Author + information
- Received July 12, 2002
- Revision received September 30, 2002
- Accepted November 19, 2002
- Published online February 19, 2003.
- Shigeto Kanno, MD, PhD*,
- Attila Kovacs, MD*,
- Kathryn A Yamada, PhD* and
- Jeffrey E Saffitz, MD, PhD, FACC*,*
- ↵*Reprint requests and correspondence:
Dr. Jeffrey E. Saffitz, Washington University, Department of Pathology, Box 8118, 660 South Euclid Avenue, St. Louis, Missouri 63110USA.
Objectives The purpose of this study was to define the role of cell–cell coupling as an independent determinant of infarct size following coronary occlusion.
Background Electrical uncoupling induced by acute ischemia enhances arrhythmogenesis, but it may also protect the heart by limiting intercellular spread of chemical mediators of injury.
Methods The left anterior descending coronary artery was ligated in wild-type (Cx43+/+) mice and Cx43-deficient (Cx43+/−) mice that are heterozygous for a null allele in the gene encoding the major gap junction channel protein, connexin43 (Cx43). Ventricular remodeling and infarct size were compared in both groups.
Results Echocardiography at 1 and 10 weeks after infarction showed that left ventricular end-diastolic volume and mass increased and ejection fraction decreased in proportion to infarct size in both Cx43+/−and Cx43+/+hearts. However, infarct size measured histologically in healing infarcts (eight days after infarction) was 29% smaller in Cx43+/−hearts (17 ± 14% of total left ventricular area, n = 30) than in Cx43+/+hearts (24 ± 15%, n = 23; p = 0.037). Fully healed infarcts were smaller than healing infarcts, owing to resorption of necrotic tissue and maturation of scar, but infarct size at 10 weeks after coronary occlusion was still smaller (by 50%) in Cx43+/−hearts (6 ± 5%, n = 9) compared with Cx43+/+hearts (12 ± 7%, n = 17; p = 0.037).
Conclusions Cx43-deficient mice develop smaller infarcts than wild-type mice following coronary ligation. New therapies designed to decrease the risk of arrhythmias by enhancing intercellular communication could lead to larger infarcts caused by persistent coronary occlusion.
Genetic engineering in mice provides an opportunity to define the role of individual gene products in complex disease processes. Previously, we analyzed mice that were heterozygous for a null mutation in the gene encoding the major ventricular gap junction channel protein, connexin43 (Cx43), to elucidate the role of Cx43 in arrhythmogenesis induced by acute ischemia. These mice (Cx43+/−mice) express ∼50% of the wild-type level of Cx43 and, under physiologic conditions, show no apparent abnormalities of cardiac structure or function other than modest slowing of ventricular conduction velocity (1–4). In response to acute coronary occlusion, however, Cx43+/−mice exhibit a significantly greater incidence and earlier onset of spontaneous ventricular arrhythmias (5), indicating that diminished coupling is a powerful, independent determinant of arrhythmogenesis in the setting of acute regional ischemia. These results suggest that if it were possible to enhance intercellular coupling in the heart, the risk of spontaneous ventricular arrhythmias would be reduced when the myocardium became acutely ischemic. However, enhanced coupling could have other consequences in ischemic heart disease. Closure of gap junction channels does not occur until late in the development of acute injury in ischemic cardiac myocytes, and previous studies have indicated that intercellular chemical transmission could promote the spread of injury (6–9). In the present study, we tested the hypothesis that diminished coupling results in smaller myocardial infarcts (MIs) in Cx43+/−mice compared with wild-type (Cx43+/+) mice after coronary occlusion. We ligated the left anterior descending coronary artery (LAD) to create MI in Cx43+/−and Cx43+/+mice, and we characterized post-MI remodeling and measured infarct size in healing and fully healed infarcts.
Founder mice (C57BL/6J-Gjaltm1Kdr) of uniform genetic background (C57BL/6) were originally purchased from Jackson Laboratories (Bar Harbor, Maine) and inbred in a standard barrier facility under veterinary supervision. Genotypes were determined by polymerase chain reaction, as described previously (5). All protocols conformed to the “Position of the American Heart Association on Research Animal Use,” adopted by the Association in November 1984, and were approved by the Animal Studies Committee at Washington University School of Medicine.
Surgery for MI
All operations were performed by the same individual (S.K.) who was blinded to the genotype. Typically, groups of 6 to 10 animals of both genotypes were operated in random order at a single setting. Animals were anesthetized with an intraperitoneal injection of ketamine (87 mg/kg) and xylazine (13 mg/kg). The trachea was exposed through a mid-line incision and intubated with a 20-gauge intravenous catheter through the oral pharynx. Respiration was controlled by a ventilator at a tidal volume of 0.7 to 1.0 ml and a rate of 130 to 150 strokes/min. The chest was opened by a left thoracotomy through the fifth intercostal space. A 7-0 prolene suture was tied around the LAD immediately distal to the origin of the first diagonal branch. Animals were weaned from the respirator and allowed to recover from anesthesia on a heating mat. Sham operations were performed in the identical manner, except that the LAD was not ligated.
