Hypoxia, electrical uncoupling, and conduction slowing: Role of conduction reserve

See article by Zeevi-levin et al.[7](pages 64–73) in this issue. Cardiac activation is based on propagation of the action potential and requires low resistance cell–cell coupling and proper excitability. The electrical coupling between myocytes is mediated by protein channels, called gap junctions, consisting of connexins (Cx, in the ventricle mainly Cx43). Sodium channels (SCN5A) are the major channel proteins involved in excitability of the cardiac cells. A third factor that determines conduction velocity is the tissue architecture, which involves cell shape [1] and interstitial collagen content (fibrosis) [2]. Many cardiac pathologies change these determinants, which give rise to slow and abnormal conduction and increases the propensity for arrhythmias. Myocardial ischemia is strongly associated with arrhythmias [3]. One of the mechanisms responsible for the arrhythmogenic substrate under conditions of myocardial ischemia is electrical uncoupling at gap junctions (for a review see [4]). It has been shown that upregulation of gap junctional conductance during myocardial ischemia reduces the occurrence of reentrant arrhythmias [5], but not of focal tachycardias [6]. The study of Zeevi-Levin (Cardiovascular Research, this issue) supports the concept of uncoupling-induced conduction slowing [7]. These investigators have used an in vitro conduction assay in which rat neonatal cardiomyocytes are cultured on top of a micro-electrode array. This experimental setup allows determination of conduction in great detail before and after intervention, during a long period. A disadvantage of cultures of neonatal rat heart cells, however, is that distribution of connexins along the membrane differs from that found in adult myocytes. In this investigation, one aspect of myocardial ischemia, i.e., hypoxia, was studied and the preparation was subjected to 1% O2 for 5 h. Conduction velocity (CV) remained unchanged until after 5 h, when CV dropped by ∼20%. Western blot and confocal analyses showed that after 5 h, total Cx43 protein was decreased by 50% and gap junction number and size were decreased by 55 and 26%, respectively. Interestingly, total Cx43 protein content was increased by ∼50% after 15 min of hypoxia, which, perhaps surprisingly, did not affect CV. As mentioned before, CV is determined by excitability, cell–cell coupling, and tissue architecture. Especially for excitability and cell–cell coupling, rather robust changes are needed to affect CV, i.e., the heart has a solid ‘conduction reserve’. This conduction reserve was found in mouse models of decreased coupling. An isolated 50% reduction of Cx43 expression in mouse hearts led to conflicting results. Some groups report a reduction in CV of 23–38% [8,9]. In other studies, however, CV was unaffected by a 50% decrease, and Cx43 levels had to be reduced to ∼10% to result in significant conduction slowing and arrhythmias [10–12]. The conduction reserve with respect to cell–cell coupling is due to the fact that in the normal heart, the intracellular resistance exceeds that of the gap junction resistance and therefore the gap junctions play only a minor role in the total resistance. The relationship between conduction velocity and gap junction coupling reveals a flat curve for values of gap junctional coupling in the range of 3–12 μS (the normal value is about 6 μS). Only if gap junctional coupling approaches values close to uncoupling (<3 μS), conduction is affected [13,14]. As such, a 50% decrease in cell–cell coupling in the adult heart might not lead to a significant reduction of CV. However, in cultured neonatal rat cardiomyocytes, like in the study of Zeevi-Levin et al. [7], gap junction expression is expected to be only 10–20% of the normal neonatal rat ventricular level [15], which might explain the slow CV under basal conditions compared to similar in vitro conduction studies [16,17]. Based on theoretical models, the low level of gap junction expression and slow CV would set the relationship between intercellular conductance and CV in the linear phase. In this linear phase, a 50% decrease in Cx43 expression would result in a decreased CV, as the authors indeed observed. A 50% increase in Cx43, however, should lead to an increased CV, which was not detected in their experiments. The latter might be due to the fact that the increased total Cx43 content is not represented by membrane-inserted Cx43 channels. Dephosphorylation of Cx43 channels due to ischemia could have triggered internalization and secondarily reduced intercellular coupling due to changes in open probability of the channels despite an increase in single channel conductance [18,19]. The study of Zeevi-Levin and colleagues show both an increase in phosphorylated and dephosphorylated Cx43 after 15 min, which excludes the latter reasoning. Alternatively, it might be that the level of coupling is at the cut-off coupling value of the curve, such that it saturates with higher levels of coupling and results in a decreased CV at reduced levels of coupling [7]. It is also possible that a second factor is involved in the changes in conduction velocity. Of the various factors involved in conduction, cell architecture is presumably stable within 5 h. Therefore, a change in excitability might be a candidate to explain the discrepancy. Analysis of extracellular unipolar electrograms cannot discriminate between changes in conduction velocity due to changes in intercellular coupling or changes in cellular excitability [20]. Computer modeling studies have shown that conduction reserve for changes in sodium current density (upstroke velocity) is presumably less solid as compared to changes in cell–cell coupling [13]. In contrast, patients haploinsufficient for the SCN5A sodium channel gene exhibited near-normal ECGs that became progressively abnormal with age, indicating a large conduction reserve for sodium current density in whole heart [21]. Model studies have shown that anoxia, modeled as the opening of IK(ATP) due to a low internal ATP concentrations, did not significantly alter the upstroke velocity of the action potential in single cells [22]. In multicellular preparations of these model cells, however, the activation of IK(ATP) caused attainment of threshold, caused by dynamic sodium channel inactivation and decreased electrotonic current flow to adjoining cells due to decreased action potential amplitude, leading to significant conduction slowing [23]. Reduction of intercellular coupling under such circumstances might reduce this effect, however [24]. Direct evidence of changes in ion currents will be needed to clarify the interference of changes in excitability with the investigated relationship between Cx43 content and CV due to hypoxia in the study of Zeevi-Levin [7]. In conclusion, several studies have shown that conduction reserve is usually large in whole hearts, which ensures impulse conduction with high safety in intact hearts, even when one of the conduction parameters is moderately impaired. Very solid isolated changes (e.g., reduction of coupling alone) are needed to impair conduction, or a combination of factors are needed to exceed the limits of conduction reserve (e.g., reduced sodium current combined with increased fibrosis). The interpretation of correlates between changes in protein expression and impulse conduction are often not straightforward, because the limits of conduction reserve are not well known. This study was supported by NWO grant 016.036.012 (TvV).