓ Therefore, the downregulation of endothelial gap junctions in the hyperlipidemic state seen in the present study might contribute to the occurrence of vasodilatation dysfunction demonstrated in hyperlipidemic animals.25,26⇓ Similarly, the upregulation of endothelial gap junctions by simvastatin demonstrated here might augment the beneficial effect of this lipid-lowering agent on endothelium-dependent vasodilatation function,27 which, as discussed earlier, is impaired in the hyperlipidemic state. On the other hand, it should be borne in mind that restoring Cx40 levels was not necessary for improving the number of gap junctions in the high cholesterol–fed mice treated with simvastatin, implying that adequate function might be maintained irrespective of connexin type. Knockin studies suggest that connexins might have shared and unique functions.28 Furthermore, in the aortic endothelium of Cx40-knockout mice, Cx37 is upregulated, and aspects of gap-junction properties differ from those of wild-type mice.29 Thus, whether recovery of Cx37 expression after simvastatin treatment satisfies all of the functional demands of gap-junctional communication here requires further investigation.
One possible factor in the differential response of depressed Cx37 and Cx40 in the hyperlipidemic mice to simvastatin is that the threshold of cholesterol level required to change Cx37 expression might be higher than that for Cx40, such that even when the cholesterol level is still 3-fold higher in the simvastatin treatment group compared with the corresponding control-fed, NC group, Cx37 expression was almost completely recovered, but Cx40 remained depressed. In any event, such a differential response as observed in this study is consistent with previous reports that endothelial connexins are differentially regulated by a variety of factors, such as aging and tumor necrosis factor-α,17,30⇓ a cytokine participating in atherogenesis. Recent studies have demonstrated that statins possess anti-inflammatory properties, including inhibition of the production of tumor necrosis factor-α in atherosclerotic diseases.31,32⇓ Whether the effect of simvastatin shown in the present study involves its action on tumor necrosis factor-α remains to be defined.
In conclusion, hyperlipidemia induced by a CED was associated with a marked reduction of endothelial gap junctions and connexins, which was largely rectified by a short period of simvastatin treatment. Hyperlipidemia induced by apoE gene deletion was also associated with an even more severe downregulation of gap junctions and connexins in endothelial cells overlying the plaque areas. Clarification of the mechanisms by which statins modulate endothelial gap junction expression in the hyperlipidemic state and whether this phenomenon exists in humans require further experiments.
This work was supported by grants NSC-90-2314-B-195-017 from the National Science Council, Taiwan, and MMH-9123 from the Medical Research Department of the Mackay Memorial Hospital, Taiwan.
- Received December 2, 2002.
- Accepted June 2, 2003.
- ↵Schwartz SM, deBlois D, O’Brien ER. The intima: soil for atherosclerosis and restenosis. Circ Res. 1995; 77: 445–465.
- ↵Wu CC, Chang SW, Chen MS, Lee YT. Early change of vascular permeability in hypercholesterolemic rabbits. Arterioscler Thromb Vasc Biol. 1995; 15: 529–533.
- ↵Kwak BR, Mulhaupt F, Veillard N, Gros DB, Mach F. Altered pattern of vascular connexin expression in atherosclerotic plaques. Arterioscler Thromb Vasc Biol. 2002; 22: 225–230.
- ↵Van Rijen H, van Kempen MJ, Analbers LJ, Rook MB, van Ginneken AC, Gros D, Jongsma HJ. Gap junctions in human umbilical cord endothelial cells contain multiple connexins. Am J Physiol. 1997; 272: C117–C130.
- ↵Yeh HI, Rothery S, Dupont E, Coppen SR, Severs NJ. Individual gap junction plaques contain multiple connexins in arterial endothelium. Circ Res. 1998; 83: 1248–1263.
- ↵Elfgang C, Eckert R, Lichtenberg-Frate H, Butterweck A, Traub O, Klein RA, Hulser DF, Willecke K. Specific permeability and selective formation of gap junction channels in connexin-transfected HeLa cells. J Cell Biol. 1995; 129: 805–817.
- ↵de Wit C, Roos F, Bolz SS, Kirchhoff S, Kruger O, Willecke K, Pohl U. Impaired conduction of vasodilation along arterioles in connexin40-deficient mice. Circ Res. 2000; 86: 649–655.
- ↵Liao Y, Day KH, Damon DN, Duling BR. Endothelial cell-specific knockout of connexin 43 causes hypotension and bradycardia in mice. Proc Natl Acad Sci U S A. 2001; 98: 9989–9994.
- ↵Goodenough DA, Simon AM, Paul, DL. Gap junctional intercellular communication in the mouse ovarian follicle. Novartis Found Symp. 1999; 219: 226–235;discussion 235–240.
- ↵Haefliger JA, Castillo E, Waeber G, Bergonzelli GE, Aubert JF, Sutter E, Nicod P, Waeber B, Meda P. Hypertension increases connexin43 in a tissue-specific manner. Circulation. 1997; 95: 1007–1014.
- ↵Haefliger JA, Meda P, Formenton A, Wiesel P, Zanchi A, Brunner HR, Nicod P, Hayoz D. Aortic connexin43 is decreased during hypertension induced by inhibition of nitric oxide synthase. Arterioscler Thromb Vasc Biol. 1999; 19: 1615–1622.
- ↵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.
- ↵Yeh HI, Lai YJ, Chang HM, Ko YS, Severs NJ, Tsai CH. Multiple connexin expression in regenerating arterial endothelial gap junctions. Arterioscler Thromb Vasc Biol. 2000; 20: 1753–1762.
- ↵Yeh HI, Chang HM, Lu WW, Lee YN, Ko YS, Severs NJ, Tsai CH. Age-related alteration of gap junction distribution and connexin expression in rat aortic endothelium. J Histochem Cytochem. 2000; 48: 1377–1389.
- ↵Takemoto M, Liao JK. Pleiotropic effects of 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. Arterioscler Thromb Vasc Biol. 2001; 21: 1712–1719.
- ↵Qiao JH, Xie PZ, Fishbein MC, Kreuzer J, Drake TA, Demer LL, Lusis AJ. Pathology of atheromatous lesions in inbred and genetically engineered mice: genetic determination of arterial calcification. Arterioscler Thromb. 1994; 14: 1480–1497.
- ↵Blackburn JP, Peters NS, Yeh HI, Rothery S, Green CR, Severs NJ. Upregulation of connexin43 gap junctions during early stages of human coronary atherosclerosis. Arterioscler Thromb Vasc Biol. 1995; 15: 1219–1228.
- ↵Yeh HI, Lai YJ, Lee YN, Chen YJ, Chen YC, Chen CC, Chen SA, Lin CI, Tsai CH. Differential expression of connexin43 gap junctions in cardiomyocytes isolated from the canine thoracic veins. J Histochem Cytochem. 2003; 51: 259–266.
- ↵Osborne JA, Siegman MJ, Sedar AW, Mooers SU, Lefer AM. Lack of endothelium-dependent relaxation in coronary resistance arteries of cholesterol-fed rabbits. Am J Physiol. 1989; 256: C591–C597.
- ↵Verbeuren TJ, Jordaens FH, Zonnekeyn LL, Van Hove CE, Coene MC, Herman AG. Effect of hypercholesterolemia on vascular reactivity in the rabbit, I: endothelium-dependent and endothelium-independent contractions and relaxations in isolated arteries of control and hypercholesterolemic rabbits. Circ Res. 1986; 58: 552–564.