All echocardiographic studies and interpretations were performed by a single observer (A.K.) who was blinded to the genotype. Echocardiographic measurement of infarct size and characterization of post-MI remodeling by measuring left ventricular (LV) end-diastolic volume, LV mass, and ejection fraction were performed in subsets of Cx43+/−and Cx43+/+mice with MI, using methods described and validated previously (10). Two groups of animals were studied. In the first group, 13 Cx43+/−and 10 Cx43+/+mice underwent echocardiographic analysis one and seven days after MI. This approach permitted analysis of the relative change in LV structure and function within individual hearts during the first post-MI week. These animals were sacrificed on day 8 after MI, and infarct size was measured histologically, as described subsequently. Echocardiographic findings in the 10 control mice (Cx43+/+) already described have been reported in a previous study designed to validate the echocardiographic methods (10). A second group of eight Cx43+/−and 10 Cx43+/+mice underwent echocardiographic assessment of infarct size and post-MI remodeling 10 weeks after MI to allow comparison of remodeling at 1 and 10 weeks after MI. These animals were sacrificed, and infarct size was measurement histologically 10 weeks after MI.
Histologic analysis of infarct size
Infarct size was measured in 30 Cx43+/−and 23 Cx43+/+hearts containing healing infarcts (8 days after MI surgery) and in nine Cx43+/−and 17 Cx43+/+hearts containing fully healed infarcts (10 weeks after MI surgery). These groups included hearts of mice analyzed by echocardiography and others not analyzed by echocardiography. All investigators were blinded to the genotype. Excised hearts were fixed in formalin, cut transversely into three roughly equal short-axis ventricular slices (apical, middle, and basal portions), embedded in paraffin, and sectioned in their entirety at a thickness of 5 μm. Every 20th section was collected on an individual glass slide yielding 24 to 30 slides per heart. Sections from the approximate mid-points of the three short-axis ventricular slices were stained with Masson’s trichrome stain for quantitative analysis of infarct size and structure. Infarct size was measured using a computerized morphometric system and National Institutes of Health Image software. The portions of each of the three short-axis ventricular sections occupied by infarct scar and the non-infarcted LV free wall and interventricular septum were traced; areas were measured and summed; and infarct size was calculated as the total area occupied by scar divided by the total ventricular area (infarcted plus non-infarcted areas). In previous validation studies (10), morphometric measurements were made in nine different sections (three sections each from the basal, middle, and apical ventricular slices). No significant differences were observed in infarct size based on analysis of three as opposed to nine separate sections. Thus, infarct sizes in the present study were determined in three ventricular sections.
All data are expressed as the mean value ± SD. Differences in infarct size between genotypes were determined by the Mann-Whitney rank-sum test, which was used because the infarct size data were not normally distributed.
Pathologic features of MI
Discrete infarcts were identified in all mice that underwent successful LAD ligation. As shown in Figure 1, some were small, localized infarcts, but most involved much of the anteroapical LV. The largest infarcts also included the lateral and posterolateral free walls, but the interventricular septum was generally spared. The subendocardium was involved in all infarcts, including the smallest. Sham-operated animals showed focal epicardial fibrosis localized immediately to the site of LAD manipulation, but none showed subendocardial fibrosis. Thus, MI was related to occlusion of the LAD rather than to other effects of surgery. When examined eight days after MI surgery, infarcts were composed of granulation tissue containing newly deposited collagenous matrix, residual mononuclear inflammatory infiltrate, and numerous small blood vessels (Fig. 1). Fully healed infarcts examined 10 weeks after MI surgery were composed of dense scar tissue virtually devoid of cells and blood vessels (Fig. 1).
