Cells from the midmyocardial ventricular wall (M region) have an action potential duration at 90 % of repolarization (APD₉₀) of about 400 ms if paced at 1 Hz. Action potentials are about 100 ms shorter when isolated from the endocardium or the epicardium. Such differences can also be appreciated when action potentials are recorded from a transmural tissue slice from the (normal) left ventricle from patients with cystic fibrosis as underlying disease. Therefore it seems justified to conclude that M cells exist in the human ventricle as they have been demonstrated in several other mammalian species with the dog as the most prominent example. Whether transmural repolarisation gradients suggested by these in vitro studies are manifest in vivo at physiological heart rates remains a debated issue. Two important issues play a role in this controversy. The difference in action potential duration between M cells and epi-or endocardial cells is a function of heart rate. At very low heart rate the difference can be very large, whereas at normal heart rate the difference is much smaller. Nevertheless, in man a difference in intrinsic APD₉₀ of about 100 ms seems to be present within both the right and left ventricular free wall. A second important issue, however, remains. In a normal heart cells are coupled to each other by gap junctions, providing the low-resistance pathway for intercellular current flow, a prerequisite for conduction of the cardiac impulse. These gap junctions also synchronize the repolarization of the myocytes. Although cells may have substantial intrinsic electrophysiological differences, this does not necessarily imply that these intrinsic differences are overt when the cells are coupled to each other in the ventricular wall.

In the hearts of patients undergoing coronary artery surgery no midmyocardial prolonged activation-recovery intervals (ARIs) could be demonstrated. ARIs are measured in local electrograms by taking the time interval between the moment of the steepest negative deflection in the QRS interval and the moment of the steepest positive deflection in the T wave, and they are considered as the extracellular counterpart of the local transmembrane potential duration. These data allow two possible explanations: (i) in these patients, no intrinsic differences are present, or (ii) intrinsic differences are present, but are reduced by intercellular coupling. We have addressed this issue further by computer simulations of a heterogeneous strand of ventricular cells with variable coupling. In a model strand of 90 ventricular cells, the first 30 cells were endocardial cells. For these cells, the densities of I_to, I_Kr, I_Ks and I_K1 (all outward currents relevant for repolarization) were reduced relative to the current densities in the model epicardial cells by 75 % for I_to, based on data in man, and by respectively 0, 8 and 11 % for the other three currents based on data in the dog. The next 30 cells were M cells with current densities reduced by 13, 54, and 26 % for I_to, I_Ks and I_K1. Finally, the last 30 cells were epicardial cells with all current densities at 100 %. By gradual increase of the intercellular coupling the dispersion in APD₉₀ values had already disappeared at about 1 mS, which is less than the ‘normal’ value for intercellular coupling estimated at 3-12 mS. This suggests that even when M cells are prominent in human ventricle, intercellular coupling may be sufficient to abolish intrinsic differences almost completely.

It is emphasized that in in vivo studies in the ‘model species’ for M cells long action potential durations were absent in the midmyocardium, even in the presence of I_Ks blockers, which prolong the intrinsic action potential duration of M cells even further. Also, in porcine heart, in which transmural refractoriness was assessed, refractoriness was not longer in the midmyocardium than in the subendocardium or subepicardium.

In summary, the absence of a transmural repolarizing gradient in patients with coronary heart disease does not allow the conclusion that intrinsic differences are absent or reduced in these patients. A moderate degree of cellular coupling may effectively reduce these gradients and thus can be regarded a potent anti-arrhythmic factor. Exposure of the intrinsic heterogeneities is only possible by a high degree of electrical uncoupling. Ultimately, this is demonstrated in isolated, single myocytes.