Changes in LV structure and function were monitored one and seven days and 10 weeks after MI in selected Cx43+/−and Cx43+/+mice, all of which were subsequently analyzed by histologic measurement of infarct size. As reported previously (10), there was an excellent correlation between infarct size estimated by echocardiography and that measured by histology (correlation coefficients 0.92 and 0.77 for 1- and 10-week time points, respectively). There were also no significant differences in infarct sizes between groups of Cx43+/−and Cx43+/+mice that were subjected to echocardiography and those analyzed only histologically. Structural and functional evidence of LV remodeling was observed during the first week of infarct healing in hearts with moderate to large infarcts. As shown in Figure 2, both Cx43+/−and Cx43+/+mice showed a direct relationship between the relative increase from one day to one week after MI in LV end-diastolic volume and LV mass and infarct size. Decreasing ejection fraction was also observed with increasing infarct size in both groups during this interval. Echocardiographic analysis was also performed in Cx43+/−and Cx43+/+mice with fully healed infarcts 10 weeks after MI. Sham-operated mice of both genotypes (n = 5) had a LV end-diastolic volume of 34 ± 9 μl, LV mass of 57 ± 11 mg, and ejection fraction of 70 ± 2%. These values were the same as those obtained from hearts with very small infarcts at either 1 or 10 weeks after MI. Figure 3compares LV end-diastolic volume and ejection fraction as a function of infarct size measured 1 and 10 weeks after MI in Cx43+/−and Cx43+/+mice. These echocardiographic findings indicate that post-MI remodeling varies as a function of infarct size in both Cx43+/−and Cx43+/+mice, and this relationship is not affected by diminished Cx43 expression.
Relationship between Cx43 expression level and infarct size
As shown in Figure 4, infarcts were smaller in Cx43+/−than in Cx43+/+hearts. In hearts excised eight days after MI surgery, the mean infarct size, expressed as a percentage of total LV area, was 29% smaller in Cx43+/−hearts (17 ± 14% of total LV area, n = 30) than in Cx43+/+hearts (24 ± 15%, n = 23; p = 0.037). Absolute infarct size (total area of scar in the apical, middle, and basal sections) was 28% smaller in Cx43+/−hearts (5.4 ± 4.4 mm2) than in Cx43+/+hearts (7.5 ± 4.4 mm2; p = 0.049), whereas the total area of infarcted plus non-infarcted segments summed across the three separate sections was the same in Cx43+/−(32.5 ± 4.8 mm2) and Cx43+/+(31.9 ± 4.3 mm2; p = 0.627) hearts. These observations are consistent with echocardiographic findings and indicate that smaller infarcts in Cx43+/−hearts at this phase of remodeling during infarct healing were the result of a smaller amount of tissue that became necrotic, rather than a larger amount of post-MI hypertrophy of non-infarcted segments. In hearts excised 10 weeks after MI surgery, infarcts were smaller (by 50%) in Cx43+/−hearts (6 ± 5%, n = 9) than in Cx43+/+hearts (12 ± 7%, n = 17; p = 0.037). Overall, the size of fully healed infarcts was considerably less than the size of healing infarcts, owing to complete resorption of necrotic muscle, a decline of inflammatory cells and blood vessels in the granulation tissue seen in healing infarcts, and maturation of the granulation tissue matrix into dense fibrous scar tissue. The reproducibility of intraobserver and interobserver measurements of infarct size was good (reliability coefficient = 0.98).
The major finding in this study is that Cx43+/−mice developed significantly smaller infarcts than wild-types when the LAD was ligated at the same point in both groups. This result indicates that a reduction of ∼50% in Cx43 expression is an independent determinant of the size of an infarct created by persistent coronary occlusion. The underlying mechanisms may be multiple. One possibility is that Cx43+/−mice developed a smaller region at risk for infarction because of differences in the structure and/or function of the coronary circulation. In a previous study (5), no differences were seen in the size of the hypoperfused zone measured by a dye-perfusion method in isolated, perfused Cx43+/−and Cx43+/+hearts 60 min after ligation of the LAD (53 ± 9% and 58 ± 15% of total LV area was unstained in Cx43+/−and Cx43+/+hearts, respectively; n = 6 for each; p > 0.5). Thus, major differences in collateral vessels at the onset of LAD occlusion or blood flow patterns after occlusion probably do not account for the smaller infarcts observed in Cx43+/−mice. A more likely reason may be decreased intercellular transmission of biochemical mediators of ischemic injury. This hypothesis is supported by the results of previous studies.
Garcia-Dorado et al. (6)have reported that hypercontracture induced by microinjection of Ca2+into one cell in a pair of isolated myocytes was transmitted to the other cell, and this process was blocked by the gap junction uncoupler heptanol. Adding heptanol during reoxygenation of hypoxic rat hearts improved recovery of contractile function and decreased necrosis (6). The addition of heptanol during reperfusion also reduced the infarct size in pig hearts subjected to coronary occlusion in vivo (6). Lin et al. (7)showed that forced expression of bcl-2 in C6 glioma cells increased cellular resistance to injury, but this effect was diminished when bcl-2–expressing cells were coupled by gap junctions to nontransfected cells. Yasui et al. (8)reported that inhibition of Cx43 synthesis by antisense oligonucleotides in cultured neonatal rat myocytes decreased dye coupling and increased the progression of apoptosis during five days in culture. Ruiz-Meana et al. (9)showed that intercellular passage of the dye Lucifer yellow, a functional marker of gap junction coupling, still occurred 10 min after the onset of rigor in isolated myocyte pairs subjected to simulated ischemia, as well as 30 min after rigor developed in isolated rat hearts subjected to no-flow ischemia. Taken together, these studies suggest that gap junction communication persists for some time after the onset of myocardial ischemia and that intercellular transfer of molecules may either limit or promote injury depending on experimental conditions.
It is difficult to achieve a defined steady-state level of uncoupling with the use of chemical uncouplers or antisense oligonucleotides and to avoid potential nonspecific effects of chemical uncouplers such as heptanol. Cx43+/−mice have the advantage of a defined reduction in Cx43 expression, but because cardiac myocytes are so well coupled, it has been suggested that a reduction of ∼50% in Cx43 expression may have little functional impact under basal physiologic conditions (11). However, under pathophysiologic conditions, a more dramatic phenotype can be elicited in Cx43+/−mice. For example, LAD occlusion in isolated, perfused hearts producing acute regional ischemia leads to a significantly greater incidence and earlier onset of ventricular tachyarrhythmias in Cx43+/−compared with Cx43+/+mice. Of particular relevance to the present study is a recent report by Suishansian et al. (12), who observed significantly largercerebral infarcts in Cx43+/−compared with Cx43+/+mice subjected to middle cerebral artery occlusion. Unlike the situation in the heart in which cardiac myocytes are extensively coupled by gap junctions, neurons within the brain are not connected by gap junctions. However, glial cells are interconnected by gap junctions, and it is thought that one mechanism by which glia protect neurons at sites of acute injury is uptake of extracellular glutamate and other biochemical mediators of injury and dissipation via intercellular transfer through gap junctions. If this mechanism is operative, then diminished expression of Cx43 would increase neuronal injury and lead to larger cerebral infarcts. Although the role of gap junction coupling as a determinant of infarct size may differ in the heart and brain, the results of the present study and the study by Suishansian et al. (12)indicate that an ∼50% reduction in Cx43 expression has a clear impact on the extent of tissue injury in both coronary and cerebral artery occlusion models.
Study limitations and clinical implications
Considerable circumstantial evidence suggests that remodeling of gap junctions in patients with healed infarcts causes slow conduction and uni-directional conduction block, which contribute to reentrant arrhythmogenesis (13–16). Although reduced Cx43 expression enhances arrhythmogenesis in response to acute myocardial ischemia (5), the results of the present study add considerable weight to the idea that reduced gap junction communication also diminishes infarct size. Uncoupling during acute ischemia may therefore serve multiple adaptive roles. Echocardiographic quantification of post-MI remodeling in mice is technically challenging, and we did not independently measure changes in ventricular volumes and ejection fractions. It must also be acknowledged that surgery to produce MI in mice is difficult. Although infarcts were consistently smaller in Cx43+/−hearts, infarcts occurred over a considerable range of sizes in both genotypes, perhaps related to technical factors. The mechanisms responsible for smaller infarcts in Cx43+/−mice remain a matter of speculation. Although it is likely that reduced coupling was responsible, secondary effects of Cx43 deficiency could have contributed indirectly. It may be argued, however, that attempts to reduce lethal arrhythmias in patients by improving gap junction communication may carry certain risks with regard to the extent of subsequent ischemic injury and post-MI remodeling.
We thank Karen Green for technical assistance and Kenneth Schechtman, PhD, for assistance with statistical analysis.
☆ This work was supported by Grant HL58507, of the Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland.
- mice with two wild-type alleles for the Cx43 gene
- mice with one null allele for the Cx43 gene
- left anterior descending coronary artery
- left ventricle or ventricular
- myocardial infarct or infarction
- Received July 12, 2002.
- Revision received September 30, 2002.
- Accepted November 19, 2002.
- American College of Cardiology Foundation
- Thomas S.A.,
- Schuessler R.B.,
- Berul C.I.,
- et al.
- Eloff B.C.,
- Lerner D.L.,
- Yamada K.A.,
- et al.
- Lerner D.L.,
- Yamada K.A.,
- Schuessler R.B.,
- et al.
- Garcia-Dorado D.,
- Inserte J.,
- Ruiz-Meana M.,
- et al.
- Yasui K.,
- Kada K.,
- Hojo M.,
- et al.
- Ruiz-Meana M.,
- Garcia-Dorado D.,
- Lane S.,
- et al.
- Jongsma H.J.,
- Wilders R.
- Peters N.S.,
- Green C.R.,
- Poole-Wilson P.A.,
- et al.
- Peters N.S.,
- Coromilas J.,
- Severs N.J.,
- et al.
- de Bakker J.M.T.,
- Janse M.J.